Why avoid panel data in examining social mobility?

Last week at Politics and Prose, I had the opportunity to hear Robert Putnam’s book talk for Our Kids: The American Dream in Crisis. In his book, he focuses on data that purport to show a growing divide in income inequality and social unraveling.

Putnam told a personal anecdote about his deteriorating home town in the Rust Belt. If Putnam were able to prove that the shrinking economy from the loss of manufacturing jobs in the Rust Belt proportionately reflected the larger economy in all other sectors, he might have a strong data point, but a personal anecdote is not enough. The availability heuristic is not strong evidence.

So much of the talk concerning income inequality pertains to an unstated premise about social mobility. The widespread fear is not just that the rich are getting richer, but that the rich are getting richer at the expense of the poor. The mental model assumes some fixed share of wealth that exists in the world should be divvied up fairly so as to avoid predation by the strong on the weak. Often, the evidence presented for a “fixed pie” theory is to show the shrinking share of income among the lower quintiles and the growing share of income among the higher quintiles. The problem with this methodology is that it doesn’t actually account for social mobility. To prove that capital is flowing from individuals in the bottom quintile to individuals in the top quintile, we need panel data.

If we don’t analyze with panel data, we might observe the top quintile, profiting from some entirely new high-tech sector, drastically increasing their income by 20% while lower quintiles still increase at 2%. More wealth generated at the top wouldn’t imply material loss for the bottom quintile.

In the Q&A, I pressed Professor Putnam on his methodology, specifically, to what extent he used panel data to show decreased social mobility. After his book signing, he elaborated for me.

Putnam claimed that there was some good panel data for income, but that it couldn’t be used to show current trends in social mobility.

We might suppose that the relevant panel data to measure social mobility would include income, y_{it} for i=Poor, i=Rich, t=20, t=40.

At t=40 or so, we might expect people to be generating the most amount of income for their life.

The methodological issue Putnam pointed out was that individual incomes over a lifetime are highly nonlinear. If you were to track a random sample of individuals starting at t=20, very different kinds of individuals would look very similar, but both would appear in the bottom quintile. Specifically, y_{Rich,20} could be -$100,000 for a Harvard pre-law student who’s taking out student loans, and y_{Poor,20} could be $16,000 for a minimum wage job. However, y_{Rich,40} might be $450,000, while y_{Poor,40} might be something like $25,000.

Putnam pointed out that because of how this panel data is measured, the data is always intrinsically 30-40 years out of date. Wait, is this a cop out? Is this methodological laziness?

Just because any one particular study requires 40 years doesn’t mean that we couldn’t observe multiple concurrent staggered studies, with different individuals to show panel data over time. We can imagine Study A starting in 1945 with a batch of individuals at t=0, Study B that tracks t=0 at 1950, Study C that tracks t=0 at 1955, and so on and so forth. Then, despite nonlinearity in lifetime earnings, we would still be able to see trends in how individuals are or aren’t moving up, out of their birth quintiles.

Did F. A. Hayek understand optionality?

Nassim Nicholas Taleb writes in Antifragile,

We may be drawn to think that Friedrich Hayek would be in that antifragile, antirationalist category. He is the twentieth-century philosopher and economist who opposed social planning on the grounds that the pricing system reveals through transactions the knowledge embedded in society, knowledge not accessible to a social planner. But Hayek missed the notion of optionality as a substitute for the social planner. In a way, he believed in intelligence, but as a distributed or collective intelligence— not in optionality as a replacement for intelligence.1

Au contraire, Hayek wrote in The Constitution of Liberty, in Chapter 2: “The Creative Powers of a Free Civilization,”

It also follows that the importance of our being free to do a particular thing has nothing to do with the question of whether we or the majority are ever likely to make use of that particular possibility. To grant no more freedom than all can exercise would be to misconceive its function completely. The freedom that will be used by only one man in a million may be more important to society and more beneficial to the majority than any freedom that we all use.2

In a footnote for the paragraph, Hayek cites a comprehensive scholarly history for the idea:

