In their seminal study, Jegadeesh and Titman (1993) document the existence of the momentum effect in stock returns. Using data from 1965 to 1989, they show that when stocks are ranked in deciles based on their immediate past returns, the past winners (the best performing stocks in upper tenth decile) continue to outperform the past losers (the worst performing stocks in lower first decile) over the next several months. Jegadeesh and Titman (2001) confirm that momentum profits persist for the U.S. market in the period of 1990–1998 following their initial sample period. Even though the academic literature agrees in general on the existence of the momentum effect in stock returns, there are significant debates in the literature on what causes it and how it is generated.
The two main explanations for the momentum effect provided in the literature are the behavioral and rational risk-based explanations. According to the behavioral explanation, momentum profits result from investors’ irrational underreaction to firm-specific information or delayed overreaction. In turn, according to the rational risk-based explanation, momentum profits are realizations of risk premiums because winner stocks are riskier than loser stocks. In any event, the existing literature does not have a definite answer as to how the momentum effect in stock returns is generated. In their review article on the momentum literature, Jegadeesh and Titman (2011, p. 507) conclude that “financial economists are far from reaching a consensus on what generates momentum profits, making this an interesting area for future research.”
In our recent study, we contribute to the above debate by providing an explanation on how the momentum effect is generated by linking it to corporate investment. We show, both theoretically and empirically, that the momentum effect in a firm’s stock returns is likely to be generated by a series of information exchanges between stock market investors and firm insiders regarding the firm’s investment opportunities. Given that these investment opportunities are rather uncertain both for outside investors as well as firm insiders, to be able to invest in these investment opportunities and realize their value, firm insiders need to reduce the uncertainty by gathering additional information about these opportunities from outsiders. This is important because, first, outsiders may possess information about the firm’s investment opportunities that insiders do not have (which may be useful in pricing such investment opportunities) and second, the firm needs to assess how outsiders value the firm’s investment opportunities since this valuation will affect the firm’s stock price as it moves to invest in these opportunities. While outside investors usually do not communicate with the firm directly, they can convey their information about the firm’s investment opportunities to firm insiders by bidding up the firm’s stock price. The extant literature, both theoretical and empirical, has shown that the information conveyed in stock prices affects many corporate decisions including corporate investment.
We also design an implementable momentum trading strategy making use of firms’ predicted investment and demonstrate that this trading strategy generates superior returns compared to a simple momentum trading strategy without investment.
Theory and hypotheses
In our theoretical setup, in the first stage of the abovementioned series of information exchanges between outside investors and firm insiders, as outsiders learn about the firm’s new investment opportunities, they start to incorporate the value of these investment opportunities into the firm’s stock price. Outsiders, however, bid up the firm’s stock price only by a portion of what they assess the value of these investment opportunities to be given the uncertainty regarding whether the firm will invest in these opportunities. In other words, if outsiders assess the value of a firm’s investment opportunities to be V and they expect these to be realized with some probability p, then the expected value of these investment opportunities for outsiders will be V´p, and this is the amount by which outsiders will bid up the firm’s stock price initially. In the second stage, after they observe an initial increase in their firm’s stock price, firm insiders use the information conveyed in the stock price to update their own information about their firm’s investment opportunities and decide on the amount of capital to invest. Finally, in the third stage, outsiders use the information conveyed by the firm’s decision to invest (and how much to invest) to update their own information set about the firm’s investment opportunities and to bid up the remaining value of these opportunities into the stock price (the value that has not been incorporated in the stock price in the first stage). As a result of these information exchanges between firm insiders and outsiders regarding the firm’s investment opportunities, positive momentum in stock returns tends to be generated.
Negative momentum in stock returns is likely to be generated in a similar fashion through a series of information exchanges between outside investors and firm insiders when the firm’s investment opportunity set contracts and the firm must disinvest capital. Again, there is a certain degree of uncertainty associated with the potential loss in the firm’s value due to such contraction in its investment opportunity set. Thus, in the first stage of this series of information exchanges, outsiders bid down the firm’s stock price by a portion of what they assess the loss in the firm’s value to be. In the second stage, firm insiders use the information conveyed in their firm’s stock price to assess the magnitude of the loss and use that updated information to determine the amount of capital to disinvest. Finally, in the third stage, outsiders use the information conveyed by the firm’s decision to disinvest (and how much to disinvest) to update their own information about the loss in the firm’s value and to bid the stock price further down.
Our theoretical setup generates several testable predictions. First, we expect past winners (losers) to significantly increase (decrease) their net investment after realizing positive (negative) returns over the recent past. We define net investment as the sum of capital expenditures and acquisitions net of sale of property normalized by assets. Second, the more (less) past winners invest in response to their past positive returns, the greater is the likelihood for them to realize positive (negative) returns in subsequent months. Finally, the more (less) past losers disinvest in response to their past negative returns, the greater is the likelihood for them to realize negative (positive) returns in several subsequent months.
Empirical findings
We empirically test our hypotheses using data on a large sample of firms in the period of 1984–2017. We find the following: first, as expected, past winners (losers) respond to their positive (negative) stock returns by significantly increasing (reducing) their net investment over the next four quarters. Second, as expected, we find that the more (less) past winners invest, the more likely they are to realize positive (negative) returns in subsequent periods (up to 12 months). Finally, as expected, we find that the more (less) past losers disinvest, the more likely they are to realize negative (positive) returns in subsequent periods (up to 12 months). Overall, these findings indicate that there is a significant link between corporate investment and the momentum effect in stock returns, and the latter is likely to be generated as a result of a series of information exchanges between firm insiders and outside investors regarding a firm’s investment opportunities.
