Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

The Valuation Mirage

Chapter 2 Financial Modeling Practical Development of Cash Flow Forecasts for Valuation

Chapter 2 Financial Modelling

Page 1

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

Chapter 2: Financial Modeling: Practical Development of Cash Flow Forecasts

Introduction
One of the inevitable steps in any valuation analysis is making a projection of the future. Projections may be for earnings per share, dividends or capital gains from the purchase of a stock; cash flows to equity and debt holders from investment in a project finance venture; or cash flows that can be obtained from acquisition of an entire corporation. In most business applications, making the cash flow projections requires some kind of financial model. A well structured financial model, the subject of this chapter, facilitates realistic and effective risk assessment (chapter 3) as well as the application of different valuation techniques (chapters 4 and 5). To many people, mechanical issues addressed in this chapter that deal with how to structure, build and evaluate deterministic financial models (that is, models which do not incorporate probability distributions) are far more mundane topics than issues such as developing Monte Carlo simulation, real options, debt capacity and making forward price projections discussed in subsequent chapters. If your primarily interest is in more theoretical valuation issues or if you already consider yourself to be proficient in building models, this chapter can be skipped without affecting understanding of the subsequent chapters. On the other hand, if you have not developed models or if you are an experienced model-builder and would like see some different modeling ideas, the techniques described in this chapter can hopefully make your life easier. One of the most influential and lasting ideas in finance has been the work of Modigliani and Miller in 1958 who demonstrated that the focus of valuation should be on free cash flow that can be distributed to debt, equity any other investors rather than the way the cash flow is split up between alternative investors (chapter 4). If you are a true believer in the theory developed by Modigliani and Miller, you should then stop financial models after computing operating cash flow (EBITDA) capital expenditures, working capital changes and taxes on operating earnings (EBIT). These are the components of the typical definition of free cash flow, namely operating cash flow less taxes that must be paid before money goes to investors, less capital expenditures that must be made to sustain operations and less working capital changes. Although calculating prices, demand, cost structure and the cost of capital equipment the drivers of free cash flow -- is the most important aspect of any model, almost all of the valuation techniques described later in the book require computing earnings, debt and equity cash flows and financial ratios that require the balance sheet. Therefore, much of the discussion of financial models in this chapter consider the financial structure of the company and the distribution of free cash flow to debt investors, equity investors and income taxes. After the financial crisis that was in large part caused by the undisciplined and often fraudulent sub-prime lending practices along with the associated structured finance products such as collateralized debt obligations, many suggested that these toxic investments are simply too complex and opaque for investors to understand and model. Investment bankers had supposedly created dangerous overlycomplex products that could not be modeled and should have been tightly regulated. To model the value of these structured investments (so named because they split-up operating cash flows to different investors) financial models had to be created that not only measure operating cash flow, but who gets the cash flow in what order in alternative states of the world. Modeling of CDOs is famous for being confusedly elegant and relying on complex statistics that are all but impossible to interpret. By working through the financial modeling mechanics in this chapter, you should see that valuation errors made because models or studies are incomprehensible are no excuse for poor risk assessment. Building a financial analysis were one sees who receives cash flow in what order and then performs risk analysis even for a toxic collateralized debt obligation is not difficult if you are careful with the structure of the model. Even if you are an excel phoebe you should be able to understand financial models and make informed decisions about the risk of a security.

Chapter 2 Financial Modelling

Page 2

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

Probably the only real way to learn financial modeling is working late at night with a tight deadline under the pressure of a transaction. Notwithstanding this, outlining the structure of models and presenting real world examples in this chapter is intended to provide a head start for those who have not build models and will have to learn the hard way. Further, some of the concepts presented below can be helpful even to experienced model builders in designing and structuring more efficient, stable, transparent and accurate models. The chapter provides practical instructions on how to build a well structured financial model that clearly delineates inputs, effectively presents key value drivers, uses separate modules to organize various components in a logical manner, accurately computes cash flow that is available to different debt and equity investors, and presents results of the analysis that effectively represents risks of the investment. Unlike other chapters where statistical, economic and financial theory provides the basis for making forecasts and assessing risk, the modeling issues discussed here can generally be explained using a cook-book style step by step approach. A bit of theoretical discussion of how different types of models can be used to establish value is included, but the main objective of this chapter is simply to provide details on how to build better models. In discussing the process for building an efficient financial model, this chapter is organized into the following sections which correspond to general structure of a financial model: Model objectives and the general notion of keeping models simple Structure and layout of alternative types of models Avoiding bad programming practices Organizing time periods in a model Projecting revenues, expenses and capital expenditures in a working analysis Programming the debt schedule with alternative loan structures Developing fixed asset modules with depreciation and amortization Creating the income statement and tax schedule Establishing the cash flow statement and a cash flow waterfall Generating a balance sheet and other auditing tools Presenting key valuation outputs of a model Computing the value of investments from cash flow forecasts in alternative models

In addition to these subjects, the chapter includes four appendices that address specialized topics. The first appendix reviews the idea of valuing an investment from free cash flow and the weighted average cost of capital. The second appendix addresses terminal value calculations that are an important issue in corporate models. The third appends describes how to adjust project finance models to model real estate projects. The final appendix addresses modelling issues that arise in valuing financial institutions. Some of the subjects discussed in this chapter such as organizing time periods of the model, using techniques to verify the accuracy of mechanical calculations and computing tax depreciation are not very glamorous. While these topics may not have dramatic effects on valuation, use of good modeling practice can improve the efficiency of the process and allow you to spend more time on the important issue of risk analysis (chapter 3). Moreover, there are many practitioners who have created models the wrong way for a long time who can attest that a few simple ideas regarding structuring and programming models can dramatically improve the operation of a model and ultimately improve valuation analysis. In addition to describing the idea behind various modeling concepts, each section of this chapter includes a few practical programming tips that that can make financial models more efficient. Further, a set of model building exercises are included on the CD that address each modeling issue if you would like to practice. The exercises are designed so that you can work through each of the modeling techniques and be able to master the programming approaches when applying them to any financial model. Finally, complete template models are also included on the CD so that you can see what a completed model may look line and steal some modeling ideas.

Chapter 2 Financial Modelling

Page 3

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

Model Objectives and Keeping Models Simple

As emphasized in the case studies of valuation mistakes from the first chapter, effective risk assessment is the centerpiece of valuation. Depending on the valuation approach, risk assessments address risks to equity holders, risks to senior debt providers or risks to other parties such as contract counterparties. The central idea of building a model of future cash flows is to understand risks in a transaction, whether the transaction is simply purchasing a stock, borrowing money, investing in an airport, acquiring a company or signing a contract. Assessing risks can mean evaluating to what extent a particular covenant reduces risk of default to determining how demand growth affects cash flow. In one way or another, the risks must be gauged against the returns. Some seemingly complicated modelling techniques presented in this chapter. Yet despite working through a few somewhat complex modelling issues such as net operating loss, cash flow waterfalls, accelerated depreciation, debt defaults and resolving circularity, it is generally a good recommendation to keep the models as simple as possible. Instead of worrying obsessively about details, it is often better to concentrate on assuring that risks of key variables are effectively presented alongside the value statistics. After the financial crisis, there has be a movement back to fundamentals and creation of simple models which assure that the most important variables do not get lost in overly complex models and transactions. For example according to an article in the Financial Times, Moodys complex financial models for 1 evaluating a certain type of CDP, a CPDO, contained a bug had a big impact on ratings : A single small error in the computer coding that Moodys used to run its CPDO performance simulation had thrown the results way off. When the error was corrected, the likelihood of CPDO default increased significantly. CPDOs, it turned out, werent triple-A products at all. Preliminary results suggested the error could have increased the rating by as many as four notches. Instead of making the models overly complex, when one is evaluating the risk of subordinated debt in a structured finance transaction, a simple but effective way to present risk is to see how low or high a variable can move before defaults occur the break-even point (Chapter 3). Once the break-even point is known, then the focus of the analysis can be on what is the possibility that the variable can reach such a pessimistic level. Monte Carlo simulation with sophisticated probability distributions may not be nearly as useful in evaluating the investment decision as a simpler break even analysis. Much of the time, the most effective models are relatively simple with a good presentation of risk. Models should not become impossible to understand large black-boxes with thirty different 2 spreadsheets. As well stated in a document named The Financial Modelers Manifesto: Our experience in the financial arena has taught us to be very humble in applying mathematics to markets, and to be extremely wary of ambitious theories, which are in the end trying to model human behavior. We like simplicity, but we like to remember that it is our models that are simple, not the world. Financial models used to value investments can be broadly categorized into three different types deterministic models, stochastic models and back of the envelop models. Deterministic and stochastic models receive the most attention in discussion of valuation, but the third type, the back of envelope models should be part of the process: Deterministic Models: Deterministic models are the kind of models most of us are familiar with. Even with all of the methods presented below to make the models transparent and logically structured, these deterministic models can become large and difficult to audit. Risk analysis in deterministic models is generally computed using judgmental assessments about how selected variables can change relative to

FT: When junk was gold (rating agencies and sub-prime), By Sam Jones. Illustrations by Balint Zsako, Financial Times, Published: October 17 2008. 2 Emanuel Derman Paul Wilmott, January 7 2009 The Financial Modelers' Manifesto

Chapter 2 Financial Modelling

Page 4

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
baseline case assumptions. For example in valuing a hotel, one would develop projections of the room price per night, the occupancy rate, operating expenses and so forth to compute projected cash flow and value. Stochastic Models: Stochastic models build on deterministic models but include probability distributions around key variables. The probability distributions depend on relatively sophisticated mathematical analysis of economic variables and their correlation with one another. (Chapter 3) After the stochastic variables are added to a financial model, one can compute probability distributions associated with key valuation measures such as rate of return or probability of default. For example in analyzing a hotel, one could assess the probability distribution of future occupancy rates and convert those to a distribution of rates of return. Back of the envelope models are very often more important than the other two model types. These models may involve developing some kind of metric to check whether the model results are reasonable or simple statistical checks of the valuation. For example, in valuing a hotel one could calculate the value per room and assure it is reasonable relative to the costs of other similar hotels. Alternatively you may compute the return on investment on a year by year basis and think about whether it is really possible to earn the returns given the competitive structure of the industry. Needless to say, coming up with effective ways to independently check a model can require more creativity and be more difficult than the other models.

Back of the Envelope:

The description of modeling methods in this chapter and the next address the first two model types: deterministic models and stochastic models. This does not mean that the third model type, the back of the envelope model, should be considered less essential in the valuation process. The importance of this type of model is demonstrated by a case where electricity plants were sold by one utility company named Central Main Power Company (CMP) to another utility company named FPL Energy for $886 million. The plants being sold were old and consisted of 1,185 MW; the largest plant in the portfolio was an oil fired plant in Maine named the Wyman plant (594 MW) which had entered service about twenty years earlier. When selling the plants, CMP used an auction process which of course came along with a deadline. FPL was able to buy the Wyman Plant and others by making a bid that was higher than the bids submitted by other companies and it is safe to say the FPL experienced a case of buyers curse. FPL most probably used very sophisticated projections of future electricity prices, transmission constraints, and natural gas prices along with a big deterministic financial model in coming up with its bid. As is common in this sort of analysis, the modeling was likely finalized late at night and everyone was probably very tired. In the end, the FPL bid for the old oil plant and the other used plants of $748/kW was a lot more than the cost of new and much more efficient natural gas combined cycle plants which were being constructed at a cost of less than $600/kW. According to rumor, FPLs sophisticated models contained errors that resulted in the high valuation. A back of the envelope analysis would have developed an approximate valuation by discounting the cost of a new plant to consider age and efficiency factors. Little more than a year after the purchase, FPL, the buying company took a $176 million write-off 3 as a result of paying too much for the CMP assets. In this situation, a much simpler model that adjusted the cost of a new plant for the age and other factors would have been more important than the sophisticated deterministic or stochastic models and it could have caused management to question whether there is something fundamentally wrong with their complex techniques.

Model Layout for Four Types of Valuation Analysis

Before beginning to enter data developing any formulas or making valuations, you should think about the architecture of a model. The architecture of any computer model involves structuring the inputs,
3

2001 FPL SEC 10-K, the official reason given by FPL for the write-off was a change in regulations regarding transmission pricing.

Chapter 2 Financial Modelling

Page 5

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
constructing mathematical calculations and presenting outputs. When pondering the general design of a financial model, one must decide how to organize the inputs from various information sources in an organized manner; how to formulate the mechanical calculations in a transparent manner that is easy to audit; and, finally how to present the outputs so that purposes of risk assessment and valuation. Other than thinking about the inputs, calculations and outputs of a model, other subjects that should be considered in structuring the model are the programming of time periods, the methods for verifying model accuracy and the setting-up of alternative scenarios. An essential part of constructing effective models is simply putting things in order. The layout and ordering of items in a financial model to a large extent depends on the type of investment being assessed. Most financial models can be classified into four general categories -- corporate models, project finance models, acquisition models and merger integration models. Because of different data sources and alternative valuation techniques, the layout of financial models is different for each of these four types of models. In addition to the four general models, specialized models can be constructed for real estate development projects as well as for banks and other financial institutions. Many of the issues for these specialized models are a subset of the four general model categories discussed below. However there are some unique issues associated with the models including structuring portfolios of real estate projects where each project has different start dates and programming maintenance of adequate capital ratios for bank models. These specialized issues are discussed in Appendix 2. Each of these models has different objectives and a different structure. The valuation techniques, data sources and outputs of these models can be summarized as follows: Corporate Model: The distinguishing feature of the first and most common model type, a corporate model, is that a corporation has a history and it is assumed to last indefinitely (although virtually all will either end up in bankruptcy or eventually be purchased.) This means that valuation of a corporation begins with historic analysis and the models must include some kind of terminal value assumption. The terminal value calculation is necessary because it is not reasonable to make detailed forecasts of cash flow item for the indefinite life of the corporation which would require forecasts for fifteen to fifty years into the future.

Project Finance Model: The second type of model, a project finance model, differs from a standard corporate model because the investment is characterized by different phases with different risks; the fact that no history on cash flows exists for the investment (no matter how many times a similar new combined cycle plant is built, you dont know how it will work until you switch it on); and the isolation and quantification of particular risks. The project finance models focus on cash flows accruing equity holders and lenders rather than earnings and the projections generally cover the entire defined lifetime of the project. Acquisition Model: The third type of model, an acquisition or leveraged buyout model, measures the returns earned from an acquisition transaction which is characterized by the consideration paid for the equity of the company, the holding period of the investment and exit price as well as the manner in which the acquisition is financed. To compute equity returns, acquisition models measure the manner in which alternative financing sources are repaid and ultimately compute the return earned by equity investors. The final type of model, an integrated consolidation model, computes earnings per share and credit quality measures on a standalone and a consolidated basis before and after a merger or acquisition. This type of model considers the specific financing and accounting of the transaction as well as cost savings or synergies generated by the transaction. A common application of this type of model is to evaluate how much can be paid for a company before earnings dilution will occur and bond ratings can be maintained.

Merger Model:

Chapter 2 Financial Modelling

Page 6

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

Structure of a Corporate Model

Corporate models are used to compute free cash flow for making valuations using the DCF approach. They are also necessary for making projections of earnings per share; assessing the ability of a company to repay debt; and evaluating the reasonableness of future financial ratios such as return on equity and return on invested capital. The first step in creating a corporate model should be evaluating historic data so as to develop assumptions and understand the financial performance of a company. As part of the historic evaluation, the models can use financial ratios such as return on invested capital, EBITDA margin and credit ratios to tell a story about what has happened to the company. This historic analysis of financial ratios also provides an important basis for comparing projections with actual results. For example, one can compare the historic return on investment and the historic EBITDA margin with ratios in the projected period to evaluate whether forecasts are reasonable. Different ratios may be appropriate to paint a picture of different companies: sometimes it may be trends in return on investment; in other situations it may be the ratio of capital expenditures to depreciation; in yet others it may be the interest coverage ratio. A big problem for investors in the case of Constellation Energy discussed in Chapter 1 was that there was no way to really dissect the financials of the company and compute effective metrics that told a story about how the company was taking excessive risks in order to grow. The example below shows how graphs of historic data in a historic model can provide insight as to where value comes from. The company, Nestle of Nigeria, earned returns that were above the cost of capital (the easiest way to compare the return on investment with the cost of capital is to compare returns to the interest rate rather than trying to make sophisticated calculations of weighted average cost of capital.) At the same time, revenues were growing and the capital expenditures were greater than depreciation. This resulted in large increases in the stock price shown in the final panel of the figure.
Return on Invested Capital:Nestle Nigeria
80% 72% 70% 60% 50% 40% 30% 20% 10% 0%
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
4,000,000

Capital Expenditures and Depreciation

71% 67% 67% 59% 58%

3,500,000

55%

58%

3,000,000

Capital Expenditures Depreciation

2,500,000

37% 29%

2,000,000

1,500,000

1,000,000

500,000

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Revenues: Nestle Nigeria

30.85% 30.00% 29.03%
45,000,000

Stock Price
35.00%

27.83%

300.00 250.00 200.00

Growth Rate

25.24%

25.14%

40,000,000

26.90%

21.22%

35,000,000

23.02%

30.00% 25.00% 20.00% 15.00% 10.00%

30,000,000

Revenue

25,000,000

20,000,000

150.00 100.00 50.00 0.00

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

15,000,000

10,000,000

5,000,000

5.00% 0.00%
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

After painting a picture with historic financial statement analysis, structuring a corporate model includes defining how one incorporates history, uses an opening base year balance sheet and connects interest expense and interest income in the income statement to the balance sheet debt through evaluating the cash flow. The process of beginning with a balance sheet and then working through each component of the balance sheet in distinct parts of the model produces a structure which should be transparent and mechanically accurate. In a corporate model, two very simple ideas can dramatically improve the structure of the model. The first is simply realizing that all of the fixed asset accounts, debt accounts, working capital accounts, deferred tax accounts, accumulated depreciation accounts, surplus cash

C o m p o u n d G r S a l e s L e v e l

Chapter 2 Financial Modelling

Page 7

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
accounts, and other items come straight from the initial balance sheet. The second is setting up accounts for all of these items where the initial closing balance comes from the balance sheet. For example, consider the fixed plant asset account. The closing balance for the account should come straight from the balance sheet. In the subsequent period, the opening balance is the same as the prior period closing balance. Capital expenditures are added to the opening balance (and retirements can be deducted from the balance) to yield the closing balance for the next period. Once the closing balance is established, the same process is repeated for the remaining periods of the model. The diagram below is intended to illustrate some of the important points about structuring a corporate model. The process begins with analysis of history and the initial balance sheet. The working analysis, the fixed asset balance and the debt schedule shown in the middle column are the essential intermediate steps that should be completed before constructing financial statements. Developing revenues, operating expenses and capital expenditures in the working sheets is just about always the most important part of the analysis. It is generally a good idea to graph history and projections of key variables that drive these three things such as prices, capacity utilization, variable costs and capital expenditures per unit. Once these preliminary parts of the model are complete, financial statements can be constructed. As most of the components of financial statements have already been computed, this part of the analysis, shown in the third column of the diagram, should be quite simple. The main computational challenge in a corporate model is to determine how surplus or deficit cash should be deployed and to develop projections of depreciation and deferred taxes that account for the lifetime of different asset classes. These issues are discussed in detail below. Since a corporation is assumed to last for an indefinite period, the valuation section of a corporate model typically includes an assumption about the value of the company at some terminal date (see Appendix 1 to this chapter). In a corporate model, capital expenditures are made on a continual basis to sustain operations, implying that on-going assumptions with respect to new financing must be developed. The diagram illustrates that connection of the profit and loss statement with the interest expense and interest income that is a typical characteristic of a corporate model (which can create circularity). The arrow on the right of the diagram is meant to show that interest expense on new debt calculated from the model is not known until the debt balance is computed from the cash flow and interest income on surplus cash is not known until the amount of surplus cash is derived. Unlike project finance models, dividends are determined from an algorithm such as a payout ratio and are not the result of the cash flow process.
Structure of a Standard Corporate Model

Inputs: Historic Financials Operating Drivers, Working Capital Analysis Financing, Tax Debt Schedule of Existing Issues Taxes Paid, Taxes Paid and Taxes Deferred Revenue, Expense and Capital Expenditure Analysis Profit and Loss Fixed Interest Changing Interest

Balance Sheet Free Cash Flow

Cash Flow Statement

Fixed Asset Schedule Book and Tax Depreciation Initial Balance Sheet

Cash Balance, Debt Balance Surplus Cash Balance Equity Balance

The final column of the diagram on the right shows that the balance sheet is an output rather than part of the mechanical calculations. To make the balance sheet, the common equity balance can be calculated like all of the other balance sheet accounts. As with the other accounts it begins from the opening balance of the equity (in the first year from the initial balance sheet). To compute the closing balance, the net income and equity issuances are added and the dividends and equity buybacks are subtracted. A

Chapter 2 Financial Modelling

Page 8

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
similar account can be computed for minority interest. With all of the accounts completed including the equity balance, the balance sheet can be computed by simply gathering together the closing balance of all of the accounts. If you want to work through a corporate model structure in excel, complete the first exercise in the corporate model folder.

