Today, there is essentially one accepted narrative of the economic crisis that began in late 2007. Overly optimistic homebuyers and reckless lenders in the United States created a housing price bubble. Regulators were asleep at the switch. When the bubble inevitably popped, the government had to bail out the banks, and the United States suffered its deepest and longest slump since the 1930s. For anyone who has seen or read The Big Short, this story will be familiar.
Yet it is also wrong. The real cause of the Great Recession lay not in the housing market but in the misguided monetary policy of the Federal Reserve. As the economy began to collapse in 2008, the Fed focused on solving the housing crisis. Yet the housing crisis was a distraction. On its own, it might have caused a weak recession, but little more. As the Fed bailed out the banks at risk from innumerable bad mortgages, it ignored the root cause of serious recessions: a fall in nominal GDP, or NGDP, which counts the total value of all goods and services produced in the United States, not adjusted for inflation. Such a fall began unimpeded in mid-2008, and once that happened, much of the damage had been done.
The Fed can control NGDP through its monetary policy, and as NGDP fell in 2008, the Fed should have lowered interest rates rapidly. If that proved insufficient, it should have increased the money supply through quantitative easing. Instead, the Fed, terrified of inflation, kept interest rates too high for too long—causing NGDP to fall even further.
To prevent such errors in the future, the Fed should switch from targeting inflation to targeting the level of NGDP. When a recession hits, NGDP tends to fall before inflation, which means that a central bank focused on targeting inflation will be too slow to respond. Throughout mid-2008, U.S. inflation remained positive, as NGDP began falling. Had the Fed targeted NGDP, it might have acted much sooner to boost growth—staving off the Great Recession and the suffering that came with it.
Most pundits blame the housing market for the Great Recession. But their argument doesn’t hold up to scrutiny. For one thing, the United States was not the only country to experience a housing boom; Australia, Canada, New Zealand, and the United Kingdom also did. In all four countries, house prices rose sharply in the first decade of this century, just as they did in the United States, but in all four, they have yet to fall. In fact, for a decade now, real house prices in these countries have remained close to 2006 levels, or even moved higher.
For another thing, theories connecting the Great Recession to the housing bust have a timing problem. Between January 2006 and April 2008, housing construction in the United States plunged by more than 50 percent. Yet unemployment moved only from 4.7 percent to 5.0 percent. The big problem occurred later, as unemployment doubled, to ten percent, by October 2009. During the first 27 months of the housing slump, capital and labor were reallocated to other growing industries, such as commercial construction, exports, and services, mitigating the worst effects of the housing collapse.
This makes sense: classical economic theory predicts that when one sector declines, capital and labor will shift to other sectors. Contrary to popular belief, real shocks—such as the bursting of a housing bubble, a devastating natural disaster, a stock market crash, or a terrorist attack—do not cause deep recessions. The stock market crash of 1987, comparable to the 1929 crash, had no effect whatsoever on U.S. unemployment. The earthquake and tsunami that struck Japan in 2011, devastating part of the country and shutting down the entire nuclear industry for almost two years, caused a temporary dip in industrial output but were barely noticeable in unemployment figures.
Instead, major recessions are caused by monetary policy failures. In order for capital and labor to shift easily from a declining sector to a growing sector, the total spending in an economy must continue to rise at a reasonable pace. The best measure of total spending is NGDP, since it measures changes in the total amount of money spent on all goods and services.
Central banks can control this flow of money through monetary policy. They can expand the supply of money by purchasing small quantities of government bonds in order to lower short-term interest rates, or, if rates fall to zero, by quantitative easing: the purchase of large quantities of government bonds. And they can contract the supply of money by selling bonds, which will raise short-term interest rates.
What triggers recessions are abrupt drops in NGDP. If NGDP grows too quickly, monetary policy is too loose. With too much money chasing each transaction, prices rise, and the result is inflation. If NGDP falls abruptly, on the other hand, monetary policy is too tight—there is no longer enough money to pay everyone who wants to work or to fund all the transactions that would otherwise take place, and the economy starts to contract.
Major recessions are caused by monetary policy failures.