It is worth noting that the existing literature documents lower stock returns following increases in firm investment. Our theoretical prediction and empirical finding of a positive association between corporate investment and subsequent returns of past winners are at odds with this literature. However, while the existing studies analyze the effect of investment on long-term returns (between one to three years) in all firms regardless of their past returns, our theoretical prediction is about the (contemporaneous) effect of investment on short-term returns (up to a year) in a subset of firms only, namely, those firms that have realized significant positive returns in recent past. Our theoretical setup does not preclude a situation where a firm realizes negative returns as a result of an increase in its investment level if, for example, such a firm does not experience an expansion in its investment opportunity set prior to investing, or the firm significantly underinvests or overinvests contrary to the stock market investors’ expectations. Thus, our findings in this study are novel, as we show that an increase in net investment is likely to be associated with positive short-term returns if such increase in net investment follows significant positive stock returns, and the magnitude of such increase in net investment is in line with the stock market investors’ expectations.
Momentum trading strategies using predicted net investment
We provide further evidence on the link between corporate investment and the momentum effect in stock returns by designing an implementable momentum trading strategy which makes use of firms’ predicted net investment. Given that our theoretical setup predicts that past winners with larger (smaller) increases in net investment are likely to realize positive (negative) returns in subsequent periods, and that past losers with larger (smaller) decreases in net investment are likely to realize negative (positive) returns in subsequent periods, we conjecture that a trading strategy which buys past winners with a positive predicted change in net investment over the next quarter and sells past losers with a negative predicted change in net investment over the next quarter should generate superior returns compared to a simple momentum trading strategy which buys all past winners and sells all past losers regardless of their predicted changes in net investment. We make use of an estimation model by Beatty, Riffe, and Welch (1997) to predict next quarter’s net investment and use it in our trading strategy. In this model, we use four lags of external equity financing, stock returns, internal accounting information (income, dividends, changes in cash, changes in inventory, and depreciation), taxes, net investment itself, several firm size variables, as well as financial market variables (returns on the stock market, on short-term T-bills, and on long-term T-bonds), and industry dummies to predict next quarter’s net investment.
We estimate this model separately for each fiscal quarter using all data available for that fiscal quarter. In particular, for each fiscal quarter we use data from September 1983 (the first fiscal quarter with available data in our sample) up until the fiscal quarter for which we estimate the model and then using the coefficient estimates of our regressions we predict the net investment for the subsequent fiscal quarter. Next, we generate predicted changes in net investment by subtracting current quarter’s net investment from next quarter’s predicted net investment. Finally, we identify past winners and past losers on the day of these firms’ quarterly earnings reports (which allows us to observe current quarter’s accounting information and use it in our estimation model) by allocating them in quintiles based on their stock returns in the immediate past 3, 6, 9, or 12 months including the day of the earnings report. On the next day after the earnings report, we buy past winners with positive predicted changes in net investment and sell past losers with negative predicted changes in net investment and hold these portfolios over the next 3, 6, 9, or 12 months. We show that this trading strategy generates significantly superior profits compared to a simple momentum trading strategy which buys all winners and sells all losers regardless of their predicted net investment. In particular, we show that the average monthly returns of this trading strategy exceed the corresponding average monthly returns of a simple momentum trading strategy by 0.17% to 0.53% (depending on a particular portfolio holding period), which is equivalent to 2.04% to 6.36% in simple annualized terms. Our findings on this trading strategy also show that, in general, past winners with positive predicted changes in net investment earn higher returns compared to all past winners as a group, and past losers with negative predicted changes in net investment earn lower returns compared to all past losers as a group.
Further, given the predictions of our theoretical setup, we design two additional trading strategies using predicted changes in firms’ net investment. One of these trading strategies buys past winners with a positive predicted change in net investment and sells past winners with a negative predicted change in net investment, and the second strategy buys past losers with a positive predicted change in net investment and sells past losers with a negative predicted change in net investment. We show that these two trading strategies generate positive returns which, in general, are larger than those generated by a simple momentum trading strategy. Overall, our findings on the three trading strategies described above provide further support for the existence of a significant link between corporate investment and the momentum effect in stock returns.
Contributions of our study and conclusion
Our study makes several important contributions to the literature. First, while the extant literature on the momentum effect has focused on various behavioral and rational risk-based explanations, our study links the momentum effect to corporate investment. We show that the feedback relationship between the firm and financial markets generates the momentum effect. Second, we investigate corporate investment decision in the context of the momentum effect. While many existing studies have analyzed separately the effect of past stock returns on various corporate variables (corporate investment, in particular) and the effect of various corporate variables on future stock returns, we study how past stock returns through their effect on corporate investment are related to future stock returns. Third, we contribute to the literature on momentum trading by showing that momentum trading can be more profitable if buy and sell portfolios are constructed using a firm’s predicted net investment. This provides further evidence on the link between corporate investment and the momentum effect. Finally, our study adds to the literature on managerial ability to respond to market developments. In this paper we show that stock market developments contain valuable information that managers may use when making corporate decisions: we show that past stock return performance is an important determinant of subsequent corporate investment, which, in turn, affects future stock returns.
Ki C. Han is a Professor of Finance at Suffolk University
Abu Jalal is a Professor of Finance at Suffolk University
Karen Simonyan is an Associate Professor of Finance at Suffolk University
This post is adapted from their paper, “Corporate Investment and Stock Return Momentum” available on SSRN.