MACROECONOMIC RISKS FOR INDIAN MUTUAL FUND

MACROECONOMIC RISKS FOR INDIAN MUTUAL FUND
INDUSTRY – AN ECONOMETRIC STUDY
Dr. Sachchidanand Shukla1
Abstract
The Indian Mutual Fund Industry has been growing at a robust pace with increasing
number of global fund managers operating in Indian markets with new products and
services. The volatility in stock markets and financial crisis of 2008 had resulted in a
steep fall in returns as well as AUMs managed by the equity funds. Some of the
portfolios suffered huge losses and adversely impacted investor confidence of investing
in equity funds even when the markets are at low. Some of the reasons could be the
stickiness in portfolios and inability to predict and hedge macro risks led to their inability
to churn portfolios swiftly resulting in less than returns compared to benchmark returns.
This brings forth the importance of quality research to help model the macroeconomic
risks associated with this mutual fund industry.
With the Bull Run in the Indian capital markets, the assets under management (AUM)
has quadrupled from the level of Rs 1,31,000 crore in early 2003 to a little over Rs12
trillion in FY15. The increased openness of the economy to global markets brings in
more uncertainties in terms of both internal risks and policy impacts on the markets and
participants of capital markets. The unprecedented volatility is to be tackled by the fund
manager by using tools that help him to assess and measure the macroeconomic risks.
Current market returns benchmarks of BSE and NSE are typically used for all equity
mutual funds. The portfolio composition of these funds needs to be correctly mapped
with the benchmarks to be used in estimating the performance of the funds. The inability
of these benchmarks to capture macro risks makes it difficult to attribute the fund
performance to the various macro risks.
This paper made an attempt to create alternate benchmarks that capture macro risks and
quantified the equity fund performances.
Introduction
A fundamental principle of finance is the trade- off between risk and return. It is also well
established in the finance literature that macroeconomic variables are sources of
systematic risk in stock returns. The idea that changes in macroeconomic factors may
also affect stock or MF scheme returns follows logically. An understanding of the
1
Sr.Vice President, Axis Capital Pvt Ltd. Views expressed by the author in this article are personal.
1
macroeconomic risk can help lay investors, fund managers and regulators understand the
underlying risks and sources of returns. If the factors affecting the time variation of
macro risk can be properly understood, such factors can be incorporated in capital market
models with variable betas used for predictive purposes, thereby enhancing the power of
these models.
This research work was undertaken with the broad purpose of understanding the linkages
between macroeconomic variables and equity mutual fund returns. The other key
objective was to identify and quantify these macroeconomic risks for portfolios. To
address the key research question mentioned below that was analyzed with Indian
empirical data.
What are the key macroeconomic risks for Indian Mutual funds? Can these be identified
and quantified? Can these be used for constructing superior portfolios which can help
identify sources of macroeconomic risks and portfolio returns?
In the process of analyzing the above key research questions, the following concerns
emerged:





In the current literature there were very few studies on Indian mutual funds
and their macroeconomic linkages
Could all equity MF portfolios be treated as having the same macroeconomic
risk profile despite having varied nature and number of stocks?
Is there any suitable index portfolio for comparison of individual risk profiles
of mutual fund portfolios? If not how could one compare attendant risks and
returns?
What kind of an index portfolio will be most appropriate for this study?
Could a suitable index portfolio and identification of macroeconomic risks
help assess risk and returns and could it help create superior portfolios?
There is hardly any research work which addresses the issues raised in the central
question comprehensively. Following points were the highlights of the literature review
and consequent research gap arising out of it:



The issue of research on macroeconomic risks for equity mutual funds in the
Indian context was limited.
Mutual fund portfolios in India have a high churn with ~30% of the stocks in
portfolio changing within 3-4 months thus rendering a time series comparison
extremely difficult.
Given such high churn ascertaining the macroeconomic risk profile of
individual schemes was not undertaken before.
2


The current benchmark indices ie the Sensex and the Nifty or BSE 100 etc
were not found suitable for comparison of macroeconomic risk profiles of
equity mutual funds.
The current risk measures used for assessing risks of mutual funds ie portfolio
Beta, Sharpe, Treynor and Information ratios do not capture macroeconomic
risks adequately In India.
