(For my feathered friend)

There is something mildly insane about the current econo-pundit debate on the state of the US economy and what should be done about it. They are discussing to what extent the free-market structure of the economy despite current problems is nonetheless quite broadly a success, and hence to what extent it should be reformed (if at all).

What's strange about this debate is the standard of 'success' that all parties seem to agree on: GDP, or GDP growth. The right claims that US growth has outstripped that of other less free-market based economies, and the left claims that it hasn't. The right somewhat misleadingly points to absolute GDP growth whereas the left points to GDP growth adjusted for population growth (the US population is growing faster than the EU or Japan). On the former measure the US leaves the competition far behind, but on the latter the gap narrows significantly. But on either measure the US comes off looking quite good:


Conservatives point to an economy that has almost tripled in size over 35 years and Liberals point to an economy that has almost doubled relative to its population over that period - more or less in line with other developed countries. In short, the liberal end of the spectrum of respectable opinion is that free-market economics is successful, but just not uniquely so.

So, why should the consensual focus on GDP be 'strange'? The idea behind using GDP is that

1.    GDP constitutes a rough measure of how much stuff a society produces, and
therefore
2.    the more stuff people produce, all other things equal the more they have, and
3.    the more stuff they have, all things equal the better off they are.

Now, the inference (3) here is really a matter of moral philosophy which you can argue over until your head falls off. And the premise (1) about GDP being a decent measure of the value of stuff produced can also be picked apart in all kinds of ways.

But inference (2) is really a pretty basic empirical matter: GDP is only a very approximate indicator of the level of prosperity within a society since a lot of other things factor in, and these 'other things' are very rarely 'equal'. If you want to measure the prosperity of a society, you measure the prosperity of the people that compose the society.

So if you're using the amount of stuff people have as a standard of well-being, you should look at household incomes, which will tell you how much stuff people can buy - or their 'purchasing power'. If you point this out to them, economists will oblige and direct you to look at the evolution of average, or 'mean', household incomes, where things look less impressive:


 Over 35 years, American households have seen an average increase of 33% in the amount of stuff it can buy. That's still good, though not quite brilliant. Right-wingers, eager to defend the wild success of this economic model, will point out that this income figure only includes cash income, leaving out non-cash benefits such as health care coverage whose cost in absolute terms has increased exponentially over this period. The partial rejoinder here is that you can measure the value of the said health care services independently of their cost: by looking at the health care systems of other countries with comparable outcomes which price those services at half that of the US system. So if the average household gets 15'000 dollars in health care services, those services are worth only 8'000. It's the relative inefficiencies of the US system that in various ways inflate the price of such services. Moreover, the non-cash 'benefits' in other industrialized countries outside of health care, such as job security, labor rights, and unemployment insurance, are hard to quantify though they are clearly of greater value than in the US (otherwise why do we call them 'social welfare states'?).

But, going back to the broader issue, anyone who remembers their high-school statistics class will at this point be reminded of the difference between averages and medians. To jog your memory: if you put Bill Gates in a room full of Mongolian goat-herders, the 'average' wealth in that room may be very high though most people there are nevertheless very poor. If Gates then goes on to steal all their goats, most people in there are even poorer (and likely pissed off...) though the 'average' wealth remains the same. In short, talking in terms of averages can cover up a lot of iniquity.

To get a picture of the overall level of prosperity in a given group, you need to look at how prosperous an ordinary member of that group is, since the wealth might be concentrated amongst some small subset of the group. You need to look at median household income, which you get by ordering the set of households according to their level of income and picking out the income of the household in the middle of that series. So how much better off is the typical middle of the road middle class American family?
 


Now things start looking really embarrassing for all those economists across the existing ideological spectrum who all agree that free markets, free trade, and even to some extent low taxes, are supposed to be wonderful engines of prosperity for at least most members of society. A fifteen percent increase in income for ordinary Americans over a 35 year stretch is hardly something to shout about. It's so embarrassing that economists understandably prefer to talk about that nice abstraction they call GDP growth. Nevertheless, in their defence, fifteen percent growth is still better than nothing, isn't it? Whatever the economic elites are doing, they seem on this measure to at least be making things marginally better. And they will tend to say, in their own defence, that this number is actually skewed by the fact that the size of households has shrunk over that time (from 3 people per household to the current 2.5).

