What Is Style?

January 01, 1999

Defining growth stocks vs. value stocks by measures like price-to-book is often unsatisfactory. Rather than impose a theory on the real world, Pradhuman used a pragmatic test: What do recognized and respected money managers known to use growth or value philosophies pick, and what criteria fit their behavior? The results provide food for thought: What does it mean that a growth stock may simply be one with high earnings projections and a wide following?


Investment 'style' has become a commonplace of the investment landscape in the last decade. Most research has concluded that a significant amount of variation in performance results from underlying investment styles. Furthermore, this demarcation between growth and value styles has become more institutionalized as active managers attempt to distinguish their approach from others', and the consultant community struggles to categorize and evaluate investment managers.

As a result, the search for a way to distinguish performance differentials between growth and value has become especially important.

Most existing style benchmarks tend to use some measure of valuation, most famously price-to-book but also other measures, to stratify growth versus value. Even though the general approach of using valuations is reasonable and has almost become an industry standard, such methodologies only attempt to model symptoms of growth and value.

Our research suggests that earnings forecasts and information flow account for a significant amount of the variation in investment style. Specifically, we have found that companies with high expected earnings growth and high analyst coverage appear to be a fair representation of growth stocks, while those companies that offer slower or sluggish growth and are under-followed by the 'Street' tend to characterize value stocks.


In order to better model style benchmarks we need to identify traits that can uniquely discriminate style characteristics. A priori, we expected growth stocks to exhibit higher growth expectations, more systemic risk, higher visibility, and higher valuations. Conversely, value candidates were considered more likely to reflect lower growth expectations, less systemic risk, lower visibility and lower valuations.

Higher growth expectations can generally be described by some measure of profits growth. We felt simple measures of historical or forecast earnings and sales growth rates should suffice. Furthermore, long-term growth rates should minimize the cyclical influences of short-term estimates. Systemic risk reflects the level of riskiness of some asset versus the equity market. In theory, if a company has higher expectations than the market, then it will most likely reflect above-average market risks - a fair measure of this might be beta. In short, growth stocks should reflect higher betas. If value ideas are truly out of favor, then they are more likely to reflect lower systemic risk and therefore exhibit lower betas.

In addition, growth stocks tend to be more visible than their value counterpart.

Higher expectation growth stocks tend to be the ideas that are more often discussed at cocktail parties. Wall Street finds it easier to raise capital for fast growing firms. To no-one's surprise, growth stocks tend to be more followed by sell side analysts than value ideas. Hence, we theorized the level of information flow should allow one to better identify growth from value candidates.

Finally, valuations can certainly allow one to distinguish a cheap, low expectation idea from an expensive high expectation candidate.

Hence, we tested these four elements for real-world relevance.


Most existing benchmarks and measures for investment styles, as noted, typically concentrate on valuation. That is, they rank stocks from cheapest to most expensive. Those that have lowest valuations can be thought of as 'value' ideas and those that are most expensive or 'non-value' can be called growth. The basic premise is logical - value candidates are likely to be cheaper, while growth stocks tend to be more expensive. A report by Steve Kim (also of Merrill Lynch) 'Global Equity Derivatives Monthly: Style Indices and Derivatives: An Initial Overview' (12/96) presents a good background for some of the most common and publicly used style benchmarks.

Even though this general approach is reasonable and has almost become an industry standard, such methodologies only attempt to model symptoms of growth and value. That is, low valuations are symptomatic of low expectations, not necessarily a primary driver of value. Valuation factors can lead to false signals. Is a cheap stock temporarily oversold? Or is a beaten-down stock that appears cheap anticipating greater future problems? If the logic points to the former, then a valuation variable may appropriately lead to a value opportunity. However, if the latter scenario is more descriptive, then a valuation-based signal could lead one into a deteriorating situation.

For the most part, investment managers are not satisfied with the popular definition of investment styles. For example, most growth managers will adamantly argue that in searching for good growth stocks, they are not necessarily searching for expensive ideas. While higher valuations are generally a result of faster growing firms, they are not necessarily a precursor of growth.

Our work supports their concern that valuations do not adequately describe investment style.


In most simple terms, growth stock investors look for companies with above average, sustainable growth. The search for growth typically places the investor in a higher multiple subset of the market. Value investors look for companies that offer compelling valuations. Such investors hope that the market has overpriced existing concerns in a company. Their belief is that market participants cause security prices to overshoot intrinsic values in both directions. If a company is experiencing above average growth, the market chases the stock above and beyond reasonable levels. Similarly, as a firm discloses adverse business conditions to the marketplace, the market tends to severely punish such shares. The value investor looks for such companies that are severely down in price terms - so low that a great deal of adverse news could not cause further price erosion. In short, stocks that have overpriced risks are more likely to offer more reasonable risk-reward comparisons.

