1.

From Efficient Markets Theory to Behavioural Finance – Robert Shiller, Journal of Economic
Perspectives, 2003
a. The efficient markets theory reached its height of dominance in academic
circles around the 1970s. At that time, the rational expectations revolution in
economic theory was in its first blush of enthusiasm, a fresh new idea that occupied
the center of attention.
b. Robert Lucas published “Asset Prices in
an Exchange Economy” in 1978, which showed that in a rational expectations
general equilibrium, rational asset prices may have a forecastable element that is
related to the forecastability of consumption

Behavioural critique – investors do not process information rationally (leads to incorrect prob dist of
expected future returns) and even if they do, they might make incorrect decision due to behavioural
biases (anchoring, framing, loss aversion, regret aversion, mental accounting, self control). Alpert and
Raiffa 1982 show that people are poorly calibrated in estimating probabilities. All of this means that
security prices can atleast be partially oredicted.

EMH and its critiques – Malkiel, Journal of Economic Perspectives, 2003

Markets can be efficient even if many market participants are quite irrational.
Markets can be efficient even if stock prices exhibit greater volatility than can
apparently be explained by fundamentals such as earnings and dividends. Lo and MacKinlay (1999)
finds that short-run serial correlations are not zero and that the existence of “too many”
successive moves in the same direction enable them to reject the hypothesis that
stock prices behave as true random walks. There does seem to be some momentum
in short-run stock prices. Moreover, Lo, Mamaysky and Wang (2000) also find,
through the use of sophisticated nonparametric statistical techniques that can
recognize patterns. Such patterns are due to short run momentum generated by irrational
exuberance (Shiller, 2000)

This only proves that stock s are not a perfect RW, need to differentiate between statistical
significance and economic significance. Anyone who pays transactions costs is
unlikely to fashion a trading strategy based on the kinds of momentum found in
these studies that will beat a buy-and-hold strategy. Indeed, Odean (1999) suggests
that momentum investors do not realize excess returns. The so-called “January effect,” for example,
in which stock prices rose in early January, seems to have disappeared soon after it was
discovered

There is indeed considerable support for long-run negative serial correlation in stock returns.
However, the finding of mean reversion is not uniform across studies and is quite a bit weaker in
some periods than it is for other periods. Campbell and Shiller (1998) conclude that P/E ratios
explaines as much as 40% of future expected returns (but investors using that strategy couldn’t
profit). Fama and Schwert (1977) found that short-term interest rates were related to future stock
returns. Campbell (1987) found that term structure of interest rates spreads contained useful
information for forecasting stock returns. Keim and Stambaugh (1986) found that risk spreads
between high-yield corporate bonds and short rates had some predictive power.
Again, even if some predictability exists, it may reflect time varying risk premiums
and required rates of return for stock investors rather than an inefficiency. More-
over, it is far from clear that any of these results can be used to generate profitable
trading strategies. Whether such historical statistical relations give investors reliable
and useful guides to appropriate asset allocation is far from clear.

Schwert (2001) points out that the investment firm of Dimensional Fund
Advisors actually began a mutual fund that selected value stocks quantitatively
according to the Fama and French (1993) criteria. The abnormal return of such a
portfolio (adjusting for beta, the capital asset pricing model measure of risk) was a
negative 0.2 percent per month over the 1993–1998 period

Nothing is ever as clear in prospect as it is in retrospect Palm Pilot from its parent 3Com Corporation
during the height of the Internet boom in early 2000. Initially, only 5 percent of the
Palm Pilot shares were distributed to the public; the other 95 percent remained on
3Com’s balance sheet. As Palm Pilot began trading, enthusiasm for the shares was
so great that the 95 percent of its shares still owned by 3Com had a market value
considerably more than the entire market capitalization of 3Com, implying that all
the rest of its business had a negative value. Other illustrations involve ticker symbol
confusion. Rasches (2001) finds clear evidence of comovement of stocks with
similar ticker symbols; for example, the stock of MCI Corporation (ticker symbol
MCIC) moves in tandem with an unrelated closed-end bond investment fund Mass
Mutual Corporate Investors (ticker symbol MCI). In a charming article entitled “A
Rose.com by Any Other Name,” Cooper, Dimitrov and Rau (2001) found positive
stock price reactions during 1998 and 1999 on corporate name changes when “dot
com” was added to the corporate title. Finally, it has been argued that closed-end
funds sell at irrational discounts from their net asset values (for example, Shleifer,
2000)
 Thaler and Fama – The Question is ‘are prices correct’ and ‘can one beat the market’ ,
definition of a bubble (timeline issue)

