The Ambrosini Critique

Prof. Kenworthy responds to Wilkinson’s essay on inequality and mostly punts. He goes after the idea that consumption inequality should matter and not income inequality. His argument basically boils down to the idea that he doesn’t think we measure consumption well. He’s right about consumption surveys missing outliers like the very very rich because they use random sampling. Also, we don’t include walks on the beach and leisure time in consumption statistics ((Wilkinson actually addresses this issue, if I remember correctly, in his recent policy paper regarding inequality. He shows data that suggests richer folks have less leisure time these days than poor.)). So, like always, we need better data.

Kenworthy’s reply doesn’t address Wilkinson’s main point about the importance of considering underlying mechanisms. He does list some bad things that could, might be, maybe be related to income inequality, but he doesn’t give any reason — besides listing them in a article about inequality — to make us think they are related to inequality.

And in a strange twist, he ends his comment saying the real problem is “poverty” and not income inequality at all:

Imagine an America in which high-quality public services raise the consumption floor to a high level: most citizens can put their kids in high-quality child care followed by good public schooling and affordable access to a good college; they have access to good health care throughout life; they can get to or near work on clean and efficient public transportation or roads with limited congestion; they enjoy clean and safe neighborhoods, parks, roads, museums, libraries, and other public spaces; they have low-cost access to information, communication, and entertainment via reliable high-speed broadband; they have four weeks of paid vacation each year, an additional week or so of paid sickness leave, and a year of paid family leave to care for a child or other needy relative. Even if the degree of income inequality were no less than today and we still had CEOs, financiers, and entertainers raking in tens or hundreds of millions of dollars in a single year, that society would be markedly less unequal than our current one.

I put poverty in scare quotes because its clear he’s not talking about Dickensian/sub-Saharan African poverty, but a relative sort. Its true that I like it better, and apparently so does Kenworthy, when the folks in my community consume at a similar level of quantity and quality as me, but that they don’t isn’t necessarily, in such a rich society, a sign that they aren’t able to do so. People that aren’t me have weird tastes and that drives them to do weird stuff that I wouldn’t do. They’re not poor because of it, just weird. And any case, what does “poverty” or poverty have to do with inequality?

As dictator of the universe, I would force everyone to read the second part of Will Wilkinson’s essay; the part on why mechanisms generating income inequality, not inequality itself, are what matter. This is mostly because I’m writing a paper on one such mechanism and, hey, publicity!, but also everyone should read it because, well, what Wilkinson says is true.

You like truth, don’t you?

Income inequality is one of those topics that makes smart people dumb. When asked why income inequality matters to him, a once proficient and smart (he’s still smart, no doubt) internet commentator wrote on this blog “I don’t really care why, I just do.” A short-circuit of thought: income inequality is bad, QED.

Inequality matters is incorporated into the very structure of the mind ((This line and the post title are taken from this essay. While the choice might be politic, Wilkinson dwells on correctable mechanisms generating inequality like unequal education, taxes or even norms regarding renumeration of CEOs. We should be prepared to confront innate inequalities, too. )).

Alas, I’m not yet the dictator of the universe. Go read that section (or heck the whole thing) anyway. Its not a nitpick of the empirics. Its not apologetics for the status quo. Its not a tired retread of right-wing talking points. At the very least, read it as the most advanced arguments your ideological opponents can throw at you so you might prepare your own counterarguments.

Read it!

That’s not the question Will Wilkinson asked, but until more disaggregate data becomes available we won’t know the impact of the recession on inequality. What’s more, we only have data at the State level through 2008 so we don’t even see the brunt of the badness. Here’s State income levels in 2007 plotted against income growth over the next year:

There’s a statistically insignificant and small negative relationship between initial income level and growth rate.

Beware the ecological fallacy (in reverse).

Mike asks for median wage-age profiles. I don’t know how to do quantile regressions for panel data, but I have a second best for him. Here’s wage-age profiles with the top 90th percentile wage earners removed:

And here’s the wage-age profile with both the bottom 10th percentile and the top 90th percentile removed:

Is this the best way to determine optimal policy?:

But the current magnitude of inequality in America strikes me as unfair.

Prof. Kenworthy goes on to say:

What’s the proper amount of income inequality? I don’t have a precise answer, but that doesn’t mean it’s wrong to feel that our current level is excessive.

With all due respect, except in dictatorships one person’s gut feelings about something doesn’t translate at all to optimal policy. In fact, there’s impossibility theorems that say one person’s gut feelings don’t have to monotonically effect aggregate gut feelings.

(Irony alert: to play with vote aggregation examples in writing that last sentence, I was, without realizing it, doodling on my just arrived voter information guide.)