Structure of a Project Finance Model

A project finance model contrasts with corporate models that are founded from historic data. Key outputs from the project finance model are related to cash flow: the growth rate of equity cash flows which is called the equity IRR (discussed in detail below); and, the ability of cash flows to cover debt service in each period which is called the debt service coverage ratio (Chapter 4). Project finance models are typically used for two things. The first is to structure the size of the debt, the tenor of the debt and the manner in which debt will be repaid. The second is to assess the risk of the project given the financial structure. Unlike corporate models, the information base for project finance models is a series of contracts, construction budgets, engineering data and possibly projected commodity prices. Given that there is no historical record for a project finance corporation called a special purpose vehicle (SPV) and therefore no historic balance sheet, the launching point for projection of financial statements is a sources and uses of funds during the construction period rather than the balance sheet. An example of outputs from a project finance model is shown in the diagram below. Key drivers of the project such as the capital cost per unit of output, the product price and the utilization are shown together with the IRR and various different ways to measure the debt service coverage.

The structure of a project finance model is illustrated on the diagram below. One of the essential elements of a project finance model is that different calculations are made for distinct phases of the project the development phase, the construction phase and the operation phase and the debt repayment phase. The sources and uses statement is computed during the development and the construction phase. While the things that go into this part of the model are not complicated to think about what ones spends money and how one raises the money the source and use statement provides a good picture of what the project is about (recall Eurotunnel from Chapter 1). From a mechanical perspective, the sources and uses statement replaces the balance sheet as the starting point for the balance sheet accounts such as the fixed assets, the debt balance, the debt service reserve balance and so forth. This is illustrated in the diagram where arrows from the source and use statement launch the fixed asset schedule and the debt schedule. The working module that computes revenues and expenses is similar to the corporate model as is the fixed asset schedule and the debt schedule. Part of the debt schedule is computing the interest during construction that is capitalized to the cost of the plant. Once the debt schedule is computed, the profit and loss statement and the tax analysis are simple to construct. The most complex part of making financial statements is the cash flow waterfall that defines

Chapter 2 Financial Modelling

Page 9

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
which investors receive the highest priority of cash flow is often the most complicated part of the process. Since a project is generally a one-off investment where debt and equity investors focus on cash flow rather than accounting earnings, the structure for computing cash flow is also different in a project finance model than a standard corporate model. The final part of the cash flow waterfall is the dividends paid to the owners of the SPV (often called sponsors) meaning that dividends are not defined by a dividend payout, dividend per share or some other algorithm, but rather are the residual cash flow not paid or reserved elsewhere. As illustrated in the diagram below, effective modeling of cash flows involves integration of the debt scheduling with the cash flow waterfall in the cash flow analysis and launching from a sources of funds analysis. Risk analysis of a project finance model is also different phases as the construction period typically involves technical risks and the operation period includes economic risks. Finally, as with the corporate model, the balance sheet is part of the output in which all of the accounts such as plant balance, debt service reserves, senior debt balance, subordinated debt balance and common equity balance are already defined. If you want to get your hands dirty and work through the structure of a project finance model, complete the first exercise in the project finance model folder.

Structure of a Project Finance Model

Revenue, Expense and Capital Expenditure Analysis Inputs: Working Capital Analysis Operating Drivers from Contracts and Other, EPC Contract, S-Curve, Interest Rate Tax Sources and Uses of Funds During Construction Including Interest Roll-up Debt Schedule

Profit and Loss

Taxes Paid, Taxes Paid and Taxes Deferred

Fixed Assets Interest Capitalized Fees and Other

Cash Flow Statement With Waterfall, Debt Defaults, Sweeps etc. Cash Balance, Debt Balance Equity Balance

Balance Sheet Equity IRR DSCR, LLCR

Structure of an Acquisition Model

An acquisition model is typically developed to determine how much to pay for a target company and how much of the purchase price can be financed with different types of debt including amortizing debt, debt with a bullet maturity and debt with capitalizing interest (this is sometime known as ABC.) Key assumptions in an acquisition model include how much operating cash flow can grow with new owners and a new strategy as well as the how much the company can be sold for after the holding period. Financial ratios used in assessing an acquisition are often related to the EBITDA. These include the enterprise value (debt plus equity value) to the EBITDA; the senior debt level to the EBITDA; and the total debt level to the EBITDA. As with a project finance model, an acquisition model can be used for both structuring the amounts of different type of debt as well as risk assessment. The figure below illustrates output from an acquisition model that could include an illustration of how cash is distributed, break-even points and a J-Curve that illustrates optimal holding periods and equity cash flows that drive the equity IRR.

Chapter 2 Financial Modelling

Page 10

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

The structure of an acquisition model is illustrated in the diagram below. The diagram illustrates that modeling a leveraged acquisition transaction involves combining some aspects of corporate models and other aspects of project finance models. As with a corporate model, the history of the company should be studied when developing acquisition models. However, as with a project finance model, an acquisition model should begin with a sources and uses analysis that shows how much cash is used for the transaction and where the cash comes from. After the sources and uses map is established, a goodwill analysis should be added that allows construction of a pro-forma balance sheet. The general structure of computing a pro-forma balance sheet through mapping the sources and uses of funds along with a goodwill analysis that incorporates the accounting aspects of the transaction is shown in the second column of the diagram. Once the pro-forma balance sheet is created, the modeling process contains similarities to both a corporate model and a project finance model. As with a corporate model that works through different asset and liability accounts on the balance sheet, an acquisition model works through accounts where the first year closing balance comes from the pro-forma balance sheet. An acquisition model also has phases as does a project finance model in particular, the transaction period should be distinguished from the holding period and the terminal period. In addition, an acquisition model also has phases as does a project finance model the transaction period should be distinguished from the holding period and the terminal period. The right hand side of the diagram shows that the cash flow modeling process is analogous to the methods described for a project finance model where a waterfall progression measures the priority of cash flows to the various sources of funds and ultimately the equity holders.

Structure of an Acquisition Model

Inputs: Operating Drivers from History, Acquisition Price and Financing Sources, Tax

Revenue, Expense and Capital Expenditure Analysis

Profit and Loss

Working Capital Analysis

Taxes Paid, Taxes Paid and Taxes Deferred Debt Schedule Cash Flow Statement With Waterfall, Debt Defaults, Sweeps etc. Cash Balance, Debt Balance Equity Balance

Sources and Uses of Funds

Fixed Assets Fees and Other

Goodwill and Purchase Price Allocation Balance Sheet Equity IRR Debt IRR

Pro-Forma Balance Sheet

Chapter 2 Financial Modelling

Page 11

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital Structure of an Integrated Merger Model
The idea of a merger integration model is evaluate how much to pay for a company and how to structure the financing of a merger. To do this an integrated merger model can be developed that compares earnings per share with a merger and without a merger along with the credit quality of the combined company. The prospective earnings of the combined company depend on how much is paid for the acquisition and how many synergies are generated from changing management. When measuring the cost and benefits of a merger, the information base is the historic operations, projected cost savings and/or revenue increases and transaction terms. The valuation section below illustrates the output of an integrated merger model and explains how the valuation of a target company can be accomplished without worrying about the discount rate or the terminal value. The diagram below of an integrated model demonstrates that this structure mixes elements of an acquisition model and a standard corporate model. As with the leveraged acquisition model, the starting point of a merger integration model is a sources and uses analysis of the transaction and the pro-forma balance sheet after goodwill and other adjustments. The only difference is that the pro-forma balance sheet begins with existing balance sheets of both the target company and the acquiring company. For an integration model, the transaction assumptions incorporated in the sources and uses analysis may include a share exchange, multiple debt issues and new equity offerings. Once the starting balance sheet established, the remaining calculations of the cash flow process are similar to the corporate model where computation of net cash flow leads to the accumulation of a net cash balance and the net cash balance is separated between short-term debt and surplus cash liquid investments. The debt schedule shown in the middle column of the diagram includes issues retired in the transaction (shown in the sources and uses analysis) and new issues to finance the merger as well as the associated fees and breakage costs. Shares issued in the transaction could come from offering new shares or from the shares issued to target company shareholders as part of a share exchange transaction. The diagram shows that instead of computing a working analysis to derive revenues, expenses and capital expenditures these numbers come from individual corporate models for the target company and the acquiring company. The most difficult item to project is the synergies that arise from management changes that occur with the transaction. In the real world, these synergy projections must be made with only public information and no time.
Structure of an Integrated Consolidation Model

Profit and Loss Target Company Financials Inputs: Operating Drivers from History, Acquisition Price and Financing Sources, Tax Goodwill and Purchase Price Allocation Balance Sheet Pro-Forma Balance Sheet EPS Accretion Credit Measures Acquiring Company Financials Taxes Paid, Taxes Paid and Taxes Deferred Debt Schedule Cash Flow Statement Sources and Uses of Funds Fixed Assets Fees and Other

Cash Balance, Debt Balance Equity Balance

A comparison of the elements that are included in the layout of different types of financial models is shown in the table below.

Chapter 2 Financial Modelling

Page 12

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Project Finance Corporate Model LBO Model M&A Integration Model

Information Base Model Starting Point

Contracts and analysis of Historical financial Commodity Prices and statements; Analysis other value drivers of value drivers Sources and Uses Analysis

Historical financial statements; Analysis of value drivers; Transaciton Terms

Historical financial statements; Analysis of value drivers; Transaciton Terms

Historic Balance Sheet

Sources and Uses and Sources and Uses and Pro-Forma Balance Pro-Forma Balance Sheet Sheet

Cash Flow Process

Cash flow waterfall that ultimately measures dividends paid to equity

Net cash flow after Cash flow changes Cash flow waterfall dividends that result in that result in changes that ends in dividends changes in short-term in short-term debt or paid to equity debt or surplus cash surplus cash Existing Debt Issues; Retired Debt Issues; New Debt Issues EPS analysis period

Debt Analysis

New Debt Issues from Transaction

New and Existing

New Debt Issues from Transaction Transaction holding period

Model Termination

End of project life

Arbitrary terminal period

Model Complexities

NOL; cash traps and sweeps; construction NOL; target capital period issues; debt structures; circularity; service reserves; debt depreication vintage sculpting DCF Valuation; EPS projection; Implied P/E; Credit Quality

NOL; cash sweeps; Pro-forma balance interest capitalization sheet; minority interest on sub debt; debt changes; new debt service reserves; issues terminal period Project EPS and Other Equity and Debt IRRs; Ratios on Standalone Debt/EBITDA vs Combined Basis

Model Output

Equity IRR; Project IRR; DSCR

Subsequent sections of this chapter discuss issues in creating each of the major model components. In addition to the general description of what should be included in each section, a discussion of some practical programming tips for those who are involved in construction of models on a regular basis.

Avoiding Bad Programming Practices

For purposes of valuation and investment analysis, models do not have to be overly complex with numerous different spreadsheets and tedious detail of items that are relatively insignificant. Most peoples reaction to receiving a model with thirty different sheets and excruciating detail of operating expenses is to put the model in a drawer and to do no risk analysis with the model. While effective valuation models may be relatively simple, creating a model that is flexible enough to handle different risks; that accurately measures cash flows; that presents the key value drivers and important outputs in an easily understandable and transparent manner; and that does not crash excel because of circular references requires a disciplined approach to excel programming. From first-hand experience in developing models that are not well designed, the benefits of setting-up financial models in a careful structured manner and using practices that make the models more accurate and stable become obvious. A well structured model can maybe avoid you having the ghastly feeling of, after completing a model, being asked by a supervisor to open and revise a model that was created a few weeks earlier. This feeling often comes from knowing that you will have to review a long and complex formula with multiple if statements that can no longer be interpreted. These types of formulas are generally incomprehensible without completely re-constructing the formula and splitting it up into many different rows.

Chapter 2 Financial Modelling

Page 13

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

To avoid incomprehensible, un-transparent and unstable models that are virtually impossible to audit, some companies require use of a set of best practices when programming models. While these practices can help, many people become very obsessed with the programming practices and lose sight of the ultimate objective of a model, which is to effectively measure the future cash flows and risks of those cash flows. For example, in order to keep formulas the same across excel columns for the transaction period and subsequent periods, complicated if statements may be developed that are difficult to interpret. Instead of suggesting there is such a thing as best practice, a list of practices that should make a model easier to interpret and modify is presented below: Divide the model into separate modules, beginning with input modules and make the inputs a separate color (the word module simply refers to a separate part of the model -- it could be a separate spreadsheet page or simply a segment of one sheet.) When entering inputs in one or more modules, operating inputs should be separated from financial inputs. Make calculations of operating cash flows in a separate module to measure how the value drivers result in operating revenues, operating expenses and capital expenditures in a working analysis. This means that elaborate revenue or expense calculations should not be directly in financial statements. The calculations in the working sheet should be as transparent as possible through repeating inputs and including all of the factors that are used in making various computations. Include separate modules for debt issues, fixed plant assets, working capital and cash balances. The debt module should include the interest rates, interest expenses, repayment mechanics, prepayment possibilities, financing fees and potential for default. The fixed plant asset schedule should include calculation of book and tax depreciation. Carefully specify the time period of the model using time period codes that define alternative phases of the analysis. Work through every single balance sheet item showing the opening balance, additions and subtractions from the account and the closing balance for each the accounts. This analysis should be made for everything single title in the balance sheet ranging from cash accounts to common equity. Limit or avoid the use of macros and iterations to resolve circular references as circular references are not present in the real world and fixing circularity makes many risk analysis programming techniques more difficult. Use the balance sheet and other items as auditing tools and include a separate integrity page of the model to present verification checks. Include a dashboard at the top of each page of the model to monitor the integrity and key outputs of the model. Assure that no formulas in the output module of a model affect anything in any other section of the model. This means that you should be able to delete the output section without causing any reference errors in the mechanical calculation parts of the model. There should be no inputs in any part of the model other than the input page. One of the worst and most obvious problems is to include inputs as part of a formula; these partial inputs are difficult to find and make the models inflexible. Make sure that spreadsheet columns are consistent throughout the model and that the formulas for each column are identical (with the possible exception of the very first period in corporate models and acquisition models.)

Chapter 2 Financial Modelling

Page 14

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Keep formulas in the model as simple as possible and clearly delineate how each formula is derived from the inputs (this is often a problem with long if statements). Long formulas can be avoided by splitting formulas into multiple different rows and by using TRUE/FALSE switches. Use the positive number convention which means that individual items are shown in the model as positive numbers and subtotals are explicitly shown as additions or subtractions. For example, capital expenditures are listed as positive in the cash flow statement with the caption less: capital expenditures and cash flow after capital expenditures subtracts the capital expenditures. The above principals seem obvious when written down in a book. However, in the midst of creating a model to meet a deadline, it is very easy to ignore these principals as they seem to slow you down when working under a tight deadline. Examples of bad practices include entering input data in the fifth sheet of a model; including complicated revenue and expense calculations with long if statements directly in the income statement; forcing the balance sheet to balance using a cash item; computing interest expense without separately listing debt issues and not integrating financial statements. The single worst practice that is often made by otherwise very good modelers is to make formulas that are too long. An example of a formula (for projected prices) that is almost impossible to verify is shown below: =IF(AB5<14,IF(AB2=6,AA39*(1+inflation),AA39),IF(AB5=14,'OperatingInputs'!$E$103/'OperatingInputs'!$ E$104,IF(LEN(AB5)=2,AA39,AA39*(1+inflation)))) This formula, taken from an actual model, contains a number of bad practices. One problem is that fixed numbers are included in the formula (i.e. the number 14 and the number 6.) The larger problem is that the formula is far to complex to easily verify and audit. This formula could be vastly improved if one would split it up into a number of separate rows where one would show the inflation rate in a separate row, the tests in separate rows and the alternative results of different conditions in various rows. Indeed, when asked to review somebody elses model, it is a good idea to split up formulas like this one. This formula took about fifteen rows and once the more transparent separate rows were presented, several obvious errors were apparent. A few practical tips for creating a well structured model are described below. Other practical tools that can be helpful in structuring a model are included in the file named fm.xls included on the accompanying CD.

Practical Tips for Good Modeling Practice

The next section walks through details of how to implement some of the above ideas. If you are interested in general modeling concepts or you want to be able to read models but not to program them, you should skip the detailed discussion of practical programming and read the next section on time periods and valuation. Short-cut keys and Setting-up the Model Area In structuring a model it is helpful to set-up the model so that you can work with the model in an efficient manner. The more a model can be created with excel short-cuts rather than using the mouse, the faster you will be able to develop and modify the model. (In some financial modeling courses imposed on young bankers, the mouse is removed from computers so students learn how to navigate around spreadsheets much more quickly, lest they waste a minute in creating collateralized debt securities out of sub-prime loans.) Use of short-cut keys may seem a bit odd when you first use them instead of the mouse, but if you practice, they really can make your work on the model easier. One of the most helpful shortcuts is the combination of the following two keystrokes that allows you to very quickly copy and paste rows in a financial model: 1. 2. Press the SHIFT,CNTL, (right arrow) at the same time to mark a row, and Press CNTL,R to copy the formula or data to the right.