2. Stefan 2009, UC Berkely, TESTING THE EFFICIENT MARKETS HYPOTHESIS:

A BEHAVIORAL APPROACH TO THE CURRENT ECONOMIC CRISIS

The Weak Form of the EMH states that prices incorporate only past information about
the asset. An implication of this form of the EMH is that one cannot detect mis-priced assets
and
consistently outperform the market through technical analysis of past prices.
The Semi-Strong Form of the EMH asserts that stock prices reflect all publicly available
information. This information includes past prices and returns as well as a companyís
financial
statements, accounting practices, earnings and dividend announcements, and competitorsí
financial situation.
The Strong Form of the EMH states that the current price of a stock incorporates all
existing information, both public and private. In this case, one should not expect to
systematically outperform the market even if trading on insider information. According to
this
form of the EMH, the market anticipates future developments and asset prices adjust to
incorporate this information.

EMH misinterpretations:

Although the EMH claims investors cannot outperform the market, analysts
such as
Warren Buffet have done exactly that. Hence the EMH must be incorrect. This
interpretation is incorrect because the EMH implies that investors cannot
consistently
outperform the market; there will be times when, out of luck, an asset will
outperform
the market. Additionally, it is possible for one to consistently outperform the
market by
5
chance. Suppose a fund manager has a 50% chance of beating the market next
year.
His probability of beating the market two years in a row is 25%; of beating the
market
eight years in a row, 4%. Consequently, out of 1000 fund managers, 4 will
consistently
outperform the market eight years in a row. Hence not only is it possible to
outperform
the market on occasion, it is possible to consistently outperform the market by
luck.
2. According to the weak form of the EMH, technical analysis is useless in predicting
future stock returns. Yet financial analysts are not driven out of the market, so their
services must be useful. Hence, the EMH must be incorrect. This statement is incorrect
because a financial analyst can put together a portfolio that matches the risk-tolerance
of each client, whereas a random collection of stocks will most likely not cater to
particular risk preferences. Secondly, financial analysis is essential to the efficiency of
markets because it allows investors to take advantage of new information to identify
mis-priced stocks. When there is competition among many investors, arbitrage
opportunities vanish, i.e. stock prices adjust immediately to incorporate any new
information, leading to market efficiency.
3. The EMH must be incorrect because stock prices are constantly fluctuating randomly.
The fact that stock prices are constantly changing is evidence in support of the EMH,
because new information appears almost continuously in the form of opinions, news
stories, announcements, expectations, and even lack of news. The constant arrival of
new information causes the continuous adjustment of prices, as the EMH claims.
4. If the EMH holds, then all investors must be able to collect, analyze and interpret new
information to correctly adjust stock prices. However, most investors are not trained
financial experts. Therefore, the EMH must be false. The EMH does not require that all
6
traders be informed. A relatively small core of experts is necessary to analyze new
information and adjust stock prices correctly, after which other investors can trade on
the new prices without having followed the underlying causes for the price change.

3. The existence and severity of the forward premium puzzle during tranquil
and turbulent periods: Developed versus developing country currencies –
Shehadeh, Li, Vigne et al – International Review of Financial Analysis

