The Agenda

Disparities and Public Sector Reform

Thinking about euvoluntary exchange has led me to think about how we think about income and wealth disparities in American life. 

Conservatives, myself included, often point to the rise in the number of single-parent families as an important issue. The deterioration of marriage among the less affluent does seem to have contributed to increasing poverty, etc. Yet it is also not susceptible to the same kinds of intervention as disparities in pretax money income. And so the at least some on the left see an emphasis on family structure as a dodge — an effort to “blame the victim,” to distract from the more urgent question of how much we should increase cash transfers.

Recently, I was perusing U.S. poverty statistics, and I was struck, though not exactly surprised, by the relatively low level of poverty among non-Hispanic whites, and in the incredibly low level of poverty among married-couple non-Hispanic white households.  

The moral case for the urgency of redistribution in the U.S. context rests heavily on the historical legacy of enslavement and segregation. In my next column for The Daily, I briefly reference the obvious point that intergenerational economic progress is cumulative. Having a great-grandfather who was a prudent saver had benefits for your grandfather, and so on down the chain. In contrast, having a great-grandfather who was systematically excluded from participation in the market will have consequences down the chain as well, e.g., you’re probably going to have less inherited wealth and social and cultural capital. 

A parallel issue is that the United States has fared unusually well over the last two hundred years, and people who’ve arrived in this country with modest levels of schooling from societies that haven’t been quite so fortunate are going to suffer the lingering impacts of all kinds of collective trauma, e.g., family fortunes wiped out due to looting and other forms of dislocation, countless generations of illiteracy and poverty, etc. 

One important question is whether the particular ways in which we redistribute in the United States are well-tailored to meeting the nature and scale of the actual disparities. Are more generous cash transfers the best way to deal with the legacy of generations of entrenched rural poverty in central Mexico or Haiti or Guatemala or Bangladesh? Or are other interventions more likely to address the underlying problem?

With regards to the legacy of the historical experience of African Americans, it seems plausible to suggest that mass incarceration is actually a more pressing problem than building in a default towards more generous cash transfers or preserving high levels of public employment. But consider how much time and energy mainstream news outlets — or for that matter ideological media, on the left and right — devote to issues surrounding the taxation of high earners as compared to the social and economic impact of mass incarceration.

Last month, Gregory Acs of the Pew Trusts published a fascinating report on downward mobility in the United States. It is eye-opening. Early on, the report notes that “marital status, education, test scores and drug use have a strong influence on whether a middle-class child loses economic ground as an adult.” Basically, you’re much more likely to fall out of the middle class if you’re divorced or widowed, you don’t have any education beyond high school, and you’ve at some point been a user of heroin or crack cocaine. One finding I found particularly depressing was the following:

Black men raised in middle-class families are 17 percentage points more likely to be downwardly mobile than are white men raised in the middle. Taking into account a range of personal and background characteristics—such as father’s occupational status, individual educational attainment and marital status—reduces this gap, but still leaves a sizable portion unexplained. However, taking into account differences in AFQT scores between middle-class white and black men reduces the gap until it is statistically indistinguishable from zero. [Emphasis added] 

The AFQT test, Acs explains, “measures reading comprehension, math knowledge, arithmetic reasoning and word knowledge,” all of which are heavily influenced one’s home environment. Acs elaborates on this theme later in the paper:

Studies have shown that AFQT scores correlate well with performance in the military, as well as with adult wages. It is important to note that beyond measuring a person’s human capital in the form of knowledge or cognitive skills, AFQT scores likely reflect a host of other factors that affect test performance not included here, such as motivation and self-confidence, which could also influence downward mobility.  Furthermore, AFQT scores are likely to be influenced by a range of outside factors, such as school quality, which are not accounted for in the models here because measures are unavailable.

If we care about persistent disparities between between at the 10th percentile and people at the 50th percentile, a few policy initiatives come to mind, many of which would have to be implemented at the state and local level:

(1) a public health program devoted to minimizing environmental hazards and social practices that pose a risk to infant brain health;

(2) home visitation programs focused on the parents who are considered most likely to abuse their infant children (e.g., parents who were victims of child abuse themselves);

(3) a large-scale effort to improve school quality, which will require giving administrators the flexibility they need to deploy workers as they see fit to the meet the needs of a diverse student population and to draw on a wide variety of specialized service providers;

(4) a grand shift in criminal justice from an emphasis on punishment to an emphasis on deterrence.

There are many other things we might contemplate as part of our campaign. But part of what I find interesting is that while (1) and (2) might cost money, it’s not unimaginable that they would cost less money that what we might save by embracing (3) and (4). This would require a shift in the personnel employed by the public sector, particularly at the state and local level. It’s worth noting that (3) has been treated as a kind of Holy Grail, and the conventional understanding is that it will cost a great deal of money to improve school quality. There is reason to believe that the rise of personalized education has the potential to “bend the cost curve down,” provided incumbent educational providers are disempowered. 

My sense is that many on the left would endorse (1), (2), (4), and a different version of (3). The difference is that the version of (3) they’d (likely) champion is, in my view, extremely counterproductive, and they’d presumably go for a heavier footprint version of (1) and (2). Perhaps more importantly, they’d be somewhat more likely to want to layer (1)-(4) on top of existing public institutions, work rules, etc., whereas the approach outlined here is premised on maintaining a high degree of flexibility in how public workers are deployed, how they’re compensated, etc. 

So maintaining rigidity in terms of how the public sector is organized might be described as the dependent variable. A rigid public sector can take on new tasks and perform them at very high cost, thus necessitating a permanent ratchet towards higher taxes. This ratchet will shape patterns of attachment to the labor market, e.g., it might incline more people to shift from maximizing pretax money income to maximizing psychic income, which might in turn mean less outsourcing of household production and thus less demand for less-skilled labor, particularly less-skilled non-native labor. 

A flexible public sector, in contrast, might take on new tasks, abandon old tasks it doesn’t do terribly well, pay the best employees really well, shed ineffective employees, etc. This model of public sector institution — which might indeed be a pipe dream, by the way — would be a learning organization: one that could go out of business and that must adapt to survive. In theory, this kind of public sector institution (or firm) could deliver a higher quality of service at lower cost, in part because it shifts away from a Baumol’s dynamic and towards heavier reliance on technology.

All of this is why collective bargaining reform might actually be the hinge issue when it comes to the future of U.S. political economy. Remember Doug Schoen’s survey on public sector employment, which he discussed in the Wall Street Journal some months ago? (It was funded by Economics 21′s partner organization, the Manhattan Institute.)

Yet the reason for this apparent movement against collective bargaining reform is that unlike reducing state spending and benefits, voters nationally and in those two states are not convinced that clear savings will result from reforming the labor relations process. By 56% to 33%, voters nationally say “it is unclear how much money will actually be saved by limiting” collective bargaining rights. 

Suffice it to say, one of the key issues shaped by collective bargaining is staffing. As Alon Levy wrote a while back:


Repeating my comment on the Austin Contrarian, and similar comments I’ve made on Second Avenue Sagas: the problem is more staffing than salaries. At New York City Transit, salaries are the same as at Toei – a little more than $100,000 in total compensation per employee. (No data for Tokyo Metro, alas.) The difference is that NYCT has 47,000 employees and Toei 6,400, a factor-of-7 difference, even though NYCT only carries twice as many passengers, and provides only four times as many train revenue-hours and stations. A train driver on Toei spends 700 hours a year driving a revenue train, versus about 450 in New York. And Toei isn’t even the most efficient agency: Tokyo Metro is three times as big as Toei but has only 8,400 employees.

Republican outfits advocating pay cuts likely do not know anything about staffing levels. I’ve never seen the Manhattan Institute, which is arguing for union-busting and pay cuts in New York, say a single thing about staffing levels. 

Actually, Alon was wrong. Nicole Gelinas of MI has discussed staffing levels, and several other MI scholars have done so as well. But Alon’s point is a good one: staffing matters. What is confusing is that Alon’s comment seems to suggest that public sector unions care only about compensation levels, not about the number of public employees. It is safe to assume that most public sector unions do indeed care about maintaining high staffing levels, if only because every additional unionized public employee represents an increase in resources and political heft.