The Agenda

Andrew Gelman on ‘Threshold Earners,’ and More Thoughts on the Same

Andrew Gelman has a different take on “threshold earners.” It is definitely well worth your time. Part of the problem is that he is using my examples to address Tyler’s characterization. My own characterization of the phenomenon would have been somewhat different, but I did quote Tyler approvingly, which could account for at least some of the confusion.

Bottom line: I see the “threshold earners” phenomenon as yet another reason to care less about measured inequality, irrespective of whether it increases or decreases measured income inequality. Here’s Tyler:

What isn’t so widely advertised is that those same critics have basically been telling us, without realizing it, that we should be acting in such a manner as to increase measured income inequality. Not only is high inequality an inevitable concomitant of human diversity, but growing income inequality may be, too, if lots of us take the kind of advice that will make us happier.

I think Gelman is weighing this last sentence more heavily than he should. The part that I strongly endorse is the notion that high inequality is an inevitable concomitant of human diversity. Tyler is far more cautious about whether growing income inequality follows as well. And insofar as he thinks this might be true, it really might be due to scenarios in which market income for people with medium-to-high potential incomes see their reported market incomes fall near zero. (See dumpster-diving, “freeganism,” and other fringe phenomena.) But is this the heart of Tyler’s thesis regarding “threshold earners”? I don’t think it is, though I hope Tyler will elaborate. 

Per Robert Gordon, it’s not obvious to me that our approach to measuring income inequality is very useful regarding the things we, or perhaps I should say I, actually care about, e.g., access to high-quality goods and services and the distribution of leisure and public esteem. In my view, the fixation on measured income inequality derives from a normative gut instinct that potential income and realized income are very similar. The measured stagnation in wages and to a lesser extent in compensation is thus seen as an artifact of political economy rather than a phenomenon that is driven in no small part by changing preferences.

If Gelman is right and “threshold earners” actually compress measured income inequality, as I think he might be, at least one straightforward implication is that some of this “stagnation” is a healthy reaction to the falling cost of an attractive standard of living. 

Gelman writes:


I don’t see how you can get from threshold earners to increased income inequality. It seems to me that threshold earning would decrease inequality, in both the examples that Salam gives. First, you have a well-educated middle class person who could be a high earner but instead gets a more moderate income (perhaps by writing for a political magazine, for example). You’re taking someone who could, perhaps, with sufficient effort make $200,000 a year but instead makes $50,000. This will tend to compress the income distribution, pulling some of the moderate high tail toward the median. Second, you have some who could possibly become a billionaire who decides to be a mere multi-millionaire instead. Again, this compresses the income distribution.

The only way I see threshold earning as increasing income inequality is someone who could earn $20,000 a year instead slacks off and earns $5,000 a year instead. I’m sure there are such people, but given the cost of living, I’m imagining there are a lot fewer people like that now than in the 1960s and 1970s. And these don’t seem to be the people that Cowen and Salam are talking about. Cowen in particular was focusing on inequality in the top 1 percent of the income distribution.

Slacking off is, in my view, the wrong way to think about this phenomenon. When I wrote about the “threshold earners” in my life, I was actually not referring to people who primarily make a living by writing for political magazines — I don’t have any close friends who primarily make their living by writing for political magazines, though I’m fond of many people who do! I do have one close friend who is a newspaper columnist, but he’s not one of the friends I had in mind. Rather, I was thinking about:

* A friend who was recently offered a job by an elite law firm after working for several years for a small urban non-profit, and instead chose to work for another small urban non-profit devoted to building affordable housing. This friends has built and restored houses while holding down an extremely intellectually and at times physically demanding job since graduating from the most selective law school in the United States. I don’t think of him as a “slacker.” I think of him as someone for whom psychic income weighs more heavily than market income.  This friend is unmarried, and so he has more freedom to indulge this preference. It is also true, however, that his partners have generally been women with a similar set of preferences regarding psychic vs. market income. (See Buddenbrooks.)

* Another friend with similarly impressive credentials who leads an itinerant lifestyle, funded by a combination of non-fiction writing, documentary work, and various odd jobs, including a brief spell delivering packages in war zones. At present, he has no fixed address, though he spends the bulk of his time, if I understand correctly, in Peshawar and Bavaria and Paraguay. Given the extent to which he exerts himself in his core pursuits, most of them intellectual but some involving feats of physical endurance, it’s not clear to me that “slacker” is the right word for him. This friend has a partner, a woman who does scholarly work in central Africa. 

* Then there is my friend who recently graduated from an elite school of management — he received a fellowship, and thus made money off of the transaction if we leave opportunity cost to the side — yet has decided to return to acting and writing plays, which had been his primary focus after graduating from a selective undergraduate program, forgoing a number of lucrative job opportunities.