Cf. Rev. Hastings Rashdall, “The Philosophical Theory of Property,” in Property; Its Duties and Rights: Historically, Philosophically, and Religiously Regarded, Charles Gore and Leonard Trelawney Hobhouse, eds. (new ed.; New York: Macmillan, 1915) pp.61-62: “The plea for liberty is not sufficiently met by insisting, as has been so eloquently and humorously done by Mr. Lowes Dickinson (Justice and Liberty: A Political Dialogue, e.g. pp. 129 and 131), upon the absurdity of supposing that the propertyless labourer under the ordinary capitalistic regime enjoys any liberty of which Socialism would deprive him. For it may be of extreme importance that some should enjoy liberty—that it should be possible for some few men to be able to dispose of their time in their own way—although such liberty may be neither possible nor desirable for the great majority. That culture requires a considerable differentiation in social conditions is also a principle of unquestionable importance.” [The full citation of the book quoted by Rashdall is: Goldsworthy Lowes Dickinson, Justice and Liberty: A Political Dialogue (London: J. M. Dent, 1908).—Ed.] See also Bennett E. Kline and Norman H. Martin, “Freedom, Authority, and Decentralization,” p. 69: “If there is to be freedom for the few who will take advantage of it, freedom must be offered to the many. If any lesson is clear form history, it is this.”3

1 Nassim Nicholas Taleb. Antifragile (New York: Random House, 2012), Kindle edition.

2 F. A. Hayek, The Constitution of Liberty: The Definitive Edition (Chicago: The University of Chicago Press, 2011), 83.

3 Ibid.

Racist ignorance of prior probabilities, explained with Venn diagrams

Last year, in reference to the Trayvon Martin case, Matt Yglesias wrote a great short explanation of how racism is a kind of ignorance of Bayesian updating. In short, racists justify their racism by neglecting base rates. It’s a fallacious form of the availability heuristic.

Bayesian reasoning can be difficult to understand when it’s presented formally with equations and formulas, so let’s illustrate the concept with Venn diagrams.

Consider two populations, white and black. The white population is larger than the black population.


Assume that the amount of crime that occurs in the two populations is roughly proportional to the size of each, represented by the red circle below.


Assume that police attention, due to institutional racism, is disproportionately focused toward black criminals, as represented by the blue ellipse below.


The population of convicted criminals would be represented by the purple shading below.


A racist observing racial discrepancies between the inmate population and the general population is myopically only seeing the purple shading. The racist sees that blacks make up a disproportionately large fraction of the purple shading while falsely assuming the inmate population is an unbiased sample. The racist neglects what the true parameter is.

Disproportionately punishing criminals in minority communities would be horrible in and of itself, but the War on Drugs has subverted the criminal justice system in an even worse way, and exacerbated institutional racism. How?

Conducting the War on Drugs requires the violation of civil liberties. Why? Whereas victims of crimes cooperate with police and offer evidence to bring criminals to justice, victimless crimes produce no such cooperative victims. Without victims pointing toward any kind of offender, the primary method to catch violators of victimless crimes is to preemptively assume some fraction of a population is criminal and use sweeping powers to arbitrarily detain and search.

Without any victims, from where would probable cause originate? Terry v. Ohio paved the path for Arizona v. Johnson, and now the police act on “reasonable suspicion,” which in practice has turned into arbitrary officer discretion, far beyond the original scope of the standard to ensure officer safety. “Reasonable suspicion” is a lesser degree of certainty than probable cause, and as such, was always obviously unconstitutional.

If a police department were already predisposed to target a black community, instructing them go after victimless crimes would intensify their biased policing, giving them cover to target whomever they already were going to target.

Not only does the War on Drugs erode the potency of the Constitution, it erodes the trust between the public and law enforcement. Whereas the public might, in ideal theory, primarily rely on law enforcement for protection from criminals, the War on Drugs has subverted the relationship, and given the public a reason to fear the police. The War on Drugs distracts the police with incentives to maximize drug arrests, drawing their focus drawn away from putting away the harmful elements of society.

The War on Drugs produces a more disturbing Venn diagram.

Assume a majority white population and a minority black population.


Assume that victimless crimes in the two populations are occurring proportionally to their populations, because the drive to alter consciousness is a human universal.


Assume that a smaller amount of crimes with victims are occurring in the two populations proportionally.


The encouragement of police attention to victimless crimes gives the police cover to disproportionately target the black community.


Racists incorrectly infer biased police attention as a proxy for societal harm, failing to distinguish between malum in se and malum prohibitum. The purple shading below represents an entire group of people who are being oppressed by a criminal justice system that is consistently and repeatedly violating Mill’s harm principle.


How does one person’s wealth affect another’s?

Matt Bruenig attempts to disprove a conservative talking point about income inequality, that one person’s wealth doesn’t have an effect on another person’s:

First, even if you wrongly think of wealth as a store of money and property created long ago, the distribution of it still impacts people’s lives, especially in America.

I’m already confused. Wealth, as far as I understand it, is the net value of the entity being examined. Net income is a measure of wealth over time. Is there some kind of obfuscation going on here? It seems that there’s imprecise jargon here. “Wealth” isn’t a store of “money,” but “money” is a store of value.