Chapter 2 Financial Modelling

Page 15

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

The problem with this short cut is that it copies the contents of the first cell to each column of the sheet, all the way to the right end of the sheet. To limit the copying of columns to the maximum rows of your sheet the maximum number of periods that you may forecast, you can hide the columns for which you do not want the item to be copied. To hide columns to the right of the model periods, you can group all of the columns to the right of the last column in your program. To use this method, use the SHIFT,ALT, combination to group columns and press the number 2 on the top left corner of the spreadsheet as summarized below: 1. 2. 3. 4. Shade the entire column of the first column to be hidden (you can used the CNTL, Spacebar combination) Press the SHIFT, CNTL, to shade the remaining columns in the sheet Use the SHIFT, ALT, combination to group the remaining the selected columns. Press the Number 1 in the square box at the top left of the sheet.

If you need the columns to the right, you can press the number 1 button instead of the number 2 key. Results of the grouping approach are shown on the diagram in the next section where the right part of the sheet after the columns have been grouped is grey. Many other short-cut keys and hints are included in the excel background folder of the accompanying CD. Color Conventions In a well structured financial model, colors should be used as a guide to what is happening in various cells of the model. This means the colors should not be used to make the model into an attractive piece of art, but it is better that they provide readers of the model with a quick understanding of where the numbers in a cell come from. One principle is that input cells should be colored differently from other cells, generally through using the fill color and the background color. There are few methods to color inputs. A simple method is to use the F5 key and then the press the special key. An alternative method is to use a macro provided in the CD (a set of macros is included on the accompanying CD that allows you to automatically find each input set the color.) It is a good idea to show a color key in the input section or the table of contents of a model. A second principle is that a different font color should be used for cells in which data comes from another sheet. For example, if price data is transferred from the input sheet to the workings sheet, that cell should have a color to notify users in what sheet they can quickly find the source of the data. A macro that sets the font color of a cell from the tab color of the sheet included with the materials. Other colors can be used for cells that have been computed from the goal seek or the solver and for cells that are computed through the operation of a macro. Finally, cells in which calculations are made from information in the same sheet should not have a font or a fill color. The example below illustrates a couple of the excel set-up ideas discussed above. The sheet (from an integrated consolidation model) is set-up using the SHIFT, ALT, method to limit the size of the sheet with the grouping method (when you press 1 at the top left, the columns are hidden, while when the number 2 is pressed, the columns are unhidden.) The red cells come from the target company financials which has a red tab color while the blue cells come from the acquirer sheet that has a blue tab color. Finally, the example illustrates that it is a good idea to make a few small columns on the left of the spreadsheet so that it is clear what rows are sub-totals and headings.

Chapter 2 Financial Modelling

Page 16

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

Use of TRUE and FALSE Switches and Model Integrity Page Setting up a model where a series of mechanical checks can be automatically monitored is effective when building and using a model. A useful technique is placing a dashboard on each page of the model that shows key outputs, inputs and sensitivity analysis switches. In addition, the dashboard on each page should show an overall check on the integrity of the model does the balance sheet balance, is the cash balance positive, are the dividends non-negative, do the sources of cash equal the uses of cash and so forth. Sometimes it is useful to include a series of checks as to whether debt is in default as well as mechanical checks. An effective way to set up the integrity check of a model is to use TRUE and FALSE logical variables. The use of TRUE and FALSE switches is helpful in many parts of the model. For example, it eliminates the need for painful if statements that can be very difficult to audit and the variables can be used to test covenants and run the model in different modes of operation. A TRUE/FALSE variable can be created by simply using and equal sign (for example, =1=1 is TRUE). Variables that have a value of TRUE or FALSE are helpful because the =AND function can be applied with a series of logical variables to test if the overall value is TRUE (analogous to the sum command for number.) Further, when if statements are used, they can be used with the TRUE or FALSE as in the logical section of the statement. A prominent verification check is to test whether the balance sheet balances in every period and it is good practice to show each balance sheet item in a different module of the model. The following step by step process illustrates verification with the balance sheet test. Step 1: After computing the balance sheet, subtract the assets from liabilities Step 2: In a separate row, use a TRUE/FALSE logical variable (created by setting one cell equal to another) to test whether the difference, after rounding, is equal to zero. (The rounding is necessary because the difference is often not precisely equal to zero if, for example, there is a division by three somewhere in the model).The formula is =ROUND(difference,0) = 0. Step 3: Once such the TRUE/FALSE result is established for each period, the =AND function can be used to test whether the balance sheet balances in every single one of the periods. Specifically, the AND function is used on a range of TRUE/FALSE logical variables and placed to the left or the right of all of the year by year tests.

Chapter 2 Financial Modelling

Page 17

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Step 4: Link the TRUE/FALSE result from the AND statement to a page of the model that contains other verification checks (for example, in a project finance model or an acquisition model, check whether the sources of funds equal the uses of funds.) Step 5: Create a separate verification page that combines the balance sheet test with other tests (again using the =AND function) to develop a test for the overall integrity of the model. Step 6: Place the overall TRUE/FALSE test in the dashboard that monitors each page of the model. The figure below illustrates the verification of the balance sheet that feeds into an aggregate balance sheet test. The row labeled test uses a switch and the ROUND function to make sure the balance sheet balances in each period. If all of the tests are TRUE, then the aggregate test computed with the AND function is also TRUE.
8-Sep-08 31-Dec-08 31-Dec-09 Balance Sheet Plant Assets Less: Acc Depreciation Net Plant Senior Debt Mezz Debt Equity Total Difference Test Aggregate 990.00 990.00 306.90 336.60 346.50 990.00 TRUE TRUE 1,030.00 15.40 1,014.60 222.51 369.92 422.16 1,014.60 TRUE 1,071.60 66.90 1,004.70 140.33 406.55 457.83 1,004.70 TRUE 31-Dec-10 31-Dec-11 1,114.86 120.48 994.38 41.04 446.79 506.55 994.38 TRUE 1,159.86 176.22 983.64 348.01 635.63 983.64 TRUE 31-Dec-12 1,206.65 234.22 972.44 165.50 806.94 972.44 TRUE 31-Dec-13 31-Dec-14 1,255.32 294.55 960.77 960.77 960.77 TRUE 1,305.93 357.31 948.62 948.62 948.62 TRUE

In addition to mechanical tests of the model, tests of whether a model is in default on any debt issues should also be developed. In a project finance model and/or a leveraged acquisition model, one of the key questions is whether cash flow is sufficient to pay all debt service and make required payments into debt service accounts and other requirements. If the cash is not sufficient, then defaults occur and ultimately with low enough cash flow, debt cannot be repaid. When the debt cannot be paid, the model should clearly present default on a dashboard. These issues can be addressed with the same type style of programming as the verification tests. For example, a test can be established for default on debt service reserve accounts and a test for negative cash flow. An example of a page that shows verification checks is illustrated in the figure below.
Aggregate Test FALSE Problem with Balance Sheet

Town Balance Sheet Sources and Uses Negative Dividends Repayment Period Debt Balance Commitment Test Test of Model TRUE TRUE TRUE TRUE TRUE TRUE

Partnership TRUE TRUE TRUE

Private FALSE TRUE TRUE TRUE TRUE TRUE

Town - Hybrid TRUE TRUE TRUE

PRIVATE Private Balance Sheet Does not Balance Sources and Uses OK Private Dividends OK Debt Repayment Period OK Debt Repaid OK Debt Sizing OK

TRUE

FALSE

TRUE

Use of Range Names to Structure and Document Models When entering data, the cell references such as B2 can be renamed using range names such as holding period. For most people in the world, the question of whether or not to use range names in a financial model is not very interesting. But amongst financial modelers, the issue of when to use range names for data can become a heated debate. Some suggest that every row in a model should have a range name such as revenues, costs and EBITDA. This can be accomplished using the SHIFT, CNTL,

Chapter 2 Financial Modelling

Page 18

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
F3 short-cut which forces the range names to be documented and allows one to revise the range name. With range names in the spreadsheet, the model formulas are documented (for example, the formula in excel is EBITDA = revenues costs rather than C5 = C4 C3.) Further, the use of range names forces all formulas in the sheet to be consistent across columns because if the row name is used in a column (the formula is defined using SHIFT, CNTL, ENTER) then once cannot change the formula definition across a single row. Finally, if the model is structured with range names, one cannot mix up the years by putting the year 2010 in one column in one place in the model and then put another item for 2010 in another column in the model. While some insist on range names in models, others argue that over use of range names is not helpful because the original source of formulas (e.g. B5 x AB92) is lost (e.g. int_rate x debt_bal). This means that without cell references in formulas, the model may be somewhat more difficult to trace. Further, if a named range is deleted, the model becomes unstable as the formulas can no longer be computed. If one copies a sheet that contains range names to another sheet, the range names do not transfer. Similarly, if one combines files two different files together that each contain the same range name, then excel must decide which range name to use. Finally, using range names can make the model more tedious to program, particularly for the opening balance sheet period in a corporate model and the transaction period in an acquisition model. While the range name debate is not an exiting subject to most people, a compromise is probably best for most financial models. Some inputs such as the holding period, the plant life or the acquisition price may be useful to name. Further, use of range names is very beneficial, if not essential, when the inputs are used in macros, when inputs are used in forms (combo boxes, spinner boxes and so forth) Setting-up a Dashboard When there are multiple sheets in a model, it is nice to show a dashboard in which you can change different items and you can show outputs as shown on the example below. It is also helpful to show the model verification switch in the dashboard. This is very simple to add to each sheet after you have worked on the model if you use range names for the output variables and you have structured forms in the model to work with multiple sheets. The simplest way to create forms is to make an extra working sheet and refer the cell links to different sheets. In this way the forms can be copied from one sheet to another while retaining the cell links in the appropriate sheet. When creating a dashboard, it is useful to use range names for all of the inputs and outputs shown on the page so that the dashboard can easily be copied to added spreadsheet pages. The example below illustrates use of a dashboard for a leveraged acquisition transaction. The drop-down boxes and spinner boxes may seem a bit like a gimmick, but they can be used to effectively summarize the key inputs and the key risks in a transaction. With these buttons, one can evaluate different capital structures, different operating cash flow levels in terms of return as well as risk. In this case, the risk is measured as the break-even cash flow that can support the level of senior and mezzanine debt before the return falls below the risk free rate.

Chapter 2 Financial Modelling

Page 19

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Waterfall Senior Junippr
Mezz Sweep Full Sweep

Check

TRUE

Senior Mezz Equity Mgmt Total

Percent 31% 34% 35% 1% 100%

IRR 7.42% 19.03% 27.60% 36.02% 22.81%

BE 13% 33% #N/A 57%

To EBITDA 2.05 2.24 2.28 6.57

EBITDA Sensitivity Year to Apply Sens Holding Period Entry Multiple Exit Multiple

130.0% 4 7 6.60 4.00

Cash Flow Waterfall

350.00 300.00 250.00 200.00 150.00 100.00 50.00 -

Cap Exp Senior Mezz Equity EBITDA

2008

2009

2010

2011

2012

2013

Time Period Structure and Valuation

A seemingly trivial but important part of the process of developing a model is setting-up dates and time period codes that classifies different various phases of a transaction. Efficiently structuring the time periods assures that you can you to gauge the effect of issues such as the length of the holding periods of an acquisition, delays in the construction of a large project, lengthening of the life of an investment, or the amount of time before which a stable growth rate is achieved. From a mechanical perspective, carefully setting-up time periods allows calculations of items such as interest during construction, terminal value, depreciation and amortization, debt service reserve repayments, gain on sale of assets and other items.

Corporate Models and Terminal Period

A corporate model encompasses a historic period, an explicit forecast period, a terminal period and sometimes a fade period. For a corporate model, the structure of time periods should begin with a base period which is the balance sheet of the final historic period. It is a good idea to label this base period as zero so that the explicit forecast period begins with one and increase thereafter. Given the subjectivity of estimating future growth, it is useful to be construct the model so as to be able to evaluate different periods before which the explicit forecast period ends. Finally, corporate models can be structured to include a fade period in which cash flow growth declines from the rate achieved in the terminal cash flow period until a stable growth rate is obtained. The diagram below illustrates different time periods in a corporate model and the importance of the assumptions developed for the final forecast period.

Chapter 2 Financial Modelling

Page 20

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Terminal Value in Corporate Model
Year after explicit period to establish stable cash flows

Gordons CFt x (1+g)/ (WACC-g)

History

Explicit Forecast Period

PV of Cash Year t Step 2: PV from Year t to current year Make sure Cash Flow in period t reflects realistic ROIC, Working Capital that reflects long-term growth and realistic capital expenditures

Terminal Value
Infinity Step 1: PV to year t End of period t, so Gordons method must use t+1 cash flows

EV/EBITDA that Reflects Long-term Growth Rate and ROIC

Valuation Date

Market to Book Ratio that Reflects Long-term ROE

Project Finance Models and Different Phases

Structuring the dates is particularly important in project finance models where different accounting, financing and risks occur in the different periods. These different stages that occur when making a new investment typically include the time period when the project is being developed before construction the development period; the period when construction occurs and investors are putting money into the projects; the period of operation of the project, and finally, the period in which the project terminates. The mechanical calculations in a model and the risks are different for these different time periods. Further, the length of the periods in a model often varies for the different phases of a project. In many models, the construction period is presented on a monthly basis in order to accurately measure interest during construction since the calculation requires accumulation of debt for each month. On the other hand, when the plant begins operation, the model switches to semi-annual periods because of debt is repaid on a semi-annual basis. In order to construct time periods, one can create switches that define each phase of the project as well as the important millstone dates such as the financial close and the commercial operation date. The programming of different phases can become quite complex if the millstone dates do not start at the beginning of a month and if there are different time periods modeled in each phase of the project. The essential practice is to make a period code which is one at the commercial operation date, negative during operation and development and measures the age of the project. The figure below illustrates how various stages of a project can be modeled with a period code. The mechanics of developing the various switches is discussed in the section below.

Set-up of Project Finance Model with Different Phases

Financial Close

Commercial Operation Date

Debt Repayment Date Retirement

Develop

Construction

Debt Repayment

Large Dividends

Negative of Construction + Develop Periods less 1

Period 0 Start + Develop

Period 1

Periods of Operation e.g 20 x 2 periods per year = 40

Chapter 2 Financial Modelling

Page 21

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

Acquisition Models Transaction Period and Exit Period

In the case of acquisition models, it is also important to spend some time setting-up time periods. In this case, separate switches can be developed for the acquisition period, the holding period, and the terminal period. The acquisition period should be a separate period in the model that only lasts one day so that the mechanics of the on-going calculations are not mixed up with the transaction assumptions. The period subsequent to the acquisition period can be a fraction of a year followed by annual periods until the final sale period. The exit period should also be separate from the holding period in order to isolate on assumptions with respect to the sale of the company. Developing periods in this manner allows the model to handle movements in the transaction date as well as alternative holding periods of the investment. With the holding periods defined, a J-curve illustrating how the rate of return is affected by the length of time between the purchase and sale can be established as illustrated in the graph below.

J-Curve Entry Multiple 11.60 Exit Multiple 9.00 Senior Debt/Capital 62.7%
35.00%

30.00%

25.00%

Equity IRR

20.00%

15.00%

10.00%

5.00%

0.00%

10

Holding Period

Use of TRUE and FALSE Switches for Modeling Time Periods

Mechanically, the definition of periods can be developed by defining the various dates and then programming separate variables that contain switches that define the phase of the project. After defining the dates and periods of various phases, logical variables (TRUE or FALSE) that define each phase of the model with switches can be created. For example, the terminal period in a corporate model would have a value of FALSE for each period except for the terminal period, which has a value of TRUE. If the model is set up beginning with zero for the last historical year, then the logical variable for the terminal period is simply programmed to be when the terminal period equals the time period. Use of the logical variables that have a value of TRUE or FALSE is helpful because TRUE is equal to one and FALSE is equal to zero. A switch can be created by simply using and equal sign (for example, = period <= 0). It can also be created using and AND, OR, or NOT function (for example, AND (period > 0, period <= retirement date) defines a switch for when the plant is operating. Establishing the TRUE and FALSE statements is also effective with if statements. One can simply use the TRUE or FALSE in if statements and make the statements much easier to read. The step by step process for setting up dates at the beginning of the sheet includes:

Chapter 2 Financial Modelling

Page 22

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Step 1: Input months per period for various phases of the project typically one month for construction period and six months for operating period. Step 2: Develop period code this is an essential part of the process, the last date before operation is period zero and the project begins its life in period one. It does not matter at all that the length of some periods is different than other periods. The start period of construction is the negative of the number of construction months minus one. For example if the start date of construction versus the completion date yields 40 months, the first period is negative 39. Then, the first period of construction will be zero as desired. (Note: it is a good idea to hide remaining columns) Step 3: Develop switches (TRUE/FALSE) for different phases and important dates construction period, operation period and terminal period. Step 4: Correspond the months per period from the switches. For example, for the construction period, the months per period is one. Step 5: Enter the dates using the EDATE function. Show both the beginning date and the ending date. The first beginning date is the established date and the ending date is the EDATE function with using the months per period. The beginning period in the second period is the ending period plus one day. Step 6: Compute the number of days per period and the days in 360 day year (first period next year) Once the switches are established, a number that defines the number of months per period can be computed that differentiates the construction period from the operation period through multiplying the switch by the twelve divided by the periods per year multiplied by the switch. For example, if the construction period has is monthly, then the periods per year is twelve divided by one or twelve and if the operating period is semi-annual, the periods per year is twelve divided by six or two. If the dates do not begin at the first day of the month, then the EOMONTH function should be used rather than the EDATE function. Further, the ending month must change fir different millstone dates including the start of the project development, the financial close and the commercial operation. For all other months, the ending date is the end of the starting month that can be computed with the EDATE function. A final complication results because the time periods are not constant and must be adjusted if the start date is more or less then the middle of the month. An example of setting up dates with different phases for a project finance model is included in the reference CD.

