Statistical

Variables
Methodology
Statistical Variables Methodology

Some people monitor only their funds’ returns, while others need a comprehensive statistical
analysis. For the latter, we publish several statistical measures on Value Research Online’s fund
pages and the Fund Screener tool. These measures help in analysing a fund’s risk and risk-
adjusted returns. Here’s a quick explainer of these measures and how we calculate them.

Mean return (%)

This is a fund’s average monthly return (annualised) over the trailing three years.
Mathematical formula:
∑Ri
Mean return=
n
where,
∑ is symbol for summation of each instance of the fund returns considered
Ri is each instance of monthly return of the fund portfolio
n is number of monthly instances considered
A fund with higher value for mean returns has performed better than others.

Standard deviation (%)

Standard deviation measures the total volatility of a fund and is based on the
fund’s monthly returns over the last three years. If you add and subtract standard
deviation from the mean return, then that would provide you the range within
which the fund’s returns have largely varied.
Mathematical formula:
2

Standard deviation (SD)= Ri - Rp

∑
n-1
where,
∑ is symbol for summation of each instance of the fund returns considered
Ri is each instance of monthly return of the fund portfolio
Rp is the mean return of the fund portfolio
n is number of monthly instances considered
A fund with low standard deviation indicates less volatility in fund returns, whereas
a high standard deviation implies higher volatility in returns, as it has varied over a
larger range of returns.

September 2023
Statistical Variables Methodology

68-95-99.7 rule

This is an empirical rule in statistical science that determines the percentage

of values that fall within an estimated interval for a normal distribution.
Application - Given a normal distribution of funds returns, 68% of the times
the fund returns are within one standard deviation (i.e., both +1 and -1 x
SD) of the mean, 95% are within two standard deviations (i.e., both +2 and
-2 x SD) of the mean, and 99.7% are within three standard deviations (i.e.,
both +3 and -3 x SD) of the mean.

Variance (%)

Variance is nothing but the square of standard deviation value. Both are ways of
measuring how much the fund’s returns vary from their average value.
A low variance indicates low volatility in fund returns, whereas a high variance
implies higher volatility in returns.

R-squared
R-squared is a measure of a fund’s correlation to the market. It can be interpreted
as the percentage of a fund’s movements that can be explained by the
benchmark’s movements. We calculate R-squared by comparing monthly returns
over the trailing three years to those of a benchmark.
The R-squared ranges between 0 and 1. A score of 1 indicates a perfect
correlation with the benchmark, i.e., the returns of the fund closely trace those of
the respective index.

Example - If the R-squared of a fund is 0.50, then about half of the observed
variation in the fund’s performance can be explained by the benchmark’s
performance.

Beta
Beta measures how much a fund’s returns are sensitive to the market

September 2023
Statistical Variables Methodology

movements. It helps you understand how much a fund can gain or lose when
the market moves up or down. Thus, it tells you how risky or volatile a fund is
as compared to the market. It is calculated based on the trailing 3-year monthly
returns of the fund and the benchmark.
The beta of the benchmark or (ideal) index fund is 1. A higher beta (closer to
1 or more) indicates that the fund’s movements are sharper than the market.
However, a low beta (closer to 0) does not necessarily mean lower volatility – it

Beta should always be used in conjunction with R-squared

For example, suppose you have two assets with different R-squared and beta
values:
Fund A: R-squared = 0.9, Beta = 1.5
Fund B: R-squared = 0.3, Beta = 1.5
Both funds have the same beta, but Fund A has a higher R-squared than Fund
B. This means that Fund A is more closely related to the benchmark than Fund B.
Fund A’s beta is more reliable and useful than Fund B’s beta because it reflects
how Fund A actually behaves compared to the benchmark.

only indicates that the fund does not have a high correlation with its benchmark.
A negative value of beta means that the stock is inversely correlated to the
benchmark, i.e., it moves opposite to the movements of the benchmark.

Alpha (%)
Alpha is a measure of a fund’s risk-adjusted return. It is the excess return of
the fund above the risk-adjusted market return, given its level of risk as
measured by beta.
The mathematical formula for calculating Alpha (theoretically also referred to as
Jensen’s Alpha) is as below:

September 2023
Statistical Variables Methodology

Alpha= Rp-(Rf+Beta (Rb-Rf ))

where,
Rp is the mean return of the fund portfolio
Rf is the risk-free rate of return for the time period considered
Rb is the mean return of the benchmark

Beta is the beta of the fund with respect to the benchmark

A positive alpha indicates that the fund has performed better than expected, given
its beta. And a negative alpha indicates the fund has under-performed given the
amount of risk taken by the fund manager.

