We extend the neoclassical investment model (Hayashi, 1982) to allow for limited commitment on compensation contracts. We consider three types of limited commitment: (i) managers cannot commit to compensation contracts that provide lower continuation utility than their outside options; (ii) shareholders cannot commit to negative net present value (NPV) projects; (iii) both the managers and the shareholders cannot commit. We characterize the optimal contract under general convex adjustment cost functions and provide examples for which closed-form solutions can be obtained. We show that, as in the data, small firms invest more, grow faster, and have a higher Tobin׳s Q than large firms under the optimal contract. In addition, the pattern of the dependence of chief executive officer (CEO) compensation on past performance implied by our model is also consistent with empirical evidence.

Abstract We develop an equilibrium model with moral hazard, which arises because some productivity shocks are privately observed by firm managers only. We characterize the optimal contract and its implications for firm size, growth, and managerial pay-performance sensitivity and exploit them to quantify the severity of the moral hazard problem. Our estimation suggests that unobservable shocks are relatively modest and account for about 10% of the total variation of firm output. Nonetheless, moral-hazard-induced incentive pay is quantitatively significant and accounts for 50% of managerial compensation. Eliminating moral hazard would result in about a 1% increase in aggregate output.

ABSTRACT We develop a unified theory of dynamic contracting and assortative matching to explain firm dynamics. In our model, neither firms nor managers can commit to arrangements that yield lower payoffs than their outside options, which are microfounded by the equilibrium conditions in a matching market. The model endogenously generates power laws in firm size and CEO compensation, and explains differences in their right tails. We also show that our model quantitatively accounts for many salient features of the time‐series dynamics and the cross‐sectional distribution of firm investment, dividend payout, and CEO compensation.