Modern Portfolio Theory

Dr. Sung Won Kim

Modern Portfolio
Theory

Introduction

Dr. Sung Won Kim

Introduction to Modern Portfolio Theory
(MPT)
Introduced by Harry Markowitz in 1952
The investment theory is based on the risk aversion of individuals
The goal is to maximize expected returns based on a given level of
risk
The theory emphasizes risk as an inherent part of higher returns
MPT attempts to analyze the interrelationship between different
investments
Modern Portfolio
Theory

Expected Returns

Dr. Sung Won Kim

Expected Returns of one Security

The amount of profit/loss an investor can anticipate on an

investment given the investment’s historical returns
It is calculated as the historical average of an assets
returns.
However, it does not take risk into account.
INSERT Example of expect returns one security
Expected Returns of a Portfolio

Expected return for an investment portfolio is the weighted

average of the expected return of each of its components
Weight of a security is the ratio of the investment in the
security to the total investment in the portfolio
Expected Return Calculation

The formula for the expected return on a portfolio is

given by
𝐸 𝑟 = (𝑤& ∗ 𝑟& ) + (𝑤* ∗ 𝑟* )
where,
𝑤 is the weight of a security
𝑟 is the return of the security
Expected Returns of a Portfolio

Expected return = (weight of A)×(expected return of A)

+ (weight of B)×(expected return of B)
+ (weight of C)×(expected return of C)
Expected Return Example

Let us consider an investment in a two asset portfolio.

• Asset A – 12% return with 20% standard deviation
• Asset B – 16% return with 35% standard deviation

Given $10,000, let us invest $6000 in Asset A and $4000

in Asset B.
Expected Return Calculation

The expected return is calculated as follows:

6 4
𝐸 𝑟 = ∗ 12 + ∗ 16 = 7.2 + 6.4
10 10

𝑬 𝒓 = 𝟏𝟑. 𝟔%
Add example using Real
data
Modern Portfolio
Theory

Risk of a Security

Dr. Sung Won Kim

Risk of a Security

Risk represents the chance that the actual return of an

investment will be different than expected.
Risk is synonymous with volatility—the greater the
portfolio volatility, the greater the risk.
Volatility refers to the amount of risk or uncertainty related
to the size of changes in the value of a security.
The most common measures of volatility of a security’s
return are variance and standard deviation.
Risk of security
example

Find two graphs one not

volatile and one very volitle.
Volatility of a Portfolio

Volatility of a portfolio can be calculated by first

determining the following parameters:
The variance of expected return of a security
The covariance of a portfolio of securities
Observations indicate that the variance of a portfolio
decreases as the number of portfolio assets increases –
diversification.
Standard Deviation
The standard deviation of returns is calculated as the
square root of the sum of squares of the difference
between each return data point and the mean of returns
divided by the number of returns considered.

∑(r − E(r))G
Standard Deviation (𝜎) =
n

The variance of a portfolio is the square of the standard

deviation of the portfolio (i.e., 𝑉I = 𝜎IG).
Correlation and Covariance

The correlation indicates the strength of the relationship

between the returns of two assets.
∑(MN,O PMN )(MQ,O PMQ )
Correlation: 𝜌&,* =
∑(MN,O PMN )R ∑(MQ,O PMQ )R

Covariance is a measure of the joint variability of two

random variables.
Covariance: 𝐶𝑜𝑣(𝑎, 𝑏) = 𝜌&,* ∗ 𝜎& ∗ 𝜎*
Volatility Calculations Example
Let us consider an investment in a two asset portfolio
• Asset A – 12% return with 20% standard deviation
• Asset B – 16% return with 35% standard deviation

The correlation between the returns on the two assets is

0.6.
Given $10,000, let us invest $6000 in Asset A and $4000
in Asset B.
Volatility Calculations

The volatility of the portfolio is given by:

𝑉I = 𝑤&G ∗ 𝜎&G + 𝑤*G ∗ 𝜎*G + 2 ∗ 𝜌&,* ∗ 𝜎& ∗ 𝜎* ∗ 𝑤& ∗ 𝑤*

= (0.62)(0.22) + (0.42) ∗ (0.352) + 2(0.6)(0.2)(0.35)(0.6)(0.4)
= 0.0144 + 0.0196 + 0.02016 = 0.05416 => 𝝈𝟐𝑷 = 𝟎. 𝟎𝟓𝟒𝟏𝟔
Modern Portfolio
Theory

Efficient Frontier

Dr. Sung Won Kim

Risk-Return Trade-Off

Basic principle that the riskier the investment, the greater

the required potential return
If investors are willing to bear risk, then they expect to
earn a risk premium.
The risk-return trade-off refers to the possibility of higher
return on investments but does not guarantee higher
returns.
Risk-Return Trade-Off
Sharpe Ratio
Sharpe ratio is the measure of risk-adjusted return of a
financial portfolio:
𝐸(𝑟b ) − 𝑟c
Sharpe ratio =
𝜎b
𝐸(𝑟𝑝) - Expected return of a portfolio
𝑟𝑓 - Risk-free return
𝜎𝑝 - Standard deviation of the portfolio returns
The ratio is the average return earned in excess of the risk-free rate
per unit of volatility or total risk.
The greater a portfolio's Sharpe ratio, the better its risk-adjusted
performance.
Sharpe Ratio Calculation
Let us consider the example used to calculate the portfolio
parameters.
We have calculated the expected return (𝐸(𝑟) = 13.5%)
and variance of the portfolio ( 𝜎IG = 0.05416).
Let’s assume the risk free rate (𝑟𝑓) to be 0%.
The Sharpe ratio is calculated as:
0.135 − 0.000 0.135
Sharpe ratio = = = 𝟎. 𝟓𝟖
0.05416 0.233
Portfolio Weights

Portfolio weight is the percentage composition of a

particular holding in a portfolio.
The degree of risk reduction depends, among other
things, upon the weight of the assets in a portfolio.
In this class, we assume only positive weights in all assets
(i.e., long-only portfolios, or mutual funds).
Efficient Frontier
Represents the best combination of expected returns and
risk.
The efficient frontier is the lowest risk for a given expected
return or the highest return for a given level of risk.
Portfolios below the efficient frontier are not desirable.
Usually depicted in graphic form as a curve on a graph
comparing risk against the expected return of a portfolio
Efficient Frontier – Tech Stocks
Recall

The volatility of the portfolio is given by:

𝑉I = 𝑤&G ∗ 𝜎&G + 𝑤*G ∗ 𝜎*G + 2 ∗ 𝜌&,* ∗ 𝜎& ∗ 𝜎* ∗ 𝑤& ∗ 𝑤*

Shape of the Frontier

𝜌 = −1 𝜌 = −0.5

𝜌= 0
𝜌=1
Modern Portfolio
Theory

Minimum Variance Portfolio

Dr. Sung Won Kim

Minimum Variance Portfolio

This indicates a portfolio, comprised of risky assets, with

the least amount of variance (risk).
The minimum variance portfolio can be pointed out on the
efficient frontier (which we will see in our R code).
The optimal minimum variance portfolio will decrease
overall volatility with each investment added to it, even if
the individual investments are volatile in nature.
Capital Allocation Line (CAL)

In constructing portfolios, investors often combine risky

assets with risk-free assets to reduce risks.
A complete portfolio is defined as a combination of a risky
asset portfolio, with return 𝑅𝑝, and the risk-free asset, with
return 𝑅𝑓.
The optimal risky portfolio is found at the point where the
CAL is tangent to the efficient frontier.
CAL (cont’d)
The expected return of a complete portfolio is given as:
𝐸 𝑟h = 𝑟c + 𝑆I 𝜎h Equation for CAL

𝑟𝑐 is the expected return of the complete portfolio of the risk-free

asset and risky portfolio.
SP is the Sharpe ratio of the risky portfolio.
𝜎c is the standard deviation of the complete portfolio, which is
given by 𝜎𝑐 = 𝑤𝑃 ∗ 𝜎I + 𝑤c ∗ 𝜎c . 𝑤𝑃 is the allocation in the risky
portfolio and 𝑤c in the risk free asset.
Since 𝜎c = 0, we have 𝝈𝒄 = 𝒘𝑷 ∗ 𝝈𝑷
Tangency Portfolio

