Signaling Equity Performance Shifts by Size and Style

Print pagePDF pageEmail page

Mr. Richard Rush is a Vice President at PNC. He has over 24 years of financial services experience including serving as managing director at Wi Trust; divisional director at Alliance Bernstein; portfolio manager and director of research at Fox Asset Management; and national director of institutional consulting at Prudential Securities. Richard currently serves a Investment Advisor at PNC Wealth Management. He is an invited contributor at The Business Thinker, llc.

Leadership does change. Whether it’s in a car race, in a classroom or even in a country — whoever or whatever leads — eventually changes.  That’s history.

In this context, it is easy to understand then, that changes or shifts occur in all things.  By virtue of this certainty, and the ability to exploit that change, does it not warrant some important considerations in investing behaviors as well? For example, the premise that there aresignificant and sustained leadership changes in the stock market’s three primary equity capitalization groups called: Large, Medium and Small Capitalization stocks, is a given. That’s also history.  Each capitalization group has led the market, and each has trailed the market, in sustained fashion but always in random rotation.  Clearly, fertile territory exists within which to exploit this certain but random change.

Since Graham and Dodd’s pioneering work in 1984, and further refined by other academic and portfolio practitioners in the early 1990’s, a commonly held belief was that differentiators of classification existed between Value and Growth style segments.  And they, too, have rotated, or changed leadership despite the ambiguity (but mostly evolving agreement) on their respective market interpretations.
Like the late 1990’s “growth” stock rallies and the current favorable disposition to buying “value” stocks as examples; performance shifts have repetitively occurred, as demonstrated in Table I.  By consequence, performance shifts in Sizes (capitalization) and Style bias (value and growth) have become a hallmark of capital markets history, as illustrated by Table II.  Reliably anticipating when, and into what
classes these rotations will result, has been a perennial challenge.  Exploration into style momentum investing may diminish that challenge and guide exploitation of these shifts.






*The annual Value returns are represented by the Russell 3000® Value Index.  The annual Growth returns are represented by the Russell 3000® Growth Index.  These
indexes are unmanaged and cannot be invested in directly. Source: Crandall Pierce











**The annual Large Cap returns are represented by the Russell 1000® Value Index.  The annual Midcap returns are represented b the Russell Midcap Index.  The annual Small Cap returns are represented by the Russell 2000® Index. These indexes are unmanaged and cannot be invested in directly.  Source:  Crandall Pierce



While it has been historically true that class rotational swings do occur, and do so in a sustained manner, would it not be enormously advantageous for investors if one could discover how and when these shifts would occur? Would it not be distinctly favorable to anticipate from where better returns could be generated so a portfolio could be optimally composed to capture (by forecast) the best performing segments (and avoid the worst) during the next investment measurement period? Wouldn’t investors benefit if a system could be devised, repeatedly tested, and eventually implemented, that made simultaneous size and style predictions with statistically significant results?

Such are the underlying questions that predicate the research for a size and style momentum forecasting discipline.  According to the current literature to date, there has never been a recorded systematic tactic or strategy to identify and capture the best performance of the shifting size and style segments of the equity markets; only explorations of its potential.  There has never been a means to reliably forecast where and when the market segments were “about to perform” – only how they have performed.  At best, predicting how a conventional portfolio would perform was an in-exact, wish-based practice founded on an individual portfolio manager’s belief in his aggregate stocks’ positive contribution to portfolio performance. And it remains, yet, a discomforting, insufficient means to insure a portfolio compositions’ intent on out-producing selected benchmark measures.  There clearly has to be an emergence of a more reliable way.

William Sharpe, a Nobel prize winner in his seminal research in Economics and Capital Markets Theory, undertook a trailblazing effort in the concept of equity style classifications in 1992. His work explored returns-based investment management style in the context of asset allocation and the eventual optimization (maximum return for the least amount of risk) of equity portfolios.  While conventional portfolio managers had
undertaken to exploit size and style anomalies, the critical mass of academic researchers and portfolio managers combined, had achieved, at least, one consensual objective:  to group stocks into homogeneous groups based on key, fundamental performance drivers.  Until today, these classifications remain much the same as they were then.  That is, classes based on capitalization and valuation: Large Cap stocks, Mid Cap stocks, Small Cap stocks, and Growth and Value styles.

By nature, strategic allocation – the sometimes immutable proportionate exposure to differing asset classes in simultaneous combinations, tends toward greater stability of return than does its not-so-distant cousin – tactical allocation. Yet, the latter is considerably more reflective of short term performance characteristics and indeed remarkably dynamic. The advantages to both are obvious. But, so are the costs to the
implementation of either.  Strategic allocation, in its unalterable form, may miss the nuances of a rotation favoring asset classes not present or grossly underweighted in the strategic composition.  By contract, the tactical allocation may mistakenly configure a portfolio to an unintended, untimely exposure.  The risks of volatility, opportunity loss or compromise of portfolio optimization are nearly identical,but nonetheless, always present.  So what method seems to work best?