So let us grant them that, and turn away from talk of household income. Let us instead look at individual working adults, the median American worker's income, and see how much more they are taking home if they work full-time:
 

Now you see why I had to use a fancy-pants three-dimensional bar graph: the growth in median wages is pretty much invisible to the naked eye from a horizontal view-point. The median income of full-time workers has not improved at all over the last 35 years: a total of 0.9% in aggregate wage growth. And, no, that's not a typo. The Homer Simpsons of 2007 are making no more than the Archie Bunkers of 1973 (he's the seventies, right...?).

This is pretty shocking. All the increase in income for ordinary families comes from working longer hours. None of the added wealth comes from an increase in the level of wages. Even the Stalinists had figured out how to produce that kind of 'economic growth'.* The fine art of economics was supposed to be about increasing efficiency such that people had more stuff for the same amount of effort. That's what I thought we were paying these people the big bucks for. But all we've got is more stuff through more effort. Not exactly rocket science...

One assumed, or at least I did, that economists - and, on their cue, the media - focused so intensely on GDP numbers because these were a rough but adequate indicator of the level of prosperity in society, that they ignored the actual direct measure of prosperity for ordinary members of society simply because GDP growth approximately reflected the growth in prosperity across society. Sure, you hear every now and then about rising inequality, yet you assume some of those trillions of dollars must 'trickle down'. But these implicit assumptions are wrong. Our 'dismal scientists' actually ignore the real numbers because the real numbers are beyond dismal: they're awful. It's a problem for economists, because they can't keep going around claiming they know what they are doing when they mold society according to the free-market model and show no improvement for ordinary Americans after 35 years of such an experiment. Consequently, they flaunt a number that looks great, yet that number does not reflect what they imply it does. I don't know where these people come from, but where I come from we call that 'lying'.

So the next time you run into an economist, ask him or her whether they think this economic model is a success. And, if so, according to what 'utility function' does

-    throwing another 8 million children under the poverty line**, 
-    making a few thousand old white guys insanely rich, and
-    pretty much leaving everyone else where they were,

amount to economic 'success'?

Or, if he prattles on about the virtues of self-regulating markets, labor flexibility, comparative advantage, just work up your best Cuba Gooding Jr. impression and say

Show Me The Money!

(If you're bored, you can stop reading here. If you have some questions about the numbers presented, keep going...)


Coda. - preemptive responses to some possible quibbles with my numbers

When I first heard these numbers, I found them unbelievable. It just seems impossible to have an economy where workers are 50% more productive than 35 years ago***, and see absolutely no wage gains over that time. So I double-checked (see U.S. Census Bureau, Current Population Reports, P60-236, Income, Poverty, and Health Insurance Coverage in the United States: 2008, cf. p. 43 on median full-time worker compensation which gives you a median worker income in 1973 of $42'088 in adjusted 2008 dollars, and a median in 2007 of $42'480)****. In general, all the numbers cited come from this report or from the St. Louis Fed, unless otherwise specified.

Since I find these numbers hard to believe, I'll assume you find them hard to believe too. So let me just clarify what I have been careful to do to avoid distorting the figures and the trend in any way.

Firstly, it's easy to misrepresent a trend with graphs, i.e. you can draw graphs which make a trend look great if you start your series at the bottom of a business cycle - in the trough of a recession - and end it at the top of some subsequent expansion. By the same token, it's easy to make things look overly bad by starting at the top of a cycle and ending your series at the bottom of a recession. To avoid misrepresenting things, I've chosen the top of the business cycle which occurred in 1973 (the median worker income for 1972 and 1974 were both below $41'000) as a starting point, and the top of the last business cycle before the start of the current recession as an end-point.

Yet note that, if the current recession is as bad as the 1974 and 1981 recessions, then median worker (and household) incomes can be expected to fall 6% from the peak in 2007. If it turns out to be worse and more prolonged than those previous recessions, then they could fall up to 10%, to wage levels last seen in the 1960's.