The "tangency portfolio" is the point on the efficient

frontier that has the highest Sharpe ratio.
The point of tangency between the capital allocation line
(CAL) and the efficient frontier
CAL makes up a combination of risk-free assets and risky
portfolio.
At the point of tangency, the investor fully invests in the
risky portfolio.
Efficient Frontier – Tech Stocks
Capital Allocation Line

Asset Weights

Tangency Portfolio

Minimum Variance Portfolio

Frontier Weights
Portfolio Optimal Weights – Tech Stocks
Further Analysis

Portfolios with provision for short positions can be

generated.
Other constraints such as risk targets, returns targets, or
alpha targets can be set using certain functions.
The benefits of diversification can be studied by varying
the number of stocks in Portfolio 3 from 5 to 10 or more.
Modern Portfolio
Theory

Diversification

Dr. Sung Won Kim

Diversification
Diversification is the process of mixing a variety of
investments in a portfolio with the intention to reduce
overall risk.
Diversification strives to smooth out unsystematic risk
events in a portfolio, so the positive performance of some
investments neutralizes the negative performance of
others.
Diversification limits portfolio risk but can also mitigate
performance, at least in the short term.
Example: Diversification Benefits

We will construct two portfolios:

Portfolio A: 5 random stocks
Portfolio B: 10 random stock
According to the MPT, we should observe a lower risk on
portfolio B.
Correlations of assets drive diversification.
Portfolio A vs. Portfolio B
Observation
We can see that the risk decreases significantly when the
number of stocks in the portfolio is increased.
Studies have shown that using 30 stocks in a portfolio
provides the most benefits of diversification.
From the formula for the variance of a portfolio, we can
see that the variance is directly related to the correlation
between the assets – the lower the correlation, the lower
the variance.
Recall

The volatility of the portfolio is given by:

𝑉I = 𝑤&G ∗ 𝜎&G + 𝑤*G ∗ 𝜎*G + 2 ∗ 𝜌&,* ∗ 𝜎& ∗ 𝜎* ∗ 𝑤& ∗ 𝑤*

Dataset Selection
Portfolio 1: five tech hardware stocks
• Apple Inc
• Intel Corp
• Cisco Systems
• Nvidia Corp
• Texas Instruments
Portfolio 2: five stocks from different sectors
• Microsoft Corp
• Johnson & Johnson
• JP Morgan Chase & Co
• Exxon Mobil Corp
• Home Depot
Portfolio 3: n number of random stocks across sectors
R Packages Used

fPortfolios – generates portfolios

Quantmod – obtains historical stock price data
ggplot2 – plots graphs
batchGetSymbols – downloads current components of
S&P500 index from wikipedia
timeSeries – converts numerical data to time series data
fPortfolio Package
The package is specially geared toward portfolio optimization
Key functions from the package used during the class are–
portfolioFrontier – generates portfolios (set at 100) for the
given stock returns data
getWeights – extracts the weights of each stock in all the
generated portfolios
minvariancePortfolio – returns the portfolio with the minimal
risk on the efficient frontier
tangencyPortfolio – returns the portfolio with the highest
return/risk ratio on the efficient frontier
Introduction to Algorithmic
Trading
Introduction

Dr. Sung Won Kim

What Is Algorithmic Trading

Algorithmic trading is a style that utilizes a computer’s

ability to process data quickly and react faster.
This approach is almost always based on hard data rather
than forecasts or opinions.
The main objective of algorithmic trading is not necessarily
to maximize profits but rather to control execution costs
and market risk
Algorithmic Trading Strategies

These are some of the major algorithmic trading

strategies:
Mean Reversion
Momentum Strategy
Statistical Arbitrage
Trend Following Strategy
Mean Reversion

Mean reversion strategies exploit the tendency of many

asset prices to revert to the mean after periods where they
are oversold or overbought.
Mean reversion strategies try to capitalize on extreme
changes in the price of a security, assuming that it will
revert to its previous state.
Relative Strength Index (RSI) is the most commonly used
indicator for mean reverting strategies.
Momentum Strategy