The goal is simply to optimize the weights of the six (6) size and style subsets of the Russell 3000 (the chosen benchmark) on a monthly basis.  For the practice then, the six
(6) Russell size and style benchmarks are selected, representing growth and value for large, mid and small cap stocks. (See Table III)

Table III

Small Cap

Mid Cap

Large Cap


Russell 2000® Growth

Russell Midcap®

Russell 1000® Growth


Russell 2000® Value

Russell Midcap®

Russell 1000® Value


This mirrors 98% of the investable domestic equity markets, with micro-cap stocks as its single non-represented component.  These indices are easily investable via exchange traded funds (or ETFs).  These instruments facilitate the implementation around issues surrounding a tactical size/style allocation.

First utilized is a traditional Markowitz-like, mean-variance optimization to maximize an investor’s utility.  Following the pair-wise spread strategy of Qian (2003), the portfolio of six (6) size/style indices is subsequently segmented into fifteen (15) unique pairs (see Table IV).

Table IV

Large Cap Growth

Large Cap Value

Mid Cap Growth

Mid Cap Value

Cap Growth

Small Cap Value

Cap Growth
Cap Value

Pair 1

Cap Growth

Pair 2

Pair 3

Cap Value

Pair 4

Pair 5

Pair 6

Cap Growth

Pair 7

Pair 8

Pair 9

Pair 10

Cap Value

Pair 11

Pair 12

Pair 13

Pair 14



Since empirical size/style rotation studies support the ease of forecasting spreads versus forecasting total returns, the spreads are then examined.  The forecasted weights across each of the fifteen pairs are aggregated, then, scaled to find the desired total portfolio weights.

Subsequently, factor-mimicking portfolios are created by calculating the decile return spreads on a monthly basis.  After aggregating all potential factors one is able to forecast the key fifteen (15) pair-wise spreads (Chen (200)).  The goal is to select the 10 best factors for each spread as these factors should contain critically important information for a forecasting process across multiple pairs.  Investor sentiment measures such as short-interest, and money flow appear most often.  Earnings surprise related factors also occur frequently in a majority of pairs. Predictably, momentum-type and short term factors succeeded most often. Armed with these multiple signals, one can facilitate exposures to each of the six size and style boxes on a monthly basis in an optimally fashioned portfolio.

The results are noteworthy.  After generating spread forecasts; as the risk aversion is diminished, average return increases in line with the portfolio variance. The tactical style allocation no longer mimics the lesser
returns of a constrained equal weighted portfolio. The tactical allocation actually adds incrementally (in the double-digit range) over the equal-weight style portfolio monthly, at a lower standard deviation.  A workable size/style tactical optimization is achieved on a bottom-up basis.  The portfolio is rebalanced monthly to reflect weighting alterations. Initial and subsequent excess cash from prospective investor purchases remain invested in the ETF tracking the Russell 3000, for the purpose of diminishing any negative cash affects.


The systematic size/style rotational tactic described here has many compelling advantages.  First among them: the nearly complete representation of the investable equity markets.  The selected benchmark, the Russell 3000 Index (from which the subsets were created), measures nearly all equity capitalizations with the exception of micro-caps.  Second, high liquidity and transparency, as ETFs are valued and traded intra-day like individual stocks.  Third, elimination of significant variations related to stock weighting and stock selection factors as the entire equity segment is bought.  Fourth, the application of anticipation-signals for monthly shifts across size and style groupings. This dictates where and when the portfolio must be re-allocated to capture the best, and avoid the worst, of the market.  And last, the resulting attractive information ratios, active return products, as well as, extraordinarily competitive up and down market capture ratios versus the portfolio return universe.

Style momentum investing is emerging as a winning discipline in today’s capital markets. Over complete market cycles, this iteration of quantitative signaling shows promise in providing attractive risk-adjusted returns. Doubtlessly, improvements and alternative strategies will emerge as few conventional portfolios seem to offer the potential reward that this tactic does.  As embryonic as this composition appears, at minimum, it portends to provide a foundation that fosters subsequent exploration.  How exciting it should be to render an already efficacious application of signaling size and style
rotations even better.



Ahmed, Parvez, Larry Lockwood &
Sudhir Nanda.  “Multistyle Rotation
Strategies.”  The Journal of Portfolio
Management, 2002. Chen, Hsiu-Lang. “On Characteristics Momentum.”  Working Paper, 2002. Graham, Benjamin &
David Dodd, Security Analysis, The Classic 1934 Edition, 1934. Markowitz,
Harry.  “Portfolio Selection: Efficient
Diversification of Investments.”  Journal
of Finance, 1959. Qian, Edward.
“Tactical Asset Allocation with Pairwise Strategies.” The Journal of
Portfolio Management, 2003. QSG. “A Stylistic Investigation of Money Flow.” QSG
Investment Insights, 2003. Sharpe, William. “Setting the Record Straight on
Style Analysis.”  Dow Jones Fee Advisor,


Leave a Reply

Your email address will not be published. Required fields are marked *