Kind of an interesting little chart. Is it conceivable that most of our political narratives are nonreflective of economic reality? Don’t answer that.
-
Archives
- January 2010
- November 2009
- October 2009
- September 2009
- August 2009
- July 2009
- June 2009
- May 2009
- April 2009
- March 2009
- February 2009
- January 2009
- December 2008
- November 2008
- October 2008
- September 2008
- August 2008
- July 2008
- June 2008
- May 2008
- April 2008
- March 2008
- February 2008
- January 2008
- December 2007
- November 2007
- October 2007
- September 2007
- August 2007
- July 2007
- June 2007
- May 2007
- April 2007
- March 2007
- February 2007
- January 2007
- December 2006
- November 2006
- October 2006
- September 2006
- August 2006
- July 2006
-
Meta
I think those charts are a bit misleading for two reasons. First, one has to consider the fact that in the lowest income quintile, there will be many college students who are working part time while going to school. They may come from middle class or upper class families, so they were never really part of the so-called permanent underclass. A similar comment could be made concerning a spouse who works part-time. Perhaps he/she starts working full-time, and this vaults him/her up to a higher earning power. In other words, the unit of analysis that is relevant here is the family, not the individual. It would be interesting to see how much mobility remained at this level of analysis.
Second, the skew of growth is not as fantastic as one might initially suppose on seeing the big percentage point differences. The lowest quintile saw a 90.5% growth in median income, but from a lower base median income than the next highest quintile. If measured by the Gini coefficient, I believe inequality worsened during this period.
Finally, the whole notion that a progressive tax system is Marxist (as is repeated multiple times in the IBD op-ed) is simply baseless slander.
Ryan,
A better point using what you say is the Gini coefficient is misleading. Your evidence shows that when we speak of quintiles by median income we are not talking about the same people over time. That is what these charts show. The term “permanent underclass” is used ironically here, because it hardly exists. Let me give a demonstration.
We can imagine a society where every person makes the same salary. There would be no quintiles.
Now, we imagine a society where every person makes the same income based on age. The income goes up every year as one gets older. As soon as one does this inequality increases dramatically, even though nobody earns any more than anyone else over their lifetime. Perfectly equal, but the gini coefficient increases by a large amount.
Now assume another society where incomes as one increases in age pay far more relative to the other age based pay society. The gini coefficient worsens dramatically yet again, though it is still a completely equal distribution of income overall.
As your examples show, this effect, college students, people becoming wealthier over time, etc., has a large effect on income. The evidence that the charts above show is that people at the lower end move up from being college students, recent immigrants or just young, stupid and irresponsible (been there, done that.) Over my lifetime I might move from living with four of my buddies in an apartment on Chimes street, to a struggling head of household, to successful business person, to extremely wealthy. To the extent that incomes over time increase (in a really successful economy one would expect that the returns on ones labor and human capital would increase relative to less succesful economies) as people move up due to ending their education, not living a decadent lifestyle on Chimes street, becoming more responsible, more skilled, have the opportunity to hit it big in a business, learn English, assimilate, network etc., that would pretty much bake into the cake an increasing gini coefficient.
Looking at income distribution in terms of quintiles and gini coefficents actually penalizes societies which have high upward income mobility such as ours. The charts above show just what I am taking about. Talking about inequality in the way is typically done is looking at the question completely wrong.
Given that there are only roughly 2.5 million college students in the US at any given moment… and that this would represent about 8% of a fifth of the total US workforce, this seems to be a relatively unhelpful distinction. A qualitatively selective one at that. Because if you cut them, I’d immediately motion on identical grounds that we also cut structurally unemployed skilled workers, who are undergoing commercial workforce retraining. And we can play that game all day until we are left with almost no one at all in the lower income bracket. I think you have to take the aggregate to see any trend that isn’t quality biased.
Or you could just conclude as Lance did, that in an open mobile society, everyone qualifies as an exception.
Lance, I appreciate the response. I agree that the Gini coefficient is an imperfect measurement, as it doesn’t take into account mobility over time, which is usually considered to be quite high in the US. However, I would suggest that a high Gini coefficient in and of itself – even with a decent amount of mobility – is problematic. There is often a lot of attention on getting the incentives right for high levels of growth, thus the concern that raising income taxes on higher income brackets will hurt growth. However, there is some evidence – although it is inconclusive – that high inequality as measured by the Gini coefficient is bad for growth. Francois Bourguignon sums it up in this article. Now, if you think on principle that the notion of redistribution is wrong, then it doesn’t matter what effect it has on growth. However, if your primary concern is about growth – and I will admit that this is my primary concern – then you can’t be completely indifferent to Gini-style inequality.
The second point I’d like to make is that the U.S. does not in fact have as much mobility as it used to, if one looks at the unit of the family. If you look at table one from this paper from Brookings, you’ll see that intergenerational growth in incomes was heavily skewed toward the wealthy. The data are from the Panel Study of Income Dynamics and compare income of parents and their children across time. Generally, the data suggest that intergenerational mobility is rather moderate. I think if we use the phrase “permanent underclass”, then we have to consider intergenerational economic mobility and not just an individual’s mobility over time. But that’s obviously my particular viewpoint.
