The American incomes debate

This is my Perspective column for this week’s issue of Fund Strategy.

A strong challenge has emerged to the widely held view that living standards for the average American stagnated for many years even before 2007. This is an important debate to examine because it suggests that most people did not become more prosperous even when America was growing.

The standard argument concedes incomes rose roughly in line with economic growth during the boom years from the late 1940s to the early 1970s. Naturally, a gap between rich and poor remained but it did not significantly widen during this period. Most Americans enjoyed rising prosperity as the economy grew even though some were much richer than others.

It is then argued that things went horribly wrong from the mid-1970s onwards. Not only did economic growth rates slow – a point often downplayed – but a small elite captured the gains from rising prosperity. Inequality widened while average incomes were squeezed. From this premise, it is often argued that America entered a new “gilded age”; a winner-takes-all society in which the wealthy enjoy the spoils of success while others get next to nothing.

Barack Obama has broadly endorsed this view several times. For instance, in a keynote speech in Osawatomie, Kansas, last December he said: “Over the last few decades, the rungs on the ladder of opportunity have grown farther and farther apart, and the middle class has shrunk.”

Similar views about the patterns of American prosperity are common in Britain. In November 2011 The Guardian argued: “The US economy is now almost thrice as big as in the early 1970s – and yet the typical working man finds not a dime of this transformative growth in his pay packet.”

There are many sources of data on this question but the work by Thomas Piketty and Emmanuel Saez, two French economists, has become an important benchmark. Piketty and Saez would no doubt dispute some of the more extreme political conclusions drawn from their data. But it is necessary to examine the figures as dispassionately as possible to try to identify its strengths and weaknesses.

I was therefore fascinated to hear a recent interview with Professor Richard Burkhauser of Cornell University in Ithaca, New York, on the invaluable weekly Econtalk podcast series (available at www.econtalk.com ). Burkhauser has re-examined the official data on the subject and has concluded that on different – he would argue more realistic – assumptions, median incomes in America from 1979 to the onset of crisis in 2008 did not fare nearly as badly as Piketty and Saez suggest.

The accompanying table (see link here ) neatly sums up Burkhauser’s point about assumptions. Median income for each sharing unit between 1979 and 2007 is estimated to have risen between 3.2% and 36.7%, depending on the assumptions made. One set of assumptions relates to what is included as median income. The lowest estimate for each tax unit looks at the figures before tax and without including any other transfers, such as social security payments. Look at the figures after tax and transfers, throw in health insurance, and the rise is considerably more.

The other set of assumptions relates to the “sharing unit” measured. For instance, a couple who live together and both work would be counted as two tax units but one household. Adjusting for the size of the sharing unit also makes a difference.

There are also more specific technical factors to be taken into account. Evidently the Current Population Survey, the American Census, was redesigned between 1992 and 1993 to capture more data. Health insurance information was also unavailable before 1988. Burkhauser therefore assumes that the value of health insurance from 1979 to 1989 rose in line with post-tax, post-transfer income.

A debate on the relative merits of the Piketty-Saez and Burkhauser data alone would be a valuable enterprise. Hopefully, the two French economists will in time respond to the criticisms and the outcome will help to get everyone closer to the truth of the matter.

It would certainly be wrong for people to let their political views colour their view of the data. Many liberals (in the common American sense) will no doubt feel more of an emotional affinity with Piketty-Saez while conservatives will favour Burkhauser. But it would be far better if the debate focused on getting the best possible picture of what is happening to American living standards.

Important as this discussion is, there are also other dimensions to measuring living standards. For instance, does the consumer price index accurately reflect inflation? Clearly, the inflation rate is necessary to convert nominal incomes (their face value) into real (that is inflation-adjusted) incomes. Another question is whether the various surveys of income accurately capture all the necessary data.

Other debates go beyond the correct statistical measure of income. Some experts argue that what people consume, rather than the money they receive, is a better indicator of living standards.

In my view, it might also make sense to recalculate the economic growth rate over the years of apparent boom. In a sense the debt-fuelled growth in the years running up to 2007 can be understood as based on borrowing from the future. It might therefore be better to look at the average growth rate from the 1970s until now – that is including the recent downturn – to measure the real trend rate of growth.

Such a procedure would no doubt show slower growth of both the economy and income. But perhaps that would be a better description of the trend than the assumption that the two variables were dramatically out of line.

The more constructive the debate, the closer it will be possible to get to the truth on these vital questions.