Investors should be able to immediately recognize how CRSP's style placement decisions make the indexes meaningful measures of manager style performance. To clarify the intuition: CRSP understands that a large-cap value manager may only choose securities that look like value stocks within his or her universe; the use of a relative value score specific to the large-cap universe provides a good representation of his or her opportunity set. It also means that the same security may have a different style assignment in our midcap or large-cap portfolios, which use their own relative scores.
Specifying what can reside within the opportunity set is a problem different from determining the best way to map to what investors see. The investor's vantage point must be inferred empirically.
Even a question as simple as "What is a domestic company?" has a largely empirical answer. Companies that all of us would consider American are incorporated overseas and recognize revenue abroad for tax purposes. Others, clearly foreign, go public through reverse mergers with U.S.-listed shell companies. We turned to data in an attempt to develop a transparent methodology. The domicile scheme we rely upon is a product of testing thousands of models that examine hundreds of company-level variables.
Other areas are less clear still—cap breakpoints, for example. As mentioned before, large and small stocks behave differently, but even after significant research into potential breakpoints, no clear statistical discontinuities emerge. CRSP's conclusion: Cap segments are a matter of convention. In an effort to stay practical, we adopted cumulative cap breakpoints at levels that should look very familiar to practitioners. A comparison with other major indexes is provided in Figure 3.
Importantly, though, CRSP also included "bands" around these breakpoints and a migration plan called "packeting." We studied banding and migration in depth and let the data provide guidance as to scale and mechanics. We require that a security pass a threshold beyond the breakpoint before a 50 percent "packet" of the security's weight is moved to the adjacent index. It follows that a small-cap stock must move beyond the mid/small breakpoint and the mid/small band threshold before triggering the first 50 percent move to the midcap index; the migration of the remaining 50 percent depends on remaining beyond the threshold next quarter. Among all index providers that rely on banding, turnover reduction is touted as a benefit. We agree, but we believe there is also a deeper purpose. As the indexes age, banding and packeting capture something important: There is no unanimity of opinion among managers as to which marginal companies qualify as large or small. Beyond the reduction in turnover, our migration strategy improves the fit of the index to manager behavior.
If there is a lack of unanimity on cap breakpoints, value and growth styles are downright fuzzy. CRSP's work suggests that, in contrast to the model in Figure 2, value and growth managers do something other than hold portfolios of single-dimension "value" or "growth" stocks. Value managers typically describe their process as one that involves trying to buy assets or cash flows at inexpensive prices. Growth managers, on the other hand, look for fast-growing firms. These managers follow separate, though related, processes. Accordingly, we treat value and growth separately. This two-dimensional method allows us to generate a richer description of the style-specific investment opportunity set.