Tips for Computing Start and End Dates for Each Period
Once the period length for project phases, the switches, and the period codes have been defined, the dates of the model can be established. The dates cannot be incremented by simply adding numbers to the starting date. For example, you cannot add 30.5 to the last date, and have each month begin at the first day of the next month. Instead, you can use the EDATE function or the DATE function. If the EDATE function is used, then the excel analysis tool pack must be added in and the formula is simply EDATE (prior period, periods per year). If the DATE function is used, the YEAR, MONTH and DAY arguments of the function should refer to the previous date and the period per year should be added to the MONTH argument. Given the definition of periods for year from the switches above, the DATE function would be DATE(YEAR(prior date), MONTH(prior date) + periods per year, DAY(prior date)). After the dates have been computed, it is a good idea to explicitly show the start date and the end date for the period and the number of days in the period. In addition, the number of days in a 360 day period should be shown using the DAYS360 function so that interest expense or interest income can be computed per the appropriate convention. To set-up a model that accounts for different project phases, a few excel techniques are helpful. First, the total construction period should be defined by subtracting the completion date from the date at which construction begins (the construction start date and the completion date should be defined in the model along with the number of periods per year for construction and the number of periods per year for operation.) Through defining dates in this manner, the model can be flexible enough to account for delays in construction of the project. To compute the start date in this manner on a monthly basis, use the DAYS360 function in excel and enter the start date of construction and the completion date

Chapter 2 Financial Modelling

Page 23

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
(DAYS360(start date, completion date), then divide this number by 30. To illustrate this process consider the following example: Start date of construction: 1-March-2010 End date of construction: 1-December-2015 DAYS360 (begin, end) : 270 Divide by 30 : 69 The issue of developing dates is more complex when the periods do not occur at the beginning of the month. In this case, you should use the EOMONTH rather than the EDATE function. Much of the other process is the same you should still define the number of months per period and the start and the end of the period. In this case, the EDATE function is defined with the opening date minus one month as illustrated below.

Tips for Converting Periodic Data into Annual Data

It is simple to annualize periodic cash flows once the fiscal year has been established. This can be accomplished by listing the year below the date and then using the SUMIF function to sum the periodic cash flows within the year. To demonstrate the process, the establishing the fiscal year is discussed and then the process of using the conditional sum is described. Establishing the fiscal year can be demonstrated by considering the example of Microsoft Corporation. For this company, the fiscal year 2007 ends at June 30, 2007 and covers the period July 2006 through June 2007. Here, one would define the fiscal year as June. For programming purposes, January through June correspond to the current calendar year (2007) and the later months, July through December, correspond to the prior fiscal year. In project finance models, for convenience, the fiscal year should in the month just before commercial operation. With this assumption, the first year covers a full year of operation. For example, if construction ends in May and the project begins in June, the fiscal year should end in May. Then, the first fiscal year would cover the period from the beginning of June at commercial operation to the next June. Once the month of the fiscal year is established, one can use the YEAR function along with the MONTH function along with an if statement to determine the fiscal year of alternative periods. For the example above where the fiscal year ends in May, if the month is less than to June, the fiscal year equals the calendar year. On the other hand, when the month is greater than June, the fiscal year is the next calendar, or the calendar year plus one. Once the fiscal year has been established, the periodic cash flows can be converted into annual cash flows for balance accounts that record the balance at end of the year and for cash flow or profit and loss accounts that accumulate amounts into the annual periods. The following step by step process walks through the process and allows one to very quickly convert periodic cash statements to annual statements. Step 1: Compute Fiscal year If the month of the opening date is less than the fiscal month, (e.g. June) then the fiscal year is the same of the calendar year (e.g. 2007). (For example, for months of January through May, the year is 2007.) If the year of the opening date is greater than or equal to the fiscal month (e.g. July), the year is the calendar year plus one (e.g. 2007+1). (In the example above, June through December are set to 2008.) Step 2: Copy titles of cash flow, balance sheet and other items to a separate section or a different sheet. Enter the year of for the report on the top of the sheet as well as the month number of the fiscal year. Then use the date function to create the ending date of the fiscal year (see the figure below).

Chapter 2 Financial Modelling

Page 24

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

Step 3: Use a TRUE/FALSE switch to identify codes that are accumulated over the year and the accounts in which balances are taken from the end of the fiscal year period. Step 4: Apply the SUMIF function, where the initial range is the fiscal year definition, the criteria is the year and the sum range is the cash flow item. To make this work shade the entire row of the fiscal year in the periodic section of the model (click on the row number to shade the entire row) and then fix the references (press the F4 key). Next, refer to year entered for the report in step 2 above and use a relative reference to lock-in the row number but not the column name (press the F4 key twice). Finally, refer to the range to be summed by clicking on the entire row without locking in the row number or the column number. This process will work for the cash flow items to be accumulated over a period. Step 5: Use a similar process using the SUMIF function to apply to the balance accounts. In this case the sum range is the date itself and the criteria is the ending day of the fiscal year. This means that the only items are summed when the fiscal year data used in the criteria of the SUMIF formula is the same as the date in the periodic section (sheet of the model.) To make this work, shade the entire row of the end date in the periodic section of the model (click on the row number) and then fix the references (press the F4 key). Next, refer to fiscal year end entered for the report (see the diagram below) and use a relative reference to lock-in the row number but not the column name (press the F4 key twice). Finally, refer to the range to be summed by clicking on the entire row without locking in the row number or the column number as with the accounts to accumulate. Step 6: Use an if statement with the TRUE and FALSE switches (create a relative reference on the column name of the switch) and copy the same formula for each row in the sheet.

Financing Assumptions Year Month of Fiscal Year Fiscal Year End Day Accumulate/ Balance FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE TRUE TRUE TRUE

9 9 9 9 9 9 9 9 9 2007 2008 2009 2010 2011 2012 2013 2014 2015 30-Sep-07 30-Sep-08 30-Sep-09 30-Sep-10 30-Sep-11 30-Sep-12 30-Sep-13 30-Sep-14 30-Sep-15

Power Output Plant Capacity (kW) Capacity Factor Annual Degredation Factor Daily Degredation Factor Degredation Index Capacity Factor with Degredation Availability Period Plant Availability Number of Hours in Period Gross Generation (MWH) Paracetic Load Total Net Generation (MWH)

600 600 600 600 600 600 600 600 600 29.7% 29.7% 29.7% 29.7% 29.7% 29.7% 29.7% 29.7% 29.7% 0.0% 1.0% 1.0% 1.0% 1.0% 1.0% 1.0% 1.0% 1.0% 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.00 0.99 0.98 0.97 0.96 0.95 0.94 0.93 0.92 29.7% 29.4% 29.1% 28.8% 28.5% 28.3% 28.0% 27.7% 27.4% 1.00 1.00 2.00 3.00 3.00 3.00 3.00 3.00 3.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 8,784 8,760 8,760 8,760 8,784 8,760 8,760 8,760 1,549.77 1,530.23 1,515.08 1,500.08 1,489.25 1,470.48 1,455.92 1,441.51 20.00 20.00 20.00 20.00 20.00 20.00 20.00 20.00 1,529.77 1,510.23 1,495.08 1,480.08 1,469.25 1,450.48 1,435.92 1,421.51

Inputs, Workings Analysis and Value Drivers

The most important part of the modeling process is to accurately define and analyze input items that drive the value of an investment and effectively present how the risks of the value drivers affect the ultimate value. Indeed, the general objective of programming tips presented in this chapter is that you will hopefully have more time to study the drivers of value. The drivers of cash flow generated from any asset are things like forward prices, capacity utilization ratios, plant efficiency ratios, input prices, fixed and variable cost and the capital cost per unit of capacity to build the plant. The modeling and valuation mistakes discussed in the last chapter were not due to incorrectly structuring a model or making incorrect calculations in the model, they were the result of not using valid economic and financial principles in developing inputs to the valuation process. Other parts of the book deal with how to develop inputs for prices, discount rates, volumes sold, operating costs and other inputs. These inputs can be studied by

Chapter 2 Financial Modelling

Page 25

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
reviewing historic data, performing statistical analysis, applying marginal cost concepts, and considering industry structure. In addition the value drivers should use judgment as to whether sudden non-linear changes can occur, industry expertise and perhaps mathematical simulation. This section describes more mundane issues concerning how to structure, document and present inputs in a financial model.

Structure of Inputs
The most important point about inputs is of course the assumptions that are behind the numbers. However a couple of notes with respect to the structuring of inputs in a model is worth summarizing. The most important element in setting-up inputs is that the data should be classified in a manner so that the data can be easily found. The inputs should be on the same page (or set of pages) and the relevant input categories should be grouped together. If scenarios are developed, a separate page or area should contain all inputs for the scenarios including the base case and alternative scenarios. First, the inputs should be set-up in a way that anybody can easily find the inputs and understand exactly what each input means. For example, a model has a input labeled development percent would be confusing because one has no idea what is the basis for the percentage. Second, the inputs should allow the model to be adjusted and not be restrictive. For example, when there is one inflation rate applicable to all future years, the model obviously has limited flexibility. Third, the inputs should be arranged in a logical manner and grouped together by categories. Separate sections can be shown for inputs for the general transaction dates, purchase prices and so forth; operating revenue and expense drivers; capital expenditures; financing parameters; and tax assumptions and other items. Fourth, no matter how tempting, outputs from the model should not be mixed together with the model inputs. Finally, when using the inputs in other sections of the model, the inputs should be repeated and not used in the middle of the calculations.

Value Drivers; Supply Driven Models versus Demand Driven Models

In setting-up a financial model it is best, whenever possible, to begin with the capacity or demand of the company or project. The starting point should generally be the capacity of the company or the project in a supply driven model or the demand for the product in a demand driven model. To demonstrate the difference between demand and supply driven models consider an upstream oil project. The most likely place to begin the upstream model is with the capacity of the oil fields in terms of oil and gas reserves implying the model would be supply-driven. You would not begin with an analysis of the worldwide demand for oil because oil prices the company operates in a commodity business and it will sell whatever it can produce. Alternatively, for a company operating in a limited market region that may have market power, the model would begin with demand for the product and market share of the company. that has a The diagram below illustrates how the value drivers should be connected to one another in a model in order to produce revenues, expenses and capital expenditures for developing cash flow models and valuing investments.

Chapter 2 Financial Modelling

Page 26

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

Most time in the valuation and modeling process should concentrate on statistical and economic analysis of the inputs that are have the most significant effect on value the value drivers. The true art of modeling is quickly identifying these value drivers and then developing reasoned projections of what could happen to these inputs in the future. This analysis can range from simple graphing of historic data to complex regression analysis of mean reversion tendencies to judgment with respect to political risks. As a side note, value drivers are not items such as revenue growth, operating margins, return on investment or the ratio of capital expenditures to sales. These are statistics that are the result of the economic position of a company or a project, not the items that are the ultimate source of value. When setting-up a model, it may seem that establishing an inflation index is straightforward and simply a matter of multiplying one plus the inflation rate by the prior inflation index. One must be careful in defining the base period for which prices are defined and escalate from that period. Difficulties can arise when time period lengths change and when intervals are used for inputting the inflation rate. Discussion of looking-up data using the MATCH and FUNCTION functions is discussed later. The step by step process below illustrates how to deal with varying periods. This process involves converting annual rates into daily rates and computing the index from the number of days in the period. The procedure is analogous to verification of the XIRR discussed in the output section Step 1: Convert the annual rate into a daily rate using the formula (1+Annual Rate)^(1/365)-1. Step 2: Beginning with 1.0 for the base period, compound the index through multiplying the prior index by (1+daily rate)^(days in period) (Also cannot have positive of zero cash flow in first period in XIRR)

Excel Tips for Data Validation

When entering data, the data validation tool in excel can be useful in assuring that the data you have entered is valid. The tool can also be useful in documenting the source of inputs. You can make sure the data is entered is in the intended range (such as a number between zero and one), you can make sure the sum of a series of data inputs accumulates to one (such as depreciation rates) and you can assure that a data input is within a given set of numbers. To use the data validation tool, select the validation option from the data menu. If you simply want to make sure the data is a decimal value between zero and one, select the decimal option from the settings option. Then enter zero for the minimum and one for the maximum. To validate that the sum of depreciation rates sum to 1.0, Data validation techniques; use of match command to link data to appropriate periods in a flexible manner.

Setting-up Scenarios with Forms, and INDEX Command

One of the most important uses of models is to develop risk analysis. To prepare the model of risk analysis, one should set-up the model in anticipation of the risk analysis through entering different possible scenarios in the data input section. The task is to enter the data without inordinately cluttering up the model. This can be accomplished through entering data and then grouping the data. In addition, use of forms together with the INDEX or CHOOSE command is generally effective. The process of setting-up scenarios is summarized below: Step 1: Enter different scenarios for an input item by simply typing the different scenarios on a set of rows. Step 2: Enter a row number corresponding to one of the scenarios in an open cell somewhere in the spreadsheet. Step 3: On a blank row below the list of scenarios, use the CHOOSE or the INDEX function along with the selected row to create the chosen scenario (the row number should be fixed so the function can be copied across the rows. Step 4: Use the selected scenario as the variable in the model.

Chapter 2 Financial Modelling

Page 27

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
If you would like more specific guidance on creating scenarios, please refer to the file named scenario and sensitivity analysis on the CD.

Excel Tips for Looking up Data with MATCH and INDEX

In developing indices of inflation, growth, and loan pricing grids, inputs can be presented using a grid that shows the from date, the to date and the input item. For example, the credit spread on a loan may vary by year and also according to a financial ratio as shown in the example below where the model is created on a monthly basis, but many of the inputs are developed on a calendar year basis.
DSCR From 1.00 1.30 1.50 To 1.3 1.5 2.5 Credit Spread From 2009 To 2012 1.50% 1.30% 1.25% 2012 2015 1.60% 1.40% 1.35% 2015 2050 1.70% 1.50% 1.45%

In order to array the credit spreads in the above table with both years and the debt service coverage ratio, one can use either VLOOKUP or the HLOOKUP functions or the MATCH together with the INDEX function. To array the values where there is space between the values, the from year can be used with the lookup or the match commands. The MATCH command can use either 1, 0, or -1. The value of 0 means there must be an exact match of the value, while the 1 is used for ascending values and -1 is used for descending values.

Debt Schedules
Introduction
While most time spent in financial modeling should be studying and analyzing value drivers, it is also important to accurately reflect financial structure so the value drivers can be translated into cash flow projections that will measures value and risk. These mechanical issues that translate the key assumptions into effective analysis of the risk and value of debt and equity value discussed in the remainder of the chapter. This section describes various modeling issues that arise when incorporating debt into a corporate, project finance, acquisition or merger model. The amount of debt that can be issued and maintained on the balance sheet is central to the general idea of using the risk analysis process of lenders to derive value of an investment which is central to many of the ideas presented in this text. Furthermore, the risk to lenders in itself relative to the earned credit spread is also a critical subject in finance in itself. These issues associated with debt capacity and debt valuation require careful incorporation of a number of issues associated with debt into a financial model. Once debt is incorporated into a model, the risk of the debt can be evaluated. The cash flow to cover debt can be used to test how different levels of debt leverage affect break-even points of operating variables. Alternatively, the models can be used together with Monte Carlo simulation to compute the explicit probability of default and the required credit spread. The mechanics of computing sensitivity analysis, break-even analysis, scenario analysis, tornado diagrams and Monte Carlo simulation analysis is described in the next chapter. The importance of risk analysis together with modeling debt comes together in determining debt capacity. If the cash flow that covers debt is accurately modeled, then one can see if cash flows in a downside case can cover debt. If cash flows cannot cover debt service in a downside case (perhaps with some margin) then the amount and the terms of the debt should change. Before describing the important financial issue regarding the manner in which volatility and judgment can be used in assessing debt capacity, the mechanics of computing various features of debt must addressed. For each of the four types of models, analyzing details of the structure of the debt can be an essential component of the process of valuing cash received by equity and debt investors. One of the primary uses

Chapter 2 Financial Modelling

Page 28

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
of corporate models is to assess the credit classification of debt while the project finance models and leveraged buyout models evaluate how specific risks affect the ability to repay debt. Risks of debt and risk classification of debt is crucial to the overall valuation process if one accepts the notion of using debt capacity to evaluate risk using the analysis of lenders whose job it is to evaluate risk, who are from outside of the organization and who have money at risk. Using a model to classify risk could be accomplished with Monte Carlo simulation as described in the next chapter. Alternatively, a more realistic approach is to compute the break-even points of different variables from the perspective of debt repayment. For example, in evaluating an electricity generating plant, one could compute how low the electricity price can fall before the debt defaults. If the electricity price can fall by a wide margin, the loan is less risky than if the electricity price can only fall slightly. Assuming the electricity price is the key item that is subject to uncertainty in the model, if one knew the probability of the electricity price falling below the break-even level, one would also know the probability of default.

Debt Capacity from Cash Flows with Different Volatility

High Risk Cash Flows Low Risk Cash Flows

High Volatility of Cash Flow

Low Volatility of Cash Flow

High Risk Project has higher margin, shorter-term and declining debt service. Low risk has flat debt service, and longer-term and higher IRR on Equity

Given the importance of evaluating risks faced by lenders, reflecting the specific features of debt in a model is an essential part of the modeling process. These features include the amount of debt, the repayment structure of the debt, fees paid and spreads on the debt, covenants and required debt service reserves can be principal determinants of whether the transaction will proceed. Depending on the transaction, the covenants and credit spreads may depend on financial ratios such as the debt service coverage or the debt to EBITDA. Similarly, the nature of debt issued in leveraged buyouts is a key driver of equity returns and the financial viability of the transaction. The structure of debt repayments, new debt issues and the credit spreads is also important in corporate models and merger integration models as the amount of debt issued in a merger can be an important driver of the accretion or dilution in earnings per share. Given the importance of debt in many types of transactions, this section describes various issues associated with structuring the debt schedule in the four model types.

Basic Debt Schedule

The fundamental part of a debt schedule is defining the balance of debt outstanding and computing the interest rate from the balance of the debt. The debt outstanding should be structured by explicitly showing the opening balance, the new debt issues, the debt repayments and the closing balance on separate rows of the model. These debt balance schedules should be listed for every single existing and prospective debt facility that will be present during the forecast horizon. For corporate models, the debt schedule should include all of the debt issues that are outstanding for the last balance sheet date plus any new issues that may occur over the forecast period. In the case of project finance models, the debt issues include all of the different tranches of debt that are issued to finance construction. In addition,

Chapter 2 Financial Modelling

Page 29

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
project finance models should be set up to include debt issues that will in a re-financing. Leveraged acquisition models should include debt issues that are used in financing the acquisition as well as debt that was issued prior to the acquisition and that will be assumed by the new owners. Similarly, the debt schedules of an integrated merger model include new debt issued in the transaction as well as debt that remains outstanding for both the target company and the acquiring company. The starting point for the debt schedule differs depending on the type of model. When developing corporate models, the closing balance of each debt balance is launched from the financial inputs which should list the amount outstanding in the base balance sheet year for each debt issue. The sum of these individual issues should correspond to the total amount of long-term debt on the balance sheet (including current maturities of short term debt.) If the sum of the closing balance of all of the debt issues does not equal the balance sheet amount, the prospective balance sheet will not balance and there will be an inconsistency in debt maturities and interest expense with actual financial obligations. Therefore, a verification check to assure that the total debt on the balance sheet equals the sum of the individual debt issues can be effective in verifying the model. The process of adding a debt schedule to a corporate model is demonstrated in the following few steps: Set-up the debt schedule with separate lines for: o The opening balance o Additions from new issues o Subtractions from debt repayments o The closing balance The initial closing balance is derived from inputs that tie to the balance sheet The subsequent opening balance is equal to the closing balance in the prior period Once the opening and closing balance are computed the interest cost (whether capitalized or expensed) can be computed as a weighted average of the closing and opening balance. If the repayment occurs at the beginning of the period, the interest expense is computed on the closing balance. If the repayment occurs at the end of the period the interest expense is computed on the opening balance. If the repayment occurs somewhere between the beginning and end of the period, the interest is a weighted average of the opening and closing balance.