Modern life is fraught with very expensive risks lurking around every corner. A sudden illness or accident could render you disabled and unable to work. A recession or economic restructuring could render you unemployed and render the skills you’ve spent your life learning useless. Reaching old age with inadequate savings could mean living your golden years in poverty.

By what measure is modern life riskier than than the pre-modern era? When exactly was this golden age of low-risk living? This is not just a pedantic semantic objection. By modern standards, life in the pre-modern era was characterized by utter poverty. Until about 1800,  no society had experienced sustained growth in per capita income. So, while it may be true that there was less risk in the pre-modern era, because pre-modern people didn’t have as much wealth to lose, I don’t think Bruenig is trying to celebrate the massive gains in per capita wealth after the Industrial Revolution.

Many societies have created robust social insurance systems to protect their populations from these kinds of risks. The U.S. has done so as well, but to a much lesser extent. Because social insurance in the U.S. is so inadequate, it is incumbent upon people to self-insure against these risks. That means they need to have enough wealth to draw upon as a cushion if they end up facing hard times. But here’s where the social contract fails: When the bottom half of the country owns basically none of the country’s wealth, they can’t self-insure against these risks. Instead, they must lead a relatively perilous life in which one misstep or mistake could wreck them and their families.

I think this is confusing because there’s an unstated assumption here, that social insurance is just a state program that redistributes money from the wealthy to the poor. This isn’t Bruenig’s fault, but “social insurance” has become a kind of Newspeak term for state redistribution from the wealthy to the poor. Before the rise of the welfare state, there was a robust history of mutual aid societies. That kind of social insurance wasn’t deployed through governments, and wasn’t relevant to the tax code. In fact, that kind of social insurance often functioned to counteract government institutionalized racism.

There’s an unstated premise that redistribution from the wealthy to the poor would lessen the poor’s vulnerability to financial risks. This is true but irrelevant. Bruenig still hasn’t laid out what the exact negative externality is.

Second, wealth is not just a pile of dead value created in years past. When utilized properly, wealth ensures its owners a share of future income, too.

What is this term “dead value?” What does that even mean? Does it refer to a kind of Marxist labor theory of value, or a marginalist subjective theory of value? What is this caveat, “When utilized properly?” Show your work!

It seems that there’s another unstated assumption here, about capital accumulation. The underlying theory is that once an entity has collected a certain amount of capital, the capital will start generating returns without much attention, and that owners of capital can passively collect interest.

It seems obvious to me that there are significant costs to managing capital, but maybe it’s not so obvious. Most startups fail. Venture capitalists and investment banks are constantly assessing and lending capital from household savers to cash-poor entrepreneurs. The total economic surplus generated by the few successful companies outweighs the deadweight loss from overproduction of failed businesses. Investing isn’t simple; it’s not inevitable that once an entity has a certain amount of capital, there will be easy and steady returns.

 Most people seem to equate income with working: You go to your job, do your tasks, and get a paycheck. But this is only half of the story. At a macro-economic level, a nation’s income is divided between owners and workers, with the part flowing to the owners called “capital’s share” and the part flowing to the workers called “labor’s share.” In recent years, capital’s share of the national income has been as high as 37 percent, which is to say 37 cents of every dollar of income in a year goes to passive owners of wealth.

I had hoped when I started reading Bruenig’s article that I might find something other than rehashed social conflict theory. If Bruenig wants to abandon the neoclassical paradigm, he should say so explicitly. If not, then it’s not useful to talk of conflict between capital and labor. Both are mere inputs.

I think I can make Bruenig’s argument stronger than he formulated it here. Bruenig never explained what the negative externality was, but I can hazard a guess. The negative externality exists from whatever institutional arrangements preserve the inequality. That may sound obvious, but I’ll unpack it.

What Bruenig should have argued is not that the wealthy are only wealthy by withholding capital from the poor, but that the wealthy have historically used their wealth in the political process to calcify their social position. Why is a donation to the opera tax-deductible? Why are there home ownership deductions? If there is political power available to be purchased, it will be purchased.

Invasive monitoring for discounted health insurance policies

What is often called “health insurance” in the United States often isn’t actually health insurance, but a kind of imperfect prepayment plan for medical services.

If “insurance” companies were ever again to become actual insurance companies, seeking profit by assessing and pricing risks of payouts, how much producer and consumer surplus might be available through invasive health monitoring? If insurance companies could more comprehensively and invasively monitor their customers’ risk factors by, for instance, requiring monthly blood tests, or requiring shared access to a 23andMe profile, how much economic surplus might be available?