The nature of growth stock investing suggests that such an investment manager is much more willing to ride out the vagaries of the market if the firm's business is sound. Conversely, a value investor may be more willing to ride out the vagaries of the economic cycle to avoid the swings of the market. In short, style based investing may be defined along the lines of preference for risk taking.

It is commonly thought that growth stock bettors tend to participate in a riskier segment of the equity market. After all, investing in companies with extreme valuations can lead to experiencing more price fluctuations. At the first hint of concern, the market reaction can be swift. The caveat for holding such securities with extreme valuations is that such companies typically exhibit few signs of business erosion. In fact, the business outlook for most growth stocks tends to be quite positive.

Conversely, a value candidate may reflect less market volatility than its growth counterpart, yet reflect greater business risks. The market is fairly efficient at pricing concerns. If a stock is trading at a significant discount to its peers then it is likely that there are concerns regarding the business outlook of the firm. By purchasing a cheap stock, an investor might not only minimize the potential for market risks, but also increase his exposure to the business risks of a firm about which the market is pessimistic or discouraged to begin with.

Companies do not necessarily have to remain in either category, growth or value, over the long run. Depending on perceptions, market direction, and company strategies, a full-fledged growth stock can quickly become value. For instance, an emerging growth company that has fallen on tough times may stumble with a product rollout and raise serious concern in the marketplace. Investors may re-evaluate the firm and question its potential for growth. If the expectations fall causing the selloff of the stock to levels well below intrinsic value, such a firm that was only recently perceived as growth can quickly become a value candidate. At this point, value investors may assess the damage and question whether the market's wrath was an over-reaction and decide whether this 'fallen angel' is of good value.

Similarly, a beaten-down value idea, by design or accident, can launch a successful product or campaign and become a high expectation candidate in a fairly quick manner. As a result, a price level may not be the best indicator of the inherent style of a stock. Yet, most simple valuation parameters of style benchmarks are based on price. While most simple measures tend to contain some measure of value from the income statement, the balance sheet or both, the share price of the value measure is by far the most volatile component. Therefore, as price swings occur, the style assignment of a stock, call it growth or value, can bounce anywhere between deep value situations to severely overbought conditions.


The MLSCR Style benchmarks were arrived at only after numerous tests. The initial assumption, or the naive approach, placed equal emphasis on each of the four traits we selected: expected earnings/sales growth, beta, analyst coverage and valuations. The equal weighted results correlated well. The chart in Figure 3, Panel A, depicts the results of the naive model versus the relative performance of active managers.

Even though the equal weighted model appears to properly capture key inflection points, the results can deviate severely from actively managed portfolios. The last two years are a good example of the deterioration between the simulated data and active returns. In fact, the modeling results were improved by leaving out the inferior variable. The simple backtest results appear to suggest that expectations of growth, rather than valuations, may be more relevant in defining style.

Similarly, systemic risk, or beta, may already be accounted for by the amount of information flow and outlook for profits. After all, higher expectation stocks can generate significant price swings when earnings come in, even if only pennies, above or below expectations. Furthermore, stocks more in the spotlight are more likely to fluctuate with the ebb and flow of the overall market. As a result, beta did not appear to add much to one's decision framework for growth and value.

The model results improved by simply removing a valuation parameter. And in fact, the systemic characteristic, or beta factor, also appeared to add little benchmarking value.


We applied a similar framework to large cap stocks. For the most part, the results appear to be quite similar. That is, valuations and beta did not appear to enhance or better define style results. Acombination of expected growth and information flow or analyst coverage handily models the twists and turns of large cap growth and value investments.

It is interesting to note that even though the results are fairly strong, they are not as compelling as the smaller capitalized sector results. Some of this is explained by more subtle style distinctions that may exist among large companies. Pradhuman and Crosby have argued that larger companies, due to greater access to financing and awareness, are more able to re-invent their corporate identity. As a result, a large cap value idea can more easily recover and create interest among traditional growth managers.

Note, the distinction of companies by size is significant. To better model style, one needs to segment the various sizes that exist in an equity market. By segmenting the equity market by size then style, many difficulties in categorization can be avoided. The range of small cap data tends to be much more varied than large caps. Furthermore, small caps tend to have less available data. To combine such varied data can lead to mixed results.


A non-parametric approach, using a ranking scheme, to separate growth from value was utilized. This approach assigns a rank for each of the relevant variables. We then combine the factors by assigning a certain weight or probability. This leads us to the iterative or optimized portion of the analysis.

What is the optimal combination or optimal weighting for each factor?

The table on page7 relates a series of basic combinations examined. See the appendix for a description of the data. Panels Aand B portray the performance spreads of many of the tested combinations.

The two best approaches, #10 and #12, are defined as combinations of expected growth and analyst coverage. The analyst coverage variable or #21 also appears reasonable but does not appear to track the fund manager results as closely. There is also an implementation issue with analyst coverage - the data are clustered.

Much of the data is clustered around no analyst coverage.

