Did Economists Doom Obama's Presidency?
If President Obama loses the election in November, economists may well end up taking a share of the blame - for good reason. Their models misled him into applying ambitious stimulus therapies to jump start the economy and boost employment that haven't worked, vastly undermining his re-election prospects.
Back in January 2009, a now infamous study coauthored by Christina Romer, the future chair of the President Obama's Council of Economic Advisors, and Jared Bernstein, the future chief economist for the Vice President, predicted that an $800 billion economic stimulus targeted toward boosting consumer demand would stave off a severe recession and hold unemployment below 8 percent by the end of 2009.
What was so compelling about their study was the illusion of precision. The Obama administration used their statistical analysis to aggressively promote specific policy proposals, including the package of tax cuts and discretionary federal spending embodied in the so-called stimulus package, the American Recovery and Reinvestment Act of 2009.
But little of what they predicted has panned out.
Unemployment remained over 8 percent for three-and-a-half years, finally dropping to 7.8 percent in September. But that's only because we don't include "discouraged workers," those who have given up looking for work, or those unwillingly working in jobs below their skill and education levels. Moreover, job creation remains lackluster, barely adding enough to meet demand in a still anemic economy three years after the recession officially ended.
The fact that the economy is still struggling with legions of discouraged and underemployed workers after the federal anti-recession efforts started (under Republican George W. Bush) should humble policymakers and professional economists, and prompt them to reconsider their use of macroeconomic models in making major policy decisions.
Romer and Bernstein claim their predictions were off because the published data at the time didn't capture the true magnitude of the economic crisis. But the fundamental problem runs much deeper. The model missed the severity of the recession, to be sure, but, more importantly, Romer and Bernstein (and other forecasters) failed to understand the nature of the recession.
Industry "bubbles" - growth unsustained by market fundamentals such as household income or demand - are hard to identify, let alone predict, regardless of whether it's a tech bubble (1990s) or a housing bubble (2000s). Investors and businesses have difficulty determining whether the consumer demand that seems to be fueling purchases of their product are the result of real demand from, say, higher incomes, or false signals based on distortions in lending practices or inflation. During the height of the housing bubble, for example, home builders often voiced skepticism about the sustainability of demand. But if a potential home buyer was able to secure a mortgage, they sold her a house anyway. Moreover, even with lots of knowledge about the economy, macroeconomic models are notoriously poor at predicting downturns in the economy as well as the upturns.
Thus, even though Romer and Bernstein's models were fairly sophisticated by mainstream standards, they failed to capture the real world complexities of the economy. More specifically , their models (and others) failed to understand the way housing finance and risky mortgages had become inextricably linked to broader financial markets. The homebuyer may have secured a mortgage, but lax lending rules had greatly increased the likelihood of foreclosure, and these risks were spread throughout the financial industry and hard to quantify in a model.
Romer and Bernstein's forecasts started missing their mark almost immediately. Their model predicted unemployment rates with the federal stimulus should have peaked at 8 percent in August 2009 and continued a steady decline. By the end of the 2012, Romer and Bernstein's forecast had predicted unemployment rates with or without the stimulus would have converged to near 6 percent. In reality, unemployment continued to rise, peaking at 9.9 percent at the end of December 2009 (even with the stimulus) and didn't dip below 8 percent until last month.
But Romer and Bernstein missed more than the actual unemployment predictions: they also missed its trajectory. Their model predicted a steady decline in the unemployment rate under the stimulus and non-stimulus scenarios. The actual path of the economy has been far bumpier, with unemployment rates (and employment growth) rising and falling unevenly throughout 2010 and 2011. Even with the September decline, most of 2012 has seen the unemployment rate fluctuate unpredictably between 8.1 and 8.3 percent
In short, in what is perhaps the most important exercise in economic policy modeling since the Great Depression, two of the nation's foremost economists failed. And the failure was an epic one. They predicted that unemployment would peak at 8 percent after the stimulus. In fact it peaked at 9.9 percent. So it's unclear whether trillions of dollars of stimulus spending bought the country any reduction in unemployment whatsoever. It's also hard to escape the conclusion that it would have been better to do nothing and let the economy run its course.
Indeed, this failure is particularly notable because Romer and Bernstein's effort was well within accepted mainstream practice of the profession, not an exception.
Professional economists created benchmarks and expectations that were incapable of being matched in the real world or through public policy. The influence of these forecasts at one of the most critical times in the modern history should give elected officials, economic policymakers, and professional economists pause and encourage serious soul searching about the meaning, value, and relevance of macroeconomic models in national policymaking.