Literature review
These studies started with Treynor (1965) who made an attempt to find an appropriate
measure of portfolio performance which considered the risk involved in a portfolio. He
concluded that managed portfolios carry market risk, i.e.; the aggregate value of the
portfolio is dependent on the market trends. During bull phase of the market, the value
may go up and during bear phase, the portfolio value may go down. It was Treynor who
introduced the concept of 'beta'' which indicates the degree of variation in the portfolio
value compared to the market portfolio. Hence, the appropriate measure of portfolio
performance is risk premium per unit of 'market risk' generated by the portfolio.
Risk premium is defined as excess portfolio return over risk-free return. Higher value of
Treynor's index denotes better performance of portfolio and vice versa. Treynor's
measure of portfolio performance is a relative measure that ranks the funds in terms
(market risk) and return. This is also known as reward to volatility ratio.
Sharpe (1966) followed suit and propounded another measure of portfolio performance
evaluation. He replaced 'market risk' (beta) in Treynor's equation with 'total risk' i.e.,
'standard deviation and measured performance in terms of risk premium generated per
unit of 'total risk'. The higher the value of Sharpe ratio, the better it is as a higher value of
Sharpe's index indicates better performance of portfolio and vice versa. The Sharpe's
measure of portfolio performance is also a relative measure that ranks the funds in terms
of risk (total risk) and return. The ratio is also known as reward to variability ratio.
Comparing the performance of 34 open-ended mutual funds during the period from 1954
to 1963 with Dow-Jones industrial average (market portfolio) in terms of reward to
variability ratio, it was concluded that overall performance of mutual funds in terms of
Sharpe's index was inferior to Dow-Jones portfolio (market portfolio) during the period
of study. Only 11 out of 34 funds had posted better performance than the market portfolio
(Dow-Jones industrial average).
Jensen (1968) studied the performance of 54 open-ended US mutual funds for the period
1945-64 and found that the returns of mutual funds before the load fees and after
management and other expenses were on average 1% per annum below the benchmark
return. Jensen used return on the S&P 500 index as the benchmark and the yield on one
3
year US Treasury bill as the proxy for the risk-free rate of return. Returns on more than
half of the funds were below the benchmark.
Merton (1973) introduced a macro variable into asset pricing theory, by placing interest
rates into an inter-temporal model. He averred that changes in interest rates produce
changes in the investor’s opportunity set given the belief that current demands are
affected by the possibility of uncertain changes in future investment opportunities.
Merton’s theorem was that all risk-averse investors will be indifferent between portfolios
with a mean-variance efficient frontier, plus a third one to hedge against unfavorable
inter-temporal shifts in the frontier.
McDonald (1974) analyzed the performance of 123 mutual funds in the USA during
1960-69 and used the New York Stock Exchange (NYSE), index as market proxy. He
found that 54% of the funds had posted better performance than the market in terms of
Treynor's measure, whereas only 32% of the funds performed superior to NYSE index in
terms of Sharpe's measure.
Kon and Jen (1979) in their study found that on average, the mutual fund sample was
able to predict security prices well enough to outperform the naive policy, given their
selected levels of risk, and to recoup' all management fees and brokerage commissions.
Ippolito (1989) examined 143 mutual funds for the period 1965-84 and found that on an
average, mutual funds provided 0.83% higher returns per annum over the benchmark
before load fees but after management expenses and other fee. The benchmark was taken
to be risk-adjusted return on the S&P 500 portfolio and the risk-free returns were taken to
be the yield on one year US T-bills.
Elton, Gruber and Blake (1996) found that that there is no evidence supporting stock
picking abilities of the portfolio managers, even if their services are provided free of cost
to the investor. Using relative pricing model they establish that bond funds
underperformed the returns on average by the amount of cost expenses of those funds.
Elton, Gruber and Blake (1996) examine predictability for stock mutual funds using riskadjusted returns. They fund that past performance is predictive of future risk-adjusted
performance. They were also able to recreate portfolio with the same level of risk as the
index funds however with a higher return.
Lockwood (1996) examined equity fund betas using the same variables as Chen, Roll and
Ross (1986). Lockwood used monthly returns for 171 mutual funds for the period 1978 to
1991 and applies a nonlinear factor model. He finds that on average equity fund betas are
negatively related to inflation changes and default risk premia while bond fund betas, on
average, are negatively related to changes in risk-free rates, industrial production growth,
and the term structure. However this study does not show the sensitivity of beta to the
change of the economic variables.
