Chart 1: Time Heals

Source: Ibbotson Associates Time periods really do matter. Growth soundly defeated Value in the late 90s, but as the measurement period grows, the two converge to the S&P 500.

In an effort to compare the styles over the long term, we examined the Ibbotson data for the S&P Barra Growth and Value Indexes for the twenty-five years starting from their inception (January 1, 1975 - December 31, 1999). As Chart 1 shows, growth led value in the latter 90s yet

Chart 2: Convergence

Source: Ibbotson Associates Despite varying short-term performance, at the end of the 25-year period Growth, Value and the S&P 500 Index ended up at the same level.

over longer time periods, the style indexes and the S&P 500 all had roughly the same returns.

Chart 2 which shows both style indexes as well as the S&P 500 with roughly the same cumulative return over the twenty-five years further illustrates this point. As you’d expect from these results, the risk/return profile of all three is quite similar (see table below).

This isn’t to say that each style doesn’t dominate over shorter periods of time. When the 25-years are divided into five-year (Chart 3) and ten-year periods (Chart 4), value outperformed

Chart 3: Five-Year Periods

Source: Ibbotson Associates The best performer depends on the time period chosen. Variances are larger than over 10-year periods in Chart 4, below.

in the late 70s and early 80s while growth led the way in the late 90s. The S&P 500 dominated in only one 5-year period (1985-1990) although it was often near the lead.

Death to Indexing

This might be taken as an argument against indexing although we’d argue that at least for long-term investors, the exact opposite is true. The fact that returns converge over the long-term gives little reason to deviate from the index (see table, below). In addition, the Sharpe ratio (a measure of risk adjusted return) and standard deviation (a measure of volatility) are also favorable for the 500. Given this, why should the long-term investor ride through the periods of underperformance when the index itself has steadier returns?

Chart 4: Ten-Year Periods

Source: Ibbotson Associates Leading styles fluctuate over 10-year periods, but not to the extent as in the 5-year periods shown in Chart 3, above.

An alternative, of course, would be to switch between value and growth as the situation dictates. Indeed, if you could successfully follow this approach, you would handily outpace the style indexes as well as the overall S&P. Chart 5 illustrates this best-case scenario. We calculated the best-case returns by taking the highest return for the Value and Growth Indexes for each month in the measurement period. We then looked at five-year periods as before. Not surprisingly, the best-case scenario performed much better than any of the other styles (compare with Chart 3).

Unfortunately, no investor has the omniscience to use this approach. In order for it to work, you’d have to know beforehand which style would outperform in any given month. Real investors can’t do that. Instead, the closest they come is to move from one style to another only after it

Can't Beat the Index

Geometric Mean Arithmetic Mean Sharpe Measure Standard Deviation S&P/BARRA 500 Growth 16.00% 17.49% 0.29 18.72% S&P/BARRA 500 Value 16.31% 17.47% 0.33 16.45% S&P 500 Index 16.37% 17.63% 0.32 17.16% Source: Ibbotson Associates

begins to outperform. To see how this affects results, we went back to the monthly numbers. This time we assumed the investor would stick with the same style until there was a month when the other outperformed. When that happened, the investor would switch to that style the next month. In essence, the switch was always one month late.

These "one-month-late" results are also shown on Chart 5. In every instance these results are clearly inferior to those of the best-case scenario. Secondly, with the exception of the 5%+ difference in the 1975-1979 period, results don’t differ that significantly to those of the S&P 500. Once trading costs and taxes are considered, any benefit over the broad index is greatly diminished if not totally eliminated.

Just a Matter of Time?

Would different time periods change the results? Sure they would, at least in the short-term.

Chart 5: Outguessing the 500

Source: Ibbotson Associates If you could jump from Growth to Value and back again before the fact, you'd do significantly better than the index. However, if you miss the shift by just a month, your benefit decreases to the point that taxes and transaction costs would leave you right back with the index return -- maybe even worse.

In trending markets, style bets can outperform. This is clearly demonstrated in Chart 3 in the 1975-1979 (Value) and 1995-1999 (Growth) periods. But the one thing that doesn’t change is the convergence of styles over time. The ten-year periods illustrated in Chart 4 show smaller effects of short-term trends as well as stable performance from the S&P 500 across all periods.

We would suggest the debate between value and growth is inconsequential for long-term investors. While it may be the basis of interesting speculation in a world with perfect information and no transaction or tax costs, real investors are better served with an approach that closely tracks specific benchmarks. Better yet, if returns can incrementally exceed the benchmark -- even by a small amount -- over time such a "core" approach should exceed the best style based management.