Sharpe ratio
Sharpe ratio measures the fund’s returns per unit of risk assumed. It is calculated
simply by deducting the risk-free rate of return from the average monthly return
and dividing it by its standard deviation for the time period considered.
Normally, the higher the Sharpe ratio, the better the fund’s historical risk-adjusted
performance. A value greater than 1 is generally considered good, and anything
less is considered to be sub-optimal.
Mathematical formula:
Rp-Rf
Sharpe ratio =
SD
where,
Rp is the mean return of the fund portfolio
Rf is the risk-free rate of return for the time period considered SD is standard
deviation of the fund’s returns for the time period considered
Pitfalls: Sharpe ratio is calculated based on the assumption that returns are
normally distributed, but it may not be so in the real-world markets. Also, the
ratio does not distinguish between returns on the upside or downside and
focuses only on volatility irrespective of its direction. Hence, it is better to look
at it along with other statistical measures such as Sortino ratio, maximum
drawdown, etc., to better comprehend the fund’s risk-return character.

September 2023
Statistical Variables Methodology

Treynor ratio
It is a measure of the fund’s returns (risk-adjusted) per unit of market risk
undertaken. It is calculated by deducting the risk-free rate of return from the
average monthly return and dividing by its beta. The higher the ratio, the better
the fund’s historical risk-adjusted performance.
Mathematical formula:
Rp-Rf
Treynor ratio=
Beta
where,
Rp is the mean return of the fund portfolio
Rf is the risk-free rate of return for the time period considered
Beta is the beta of the fund with respect to the benchmark
The formula is quite similar to the Sharpe ratio in the numerator, however the two
are different. Sharpe ratio helps to understand a fund’s return compared to its
own portfolio risk, while the Treynor ratio explores the fund’s return generated for
each unit of systemic risk of its portfolio.

Sortino ratio
Sortino ratio is a modified version of the Sharpe ratio, wherein it takes into
account only the negative volatility, while Sharpe ratio considers the total overall
volatility. It achieves this by utilising the asset’s downside deviation, which is the
standard deviation of only negative portfolio returns, instead of the total standard
deviation of portfolio returns.
Mathematical formula:
Rp-Rf
Sortino ratio=
SDd
where,
Rp is the mean return of the fund portfolio
Rf is the risk-free rate of return for the time period considered
SDd is downside standard deviation of the fund’s returns for the time period
considered

September 2023
Statistical Variables Methodology

Again, much like Sharpe ratio, a higher Sortino ratio is better as it helps evaluate
a fund’s return for a given level of only the “bad” risk, as deviation on the positive
side is not really intuitively considered as a risk by investors.

Information ratio
Information ratio measures by how much a fund outperforms its benchmark,
while taking into account the incremental risks (compared to benchmark) involved
in achieving those higher returns. It is calculated by deducting the return of the
index from the return of the portfolio and dividing it by the tracking error (standard
deviation of the differences between each instance of the fund’s returns and the
benchmark’s returns for the period considered).
Mathematical formula:
Rp-Rb
Information ratio =
TE
where,
Rp is the mean return of the fund portfolio
Rb is the mean return of the benchmark
TE is the tracking error
A high information ratio indicates that a manager has consistently generated
better returns than the benchmark index, after adjusting for risks.

Covariance
Covariance measures the directional relationship between the returns on the fund
and its benchmark.
Mathematical formula:
(Ri-Rp)X (Rm-Rb)
Covariance=
(n-1)
where,
∑ is symbol for summation of each instance of the fund returns considered
Ri is each instance of monthly return of the fund portfolio
Rp is the mean return of the fund portfolio
Rm is each instance of monthly return of the benchmark

September 2023
Statistical Variables Methodology

Rb is the mean return of the benchmark

n is number of monthly instances considered
A positive covariance shows that the fund and its benchmark generally move
up or down together, while a negative value for covariance indicates that they
have an inverse relationship, i.e., they tend to move in opposite directions. If
covariance is closer to zero or zero, then it means that there is very little or no
clear relationship in a particular direction between the fund and the benchmark.

Upside ratio (%)

The upside ratio measures how well a fund performed during periods when the
benchmark return was positive. This statistical measure helps to evaluate the
overall performance of the fund during up-markets. It is often compared with the
downside ratio to get a comprehensive understanding of the fund’s performance in
both, up and down markets. It is sometimes also referred to as the Upside capture
ratio.
Mathematical formula:
Rpu
X 100
Upside ratio = R
bu

where,
Rpu is the mean return of the fund portfolio during instances of positive
benchmark returns
Rbu is the mean of positive benchmark returns

Example - If a fund has an upside ratio greater than 100, it means that it
has performed better than the benchmark when the latter delivered positive
returns. For instance, an upside ratio of 130 indicates that the manager
outperformed the market by 30% the period under consideration.

September 2023
Statistical Variables Methodology

Downside ratio (%)

The downside ratio measures how well the fund was able to outperform (i.e.,
stem the fall of returns) as compared to its benchmark when the benchmark
delivered negative returns. It is often compared with the upside capture ratio to
get a comprehensive understanding of the fund’s performance in both, up and
down markets. It is sometimes also referred to as the Downside capture ratio.

Mathematical formula:
Rpd
Downside ratio = X 100
Rbd
where,
Rpd is the mean return of the fund
portfolio during instances of negative benchmark returns
Rbd is the mean of negative benchmark returns

Example - When a fund has a downside ratio of less than 100, it means it
has performed better than the index during periods when the benchmark
has had negative returns. For instance, if a fund has a downside ratio of
75, it suggests that the fund’s portfolio declined only 75% as much as the
benchmark during the period under consideration.

Downside risk
Downside risk is calculated as the standard deviation of the fund’s performance,
but only on the negative side, i.e., in case of losses. Thus, the mathematical
formula for downside risk is the same as for standard deviation, just that is
calculated only for instances when the fund delivered negative returns.Downside
risk is sometimes also referred to as semi-deviation.

September 2023
Statistical Variables Methodology

Example - Suppose an investment has 10 annual returns, as follows:

5% -2% -5% 1% 9% 8% -3% 8% -8% 12%

The standard deviation of the data set is 6.82%, and its downside deviation
(i.e., only for negative numbers) is 2.29%. This means that roughly 33% of
the total volatility is due to negative returns, while the remaining 67% is due
to positive returns. This breakdown reveals that the investment’s volatility is
mainly due to “good” volatility.

Maximum (Max) drawdown (%)

Maximum drawdown is the highest loss that a fund experiences from its highest
point to its lowest point before it reaches a new high during a given period.
It is calculated based on observations of peaks and troughs of the fund’s
monthly return data when plotted on a graph for a given period. Thus, it can be
represented as the following equation.
Max drawdown = Trough Value-Peak Value
Peak Value
This measure is a way to evaluate the downside risk of a portfolio over a
particular period of time. It is a useful tool for evaluating the comparative risk level
of one fund versus another.
Max drawdowns are always negative. It is a useful tool for evaluating the
comparative risk level of one fund versus another. Thus, if one were to compare
risk between two funds using max drawdown, then a fund with a lower value
of maximum drawdown should be preferred as it implies that losses from their
investments were minimal. On the other hand, a maximum drawdown of -100%
would indicate that the investment has no value at all, which is the worst possible
outcome.

Example - Maximum drawdown of Fund A and Fund B are -66% and -40%
respectively, for a given period. The Fund B has a lower maximum drawdown
and hence is less risky than Fund A for that period.

September 2023
Statistical Variables Methodology

It’s worth noting that maximum drawdown only considers the magnitude of the
most significant loss, and it does not take into account how frequently significant
losses occur. This metric concentrates on capital preservation, which is a vital
consideration for many investors.

Maximum (Max) gain (%)

Maximum gain refers to the highest total return over consecutive periods of
positive returns within a given timeframe. Let us understand this with an example.
Consider the following case where we have the month-end NAV given
for a fund.

Total positive returns

Date Fund NAV Monthly return
for the consecutive months

31-12-22 10
31-01-23 10.5 5.0%

28-02-23 10.7 1.9% 7%

31-03-23 10.4 -2.8%

30-04-23 10.6 1.9% 5.77%
7%
Max gain

Here, we see that the fund has delivered positive returns in the first two months
consecutively, i.e., in Jan-23 and Feb-23, where the total return delivered by
the fund was 7%. Subsequently, it had a poor run delivering negative returns in
Mar-23. Thereafter, for the next three months, it delivered positive returns, which
aggregated to 5.77%.
Thus, when we see the highest returns delivered by the fund where it had a
consecutive streak of positive returns, it is 7%, which is the max gain of the fund
for the six months considered.

Maximum (Max) loss (%)

Maximum loss refers to the lowest total return over consecutive periods of
negative returns within a given timeframe. Let us understand this with an example.

September 2023
Statistical Variables Methodology

Consider the following case where we have the month-end NAV given
for a fund.
Here, we see that the fund has delivered negative returns in the first two months
consecutively, i.e., in Jan-23 and Feb-23, where the total return delivered by the
fund was -4%. Subsequently, it had a sudden rally where it delivered high positive
returns in Mar-23. Thereafter, for the next three months, it delivered negative
returns, which aggregated to -7.55%.
Thus, when we see the lowest returns delivered by the fund where it had a
consecutive streak of negative returns, it is -7.55%, which is the max loss of the
fund for the six months considered.

Date Fund NAV Monthly return Total positive returns

for the consecutive months

31-12-22 10

31-01-23 9.8 -2.0%

28-02-23 9.6 -2.0% -4%

31-03-23 10.6 10.4%

30-04-23 10.3 -2.8%
31-05-23 10.2 -1.0%
30-06-23 9.8 -3.9% -7.55%

-7.55%

Maximum (Max) loss

September 2023