Lee points out that my example of college students was selective – and indeed it was. I was simply trying to point out that I didn’t find that first table particularly helpful – we would need to know a bit more about the people in that lowest income quintile. Perhaps Lee is right that college students or part time workers make up too small a number to make any difference or that they are outweighed by structurally unemployed workers, but there’s no way to tell from that table.
Nice response Ryan. Please keep coming back.
First the study from Brookings. I will get back on it after I have had more time to confirm my impression, but it seems to make the same compositional error that I discussed earlier. However, I am quite willing to believe that its main thrust is correct anyway. Generally, I believe such things will vary over time, and it doesn’t show a systemic problem (granting for the sake of discussion that it is a problem.) I’ma just not ready without more time to examine this paper to say it shows that to be true. I am personally of the opinion that most studies I have seen suffer serious methodological issues that may not be able to be overcome. It may not be possible in a country such as ours, or possibly anywhere, to say anything definitive about the issue. What I do think the chart above, with all of its limitations, shows is that speaking about quintiles as static entities looks at the problem in ways that are very misleading, and that is 99% of what drives the debate about inequality. I appreciate that you are looking for a more nuanced and realistic apporach.
That however isn’t a criticism of the chart, which was not attacking that more nuanced approach, but the naive way that this discussion is usually approached, generally by left of center academics, researchers, politicians and pundits.
Francois Bourguignon’s paper is another story, as its empirical conclusions are ones I would assert are true. However, the reasons for it being true do not support directly attacking income inequality in the US, assuming your interest is growth. I am not as concerned with growth, I think the actual lived condition of the majority of the people, including especially the poor, is worth sacrificing a little growth. Working in Finance myself, I see a reason to limit hedge fund managers and other highly paid people if it does lead to the previous stated goal. I just don’t think it does. Or, I should say to the extent it is due to regulatory and other barriers that give a certain amount of rent seeking opportunities that should be addressed (such as licensing) as opposed to mere redistribution.
Which goes to why the relationship in the paper is true. Highly unequal societies in the developing world often have enormous barriers to mobility, not a lack of redistribution. For a more overarching look at the research in this field, and possible rationales for why it is true, and an examination of difficult to explain data if inequality itself is the probelm, go here.
I am very sympathetic to this interpretation of the data and research:
To put it another way, correlation is not causation. That is important in understanding developed economies. However, in developing countries, which is what the study is talking about, inequality is caused by rigidity and poor institutions, and poverty in such an unequal society restrains growth due to lack of ability to overcome barriers set up by economic and political elites. Hernando De Soto documents this impressively in his work.
So I think it is highly debateable whether such work applies to the US. It may though.
One way it may is that what may explain inequality growth in the US (assuming “bad” inequality, vs. the kind I discussed earlier, has risen) is an increase in institutional rigidity and poorly performing maket mechanism’s. Ironically, most of that has in this country been by attempts of the government to regulate industries and protect them (the most newsworthy of the moment being the long time subsidized rent seeking for benfit of shareholders, and especially management, of Fannie and Freddie) for “our benefit.”
Hello Lee,
Yes, I would like to chime in! A couple points. One is that I agree very much with your discussion regarding the Gini coefficient and earnings over a lifetime.
Similarly, the difference between absolute and relative income mobility is often sidestepped although it makes a huge difference. True relative mobility studies just look at changes in income classes (first quintile to fifth quintile) and those quintiles change over time and between countries. If incomes are more equal overall then relative mobility is going to be much higher, even if absolute earnings increases are equal. Advancing from the bottom quintile to the top quintile in a very equal society, like Sweden, may only advance you to the third quintile in a less equal society like the US.
Second, mobility over a lifetime is generally not counted in income mobility studies, because the studies compare incomes to the “peers” of the parties being surveyed. This is because mobility is usually defined in relative terms, as “class” or “social” mobility – as in, how well someone is able to move from the worst off to the best off, compared to others around him.
So, if all Americans can increase their earnings over a lifetime equally, with those starting out under the Census poverty line at the start of their career earning $200,000 10 years later, but those starting out at $200,000 at the same age also earn more 10 years later, there may be no mobility recorded. It is only a relative advance that counts.
Taken together, this explains much (or all) of the higher mobility seen, for example, in Sweden.
I wrote a little story about this some years ago here.
The rest of my post was considered spam by your blog server. I will try again.
The second point is that we use inflation in these studies. But, it is worth noting that inflation is a flawed measure, as it is just an aggregate of price changes in the economy, and all income groups don’t spend their money on the same thing. This was discussed in a recent NBER paper, discussed by the Economist here (the paper is linked there). I also discussed the difference between absolute and relative poverty measures, and alternative poverty definitions for international comparisons in this working paper.