Momentum strategies are based on the assumption that

the price of financial instruments has some inertia.
Momentum trading is generally used to capture strong
moves in short time frames.
Momentum is the acceleration in a stock’s price that can
be due to earnings, sentiment, news, greed, or fear.
Momentum strategies are a particular case of trend-
following strategies.
Statistical Arbitrage

This strategy identifies pricing discrepancies between

securities through mathematical modelling techniques.
This strategy is market neutral as it involves a long
position and also a short position in different securities.
The end objective of such strategies is to generate alpha
(higher than normal profits) for the trading firms.
Trend Following Strategy

The goal of this strategy is to buy an asset when its

price trend goes up, and sell when its trend goes down,
expecting price movements to continue.
Trading decisions are made based on technical analysis,
market patterns, and indicators.
The most commonly used indicator is the Moving
Average, which we will study in this class.
Data – Advanced Auto Parts Stock
Introduction to Algorithmic
Trading
Trend Following Strategy

Dr. Sung Won Kim

Basic Trading Strategy
The strategy used in this study is called a trend-following
strategy.
It works on the basis of historical average price
movements of the asset.
We will compute what is known as a Simple Moving
Average for two time periods – the average price over the
n previous days.
If the smaller time period line crosses the longer time
period line for below – we have a buy, and a sell if it
crosses from above.
Signal Generator

The trade signal is generated as follows:

If the fast moving average crosses the slow moving
average from below – buy, else sell/hold.
The position once bought/sold will be held until there is
change is signal direction.
When the signal goes from sell to buy, the sold assets
are bought back and more assets are bought.
Trade Strategy Example

Black -> Prices

Green -> Slow MA (26 day)
Red -> Fast MA (12 day)

The fast moving MA crosses the

slow moving MA from above–this indicates a SELL.

The fast moving MA crosses the

slow moving MA from below–this indicates a BUY.
Introduction to Algorithmic
Trading
Backtesting

Dr. Sung Won Kim

Backtesting a Trading Strategy

Backtesting is a key component of developing an effective

trading strategy.
We reconstruct, using historical data, the trades that would
have occurred using the rules established in the trading
strategy.
The result we obtained offer statistics to gauge the
effectiveness of the strategy.
Backtesting the Trend-Following Strategy
To backtest the moving average strategy, we chose the
stock price data for Advanced Auto Parts.Inc (AAP) and
carry out the following steps:
We obtain the simple moving averages for 12 days and
26 days.
We convert the stock price data to returns.
We check the position of the 12-day MA (fast) with
respect to the 26-day MA (slow).
If the fast moving average is above the slow moving
average, it is a buy, and sell if it is below.
Backtesting the Trend-Following Strategy
We denote a long position as +1 and a short
position as -1.
Now, we have the actual returns based on historical
data and the position initiated by the strategy.
To obtain the performance of the algorithm, we multiply
the position initiated on date with the returns for that
day.
A positive return on a long day is profitable and so is a
short on a negative return day.
In any other scenario, we end up with a loss day.
Backtesting the Trend-Following Strategy

Executing trades as such, we will end with the

daily returns on each trade.
The average of the daily returns will give us the
expected daily return from this strategy.
We can study the volatility of the returns and the
Sharpe ratio to get a better understanding of the
performance of the strategy.
Results
For the given backtest data set of Advanced Auto Parts
Stock returns from 2010 to present, we have the following
parameters:
Expected Return – 13.5%
Standard Deviation – 1.86%
Sharpe Ratio – 6.2
Looking at the results above, we can see that the strategy
has performed considerably well with a very high-risk
adjusted return.
Results Table
Introduction to Algorithmic
Trading
Conclusion

Dr. Sung Won Kim

Conclusion

The basic trend-following strategy was

considerably profitable on an annualized basis.
While this was a simple strategy, we can also
employ more advanced predictive models to
predict the movement of the stock prices.
Coupling multiple strategies is also a good way to
minimize bad loss days.