Through our various analysis procedures, we find that the existence

and severity of the FPB vary significantly across different currency
samples and time periods. Over the calm period the FPB is found to
evidently exist for both developed and developing currencies. In addi-
tion, the bias is found to be more pronounced for developed currency set
than for developing currency set. In contrast, the results are significantly
different over the turbulent period. Specifically, our result show that the
2007–2008 global financial crisis constituted a turning point for the
variation of the existence and severity of the bias. We find that coin-
ciding with the peak of the crisis the FPB disappeared prominently and
the UH turned out to hold especially for the developed currency set.
However, for some high-yielding developing currency set the FPB sur-
vived during the crisis period and yet became more severe compared to
the calm period.
Our results and findings can be linked to two important strands of
recent literature. On the one hand, the disappearance of the FPB over the
turbulent period is consistent with the findings of recent research on
currency carry trades. Specifically, currency carry trades are found to perform poorly
and even generate negative returns during turbulent
periods. In addition, it is found that investors tend to unwind their carry
trade positions in periods of low market liquidity, resulting in substan-
tial losses in currency carry trades. Note that the poor performance and
negative returns of currency carry trades is just the flipside of the
disappearance of the FPB. In this context, the vanishing of the FPB over
the turbulent period gives rise to the branch of literature that attributes
the FPB anomaly to the existence of time-varying risk premium in cur-
rency forward rates. Examples of those risk factors, which are found to
significantly explain the excess return of currency carry trades, include
currency crash risk and high liquidity constraints (Abankwa, 2020;
Brunnermeier, Nagel, & Pedersen, 2009; Kumar, 2020), realized and
implied exchange rate volatility (Clarida, Davis, & Pedersen, 2009),
monetary policy volatility (Berg & Mark, 2019; Moore & Roche, 2012),
global FX volatility (Menkhoff, Sarno, Schmeling, & Schrimpf, 2012)
and downside market risk (Dobrynskaya, 2014). This, in turn, suggests
that the specific characteristics of the sample period under consideration
can play important role in the existence of the FPB phenomenon.
On the other hand, the survival of the FPB with more severity for
some high-yielding emerging market currencies over the turbulent
period can be associated with the effects of carry trade flows.13 In the
study of Ahmed and Zlate (2014) the relatively high interest rates are
found statistically and economically significant driver of the capital
inflows received by emerging markets over the period followed the
global financial crisis. Note that carry trade flows create a demand
pressure on high-yielding currencies leading to rising their values. This
just contradicts the predictions of the UH. Consequently, as shown by
Gagnon and Chaboud (2007) and Bacchetta and Van Wincoop (2010),
the gradual building up of carry positions along with the infrequent
changes in portfolio positions can result in “delayed overshooting”,
which in turn can lead to continuous appreciation of the high interest
rate currencies for a prolonged time. This suggests that carry trade flows
and currency-specific factors can impact the existence and severity of
the FPB.14
Overall, the results suggest that the FPB anomaly can be related to
the existence of risk factors, which can affect the predictions of forward
exchange rates, and to the intensity and movements of currency carry
trades. The results also show that the FPB is not a constant and per-
manent finding. It rather varies over time, and across and within cur-
rency sets depending on the sample period. Accordingly, time-period
characteristics and currency-specific factors play important role in the
existence and severity of the bias. In other words, the existence and
severity of the FPB are closely related to both period-specific charac-
teristics as well as currency-specific factors. Examples of those charac-
teristics and factors may include volatility, liquidity, capital movements,
rare events, monetary policy and forward premium sizes. Investigating
the effect of such characteristic and factors on the existence and severity
of the downward bias in the predictions of forward exchange rates are
sensible topics for future research
 Hodrick and Srivastava (1986)

- The primary purpose of this paper has been to investigatethe

- covariation of the risk premium in the forwardforeign exchange market and
- the expected rate of depreciation of the U.S.dollar relative to five
- other currencies. Using alternative statisticaltechniques, we confirmed
- the findings reported in Fama (19814). If one views the forward premium as
- the sumofthe expected rate of depreciation of the currencyplus a risk
- premium, then our evidence indicates thatthe risk premiumisnegatively
- correlated with the expected rate of depreciation. The risk premium, when
- defined this way, is the expected return to purchasingdollars in the:
- forward market. The expected return to sellingdollars or buying foreign
- currency forward therefore covaries positivelywith the expected. rate of
- depreciation of the dollar relative toall fiveforeign currencies..
- Although Farna (19814) found such a covariation puzzlingand potentially
- inconsistent with economic theory, we havedemonstrated that it is
- intuitively plausible and consistent with the predictionof theLucas
- (1982) model.

Of course, this analysis as well as Fama's and all modern rational

expectations time series analysisrelies on the statistical assumptions of

stationarity and ergodicity. Krasker(1980)hasarguedthatthese

assumptions may be incorrect in such analyses. Agents may care about

events that have not occurred in the sample, and the probability of these

events may fluctuate.9 Developing estimation methods to handle these

problems may riot be as critical as determining what the factors actually

are.

