The Leverage–Profitability Puzzle Resurrected
B Espen Eckbo, Michael Kisser
Review of Finance, Volume 25, Issue 4, July 2021, Pages 1089–1128, https://doi.org/10.1093/rof/rfaa032
The leverage-profitability puzzle is the tension between the positive leverage-profitability correlation predicted by classical tradeoff theory of capital structure choice and the robust negative correlation observed for industrial companies. In this paper, we attempt to resolve this tension by exploiting quarters in which firms undertake large leverage rebalancings by issuing debt and distributing the proceeds to shareholders. Dynamic tradeoff theory holds that fixed debt-issuance costs cause firms to optimally refrain from this type of rebalancing until the benefit covers the transaction cost. While waiting, positive profitability shocks drive down leverage, which may be the driving force behind the negative unconditional leverage-profitability correlation in the data. When actively rebalancing, however, firms move close to their optimal leverage ratios, which is when the theory implies a positive leverage-profitability correlation.
Our main finding is that the leverage-profitability correlation is significantly negative also in periods when firms undertake large capital structure rebalancings financed by debt issues. This finding, which fails to support dynamic tradeoff theory, is robust to variations in the debt-issue size-threshold used to define the rebalancing event, and to the specific definition of leverage and profitability. Moreover, while the class of dynamic inaction models treats investment financing as exogenous, we show that the conditional leverage-profitability correlation is negative also after controlling for the level of same-period investments. In sum, our evidence rejects the notion that the negative unconditional leverage-profitability correlation empirically overrides a positive correlation in periods with significant rebalancing events.
We also contribute by clarifying Danis, Rettl and Whited (2014) ‘s finding of a positive conditional leverage-profitability correlation, which they suggest is “good news for the class of dynamic trade-off theories in which cash flows are exogenous and leverage adjustments are infrequent”. We show that their positive correlation estimate is caused by their inclusion of internally financed rebalancing events – events financed by cash draw-downs rather than by new debt issues. This inclusion is problematic for their empirical test because events financed by cash draw-downs are neither costly nor do they increase the debt-related corporate tax shield determining the leverage target. With zero financing cost, there is no rebalancing inertia, and hence those events do not inform on the dynamic tradeoff theory being tested (which itself relies on costly adjustment).
Furthermore, we show that cash-financed events are associated with a positive leverage-profitability correlation only when the firm’s pre-event cash-balance exceeds an estimated cash target. Since it is well known that firms tend to disgorge excess cash following high profitability, this further suggests that the positive leverage profitability correlation associated with internally financed rebalancing events has more to do with managing cash balances than leverage targets. Of course, irrespective of the true role of internally-financed rebalancing events, our evidence of a negative conditional correlation across debt-financed rebalancings cannot be reconciled with dynamic trade-off theory of capital structure. Thus, our results effectively resurrect the leverage-profitability puzzle.