On Inequality: There’s no such thing as a Pareto Improvement

I don’t really want to be writing about this now. I had hoped to spend today theorizing obscurely about liquidity. But, after reading (via Felix Salmon) the apologetics of a bunch of prominent economists for inequality (here, here, and especially here — or here via Mark Thoma, to get around Times Select), I just have to go off half-cocked a bit.

Earth to Economics Professors! Earth to Economics Professors! In this, the real world that we inhabit, There Is No Such Thing As A Pareto Improvement. An increase in inequality (or an increase in equality for that matter) always helps some people and hurts others in a variety of ways. Period.

You can claim, if you like, that the harms people experience by increased inequality are outweighed by other benefits, even to the harmed. But the dimensions along which people are harmed and helped are incommensurable. Conventional economic simplifications like “real income” do not capture the full effect of the changes. And we need not and should not, resort to emotional fairy tales like “spite and envy” (as in a previous blogospheric inequality controversy) or “inequality of happiness” (an unfortunate addition today by the usually excellent Tyler Cowen).

Increases in wealth to the wealthy can harm the less wealthy in many ways that don’t show up in “real income” stats. Here are two important but often overlooked examples:

1) Inaccessibility of public goods for which superior private substitutes can be purchased

Consider almost anything that is arguably a “public good” — parks, well-paved roads, transportation in general, schools, medical care, social insurance, personal security. Private substitutes are available for nearly all of these goods, to those who are sufficiently wealthy. Who needs a park, when one can buy a home with a lovely back yard? Is it worth paying high taxes to keep the roads smooth, when one can purchase an SUV, or a helicopter, that is not bothered by potholes, and is more comfortable and functional than an economy car anyway? Speaking of cars, should we just invest in an excellent streetcar and/or subway system, and not spend so much money on roads? Should we accede to higher taxes to support well appointed public schools, or do we prefer strictly private education? Should I support taxes to fund the police, when in any case I am going to require expensive private security?

Rational individuals will answer these sorts of questions very differently, depending on their circumstances. To those for whom the marginal utility of a dollar is low, buying superior private substitutes for public goods will appear to worthwhile. They will oppose government purchase of these goods and the taxation required to support it. But, poorer people will find the dollar cost of private substitutes to be burdensome, and will rationally choose less expensive state provision via inferior public goods. Choices have to be made. We build a comprehensive subway system, or we don’t; we buy and maintain a new park, or we don’t; we tax to fund more policing, or we don’t. Increases in inequality (particularly increases in the numbers and the political influence of the relatively wealthy) shift the likelihood that we opt for private provision of what could be provided in an inferior manner, but at lower cost-per-person, as a public good. Poorer people are forced to pay more for a service than they would have under the policy they would have chosen, or to do entirely without services that might have been provided them inexpensively by the state.

2) Goods and services become expensive, or fail to be produced at all, under inequality, due to reduced economies of scale and increased resource prices.

The mix of goods and services produced by a society is affected by the distribution of wealth in that society. It is easy to see that luxury goods might fail to be produced in very equal societies. Suppose, counterfactually, that we could have an economy in which goods and services are produced as efficiently as they are currently, but with no inequalities in wealth or income, everyone earns the same, current US-average income. The market for Lamborghinis would dry right up. But the converse holds as well. Suppose that the same society is divided into two groups, one wealthy enough such that the superior performance of a Lamborghini is clearly worth the extra expense, and another group whose wealth is roughly the current US-average. The poorer group will be adversely affected by the sudden good fortune of their neighbors. As the market for economy cars will be cut in half, economies to scale in auto production will be adversely affected. Car companies might be the first to take a hit, as they will have already sunk costs based on now violated scale expectations. But new lines of economy cars will have to be designed more sparely, or else priced more expensively, in order for car companies to build profitable econoboxes. The scale effect is exacerbated if, plausibly, Lamborghini production utilizes more non-human resources (metal, energy, etc.) than econobox production. Goods prices will be bid up by the wealthy half of the world, and price and quality of the econoboxes will thereby suffer as well. There may be dynamic effects that mitigate the ill effects of half the world’s sudden good fortune on the other half, but even in this most contrived case, you can’t claim a simple “Pareto improvement”.

The above is not to suggest that inequality is inherently bad. On the contrary, my intuition is that most societies I’d want to live in would tolerate a great deal of inequality. But “a great deal” does not mean unlimited, and tolerance does not imply unconditional cheerleading. At this moment in the United States (and throughout the world), a lot of people (myself included) see a disturbing degree of inequality whose growth, we believe, is insufficiently attached to the kind of positive effects that sometimes persuades us that inequality is a good thing. We may be mistaken. But today’s crop of justifications, by authors whom I usually admire, struck me as shallow, glib, and unserious.

 
 

7 Responses to “On Inequality: There’s no such thing as a Pareto Improvement”

  1. bsetser writes:

    I agree

  2. Andy writes:

    In terms of public goods, couldn’t it be the case that increased inequality actually enhances their availability? If fewer rich people are hiking in Yellowstone, then that a) reduces wear and tear, and b) lessens crowding. Regular folks may actually have more access. Same thing for public transportation. You can make the argument that it is likely that funding for these things will but cut — but that isn’t necessarily given. Maybe rich people will actually appreciate public goods more, and want to improve them, and poor people can free-ride.

    Also, it’s not likely that the market for cheap cars will be cut in half — more likely that it will be just slightly reduced, probably not enough to cut down on economies of scale. In fact, spending on services probably goes up much more than spending on goods, and services are labor-intensive. What about the Baumol effect? All of a sudden, there’s a market for $200 haircuts, expensive massages, restaurants, and all sorts of other personal services.

  3. Andy writes:

    Just as a quick check, the CPI for food has risen by 12.7 percent over the last five years, whereas the CPI for all other items is up 14.6 percent over the same time — i.e., food is getting relatively cheaper. In general, inflation for services is much higher than for commodities over a period of rising inequality.

  4. Nels Nelson writes:

    I second the comment by bsetser.

    Furthermore,I will add that the high standing among economists of hypotheses which are important in economic history and which have been rejected by the data is one of the reasons many current debates in economics are so insane and reinforces my low level of expectations from the majority of professional economists.

    In my freshmen economics course “The Worldly Philosophers” by Robert Heilbroner was required reading. He updated the book some years ago and therein called on the economics profession to reformulate “the economic problem.” In the updated edition he said: “We live in a period in which much of the conventional wisdom of the past has been tried and found wanting. Economics is in a state of self-scrutiny, dissatisfied with its established premises, not yet ready to formulate new ones. Indeed, perhaps the search for a new vision of economics, a vision that will highlight new elements of reality and suggest new modes of analysis, is the most pressing economic task of our time… Perhaps in a different society of the future, another hypothesis about behavior would have us serve as our starting point. People might be driven by the desire to better the condition of others rather than themselves… A story about heaven and hell is to the point. Hell has been described as a place where people sit at tables laden with sumptuous food, unable to eat because they have three-foot long forks and spoons strapped to their hands. Heaven is described as the very same place. There, people feed one another.”

  5. Aaron Krowne writes:

    I agree with the general sentiment of this post — that inequality is being excessively apologized for — but I do not agree with the specific foundational claim that there there is “no such thing as a Pareto increase”.

    Pareto increase is founded on the injection of new value into society, not specific transfers and allocations of money, which are secondary. A transaction in which both parties benefit — which is the kind of transaction most common in a free market — generally produces such a Pareto increase. I genuinely believe that. I think the proof is essentially that gradual deflation will be produced in a society with no upwards liquidity bias; all the Pareto increases accumulate and you end up with increasing value relative to money.

    Now, all things go out of whack when you mess with credit and money. Money is supposed to track this injection of wealth and value into society, but in our system it often does not. You are quite likely to get more money purely because you already had money, and get it without producing at least that much value, which means you’re getting the money at someone else’s expense (even if the money/credit is created ab initio, the expense on society is precisely that much more inflation or asset-bubbling).

    To me this is a more dominant phenomenon than “classical” redistributive transfers of the extant money around society, of the kind lacking in the Pareto property. Instead we’ve got an out of whack money and credit system, which causes another kind of decoupling of money from value creation. The lack of Pareto-type benefit isn’t the problem here; it’s the lack of tracking such benefit closely with monetary reward.

  6. All — Sorry to be so out of touch. My time is not my own these days, and I’m putting a lot of work into the next post (which I suspect will be of little interest to anyone but me).

    bsetser — (blushing) I’m just flattered that you dropped by to read, let alone to agree. You’re my original econoblogging hero.

    Andy — So, Yellowstone is the kind of public good for which there are not, exactly, superior private substitutes. I’d guess that Yellowstone might not be harmed by inequality. But despite your hopeful conjecture about public transportation, my sense is that in the real world, it just doesn’t work that way. People with cars quickly cease to support public transportation spending (except on roads), and become actively hostile if they find that buses cause them any sort of inconvenience. I’ve done no empirical study, but I’d be willing to wager on the overall direction. In less developed countries with, as often happens, very many poor people and a wealthy class rich even by Davos standards, it is an sad though amusing, and very observable, phenomenon that rural and mountain roads are poorly maintained while wealthy people with vacation homes travel in comfort with SUVs. If the SUVs cold not be purchased, my guess is that the roads (at least to the lovelier places) would be better maintained. I’m willing to hold to my general proposition: the availability of public goods for which superior private substitues can be purchased is nearly always diminished with increasing inequality.