An example of a debt schedule in a corporate model is shown below. The quantification of interest expense on a fraction of year basis must be treated with care in a financial model. If the debt is refinanced or results in equity cash flow, this re-financing or equity cash flow must also be modeled on a partial year basis. In the case of project finance models, the new debt draw downs that define how much debt will be issued in a project are computed from the source and use statement and from some criteria that defines how re-financing will occur. Leveraged acquisition and integrated merger models combine information from the sources and uses analysis and the existing debt balance sheet to determine how much debt will be outstanding after the transaction. The manner of repaying debt depends on the type of debt issue and the type of the model. In some models, the repayment calculation can be derived from the opening balance of the debt. For example, in a corporate model where debt issues may be repaid on a single date bullet repayments a simple test can be created from the repayment year and this is used to assure that repayment only occurs on the repayment date. Specifically the programming involves: Add a line for the repayment of the debt after the opening balance Create a formula that compares the model year with the repayment year to create a logical variable (year = repayment year) Multiply the (year = repayment year) by the opening balance of the debt

The manner of debt repayment in a project finance model or a leveraged finance model is often tailored to the expected cash flows generated by the investment. For example, a project finance model may have a pre-defined set of repayment percentages that vary over the lifetime of the debt. Because of the percentages are applied to the aggregate amount of debt issued rather than the closing or opening

Chapter 2 Financial Modelling

Page 30

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
balance it is generally a good idea to show the total accumulated amount of the debt on a separate line item above the debt balance. This accumulated balance can then be multiplied by the repayment percent or a tailored repayment schedule to establish the periodic repayments. Because of early debt repayments that can occur if covenants or cash sweeps are triggered, it is important to make sure that the amounts in the repayment row do not exceed the amount of debt outstanding. To program this, the MIN function should be used to assure that the amount being repaid does not exceed the opening balance of the debt. For leveraged acquisition models and/or integrated merger models, a combination of the corporate finance and the project finance approaches can be applied depending on the type of debt used to finance the acquisition.

Computation of Interest Expense with Credit Spread Grids

The credit spread can change over time and it can also depend on financial ratios. An example of a grid is shown below.

Spread No

3 From Year To Year

1.5%-1.7% Spread 2008 2011 1.70% 1.60% 1.55% 1.50% 2011 2015 1.90% 1.70% 1.65% 1.60% 2015

DSCR

From 1.20 1.50 1.80

To 1.20 1.50 1.80 -

2.00% 1.80% 1.75% 1.70%

In developing the interest rate, two technical issues arise, the first is finding the rate to apply and the second is working through circularity that can occur. Use the INDEX command with the MATCH command. As explained above use the from column. One can show the row number and the column number from the table using two MATCH statements the row number matches the DSCR and the column number matches the year (you can group and hide the row and column number lines using the hide command SHIFT then ALT then right arrow.) Then use the INDEX function with the row and the column. To get around circularity, use the previous year DSCR in the test. If use current or forward DSCR, then must resolve circularity which can create dangerous circularity in a model. Note: Also use the DAYS360 function if the model does not have annual periods DAYS360/360 multiplied by the interest rate.

Defaults on Debt and IRR on Debt

In using a model to assess an investment with debt financing, one of the essential tasks is to determine at what point a loss on the loan occurs. When a loss on debt occurs, the interest rate or the IRR realized by lenders is less than promised interest rate and attempts to restructure the debt have not allowed temporary defaults to be repaid. The IRR earned on debt can be computed by measuring the cash flow dispersed to lenders and then the debt service received by lenders. Modeling defaults on debt is useful in credit analysis because the ultimate task in credit analysis is to measure the probability of default as well as the loss given default relative to the credit spread that compensates for risk. Through explicitly modeling defaults, one can analyze at what point the cash flow is insufficient to repay the debt after restructuring. Once the realized IRR on debt is computed through accounting for defaults, the risk of debt can be evaluated through break-even analysis, scenario analysis and Monte Carlo simulation. For example, by incorporating defaults on debt in a model, the break-even level of EBITDA can be computed. The probability of achieving this break-even level of EBITDA can then be used to assess the credit spread of the loan. The process of incorporating defaults into a financial model involves linking the cash flow statement with the debt schedule. The amount of defaults is determined after the cash flow is computed and this amount affects the debt balance that is outstanding. The following process can works through how to compute the defaults:

Chapter 2 Financial Modelling

Page 31

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

Step1: Set up the debt balance schedule with line items to incorporate defaults and re-payment of defaults. Step 2: Fill in the default row from a sub-total line in the cash flow waterfall that computes the default on debt from the inability to meet debt service (this should be computed using the formula MAX(cash flow,0). Step 3: Set the re-payment of default to be the previous years default amount implying that all defaults are paid of in the subsequent year. This means the model attempts to fully repay the default in the year immediately following the default. If there is no cash flow to repay the default, the default increases by the amount of the default. Step 4: Enter the repayment of default as a cash outflow item in the cash flow waterfall. Step 5: Assure that the repayment of default is less than the opening balance of the debtless the scheduled repayment so that too much debt is not repaid in the final year of the model. Step 6: Compute the cash flow realized by lenders through deducting defaults and adding re-payment of defaults to the scheduled debt service. A diagram that is intended to illustrate the link between the cash flow statement and the debt balance is presented below. The diagram below illustrates that the default comes from the cash flow statement, the repayment of default comes from the earlier default, and the repayment of default is put back into the cash flow statement.
Debt Schedule Opening Balance New Issues Repayments Default Repayment of Default Repay after default

Cash Flow Statement Operating Cash Flow Plus Interest Net Cash Flow to Pay Debt Service and Dividends Attempt to pay all debt service including repayment of default If positive, flows to next section of the cash flow If insufficient cash after debt service, default

Covenants and Sweeps

In negotiating with lenders, one of the subjects that must be analyzed is the level of covenants and cash flow sweeps. Covenants can range from mandating that equipment is in good working order to strict limitations on payment of dividends. For modeling purposes, it is the latter type of negative financial covenants that should be incorporated in a model. These covenants can affect returns to equity investors because they affect the timing of cash flows that are allowable to be paid out as dividends. The covenants can take the form of cash flow traps that limit dividends when times are bad (and there is not much cash flow available for dividends anyway.) Alternatively, the covenants can be in the form of cash flow sweeps which limit the cash flow that limits the cash than that can be distributed when times are good. Covenants do not change the operating cash flow of an investment and can only change the timing of who receives the money and in what order. With more restrictive covenants, the equity holders must wait while without the covenants, the equity holders can receive cash flows earlier. The timing of equity cash

Chapter 2 Financial Modelling

Page 32

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
flows can have relatively large effects on the equity IRR and assuring that cash flows accrue to lenders before equity holders can reduce the risk to holders of debt. A financial model should be able to assess the tradeoff between risk and return of covenants and cash sweeps. For example, the model can be used to assess how much the operating income can decrease with and without a covenant to evaluate the effectiveness of the covenant. If the covenant allows the operating to decline by a wide margin more than the operating income can decline without the covenant, than the covenant is effective in reducing risk. The efficacy of the covenant depends on the structure of the cash flow. If cash flow is increasing over time, the covenant will not be effective in reducing risk. If the cash flow falls off a cliff after a few years, the sweep and covenants can be effective. A summary of the analysis of covenants and sweeps is illustrated below.
Effect of Cash Sweep With Declining Cash Flows
No Enhancements
Without enhancements, the break-even is 78%
On-Shore Wind PPA, Wrapped EPC, O&M Contract, Fixed Interest Rate 1.3 Equity IRR 5.9% Minimum DSCR - Leverage 86.4% Capacity Factor Sensitivity 78% Price Sensitivity 100%
30,000.00
30,000.00

With Cash Flow Sweep

With a sweep, the breakeven is 68%
On-Shore Wind PPA, Wrapped EPC, O&M Contract, Fixed Interest Rate 1.3 Equity IRR #NUM! Minimum DSCR - Leverage 86.4% Capacity Factor Sensitivity 68% Price Sensitivity 100%

25,000.00
25,000.00

20,000.00

20,000.00

15,000.00

15,000.00

10,000.00

Dividends

10,000.00

Dividends Junior Debt Service Trap and Sweep Senior Debt Service
5,000.00

Cash Sweep

Dividends Junior Debt Service Trap and Sweep Senior Debt Service DSRA Flows

5,000.00

2017

2018

2019

2020

2021

2022

2023

2039

2017

2018

2019

2020

2021

2022

2023

2039

2040

2012

2013

2014

2015

2016

2024

2025

2026

2027

2028

2029

2030

2031

2032

2033

2034

2035

2036

2037

2038

Cash From Operations

(5,000.00)

(5,000.00)

With declining cash flows, the break-even point reduces significantly

The mechanics of modeling covenants are analogous to calculating defaults as described above. In particular, the debt schedule is tied to the cash flow statement. In the case of covenants, the cash that is not allowed to be paid in dividends can be either used to pay of debt earlier than scheduled or alternatively, it can be placed in a cash reserve account depending on the language of loan agreements. If cash is used to pre-pay debt, the cash that is trapped should be included in a line item in the debt schedule. If the cash is applied to a reserve account, a separate account should be set-up the debt module of the model. The step by step process below describes how to incorporate a cash trap or a cash sweep covenant into a model. Step 1: In the cash flow statement, add sub-total rows that show the cash flow that is available for a cash sweep or a cash covenant. Step 2: Include a test in the model that evaluates how much of the cash that is available for the cash sweep or cash trap will be unavailable to pay dividends or other subordinated debt. This may be a test against the debt service coverage ratio in the case of a cash trap covenant or a test against a debt to EBITDA ratio in the case of a cash sweep covenant in a leveraged buyout. Step 3: Use a covenant test to determine the amount of cash trapped or swept in the cash flow statement. The test could be whether the DSCR is below the covenant in the case of a cash trap, or it could be a multiplication of a cash sweep percentage. Step 4: In the cash flow statement where a line item lists the cash flow sweep or cash flow trapped, the line item should include a minimum function to assure the most cash that is trapped is the amount of debt outstanding. Step 5: The cash flow that is trapped or swept must be applied to something. One option is to use the cash to pre-pay debt and the second is to build-up cash in a reserve fund. If cash is used to pay down debt, then as long as the minimum function is used in the repayment analysis, then the debt will appropriately be paid earlier. If the reserve

Chapter 2 Financial Modelling

2040

2012

2013

2014

2015

2016

2024

2025

2026

2027

2028

2029

2030

2031

2032

2033

2034

2035

2036

2037

2038

DSRA Flows

Cash From Operations

Page 33

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
account is used, then an item must be added that releases cash from the reserve when the covenant is met or the debt is repaid. The diagram below illustrates how process for covenants or cash flow sweeps works in terms of linking the cash flow statement to the debt schedule. (Note that if the cash flow from the cash sweep or the cash trap is applied to a reserve account, the proceeds are used as cash inflows into the cash reserve.) The intermediate box in the diagram demonstrates that evaluation must be included somewhere in the model to assure that the cash flow sweeps are not pre-paying more debt than is outstanding if all of the debt is paid, there is no longer a reason to sweep any cash. If a cash sweep or a covenant is implemented, then the cash that is prevented from paying dividends can either be used to pre-pay debt or put into a reserve account. For debt holders, prepaying debt is more favorable from a credit perspective because it permanently reduces debt while money in a reserve account can be used to pay dividends before debt is repaid. Mechanically, the programming of sweeps and covenants is similar to modeling of defaults in terms of connecting the debt schedule to the cash flow statement. Structuring provisions on debt such as cash lock-ups and sweeps limit distributions to equity in different ways. Discussion of the reason for covenants and sweeps in alternative transactions. How much they really protect the debt holder and harm the debt holder. The effect in alternative transactions project finance and leveraged buyouts. Modeling issues.
Debt Schedule Opening Balance Less: Scheduled Repay Prepayments - Covenant Closing Balance Remaining Debt for Sweep Opening balance les repayment

Cash Flow Statement Operating Cash Flow Plus Interest Net Cash Flow to Pay Debt Service and Dividends Attempt to pay all debt service including repayment of default If covenant is triggered, use trap or sweep cash (subject to test) If covenant is not triggered, allow cash to flow to equity or next level

As with the discussion above, to avoid circularity, use the prior year debt service coverage ratio. When developing the model use the prior. If the terms require forward DSCR or current DSCR, then compute a fixed line and use a circularity resolution method.

Subordinated Debt
The most basic issue in finance is assessing risks and returns from an investment. This issue is highlighted in analysis of subordinated debt where the higher credit spread on subordinated debt must be evaluated relative to the higher risk of subordinated debt relative to senior debt. To evaluate the risk of subordinated debt, a financial model much be able to determine the point at which the loss occurs and the risk and return characteristics of the subordinated debt. In modeling subordinated debt, the first step is to include the debt in the sources and uses analysis. Using the sources and uses analysis, the subordinated debt should be added to the debt schedule. When developing the debt schedule, it is possible that interest is capitalized. To model the interest capitalized, the interest is added to the balance of the debt. In the final period when the subordinated debt matures, the amount of the debt repayment is the sum of the opening debt balance and the interest capitalized for the final year.

Chapter 2 Financial Modelling

Page 34

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Once the debt schedule is established for subordinated debt, the cash flow statement must reflect the priority of the subordinated debt relative to other debt service. If debt with differing priority is included in the cash flow statement, then a cash flow waterfall should be modeled that reflects the specific provisions including the interest, repayment of debt service, covenants and sweeps of the alternative debt. When structuring the cash flow statement with alternative priorities, it is essential that the ordering of titles in the cash flow analysis conform the ordering of cash flows in the loan documents. For example, if there is a cash flow sweep for senior debt and junior debt interest is not capitalized, then the senior debt sweep must occur after the junior debt interest payments. An example of how junior debt is incorporated in a cash flow waterfall is illustrated below. Notice that there are many sub-total lines.

Cash Flow EBITDA & Terminal Less: Senior Interest Less: Repayment of Senior Cash after Debt Add: Defaults on Debt Cash After Default Less: Repayment of Default Cash After Default Repayment Less: Interest Paid on Junior Cash After Interest on Junior Add: Default on Junior Interest Cash Flow for Sweep Less: Senior Cash Sweep Less: Junior Repayment Cash after Junior Repayment Add: Default on Junior Debt Less: Junior Sweep Equity Cash Flow - Dividends Cash Flow Test

TRUE

950.00 201.60 480.00 268.40 268.40 268.40 384.00 (115.60) 115.60 TRUE

1,100.00 168.00 480.00 452.00 452.00 452.00 395.56 56.44 56.44 56.44 TRUE

1,300.00 130.45 480.00 689.55 689.55 689.55 395.56 293.99 293.99 293.99 TRUE

1,500.00 76.27 480.00 943.73 943.73 943.73 395.56 548.17 548.17 548.17 TRUE

1,600.00 4.30 61.40 1,534.30 1,534.30 1,534.30 395.56 1,138.74 1,138.74 1,138.74 1,138.74 TRUE

14,688.00 14,688.00 14,688.00 14,688.00 281.69 14,406.31 14,406.31 2,816.86 11,589.46 11,589.46 TRUE

To illustrate how a model can be used to assess the risk of senior versus subordinated debt, a sensitivity analysis can be performed of the EBITDA in terms of senior IRR, subordinated IRR, equity IRR and the overall project IRR. The difference in points at which the senior IRR and the junior IRR crosses is a measure of the risk. In the diagram below the case with more subordinated debt has more risk as demonstrated by the break-even points.

IRR on Senior versus Junior Debt with Different Capital Structures

More Senior Debt More Subordinated Debt
Entry Multiple 11.60 Exit Multiple 9.00 Senior Debt/Capital 27.0% Mezz Debt 43%
20%

Entry Multiple 11.60 Exit Multiple 9.00 Senior Debt/Capital 43.3% Mezz Debt 27%
18%

18%
16%

16%
14%

E q u i t y I R R

Difference in break-even for senior and mezzenine debt

12%

10%

Equity IRR
8%

E q u i t y I R R

14%

Difference in break-even for senior and mezzenine debt

12%

10%

Equity IRR Junior IRR Senior IRR Project IRR

Junior IRR Senior IRR Project IRR

8%

6%

6%

4%

4%

2%

2%

0%

0%

80% 75% 70% 65% 60% 55% 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% EBITDA Sensitivty

80% 75% 70% 65% 60% 55% 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% EBITDA Sensitivty

Debt Service Reserve Accounts

Chapter 2 Financial Modelling

Page 35

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Project finance transactions and some leveraged buyout transactions include requirements to put cash aside in a bank account to assure that prospective debt service requirements can be made. A typical requirement is that the next semi-annual debt service must be held in an account. Such an account assures that temporary blips in cash flow will not cause a default. The problem with locking up cash in this manner is that holding cash on the balance sheet and earning a return much lower than the overall equity return can be very expensive in terms of the rate of return on equity. Further, if there is a long-term problem, the cash reserve does not mitigate much of the risk. The amount of risk reduction is questionable because all the funds really do is to hide problems for six months. Debt service reserve accounts also may be present in the construction phase when borrow more money than is needed and the surplus is placed into a reserve account. Debt service reserves can present tricky modeling issues from a technical standpoint. Mechanical issues include: (1) computing changes in the debt service reserve account that arise from changes in debt service; (2) calculating uses of the debt service reserve account when there is deficit cash flow; (3) building-up the debt service reserve account from cash flow in a project; (4) withdrawing amounts from the debt service reserve account when debt matures; (5) adding amounts to the debt service reserve that arise from a cash flow sweep; and (6) transferring amounts from reserve built up during the construction period. In general, a debt service account should be modeled in an analogous manner to other debt issues with an opening balance, additions and subtractions, a closing balance and interest income. A technique that simplifies modeling of debt service reserves is to first establish the required amount that should be in the debt service reserve at the end of the period. (Note that if the convention is that payments and are inflows occur at the end of the period, and the reserve is built up during the period, an argument could be made that the ending balance should be zero.) Since the ending balance must be equal to the required balance, the net inflows to the account (a negative amount on the cash flow statement) can be established where by the inflows are the opening balance less the are made at the end of the period. Computing the net amount in this manner is a short-cut method and one could separately show the various components including removal of the closing balance and initial funding amounts. A separate line item to withdraw from the debt service reserve for cash flow shortfalls should be shown. As with the discussion of cash flow sweeps discussed above, the amount of withdrawn from the debt service reserve comes from the cash flow statement. This item is limited to the amount of the balance in the debt service reserve. These steps are described below: Step 1: Compute the amount of the required debt service reserve. Step 2: Set-up the balance of the debt service reserve by including a line item for net inflows into the account to derive the required balance. Step 3: Include a line item for amounts withdrawn from the debt service reserve account to meet cash flow short-falls. Step 4: Include a line item below the debt service reserve balance that shows the net debt service reserve amount before cash short-falls. This amount is necessary to test how much cash flow is available in the debt service reserve when making the calculation of the amount to remove from the reserve to meet cash flow short-falls. Step 5: Include a line item in the cash flow statement for net inflows into debt service reserve (note that this amount can be negative when amounts are withdrawn because the required balance declines and/or falls to zero at the end of the debt life.) Step 6: Include a subtotal line after the payment of debt service to reflect the potential for negative cash flows that arise. Once the negative cash flows are computed, evaluate whether the amount can be met from the reserve balance. (Use the MIN(available reserve balance, MAX(-cash flow,0)) to make this calculation. Because the calculations can become messy with circular references it is helpful to compute the committed amount of debt. The notion of computing the committed debt applies to the financing fees below.