Surely there’s potential producer surplus, because insurance companies would be able to keep more money if they knew certain kinds of healthy customers would require fewer expenditures. Surely there’s potential consumer surplus, because healthy customers would be rewarded with lower prices for their good health. Pricing could even be dynamic, depending on the particular monitoring technology.

Aside from gains in producer and consumer surplus, there would be an even greater benefit. Prices would serve as a kind of check on biased medical research. Medical academics politicking for research money might continue to make wild and untrue claims about different pathologies, but insurance companies would have skin in the game to evaluate medical research.

As far as I know, privacy regulations and price regulations make this idea completely impossible today.

Why do women wear makeup?

Why do women wear makeup? SSSM advocates incorrectly consider the use of cosmetics an arbitrary social construct.

The practice of applying cosmetics is so ubiquitous that its strangeness is underappreciated. What other animals deliberately modify their appearance to enhance their attractiveness? What other animals would actually find such modifications attractive?

In his paper, “Why Cosmetics Work,” Richard Russell proposes that there are evolutionarily determined universal factors of facial attractiveness, and the different practices of applying cosmetics, though quite varied across cultures, all attempt to enhance such factors of facial attractiveness. The practice has been convergent across cultures.

The paper explains that cosmetics are primarily used to exaggerate sex differences. Males actually have perceptibly redder skin than females, likely from higher levels of hemoglobin. Because of this, females have a higher luminance contrast surrounding the eyes and lips. Dark eye makeup and lipstick enhance this luminance contrast. Recall how Snow White’s lips were as red as blood, and her was skin was as white as snow? Her feminine beauty was signaled by an extreme facial luminance contrast.

The paper also mentions that fairer female skin might have come about from natural selection for increased cholecalciferol and calcium production for pregnancy and lactation.

Further questions:

  1. Why wouldn’t men use makeup to decrease facial luminance contrast? Is this a historical accident?
  2. The paper only used Caucasian and East Asian faces. Would subjects from ethnic groups with darker skin corroborate this research?

How do conflict-generated diasporas affect homeland conflicts?

Edward Said claimed,1 and Noam Chomsky agreed, that the acceptable range of political attitudes about the Israeli-Palestinian conflict are much more diverse among Israeli Jews than American Jews. I’ve been searching for some hard evidence that American Jews are more hawkish toward Palestinians than Israeli Jews, but haven’t found any.

The claim is plausible, though. Conflict-generated diaspora groups might feel a stronger urge to demonstrate in-group loyalty to compensate for their absence. Terrence Lyons explains in his 2004 paper,

One dynamic that tends to make conflicts in the homeland more protracted, therefore, is the existence of certain types of diaspora groups with strong symbolic attachments to a territory and uncompromising views on how conflict there should be understood and contested.

The hawkishness from without actually seems to aggravate the homeland conflict. Paul Collier and Anke Hoeffler hypothesize in their 1999 paper,

A further potentially important source of start-up finance for rebellion is a diaspora living in OECD countries. Such diasporas are usually much richer than the population in their country of origin. They are better-placed for collective action: emigrants have a cultural incentive to create diaspora organizations which can then discipline free-riding. They do not suffer the consequences of the conflicts they finance. As with grievance among the local population, in the greed-model grievance among the diaspora is assumed to be manufactured by the rebel organization rather than being an original cause of conflict. Hence, the diaspora increases the risks of conflict renewal, but not the initial risk of conflict. We measure the size of diasporas in the USA relative to the population in their country of origin.

They find,

By far the strongest effect of war on the risk of subsequent war works through diasporas. After five years of post-conflict peace, the risk of renewed conflict is around six times higher in the societies with the largest diasporas in America than in those without American diasporas. Presumably this effect works through the financial contributions of diasporas to rebel organisations.

Collier and Hoeffler also controlled for how large diasporas might be because of the size of conflicts themselves.

The case of the Israeli-Palestinian conflict in particular isn’t so straightforward. Jonathan Rynhold, writing in 2008, explains, “Prior to the 1980s Diaspora identification with Israel was expressed in unwavering support for Israeli policies. Since then Diaspora support for Israeli policies cannot be taken for granted.” Rynhold holds that the 1982 Lebanon War dissolved the unanimity of opinion in the diaspora, due to the media coverage of Sabra and Chatilla. The 1980s brought the Jonathan Pollard affair, as well as the First Intifada, which impelled liberals to contemplate the policies in the Palestinian territories.

It would be useful to collect opinion data from emigrant communities connected to current ongoing conflicts, and compare them to opinion data in the homelands. Where can I get some good, current research on this?

1 Edward Said, and Christopher Hitchens, Blaming the Victims, (London: Verso, 1988), 9-10.