Drops in NGDP are particularly damaging for two reasons. First, wages are what economists describe as “sticky downward”: when spending throughout the economy rises, employees are able to negotiate pay increases, but when spending falls, employers would rather fire a few people than negotiate pay cuts with all their employees. Drops in NGDP kick off something akin to a game of musical chairs. Just as removing several chairs will leave some players sitting on the floor when the music stops, removing several percentage points of expected NGDP growth will leave too little revenue to employ the existing work force at the wages they have negotiated. The result: rising unemployment.
Consider what happened in the U.S. labor market in the 1970s. From 1971 to 1981, NGDP grew at an average of 11 percent per year, and workers negotiated large pay increases. But in 1982, after Paul Volcker, then chair of the Federal Reserve, had tightened the Fed’s monetary policy to fight inflation, NGDP growth fell to less than five percent, and unemployment soared. Companies had committed to pay workers based on revenue forecasts that proved inaccurate.
A similar problem occurs in the credit market. Even sophisticated Wall Street firms issue long-term contracts in nominal terms, such as 30-year bonds with fixed nominal interest rates. Such contractual obligations are more difficult to meet when monetary policy allows NGDP growth to slow sharply.
This leads to the second major problem associated with NGDP shocks: financial market instability. When NGDP growth falls sharply relative to expectations, economies tend to suffer financial crises. A decrease in nominal income means there is less money to pay back loans, so defaults become more common and banks come under increasing strain. This is a familiar phenomenon. NGDP in the United States fell by half during the early 1930s, and there were debt crises all over the world. NGDP growth fell to roughly zero in Japan after 1993, triggering severe banking problems, and it plunged in the late 1990s in Argentina, leading to a serious financial crisis in 2001. And something similar happened in the United States and the eurozone during the Great Recession.
THE ROAD TO RECESSION
When the housing crisis hit at the end of 2007, defaults on reckless subprime mortgages put the U.S. banking sector under stress. The Fed stepped in to rescue the financial system, bailing out the investment bank Bear Stearns and lending money to banks. Such actions might have been sufficient if the problem had been contained to turmoil in the financial sector.
But in mid-2008, two years after the housing market began to collapse, a much more serious problem emerged. The Fed did not cut interests rates quickly enough to offset the drag caused by the housing crisis, perhaps out of fear of high inflation resulting from rising oil prices. As a result, NGDP fell sharply. Until 2008, NGDP growth had averaged about five percent per year. Starting in June 2008, however, NGDP fell by roughly three percent in 12 months, to about eight percentage points below the pre-recession trend line.
As NGDP fell, unemployment rose and spread from the housing sector to almost every part of the economy. And the financial crisis, initially triggered by the housing slump, became much worse. As a result, what had initially been just a financial crisis turned into a full-blown macroeconomic crisis.
Yet policymakers initially ignored the fall in NGDP growth. Throughout 2008, they continued to assume that the problem was banking distress, rather than a contraction in nominal spending. Worse, they thought that the risk of inflation was just as great as the risk of a recession, even after Lehman Brothers failed in September. It is true that inflation had been quite high for the previous 12 months, thanks to high oil prices. But the markets thought inflation would fall sharply over the next few years. The Fed chose to ignore those market forecasts. Instead of expanding the supply of money to boost NGDP, it refused to touch interest rates between April and October 2008, keeping them at two percent. Even on September 16, 2008, the day after Lehman Brothers filed for bankruptcy, the Federal Reserve Board voted not to cut interest rates, a decision that Ben Bernanke, at the time the Fed’s chair, now concedes was an error.
Normally, the Fed’s aggressive moves to inject money into the banking system would have immediately pushed interest rates to zero. But because the Fed did not want to boost nominal spending, in early October, it introduced a new policy: it started to pay interest on reserves that banks hold with the Fed. The move prevented interest rates from falling to zero and encouraged banks to keep their money at the Fed rather than move it out into the wider economy: a contractionary move at a time when monetary stimulus was essential.
A cynic might say the Fed was trying to rescue Wall Street without rescuing Main Street: it was saving the banks but not allowing the interest rates that affect the wider economy to fall enough to boost NGDP. A more likely explanation, however, is that the Fed made a misdiagnosis. There were two distinct problems: banking distress caused by defaults on subprime mortgages and a much more serious macroeconomic crisis caused by the shortfall in spending. The Fed recognized the first but missed the second.