4
Chin Shan Wong and LakshmanAlles (2001) studied the stability of Australian betas.
Exchange rates, current account balance, interest rates, visible trade balance and
unemployment were among the more important macroeconomic descriptors that are
found to influence significant variation in betas of some industries.
Otten and Bams (2002) analysed a sample of 506 funds from 5 different countries
(France, Germany, Italy, Netherland as well as UK) during January 1991 to December
1998. They investigated the performance of the mutual funds based on the Carhart (1997)
4-factor model. They found that the expense ratio and age of the mutual fund are
negatively related, and fund assets, positively related to the risk adjusted performance.
They have also found that in the UK the mutual fund market exhibits significantly
positive net of expenses as well as for gross returns.
Some of the more recent findings on the mutual funds presented by Nitsche, Cuthbertson
and O’Sullivan (2006) show that in cases of US and UK mutual funds only in cases of 25% there is statistical ex-post evidence the passive benchmark is outperformed. However,
there is also substantial evidence that 20-40% of funds underperform. Nitsche,
Cuthbertson and O’Sullivan also oppose Ippolito’s (1989) argument by stating that the
key drivers of relative performance are load fees, expenses and turnover.
More extensive recent analysis was also done by Ferreira, Miguel and Ramos (2006)
where they have conducted analysis on 10,568 actively managed funds from 19 countries
over period from 1999 to 2005. They have used 4 alternative benchmark models and used
economic factors in their analysis such as fund age, size, fees and management structure.
Their findings amongst many comprise of: mutual fund performance being better in
countries with stronger legal institutions; when investing abroad, young funds are more
able to obtain better performance; performance is higher in funds with higher fees and if
the managed by an individual manager with more experience.
Humpe, A., et al. (2009) tried to relate the macro economic variables with long term
stock market movements in US and Japan within the framework of a standard discounted
value model by using monthly data over 40 years. A cointegration analysis was applied to
model the long term relationship between the industrial production, money supply, the
consumer price index, long term interest rates and stock prices in US and Japan. The
authors found a significantrelation between the macro economic variables and stock
market.
In the Indian context, Gupta (1981) laid the foundation of performance of Indian equities.
Immediately thereafter, Jain (1982) had pioneered the work on financial performance of
investment schemes of Unit Trust of India (UTI) during the period 1964-65 to 1979-80,
His work is considered as the first notable work on performance evaluation of mutual
funds in India, In 1986, UTI had floated first equity fund in India, namely, Mastershare
5
under the banner of its subsidiary; UTI (Mutual Fund) subsidiary, 1986. Thereafter
considerable interest has been shown by analysts, academician researchers to examine the
performance of equity mutual funds and financial funds in India from the perspective of
investors and fund managers.
Barua and Verma (1991) had provided empirical evidence of equity mutual fund
performance in India by studying the performance of India's first 7-year close-end equity
mutual fund, Mastershare. They concluded that the performance of the fund was
satisfactory for large investors in terms of rate of return.
Gupta and Sehgal (1997) evaluated mutual fund performance over a four-year period
from 1992-96. The sample consisted of 80 mutual- fund schemes. They concluded that
mutual fund industry faired reasonably well during the period of study.
Sondhi and Jain (2006) examined the performance of 36 mutual funds for the period 1993
to 2002 and concluded that performance of their, sample funds remained far from
satisfactory in terms of rates of return and risk-adjusted returns. All these studies except
Gupta and Sehgal (1997) have reported less than satisfactory performance of mutual
funds in India.
Sarkar, P. (2007) examined the relation between growth and capital accumulation exists
in case of India. Annual data on various variables like nominal and real share price, share
market turnover ratio, number of listed firms in the stock market, fixed capital formation
and growth ofreal GDP and industrial output. But the study concluded that no positive
relationship exists between real and stock market variables either in short run or long run
during 1950-51 to 2005. Sarkar also individually studied the trends over the period of
time in all the said variables and found that most variables became volatile and had
usually an upswing trend during and after mid 1970s.
Kanakaraj et al. (2008) examined the trend of stock prices and various macro economic
variables in India between the time periods 1997-2007. They concluded that a strong
relationship exists between economic variables and stock prices. The study observed that
GDP growth in India has grown consistently at high levels touching the highest average
from 2003-04 to 2006-07 since Independence, and was strongly backed by manufacturing
sector growth and services sector growth.