Chapter 2 Financial Modelling

Page 36

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital Circularity Created by Interest Expense
In many models, a problem of circularity arises because interest expense drives cash flow, but the debt balance or the interest expense is affected by cash flow itself. The most common circularity problem in corporate models comes from the assumption that cash flows and therefore interest expenses occur in the middle of the year. (If monthly or even a daily model is constructed, the circularity would not arise because interest is paid on the opening balance of the debt.) If the assumption of interest expense on the opening balance cannot be applied, then the interest expense affects cash flow, but the cash flow affects the debt balance and interest expense, the interest expense requires iterative calculations. In the real world, bankers do not require interest expense to be paid using a circular formula where they first compute interest expense and then re-compute the interest expense because the debt is increased by interest expense. The problem with circular calculations is that the iterative calculations can cause models to be unstable. There are various ways to resolve the circular references, but the solutions can be worse than the original problem. The problems can be resolved with macros or the solver function. The simplest and most common method involves creating a macro using the following steps: Step 1: Make sure there are two calculations of the thing that is causing the circularity (e.g. interest expense computed in the profit and loss statement and interest expense calculated in the debt schedule). Step 2: Compute the year by year difference between the two calculations and then sum all of the differences. Step 3: Record a macro and then copy and paste special as values from one of the rows to the other. For example, in the case of interest expense, copy the interest expense line below the debt schedule and paste it special to the income statement. Step 4: Modify the macro to include a while and end loop that continues until that sum of the differences equals zero as illustrated below. While Range(sum_of_difference) <> 0 Copy and paste special interest expense from debt schedule to profit and loss statement Wend

Circularity and Debt Commitments in a Project Finance Model

When a company goes to a bank to ask for a loan in a project financing, the company has already created a model and evaluated the sizing of debt, presumably through evaluating the debt service coverage or some other criteria. Some suggest from a philosophical standpoint that circularity is not a logical notion. The sponsor does not go to the bank and ask for fees and then ask for another loan as the fee changes and then, when the fee changes again, ask for yet another loan. Instead, the sponsor asks for a loan commitment that will define the fees, the debt service reserve account, and the interest during construction that can cause a circular reference problem. There are alternative methods to compute the debt commitment, most of which require some form of goal seek and/or a macro. Alternative approaches include: Multiplying the total uses of funds in each period (included capitalized interest) by a given leverage ratio. The problem with this approach is that creates difficult to resolve circular references for fees, debt service reserves and interest during construction.

Chapter 2 Financial Modelling

Page 37

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Using a pre-determined leverage ratio to determine debt commitment. Here, the given leverage is multiplied by the total uses of funds in each period. Then, the amount of the commitment that matches the given debt leverage is derived from an optimization technique (a simple goal seek). For example if the total uses of funds are 100 in each of the first three periods and the leverage is 50%, then the goal seek would compute the total debt commitment required at the commercial operation date to achieve the leverage ratio. This could be accomplished by fixing the commitment and then computing the difference between the commitment and the computed required leverage until the difference goes to zero: o Debt draws = leverage x uses of funds interested capitalized o Debt leverage at commercial operation = Total debt/(debt + equity) o Total debt commitment = difference o Iterate until difference is zero Using a debt service ratio to compute the commitment. Here, the commitment is spread across the construction period according to a ratio.

Debt Sizing and Debt Sculpting

Discuss the general idea of debt sizing. No simple formula. Establish bank case and margin. Establish a base case and make sure that can maintain debt service in downside case. Also discuss problems with the solver. Idea of making a debt commitment to resolve circularity When working on a project finance model, you sometimes would like to compute the debt capacity through sculpting the debt repayments such that the debt service coverage ratio achieves a targeted ratio. This is accomplished using the solver and the following steps: Setting the both debt leverage and the repayment for each period. Add a constraint that the computed debt service ratio equals the target debt service coverage ratio Add a constraint that the final debt balance must be zero.

It is helpful to set-up a macro to re-do the solver with a macro. Unfortunately, this is a bit complex because the visual basic must be adjusted to allow the solver to work. This process involves: Step 1: Press the Alt-F11 key to get the visual basic menu Step 2: Go to the TOOLS, REFERENCE option and then click on the SOLVER option. Step 3: After beginning to record a macro, re-set the solver Step 4: Make sure the first part of the solver is not blank Step 4: Re-do the solver Step 5: Add a Userfinish = FALSE after the SolverSolve

An example of the solver code with the adjusted finish is illustrated below. All of the lines of code except the last line come from simply recording the macro after running the solver. SolverReset SolverOk SetCell:="$I$5", MaxMinVal:=3, ValueOf:="0", ByChange:="$G$32,$G$46" SolverAdd CellRef:="$L$2", Relation:=2, FormulaText:="0" SolverSolve UserFinish = False

Chapter 2 Financial Modelling

Page 38

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
While solver seems to be a very nice tool, it can be very slow and sometimes it does not work very well. A more elegant solution is to compute the debt repayments using a formula and then size the debt with the more simple goal seek. This is illustrated using a bit of simple algebra as illustrated below: DSCR = Cash Flow/(Interest + Repayment) (Repayment + Interest) x DSCR = Cash Flow Repayment = Cash Flow/DSCR - Interest The problem with this formula is that although the repayment yields the appropriate DSCR, the size of the debt does not mean that the ending balance of debt becomes zero. To deal with the problem of non-zero debt ending debt, a goal seek formula can be used to determine the leverage percent or the debt commitment.

Fixed Asset Schedules and Depreciation

Once the debt schedule is established, it is useful to develop a separate module which establishes the book and tax depreciation on an asset. Other assets on the balance sheet that are explicitly considered in a model such as investments, debt issuance cost associated with capitalized fees and operating reserve accounts can also be included in this section of the model.

Portfolios of Assets with Vintage Process

Computation of depreciation can be complex when the depreciation rate is not constant for each year. In this situation, the age of the asset must retained to compute depreciation. This can make detailed computation of depreciation one of the most difficult parts of the modeling process. To do this, the following steps can be used: Step 1: Enter the depreciation rate by year for category. Step 2: Use the TRANSOPSE function (not the copy and paste special) to set up a matrix that has the year the asset was created on the column and the year of the model on the row Step 3: Compute the age of the plant and allow the age to be negative in years before the asset was created (use relative references and create a range name such as AGE) Step 4: Use the HLOOKUP function to relate the depreciation rate to the vintage of the plant Step 5: Multiply the depreciation rate by the cost of the asset being depreciated from the cost of the asset that was created from the TRANSPOSE function. Step 6: Repeat the process for assets with different vintages and with book and tax depreciation.

Computation of Major Maintenance and Other Reserve Accounts

Major issues are mechanical and not becoming confused by accounting. Reserve accounts are created because cash will not be available and cannot assume that financing will occur.

Profit and Loss Statement and Taxes

Once a working module that computes revenues, expenses and working capital is established, as well as the depreciation schedule and the debt schedule, the profit and loss statement should be simple to calculate. The EBTIDA is given by the revenues and cash operating expenses on the income statement. Depreciation and amortization (DA) from the fixed asset module is then subtracted resulting in EBIT. After computing EBIT, interest expense is subtracted and interest income is added resulting in EBT. For project finance and acquisition models, the interest expense and interest income is taken from the debt schedule. For corporate models, the short-term interest expense and the interest income cannot be

Chapter 2 Financial Modelling

Page 39

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
computed until the short-term debt and surplus cash is computed. With EBT, book taxes are subtracted by multiplying the EBT by the tax rates and subtracting minority interest overall earnings can be computed. Cash taxes are a somewhat more difficult than deriving book profit, although the calculation is not too painful if taxes are structured properly when setting-up a model. The computation of taxes can have an important effect on capital intensive projects such as renewable energy projects where rapid tax depreciation may be allowed, but the taxable income is not sufficient to use all of the tax deductions (including high levels of interest expense at the beginning of the modeling period.) As with the other mechanical issues regarding construction of a model one of the most important elements is simply not to be afraid to either read or construct the tax section of models. An effective way to compute taxes paid is to add a separate tax schedule that lays out the taxes paid from a net operating loss carryforward analysis. Incorrect computation of taxes paid and simplistic accounting for taxes can cause major problems in measurement of cash flow and valuation.

Computation of Taxes Paid and Taxes Deferred

A modeling process that will model taxes is to first compute the taxes on the books and the cash taxes actually paid. After the income statement is completed, the EBT from the income statement can be used as the first item of cash tax analysis. From the EBT, adjustments can be made that convert the EBT for books into the EBT for taxes. One of these adjustments involves adding back book depreciation and then deducting tax depreciation. Once the adjustments are made, the taxable income for purposes of computing cash taxes can be computed. After the taxable income is derived, adjustments for a net operating loss carryforward can be made. This involves determining the balance of the net operating loss and making adjustments that increase or decrease the carryforward balance. Once the cash taxes are established, the change in deferred taxes can be calculated along with the accumulated deferred tax. To compute the taxes paid and the taxes deferred, one can following the following step by step process in a module below the income statement. Step 1: Create a row that repeats the EBT from the income statement. Step 2: Adjust the book EBT for depreciation and other items that cause cash taxes to differ from the taxes reported on the book profit and loss statement. Step 3: Set-up and account that maintains the net operating loss balance including the opening balance, the additions that occur when there is negative taxable income and the deductions that occur when the net operating loss is used to reduce taxes that would otherwise be paid when there is taxable income. Step 4: Compute the amounts deposited into the account from negative taxes through converting the negative amounts into positive numbers. Step 5: Calculate the amounts removed from the operating loss balance through determining the minimum of the opening balance and taxable income. Step 6: Compute the adjusted taxable income after adjusting for inflows and outflows from the operating loss account. Step 7: Multiply the adjusted taxable income by the statutory tax rate to determine the cash taxes. Step 8: Subtract the book taxes from the cash taxes to determine the change in accumulated deferred taxes. Step 9: Accumulate the deferred taxes through adding the changes in deferred taxes to the prior years balance. If the accumulated deferred taxes are positive, put the balance on the liability side of the balance sheet and if they are negative, change the sign and put the account on the asset side of the balance sheet.

Adjusting the Tax Basis in Acquisition

Compute the additional tax depreciation that can occur if acquisition is purchase of assets rather than purchase of stock for tax purposes. If purchase of assets, will pay more the comparative multiples should account for this as they should account for the difference in tax rates for different countries. If the

Chapter 2 Financial Modelling

Page 40

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
acquisition is stock, no write-up for tax purposes and can use the exiting net operating loss. From the perspective of both the acquiring company and the target company, if there is no net operating loss, a stock transaction generally has a positive outcome because the gain on the sale is taxable as current income while the write-up is deducted on a periodic basis. Therefore, on a present value basis, the government treasury wins and the combined shareholders of the two companies lose. In situations with net operating loss, the situation is even worse.

Valuation from Financial Models

Valuation in Alternative Models
All four model types introduced in this chapter corporate models, project finance models, leveraged acquisition models and integrated merger models -- compute financial statements for the company being evaluated. However, the manner in which valuation is derived from the models and the ways risks can be measured is very different. Corporate models generally compute value from projected earnings or from forecasted free cash flow using the discounted cash flow (DCF) method. The foundation of this approach is the presumption that the value of real assets is not dependent on the manner in which the assets are financed. A simple definition of free cash flow cash flow that accrues to debt investors, equity investors and anybody else who put money into the company and expects a return is the following: Free Cash Flow = EBITDA Operating Taxes Capital Expenditures Working Capital Change + Deferred Tax Changes These valuations are highly dependent on the discount rate used to compute the present value of cash flows or the multiples applied to projected earnings. It is well known to practitioners that classic valuation techniques which consist of computing the discounted value of free cash flow or measuring relative value using multiples can be subject to a lot of bias and manipulation. All one has to do is ask somebody who has made a valuation analysis using the discounted cash flow technique whether the results can be manipulated through tinkering a bit with the terminal growth, the discount rate or one of the assumptions. When developing valuations using the alternative methods to the DCF backing into the value from required equity returns, debt capacity and/or prospective financial ratios -- computing free cash flow is not sufficient. An illustration of the classic DCF problem is shown in extract below from an Investment Bank Analysis of an acquisition. Note that the range in value driven by the weighted average cost of capital and the terminal value assumption (this case is relatively small compared to other examples where terminal growth rate is used in computing the terminal value.)

Equity Value of Corporation Terminal EV/EBITDA 7.0x 7.5x 8.0x 6.0% $1,002.3 $1,153.1 $1,309.5 6.5% 887.3 1,033.2 1,185.4 7.0% 777.7 919.0 1,067.0 7.5% 673.4 810.1 954.2 8.0% 574.0 706.2 846.5

WACC

Low High Pct Diff

574.0 1309.51 128%

In contrast to valuation approaches in corporate models that depend to a large extent on terminal value and cost of capital assumptions, project finance analysis in essence backs into the value of the prospective equity investment through assessing whether an adequate equity return can be achieved. The equity return in these models depends in turn on the amount of debt that can be lent to the project. This means that project finance models do not have the same problem that dramatically different results come from changing variables that are very difficult to estimate. Leveraged buyout models are similar to project finance models in the sense that one can back into the value of a company accounting for the amount of debt that lenders will commit to the transaction. With acquisition models value is computed through measuring the maximum entry multiple or premium that can be paid for a company such that a

Chapter 2 Financial Modelling

Page 41

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
given the equity internal rate of return can be achieved. As with the project finance model, the rate of return to equity investors depends to a great extent on how financing of the transaction is structured, but not on terminal growth and subjective cost of capital estimates. Finally, merger integration models also back into the value of the target company and the acquisition premium. This time however, the value is derived through comparing earnings per share and other financial ratios before and after a merger. Value can be derived by determining whether earnings per share for the combined company increase whether financial ratios of the new combined company remain strong enough to support the desired credit rating. Use of an integrated model to back into the value of a company is illustrated below. In this artificial example, the transaction results in increased earnings (accretion) and it also maintains the bond rating 4 within the standards for a BBB company given the assumed business category. If the equity consideration is increased or the amount of debt used in the transaction changes or the accounting treatment changes, the accretion may no longer exist.

Equity Consideration Equity Issued Debt Issued

1,700.00 466.48 1,338.00

Synergies Write-up

50.00 1,000.00

Years on Graph Business Risk Category Target Bond Rating

7
BBB

EPS Comparison: Target Combined with Acquiror Transaction Value 2,090 Shares Issued 52
4.50 4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 2004 2005 2006 2007 2008

Interest Coverage Comparison: Target Combined with Acquiror Transaction Value 2,090 Debt Issued 1,338
10 9 8 7 6 5 4 3

Bond Rating Standard Consolidated

Consolidated Standalone
2009 2010

2 1 0
2004 2005 2006 2007 2008 2009 2010

The alternative approaches to classic free cash flow valuation from project finance models, leveraged acquisition models and integrated merger models eliminate the need for debating the weighted cost of capital or the terminal growth rates. Instead, market information on the structure of the financing and direct information on the required returns for equity holders forms the foundation of the valuation process. If value is to be computed using the alternative models, the financial models must address detailed aspects of how free cash flows accrue to debt and equity investors. The cash flow process is different depending on the structure of the transaction implying that alternative starting points, different time period conventions and distinctive calculations of the manner in which cash flow is dispersed to debt and equity investors must be incorporated in the alternative types of models. The difference in valuation techniques that are implied from using the alternative models is summarized in the table below. Subsequent chapters will further discuss the advantages and disadvantages of using alternative valuations.

The bond rating band is computed from tables provided by Standard and Poors which show provide the range in financial ratios that are used to gauge bond ratings given a certain business risk (chapter 4).

Chapter 2 Financial Modelling

Page 42

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Valuation Analysis in Alterative types of Models
Corporate Model Project Finance Investment Decision and Implied Value depends on Equity IRR versus Market Hurdle Rate LBO Model Entry Multiple and Acquisition Premium Depends on Equity IRR and Hurdle Rate Senior and Subordianted Debt Financing and Exit Multiple M&A Integration Model Acquisition Premium Depends on Earnings per Share Acretion and Debt Ratios Sources of Funds Used for Tranasction and Assessment of Credit Quality

Valuation from Model

Present Value of Discounted Free Cash Flow or Multiples

Key Valuation Parameters

Weighted Average Cost of Capital, Multiples, Terminal Growth

Debt Capacity, Debt Terms

Traditional Risk Assessment from Equity Perspective

Sensitivity Analysis Sensitivity Analysis Sensitivity Analysis Sensitivity Analysis and Scenario Analysis and Scenario Analysis and Scenario Analysis and Scenario Analysis of DCF and Multiple of EPS Accrection and of Equity IRR of Equity IRR Value Credit Quality

Traditional Risk Assessment from Debt Perspective

Break-even Analysis Break-even Analysis to Determine Ability to to Determine at what Re-finance and Point Cash Flow Maintain Credit Rating Cannot Service Debt

Break-even Analysis to Determine IRR on Senior and Subordinated Debt

Break-even Analysis to Determine Prospective Credit Rating

Monte Carlo Analysis with Model

Probability Distribution of EPS and DCF Valuation

Probability Distribution Probability Distribution of Equity IRR and of Equity IRR, Senior Probability of DSCR IRR and Junior IRR below 1.0

Computation of Free Cash Flow from Financial Statements

In all types of models, free cash flow is generally computed to establish asset value, although equity cash flow is the emphasis of project finance and acquisition models. Traditional finance theory suggests that valuation of an investment should be measured from the present value of cash flows as if the plant was financed without any debt -- that is, discounted free cash flow. After tax cash is what has value to investors; cash is what lenders need in order to be repaid; cash is what customers pay and employees want. The notion of using cash flow and a discount rate that are independent of financing is intuitive. Say you are valuing a manufacturing facility. If you take a tour of the factory, its value should be a function of its efficiency, its ability to produce good products, its construction cost, its cost to operate and maintain and other physical factors. Its value should not depend on the manner in which money was raised to pay for construction. Without taxes, it is straightforward to demonstrate that free cash flow is equal to the cash flow accruing to equity holders plus the cash flow accruing to debt holders, minority owners, preferred shareholders and any other investors in a company. While it is well known to professionals familiar with valuation that enterprise value is equal to the present value of free cash flows discounted at the weighted average cost of capital, it is useful to review basic ideas underlying the notion. The net cost and benefits of alternative decisions are made by computing the present value of discounted free cash flow at the after-tax cost of capital which depends on the risk of the project. If the free cash flows (that accrue to all claimants, including debt, equity, preferred stock, minority interest and others) are discounted at a weighted average of the discount rates for each element, then the present value of free cash flows equal the sum of the present values of each claimant. The sum of the value of debt and equity and other claimants is defined as the enterprise value of a firm. This means that the present value of free cash flow at the weighted average cost of capital is equal to the present value of equity cash flows discounted at the cost of equity, plus the present value of debt discounted at the discount rate for debt and so forth. Equity value to

Chapter 2 Financial Modelling

Page 43

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
investors can be established from enterprise value by subtracting current market value of debt from the asset value.

