So far, the literature on financial constraints and misallocation has focused on a particular class of financial constraints, which prevents entrepreneurs from borrowing more than a fraction of their existing assets. Regardless of whether the constraint is binding or not, firms pay a constant, risk-free interest rate on the amount they effectively borrow, which might allow them to quickly accumulate internal funds and finance their investment without the collateral constraint being binding.
In this paper, I present some data that suggests that the heterogeneity in borrowing costs might actually be quite important, and that interest rates paid on debt correlate negatively with firm size, and positively with the ratio of debt to assets. I present and simulate a model that includes occupational choices and endogenously determined costs of credit, which could match some facts observed in the data. The main prediction of the model is that initial assets matter when it comes to firm selection, and that entrepreneurs which have not accumulated enough savings in the past face higher borrowing costs, which might significantly hinder the wealth accumulation process and force them to either get out of the market, or remain small.
I conduct numerical exercises and show that financial frictions can account for moderate TFP losses in France. My model also allows to estimate the impact of different policies designed to facilitate firms' access to credit. I show that credit subsidies and government guarantees on loans only have a moderate positive impact of TFP, since public authorities do not have better technologies to observe actual productivity and cannot discriminate productive entrepreneurs better than the private sector.