The Boom Is Part of the Risk
This series examines financial liberalization as a change in banking regulation, with particular attention to reforms that directly loosen banks’ ability or capacity to supply credit. The question is not whether finance is simply good or bad. It is why a policy can make borrowing and economic activity look healthier in the short run while increasing the chance of a severe downturn several years later. Understanding that sequence gives readers a better way to judge reforms than either celebrating the first boom or condemning every opening of a financial market in advance.
The promise and the problem
The promise of liberalization is familiar. When banks face fewer restrictions on whom they can serve, what they can lend, or how much credit they can create, more households and firms may gain access to external finance. A firm that could not previously borrow for equipment, inventory, or expansion may now be able to do so. A household may obtain a mortgage. New entrants may compete with established banks, making intermediation more available and sometimes less expensive. In this view, deeper credit markets can support investment, production, and growth.
The problem is that the same expansion can change the financial system’s exposure to loss. More lending means more debt. If credit flows disproportionately toward property and other activities whose value depends on collateral, rising asset prices can make the first wave of borrowing look safer than it is. Banks may also fund loan growth with liabilities that are less stable than ordinary deposits. For a time, low defaults and strong activity can reinforce confidence. The balance sheet is expanding while the warning signs remain muted.
That is why the title’s sequence matters. The boom is not necessarily evidence that the policy has failed; it may be evidence that the policy is working through its intended channel. But the boom can also be the process through which leverage, collateral dependence, and fragile funding accumulate. A later shock does not need to create all the vulnerability from nothing. It may only reveal what the earlier expansion made harder to absorb.
The Federal Reserve study at the center of this discussion follows banking regulation across 18 advanced economies from the postwar period through 2020. Its central result is selective rather than sweeping: reforms that directly affect the supply of bank credit are followed by stronger credit expansions, temporary gains in real activity, and higher medium-term financial-crisis risk. Other banking reforms do not show the same boom-and-bust pattern in the estimates. The distinction is essential. “Financial liberalization” names a family of changes, not a single economic treatment with one predictable effect.
The study is also a working paper, and its authors describe the evidence with appropriate caution. Historical policy changes cannot be treated as laboratory experiments. Governments may liberalize when they expect stronger growth, when borrowers are demanding more credit, or when they are pursuing other market reforms at the same time. The paper uses several strategies to reduce those concerns, but the responsible conclusion is still channel-specific and conditional: direct credit-supply liberalization is a documented source of credit expansion and later risk in the sample, not an automatic crisis law.
Why the timing is the mechanism
The easiest way to misunderstand this subject is to compare the reform with only the next few quarters. Early results can be genuinely positive. The study estimates that real GDP per person rises in the first years after a supply-side banking reform. Credit becomes more plentiful, businesses can act on opportunities that were previously financing-constrained, and households or firms may spend against higher expected income or asset values. If an evaluation stops there, liberalization appears to have delivered exactly what its proponents promised.
The risk often operates on a slower clock. Banks can originate loans quickly, but fragility takes time to build. Borrowers must take on debt, property prices must respond, and banks must decide how much of the expansion to finance through deposits, wholesale markets, or other non-core funding. During the early phase, stronger income and rising collateral can keep defaults low. Credit spreads may not widen because lenders and borrowers have not yet had a reason to reprice the system-wide exposure. The absence of an immediate crisis therefore tells us little about the risk several years into the transition.
This creates an important difference between an average path and a distribution of outcomes. In the study’s estimates, GDP per capita rises by about 0.74 percent cumulatively over the first three years, peaks around the fourth year, and then declines toward its previous trend. By roughly the eighth year, the average level has returned to trend. At the same time, the probability of severe left-tail outcomes—such as unusually weak GDP growth or unusually high unemployment—rises in the medium term. A positive average in the first phase can coexist with a larger chance of an unusually bad result later.
The crisis estimates make the lag more concrete. They show little immediate response after deregulation, but the estimated crisis probability rises from about the three-year horizon onward. For a one-unit increase in the study’s credit-supply index, the estimated increase is about 2.8 percentage points at three years and about 9.8 percentage points at eight years, relative to an economy whose regulatory framework did not change. These are estimates tied to a particular index, sample, crisis definition, and comparison—not forecasts that a specific reform will produce a crisis. Their importance lies in the timing and concentration: the reforms that expand credit most directly are also the ones associated with the later rise in crisis risk.
The welfare question follows naturally. A policymaker may reasonably value the gains from making finance more accessible, especially if credit releases productive investment. But a policymaker who cares about households and firms that bear severe losses may place a high value on the insurance provided by constraints. The relevant comparison is not “growth versus no growth.” It is higher consumption in ordinary times versus a distribution with a heavier downside tail. The sources later give that trade-off a formal, assumption-dependent treatment.
The same credit can travel farther than it first appears
The domestic bank is only one place where the story can be observed. International banking changes the route by which funds reach a borrower and the location where exposure appears. The BIS reports that international claims of banks in reporting countries grew from less than 2 percent of world GDP in 1963 to more than 60 percent in 2007, before falling to around 40 percent in early 2021. The expansion was driven by regulatory arbitrage, financial liberalization, and financial innovation, and it helped amplify credit booms before several major crises.
Some of that business was offshore: the lender and borrower used a currency foreign to both of them. Other transactions crossed a border in the borrower’s currency, or were booked through a foreign affiliate that lent locally. These distinctions matter because a credit boom can be funded in one country, originated in another, denominated in a third country’s currency, and ultimately carried by a borrower who never dealt directly with the original funder. A balance sheet that looks local can therefore depend on a global funding chain.
Innovation can extend that chain. Syndicated loans allowed more banks to participate in large loans and made those loans easier to trade. Derivatives helped banks manage interest-rate, foreign-exchange, and credit exposures, sometimes moving risk away from the originating balance sheet. Repos connected banks to money-market investors through collateralized funding. None of these instruments is inherently a crisis mechanism. But they can make the network larger, more distributed, and harder to read. A fall in visible interbank assets does not necessarily mean that the system has become less connected if important links now run through derivatives, central counterparties, or funding markets.
After the Great Financial Crisis, tighter bank capital and leverage rules increased banks’ ability to absorb shocks while also making balance-sheet expansion more expensive. Non-bank creditors then supplied a larger share of foreign-currency credit. The BIS reports that dollar credit to non-bank borrowers outside the United States more than doubled from mid-2008 to $13 trillion by early 2021, while the share supplied through bank loans fell from 60 percent to 47 percent. This is not a final escape from the problem. It is a change in the perimeter: risk can move from banks to asset managers, bond markets, and other institutions, along with new questions about transparency and liquidity.
The map is deliberately simple. It does not claim that every reform follows every arrow, or that every crisis has one cause. It identifies the relationships that need to be tested. Does the reform directly release lending capacity? Where does the new credit go? How is it funded? Which balance sheets, currencies, and institutions ultimately carry the exposure? And what happens when the economy is evaluated after the initial optimism has had time to become leverage?
The first discipline, then, is classification. Before asking whether financial liberalization “works,” we need to ask what has actually been liberalized. A rule that removes a binding limit on loan supply should not be treated as equivalent to a reform that changes market structure, international capital allocation, or risk-sharing without directly increasing the quantity of bank credit. That distinction is where the next part of the series begins, because it determines whether the apparent boom is merely a favorable outcome—or the first visible stage of a much longer risk cycle.