Tradeoff Theory and Leverage Dynamics of High-Frequency Debt Issuers
B Espen Eckbo, Michael Kisser
Review of Finance, Volume 25, Issue 2, March 2021, Pages 275–324, https://doi.org/10.1093/rof/rfaa018
Do public industrial firms manage leverage towards a long-run target? We test three core predictions of dynamic tradeoff theory using a hitherto unexplored set of US public industrial companies: high-frequency net-debt issuers (HFIs). HFIs raised the bulk of all public and private debts (net of debt retirements) over the past three decades.
The HFI selection algorithm tends to exclude firm-quarters with extremely high leverage—periods in which outright firm survival is likely most important. Also, by selection, the fraction of firm-quarters with active leverage adjustments is much higher for HFIs than for the remaining population of public industrials.
The combination of likely high debt financing benefits (given the firm characteristics) and low debt issuance costs (given the high net-debt issue frequency), makes HFIs ideal test assets for central tradeoff hypotheses to make predictions about leverage dynamics. If HFIs do not manage leverage towards a target, then who does?
We formulate and test three dynamic tradeoff hypotheses, all of which rely on the presence of debt-issuance costs to explain capital structure rebalancing inertia. The first (H1) predicts that more profitable HFIs will move to a higher leverage ratio only in quarters when they actively rebalance capital structure upward (by issuing debt and paying the proceeds to shareholders). In passive quarters, where debt-issuance costs deter rebalancing, the positive drift in profitability (which is exogenous to capital structure decisions) mechanically causes the leverage-profitability correlation to be negative. Importantly, tests of H1 do not require an estimate of the target leverage ratio itself. Notwithstanding that the HFIs provide particularly powerful test assets for this conditional cross-sectional prediction, H1 is rejected by the data.
The second hypothesis (H2) exploits the prediction that, as firms manage leverage towards a target, HFIs will reverse target-leverage deviations faster than low-frequency net-debt issuers or LFIs (higher speed-of-adjustment or SOA). Notwithstanding long spells between debt issues by LFIs, SOA estimates of HFIs and LFIs are statistically indistinguishable. This finding suggests that the SOA estimates found in the literature are driven much more by variation in equity values (the denominator of the leverage ratio) than active debt issuances.
The third and final hypothesis (H3) integrates investment finance decisions into the leverage dynamics of HFIs. Part (1) of H3 predicts that SOA increases in periods with particularly low marginal adjustment costs (low leverage and high investment) or high adjustment benefits (high leverage and low investment). Part (2) holds that investment-driven debt issues by already over-leveraged firms are transitory—they should be repurchased when investment-financing needs abate. We find little evidence to support either the two parts of H3, which, again, is problematic for the standard tradeoff model.