Does Asset Growth Predict Stock Returns?

Summary

• Over the long term, low-investment firms have outperformed high-investment firms.
• There’s a significantly positive relationship between the current year’s asset growth and future years’ crash risk, with the predictive power lasting up to three years.
• Higher asset growth is also related to lower future profitability.
• Growth in current assets, operating liabilities and retained earnings have the strongest predictive power for future crash risk.
• Combining value and profitability.

Over the long term, low-investment firms have outperformed high-investment firms. This finding has led to the investment factor (CMA, or conservative minus aggressive) being incorporated into the leading asset pricing models - the four-factor Q model (market beta, size, investment and profitability), the Fama-French five-factor model that adds value, and the Fama-French six-factor model that adds momentum.

As shown by Kewei Hou, Chen Xue and Lu Zhang in their paper “ Digesting Anomalies,” firms with lower discount rates (lower costs of capital and thus lower expected returns) invest more. Firms with higher discount rates (higher costs of capital and thus higher expected returns) face higher hurdles for investment and thus invest less. In other words, investment predicts returns because, given expected profitability, high costs of capital imply low net present value of new capital and low investment, and low costs of capital imply high net present value of new capital and high investment. Thus, all else equal, firms with higher investment should earn lower expected returns than firms with lower investment.

In addition, valuation theory predicts that controlling for a firm’s market value and expected profitability, a company that must invest heavily to sustain its profits should have lower contemporaneous free cash flows to investors than a company with similar profits but lower investment. This is what Eugene Fama and Ken French found in their 2006 paper “Profitability, Investment and Average Returns.” They also found that while there is not a direct way to measure future investment, recent asset growth is a reliable proxy for expected investment, allowing them to measure the effect.

Siu Kai Choy, Gerald Lobo and Ying Zheng contribute to the literature on asset growth with their November study “Asset Growth and Stock Price Crash Risk.” They investigated the relationship between asset growth and stock price crash risk, asking: “Does asset growth relate to crash risk and, if so, what is the mechanism?” They began by noting: “Stock price crash has drawn considerable attention from researchers and regulators. A strand of academic research attributes stock price crash to managers’ incentive to hide bad news for reasons such as career concerns and continued extraction of private benefits. Managers’ tendency to cover up bad news, hoping that the situation will improve, results in accumulated bad news so that stock price crashes when the bad news is eventually released. Research finds that this incentive to delay the release of bad news is related to earnings opacity, corporate tax avoidance, equity incentive, CEO overconfidence, regional religiosity, and other corporate and reporting characteristics.”

They noted that the “corporate finance literature suggests that managers have an incentive to empire-build and control more resources by stockpiling projects and making investments.” Their hypothesis was that the “combined effect of asset growth from empire-building and the tendency to conceal bad news leads to stock price crash. Our reasoning is that managers can extract more private benefits by controlling more resources through means such as raising capital and making capital investments, or delaying contraction by keeping inefficient operations for too long. Together with managers’ incentive to cover up private bad news, the bad news is accumulated up until a tipping point, when it is released to investors at one time and results in stock price crash.”

Their data sample covers the period 1988 through 2017. They measured stock price crash risk using the three crash risk measures: (1) an indicator for whether one or more of the firm’s weekly returns is at least 3.2 standard deviations below the mean; (2) the negative of the coefficient of skewness of weekly returns; and (3) the natural logarithm of the ratio of down-market to up-market weekly return volatility. “The intuition for the crash risk variables is that a sudden drop in stock price may result in either extreme negative returns captured by the indicator variable for a substantial price drop or asymmetry in the return distribution captured by the skewness and return volatility ratio measures.” Total asset growth is the year-over-year percentage change in total assets. Following is a summary of their findings:

• Next year’s probability of a sizable price drop increases monotonically from 13.6 percent for the lowest asset growth quintile to 20.0 percent for the highest asset growth quintile.
• There’s a significantly positive relationship between the current year’s asset growth and future years’ crash risk, with the predictive power lasting up to three years.
• Higher asset growth is also related to lower future profitability.
• Growth in current assets, operating liabilities and retained earnings have the strongest predictive power for future crash risk. Other components such as debt and stock financing also are significant in predicting future crash risk.
• Going from the 25th to the 75th percentile of size increases the probability of CRASH by 2.2 percent.
• The majority of sudden price declines happen outside earnings announcement windows.
• The results hold during the financial crisis and for different sample periods.
• Growth in retained earnings contributes to crash risk - dividends/stock buybacks reduce it.

An interesting finding was that there was a negative relationship between leverage and future crash risk. This could be attributed to the possibility that more crash-prone firms choose less debt. Alternatively, investors could underprice highly leveraged firms, making a crash less likely.