Even worse, the problem that the Fed ignored exacerbated the banking crisis—as NGDP fell, people and businesses across the economy had less money than they had anticipated to pay back debts. The financial crisis worsened, the housing market collapsed further, and unemployment soared. Only in December 2008 did the Fed cut rates close to zero. But by then, the damage had been done: a mild downturn had turned into the Great Recession.
OUT OF AMMO?
In late 2008, the Fed finally sought to reverse the shortfall in nominal spending through programs such as quantitative easing. This was better late than never. Still, throughout the crisis and the ongoing weak recovery, the Fed has been too tentative and its monetary policy too contractionary. This view flies in the face of accepted wisdom. Most pundits still think the Fed’s post-2008 policy was expansionary because it ultimately brought short-term interest rates close to zero. But they misunderstand how monetary policy really works.
A low nominal interest rate in itself does not constitute an expansionary monetary policy; what matters is its value relative to the “natural interest rate,” the rate at which inflation and NGDP remain on target. If the nominal rate is above the natural rate, monetary policy is contractionary. And if the natural rate falls, as it did in 2008 when spending slowed down and inflation decreased, then a fixed nominal interest rate will effectively rise relative to the natural rate, causing monetary policy to tighten. The supply of money will no longer be growing quickly enough to pay everyone’s wages. If nominal spending has fallen to such an extent that the natural rate is below zero, then an interest rate just above zero, although low in absolute terms, may in fact still represent a tight monetary policy.
So even though interest rates fell close to zero from December 2008 onward and many assumed that the Fed’s monetary policy was expansionary, in reality it was not. Between July and December 2008, a number of things happened that point to a contractionary monetary policy. Commodity prices fell by roughly half. Stock prices crashed. The dollar strengthened against other currencies. Real estate prices fell in states all over the country, not just in places where subprime mortgages had been common. And the financial markets expected inflation to turn negative.
The so-called zero lower bound, when interest rates are at or near zero, provides no excuse for the Fed’s refusal to employ an expansionary monetary policy when such a policy was needed. Contrary to the claims of many pundits, it is not true that central banks are out of ammunition when interest rates approach or hit zero. They can always increase the supply of money if they choose, by creating money through “unconventional” measures such as quantitative easing.
From the end of 2008 until 2014, the Fed launched several rounds of quantitative easing to boost NGDP. Although it should have made these moves sooner and more aggressively, they did help end the recession in the United States. Some argue that quantitative easing and extremely low interest rates have increased the risks of another financial crisis by creating asset price bubbles as investors search for higher returns from riskier assets, but these fears are exaggerated: the low interest rates of recent years do not reflect a loose monetary policy; rather, they indicate a new normal of slow growth in the developed world. Asset prices should in fact be higher when interest rates are low. Market indicators suggest that relatively low rates are here to stay, and so current asset prices are not necessarily a bubble waiting to burst.
Had the Fed acted decisively back in 2008, the crisis would have been far less severe.
The European Central Bank, in contrast to the Fed, avoided quantitative easing until much later and did even less to boost spending. As a result, the eurozone slid into a double-dip recession in 2011, as the U.S. economy continued to recover. This was thanks to monetary policy, not fiscal policy: after 2011, there was actually more fiscal austerity in the United States than in the eurozone thanks to sequestration, the automatic budget cuts agreed to by Congress that year.
A good example of the power of monetary policy came in early 2013. Taxes increased sharply at the beginning of 2013, and a few months later, U.S. government spending tightened because of sequestration. Several months earlier, the Fed had launched a round of quantitative easing, purchasing trillions of dollars of assets in the following two years to expand the money supply. Three hundred and fifty Keynesian economists signed an open letter warning that the budget cuts could push the U.S. economy back into recession in 2013. The economist Paul Krugman argued that an expansionary monetary policy would not be able to offset the fiscal austerity. They were wrong: the U.S. economy grew faster in 2013 than in 2012. The budget deficit fell by almost half, or $500 billion, in one year with no discernible effect on growth. This showed not only that monetary policy remained effective when interest rates were at zero but also that it was much more powerful than fiscal policy. And it suggested that had the Fed acted so decisively back in 2008, the crisis would have been far less severe.
If monetary policy, not the housing market or the banking system, was the root of the Great Recession, then well-intentioned financial regulation, such as the Dodd-Frank Act, won’t solve the problem. Instead, the Fed should reform the way it conducts monetary policy and stop targeting inflation.
