The new lease standard (ASC 842) allows firms to keep variable leases off balance sheet, in part based on the assumption that future expenses are difficult to estimate reliably. We show that variable lease expenses are both prevalent and substantial, exhibiting persistence and predictability comparable to operating lease expenses while showing limited sensitivity to revenue changes. These patterns are consistent with variable lease payments being based on stable drivers. Following ASC 842 adoption, firms report lower minimum operating lease commitments and higher variable lease expenses, suggesting a substitution from operating to variable leases. Neither equity betas nor credit ratings reflect potential variable lease liabilities. Conservative estimates show that recognition of variable lease liabilities would increase debt by 7.1% on average. Our findings provide evidence on the properties of variable leases and the potential implications of keeping them off balance sheet.

The choice between historical cost and fair value measurement is one of the most debated issues among accounting academics and practitioners. We use the election of the fair value option (FVO) to study the effects of entities’ measurement choices on accounting comparability. The FVO enables entities to use different measurement bases for similar assets and liabilities, raising questions about whether the FVO compromises or enhances comparability. Using a sample of US banks, we find that FVO elections increase comparability both across FVO electing banks and between FVO electing banks and banks that never elect the FVO but only if the FVO elections comply with the intent of the standard setters to remedy accounting mismatches. Overall our results suggest that banks elect the FVO to better present their economics, yielding higher comparability.