The Asset Growth Anomaly: Why Rapidly Expanding Firms Underperform

20 Percentage Points Per Year

Between 1968 and 2003, U.S. firms with the most aggressively expanding balance sheets delivered annual returns roughly 20 percentage points lower than firms whose total assets barely changed. The spread was not a small statistical curiosity. It was large enough to be economically transformative, persisted after controlling for market beta, size, and book-to-market ratios, and appeared in virtually every subperiod examined. The firms that spent the most acquiring assets, expanding capacity, and growing their balance sheets were precisely the ones equity investors should have avoided.

This was the central finding of Cooper, Gulen, and Schill (2008), published in the Journal of Finance. Their paper elevated asset growth from a niche accounting metric to a first-order signal in the cross-section of expected returns. Understanding why expansion predicts underperformance — and how robust that prediction is — remains essential for anyone constructing evidence-based equity strategies.

Measuring Asset Growth

The construct Cooper, Gulen, and Schill tested was deliberately simple. Asset growth is the year-over-year percentage change in a firm's total assets from the balance sheet. If a company had $1 billion in total assets at the end of last fiscal year and has $1.3 billion this year, its asset growth rate is 30%.

This simplicity was a feature, not a limitation. Prior research had examined individual components of investment: capital expenditures, acquisitions, changes in working capital, share issuances. Each showed some return predictability in isolation. What Cooper, Gulen, and Schill showed was that aggregating all of these into a single balance sheet metric — total asset growth — captured return predictability more powerfully than any individual component examined separately.

They sorted all NYSE, AMEX, and NASDAQ common stocks into deciles each year based on prior fiscal-year asset growth, then tracked value-weighted returns for the following twelve months. The spread between the highest and lowest deciles was striking:

The numbers above capture value-weighted portfolio returns in raw form. After applying standard factor adjustments using the Fama-French three-factor model, the abnormal return (alpha) from a long-low-growth, short-high-growth portfolio was approximately 8% per year — still economically large, and statistically significant with a t-statistic exceeding 4.

Why Does Expansion Predict Underperformance?

Three classes of explanation have emerged in the literature, and they are not mutually exclusive.

The Overinvestment Hypothesis

The behavioral story builds on a fundamental principal-agent tension in corporate finance. Managers of growing firms often have incentives — compensation structures, empire-building preferences, career concerns — that lead them to invest beyond the level that maximizes shareholder value. Jensen (1986) formalized this logic in his free cash flow hypothesis: when firms generate more cash than their profitable investment opportunities can absorb, managers tend to channel the excess into value-destroying acquisitions, capacity expansions, or other asset accumulation. Shareholders do not fully anticipate the extent of this overinvestment at the time assets are acquired. When subsequent earnings fail to justify the balance sheet expansion, prices adjust downward.

The overinvestment mechanism predicts several testable features. It should be stronger among firms with high free cash flow but limited monitoring (smaller boards, dispersed ownership, weaker governance). It should be associated with acquisitions as well as organic capital expenditure. And the underperformance should be concentrated in the years immediately following the asset expansion, as the market gradually learns that returns on the new assets fall short of expectations.

Cooper, Gulen, and Schill found that the return predictability from total asset growth was indeed largely symmetric — high growth predicts underperformance, not just low growth predicting outperformance — which is more consistent with the overinvestment story than with pure risk-based explanations.

The Q-Theory Channel

The rational, risk-based alternative draws from neoclassical investment theory. In the q-theory framework developed by Tobin and later formalized by Xing (2008), firms invest up to the point where the marginal cost of capital goods equals their marginal value. Firms with high asset growth are firms that recently faced strong investment opportunities — they were endowed with high expected returns on capital and therefore expanded aggressively. But high investment is equilibrium behavior in a world of diminishing returns: as firms deploy capital, the marginal productivity of additional assets falls. Expected future returns decline not because of mispricing but because the firm's risk profile has shifted.

Under this view, asset growth is a proxy for the firm's position in the investment cycle. High-growth firms have already harvested their best opportunities and therefore face lower expected returns going forward. Low-growth firms are capital-constrained or have few opportunities, suggesting higher expected returns in the future if conditions improve.

The q-theory channel received indirect support when Fama and French (2015) incorporated an investment factor (CMA, Conservative Minus Aggressive) into their five-factor model. CMA goes long firms with low investment rates and short firms with high investment rates — essentially capturing the asset growth spread through a different construction. The inclusion of CMA substantially absorbed the anomalous returns from various investment-based strategies in their model.

Limits to Arbitrage

A third mechanism explains not why the return pattern exists but why it persists despite investor awareness. High asset growth firms tend to be larger, liquid, and well-covered by analysts — characteristics that should facilitate arbitrage. However, Li, Livdan, and Zhang (2009) argued that the anomaly is partly sustained because the short side — rapidly growing firms — is disproportionately expensive to short. These firms often have high institutional ownership and limited short interest capacity at the margin. Even if sophisticated investors know high-growth firms are expensive, executing the short position at scale is costly.

Additionally, the return predictability from asset growth is distributed over multiple years, which means any single-year trading signal is modest in magnitude. Arbitrageurs face holding period risk: the mispricing may deepen before it corrects, creating drawdown risk that dissuades systematic shorting.

Decomposing the Components

One of the most important contributions of Cooper, Gulen, and Schill was demonstrating that the total asset growth effect was more than the sum of its parts. They broke down asset growth into its constituent balance sheet changes and tested each separately:

Operating accruals (changes in working capital items like receivables and inventory) showed return predictability consistent with the accruals anomaly documented by Sloan (1996). Non-operating accruals (changes driven by acquisitions and long-term investment) showed comparable predictability. Financing activities including equity issuances and debt increases also predicted lower future returns.