Robert D. Gay, Jr (2008) investigated the time-series relationship between stock market
index prices and the macroeconomic variables of exchange rate and oil price for Brazil,
Russia, India, and China (BRIC) using the Box-Jenkins ARIMA model. Although no
significant relationship was found between respective exchange rate and oil price on the
stock market index prices of either BRIC country, this may be due to the influence other
domestic and international macroeconomic factors on stock market returns, warranting
further research. Also, there was no significant relationship found between present and
6
past stock market returns, suggesting the markets of Brazil, Russia, India, and China
exhibit the weak-form of market efficiency.
Methodology and database
The APT model derived by Stephen Ross (1976) is based on no transaction costs,
investors having homogenous market expectations and the markets being completely
competitive. The APT is a multi-factor model. In our case we considered the risk factors
to be index specific, even though APT gives no guidance to which factors should be
considered. The drawback over here could be that the asset can be mispriced due to the
fact that irrelevant factors can be used. In this specific case it seems economically
relevant to include specific indices that are to be thought of as benchmarks for particular
asset portfolio. However, one can always gain explanatory power by including more
factors in the model even if some could be irrelevant. Multifactor model such as APT has
tracking properties that can be used to mimic returns of a specific portfolio. The more
assets that have been included in the model the closer it could be to mimic such tracking
portfolio. Given that the portfolio is well diversified and not company specific the closer
the model tracks the returns of such portfolio. The covariance of the factors with the
portfolio expands the measure of risk by the beta attached to each factor.
The general form of the multifactor model can be thought of as
As it can be seen from the above model the return on a specific asset
can be interpreted
as a function of various economic factors . Beta represents the sensitivity of an asset to
a particular factor, also called factor loading in context of arbitrage pricing models. The
error term
represents each asset’s idiosyncratic risk. It can be thought of as
representing company specific component of the return. However, because factors can be
viewed as proxies for new information the error values are to be with mean equal to zero.
That allows for
in the model to be interpreted as the expected return for a given asset.
The model requires that the number of assets must be significantly greater than the
number of estimated parameters.
In order to apply the model to the mutual fund analysis we will use variation of the
multifactor model, the multi-index model of the following form.
Where
is representative of an excess return on a given asset,
represents
managers’ forecasting abilities,
defines the asset’s sensitivity to non-diversifiable
market specific risk and finally
is representative of market specific risk
7
premium where
stands for the primary benchmark/index and
on other benchmarks.
represent returns
A suitable multi-factor model modified and used a priori macroeconomic variables
suitable in the Indian context. A hybrid approach has been used in which some factors
are estimated with pre-determined factor loadings in a cross-sectional manner, but the
loadings for macro factors are estimated with time-series techniques given a selected
macroeconomic series.
This approach is based on the assumption that factor loadings are linear functions of
observable characteristics. This framework fits our research objectives of developing a
useable multifactor model with time-varying (dynamic) structure for Indian Stock
Market.
Thus, with a view to accomplish the stipulated set of objectives of this study, multiple
regression analysis, and econometric techniques such as Unit root tests, Granger Causalty
tests besides attribution analysis tests were used. The key to the methodology has been
the transformation of stock market data as well as mutual fund portfolios.
This study found that the Nifty or Sensex, the most used benchmark indices were not
giving a true comparative picture because of the difference in the nature of the portfolios
or rather their constituents. For eg the returns of Birla SunLife Equity scheme were tested
with the Sensex wherein the NAV of the scheme (the dependent variable) was regressed
upon the Sensex returns but the R-squared or co-efficient of determination turns out to be
a poor 0.065.
Repeating the same exercise with Nifty returns as the independent variable yielded a
similarly poor 0.064 as co-efficient. The study conducted yet another test with BSE 100
returns which comprises of 100 stocks as compared to just 30 in the Sensex and 50 in the
Nifty. Yet, the coefficient hardly changed to 0.067. Since 1992, 53 companies have been
taken out of the index and an equal number of companies have been added. Similarly for
BSE 100, 276 companies have come in and an equal number have moved out of the
index. However, the key thing to note is that while the number of stocks being replaced
is the same, their sectoral composition has undergone a change, which means that the
sensitivity to macro factors would also have undergone a change.
The study used a sample of 55 Mutual fund schemes having a combined AUM of Rs
69,698 cr or 49% of the total equity AUM of India’s Mutual Fund Industry comprising
912 stocks over a period of 10 years. Most importantly, almost 40% of the stocks
changed within a year in most of the portfolios. Hence, a fair comparison of returns with
mutual fund NAVs and the benchmark indices was not possible which explains the poor
R-squared coefficients.
