What Chile Teaches About Crisis Credit
The subject of this series has been Chile's coordinated COVID-19 response: a public credit guarantee and a temporary employment-protection program operating at the same time. The final question is why the second instrument changed the meaning and consequences of the first. The answer is not that employment protection made firms safe or that guaranteed credit was inherently reckless. It is that the two programs altered different decisions—operating, borrowing, and screening—and their interaction determined how private debt became bank and government exposure. That is the judgment the evidence can support, along with a clear account of where it stops.
The answer lies in the interaction, not in either program alone
Crisis credit works differently when employment protection is available because a firm no longer approaches the bank with only one way to manage a temporary shock. It can borrow to keep activity running, use wage support while suspending or reducing work, or combine the two. The choices depend on what the shock has damaged. A firm with a temporary liquidity shortfall may prefer low-cost credit and continue operating. A firm under a lockdown may have little profitable work for some employees and therefore value employment protection more. A firm can also face both conditions and need a loan after wage support has covered only part of its obligations.
That choice changes the credit decision in two ways. First, employment protection can reduce financing demand. When wage support covers part of the cost of retaining workers during reduced activity, the firm has less reason to finance that cost with new debt. The evidence that firms using both programs accumulated less debt than firms using credit alone is consistent with this quantity channel. The employment program does not remove every financing need; it changes the size and composition of the gap that the credit program must fill.
Second, employment-program use becomes an observable response to the real shock. A firm that suspends contracts or reduces hours has accepted a lower operating scale. For a bank reviewing a guaranteed-credit application, that may signal current difficulty servicing debt even if the firm's pre-pandemic finances were sound. The study finds that banks were less likely to approve guaranteed-credit applications from firms using employment protection after accounting for its measures of prior risk and sales performance. This is not evidence that every user was a bad borrower. It is evidence that the operating response carried information in the bank's screening decision.
Guaranteed credit creates a different asymmetry. Its low price makes borrowing especially attractive to firms that would face expensive or unavailable ordinary credit. The study finds that riskier eligible firms were more likely to apply and, in equilibrium, more likely to use the program. Yet conditional on applying, riskier firms were less likely to be approved. Demand pulled the portfolio toward risk; bank screening pushed in the other direction. The final users were produced by both forces, not by a government guarantee acting alone.
This resolves the apparent contradiction at the center of the case. A program can attract more risky borrowers while banks still screen risk seriously. The low interest ceiling and public guarantee change the firm's incentive to ask for credit. The partial guarantee and deductible leave the bank exposed to part of the loss, so the bank retains a reason to reject or limit applications. Employment protection adds another signal and meets some of the firm's need before the bank decides. The resulting credit allocation cannot be understood from the price of the loan, the guarantee percentage, or the employment benefit in isolation.
The evidence also shows why the type of shock matters. Firms with positive and negative sales growth were both more likely to use guaranteed credit than firms with little sales change, while employment protection was used more often by firms with negative sales growth. Staggered municipal lockdowns sharpened the contrast: lockdowns increased employment-program take-up but did not significantly change credit-program take-up in the study's setting. A lockdown therefore changed the cost of reducing work more directly than it changed the incentive to borrow. A financial-risk measure and an operating-disruption measure are not competing names for the same condition.
The most defensible description is a sequence of filters.