The media is really bending over backwards to find ways to twist earnings data for average Americans to try to make the point that real income for many folks has stagnated or dropped. They are doing this to support a two-pronged legislative strategy in the next Obama administration:
- Use the power of the government to further tilt the balance towards unions and against employers in wage negotiations (this strategy having worked out so well to create prosperity in the automobile and airline industries)
- Further modify the income and Social Security tax structures to make them even more regressive than they are today.
They are firing on all cylinders behind this strategy. They are even mobilizing the neo-Keynesians to make the pitch that the Great Depression and the current financial crisis were caused by a shift in wealth from laborers to the capital classes, and that the only way to prevent future crises and depressions is to, wait for it, increase the power of unions and institute more wealth redistribution (Example here, via Kevin Drum).
I was going to do a post fisking the James Livingston article linked above on Kevin Drum's site, but Livingston's hypothesis was such a mess that it was just going to take too much of my day. But in doing some research, I found this chart from a couple of years ago in the NY Times that really caught my attention:
Talk about chutzpuh -- look at the lede on the chart and then look at the chart itself. Yes, the lede is correct, but only if you choose the totally meaningless number of "cash wages" rather than total compensation. If one looks at total compensation (or what they call "overall" compensation), the entire argument falls apart. Workers have maintained about their same "share" of the economy.
Sure, a large percentage of that is now in health care benefits, but that's a choice workers have made (and the government has encouraged through tax policy). In fact, this compensation mix has been driven in large part by the Left's beloved unions, so on what basis can folks say that these other benefits somehow "don't count?" Certainly, they cost their employers equally, whether it is cash or health care. Corporate profits are up a bit, but in line with their normal historical levels in the 1950s and 1960s, the golden age of the US economy, according to the Left. (By the way, the pattern of falling wage shares and rising profit shares after recessions is a well-documented one. Wage-earners do best at the end of an economic cycle, employers more towards the beginning. The chart cut off after 1997 would look about the same as the last several years).
I will tell you right now that every time you hear someone bemoaning the stagnation of wages, they will never, ever, ever be talking about total compensation per individual. Having, through government policy and union activity pushed the compensation mix to non-cash elements, they then play a heads-I-win-tails-you-lose game of not giving any credit for those compensation elements.
Other games that are played to try to make the case that real earnings have stagnated include:
- Time frame selection. Everyone making this argument will choose 2000 as a starting point. They justify it by saying it is the beginning of the Bush years, but 2000 is really selected because it is a pre-recession peak, and they have to measure peak-to-trough of the economic cycle to try to make their point. Just as an example, if you look at the household income numbers below, you can see there is very typically a 5-year drop after a recession followed by net gains. If we chose, say, the first Clinton term we could play the same game, showing a peak-to-trough drop in real incomes.
- Household income game. The household income numbers are fraught with peril, because companies don't pay households, they pay individuals. And household makeups are changing simultaneous to income changes. For example, imagine the economy was just my household. If my wife were to get fed up with my shtick and divorce me tomorrow, average household income would drop by 50% in one day (as our total income stays the same but we go from one to two households). If my wife were to go back to her high-paying pre-kids job tomorrow (if only it were so!) our household income would go way up, in part because the labor department does not capture the value of the labor she provides at home.Mark Perry has a lot more on the household income numbers here, but he shows that the household size number has been changing a lot, causing the metric to understate income changes per individual:
- Individuals matter. Median income looks at the middle person on the ranked list of US incomes. So, for example, if there are 100 million income earners, the median income is the income of number 50 million on this list. But whoever the person is at spot 50 million is almost certainly not the same person who was at spot 50 million last year. They might have fallen on the list, but the odds are they moved up. As folks age and gain experience and/or seniority, they tend to increase income faster than inflation. Most minimum wage earners, for example, tend to be under 25. The number of families supporting three kids on minimum wage (at least of the primary bread-winner) at the age of 45 is really, really low, despite the anecdotes we are bombarded with in the media.
- Immigration has a huge effect. The total number of foreign born people in the labor force is estimated around 21 million, of which perhaps 6.3 million are illegal immigrants. Positing that at least 10 million of these arrived in the last two decades, and that many of these folks began at relatively low, below-median incomes, means that median incomes are hugely affected by immigration. Leaving immigrants out so the comparison is close to apples and apples, to find the true median income gain over the last 20 years one would have to count up 10 million or so spots on the list. Again, as in the previous point, most individuals can be better off even if the median stagnates (presumably immigrants coming in at the bottom are also better off, even at the bottom, than where they were before, or they would not have come. We often forget that much of our bottom quartile of income in this country would be upper middle class in many other nations). This is a classic mix problem that most people, and the media, almost always get wrong. In a situation with a changing mix of multiple groups, each of the groups can be improving on some metric, but the overall metric can go down. You can see the income stats by race here. Every race group has increasing median income, but since the Hispanic group has grown 8x faster than the anglo population in the US, the total results are mixed downwards.Here is a quick example. Group A has values of 5,6,6,7. Group B has values of 1,2,3. Ten years later Group A is the same size and has values of 6,7,7,8. Group B has doubled in size, and now has values of 2,3,4,2,3,4. In these examples, every single individual has a higher value. Also, Group A's median has increased from 6 to 7, and Group B's has increased from 2 to 3. But the median for the whole combined group A+B has dropped from 5 to 4. Both medians (and averages) can do funny things when mix is shifting.
- Even the NY Times. The NY Times actually makes two of these points for me in another article, arguing that historic median income drops were concentrated in areas of high immigration, and reported drops were due to the choice of the economic peak as a starting point. WOW? Is this the same NY Times I began this post criticizing. Yes it is, the only difference is that this article ran in 2001, when they were reporting on the economy during a Democratic administration.
- Income taxes are already wildly progressive. While I would love to be in that top 1% group, I don't really begrudge them their success. Besides, who can look at the chart below, again from Mark Perry, and come to the conclusion that the top 1% are being treated unfairly generously.
- Every country that has implemented this plan (government-backed unions and wildly progressive tax policy), including most of Western Europe, is demonstrable worse off than the US on absolute measures. This is both the median, but also in every quintile, including the poorest. While it is true the poorest quintile has a bigger gap from the riches in the US vs. France for example, on an absolute basis our poorest are at least as well off (particularly when differences in immigration policy are taken into account).