As things work now, the Fed aims to hit two percent inflation each year, and if it misses that target, it simply tries again the following year. It doesn’t try to get back on track and make up for lost ground. If prices fall one year, as they did in 2009, for instance, then instead of trying to make up the difference by raising inflation above two percent the following year, the Fed just tries once again to hit two percent. But this means that if inflation is less than two percent two years in a row, as it was from 2009 to 2010, monetary policy will have become tighter: goods will cost less than they would have had the Fed hit its target each year, and the risk of a recession brought on by such a fall will be high.
The more predictable the monetary policy, the more stable the economy.
One way to avoid this would be for the Fed to switch from setting a target for inflation to setting a target path for what goods should cost over the next few years. This practice is known as “price-level targeting.” Imagine a piece of graph paper showing the price level rising along a two percent trend line from its current position. Under price-level targeting, a central bank promises to move the price level back to that trend line anytime it falls below or rises above it. Next year, the target cost of living is two percent higher than today; the following year, it is four percent higher; and in three years’ time, it is roughly six percent higher. In contrast to the current system, in this system, if prices fall one year, the next year the Fed would have to aim for a period of above two percent inflation to catch up to the trend line. Krugman has described such a strategy as “promis[ing] to be irresponsible.”
To see the advantages of price-level targeting, consider the situation in late 2008, when it was clear that the economy was entering a deep recession in which prices would fall. If investors knew that the Fed would eventually print as much money as necessary to bring prices back up to the pre-recession trend line, asset prices such as stocks, commodities, and real estate would have fallen by much less. As a result, fewer people would have defaulted on their loans, and banks such as Lehman Brothers would have been less likely to fail. That doesn’t mean that no banks would have failed, but it does mean that the crisis would have been milder.
Bernanke himself recommended that the Bank of Japan consider price-level targeting in 2003. Yet he discovered that there was a great deal of institutional resistance at the Fed to the practice, and his suggestions were dismissed at a meeting of the Federal Open Market Committee that he attended that same year. By all accounts, Bernanke governed by consensus during his term as Fed chair, and so the Fed’s actual policy might not have reflected his ideal policy.
The main benefit of price-level targeting is that it assures markets that the price level will remain predictable in the long run. The markets know that the Fed will expand the supply of money if it undershoots its target one year and contract it if it overshoots it the next. The more predictable the monetary policy, the more stable the economy.
Yet an even better policy would be to target the level of NGDP directly, because changes in NGDP tend to track changes in unemployment more closely than do changes in inflation. Under this approach, the Fed would commit to increase total nominal spending by four or five percent every year. Any decline in NGDP growth would be quickly reversed. Everyone would know that whatever happened, enough money would flow through the economy to generate the sort of growth in national income that was expected when wage and debt contracts were signed. Individual sectors would still have their ups and downs, and financial institutions would still collapse from time to time. But total nominal spending would rise at a slow yet predictable rate.
This approach has another appealing feature: it would send a clear and credible signal to the markets that the Fed would do what it took to get NGDP back on its long-term trend line. Because the Fed would target the level of NGDP, if spending were too low one year, interest rates would fall to boost spending, and investors would know that NGDP growth the following year would be higher to make up the lost ground. Their bullish expectations would themselves lead to increased current spending. A collapse in confidence like the one that accompanied the start of the Great Recession would be much less likely if central banks focused on keeping NGDP growing steadily. Australia has not had a recession since 1991 because it has kept NGDP growing along a relatively stable path.
Economists of all stripes (and not just “market monetarists,” who initially supported the policy) are increasingly starting to back NGDP-level targeting. In the aftermath of the Great Recession, the practice has racked up endorsements from some of the most respected macroeconomists in the United States, including Michael Woodford, a professor at Columbia; Christina Romer, a former chair of the Council of Economic Advisers; and Jeffrey Frankel, a former member of the same council. Many other top economists, such as Krugman and former U.S. Treasury Secretary Larry Summers, have suggested that it’s worth examining closely.
Central banks are conservative institutions and will no doubt be slow to embrace this new way of thinking. Yet perhaps their conservatism will not prevent them from learning at least some of the lessons of this painful past decade.