The Learning, Timing, and Pricing of the Option to Invest with Guaranteed Debt and Asymmetric Information

We consider a borrower who must get a loan from a lender to start a project. The loan is secured by an insurer, who takes the project and the lender's loss at default. The borrower grants the insurer a fraction of the loan (fee-for-guarantee swap, FGS) or of the project's equity (equity-for-guarantee swap, EGS) as the guarantee cost. FGSs or EGSs are similar to credit default swaps (CDS) but the former are much more popular than the latter to small businesses. We assume the borrower knows the project's expected growth rate, but the insurer merely knows it to take a high or low value with a given initial belief on the high-type borrower. The insurer learns about the growth rate and dynamically updates her/his belief. We show that the learning reduces the guarantee cost and adverse selection cost. A sufficiently high initial belief would induce a pooling equilibrium, where the learning will accelerate investment, though the learning would postpone the investment sometimes. FGSs are superior to EGSs if information is somewhat asymmetric, explaining the fact that the former are more popular than the latter in practice.