The Great Distortion

by Russ Roberts on March 4, 2011

in Data, Standard of Living

From Tyler at Marginal Revolution:

The median earnings of full-time Canadian workers increased by just $53 annually — that’s right, $53 annually — between 1980 and 2005.

Here is more, or here.  This is one reason why I do not adhere to some of the progressive or “class struggle” explanations of relative stagnation in median income growth.  Canada is not ruled by the so-called Republican Right.

Tyler argues in The Great Stagnation that the slow-down or stagnation in median income growth is due to a slow-down in technological change and innovation.

Perhaps. But the numbers on median income are distorted by demographic change. When divorce rates rise, women who were not in the labor force suddenly find themselves looking for work. The number of workers and the number of households rises faster than the population. And the median slides down as a statistical artifact–there are many new households created where the head of a household, typically a woman, earns less than the median.

So changes in median income over time do not capture progress for any one person, certainly not the person who was at the median in 1980.

Here are statistics on divorce in Canada:

Years No. ofdivorces Rates per 100,000 population. Rates per 100,000 Married Couples
1921 558 6.4 N/A
1941 2,462 21.4 N/A
1961 6,563 36.0 N/A
1968* 11,343 54.8 N/A
1969 26,093 124.2 N/A
1981 67,671 271.8 1,174.4
1985** 61,980 253.6 1,103.3
1986 78,304 298.8 1,301.6
1987*** 96,200 362.3 1,585.5
1990 80,998 295.8 1,311.5
1994 78,880 269.7 1,246.3
1995 77,636 262.2 1,221.9
* Reform of Divorce Laws
** Divorce Act (“no fault”)
*** Peak year

I didn’t make the chart. I don’t know why they left out he 1970s. But it’s clear that divorce rates surged in the 1980s if not earlier and stayed very high. My hypothesis is that many of the women who divorced found themselves in the bottom half of the income distribution and that pulled the median down.

So a snapshot of median income in 1980 and 2005 that reveals essentially no change is very misleading. They are not the same people. Not because some have died or been born in the meanwhile, exiting and entering the data sample you observe, but because there has been a radical change in the nature of the family and who is in the workforce. There has been an increase in workers and households well above the rate of population growth, and that increase is concentrated among below-median earners. That means that if you looked at the median income worker in 1980, that person could have made great progress over time but it is obscured and distorted when you look at the two medians over time.

A similar pattern occurs in the US starting in the 1970s. If you look at panel data in the US–that is if you follow the same people over time, there is growth in every quintile of the income distribution and the biggest growth occurs among the poorest households. This can be true even though the median is unchanged across all workers. These data are from the Pew project on income mobility, comparing family income from the late 1960′s to the late 1990;s, when supposedly the median stagnated and the rich got all the gains. The panel data tell a very different story relative to the snapshot of the two medians over time.

Does anyone know of panel data for Canadian workers or households?


Be Sociable, Share!



44 comments    Share Share    Print    Email

Previous post:

Next post: