 |
Index mutual funds are becoming increasingly popular in Canada. In 1993,
there were 5 equity index funds available in Canada, but by the end of 1997, the
number has ballooned to more than 25. Given this surge in popularity, it makes
sense to examine the rationale for indexing.
Indexing is - very
simply - an investment strategy which attempts to mimic the performance of a
market index. An index is a "yardstick", and a market index is a
group or "basket" or portfolio of securities selected to represent and
reflect the market as a whole. Indexing is, therefore, a passive strategy,
because it does not involve either security selection or trading. The basket or
portfolio of securities defined by the index is purchased, and held
indefinitely. |
The most widely recognized index is the Dow Jones Industrial Average, which
is composed of 30 large U.S. industrial companies. Despite its stature, the Dow
is not particularly representative of the US market and most index funds
tracking the US market are based on the Standard and Poor's 500 (S&P 500).
In Canada, the most widely followed index is the TSE 300, which tracks the
1200-plus stocks on the Toronto Stock Exchange. Other indexes of interest to
Canadian investors include the Financial Times 100, or "footsie" which
is based on the London Exchange; the Nikkei, which is based on the Tokyo
exchange; the Dax, which is based on the Frankfurt exchange, and the Hang Seng,
which is based on the Hong Kong exchange. There are also a couple of important
international stock indexes: the Morgan Stanley Capital International (MSCI)
index and the Europe, Asia and Far East (EAFE) index.
There are also indexes for bonds. In Canada, the most prominent index is
the ScotiaMcleod Universe Bond Index. There is however only one fund based on
it: the brand new CIBC Bond Index Fund.
Historically, the justification for index funds has derived much of its
strength from the "efficient market hypothesis", which maintains that
the history of stock price movements contains no useful information that will
enable an investor consistently to outperform a fully diversified randomly
chosen portfolio. This view was supported by a number of US academic studies in
the 1960s and 1970s and seemed to be confirmed by the experience of the 1980s,
when US index funds significantly outperformed most actively managed funds.
The logical argument for index funds seems simple:
-
the market is composed of either passive or active investors
-
on average, the performance of passively managed portfolios matches the
performance of the market
-
on average, the performance of actively managed portfolios, must,
therefore, match the performance of the market
-
the net returns of actively managed portfolios will be lower than
the net returns of passively managed portfolios due to the greater transaction
costs associated with active management strategies.
The first two of these propositions are uncontroversial. There is also
little doubt about the last proposition; the costs of managing an index fund are
lower. There are several reasons for this. The most costly element of active
portfolio management is security analysis, but index funds do not require
security analysis, since the securities in their portfolios are determined by
the composition of the index. In addition, the cost of buying and selling
securities can be expensive in actively managed portfolios (depending on the
degree to which the portfolio is traded). Since the composition of an index
changes relatively little, index funds incur very minor transaction costs.
Finally, most index funds are extremely efficient. In an actively managed fund,
administrative inefficiencies may go undetected because they cannot be isolated
from expenses which are incurred in the pursuit of higher gross returns. In
index funds, because gross returns are dictated by the returns of the market
index on which the fund is based, cost inefficiencies in the operation of the
fund have a direct negative impact on the fund's net returns. As a result,
competitive pressures usually operate to eliminate these inefficiencies.
The central claim in the justification for index funds is therefore the
proposition that the performance of actively managed portfolios will - on
average - match the performance of passively managed portfolios. The evidence
typically advanced in support of this proposition is that the average fund
actually underperforms the index in most years. In the US, for example,
approximately three-quarters of all stock funds underperformed the S&P 500
index during the 1980s. But a more careful consideration of the data is more
revealing. During the latter 1970s and continuing into the early 1980s, small
capitalization stocks dramatically outperformed large capitalization stocks. By
the 1980s, small cap. stocks had become significantly overvalued. As a result,
large-cap stocks began to outperform small-caps, and because the S&P 500 is
composed disproportionately of large-caps, index funds appeared to outperform
actively managed funds. The underperformance of actively managed funds
stimulated a shift of assets to index funds, which were of course invested in
the stocks of the index, driving the value of the index itself up in a kind of
self-fulfilling cycle. It took the recession of 1990 to put an end to that
cycle.
In Canada, the experience is somewhat different. The TSE 300 outperformed
the average Canadian stock fund in only nine of the past fifteen years, and in
only four years during the 1980s. Nevertheless, there is an enduring/abiding
sense that the index is somehow a 'safer' alternative than actively managed
funds.
Evidence regarding the relative performance of active and passive funds is
usually met with the argument that investors choose specific funds; they do not
choose the "average" fund. While some funds underperform, others
outperform the index. In response, advocates of indexing argue that even the
funds which exceed the index in one or two (or even three) years cannot
outperform on a sustained basis. Studies of mutual fund performance have
consistently demonstrated that funds with the highest returns in any given
period do not achieve the highest returns in a subsequent period. According to
John Bogle, chairman of the US-based Vanguard group of funds,
and one of the most forceful exponents of indexing, the top 20 stock funds funds
(out of 681 funds) in the US in any given one year period achieved an average
rank of 284 in the subsequent one year period. Similarly, the top 20 funds over
the decade from 1972 to 1982 ranked -on average - 142 (of 309 funds) in the
decade from 1982 to 1992. The fact that periods of relatively strong investment
performance are often followed by periods of relatively poor performance is
called "mean reversion" and is regarded by advocates of indexing
almost as an axiom.
There are however several problems with this argument. The first is that
regression to the mean is far from self evident. A careful look at the
information used to substantiate this view reveals that the data is drawn from
year-over-year returns. But focusing on discrete year-over-year performance
puts excessive emphasis on short-term performance. Yet mutual fund investments
in particular are long-term investments. Mutual fund investors do not typically
invest on a year to year to basis. A more reasonable assessment of "mean
reversion" would therefore focus on long-term compound returns.
In addition, it is well recognized that active portfolio management
encompasses a number of different investment approaches and styles, and that
each of these approaches and styles is more successful at some stages of the
business cycle than at others. To properly judge the performance of active
management, it is essential to evaluate returns over an entire cycle, and this
means examining long-term compound returns.
Another problem with the arguments raised by those who advocate indexing is
that there is a substantial difference between a portfolio manager's inability
to achieve top rankings in consecutive periods and an inability to outperform
the index in consecutive periods. While Bogle's analysis
contends that the probability of sustaining top quartile performance is 'only'
one in three, the important issue is the probability of sustaining performance
which exceeds that of the index. Not too surprisingly, Bogle does not offer
data which illuminates this issue.
There is evidence nonetheless. While it is unquestionably difficult to
outperform the market on a sustained basis, fully nine Canadian equity funds
have consistently outperformed the ten-year compound returns of the TSE 300 to
the end of 1996.
Risk-Adjusted Performance
There is however a final argument for indexing. Drawing on modern portfolio
theory, advocates of indexing argue that while actively managed portfolios may
(occasionally) beat the index, they assume a disproportionate degree of risk to
do so. On a risk-adjusted basis, indexing is the most efficient
strategy.
To understand the concept of risk-adjusted returns, an analogy may be
helpful. In baseball, a batter's ability is measured not by the number of hits
he gets, but by his batting average. If one player has 100 hits, while another
has 50, we can't say that the first is "better" than the second,
because we also need to know how many opportunities each has had. If the first
batter has had 400 "at bats" while the second has had only 150 "at
bats", then it is the second player who is better. The first batter has a
relatively ordinary .250 batting average (that is, he hits once in every four "at
bats") while the second batter has an impressive .333 batting average
(hitting once in every three "at bats.") In short, meaningful
comparisons can only be made on the basis of hits per "at
bat."
Modern portfolio theory applies the same principle to comparisons of
investment returns. If portfolio "x" has an annual return of
18%, while portfolio "z" returns only 12%, we cannot
immediately conclude that portfolio "x" is "better"
than portfolio "z". Instead of batting averages, what we want
to know is the return per unit of risk. If portfolio "x"
has twice as much risk as portfolio "z", then our initial
perception of portfolio "x" as the better investment is wrong:
its superior return required a disproportionate level of risk.
There are at least two different techniques for measuring risk adjusted
returns. The Treynor Index subtracts the risk-free return (typically the 90 day
T-bill rate) from the return of the portfolio, and divides the remainder by the beta (b) of the portfolio. The Sharpe Index is similar to the
Treynor Index, but the return in excess of the risk-free rate is divided by the standard deviation (s) of the portfolio rather than its beta.
The trouble with both these techniques for evaluating "risk-adjusted"
returns is that they equate risk with short-term volatility. They are utterly
meaningless in evaluating the relative merits of long-term investments.
Indexing and the Deficiencies of Active Fund Evaluation
Despite the shortcomings in the rationale for index funds, they continue to
attract investors. The reason for this paradox is that the conventional
approach to mutual fund analysis does not give investors any confidence that
they can identify funds that will consistently outperform the market in the
future. In his comprehensive analysis of mutual funds, Bogle tacitly
acknowledges the possibility that it is possible to outperform the market when
he observes that "....some of these fund managers have done such a good job
for such a long time that we can fairly assume they have unusual talents."
