Why Chile's Emergency Credit Was Also a Screening System
Chile's COVID-19 response is the subject of this series: a public credit-guarantee expansion operated alongside temporary employment protection, and the two programs jointly shaped what firms borrowed, what banks approved, and what risks remained with the state. This opening focuses on the puzzle behind that interaction. It gives the reader a working map for judging crisis credit not only by how much money it disburses, but also by who demands it, who is screened out, and how an alternative form of support changes the information in a loan application.
A shock large enough to make ordinary lending decisions consequential
The policy problem began with a real economic interruption, not with a spreadsheet of loan guarantees. Chile entered the pandemic while recovering from the social unrest of late 2019. In 2020, economic activity was projected to contract by 5.8 percent. Employment fell by 20.6 percent at its mid-year trough, and unemployment reached 13.1 percent before beginning to ease. For many firms, the immediate danger was a temporary loss of sales or the inability to operate normally. Without liquidity, a firm that might survive a short shutdown could fail before demand returned. Without help for workers, preserving an employment relationship could become more expensive than suspending it or closing altogether.
That urgency explains why the government did more than wait for banks to reassess every borrower under crisis conditions. The response combined fiscal transfers and tax deferrals with credit support, employment measures, and central-bank liquidity facilities. The IMF describes a multi-year fiscal package of about US$34 billion, or 13 percent of GDP. The Central Bank cut its policy rate to 0.5 percent and introduced funding-for-lending facilities of about US$40 billion, alongside other measures to keep markets liquid. These figures describe the policy environment around the case, not an effect attributable to the two firm programs alone. They show, however, why the question of bank screening mattered: the state was trying to prevent a sudden stop in credit while the normal signals used to assess repayment were being distorted by an exceptional shock.
The danger ran in both directions. If banks became too cautious, viable firms could lose working capital and an avoidable wave of bankruptcies could destroy productive relationships. If public support made borrowing unusually cheap and shifted too much loss away from lenders, firms with weak repayment prospects could borrow more, increasing private debt and future fiscal or banking costs. A large loan total could therefore mean successful stabilization, excessive risk-taking, or both. The allocation process—not just the headline amount—would decide which interpretation was plausible.
Chile's two responses to the same disruption
The World Bank study follows two instruments that addressed different constraints. The expanded FOGAPE COVID program encouraged banks to provide working-capital loans backed partly by a public guarantee. In the study's account, the loans could cover up to three months of sales, carried a six-month grace period, and faced a 3.5 percent interest-rate ceiling, close to a zero real rate in the period examined. The guarantee was partial, and banks had to analyze each application. They therefore retained exposure to default and a reason to distinguish among borrowers even while the state made lending faster and cheaper.
The Employment Protection Law worked through the labor relationship instead. It allowed a firm to suspend a contract temporarily or reduce a work schedule while workers received payments from the unemployment-insurance system. The firm did not face the same direct wage burden during the suspension, but it did face an opportunity cost: it had to scale down activity and forgo the profits associated with the workers placed under the program. In the credit program, borrowing at a low capped rate had little comparable cost to current operations. In the employment program, support was valuable precisely when the firm could not profitably keep all of its activity running.
This difference gives the case its explanatory power. The programs were not simply two channels for the same subsidy. A firm could need cash and still avoid employment protection if the opportunity cost of shutting down was high. Another firm could use employment protection because a lockdown or collapsing sales made continued operation temporarily unprofitable, reducing the amount of credit it needed. The choice itself could then tell a bank something about the firm's current operating scale, even if it did not prove that the firm had been a high-risk borrower before the pandemic.
The first distinction to retain is between a firm's need for liquidity and its likelihood of default. The pandemic could hit a sound hotel, restaurant, or manufacturer through an external restriction, while a less exposed firm could already have weak finances. The second distinction is between credit demand and credit supply. Riskier firms may want a guaranteed loan more strongly because the public guarantee and low rate make it cheaper than ordinary borrowing. Banks may still reject more of those applications because they bear part of the loss. The observed users are the result of both forces. That is why the word “screening” needs to be understood broadly. The government sets the outer gate: it decides which firms can enter the program and how much of a loan can be covered. A firm then screens itself by deciding whether the price and terms are worth taking, and a bank screens the firms that actually apply. The final portfolio is not a direct picture of the firms that needed help, the firms that were most exposed to the pandemic, or the firms that were most likely to default. It is the result of these filters operating in sequence.
Those filters also operate on different clocks. A bank can use information about a firm's earlier debt, size, productivity, and repayment record to form an ex ante view of risk. A lockdown or sudden sales decline is a different fact: it describes what the shock is doing to the firm's current operations, not necessarily what the firm was worth as a borrower before the shock. Ex post default comes later and is yet another observation. Keeping these moments separate matters because a sound firm can suffer a severe temporary interruption, while a firm with weak finances can seek cheap credit even when its local activity is less restricted.
The same separation is needed when thinking about the public balance sheet. A guarantee is not the same as an immediate government payment, and a large amount of guaranteed lending is not the same as a large realized loss. Exposure depends on the risk of the borrowers who receive loans, the size of those loans, the portion banks retain, and what happens after the crisis. Conversely, a modest aggregate loss does not mean that no risky firm borrowed or that no individual balance sheet became more fragile. The case requires both views at once: the distribution of decisions across firms and the way those decisions add up for banks and the state.
This is also why the employment program belongs inside the credit story. It changes the firm's alternatives before the firm reaches the bank. If wage support makes a temporary shutdown bearable, a firm may borrow less than it otherwise would. But choosing that support can also reveal that the firm has reduced its operating scale. The bank therefore receives information that is neither a conventional credit score nor a simple measure of past default. It is a signal about the firm's current response to the shock. Whether that signal improves the quality of the eventual credit pool is an empirical question, not a premise to assume.
The question this case makes visible
The central question is not whether guarantees are good or bad, nor whether employment protection is preferable to credit in every crisis. It is narrower and more useful: why does crisis credit work differently when employment protection is available at the same time? The answer suggested by Chile's evidence is that the second program changes both sides of the credit market. It meets some of a distressed firm's immediate needs, reducing the amount it must borrow. It also creates an observable sign that the firm is reducing operations, which a bank can incorporate into screening. Meanwhile, the low cost of guaranteed credit attracts demand from riskier firms, while partial guarantees preserve a lender's incentive to screen them.
That interaction turns emergency assistance into a selection system. Employment protection does not automatically make a firm safer, and a rejection does not establish that a firm was doomed to fail. But the combination can change the pool of borrowers that reaches a bank, the information attached to each application, and the amount of debt that survives the crisis. The relevant outcomes then extend beyond the firm: loan size and borrower composition determine expected loss, while guarantee terms determine how that loss is divided between banks and government.
The rest of this series follows that chain in the order needed to understand it. It begins with the rules that gave the two programs different costs, then separates who wanted credit from who received it. From there, it examines how sales shocks and lockdowns made employment protection informative, how approved loans changed leverage, and why aggregate losses could remain smaller than the micro-level evidence of risky borrowing might suggest. The final judgment is conditional: crisis credit can preserve access without eliminating discipline only when eligibility, pricing, lender exposure, and alternative support work together.