FRANKEL AND CHINN (1993)

- Cross sectional dataset of 17 currencies including developing countries

- Little more than half of the forward discount is attributed to variation in risk premium

Verdelhan, 2010

- Applies to Cochrane and Campbell (1999) habit based preference model which was
designed to match features of the equity market also applies to currency market
- The model has two key characteristics: time-varying risk aversion and pro-cyclical real
interest rates. The domestic investor earns positive excess returns in times when he is
 more risk averse than his foreign counterpart. In bad times, consumption is close to the
habit level, risk aversion is high, and interest rates are low. Thus, the domestic investor
expects a positive risk premium when interest rates are lower at home than abroad. This
mechanism reproduces the UIP puzzle

KUMAR, 2020

- data from nine major currencies for 1978:08–2019:09, I provide a novel resolution to this
enduring forward
- premium puzzle by building on the financial economics literature that explores the
economic implications of
- limited access to capital markets. A liquidity shock, or the urgent demand for liquidity by
credit-constrained
- arbitragers liquidating bond holdings, causes losses from sudden drops in bond prices.
Arbitragers require a
- liquidity premium to compensate for potential losses that vary directly with the interest
rate. It is this liquidity
- premium that explains persistent excess returns on currencies.

Froot and Frankel – 1989

- Survey data of institutional investors for 4 currencies

- forward discount into its two components—expected
- depreciation and the risk premium
- We reject the hypothesis that all of the bias in the forward
- discount is due to the risk premium
- We reject the claim that the variance of the risk premium is
- greater than the variance of expected depreciation. The reverse
- appears to be the case
- market participants system-
- atically underpredicted the strength of the
- dollar during the early 1980s

Karen Lewis, 1989

- agents in fact use all available infor-

- mation efficiently and in this sense are ratio-
- nal. In general, the paper analyzes the fore-
- cast error effects due to a change in the
- process of fundamentals that the market
- learns only over time using Bayesian updat-
- ing
-
- In particular, this framework is used to
- empirically investigate the implied impact
- upon dollar forecast errors due to learning
- about the increase in U.S. money deman
 -
- Although the systematic
- nature of forecast errors may appear more
- pronounced over some time intervals, the
- persistence in this behavior over longer peri-
- ods implies that learning about a change in
- fundamentals cannot be the only explana-
- tion. Thus, the apparent systematic behavior
- of prediction errors over longer time periods
- may arise from a combination of learning
- behavior together with anticipations of fu-
- ture policy changes and risk premia

Engel and Hamilton (1990)

- true model of the exchange rate is

- evolving over time. They show that, contrary
- to the hypothesis of a zero-drift random-
- walk model for the exchange rate supported
- by the previous literature (see e.g., Richard
- Messe and Kenneth Rogoff, 1983), a better
- representation of the exchange-rate depre-
- ciation is an uncorrelated shock around a
- drifting mean
-
- One explanation that could be suggested
- for the inability of the switching-regime
- model estimated by Engel and Hamilton
- (1990) to reproduce the forward-market and
- survey data and to forecast turning points is
- that it is assumed that investors use only
- past exchange-rate observations to make
- forecasts. This information set may not be
- enough to forecast changes in policies that
- will affect the future path of the exchange
- rate. Since nobody doubts that when in-
- vestors are trying to forecast exchange rates
- they try to collect as much information as
- they can about present and future monetary
- policy, output growth rates, and trade
- deficits, as well as public announcements of
- government officials, it is important to ex-
- tend the switching-regime model by includ-
- ing other variables in the information set of
- investo
Graciela Kaminsky (1993)

- This paper investigates whether ex-

- change-rate forecasts, although biased, are
- still formed rationally. The idea is that in-
- vestors can be rational and yet make re-
- peated mistakes if the true model of the
- exchange rate is evolving over time. This
- proposition, also known in the international
- jargon as the "peso problem" hypothesis,
- has been gaining support
-
- Some support for the peso problem hypothesis
-
- exchange-rate process has been evolving over
- time, perhaps capturing the changing
- regimes in the market fundamentals.
-
- monetary announcements
- help to predict the future path of the ex-
- change rate, I have assumed that the path
- of the exchange rate conditioned on Rt is
- independent of these announcements. This
- is not necessarily true since Fed officials'
- announcements, by providing information
- about future monetary policy, will affect in-
- vestors' behavior and thus may affect the
- exchange-rate path

Frankel and Foote (1990):

Heterogenous agents – chartists and fundamentalists. Fundamentalists expect the exchange rate to
revert to its fundamental value. Chartists, however, extrapolate the trend in the most recent
period(s). Frankel and Froot (1990) show that the chartists extrapolated an initial strong increase in
the fundamental value from 1980 to 1982 for the subsequent years, causing a strong overvaluation
in 1984. In other words, both fundamental and technical elements are important in the foreign
exchange market. In addition, their relative importance changes over time.