Appendix 1 Finance Theory and Valuation

Introduction Review of Basic Financial Concepts
This appendix describes general issues involving how to compute cash flow and calculate discount rates from traditional financial principles that can be practically applied in valuation. These concepts free cash flow, un-leveraged discount rates, and the CAPM are not generally disputed in academic circles. However for a number of reasons, application of the theories is often impractical and difficult to effect in real world applications. Therefore, rather than recounting the tired old graphs of an efficient market line and computation of risk premium using historic market returns, problems with the method are emphasized and alternative methods are presented. Valuation is directly or indirectly the central subject of virtually every course in finance and because so many books have been written on the subject, it is not necessary to recount derivation of finance theories. That stated, it is worthwhile beginning with a couple of practical valuation concepts that are now a standard part of financial analysis. These tools provide a context for some of the more complex material 5 relating to project finance and option pricing. Some of the fundamental propositions of finance theory include and valuation equations include: (1) Valuation of an asset or the value of an entire enterprise should be measured by discounting free cash flow over the life of a project. Theorists insist that using alternative measures such as comparative earnings per share, equity cash flow, payback period analysis, or the accounting return on equity are inappropriate (even though the methods are commonly used in practice). To see how the process works, begin with the definition of the value of an enterprise which is equal to the value of the debt, equity and other claims on the company. Enterprise (Asset) Value = Equity Value + Debt Value More technically, the equation can be written as: Enterprise Value = Market Capitalization + Market Value of Net Debt Where, Market Capitalization = Number of Share x Share Price (2) The value of net debt in the above equation is the discounted present value of future cash flows received by debt holders interest and principal payments measured at a discount rate that reflects the possibility of losing money from a default on the debt. The discount rate for the debt cash flows can be computed as a risk free rate of interest associated with the maturity of each debt issue plus the credit spread.

Brealy and Meyers, Principles of Corporate Finance, Chapter 12, Irwin/McGraw Hill, 2000.

Chapter 2 Financial Modelling

Page 44

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Debt Value = Present Value (Debt Cash Flow, Discounted at Risk Free Rate plus Credit Spread) (3) The value of equity is the present value of future equity cash flow which can be defined as dividends plus capital gains less new equity issuances. The discount rate used in computing equity cash flow is the cost of equity capital. Traditional ways of computing the cost of equity capital are to apply the CAPM or the discounted cash flow model which derives the cost of capital from the long-term expected growth rate in dividends (chapter 4). Equity Value = Present Value (Equity Cash Flow, Discounted at Cost of Equity) (4) Free cash flow is earnings before interest, depreciation and taxes less capital expenditures and working capital changes and it should be computed on an after-tax basis. The total free cash flow is also the sum of the cash flow received by equity investors plus the cash flow received by debt investors minus the tax shield realized from issuing debt; Free Cash Flow = EBITDA WC Change Capital Expenditures Operating Tax + Deferred Tax Change And, Free Cash Flow = Debt Cash Flow + Equity Cash Flow - Interest Tax Shield (5) The reason the interest tax shield is deducted rather than added is that taxes in the free cash flow formula are higher the actual taxes by the interest tax shield. With the tax shield, the true cash flow received by equity and debt investors is less than the free cash flow. Because free cash flow consists of debt and equity cash flows, and the enterprise value is defined as the sum of equity and debt value, the value of the overall free cash flow (the sum of equity and debt cash flow) can be computed to derive the enterprise value. This calculation works as long as the discount rate applied to equity cash flow and the debt cash flow is used in valuing the free cash flow; Enterprise (Asset) Value = Present Value (Debt Cash Flow, Discounted at Risk Free Rate plus Credit Spread) + Present Value (Equity Cash Flow, Discounted at Cost of Equity) Enterprise (Asset) Value = Present Value (Free Cash Flow, discounted at WACC) (7) In order for the above two formulas to work, the WACC must be weighted using market values and the cost of debt and the cost of equity must be incremental costs. When summing the debt and equity values to derive free cash flow, relative proportion of equity cash flow and debt cash flow that influences drives the total free cash flow and the enterprise value depends on the future values of the equity and debt, not the historic amounts of debt and equity invested. For example, if the equity value is very high (say the equity value is for Apple and the company has a very market value because of the iphone and the ipod which did not require much investment relative to the market value.) In this case, the value of equity at the market value is much more than the value of debt and the weighting must account for this fact. This means that weighting percentages used in computing the average cost of debt and equity capital for application to free cash flow must be calculated from market values and not book values. The components of the WACC include: Debt Discount Rate = Current Risk Free plus Credit Spread from Future Risk Equity Discount Rate = Required Rate that Reflects Risk of Future Equity Cash Flows Debt Percent in WACC = Market Value of Debt/Enterprise Value

(6)

Chapter 2 Financial Modelling

Page 45

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Equity Percent in WACC = Market Value of Equity/Enterprise Value

(8)

The effect of the tax deduction of interest expense on debt is accounted for in the WACC calculation rather than in the cash flow in classic application of the discounted cash flow. The WACC is lower from the WACC which increases the enterprise value. The lower WACC is the reason operating and not actual taxes are used in the free cash flow calculation;

Percent Debt Equity WACC Market Market

Cost Incremental Incremental

Tax 1-Marginal Tax Rate

Weighted Mkt Wt x Incremental Cost x (1-Tax Rate) Mkt Wt x Incremental Cost Sum

(9)

In the weighted average cost of capital calculation, the equity return increases in proportion to the amount of debt in the capital structure, keeping the overall cost of capital constant. This is often applied through computing un-leveraged betas and then re-leveraging the betas as illustrated in the equations below:

Asset Beta = Equity Beta x Market Equity/Market Capital + Debt Beta x Market Debt/Market Capital

If the debt beta is zero, Asset Beta = Equity Beta x Market Equity/Market Capital

In terms of the equity beta: Equity Beta = Asset Beta x Market Capital/Market Equity

(10)

Since the value of the enterprise is the value of debt plus the value of equity, the value of equity is the enterprise value less the value of the debt. Equity Value (Equity Cash Flow, discounted at equity Cost) = Enterprise Value Net Debt Value (Debt Cash Flow, discounted at debt cost)

Chapter 2 Financial Modelling

Page 46

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Equity Cash Flow, Debt Cash Flow, Free Cash Flow and Cost of Equity, Cost of Debt and WACC
Equity Cash Flow Dividends less Equity Issued
Value of Equity PV of Cash Flows at Cost of Equity

+
Debt Cash Flow Net Interest plus Net Debt Payments

+
Value of Debt PV of Cash Flows at Incremental Cost of Debt

= Free Cash Flow EBITDA Op Taxes Cap Exp WC Chg

=
Value of Enterprise PV of Cash Flows At WACC

Exercises on the accompanying CD contain proofs of these principals. The diagram below illustrates how since free cash flow is the sum of equity and debt cash flows and since enterprise value is the sum of debt and equity value, that the enterprise value is the present value of free cash flows.

(11)

The classic weighting in the WACC calculation that is universally taught and used in investment banking analysis is technically incorrect because of the simple mathematical fact taught to school children that: A/B + C/D <> (A+C)/[B x (A/(A+C))+D x (C/(A+C))] In the case of the weighted average cost of capital, one can think of the cash flow on debt as A and the cash flow on equity as C. The if one plus the discount rate for debt is B and one plus the discount rate for equity is D, then A/B is the value of debt and C/D is the value of equity while A+C is free cash flow and the rest of the term is the WACC. In terms of a one period equation:

Debt Cash/(1+Debt Cost) + Equity Cash/(1+Equity Cost) <> Free Cash/(1+WACC), Where: WACC = (Debt Cash/Free Cash) x Debt Cost + Equity Cash/Free Cash x Equity Cost

To see this, consider a simple model where cash flows are received in a lump sum in ten years. The example shows that the incorrect weighting with a large difference between the equity cost and the debt cost causes a large error in the computation of WACC:

Chapter 2 Financial Modelling

Page 47

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Effect of Incorrect Weighting on Value
Debt Equity Total FCF Future Value Cost of Capital 100 5% 200 30% 300 75.90 118.02 14.73% 75.90 50.67% Value 61.39 14.51 Percent 81% 19% WACC 4% 6% 9.78%

Actual Value of Debt Plus Equity Valuation from WACC: FCF/(1+WACC) Future Cash Flow/Present Value - Corrected WACC Value at Corrected WACC Percent Error in WACC

In the above example, there is a large difference between the equity and the debt cost and the cash flows are assumed to occur ten years into the future. The value of the equity and the value of the debt is the cash flow discounted at the respective rates. The value computed using the traditional WACC of 118.02 is much higher than the value of the debt and equity because of the weighting problem. The corrected WACC can be computed through dividing total future value of cash flows by the present value of cash flows and compounding over the number of years evaluated.

WACC = (Future Value of Debt + Equity Cash)/(Present Value of Debt + Equity)

The remarkable thing about this relatively minor point is that it is completely ignored in finance texts and in teaching WACC in business schools. The methodology is explained in more detail in Exercise 3 of the corporate finance exercises.

(12)

Discount rates applied to equity cash flow, debt cash flow and free cash flow depend on the prospective uncertainty associated with cash flows; To make a decision using internal rate of return, one should compare the earned return to a cost of capital percentage that reflects risks of the cash flows. Using APV, the rate of return on free cash flow is assessed through measuring opportunity cost with a discount rate that is not affected by debt. The discount rate that measures risk if no debt existed is known as the all-equity cost of capital. This is a similar concept to the weighted average cost of capital (WACC), but unlike the WACC which often changes when the debt leverage changes, the all equity cost of capital is entirely independent of from the leverage. Using the all-equity discount rate, the value of an asset does not depend on the financing of that asset because both the free cash flow and the discount rate are independent of the amount of debt in the capital structure. The cost of capital applied to future cash flows should account for contractual aspects of the investment that affect the future volatility in cash flow; The discount rate applied to free cash flow of a particular project is not the same as the discount rate applied to different investments made by the same company;

(13)

(14)

(15)

Chapter 2 Financial Modelling

Page 48

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital

In contrast to equity cash flow, the free cash flow cannot be plucked from items that are shown on the financial statements i.e. dividend distributions and equity issues. Debt financing and income taxes complicate measurement because free cash flow is computed on a hypothetical basis as if no debt were present to finance the asset if the investment is completely funded by equity then there is no tax benefit. The presence of debt increases the variation in equity cash flow for a given amount of variation in revenues and expenses. This implies that if the discount rate applied to equity cash flow should change as the amount of debt financing changes. To consider the adjustments for debt in the computation of free cash flow, first think about the case where debt exists but there are no income taxes. In this situation, three adjustments are necessary to obtain free cash flow from equity cash flow: (1) adding debt financing and interest during construction for the construction phase, (2) adding interest expense over the operating life of the project, and (3) adding debt repayments during the operational period. During the construction period, free cash flow can be computed using either of the following two formulas (where interest during construction is the amount of interest paid to lenders, but recorded as a capital expenditure rather than an expense). The first formula begins with analysis of capital expenditures: Free Cash Flow = -(Capital Expenditures - Interest During Construction) In the operating period (after the commercial operation date) free cash flows can be calculated from the cash flow statement through adding back debt service interest expense and principal payments -- to dividend payments shown on the financial statements. Alternatively, free cash flow can be computed from the income statement and the operating portion of the cash flow statement as the amount of revenues less expenses less working capital change 6: Free Cash Flow = Earnings before Depreciation, Interest and Taxes less Working Capital Change Computing the free cash flow from equity cash flow when income taxes exist and debt is present is somewhat more complex because interest expense is deductible for tax purposes and because interest during construction affects tax depreciation deductions. The value of the interest deductions must be included in the value. This is accomplished through lowering the WACC rather than increasing the cash flow when applying the standard DCF model. The reduction of the WACC is computed through using an after tax discount rate for the debt component rather than a pre-tax interest rate (i.e. interest rate x (1 tax rate).) If interest expense is added back to equity cash flow in computing free cash flow, the resulting number will be higher than the cash flow would have been had there been no debt at all because taxes are lower with the presence of debt. Therefore, when adding interest expense back to equity cash flow, an adjustment is required to account for the fact that income taxes are lower than they would have been without the existence of interest deductions from debt. Making this interest tax shield adjustment -computing the tax effect of interest expense -- can be accomplished through subtracting an interest tax shield from free cash flow as shown in the formula below (an additional adjustment for the tax shield created from depreciation on interest during construction is required in project financing transactions): Free Cash Flow = Cash Flow before Financing + Interest Expense Interest Expense x tax rate Depreciation Expense on Interest during Construction x tax rate Free cash flow can also be computed from earnings before interest taxes and depreciation. This formula that directly begins with operating income is more intuitive but also involves more adjustments.

The working capital change should not include interest payable and dividend payable items that are related to financing of the plant rather than operations of the plant.

Chapter 2 Financial Modelling

Page 49

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Free Cash Flow = EBITDA EBIT x Tax Rate Capital Expenditure Working Capital Change + Depreciation Expense on Interest during Construction x Tax Rate The table below illustrates computation of free cash flow in a financial model. In practice the indirect method is often a simpler calculation because of items such as minority interest, deferred taxes, non-cash income, accounting adjustments and so forth that have to be included in the computation of free cash flow that begins with EBITDA. The table demonstrates that after making adjustments, both the indirect method and the direct method of computing free cash flow should yield the same result:

Internal Rate of Return and Net Present Value

Decisions associated with investments should be gauged by considering whether marginal benefits of the investment exceed marginal opportunity costs. Marginal benefits of making an investment are the free cash flows that are generated when the plant operates. Marginal opportunity costs include the initial investment and the rate of return that could have been earned by investors if another investment with a comparable risk profile would have been made. This basic economic principle of assessing decisions with opportunity cost means that once after-tax cash flow is established, the value of an investment should be assessed by computing the net present value of the cash flows measured at an after-tax discount rate that reflects risk of the project. All else equal, with riskier the cash flows, the value of an asset is lower because the opportunity cost from making similar high risk investments is more. If we are investing in a risk investment and we were making another investment with comparable high risk, we would be able to achieve a higher return. Risk and value of an investment can be assessed using with a risk-adjusted discount rate using two methods: net present value or internal rate of return. One way to assess the marginal benefits relative to the marginal costs of an investment is to compute the present value of free cash flows at a discount rate that reflects risks that would be present if only equity was used in financing the plant i.e., it should not reflect the risks that are created by adding debt to a project. A second approach to compute the net economic value of a project is to compare the internal rate of return generated from the free cash flow with the cost of capital. Some finance texts distinguish between net present value and internal rate of return and suggest that a net present value criterion is superior to internal rate of return. 7 This is a distinction with little practical import in most real world financial analyses. Since the internal rate of return can be defined as the rate of return that causes the net present value of a project to be zero, comparison of the internal rate of return to the cost of capital is an effective and intuitive way to assess value. The internal rate of return tells managers what the level of the rate of return on their investment. Further, the
7

See Brealey and Meyers, Principles of Corporate Finance 4 Edition, McGraw Hill Inc., New York, 1991.

th

Chapter 2 Financial Modelling

Page 50

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
internal rate of return is very easy to compute using spreadsheet programs and it can be determined for different time frames (for example, the IRR can be computed over a period of ten years rather than over the entire life of the project).8

Computing IRR with IRR, XIRR and MIRR

The IRR assumes that one can re-invest at the IRR rate. For example, if you receive 100 in year 2, the IRR assumes that the amount can be re-invested and earn the same return. In fact, the 100 may be reinvested at the overall cost of capital. The MIRR solves this problem. If the weighted cost of capital is used as the re-investment rate in the MIRR, and if the project earns its cost of capital, then the IRR is the same as the IRR. However, the variation in the IRR over different scenarios is less. The internal rate of return can be computed with IRR, XIRR or MIRR functions. The leading and trailing zeros are different MIRR, IRR and XIRR. When using IRR, there is no problem with leading or trailing zeros. For example, if the cash flow pattern is the following: 0 0 -100 50 50 50 0 0 0 Then the IRR will correctly compute the internal rate of return of 23.38%. However, if the XIRR is used, the first number must be negative.

Verification of XIRR
Step 1: Convert the annual rate into a daily rate using the formula (1+Annual Rate)^(1/365)-1. Step 2: Then compound using accumulated days, beginning with accumulated days of zero in first period (1+daily rate)^(accumulated days) Step 3: Use the SUMPRODUCT command to compute the present value

Sharing Cash Flow and Hurdle Rates

Examples of management incentives using alternative approaches. Project finance uses revenue sharing in infrastructure transactions, flipping of the taxes in renewable energy transactions, management incentives in leveraged buyouts. Use a function because of complexity.

Appendix 2: Terminal Valuation in Corporate Models

This appendix discusses issues that arise when computing terminal value in using a corporate model. The necessity of this assumption comes from the notion that companies are assumed to last indefinitely. If companies have stable cash flow and earnings, the terminal value is much easier to calculated and the

The difference between project IRR the internal rate of return earned on free cash flow -- and the all-equity cost of capital can be used to measure economic profit. Through multiplying the IRR difference by the cost of the project a measure of the economic profit earned by the investment can be obtained. Because electricity generating projects require a lot of up front capital, small differences between the earned return and the cost of capital can lead to significant economic profit. This is very similar to the traditional positive net present value rule whereby all projects with positive net present value should be selected. If economic profit is positive, the investment is generating positive value. Given that the components of free cash flow have been established in earlier chapters, most of the remaining discussion addresses how to compute the all-equity cost of capital for an electricity generating plant.

Chapter 2 Financial Modelling

Page 51

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
valuation should be higher. If the valuation of a company depends on an aggressive terminal value assumption which assumes a high growth rate, the company obviously has more risk.