The central finding was that all components contributed, and the composite total asset growth measure unified them into a single, more powerful signal than any individual piece. This decomposition suggested that the return predictability from asset growth was not driven by one specific agency problem or one accounting artifact — it reflected something more fundamental about the economics of corporate expansion.

Titman, Wei, and Xie (2004) had earlier documented that capital expenditure growth independently predicted future return underperformance, interpreting the result through an agency lens. Cooper, Gulen, and Schill showed that capital expenditure was one piece of a larger phenomenon. Acquisitions, organic growth, and financing-driven balance sheet changes all told a similar story.

International Evidence

A natural test for any anomaly is whether it survives outside the original sample. Watanabe, Xu, Yao, and Yu (2013) examined asset growth in a broad international dataset covering 40 markets from 1968 to 2008. Their findings confirmed that the anomaly was not a U.S.-specific artifact.

Asset growth negatively predicted future returns in the majority of markets studied. The effect was present in both developed and emerging markets, though with meaningful variation in magnitude. Countries with stronger investor protection, more developed financial markets, and greater analyst coverage tended to show smaller return spreads from asset growth, consistent with the interpretation that better information environments allow faster price correction.

The international evidence also shed light on the competing explanations. If the anomaly were purely rational risk compensation, it should be of similar magnitude across markets with similar risk characteristics. The fact that the spread was larger in markets with weaker governance and less institutional ownership tilted toward the overinvestment and limits-to-arbitrage stories rather than pure q-theory.

Relationship to the Value and Quality Factors

The asset growth anomaly does not exist in isolation. It connects to several other well-documented return patterns, which is why understanding it enriches the broader factor model landscape.

The link to value investing is non-trivial. High-growth firms tend to trade at high valuation multiples — the market prices their growth prospects richly. When that growth disappoints or the assets turn out to be worth less than the market anticipated, prices decline. This is the same mechanism that drives the value premium: the market systematically overvalues glamour (high-growth, high-multiple) firms relative to value (low-growth, low-multiple) firms. Asset growth and valuation ratios capture partially overlapping information about investor overoptimism regarding corporate expansion.

The connection to quality also runs deep. Factor investing research has found that firms combining low asset growth with high profitability — the intersection of the investment and profitability factors — generate particularly strong risk-adjusted returns. The Fama-French five-factor model formalizes this by including both a profitability factor (RMW, Robust Minus Weak) and CMA as distinct dimensions of the cross-section. Empirically, these two factors are negatively correlated: profitable firms tend to be conservative investors, while rapidly growing firms often expand into lower-quality projects over time.

Post-Publication Decay and Current State

Academic publication of anomalies tends to weaken them as market participants incorporate new findings into their trading. The evidence on the asset growth anomaly's post-publication trajectory is mixed.

For large-cap stocks, where institutional ownership is high and arbitrage is relatively cheap, the spread from asset growth has narrowed meaningfully since the early 2000s. Smart-beta funds tracking investment-based factors provide systematic exposure to the anomaly, pushing prices closer to fair value.

For smaller, less-liquid stocks, the anomaly has been more persistent. Arbitrage costs remain high for small companies, and the overinvestment mechanism may be more acute in firms with weaker governance and less analyst scrutiny.

McLean and Pontiff's (2016) comprehensive analysis of published anomaly decay found that return predictors documented in academic papers typically lose roughly one-third of their in-sample magnitude after publication. The asset growth effect appears broadly consistent with this pattern. The anomaly has not disappeared, but it is smaller than Cooper, Gulen, and Schill originally measured — particularly among the large, liquid names that institutional investors can actually trade efficiently.

Methodological Considerations

Any analysis of asset growth as an investment signal needs to account for several practical complications.

Timing: Asset growth is measured using fiscal year-end balance sheets, which are not immediately available to investors. Accounting for the publication and filing lag (typically four to six months after fiscal year-end) is necessary to avoid look-ahead bias. The returns reported in Cooper, Gulen, and Schill used a conservative lag structure that assumed investors acted on publicly available data only.

Sector concentration: Industries with capital-intensive business models (utilities, real estate investment trusts, infrastructure) naturally have higher asset growth during expansion phases. Raw asset growth comparisons across sectors can be misleading. Sector-neutral implementation — ranking stocks against peers within the same industry — typically produces more refined signals.

Scale effects: Total assets include both operating and financial assets. For firms with large financial subsidiaries or investment portfolios, total asset growth can reflect changes in financial market values rather than genuine operating investment. Adjustments or subsample analyses that exclude financial firms are common in practitioner applications.

The Broader Lesson

The asset growth anomaly sits at a productive intellectual intersection. It can be told as a story about managerial overconfidence and empire-building. It can be told as a story about rational diminishing returns to capital in a q-theory world. It can be told as a story about limits to arbitrage preventing timely price correction.

What makes the finding valuable is that it does not require choosing between these explanations. For a practical investor, the signal is actionable regardless of its theoretical source: companies that expand their balance sheets aggressively have historically delivered weaker returns to shareholders, and that pattern has shown enough persistence across decades, international markets, and factor model controls to merit systematic incorporation into equity analysis.

Whether one views it as compensation for a poorly specified risk factor, punishment for managerial excess, or slow correction of investor optimism about growth, the data point in a consistent direction. Rapid expansion is not, on average, a friend to equity holders.