Bacchetta and van Wincoop, 2010:

Most models assume that investors incorporate instantaneously all n

folio decisions. To explain the forward premium puzzle, we depart fr

decisions are usually not made on a continuous basis. While there n

actively manages foreign exchange positions of investors, it develop

still manages only a tiny fraction of cross border financial holdings

is little active currency management over horizons relevant to mediu

ability. Banks conduct extensive intraday trade, but hold virtually no

funds do not actively exploit excess returns on foreign investment

a certain asset class and cannot freely reallocate between domestic

Richard K. Lyons (2001) points out that most large financial institut

own proprietary capital to currency strategies based on the forward discount bias.

Infrequent portfolio decisions in a simple two country general equilibrium model with data from 5
currencoes.

Little volume of foreign exchange positions is actively managed because welfare gain is remarkably
small because of transaction costs and their inherent unpredictability

Second, infrequent portfolio decisions lead to a delayed impact of interest rate shocks on exchange
rates. Thisan explain the phenomenon of "delayed overshooting," whereby the exchange rate
continues toappreciate over time after a rise in the interest rate. Third, the delayed overshooting
leads to substantial excess return predictability in the direction seen in the data. Fourth, even future
excessreturns continue to be predictable by the current forward discount, with the magnitude of
thepredictability declining as time goes on.

Hochradl and Wagner, 2010:

forward bias does not only exist statistically but that betting against UIP

yields Sharpe ratios higher than those of the S&P 500 Index, the MSCI World Total Return Index, and

a JP Morgan US Government Bond Index, even after adjusting for transaction costs. Bias-trading

strategies can be viewed as attractive investment opportunities per se (higher Sharpe ratios), useful

diversification devices (no systematic risk), and promising portfolio extensions for active fund

managers trying to outperform their benchmarks (higher Sharpe ratios and positive Jensen’s alphas).

Spronk, Verschoor, Zwinkles, 2013:

Heterogenous 3 agent model – fundamentalists, chartists and carry traders. Agents are allowed to
shift between all three roles.

The first is the profit from trading on the interest rate differential. This is a risk-free profit locked-in
when engaging into the carry agreement. Second, because the UIP relation typically gives a negative
beta the currency of the high interest rate country will appreciate, onaverage. This implies that when
converting the invested capital back to the low-interest currency, on average one makes a currency
profit.

Carry traders expect the high interest rate currency to appreciate and hence trade against UIP. carry
traders is that they can also explain the value of beta observed in empirical UIP regressions. It turns
out that carry traders have a directional role in estimating the UIP beta. Given that interest rate
differentials are persistent, carry traders introduce momentum effects in a currency that is picked up
and extrapolated by chartists. Hence, it is only due to the existence of chartists that carry traders can
have such a profound effect on foreign exchange markets.

Breedon, Rime and Vitale, 2016:

carry trading “lifts” high interest rate currenciesbut also increases the probability of a dramatic
reversal. Furthermore, using dataon market participants’ forecasts of future currency values, we can
decompose theforward bias into one part associated with time-varying risk premia as a function
oforder flow, and one part associated with forecast errors.We find that order flow can account for a
substantial portion of the forward bias—particularly for currency pairs typically associated with carry
trading.

We find evidence, particularly strong for the USD/EUR and USD/JPY crosses, that orderflow is related
to the forward premium (probably via carry trade activity) and that thisorder flow affects expected
risk premia that condition realized returns. This indicatesthat microstructural mechanisms contribute
to the forward premium puzzle. Thus,according to our decomposition of Fama’sβ, the portfolio-
balance effect of orderflow explains roughly 50% of the forward bias for the USD/EUR rate and 90%
for theUSD/JPY rate but only about 20% for USD/GBP, with the remainder being explainedby
expectational errors.Finally, we see that carry trading activity does not represent afree lunch, in
thatthe positive profits it is expected to gain are offset, to some extent, by the currencycrash risk it
provokes

and Baillee Bollerslev, 2000

The forward premium anomaly is a statistical phenomenon that occurs due to persistent
autocorrelation in the forward premium

Using simulations and subsample analysis, forward premium has very

persistent autocorrelation or long memory, and the slope coefficient in the famous

“anomalous” regression is extremely widely dispersed

The “anomalous” regression estimates reported throughout the literature

do not appear to provide convincing statistical evidence to reject the hypothesis of

the forward rate being an unbiased predictor of the future spot rate. If a time-varying

risk premium does exist, then the currently available evidence also suggests that it

is extremely small at the monthly level.