He then goes on to note that "...such extraordinary managers not only are
few in number but are difficult to identify in advance." This however is a
far less powerful argument for indexing than the logical argument. In essence,
it concedes that actively managed funds may beat the market, and argues instead
that indexing is a kind of insurance against the risks of choosing a fund which
may significantly underperform the market. Investors may not outperform the
market with an index fund, but they will never underperform (by very much.)
The roots of the difficulty in identifying funds with superior potential are
to be found in the excessive preoccupation of current analytical techniques with
numerical comparisons. The information most commonly mentioned is compound
returns, volatility (beta and standard deviation) and sector allocations.
Statistics such as alpha, R-squared and the Sharpe, Treynor and Jensen indexes
(together with a variety of even more arcane measurements) are also regularly
quoted. Conspicuous by its scarcity is data on the stability of management and
the way in which investment decisions are made. Yet it is precisely these
qualitative factors which provide the best clues to the probability of superior
long-term performance.
When investors choose a particular mutual fund, what they are really doing
is choosing a manager for their money. And in evaluating money managers, it is
sensible to examine not only their results but the approaches by which they were
achieved. By relying on an index to define the portfolio, the fund manager has
effectively delegated the stock picking function to some other person or agency
which has goals quite different from those of investors. The securities in an
index are intended to reflect the liquidity, capitalization and volatility of
the market, as well as its strength (or weakness) while the securities in an
investment portfolio are intended to provide superior growth, or a stable
income.
It is important to remember that the index is not entirely passive: it does
change over time. although a great deal of care goes into the selection of the
index, it nevertheless represents a specific view of the market. For example,
in the effort to be representative, there is an inevitable tendency to include "newer'
industries in an index. These industries tend however to be more actively
traded and often have higher price-earnings ratios than more established
industries. The consequence of modifying the index to include these industries
is therefore to exaggerate increases in the value of the index, and distort the
picture of the broader market.
Investors should be interested to know if outstanding performance was
achieved by loading up on "hot" sectors, or was the result of superior
selection of individual securities, regardless of sector. They should be
interested to know if the fund's decision-makers choose investments on the basis
of consistent criteria, or whether they rely fundamentally on "gut feel."
Prudent investors would presumably want to know whether the manager's
organizational structure affected its investment decisions. Many funds for
example hire portfolio managers on a contractual basis. There is considerable
pressure on these managers to pursue high short-term returns in order to secure
their employment, but the pursuit of high short-term returns can often
jeopardize superior long-term returns. And they would want to know if the
personnel responsible for strong performance is likely to remain with the fund
into the future. Unfortunately, much of this information cannot be easily
converted into numbers. As a consequence, its importance is minimized or
ignored.
The Best of Both Worlds
Recently, a number of fund companies have developed so-called "super"
index funds. These are funds which purport to outperform the index by
manipulating the relative proportion of securities in the index. While they are
promoted as funds which merely enhance the benefits of passive investing, the
reality is that they are actively managed, since decisions about which sectors
or securities should be emphasized must necessarily be made by the manager. As
a result, they are more about exploiting the marketing appeal of indexing than
they are about the merits of passive investing.
An examination of Royal Mutual Fund's new "strategic index" funds
clearly reveals the actual character of these funds. The securities held by the
funds are chosen from the S&P Compustat database, but the various fund
portfolios bear no relationship to any commonly recognized index and the
investment approach is anything but passive. In the case of the US Growth
Strategic Index Fund, the pool of eligible securities consists of all US stocks
in the database with a market capitalization in excess of $150 million, and the
portfolio includes the 50 stocks with the greatest one-year price increase which
have a price-to-sales ration of less than 1.5, and an increase in year-over-year
earnings. Unlike a truly passive fund, in which the portfolio is modified only
in response to changes in the composition of the index, the investment approach
of these funds require portfolio re-evaluation an annual basis, in order
to ensure that the holdings continue to meet the mandated criteria. The
strategy of "strategic indexing" is in reality nothing more than an
attempt to fool investors by marketing active management in the seductive guise
of passive investing.
A Final Note
At their root, index funds essentially abdicate the goal of outperforming
the market in favour of (nearly) matching it. They are, like all "single
decision systems" largely a mechanical substitute for thoughtful
deliberation. Although it is true both that many actively managed funds
underperform the market, and that the process of picking superior performers
requires considerable skill, these two facts in no way lead to the conclusion
that index funds are inherently better and that a strategy of indexing is the
most prudent approach. |