Another important finding was that accounting conservatism can counteract managers’ tendency to withhold bad news and thus improve investors’ welfare by reducing crash risk. The authors explained: “Firms with conditionally conservative accounting practices are likely to recognize bad news faster than good news, resulting in sudden large negative earnings relative to cashflows from assets being written down, accumulation of non-operating accruals, and earnings reflecting bad news in returns in a timelier manner.”

The authors also noted that a CEO’s power to influence decisions increases with tenure in the position. Therefore, a longer-tenured CEO has more incentive to empire-build and enjoy more compensation when the firm size increases. This agency risk leads to an increase in crash risk. On the other hand, if the CEO plans to leave in the near future, there is less incentive to grow assets, and agency risk declines. In addition, an increase in free cash flow increases agency risk. Their findings confirm these hypotheses.

Choi, Lobo and Zheng concluded: “These results demonstrate that the asset growth-crash risk relationship is pervasive across components of asset growth and across different sources of financing of that growth." They added: “The results for total asset growth and its components are consistent with our hypothesis that managers’ incentives to empire-build and hide bad news contribute to stock price crash risk.”

Summarizing, their results demonstrate that asset growth positively and significantly predicts future price crash for up to three years. They provide evidence consistent with the empire-building and bad news hoarding explanation, with high asset growth being related to poor future firm performance in terms of profit margin and return on assets, suggesting overinvestment or operating inefficiency. And finally, firms with more agency problems have incrementally higher crash risk related to asset growth, while firms with more conditionally conservative accounting practices have incrementally lower crash risk.

We now turn to another recent study that looked at asset growth and returns around the globe. Its findings are consistent with those we have reviewed so far.

Further Evidence

The research team at Dimensional contribute to the literature on the investment factor with their October 2019 paper “Investment and Expected Stock Returns.” Their data sample covers the period July 1974 through December 2018 in the U.S., January 1990 through December 2018 in developed ex-U.S. markets, and January 1994 through December 2018 in emerging markets. Following is a summary of their findings:

• The top asset growth quartile tends to have higher asset growth relative to the rest of the market one, three and five years into the future within small and large caps, and in the U.S., developed ex-U.S. and emerging markets.
• In general, small caps exhibit a larger dispersion of asset growth than large caps. This is driven mainly by differences between the top investment quartiles.
• As predicted by valuation theory, in large caps, spreads in annualized compound returns between the top and bottom quartiles are negative in all three markets. However, the return spreads are not reliably different from zero - the results are weak.
• In the case of small caps, in the U.S., while there was similar performance for the bottom three quartiles (annualized compound returns of 14.3 percent to 16.5 percent), the annualized compound return for the top quartile is substantially lower (7.2 percent). And the difference in average monthly returns between the bottom and top quartiles is 50 basis points and is reliably different from zero. The results for developed ex-U.S. and emerging markets are similar. In all three regions, there is a reliably positive investment premium that is driven primarily by the significant underperformance of small-cap firms with high asset growth.
• The results are not only pervasive around the globe but also persistent. For example, the top asset growth decile underperforms the rest of the small-cap market 84 percent of the years in the U.S. market, 72 percent of the years in developed ex-U.S. markets, and 76 percent of the years in emerging markets.
• The underperformance of the top decile firms persists, on average, for about two years after sorting.
• The historical investment premium is positive across different sectors in the U.S., developed ex-U.S. and emerging markets.
• There are three ways in which firms can raise capital to grow their assets: issue equity, issue debt, or retain earnings. All three contribute to the investment factor. Equity issuance is the most prominent individual driver of high asset growth among U.S. small caps, followed by debt issuance.
• The growth in both physical and intangible capital contributes to the investment effect.
• U.S. high asset growth firms have poor historical returns regardless of whether firms with merger and acquisition (M&A) activity are included or excluded, suggesting M&A activity is not the only driver of the investment effect.

The authors concluded: “The persistent and pervasive nature of the underperformance suggests that investors might benefit from incorporating investment in their equity strategies.”

Conclusion

Valuation theory predicts that expected investment is negatively related to expected returns, holding all else fixed. Using current asset growth as a proxy for expected investment, high asset growth/high investment growth firms tend to underperform the market. Providing confidence that the findings are not the result of data mining, the evidence is persistent across time and pervasive across the globe and sectors. And, as has been found to be the case with other factors (such as value, momentum and profitability), small-cap firms are the primary driver of this underperformance. In addition, it is present across the relative price and profitability segments.

The conclusion that can be drawn is that an efficient way to improve the expected performance of an equity strategy might be to systematically exclude small-cap firms with high asset growth, especially in the case of small-cap stocks.

Analyst’s Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.