Baillie and Kilic, 2005

{The theory of uncovered interest rate parity (UIP) states that the expected return, or rate of ap-

preciation on a currency equals the interest rate differential, or equivalently the forward premium.

One popular method for testing the theory has been to regress the rate of appreciation of the spot

rate on the lagged forward premium. A test for UIP is then to test if the slope coefficient is unity, the

intercept zero and the residuals serially uncorrelated. The forward premium anomaly is the wide-

spread empirical finding of a negative slope coefficient in the above regression, so that the rate of

appreciation of the spot exchange rate is negatively correlated with the lagged forward premium.

This phenomenon has been consistently found for most freely floating currencies in the current

float and appears robust to the choice of numeraire currency. Hence the forward premium anomaly

implies that the country with the highest interest rate will have an appreciating currency, and not

a depreciating currency, as implied by the theory of UIP.}

asymmetries and forms of nonlinearity that are pres-

ent in the relationship between spot returns and various fundamentals and variables associated

with time dependent risk. Many of the estimated LSTR models with forward premium, money

supply differentials and conditional variance of US money growth as transition variables, lead to an
outer regime that is consistent with UIP holding. The properties of the regimes are further

investigated with forward premium regressions being run on all observations falling into the

regimes. Much of this evidence is also favorable to the idea of UIP having a higher probability

of holding in the outer regime. Transition variables involving income fundamentals are gener-

ally found not to be of much use as transition variables. A simulation experiment also suggests

that an LSTR dgp can produce data consistent with the anomaly. However, parameter estima-

tion issues lead to considerable uncertainty with the estimated transition functions and hence

imprecise definitions of regimes

GOSPODINOV (2009)

In this paper, we argue that the forward premium puzzle (the finding of

large negative and highly unstable slope parameter estimates in the differenced

forward premium regression) appears to arise from the simultaneous presence

of a risk premium and several empirical characteristics of the spot and forward

exchange rate data that render some standard estimators highly misleading.

{The main stylized facts about the variables that enter the forward premium

regression for exchange rates are the following: (i) the forward premium is a

highly persistent process, (ii) the variability of the forward premium is only a

small fraction of the variability of exchange rate returns, (iii) excess returns and

spot returns exhibit very little persistence, (iv) spot and forward rates appear to

be unit root processes with very similar descriptive statistics, and (v) the errors

that are driving the processes for next period spot and forward rates are almost

perfectly correlated.}

Large negative values and highly unstable behavior of the slope estimates in the usual

differenced specification, reported in many empirical studies, appear to be due to

the simultaneous presence of a risk premium and a number of data characteristics

that have not been fully incorporated into the {inference procedure.

Pippenger. 2010

The forward-bias puzzle is probably the most important of several puzzles in international finance

and open economy macroeconomics because it suggests that there are serious informational ineffi-

ciencies in foreign exchange markets.1 But after 30 years of research we still do not have a generally

accepted solution.

The forward-bias puzzle begins with the crucial assumption that the forward exchange rate equals

the expected future spot rate. Relying on rational expectations, early tests of that assumption used

actual future spot rates as proxies for expected future spot rates. Those early tests regressed actual

future spot rates against current forward rates. Such tests usually produced regression coefficients

that were close to one. But the recognition of the possible effects of unit roots changed the standard
test equation. To achieve stationarity, current spot rates were subtracted from both sides of the
original test equation

For almost 30 years economists have been searching for a solution to the forward-bias puzzle.

Broadly speaking, earlier attempts to solve the puzzle fall into one of two categories: One maintains
the assumption of rational expectations and attributes the bias to a risk premium. The other
attributes

the bias to expectation errors. But the bias in the forward-bias puzzle can be explained by effective

arbitrage in international financial markets.

Covered interest parity implies that the standard test equation that produces the forward-bias

puzzle is missing two variable