Nice blog you have here!
Liberty,
I am glad you stopped by. I was the one who sought you out. I would have e-mailed you, but I didn’t see any contact information.
Ryan,
I contacted Liberty because she actually worked a bit on the study you referenced from Brookings if I am not mistaken. Even if not, this is a topic she studies. I also recommend her blog to everyone, I just discovered it today while looking into Ryan’s questions. Click her name and it will take you there.
Liberty,
Do you mind if I ask you about the Brookings study? I would like someone who works in the field, as opposed to my informal study over the years, to address what you think its particular strengths and weaknesses are. This is a subject of intense interest to me, not only as it relates to domestic policies, but to development abroad. I would even be willing to make it a guest post if you really wanted to dig into the topic.
You can e-mail me at lance at asecondhandconjecture dot com if you wish to consider a guest post rather than an extended comment.
Also, please let me know if you see any weaknesses in my own assertions, I take criticism well.
Lance,
Thanks for the offer! I worked on some supporting data and discussion for the Brookings report, for the Heritage Foundation where I work. It is actually for the new report which isn’t out yet, I think. My contribution was on international comparisons of equality, and relative versus absolute income mobility. I was not a primary writer of the report, just a contributor.
I linked a paper above that I wrote using the Luxembourg Income Study data, if you are interested in reading more of my analysis on this subject. In general, I agree with your assessment, and encourage you to keep working on this sort of analysis as it isn’t done enough. I too hope to work on it more as I have time, at Heritage.
I would not say that I am an expert on the subject, although I have some knowledge about it. Other analysts at Heritage specialize in this area, I do not. I was able to contribute to the report because I had done some work on that particular area. I am actually mostly working in the area of tax policy analysis and economic modeling. My resume is here, if you are interested.
To answer your question briefly, I think the Brookings report (at least the new one) is strong with regard to possible causes of low mobility, and is fairly strong in the area of intergenerational mobility. It was especially weak, in my opinion, in the two areas where I decided to contribute – hopefully making it less weak. In general, the report does not question the traditional measurements of inequality and mobility enough, making it weak in the same ways that many such reports are. But it is broad in its scope, and brings together a variety of viewpoints, which is refreshing.
Sorry that I don’t have more specifics for you.
Guinevere
Ha. You guys have got bloody essay contest going on in my comments thread. Holy Toledo.
I had been busy tracking you down and found it already but thanks. I especially like the Trotsky poetry. Are you related to economist Edward Nell?
Indeed, I am his daughter.
Huh. Funny, that makes sense.
A lot of libertarians probably wouldn’t see that connection, seeing Post Keynesians as being further ideologically from someone who is Hayekian and schooled in Public Choice than mainstream neo-classicism. However, in the way they frame issues, the particular critiques of neo-classicist’s, and the role of institutions, and a host of other issues I see affinity. They tend to end up in different places, but the questions are very similar in many ways. I guess what I mean is, with your obvious affection for Hayek, it was an easy jump for me to associate you with your father.
True. Both Austrian and Post-Keynesian economists have a more dynamic approach, and both, as it turns out, appreciate the complexity of economic systems. My father and I each came to agent based complexity economics separately – and from opposite points of view – and we found an interesting bond through that connection. There are a few economists who even seem to walk a fine line between the two schools.
Actually, those IBD numbers seem a little low. I recall numbers from the census bureau claiming roughly 85% churn out of the bottom quintile from the 1980 to 1990 census, and again from 1990 to 2000. And more interestingly, to me at least, they report an almost identical percentage falling out of the highest quintile over the same periods. If those numbers are true, then we’re looking at a “permanent” underclass of about 3% of US households, if of course by “permanent” we mean a given ten-year period.
yours/
peter.
…Oh, and a “permanent” upper class of about the same size…
peter, part of the reason the census and treasury numbers are different has to do with the difference in methodology we were discussing. Have a look at the treasury report. As for “the same size”, remember for the low end immigration and demographics (youth), and the high end, success. New people become successful all the time.
What I find so interesting isn’t the fact that new people are constantly becoming successful, but rather how many of “the rich” simply don’t remain that way. Now I’m sure the vast majority of the formerly wealthy don’t tumble to the bottom quintile, but still, this fact belies the common over-estimation (and core leftist economic shibboleth) of the general power conferred by wealth. And to the extent that these numbers suggest that temporary high income earners (the actual metric being quantified) aren’t necessarily wealthy in the first place, these high churn numbers also suggest that taxing high incomes and taxing “the rich” aren’t even nearly the same thing.
Lance,
Thanks for such a response. I didn’t come back to look at this post until just now, and, wow, what a number of responses there are. Plenty of food for thought, I will need a little time to digest it all…
No Ryan, thank you. You made good points, and I think we have come up with some things of merit to consider. I wish all discussions could be as amicable and worthwhile.
I also like the challenges. It leads to more thought out rationales. I don’t like an echo chamber.
I can’t believe that I missed your point, I will have to do some research on this.