ABSTRACT If firms issue convertible securities to facilitate sequential investment, the securities should be engineered to give sufficient flexibility to accommodate timing of follow‐on investment. We examine call provisions in convertible bonds and argue that firms with investment options expected to expire sooner (later) will offer weaker (stronger) call protection. We find that issues with weak or no call protection are offered by firms that invest greater amounts soon after issuance than those issuing convertibles with strong protection. Moreover, capital expenditure levels during the 5‐year period following issuance are inversely related to the length of call‐protection periods.

This paper presents a model that provides an explanation, based on regime switching in the real interest rate and learning, of why tests based on stock-adjustment models, Euler equations, or decision rules—which emphasize short-run fluctuations in inventories and the interest rate—are unlikely to uncover a negative relationship between inventories and the real interest rate. The model, however, predicts that inventories will respond to long-run movements, that is, to regime shifts in the real interest rate. Tests emphasizing cointegration techniques confirm this prediction and show a significant long-run relationship between inventories and the real interest rate.