The Incredible Shrinking Relevancy of GDP

Earlier this month, the New York Times and other media reported on a new study by Harvard financial economists Kenneth Rogoff and Carmen Reinhart. It concludes that the duration of the recovery from the 2007/08 recession and financial crisis is similar to historical recessions that were linked to financial panics. The recovery period is measured by a return to levels of GDP per capita in various countries over a long historic period. Why did the study’s conclusions warrant the reportage in the Times? The study’s comparison of historic recovery periods may be accurate, but it does not matter to the most important ills plaguing our economy.

The far more important issue is the measurement used in the study. GDP defines the beginning and ending of recessions for many purposes.[1] This has great impact on consumer and business confidence as well as polices. However, this measure—the amount of national income allocable equally to every person—has become a poor indicator of the economy for policy purposes. (The accuracy and usefulness of GDP measures has been a central focus of significant research and writing at Demos.) In conditions of high income disparity, GDP per capita fails to capture the fact that most people are receiving little or no benefit from the increase in GDP growth. GDP per capita reflects the mean income of individuals.[2] Median household income can stagnate or even fall even as the per capita GDP rises.

The central issue is that employment has continued to be weak and that the median income has stagnated even as GDP has increased. In other words, the recovery has not reduced income disparity. There are actually two economies. In the wealthy economy (which is disproportionately tied to stock and other financial asset valuations), recovery has progressed quite nicely. In fact, the economy is roaring. But the rest of the economy, driven by hourly earnings in very slack labor markets, is still depressed. In public opinion polls, the middle class sees the world that way. That view is correct.

Policymakers around the world use GDP per capita to measure the success or failure of national economies. Many policies are countercyclical by design, kicking in when the GDP shrinks and then reversing when GDP growth returns. Federal Reserve monetary policy is the most familiar example that is influenced by GDP.[3] These policies are intended to stimulate employment as GDP growth slows, and to ease off as the cycle reverses to protect against price inflation. GDP per capita is now far less related to healthy levels of employment. Persistent unemployment and underemployment, even as GDP per capita increases, has the inevitable effect of increasing income disparity.

To fight the recession, we have relied heavily on monetary policy that increases liquidity in the financial markets on the mostly false assumption that this will generate investment-driven employment. Meaningful fiscal policy, spending that would have directly invested in jobs, was politically impossible. If, instead of pegging “recovery” to GDP, the benchmark were, say, median household income, we would still be officially in a recession—now going into its sixth year—and the need for a more balanced policy would be clearer.       

Globalization has reduced the relevance of GDP, as productive activities have been outsourced to other countries to reduce costs. Salaries paid to US employees are a boon to consumption, benefiting the economy broadly. But if the jobs and salaries are sent abroad and profit is increased, share value may be higher but consumption suffers. GDP may increase, but the wealthy, whose incomes are closely tied to the value of financial assets, are disproportionately benefited.

All of this fuels financialization, increased financial activity (e.g., trading) that does not benefit the broad public interests by improving the process of job-creating capital investment. Highly speculative trading that simply increases rents to financial firms is a principal cause of financialization. Derivatives and highly automated trading (in which massive volumes of securities and derivatives transactions take place at nano-second speed) are two examples. These activities increase the incomes of the bankers and traders and their shareholders without creating meaningful jobs, all to the disproportionate benefit of the upper income echelons.

Unfortunately, the economics profession has been slow to venture into this less traditional area of inquiry. The Rogoff/Reinhart study is by no means unusual or egregious. But the general public does not perceive the analytical irrelevance of GDP and will attach more meaning to the study than it deserves, especially with the prominent coverage by the Times and other media.

In a highly unequal, highly financialized economy, GDP should not be the benchmark for measuring policy success for the majority of Americans.  Whether in terms of Fed policy or public investment in the economy, we need to establish a new context for government responsibilities linked to household well-being, not GDP.  Median income and income parity should be elevated as benchmarks. Hopefully, a second recession driven by a financial crisis will not be needed to bring such a change.

[1] The most commonly used definition of the start of a recession is two consecutive quarters of declining real GDP. The National Bureau of Economic Research uses a more complex measure of economic decline to denote a recession. GDP per capita and GDP are interchangeable for this purpose since population change is not measured in these time increments.

[2] GDP includes gross investment and government consumption as well as incomes, so GDP per capita is a larger number than mean income, but uniformly larger.

[3] Fed monetary policy is influenced by other factors as well.