8
Thus, to create a suitable benchmark for comparison, the research study transformed the
entire Sensex into four distinct themes or buckets ie 1) stocks that did well due to
domestic secular growth ie IIP or GDP growth sensitive 2) ones that derived their profits
from Exports and hence having sensitivity to exchange rate or the USD/INR rate, 3)
stocks that were Interest rate sensitive and 4) lastly those that were impacted due to
changes in Govt policy ie stocks in the Infrastructure/ Energy sector.
In this study, monthly data from 2002 onwards to 2010-11 has been used in case of all the
variables like, GDP (Gross Domestic Product), IIP growth rate, SENSEX (Sensitive
Index), SBI PLR, 10 year G-Sec yield, Wholesale price index (WPI). The major source of
data of all the above macroeconomic variables is Handbook of Statistics on Indian
Economy maintained by Reserve Bank of India (RBI) and for the SENSEX is obtained
from Bombay Stock Exchange (BSE) data. For mutual fund portfolios the data base used
is Association of Mutual Funds of India (AMFI).
Fama and French (1992; 1993) provide the most popular method of constructing multifactor models, especially in academia. These models tend to define factors first, using
portfolios or other economic variables. Then, stock returns are regressed on these factors
to estimate factor loadings. Alternately, factor models are often estimated on a purely
statistical basis using principal component analysis or factor analysis.
Among commercial equity risk models, Northfield Information Services, Inc. offers a
model that combines Fama-French type factors, macro factors and statistical factors that
are derived from principal component analysis (Northfield, 2005). On the other hand,
MSCI Barra takes a different approach in which factor returns are estimated given predetermined factor loadings with cross-sectional regressions (Barra, 1998). Another study
used a hybrid approach in which some factors are estimated with pre-determined factor
loadings in a cross-sectional manner, but the loadings for macro factors are estimated
with time-series techniques given a selected macroeconomic series.
This approach is based on the assumption that factor loadings are linear functions of
observable firm characteristics. This framework fits our research objectives of developing
a useable multi-factor model with time-varying (dynamic) structure for Indian Stock
Market.
Thus, with a view to accomplish the stipulated set of objectives of our study, we have
used multiple regression analysis, and econometric techniques such as Covariance, Unit
root tests, Granger Causalty tests besides attribution analysis tests. The key to our
methodology has been the transformation of stock market data as well as mutual fund
portfolios.
Thus, to create a suitable benchmark for comparison, we transformed the entire Sensex
into four distinct themes or buckets ie 1) stocks that did well due to domestic secular
9
growth ie IIP or GDP growth sensitive 2) ones that derived their profits from Exports
and hence having sensitivity to exchange rate or the USD/INR rate, 3) stocks that were
Interest rate sensitive and 4) lastly those that were impacted due to changes in Govt
policy ie stocks in the Infrastructure/ Energy sector.
For this transformation, we used Bloomberg’s widely used Global Industry Classification
System (GICS).
For the analysis of mutual fund portfolios we filtered 55 schemes that have a history of at
least 10 years. We then had to contend with 912 stocks that have been a part of various
portfolios at different points in time over the last 10 years and grouped these stocks into
four distinct groups as the Sensex.
Then first of all, to fulfill the research objectives, descriptive statistics like standard
deviation, coefficient of determination, mean, etc. were carried to show the nature and
basic characteristics of the variables used in the analysis. Correlation is the next step to
move towards the objectives of this study and finding any relation between the stock
market and macroeconomic variables. Then the formal investigation is carried out by
examining the stochastic properties of the variables by using Unit Root Test to test the
stationarity of the variables and causality tests were also performed on variables..
Analysis and results
The study finally culminates into quantification of macroeconomic risk factors for the
selected mutual fund schemes. The study also identifies - growth rate of the economy,
interest rates and exchange rates as the key macroeconomic variables.
It has been argued that macroeconomic factors are important sources of equity return
variation in stock markets. The extant literature suggests that a wide range of factors may
be relevant. Such variables include goods prices, money supply, real activity, exchange
rates, interest rates, political risk, oil prices, the trade sector, and regional stock market
indices. It is rather well known that the macroeconomic variables chosen by Chen et al.