Regarding Kenworthy’s point 2 about inequality causing harm. Besides the rhetoric of it, I’m not sure a consumption arms race causes harm. As for supposed “real” effects of inequality (e.g. health outcomes), I want to know how much of that is really an effect of poverty. I’m down to reduce poverty, but that’s not at all the same as reducing inequality.

The red line is the wage-age profile from 1990 to 2005 and the blue line is the wage-age profile from 1968 to 1980. I just picked those years randomly. These profiles were calculated in such a way that what you’re seeing is average “within person” wage profiles over their lives ((I estimated this equation with 18 year olds the comparison group. The y-axis are the estimated two-way panel coefficient on the age dummy plus the average log wage for 18-year olds in the appropriate time period. Everything’s waaaay significant. These are heads of households with positive wages in the PSID. R code and the data set is available but its too big for me to post my hosted account so email me if you want it.)). This means there’s no funny business with changes in demographics or whatever:

In the good ol’ days, workers ramped up their wages early in their careers and then wages flattened out for the rest of their careers. In these evil dark ages of widening inequality, it takes longer for workers to get to their wages to peak and the peak is higher than before. Also, the peak comes so late there’s never a period of stagnant growth in their wages.

Commenter slocum at CT:

The reason tax rates aren’t higher and bankers are getting bailed out on hugely generous terms isn’t because Rawlsians have outvoted Cohenites behind the veil of ignorance, or even because lots of economists believe the Laffer hypothesis. It’s because the rich and powerful are, well, rich and powerful.

No. Tax rates aren’t higher because of the widespread belief among U.S. voters that high tax rates are harmful to economic growth. In places and eras where those beliefs were less prevalent, tax rates are higher (despite the fact that those eras and places also have their own rich and powerful). That charges of ‘class warfare’ and ‘the politics of envy’ are effective is a result of those general beliefs. In other places and eras, class struggle was seen as a positive by enough voters, for it to be a vote-getter rather than the reverse.

As for the bailouts—does Brad Delong’s clear support for the latest Geitner bailout plan indicate he’s one of the self-serving rich and powerful (or is being paid to serve their interests, or suffers from false consciousness)? Or is it possible that the main motivation for the bailouts comes from the fact that those in charge are scared shitless (justifiably or not) about the potential for general collapse if the banks are not bailed out?

Lastly, what all y’all Rawls fans will never seem to grasp is the declining returns of high income in an increasingly wealthy society. Because of that, the lives of the rich and poor have been converging (pretty dramatically) in material ways—gini indexes notwithstanding. In the 19th century, the lower classes were physically smaller because they were malnourished. 40 years ago when I was a kid, the middle classes had clotheslines instead of dryers, black and white TVs, window fans instead of air-conditioning, rarely ate in restaurants, and most had never traveled on an airplane. Living in bog standard suburbia not rural Appalachia, I knew kids whose families who always drank powdered milk rather than fresh to economize.

But what kinds of material possessions do the wealthy in the U.S. now have that the lower classes do not? It’s hard to come up with much—yes, the rich have fancier versions and more prestigious brands, but that’s mostly it. Despite the name, Sub Zero refrigerators don’t keep your beer any colder. Material differences in living conditions have been reduced pretty dramatically even as nominal dollar inequality has grown, and they’ve been reduced by increasing societal wealth overall, not redistribution.

On the other hand, I have some relatives about my who live on wages that are, say, 1 1/2 to 2x minimum wage. You could double their wages, and they couldn’t live my lifestyle, whereas you could cut my wages to their level and I could (and did so for many years as grad student). In my late 40s, I enjoy the same vigorous activities I did in my 20s, and those cost very little (hiking, backpacking, biking, canoeing). Pandora, Project Gutenberg, and my $8.99 NetFlix subscription keep me well supplied with entertainment. They, on the other hand, are overweight smokers, don’t eat well, get little exercise, and so are developing the typical chronic health problems —the years have taken much more of a toll. They’re also not good at deferred gratification, tend to be impulsive shoppers, and so always have credit card debt problems. You get the picture. In other words, those differences that remain—and they’re not trivial—are not ones that could be readily ameliorated by transfer payments. Most of the difference that matter would persist. As societies get wealthier, it is increasingly the case that the quality of your life depends on who you are rather than how much money you have—what would Rawls say about that?

Now, you might say—well, if you don’t really need as much money as you make, then we should raise taxes on higher earners and redistribute the proceeds. But the problem is that because higher incomes bring diminishing returns, higher tax rates will have more of a discouragement effect than they would otherwise. Even at my existing marginal rates of ~50% (when you figure fed, FICA, state and local), more leisure looks pretty attractive. So I’d argue that the poor are better off if you keep marginal rates low enough that the wealthy keep working hard rather than kicking back and relaxing more. But if increasing tax rates make it more normal for people like me to demand additional leisure rather than additional income, that would be OK, too. Societally, I think it would be a mistake, but for me personally it’d really be OK —I can adapt.