To develop a relevant time series of low coverage, a pool of companies of zero coverage is created. That portion is far greater than the top decile or top 10% of companies with high analyst coverage.

The summary statistics in the table support our contention that style benchmarks are more robust when expectations and information flow are combined. In fact, the statistics become more robust as variables such as valuations and beta are removed.


Our model results not only appear to mimic the behavior of growth and value funds, but also reflect some of the basic traits believed to accompany investment styles. Bernstein suggests that factors such as the abundance or scarcity of earnings and economic sensitivity tend to drive the preference for growth or value. Value appears to lead growth in times of a healthy economic backdrop, while growth appears to lead value in times of sluggish or slowing growth.

We have argued that style investing tends to reflect a seasonal bias. See Pradhuman and the November 1998 issue of The SmallCap Perspective for a detailed discussion of the January Effect and style investing. The benchmarks appear to support our thesis that the January effect is not only a small cap effect, but also a small value effect. Furthermore, as investors attempt to bid for smaller stocks prior to January, they are more inclined to buy 'broken' growth stocks.

Hence, the seasonal bias is twofold - small cap growth outperforms in December, and small cap value wins in January.


Smaller capitalized firms tend to exhibit greater swings in style rotation than do large cap stocks. The period from June 1994 to May 1996 was an excellent period for growth investing - growth funds posted a whopping 66% gain compared to a 35% gain among value investors. The period immediately following April 1996 was much more favorable for value investors. Small cap value funds generated a gain of 3% from June 1996 to April 1997. While these figures appear paltry in comparison, emerging growth funds lost on average 15% over the same time frame.


• Expectations of growth and information flow offer a better approach to modeling investment styles. Valuations appear to be secondary and only symptomatic of growth and value.

• Style investing may be defined by one's preference for risks. A growth stock investor may be more willing to accept the vagaries of the market to own a 'good' company, while a value investor may be more willing to own a firm with business concerns to avoid market risks.

• Style rotations can be especially dramatic for smaller stocks.

• Better growth and value benchmarks can allow investors to manage their style exposure more adeptly.

• Because the preferences for interpreting information and making decisions may be intrinsic and part of human nature (see accompanying box), the demarcation of style investing or growth and value may be long standing and not simply a current fancy of market participants.


The data used to develop the MLSCR Style benchmarks are based on Compustat for financial statements and I/B/E/S for forecast information. The expected growth rates are based on the average of historical sales and consensus forecast earnings growth. Both variables combined offer a better sense of a company's prospects. The analyst coverage variable is simply the number of analysts that submit a forecast for the current year.

Valuation is an open-ended topic - in many ways, finding the proper valuation model is simply work in progress. Our more recent work on valuing firms concluded that a sector-adjusted price-to-cash-flow variable appeared to best value smaller companies. Conversely, a sector-adjusted book value model generated superior results among large firms. As a result, the tests developed for large cap firms used a book value model instead of cash flow.

Small Cap Funds Analyzed

• Emerging Growth Funds Small Value Funds

• Acorn Investment TR Pennsylvania Mutual

• Kemper Small Capitalization EQ

• DFAInvestment Dimensions Group Inc.

• PBHG FDS Inc. Evergreen Micro Cap FD

• United New Concepts FPACapital FD Inc.

• Putnam OTC Emerging Growth Funds Heartland Value

• Safeco Growth FD Inc. Pioneer Mid Cap FD

• Scudder Securities TR

• Prudential Small Companies FD Inc.

• T. Rowe New Horizons FDI Babson Value FD Inc.

• Vanguard Explorer FD Inc. Winthrop Focus FDS

*Note: Babson Value replaced Royce Value and PBHG (Pilgrim Baxter) replaced Keystone American Hartwell in September 1997.

Large Cap Funds Analyzed

• Growth Funds Value Funds

• American Century Select Dreyfus Co.

• TRowe Price Growth Fund Investment Co. of America

• American Capital Fund Putnam Growth and Income

• Fidelity Destiny American Mutual

• Nicholas Fund Pioneer II

• Growth Fund of America Lord Abbett Affiliated Fund

• Smith Barney Appreciation Fund Mutual Shares

• Van Kampen American Capital Pace Washington Mutual

• GE S&S Program Vanguard/Windsor Fund


Pradhuman, Satya and Sue Crosby. Equity Style Management. Chapter 10, (Editors Klein & Lederman, Irwin, 1992).

Bernstein, Richard. Style Investing. (John Wiley & Sons, 1995).

Pradhuman, Satya 'The January Effect - Potential style Bias at the Turn of the Year' Journal of Investing, Winter 1996.

Jung, Carl. Psychological Types - The Collected Works of C.G. Jung. (Princeton University Press, 1990).

Stumpf, Stephen and Thomas P. Mullen. Taking Charge, Strategic Leadership in the Middle Game (Prentice Hall, 1992).


Find your next ETF

Reset All