(1986) have been the foundation of the APT. However, in emerging markets, there is
argument that not all of these variables are either relevant or appropriate (Bilson et al.,
2000) who have given the description of selecting macroeconomic variables for emerging
equity market returns. They also give a good theoretical foundation why these
macroeconomic factors affect equity returns. The importance of other macroeconomic
variables such as inflation, national output, industrial production etc. has been pointed out
by Fama (1981).
However, these are only examples how macroeconomic variables can be chosen. The
APT does not give any formal guidance on which or how many factors should be used.
10
Thus, a researcher should decide the right factors for his specific purposes (e.g., what are
the unique features of the country that’s examined).
The number of factors ranges from zero to almost ten in our papers examined and
because of this lack of guidance in choosing the factors to the APT, the risk factors are
selected a priori according to the unique features of the Indian Economy and the
composition of the Sensex, the leading market index.
Thus, as described, the macro economic variables we chose are domestic growth (a proxy
for IIP growth), exports (proxy for USD/INR), Government Policy related variable and
Interest rate (proxy for SBI PLR) and defined macroeconomic themes.
We constructed four buckets of portfolios that represent four different macroeconomic
themes. For e.g. “Domestic growth” theme represents a part of the portfolio that has
sensitivity to domestic growth proxied by IIP growth in our estimations. Similarly, the
“Export or exchange rate related theme” represents a part of the portfolio that has
linkages with the ‘exchange rate linkages’ ie the USD/INR rate. “Govt ” theme represents
sensitivity to changes in government policies ie decisions on Infrastructure investments
etc. The last theme which is called as the “Interest rate sensitive” theme is that part of the
portfolio that is sensitive to changes in domestic interest rates.
It is worth pointing out, why these variables could affect equities’ returns:
Table 1: Risk measure comparisons
Scheme name
Birla Sun Life Equity Fund – Growth
DSP BlackRock Opportunities Fund – Growth
Franklin India Bluechip – Growth
Franklin India Opportunity Fund – Growth
Franklin India Prima Plus – Growth
HDFC Growth Fund – Growth
ICICI Prudential Top 100 Fund – Growth
ING Core Equity Fund – Growth
JM Equity – Growth
Kotak 50 – Growth
L&T Opportunities Fund – Cumulative
PRINCIPAL Growth Fund
Reliance Growth
Sundaram Growth Fund – Growth
Tata Pure Equity Fund – Growth
Templeton India Growth Fund – Growth
UTI Top 100 Fund – Growth
Overall
risk
2 .5 7
1 .3 1
4 .1
1 .1 8
2 .5 5
0 .5 1
1 .1 2
1 .2 4
0 .6
2 .1 4
0 .7
1 .8
-0 .3 3
1 .0 5
1 .4 8
2 .4 3
0 .3 4
Scheme
beta
1.4118
1.4634
0.6485
1.4066
0.9235
1.2479
1.0584
1.2424
1.3871
1.0596
1.0837
1.4792
1.4398
0.887
0.6137
0.8937
0.7984
Sharpe
Treynor
1.3113
1.3617
1.7841
1.0315
1.6873
1.6574
1.3431
0.6589
0.4972
1.2616
0.5288
0.7933
1.5075
1.1525
1.4911
1.5527
0.7465
15.1566
15.3461
30.1649
11.0419
22.9406
17.8809
15.0445
7.691
5.9149
13.8225
6.3944
9.1554
19.9793
14.2762
18.9008
19.4935
9.3332
Information
Ratio
1.3076
1.2287
1.1154
0.5811
1.6472
2.1765
1.3099
(0.2815)
(0.1620)
1.3333
(0.5183)
0.0034
1.6577
0.6865
0.1962
1.5710
(0.4235)
Jenson
5.8876
3.6383
13.4610
(2.2123)
12.4951
5.6032
4.5653
(4.5503)
(4.8459)
2.8072
(3.6860)
(6.4169)
15.1239
4.3220
5.1820
9.0086
(0.0048)
Brinson’s attribution model:
Perhaps the best-known approach to performance attribution is the Brinson method.
Brinson and Fachler in the Journal of Portfolio Management presented this model in
11
1985. However, the method goes back much further than that. One earlier description was
published by a working group of the Society of Investment Analysts in London, back in
February 1972.
Attempts to distinguish which factors are the sources of the portfolio’s overall
performance. Was it because the manager was superior at selecting securities, or did they
demonstrate superior market timing skills by allocating funds to different asset classes or
market segments?
