AT A recent congressional hearing I cautiously noted some “glimmers of hope” that the economy could stabilise and perhaps start to rebound later in the year. I was asked if this meant that we should cancel much of the remaining spending in the $787 billion American Recovery and Reinvestment Act. I responded that the expected recovery was both months away and predicated on Recovery Act spending ramping up greatly. Only later did it hit me that I should have told the story of 1937.

The recovery from the Depression is often described as slow because America did not return to full employment until after the outbreak of the second world war. But the truth is the recovery in the four years after Franklin Roosevelt took office in 1933 was incredibly rapid. Annual real GDP growth averaged over 9%. Unemployment fell from 25% to 14%. The second world war aside, the United States has never experienced such sustained, rapid growth.

However, that growth was halted by a second severe downturn in 1937-38, when unemployment surged again to 19% (see chart). The fundamental cause of this second recession was an unfortunate, and largely inadvertent, switch to contractionary fiscal and monetary policy. One source of the growth in 1936 was that Congress had overridden Mr Roosevelt’s veto and passed a large bonus for veterans of the first world war. In 1937, this fiscal stimulus disappeared. In addition, social-security taxes were collected for the first time. These factors reduced the deficit by roughly 2.5% of GDP, exerting significant contractionary pressure.

Also important was an accidental switch to contractionary monetary policy. In 1936 the Federal Reserve began to worry about its “exit strategy”. After several years of relatively loose monetary policy, American banks were holding large quantities of reserves in excess of their legislated requirements. Monetary policymakers feared these excess reserves would make it difficult to tighten if inflation developed or if “speculative excess” began again on Wall Street. In July 1936 the Fed’s board of governors stated that existing excess reserves could “create an injurious credit expansion” and that it had “decided to lock up” those excess reserves “as a measure of prevention”. The Fed then doubled reserve requirements in a series of steps. Unfortunately it turned out that banks, still nervous after the financial panics of the early 1930s, wanted to hold excess reserves as a cushion. When that excess was legislated away, they scrambled to replace it by reducing lending. According to a classic study of the Depression by Milton Friedman and Anna Schwartz, the resulting monetary contraction was a central cause of the 1937-38 recession.

The 1937 episode provides a cautionary tale. The urge to declare victory and get back to normal policy after an economic crisis is strong. That urge needs to be resisted until the economy is again approaching full employment. Financial crises, in particular, tend to leave scars that make financial institutions, households and firms behave differently. If the government withdraws support too early, a return to economic decline or even panic could follow. In this regard, not only should we not prematurely stop Recovery Act spending, we need to plan carefully for its expiration. According to the Congressional Budget Office, the Recovery Act will provide nearly $400 billion of stimulus in the 2010 fiscal year, but just over $130 billion in 2011. This implies a fiscal contraction of about 2% of GDP. If all goes well, private demand will have increased enough by then to fill the gap. If that is not the case, broad policy support may need to be sustained somewhat longer.

Perhaps a more fundamental lesson is that policymakers should find constructive ways to respond to the natural pressure to cut back on stimulus. For example, the Federal Reserve’s balance-sheet has more than doubled during the crisis, drawing considerable attention. Monetary policymakers have made it clear that they believe continued monetary ease is appropriate. Moreover, the Fed’s credit programmes are to some degree self-eliminating: as demand for its special credit facilities shrinks, so will its balance-sheet. But now may also be a sensible time to grant the Fed additional tools to help its balance-sheet contract once the economy has recovered. Some have suggested that the Fed be authorised to issue debt, as many other central banks do. This would enhance its ability to withdraw excess cash from the financial system. Granting such additional tools now could provide confidence that the Fed will be able to respond to inflationary pressures, without it having to create that confidence by actually tightening prematurely.

Now is also the time to think about our long-run fiscal situation. Despite the large budget deficit President Obama inherited, dealing with the current crisis required increasing the deficit substantially. To switch to austerity in the immediate future would surely set back recovery and risk a 1937-like recession-within-a-recession. But many are legitimately concerned about the longer-term budget situation. That is why the president has laid out a plan to shrink the deficit he inherited by half and has repeatedly emphasised the need to reduce the long-term deficit and put the debt-to-GDP ratio on a declining trajectory. In this regard, health-care reform presents a golden opportunity. The fundamental source of long-run deficits is rising health-care expenditures. By coupling the expansion of coverage with reforms that significantly slow the growth of health-care costs, we can dramatically improve the long-run fiscal situation without tightening prematurely.

As someone who has written somewhat critically of the short-sightedness of policymakers in the late 1930s, I feel new humility. I can see that the pressures they were under were probably enormous. Policymakers today need to learn from their experiences and respond to the same pressures constructively, without derailing the recovery before it has even begun.