Here is a graph of the trend lines for these GDP and income measures going back to the 60's
 


Another thing that might distort the picture of wage evolution for American workers is ignoring the possible effect of low-wage immigration, whereby what superficially looks like wage stagnation in fact hides a reality in which ordinary Americans may be seeing wage-increases and are just being replaced by new entrants on the lower end of the wage-scale thus making it misleadingly appear as though there is no progress. So I looked at that as well. If you focus only on the group called 'non-Hispanic whites' in the census figures, you should get numbers independent of this 'immigration effect'. And what do they show? Median household incomes for this group rose 16.9% over the '73-'07 period as compared to 14.5% for the general population, and if you then account for the increase in number of workers per household, you have a still insignificant increase of 3% over 35 years in wages for non-immigrants. So it doesn't look like introducing this 'immigration' excuse into the conversation changes things materially. (There are other ways in which immigration can possibly affect wage levels, but I won't get into that here).

A third point worth making is that I'm using figures for pre-tax income. So if you are to get a figure for median net income - income you can spend on 'stuff' - you need to find out how much less the median family now pays in taxes. I don't have an exact figure for that, though, roughly, payroll taxes have risen and income taxes have fallen. One source asserts that the average net effect of these changes on incomes for the bottom 90% of Americans is a tax cut of only 2% (but, again 'averages' are not 'medians', caveat emptor). So it's possible that changes in levels of taxation do not materially affect the accuracy of the picture I sketch above. But, beyond that, current tax rates are in any case unsustainable given the size of the long-term deficits the government is running. So the marginal improvement that has occurred due to tax-cuts that are not paid for (by a commensurate reduction in government spending) is largely illusory. 

A fourth way in which some uncertainty gets injected into these numbers is estimates concerning the inflation rate. Some believe the CPI overstates inflation, as, say, the improvement in our tech hardware (personal computers, mobile phones, cars) is not adequately accounted for - thereby obscuring real increases in wage levels. Others believe the CPI understates inflation, as, say, falling food quality (think faeces-infested hamburger meat or GM corn that causes organ damage) is not adequately accounted for - thereby obscuring the underlying erosion in real wage levels. I have my own half-baked ideas on this, the upshot of which is that the CPI understates inflation for the typical basket of goods purchased by lower-income brackets and overstates it for the typical basket purchased by higher-income brackets. In which case the trend towards greater inequality is worse than it appears.

And finally, if you think that wage evolution in those socialist hell-holes must surely be even worse than in the US, take a look at this graph. It's just anecdotal evidence, but if you look at the particular case of manufacturing wages (production workers, excluding management, etc) you get an inverse relationship in the rate of GDP growth a country achieves versus the rate of wage growth they achieve: the Japanese, who have had the worst GDP growth over the last 20 years, have had the greatest wage growth.

And with that, I bid you good day.
 
Footnotes.
* To the extent that most Americans have seen economic growth in any tangible way, it's been in the form of Soviet-style economic growth. The USSR in the '50s had an economy that grew at an impressive rate leading Khrushchev to say 'We will bury you' - by which he meant economically outperform us. But that growth came pretty much exclusively from rapidly bringing more people - men and women - into the workforce as a way of boosting production. Once full employment was reached, the economy stagnated. And once more and more of that economy was dedicated to producing stuff like guns and tanks that normal people can't really use, the economy fell apart as the stagnating productivity/production rate combined with less and less of that 'product' going to households to create shortages in household goods. That is what is happening here (or, ahem, at least the first part of that story): median household incomes are up only because there are more workers working more hours in most American households.

** This number reflects the increase in the number of children under the poverty line. The 1973 rate was 14.5%, and it is currently (end-year 2009) running at 21% and rising. 14.5% of 50 million in total child population in 1973 is 7.25 million, and 21% of 75 million children now is 15.75 million children. This gives you an increase of 8.5 million in the number of children in poverty. Of course that does not take into account the cumulative effect of this 35 year trend, which would make the number of children affected over the years much higher. (And, if you were wondering: no, discounting the effect of immigration does not materially affect this trend.)  I should however note that the poverty rate for seniors has been almost halved over the same time-span. See graph here for the evolution of the poverty rate for various sections of the population.

*** I.e. there are now roughly twice as many workers producing roughly three times more stuff.

**** More precisely put, I've taken the weighted average of the median male and median female full-time worker compensation: - 1973: 39,581,000 full-time male workers with a median of $48,452 and 17,195,000 females with a median of $27,440. - 2007: 62,984,000 males with a median of $46,846 and 45,613,000 females with a median of $36,451.