This method compares the total return of the manager’s actual investment holdings to the
return for a predetermined benchmark portfolio and decomposes the difference into
allocation effect and a selection effect.
The most straightforward way to measure these two effects is as follows:
Allocation Effect  i[( wai  w pi )  ( R pi  R p )]
Selection Effect  i[( wai )  ( Rai  R p i )]
Where:
wai, wpi = the investment proportions given to the ith market segment (eg. Asset class,
industry group) in the manager’s actual portfolio and the benchmark portfolio,
respectively.
Rai, Rpi = the investment return to the ith market segment in the manager’s actual
portfolio and the benchmark portfolio, respectively
Rp = the total return to the benchmark portfolio
Computed in this manner, the allocation effect measures the manager’s decision to overor underweight a particular market segment (ie. [wai – wpi]) in terms of that segment’s
return performance relative to the overall return to the benchmark (ie. [Rpi – Rp]).
Allocation Effect  i[( wai  w pi )  ( R pi  R p )]
Selection Effect  i[( wai )  ( Rai  R p i )]
The selection effect measures the manager’s ability to form specific market segment
portfolios that generate superior returns relative to the way in which the comparable
market segment is defined in the benchmark portfolio (ie. [Rai – Rpi]) weighted by the
manager’s actual market segment investment proportions. When constructed in this way,
the manager’s total value-added performance is the sum of the allocation and selection
effects.
Overall Actual Return  (0.50  0.097)  (0.38  0.091)  (0.12  0.056)
 8.98%
Overall Benchmark Return  (0.60  0.086)  (0.3  0.092)  (0.10  0.054)
 8.46%
12
Attribution test results:
Almost all the funds and the schemes in our sample showed that they could outperform
the benchmark Sensex by a wide margin when it came to the domestic secular theme.
Seven of the schemes did well when it came to government policy theme. Four funds
outperformed in the interest rate theme.
Table 4 Attribution test results summary
Domestic Secular theme
Exports/ Exchange rate theme
Govt Policy theme
Interest rate theme
DSP Blackrock Opportunities Fund
JM Equity
Kotak 50
L&T Opportunities Fund
Sundaram Growth
Tata Pure Equity
UTI Top 100
HDFC Growth
ICICI Pru Top 100
Reliance Growth
Birla SunLife Growth
Franklin India Bluechip
Franklin India Opportunities
Franklin India Prima Plus
ING Core Equities
Principal Growth
Thus it is clear that most mutual funds in our sample are susceptible to abrupt changes in
interest rates and exchange rates. Our finding is also vindicated by the fact that the
mutual fund industry has struggled to attract investors post the Lehman crisis wherein
exchange rate and interest rate volatility has gone up significantly.
Conclusion
This study concludes that there is a need to construct the appropriate benchmarks to
measure the performance of the mutual fund schemes. The current benchmarks do not
capture the portfolio churning impact and also the sensitivity of portfolios to
macroeconomic policy changes.
Hence, these four proposed benchmarks are to be computed and track changes in these
over a period of time and use them in the quantification of fund performance in addition
to the normal risk adjusted measures of CRISIL or ICRA and the ranking based on some
attributes by Value Research firm.
Limitations
Some of the key limitations of this research study are:
All variables are measured contemporaneously with NAV returns. Hence there are no
expectations in the model and there is an implicit assumption of contemporaneous
association as the focus of this research study is in understanding and explaining the
variation in realized, rather than expected returns.
13
We recognize that the selection of four macroeconomic variables is not perfect and that
cases can be made for inclusion of other variables. Moreover, the proxies we have used
may have measurement errors. But this will be true for any other variables that one may
include.
 The data has not been back tested to judge the scheme performances to portfolio
churning

Tools like Stress testing, simulations and computing the short run and long run
effects of macroeconomic policy changes could be used to further refine the
study.

Role of leads and lags in capturing the policy impacts is to be tested.
Policy Recommendation
The current research study reiterates the importance of macroeconomic risks in managing
equity mutual fund portfolios. The traditional risks measure such as the market Beta,
Sharpe and Treynor ratio etc are good at capturing only the market risk adequately.
However, given the macroeconomic uncertainty and outlook there is a need for
macroeconomic risk measures so as to enable investors and fund managers gauge such
risks.
Since the Indian Mutual Fund industry has an important role in the intermediation of
financial savings and a vehicle for channelizing household savings, the dynamics of
macroeconomic uncertainties need to be explicitly mentioned so as to enable people to
take informed investment decisions.
An important fact is that the current benchmarks do not adequately capture the
composition of sectors and portfolio churning from a macroeconomic risk point of view
resulting in inadequate understanding of risks and return.
Therefore, the current model of decomposing mutual fund portfolios based on
macroeconomic sensitivities is recommended as an additional risk tool soas to enhance 1)
appreciation of macroeconomic risks to the portfolio and 2) enhancing portfolio returns
by hedging against underlying macro risks or capitalizing on the upside potential, if any.
Therefore, investors should be cautious in considering relevant macro factors in
balancing maximizing fund returns and managing risk.
Some extensions of this research are obvious. Given that India is an important emerging
market economy, similar models for other emerging markets can be developed. This
approach can provide wider views than the current version without failing to capture
major local influences.
14
This new method fits the characteristics of Indian stock market and the Mutual fund
industry well. First, the empirical results demonstrate a-priory and high-explanatory
power of the model. This method captures the macro and global economic variables,
which are influential in markets globally in general and in emerging markets in particular.
Second, the dynamic structure of the model will enable to capture the key characteristics
and changes of the Indian stock market loaded with dynamic factors and uncertainty.
Reference
1) Abdullah, D. A., & Hayworth, S. C., 1993, Macroeconometrics of stock price fluctuations,
Quarterly Journal of Business and Economics, 32, 46-63.
2) Alles L. and CS Wong, 2001, The Stability of Australian Industry Betas, International Journal of
Finance, Vol. 13 No.1.
3) Berry M. A., Burmeister E., McElroy M., 1988, Sorting out risks using known APT
a. Factors, Financial Analyst Journal, Vol. 44, No. 2, 29–41.
b. Bhattacharya, B., and J. Mukherjee, 2002, The Nature of the Causal Relationship
between Stock Market and Macroeconomic Aggregates in India: An Empirical Analysis,
Paper Presented in the 4th Annual Conference on Money and Finance, Mumbai.
c. Fama, E.F., 1970, Efficient Capital Markets: a Review of Theory and Empirical Work,
Journal of Finance, vol. 25, no. 2, 383‐ 417.
4) Granger, C.W.J., 1969, Investigating Causal Relations by Econometric Models and Cross-spectral
Methods, Econometrica, 37, 428-43
5) Gupta O P and Sehgal S., 1997, Investment Performance of Mutual Funds, The Indian Experience,
in Indian Capital Market; Trends and Dimensions, Tata McGraw-Hill Publishing Company Ltd,
(Courtesy Institute of Capital Market, Navi Mumbai), 1-41
6) Ippolito R A., 1989, Efficiency With Costly Information: A Study of Mutual Fund Performance
1965-1984, Quarterly Journal of Economics, Vol, 104, 1-23
7) Jensen Michael G., 1968, The Performance of Mutual Funds in the Period 1945-1964, Journal of
Finance, Vol, 23, 389-416
8) Markowitz, H., 1952, Portfolio Selection, Journal of Finance, 15, 77- 91
9) Merton, R., 1973, Theory of rational option pricing, Bell Journal of Economics and Management
Science 4, 141–183
10) Mukherjee, T. K. and Naka, A., 1995, Dynamic Relations between Macroeconomic Variables and
the Japanese Stock Market: An Application of a Vector Error-Correction Model, Journal of
Empirical Research, 18, 223-237.
11) Roll, R., 1977, A critique of the asset pricing theory's tests Part I: On past and potential testability
of the theory, Journal of Financial Economics, vol. 4, no. 2, 129‐ 176
12) Ross, S.A., 1976, The arbitrage theory of capital asset pricing, Journal of economic theory 12,
341- 360
13) Sharpe W.F., 1964, Capital Asset Prices: A Theory of Market Equilibrium Under Conditions of
Risk, Journal of Finance, vol. 19, no. 3, 425‐ 442.
14) Sortino, F.A and Price, L.N., 1994, Performance measurement in a downside risk framework, The
Journal of investing, Vol 3, No 3; pp 59-64
15) Zubi and HussainSalameh, 2007, Explaining the Stock Return Via a Macroeconomic
a. Multifactor Model, Jordan Journal of Business Administration, Volume 3, No. 1, 2007
15