In a previous analyst spotlight, my colleague Ben King discussed the investor’s tendency to chase performance. That is, to buy “hot” funds with attractive three-year records or, perhaps, even less. Of course, it’s human nature to search out such strong-performing investments regardless of their investment styles, asset class or sector exposures.

When considering the purchase (or sale) of a stock fund, we examine its industry market exposures against a benchmark to identify the manager’s current and historic preferences. For example, if the fund has 10% of its assets in energy and the benchmark has only 5% exposure to energy, it is fair to state that the manager is overweight energy. It may also be fair to imply that the he or she is favoring that sector. (Managers often claim to be agnostic on a sector, but are simply overweight an area because they’ve found a handful of attractively priced stocks.)

Examining a fund’s sector exposure is a relatively simple, yet valuable, task – especially when constructing a portfolio. When trying to establish which investment styles and sectors investors (the market) are favoring, we run a similar exercise that compares all funds in aggregate against a benchmark. Creating an accurate picture of the exposures of the entire market is a little more involved than looking at a single fund.

To make the process and data more manageable, we begin by screening for domestic equity funds only. Doing this reduces 25,000 funds (including fixed income and international funds) to a more manageable 7,500 funds. In calculating an average exposure for all 7,500 funds, we must take into account a fund’s asset size and its overall equity exposure.

By considering asset size, we are not treating all funds equally. For example, let's say that Fund A and B are identical except that Fund A has $20 billion under management while Fund B has $200 million. If both funds have 20% of their assets in financials, this means that Fund A has $4 billion invested in that sector while Fund B only has $40 million in financials. An equal-weighted average would treat both funds simply as having 20% in financials. What's potentially lost in this observation is this: when Fund A has 20% of its assets in financials, it has a far greater impact on the market relative to Fund B. However, an asset-weighted average takes this into consideration.

Second, we want to consider to what extent a fund is invested in equities. For example, we run similar exercises as above, where hypothetical funds X and Y are identical except that Fund X is fully (100%) invested in equities while Fund Y only has 70% of its assets in equities. In drilling below the surface, we may discover that both funds have 15% weights in healthcare. But if Fund Y has 70% of its assets in equities, only 10.5% of its assets are really invested in this sector (15*.70 = 10.5). Using the equity-adjusted basis helps us to paint a more accurate picture of where investors and managers are investing.

Once we have an asset-weighted equity-adjusted average for style and sector exposures of the market, we need a benchmark to compare it against. For example, the data may show that financials has a weighting of 15%, while energy has a weighting of 10%. Does this mean that investors and managers are favoring financials over energy? Not necessarily. It’s possible that in the market (benchmark) financials count for 20% of the market and energy represents 5% of the market. If that were the case, then managers would actually be underweight financials and overweight energy. Even though financials were given a greater absolute percentage of assets by managers -- in relative terms -- they were putting more emphasis on energy. In our portfolios we prefer to use the Russell 3000 as a benchmark as it provides a more complete view of the market compared to the popular S&P 500.

A final consideration in this analysis is to place our data point relative to the benchmark in a historical context. The process described above provides us a snapshot of what the market looks like at a precise moment in time.

For example, lets say that the data shows that small cap stocks are overweight compared to the benchmark by 1.0%. How are we to interpret this? Does this mean that simply because small caps are overweight by 1% that investors and managers are favoring small caps and that it would be the contrarian play to move out of small caps? We cannot tell just by looking at one data point. Indeed, it’s entirely possible that investors and managers typically overweight small caps by 4-5%. If that’s the case, a drop to 1% overweight indicates that investors are moving away from small caps. Placing the snapshot exposure within a context is important for identifying trends and flows of assets.

A Concrete Current Example
Now, let’s pull all of this together in an actual example.

Back in August 2006, the asset-weighted equity-adjusted exposure given to large-cap value by investors and managers was 20%. The weighting given to large-cap value in the Russell 3000 was 28%. This means that investors and managers were underweighting large-cap value stocks by 8%. At the same time, investors and managers were overweight large-cap growth stocks by 2% compared to the Russell 3000. However, through August 31 of 2006 large-cap value stocks were outpacing their growth stock counterparts! The Russell 1000 Value Index returned 11% compared to a nearly flat return of 0.3% for the Russell 1000 Growth Index. This might seem contradictory, as one would expect investors to have been chasing large-cap value performance.

Where does this all fit in a historical context? In this particular exercise our historical data goes back 43 consecutive months to the beginning of 2003. The graph below provides one way to place the data in its appropriate context. The graph plots both the relative exposures of large-cap value and growth compared to the Russell 3000. A rolling three-month average, which looks back over the past three months, is used to help smooth out the data and identify trends.

 

Large Cap Value and Large Cap Growth Relative Exposures
Compared to Russell 3000

As seen in the graph, at the start of the period, investors and managers were underweighting both large-cap value and large-cap growth compared to the Russell 3000. More recently, investors and managers have been moving assets toward large-cap growth stocks and away from large-cap value in general. Also, for all time periods in this study, investors and managers continued to show a preference for large-cap growth despite large-cap value outperforming growth 89% to 48% over that period (as measured by the Russell 1000 Value and Growth indexes, respectively). If one desired to pursue a contrarian method and avoid following the herd, this model would suggest that an investor increase their large-cap value exposure.

Conclusion
Does the scenario above mean that we would take all of an investor’s assets and move them out of large-cap growth funds and into large-cap value funds? No. For starters, this would run contrary to our commitment to building balanced, diversified portfolios. Additionally, this is a short timeframe upon which to base drastic portfolio sub-allocation decisions (such as a move out of large-cap growth entirely). Nobody can say for certain which style or sector will outperform in the future, but this research can help give us an idea of where the market (herd) is moving.

Ultimately, this is an ongoing exercise. Our primary concern remains to interview managers to understand their strategy, philosophy, and temperament. But, our passion for the markets and desire to provide investors a product best suited for their needs leads us to explore alternative investments.

-- Jeff DeMaso