کارآفرینان، مخاطرات اخلاقی، و رشد درون زا

We analyze an endogenous growth model with agents differing in their endowments. Poor entrepreneurs with limited liability need to borrow in financial markets to participate in aggregate output production. We show that the first-best solution can either be achieved by decentralized financial contracting or by employing a project-specific subsidy policy. If additional capital market imperfections are introduced into the model, a negative link between inequality and growth emerges. Then, the impact of inequality on growth increases for a higher degree of frictions.

The purpose of this paper is to analyze the conditions that lead to optimal economic growth in an agency-model of financial contracting. In this model, a risk-neutral entrepreneur chooses an unobservable level of effort. He may employ the investment funds of a risk-neutral investor while both, entrepreneur and investor, are constrained by limited liability. We show that the first-best solution can either be achieved by decentralized financial contracting or by employing a project-specific subsidy policy. Recent empirical research, based on cross-country-regression analysis, has identified a negative relation between inequality and growth. Prominent examples include Persson and Tabellini (1994) and Aghion et al. (1999). In response to this finding, models have been constructed which predict lower growth rates as inequality becomes more severe. For surveys of the recent theoretical literature, see Barro (2000), Aghion et al. (1999) and Benabou (1996). There exists a variety of approaches which encompass political turmoils as well as voting behavior as possible transmission channels. Another strand of the literature examines the role of credit-market imperfections due to moral hazard in the inequality-growth context (e.g. Aghion and Bolton, 1997; Piketty, 1997). These contributions build on an incentive argument whereby inequality worsens entrepreneurial incentives which in turn depresses the economy’s growth rate as emphasized by Aghion et al. (1999). 1 Entrepreneurial investment projects in these models are very specific in that project returns follow a binomial distribution. The recent contribution of Forbes (2000) to the empirical literature challenges the supposed negative link between income inequality and growth. Given this observation, our model demonstrates that the existence of credit market imperfections due to limited liability in a model of endogenous growth is not necessarily inconsistent with these results. In contrast to the cited theoretical literature, we assume a general class of investment projects where revenue is discretely distributed. We find that the outcome of decentralized financial contracting can be Pareto-efficient.2 Thereby credit-market frictions due to moral hazard are overcome and inequality does not affect entrepreneurial incentives and growth anymore. If the Pareto-efficient effort level is not implementable by decentralized contracting, a project-specific subsidy policy can be employed which retains the first-best solution. The rationale for our result is as follows. Poor entrepreneurs are residual claimants of their project and its effort incentives depend on implementable contracts. The multiple state distribution of payoffs allows for the design of contracts which do not distort the entrepreneur’s effort decision. With a two-state payoff distribution, this requires repayments to coincide in both states due to offsetting marginal probabilities. Since in the lower profit state the repayment exceeds the payoff, limited liability prevents the coincidence of transfers. Therefore, the repayment in the low state is always smaller than in the high state, additional effort increases expected repayment, and the borrower’s effort choice is suboptimal. In contrast, a richer payoff distribution allows to offset the effect of unequal transfers by the possibility to condition repayments on additional states such that the marginal expected repayment can be reduced to zero. In this case, marginal effort return fully accrues to the entrepreneur implying a Pareto-efficient effort choice. The existence of repayment schemes inducing a Pareto-efficient allocation requires state-contingent contracts and costless state verification. With costly state verification, Gale and Hellwig (1985) and Townsend (1979) have shown the optimality of the standard debt contract. With standard debt contracts, entrepreneurs must share marginal effort return with investors. Hence, effort is distorted away from the first-best level such that a negative link between inequality and growth prevails. However, the magnitude of the impact of inequality on economic growth is driven by the degree of capital market imperfections, which is consistent with Barro (2000).

We have shown that if moral hazard is present in a lender-borrower relationship, the specification of the lender’s payoff function is crucial for the contracts that can be implemented. If the investor’s outside option does not exceed the maximum possible expected repayment from first-best contracts, then limited liability, and therefore incentives and inequality, are no obstacle to growth. The reason is that the repayment contract does not distort the entrepreneur’s effort choice and no static inefficiency arises. If the lender’s outside option happens to exceed maximum repayment, a project-specific subsidy is sufficient to resolve any inefficiencies due to incentive problems caused by inequality. The size of any such subsidy is always lower than the amount of redistribution arising from policies in the Aghion et al. (1999) spirit which provide the entrepreneur with the difference between the project’s cost and her endowment. We assume a discrete payoff function for the entrepreneur’s project. Innes (1990) analysis suggests that our result should also hold for a continuous payoff schedule. Our model demonstrates that the existence of credit market imperfections due to limited liability in a model of endogenous growth is not necessarily inconsistent with the new empirical literature which calls the supposed negative link between income inequality and growth into doubt. The introduction of additional capital market imperfections into the model, creates a negative link between inequality and growth. Then, the impact of inequality on growth is stronger for a higher degree of frictions, which is consistent with Barro (2000).