Working Paper: CEPR ID: DP15785
Authors: Filippo De Marco; Julien Sauvagnat; Enrico Sette
Abstract: We study how banks lend to overconfident borrowers. For identification, we exploit variation in pupils’ overconfidence across areas in Italy. We find that borrowers born in overconfident areas make larger forecast errors on future sales, pay higher loan rates and are more likely to be denied credit. Consistent with a credit market model where borrowers have biased beliefs, collateral-based banks are more likely to grant credit to overconfident borrowers, who then invest and default more than others. We estimate that bad loans in Italy would be €10 billion (8%) lower in 2017 if banks relied less on collateral when lending to overconfident borrowers.
Keywords: optimism; business expectations; loan applications; borrower default; collateral requirements
JEL Codes: G21; G41
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| pupils' overconfidence (C92) | borrowers' beliefs about their abilities (G53) |
| borrowers' beliefs about their abilities (G53) | biased forecasts (C53) |
| overconfident borrowers (G21) | larger forecast errors on future sales (G17) |
| overconfident borrowers (G21) | higher loan rates (G21) |
| overconfident borrowers (G21) | more likely to be denied credit (G51) |
| collateral pledged (G33) | banks grant credit to overconfident borrowers (G21) |
| banks grant credit to overconfident borrowers (G21) | higher rates of investment (G31) |
| higher rates of investment (G31) | subsequent defaults (G33) |
| reliance on collateral (G33) | economic impact of overconfidence on loan outcomes (G41) |