* An academic friend, tenured at an early age, with a skill set that would command a very high premium in, for example, the financial sector, yet who prefers to focus on research. Though he works very long hours, his work is, for him, indistinguishable from leisure. If we define “slacking” as the avoidance of unpleasant work, then he could indeed be a “slacker.” But I fear that stretches the boundaries of the conventional meaning of the word too far. This friend is married. He is in a somewhat different category, as he is definitely part of a high-income household. Yet the relevant consideration is that he earns more than he feels he needs to, and so he feels no sense of urgency with regard to increasing his cash income. 

* I was also thinking of my friends who work in the visual arts. Some have “won the tournament,” i.e., they’ve managed to deploy considerable entrepreneurial savvy to make a comfortable living, yet the psychic income is the thing. Naturally, not all of them have achieved the same success, at least not yet. Psychic income is also salient for my friends who work in K-12 education. 

* The reason why I referred to “the pattern of wage dispersion” among my friends is that I also have friends and acquaintances in the same age cohort with similar credentials and skills who are comfortably in the top 0.5 percent of the income distribution. This is not to say that all of my “threshold earning” friends could do the same, but at least some of them have had offers on the table that could get them there or very close. 

Suffice it to say, these individuals aren’t necessarily representative of the “threshold earning” phenomenon. But this hopefully offers some insight into my thinking. I assume that Andrew Gelman was using “slacker” in a playful sense to clarify the issues at stake, so there’s no harm here. Yet I think normative diversity is a really important phenomenon, and I want to give the post-materialist frame of mind the respect I think it deserves. 

Back to Gelman’s core contention:


The argument that Cowen seems to be making goes like this: There is some cohort of high-powered people in the workforce who, if they all worked the same amount of effort, would have some spread in their incomes. But then some of these people work 100-hour weeks (and make more money) while others work less hard (the “threshold earners”), and the result is increased variation in incomes.

In the context of the general population, though, I see this threshold earning as a source of decreased inequality (as discussed above) in that it brings some potential incomes high in the tail down toward the median.

As I noted above, this could certainly be true! My indifference to this possibility was the implicit point of my conclusion. But before I return to this theme, let’s work our way through a few other possibilities:

* Geographical arbitrage: Are people with higher potential incomes more mobile than those with lower potential incomes? This is a tricky question. Maximizing potential cash income might involve living in or near certain agglomerations, thus restricting mobility. But overall, I’d say that people with higher potential incomes are somewhat more mobile for a variety of reasons, e.g., having more geographically and culturally extensive networks is one possible avenue for increasing potential income.

So what impact does this have on measured income inequality? Consider my favorite example of two married couples. One couple lives in a high-cost region, and both partners engage in market production. Another lives in a low-cost region, and one chooses less lucrative yet more psychically rewarding work in the market while the other focuses on household production. Would the universe in which both settle in the high-cost region and both couples are two-earner imply more measured income inequality or less? 

* Labor market competition. I reject the lump of labor fallacy. But what if we reversed this (hypothesized) normative shift and we had far more people competing for high-wage employment in high-cost regions? Much depends on which slice of the distribution we’re talking about, and whether tournament theory or the “superstar effect” dominates or if increasing the supply of certain kinds of skilled profession lowers the returns to said skills. This is one of the arguments often deployed — by me, among many, many others — for relaxing various licensing requirements and for increasing the number of skilled immigrants allowed to work and settle in the U.S. I’d want to know the impact on measured income inequality, the impact on real income inequality, i.e., the purchasing power of households under this scenario, and much else. 

These are all tough questions to answer.

One of the phenomena I found most fascinating is sibling correlation of income. Our ability to diagnose and treat mental illness has had interesting, cross-cutting effects. While potential income at the highest levels has been increasing, my sense is that we have different cultural “off-ramps” for people identified as mentally ill than had been the case, in part because the cultural prestige of household production has declined. Cultural channels that had once existed for people who can’t flourish in the market — extended kinship networks, etc. — aren’t present to the same degree. Might this relate to declining labor force participation among men and rising labor force participation among women? And how might this phenomenon relate to measured income inequality? Consider that reported market income does not always correspond to actual earnings, and that eligibility standards for various public assistance programs creates high effective marginal tax rates on reported cash income at the low end of the labor market. 

For me, the phenomenon that really matters is entrenched, multi-generational poverty. I think that high incomes at the top have nothing to do with entrenched, multi-generational poverty. If anything, I think high incomes at the top can be helpful, by generating more resources, particularly intellectual, technological, and civil society resources, for addressing the interrelated problems that give rise to it. 

I’m pleased to see that Andrew Gelman and I are on pretty much on the same page (see comments). One thing we definitely agree on: this has been a pretty darn un-rigorous discussion, as we’re talking about phenomena that are hard if not impossible to measure. 


The Latest