Analysts, consultants, investment bankers and others who perform discounted cash flow analysis using corporate models generally make a surprisingly pessimistic assumption that growth in cash flow once a terminal period occurs will be limited to the projected rate of inflation. The valuations depend on the assumption that companies will somehow stabilize to a tranquil zero real growth rate in a period of somewhere between five and ten years, perhaps after a smooth transition period until this supposed harmony is obtained. While the assumption is commonly made, it is difficult to come up with any company -- or person for that matter -- that has reached this kind of tranquil nirvana or has managed such a transition to equilibrium. Further, the valuation analysts do not seem to be concerned about the basic point that if all companies somehow reach this kind of equilibrium where there is no real growth in cash flow, no companies would contribute to real economic growth and the world economy would stagnate in a never-ending recession. An illustration of the type of growth rate assumptions made in classic DCF analyses is shown in the graph below. 9 Of course, the date at which the transition from short-term to long-term growth occurs and the length of the transition period is arbitrary.
The long run tranquillity and equilibrium Long run assumptions: ROIC = Cost of capital Real growth = 0%

The short term Based on best estimate of likely outcome

The medium term transition to tranquillity Assessment of industry outlook and company position ROIC fades towards the cost of capital Growth fades towards GDP

Much of valuation involves implicitly or explicitly making growth estimates High P/E comes from high growth

The whole idea of assuming that companies reach equilibrium growth (often assumed to be zero in real terms) is questionable, much less imagining that one knows in what year the stable growth will occur and how long a fade period will last. While the general modeling of terminal growth rates is very subjective, at least when one structures a model with the base period, the explicit cash flow period and a fade period, the effect of changes in growth rates can be assessed. Given the difficulty in determining how long growth will last, the suggestion by some that the length of the explicit period does not matter is a bit 10 absurd for most practical modeling problems. The assumption of zero real growth from the final forecast period can of course be modified and improved upon. As stated above, classic application of the DCF method in corporate models involves applying a stable growth rate -- generally the rate of inflation -- to the final year free cash flows. This implies the value of the company in the terminal period is expressed by the formula: Terminal Value = Free Cash Flow in Terminal Period x (1+g)/(WACC g)

Growth rate in the transition period can be constructed from the equation: g t=gt-1 x[1/(short-term growth/long-term growth)]^(1/transition period). 10 Find quote in McKinsey

Chapter 2 Financial Modelling

Page 52

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
The primary items of free cash flow EBITDA, operating taxes, working capital and capital expenditures can all be handled in a more careful manner than this standard assumption. A summary of how assumptions can be derived for the period after the explicit cash flow period period t+1 -- is listed in the table below:
FCF Component Classic DCF Problems Improved Approach

EBITDA

Working Capital

Growing the Final Year Ignores the Implicit Return on Investment in the EBITDA and the Compute the Theoretical EV/EBITDA Driven by Grow Final Year Cash Flow at Low Relationship Between Cap Exp Growth and ROIC, Capital Expenditures/Depreciation, Tax Growth Rate EBITDA Growth Rate and Depreciation Rate If Revenue and Expense Growth Rates are Changing, the Growth Rate in Working Capital will not be Correct by Simply Applying the Model and Additional Period After the Last Terminal Growth to the Final Year Working Explicit Year, Where the Revenues and Grow at Final Year Capital Expenses Grow at the Terminal Growth There is a Theoretical Relationship between Compute the Equilibrium Capital Expenditures Capital Expenditures and Depreciation that to Depreciation by Making a Simple Model that Depends on Growth and Depreciation Rates; Extends for Many Periods. This Model is a this Growth Rate Does Not Equal 100% Unless Function of the Plant Life and the Growth in there is No Growth Capital Expenditures Compute Equilibrium Deferred Taxes as Function of Tax and Book Depreciation Rates and Income Tax Rate using a Similar Method Deferred Taxes Reach and Equilibirum Level as Above, with a Long-term Model that that Should be Consistent with Cap Exp Growth Reaches Equilibrium

Capital Expenditures

Either Use the Final Year Capital Expenditures or Set Capital Expenditures Equal to Depreciation

Taxes

Apply EBIT x Tax and Ignore Deferred Taxes

Stable Changes in Working Capital

Working capital movements must be consistent with the long-term growth in revenues and expenses established in period t+1, rather than in the terminal period which may have higher growth rates. A simple technique for taking care with respect to the terminal value is to add an additional period in the model after the explicit cash flow period period t+1. Using this technique, assumptions with respect to the on-going capital expenditure, working capital, and deferred tax assumptions in perpetuity can be made that are all consistent with the long-term growth rates.

Error from Stable Working Capital Assumption

50% 40% 30% 20% 10% 5% 70% -24.49% -18.82% -13.70% -9.06% -4.84% -2.87% 50% -20.41% -16.08% -12.10% -8.42% -5.02% -3.42% Explicit Growth Rate 40% 30% -18.50% -16.88% -14.88% -14.00% -11.51% -11.29% -8.37% -8.73% -5.44% -6.32% -4.04% -5.16% 20% -15.92% -13.84% -11.86% -9.96% -8.14% -7.27% 10% -16.54% -15.40% -14.30% -13.22% -12.17% -11.66% 5% -18.19% -17.65% -17.13% -16.61% -16.10% -15.85%

Capital Expenditures and Deferred Taxes

Assumptions with respect to capital expenditures should be consistent with growth rates in the EV/EBITDA multiple, and economic lives of equipment. One of the most common conceptual mistakes made in models is to assume that EBITDA can continue from relatively small capital expenditures made in the past few years. In fact, the capital expenditures should include provision for replacement of all capital which should be part of the depreciation expenditures. This means that in the terminal year, one could relate the capital expenditures to the depreciation expense. There are a few problems with this method however. First, if inflation and growth has occurred in the past, the depreciation expense can dramatically understate the required replacement capital expenditures. Second, the timing of retirements of existing capital can affect the future requirements. To compute the ratio of capital expenditures to depreciation, one must first develop an assumption with respect to the growth of invested capital in plant and equipment. The ratio of capital expenditure to

Chapter 2 Financial Modelling

Working Capital as Percent of EBITDA

Page 53

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
depreciation in perpetuity can computed once the growth rate in capital expenditures is input and retirements are computed from the depreciation life. Unfortunately, the calculations require some kind of assumption with respect to retirements of existing assets. (A simple assumption from depreciation rates can be made if a priori information is known.) The stable ratio of capital expenditures to depreciation is computed from the present value of depreciation and the present value of capital expenditures. Application of these techniques depends on setting up the model so as to be able to flexibly manage terminal periods, stable growth and reasonable terminal valuation assumptions. To compute the stable ratio of capital expenditures to depreciation is not 100% as often assumed in models because when assets grow, the capital expenditure occurs before the depreciation. To compute the stable level of capital expenditures to depreciation, a simple model that works through the balance of investment including additions driven by growth and retirements that are a function of the asset life. The model can be extended indefinitely, but the ratio will stabilize at the end of the investment life (the model is included as an exercise on the CD.) The example below show a case where the ratio of capital expenditures to depreciation stabilizes at 122% rather than 100%.
Model - Book Depreciaion Investment Balance Opening Balance Add: Capital Expenditures Less: Retriements of New Assets Less: Retirements of Existing Assets Closing Balance Depreciation Life Depreciation Rate Depreciation Expense Cap Exp to Depreciation 4,000.00 306.00 0.00 50.00 4,256.00 20.00 5.00% 200.00 1.53 4,256.00 312.12 0.00 100.00 4,468.12 20.00 5.00% 212.80 1.47 4,468.12 318.36 0.00 150.00 4,636.48 20.00 5.00% 223.41 1.43 4,492.18 395.84 0.00 0.00 4,888.03 20.00 5.00% 224.61 1.76 4,888.03 403.76 0.00 0.00 5,291.79 20.00 5.00% 244.40 1.65 5,291.79 411.84 0.00 0.00 5,703.62 20.00 5.00% 264.59 1.56 5,703.62 420.07 0.00 0.00 6,123.69 20.00 5.00% 285.18 1.47 6,123.69 428.47 0.00 0.00 6,552.17 20.00 5.00% 306.18 1.40 6,552.17 437.04 0.00 0.00 6,989.21 20.00 5.00% 327.61 1.33 6,989.21 445.78 0.00 0.00 7,435.00 20.00 5.00% 349.46 1.28 7,435.00 454.70 306.00 0.00 7,583.70 20.00 5.00% 371.75 1.22 7,583.70 463.79 312.12 0.00 7,735.37 20.00 5.00% 379.18 1.22 7,735.37 473.07 318.36 0.00 7,890.08 20.00 5.00% 386.77 1.22

300.00

4,000.00

The stable ratio of capital expenditures to depreciation depends on the nominal growth rate in capital expenditures and the lifetime of the asset. With higher growth rates and the longer lived investments, the ratio of stable capital expenditures to depreciation is greater. The table below demonstrates how the ratio of capital expenditures to depreciation increases with both depreciation life and with the growth rate.

Ratio of Stable Capital Expenditures to Depreciation

0% 100% 100% 100% 100% 100% 100% 100% 1% 102% 103% 106% 108% 111% 116% 128% 2% 104% 106% 111% 117% 122% 134% 159% Growth Rate 3% 106% 109% 117% 126% 134% 153% 194% 4% 108% 112% 123% 135% 147% 173% 233% 5% 110% 115% 130% 145% 160% 195% 274% 6% 112% 119% 136% 154% 174% 218% 317%

3 5 10 15 20 30 50

Stable ratios for computing deferred taxes can be evaluated in a similar manner as the rate of capital expenditures to depreciation. Analysts tend to be a little intimidated by deferred taxes which sometimes results in large mechanical errors. With a bit of patience one can work through the deferred taxes and develop reasonable estimates for valuation. From an accounting standpoint, deferred tax represents a liability would be subtracted from enterprise value as the tax depreciation will be less than the book depreciation in the future. The accumulated deferred tax on the balance sheet contains the nominal value of the liability that would be repaid if there were no more capital expenditures or if the tax life equaled the book life. If the company keeps growing, the new capital expenditures will continue to generate new deferred taxes and a stable rate of new deferred taxes to can be computed in a similar manner as the ratio of capital expenditures to depreciation discussed above. As with book depreciation, an investment balance is set up for the deferred taxes instead of the book taxes. The difference between the tax deprecation and the book depreciation multiplied by the tax rate yields the change in deferred tax that should be added to free cash flow.

Chapter 2 Financial Modelling

Investment Life

Page 54

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Ratio of Stable Deferred Tax to Capital Expenditures Assuming 40% Tax Rate and 20 Yr Life
0% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 1% 3.51% 3.12% 2.74% 2.36% 1.79% 0.88% 0.00% 2% 6.51% 5.75% 5.00% 4.28% 3.23% 1.56% 0.00% Growth Rate 3% 9.08% 7.96% 6.88% 5.85% 4.37% 2.08% 0.00% 4% 11.28% 9.82% 8.43% 7.12% 5.26% 2.47% 0.00% 5% 13.17% 11.39% 9.71% 8.14% 5.96% 2.75% 0.00% 6% 14.80% 12.70% 10.76% 8.96% 6.50% 2.96% 0.00%

1 3 5 7 10 15 20

The table implies that accumulated deferred taxes should be ignored in the valuation except to establish the base of future tax depreciation. The only case in which the accumulated deferred tax measures value is under the extreme case where there are no capital expenditures and where the weighted average cost of capital is zero. If the discount rate is positive, then the accumulated deferred tax underestimates the value of the future tax deductions. If the capital expenditures are positive and growing, then deferred taxes on the future capital expenditures offsets the liability from existing deferred taxes. For the analysis to be accurate, and estimate should be made of the exiting tax depreciation should be made. This can be established from the exiting balance of deferred tax and is illustrated in the case exercise.

ROIC, EBITDA and Multiples

An alternative way to compute terminal value using in the DCF model is to apply a multiple to the terminal EBITDA rather than to use the standard terminal growth rate formula known as Gordons model. The terminal growth rate formula applies an on-going growth forever using the formula (this formula is mathematical and can be proved with a simple spreadsheet; there is nothing to interpret): Terminal Valuet = Free Cash Flowt+1/(k-g) or Free Cash Flowt x (1+g)/(k-g) Once the terminal value is established, it is discounted to the current value at the discount rate to the value at the present time. An alternative approach is to compute the terminal value through multiplying the EBITDA in the terminal year by an EV/EBITDA multiple. If this approach is applied, the terminal value is no longer directly sensitive to the discount rate and the growth rate as illustrated below: Terminal Valuet = EV/EBITDA x EBITDAt The problem with this formula is where to terGeneral discussion of multiples. Tying multiples to growth and return. Long-term growth rate not driven by re-investment rate; work through algebra. Table with alternative EV/EBITDA multiples from different relationships with growth and return above cost of capital. Difference between using EV/EBITDA and normal growth rate formula. Also could multiply by investment. Underlying value drivers. Getting around the problem with consistency. Instead of simply assuming that the EBITDA can grow from an arbitrary number, the investment in period t+1 is multiplied by an EV/EBITDA ratio which accounts for the return on investment, the cost of capital and the growth rate. This can be accomplished through simulating the theoretical multiple from assumptions with respect to cost of capital, rate of return, capital expenditure to depreciation relationships and tax rate. It is better to be able to perform sensitivity analysis with respect to different terminal periods, different fade periods and different assumptions with respect to the ability of a company to realize profits in excess of their cost of capital over long periods. To incorporate the ability to perform such sensitivity analysis, it is a good idea to structure a corporate model so as to be able to be able to evaluate the effects of terminal period length, transition period length, terminal growth rates and other factors in the valuation analysis. This can be accomplished through using (TRUE/FALSE) switches to define the explicit period,

Chapter 2 Financial Modelling

Tax Life

Page 55

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
the first year of stable growth after the terminal period and the period in which the terminal value is computed.

Appendix 3: Real Estate Modelling Issues

Real Estate models are closely related to project finance models with a couple of exceptions that seem to make the models more complex. This appendix assumes that you have reviewed the sections that address project finance models related to setting-up project phases and time periods. Some features of a real estate model that are different from a project finance model include: In modelling a single project such as a multi-family development or a hotel, the sale of a project after a given holding period is often modeled and the features of debt are different from typical project finance models. In modelling some projects, cash flow can be generated before construction on the project is completed which adds complexity to the sources and uses analysis. In modelling real estate projects, there are often multiple different units and buildings which are completed at different times meaning that there is not a single construction phase. In modelling commercial real estate investments, the analysis should include the effects of different lease terms Some real estate projects have multiple participating tranches with different return targets and sharing mechanisms

Modelling a Single Project An effect way to understand a real estate project is first to model a single project and then later to put projects together in a portfolio described below in the section on mixed development. Modelling a single project is very similar to modeling project finance with the exception that a holding period is defined after which the project is sold. Operating inputs that drive the value of the project include the occupancy rate that varies with the life of the project, the rental rates which can be volatile, occupancy rates that are correlated with the rental rates and fixed and variable operating expenses. In selling the project a time period switch can be established for the terminal period. To define the sale price, a capitalization rate is generally used. The capitalization rate is the value of similar projects that are bought and sold in a market divided by the pre-tax cash flows. For example if a project is sold for 10,000 and its rental revenues net of operating expenses are 500, the capitalization rate is 500 divided by 10,000 or 5%. Rather than using the total project life to measure the operating period as in other project finance models, the operating period is generally defined by the holding period before which the project is sold. Then, similar to the acquisition model described above, the project is sold after the holding period. This can be accomplished through dividing the cash flow by the cap rate which is a percentage and also by a terminal value switch (the terminal value switch is simply a TRUE/FALSE logical variable that is defined with the formula = (period = holding period). The theory behind the cap rate is very similar to the ideas underlying computing the terminal value using the final year cash flow divided by the cost of capital minus the growth rate. The cap rate reflects both the cost of capital and growth rate. For example if one believes the growth rate in rents is 2.5% in a market and the cap rate is 5%, then the implied cost of capital is 7.5% (cap rate = (WACC g) which means that WACC = cap rate + g). With a higher growth rate the cap rate should be lower and if the cash flows are less risky (for example, if there are fixed leases) then the cap rate is also lower.

Chapter 2 Financial Modelling

Page 56

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Once the proceeds from selling an asset are computed (net of fees and other transaction costs) a couple of accounting and tax issues arise. For tax purposes a capital gain is generally computed as the sales proceeds less the net book value of the fixed assets. The capital gain is the sales proceeds less the Once the rate of return in a real estate model is computed the equity and project return Real estate debt can take many forms as with project finance debt, but when the project is sold at the end of the holding period, the debt must be retired. In addition, when the project is sold, the fees and taxes on the gain on sale should be accounted for.

Cash Flow Generated Before Completion of Construction In some real estate projects, projects are held for trading rather than investments. In these cases, cash flow is generated from receiving progress payments before construction is complete. This cash generated from selling a project reduces the need for additional debt. Further if the proceeds are more than the total amount of debt required, the cash is deposited into a reserve account. If there is money in this cash reserve account and future financing needs occur, then the reserve account is used for future cash needs. This cash process can be modeled in a similar way as the cash process that was described in a standard corporate model where deficit cash flow is funded by raising new debt and surplus cash goes to retiring cash and/or reducing debt. Modelling a portfolio of different projects The key is to use the time period phases and the period codes that were introduced in discussion of project finance models. Assume there are a few projects (e.g. residential developments) and that various units begin and end construction at different periods. Use the occupancy date relative to the start date of the overall model to determine the length of the construction period. For example if the start date of the overall model is 2010 and the start date of project is 2015, then the construction period is 5 years. This contrasts with project finance models that simply compute the construction period from the start of the construction period. The first period is 1 minus the total construction period -- this will make first operating period 1. If the occupancy dates are different for different projects, then the codes will be different for different project.

With the period code, use and S-curve table that defines construction expenditures as a function of the period code The S-Curve can be computed using a hlook-up table.

Simply add the cash flows for different projects using a date code that begins with the overall model start date that was used to compute the period code for each project. Consider the very simple inputs in the figure below.

Chapter 2 Financial Modelling

Page 57

Valuation Mirage: Practical Application of Modeling, Risk Assessment, Economic Driver Analysis, Debt Capacity and Cost of Capital
Assumptions Code Residential Type 1 Residential Type 2 Commecial Period Profile 1 Profile 2 Profile 3 1 2 3 -5 0% 0% 30% Cost 1,000.00 2,000.00 3,000.00 -4 0% 0% 25% Sales Price 1,500.00 2,500.00 3,200.00 -3 0% 0% 20% Constr Profile Period Finished 3 1-Jan-12 2 1-Jan-15 1 1-Jan-15 -1 50% 50% 10% 0 25% 50% 5% Total 100% 100% 100%

1 2 3

-2 25% 0% 10%

First Project Number Residential Type 1 Commercial Date Periods Prior to Operation First Period Profile Index+1 Cost Price Period Code Construction Profile Construction Profile Cost Construction Cash Proceeds from Sales Price Revenue Recognized - Pct of Completion

1-Jan-12 5 -4 3 1,000.00 1,500.00 -4 0.25 0.25 1,000.00 250.00 375.00 -3 0.2 0.2 1,000.00 200.00 300.00 -2 0.1 0.1 1,000.00 100.00 150.00 -1 0.1 0.1 1,000.00 100.00 150.00 0 0.05 0.05 1,000.00 50.00 75.00 #N/A 0 1,000.00 1,500.00 1 #N/A 0 1,000.00 2 #N/A 0 1,000.00 3 #N/A 0 1,000.00 4 #N/A 0 1,000.00 5 #N/A 0 1,000.00 6 #N/A 0 1,000.00 7

Chapter 2 Financial Modelling

Page 58