Some folks at Princeton have found, using panel data from England, that each additional inch in height is associated with 1.5-1.8% more wage income. Their suggested mechanism is having wealthy parents that feed their kids right and make sure they get more and better education leads to taller and better paid kids.

Its strange to me that to test the hypothesis that having wealthy parents leads to better economic outcomes, e.g. income, education, etc — a list of dependent variables that gets longer as Economics’ empire expands — these researchers had to use height data as a proxy for parental income. Economic historians often use height as a health, and thus an income, proxy because its very hard to find direct measures of wealth and income. This is an ok proxy because its plausible in the Malthusian period most variation in height was caused by variation in nutrition and its plausible that most of that variation was due to differences in wealth. In the modern context, most variation in height is due to genetic factors (I think… razib?) and I’m guessing most variation in nutrition are due to social factors (of which income group is one of many). Height, then, is a terrible proxy for parental income in the modern period.

For a modern test of this hypothesis, getting direct measures of parental income has to be a least as easy as getting height data. Why not directly test their hypothesis? At most, these guys have found evidence that doesn’t contradict their hypothesis. But random noise doesn’t contradict it either. Which is to say they haven’t moved me far from my priors.

Historians are bold with their findings, often telling a story of a long causal chain using a single correlation or two. I don’t mind it when they do this because they’re not working with much and not to mention, from my experience, it makes them colorful folk. Labor economists don’t and shouldn’t have this luxury.

(h/t economist)

Gabriel summarizes the debate on fiscal stimulus and comes down on John Taylor’s side (good choice). Those checks we got in the mail from the Feds didn’t increase aggregate demand; while take home pay went up, consumption didn’t. People, on average, put that money in savings (e.g. paid down their credit card debt).

This conforms to standard economic theory, but not, by the way, our cherished models ((Won’t anyone defend Keynesian fiscal stimulus?)). People that win a lottery will tend to put most of their winnings in savings and only consume a small bit at a time because they know its a one time windfall. On the other hand, people who get a raise at work will tend to consume most of the additional pay because they expect the additional pay to be in each of their future pay checks. ((Hey, what about Ricardian equivalence? In what sense aren’t all fiscal policies permanent ? Contra Taylor, “permanent” tax cuts without out spending cuts aren’t credible.))

These theoretical results depend on people having access to credit. Hypothetically, suppose you’re a poor student expecting to get a lavishly paid job when you go on the job market next year because you expect no Economic department budgets will be cut in the aftermath of the Great Depression 2.0. Let’s say. Anyway, you know you’re income will be much higher in the future and you’d like to spend some of that higher income today. There’s no reason for your future self to live high on the hog while your stuck in the present eating Top Raman. In other words, you’d like to borrow from your future self. Well, that’s not very likely because access to credit for poor students is limited.

Now, suppose the government sends you a check for $500. Even though this is a temporary increase in your income, contrary to that fancy economic theory above, you’re going to consume it all and not save any of it. You do this because you wanted to borrow money from you future self, but weren’t able to. You not saving is, in effect, you borrowing from your future self.

But darn us economist, here we are talking about efficiency again. What if government actors care about maximizing the consumption of the least well off? Sure their stated preferences, as evidenced by their reference to Keynsian stimulus in public statements, are for efficiency. But stated preferences and a smile buys you crappy happiness research. Instead the government may just care about those most hurt by recession and those most likely to be credit constrained and thus those more likely to increase their consumption after fiscal stimulus.

Does our Benovolent Dictator (with a mouse in his shirt pocket to make my title work) have maximin preferences? If his preferences reflect a typical individuals social preferences or the median voter’s typical social preferences, then my guess is yes.

Prof. Peri spoke at GMU today about immigration. There’s an economic argument against immigration that says because most immigrants are unskilled, they reduce the wages of native unskilled workers. These are the most economically vulnerable members of society, so we should care about the effect immigration has on these folks.

Well, one solution would be to tax immigrants and give the proceeds to native unskilled workers. This would make everyone better off because the natives would be compensated for lost wages and the immigrants would get to work in America where their wages are much higher than in their sending country. Besides being a little perverse, this policy would be hard to implement. You’d have to figure out who exactly is “low-skill” and you’d need to spend money to collect taxes from the immigrants. Plus, you’d get some inefficiency from the fact that people would switch from “high-skilled” to “low-skilled” jobs so they could get in on the action.

Apparently, Prof. Peri offered another solution in today’s talk. (GMU is suppose to be all internet friendly… where’s the video feed?!) Let more high skilled workers into the country. Open the flood gates. These engineers from India would compete with high-skill native Americans, thus increasing the (relative) wage of low-skill workers. Problem solved!