About Morningstar's Fund Classifications

From Names to Holdings A reader asked me how Morningstar determines its fund categories. Not the definitions, but instead the underlying process.

The first decision when creating a fund-classification system is whether to group funds by their names or by their holdings. For many years, names prevailed, because no other approach was possible. This often proved confusing. Equity-income funds might hold solely dividend-paying stocks, or they might blend stocks and bonds, so that they most resembled balanced funds. The two types of funds did not belong in the same category--but were so placed.

By the early 1990s, advancements in technology permitted a new approach. Armed with funds' portfolio holdings and the computational power to evaluate them, Morningstar jettisoned fund companies' definitions, opting instead to impose its own structure. Initially, that decision placed some funds into categories that belied their labels. For example, several investment-grade bond funds were classified as high-yield bond funds. Displeasure followed. Eventually, though, the affected funds either changed their names or their investment practices.

The result has been increased unity among fund companies. They use terminology more consistently than they did before.

Principle No. 1: When possible, control fund-classification schemes from within. Eventually, that will impose more discipline on fund marketing than the industry will accomplish on its own.

Establishing the Framework When switching to its new system, Morningstar appeared to have a clean slate. That was not necessarily so, however. An example will illustrate.

In the 1990s, institutional consultants commonly sorted domestic-stock funds by size and price--that is, into 1) large- and small-company approaches, and 2) value and growth styles. In addition, Eugene Fama and Ken French had recently pegged those attributes as being important determinants of future stock market performance. Morningstar, therefore, was highly likely to adopt such an approach for its own domestic-equity framework. To do otherwise would have confused the marketplace.

As it turned out, cataloguing stocks by size and price was not the only reasonable strategy. The Fama-French paper acknowledged a third factor, beta (that is, stock market sensitivity). A few years later, researchers added momentum to the list of stock-price influencers. The four-factor model was released--and Morningstar (along with other fund databases) only used two of the factors. Morningstar could just have reasonably sorted stocks on beta and momentum, as on size and price.

But the die had been cast. The marketplace had been informed that U.S. stock funds would be measured on size and price. Fund companies would create funds to fit that design; financial advisors would grow accustomed to measuring and comparing funds in that fashion; and individual investors would organize their portfolios with that structure in mind. To switch now would be disruptive.

This does not mean that the equity style-box scheme--or other aspects of the fund-classification structure--is forever locked. If significant evidence accumulates that rival approaches would be clearly superior, rather than merely equally sensible, then Morningstar would seriously consider an overhaul. But such a decision would not be made lightly.

Principle No. 2: Time matters. Morningstar's fund-classification decisions are influenced by the prevailing conditions.

Principle No. 3: Fund taxonomies do not occur in isolation. They affect fund sponsors, financial advisors, and investors. Consequently, inertia is powerful. Changes should lead to substantial improvement, not incremental benefits.

Few or Many? After determining the frameworks, the next issue is how many categories to create. Fund companies prefer more. The fewer the competitors, the better chance that a fund that can top its category, thereby attracting attention--and investor dollars--for finishing number one.

Investors, on the other hand, tend to take the opposite view. The more categories that exist, the more complicated the mapping becomes for their asset allocations. If Morningstar has 20 taxable-bond categories (as indeed it does), which to choose for client portfolios? After all, nobody needs anything close to 20 taxable-bond funds.

Both views have merit.

Dividing categories into ever-finer partitions makes for cleaner investment views. That is, offering 16 diversified domestic-stock fund categories, via a four-by-four grid, provides more-accurate comparisons than does Morningstar's current nine-box system. Twenty-five categories would be better yet. But adding categories reduces comparability; funds that once would have been measured against each other become separated. The language becomes that of Babel.

Reversing the process by forcing all funds into a few buckets makes for ready comparisons and simplifies the asset-allocation task. This is how institutional consultants once operated, many decades back. All large-company stock funds here, all long taxable-bond funds there, and so forth. Unfortunately, as the consultants discovered over time, evaluating such heterogeneous groups frequently leads to false conclusions. The "best" funds tend not to be those that are most skillfully managed, but instead those that have ridden the prevailing investment winds.

Morningstar splits the difference. It has just over 100 mutual fund categories--fewer than fund companies would like (they constantly request additions) but more than investors can easily use. Sometimes, Morningstar is parsimonious--as with carrying but nine diversified U.S. stock categories to cover several thousand funds. Sometimes, it is not. Thus, every target date receives its own category, although, (for example) 2050 and 2055 funds contain almost identical portfolios.

Principle No. 4: Whether to establish a large, medium, or small fund-classification scheme is a matter of taste. Overall, Morningstar occupies the middle ground.

Wrapping Up This column should be regarded as a description, not a prescription. I do not believe that a fund database can succeed by disobeying the first of Morningstar's four principles, to classify funds by their investment holdings rather than their names. The drawbacks for that arrangement are too severe to overcome. But the other three principles are not ironclad. They explain how Morningstar arrives at its fund-classification decisions, but other databases could justifiably dissent.

John Rekenthaler has been researching the fund industry since 1988. He is now a columnist for Morningstar.com and a member of Morningstar's investment research department. John is quick to point out that while Morningstar typically agrees with the views of the Rekenthaler Report, his views are his own.