    Nels — Interesting point. I love economics and spend much of my free time reading what economists have to say. But the word hubris should not be lost to the dismal science. As an epistemological matter, it remains an open issue to what degree progress is possible in the social sciences. No theory or prescription of economics is so certain that we should accept its imposition without a lot of cross-checks and smell tests, and not only the sort of tests performed by econometricians. Sometimes even that oldest of charnel houses, the parliament, has something valid to say, despite anybody’s t-tests. Power corrupts academics as easily as men of politics, and no man was ever more corruptible than the one with an ideology or science proclaiming what happens to be his own interest just and right.

    Aaron — So, we agree on many things, but we’re likely to argue on technicalities. I stand by my claim that “there’s no such thing as a Pareto improvement” in any system in which two preconditions hold: 1) relative wealth has nonzero positive utility; and 2) the benefits of absolute or relative wealth are incommensurable. If this is true, then the private exchange you describe — which makes both parties richer — is not a Pareto improvement because it harms the relative wealth of others. Now of course, as Tyler Cowen argued and I wholeheatedly agree, I’ve set up this game badly. I’ve made life into a zero-sum game when it needn’t be. By construction, there is no such thing as a Pareto improvement, thanks to my preconditions. (Technically, there is one possible Pareto improvement — a proportionate increase in the wealth of all players.) I agree with this criticism, and see no point in making life more zero-sum than it has to be. So what was wrong with my preconditions? Precondition 1 is absolutely true: In real life, relative wealth matters, as a matter of real-world expected outcomes as well as the fuzzy psychological stuff. I give no quarter with respect to Precondition 1. Precondition 2, however, is not necessarily true. If you get rich by inventing a better nose-hair clipper, I might lose out in terms of my relative wealth, but my absolute gains in nose-clipping utility (I have hairy-ass nostrils, let me tell you) might in some sense overwhelm that relative demotion. If we let absolute and relative gains be commensurable, and define a conversion to common units (dollars? utils?) such that the cost of the fall in relative stature is less than the gain in easy nose clipping, then your invention (and the subsequent voluntary private transactions whose result in your fortune) might indeed represent a Pareto improvement (at least in a society in which all people have hairy noses). But note how many assumptions and definitions we have to make to find a Pareto improvement: we have to define the cost of relative loss, the gain of the general “increase in wealth”, and some transmission mechanism whereby that increase is general (that is, not only the hairy-nosed benefit).

    In practice, I agree with you that there are many activities that “add real value”, and for which secondary costs (like changes in relative status) should not be much considered, and that societies that support the “production of real wealth” will be better societies than those that do not (and especially better societies that place too high a weighting on relative wealth compared to some measure of “absolute wealth”). Therefore, in practice I both agree with you, and support a the generation of a great deal of economic inequality, so long as the unequal wealth is somehow tethered to “real value creation”.

    But that doesn’t mean that economics professors who pull fairy tales out of game theory textbooks and try to claim inequality is nothing to worry about because it’s all Pareto improvement should be treated as though they are authorities on the matter. Perhaps it is an overstatement to say there is no such thing as a Pareto improvement. But I can confidently claim that no occurrence in the real world can be identified as a Pareto improvement without a lot of judgement calls between the event and the identification.

  7. Greg writes:

    Good Discussion. I find it very frustrating to read the superficial analyses that attribute inequality to ‘natural’ economic forces such as globalization and technology while ignoring the very unnatural (or perhaps it is human nature?) force of inflation that in my mind is the primary driver of the outrageous growth in wealth and income of the global elite. The Economist has been among the larger contributors to this theory, most recently only a couple of weeks ago, where they basically argued that CEO income was justified by stock price increases and policymakers should be careful not to intervene in the ‘free’ market by designing policies to redistribute incomes more fairly. I wrote to the editor in response, copied below. Think they’ll print it?

    Sir,

    In your leader on the increasing inequality in recent decades you argue forcefully against redistribution by taxation but completely ignore the much more insidious and pernicious redistribution by inflation. Inflation represents an alternative to the consensus explanations of technology and globalization for the rapid rise in incomes of the top fraction of the population. Since the early 1980s the financial sector has grown much more rapidly than the rest of the economy in sympathy with accelerated growth in credit and the supply of money. The managers and financiers whose pay packets have soared in recent years are the gatekeepers of newly issued money and have been diverting an increasing proportion of it to themselves at the expense of the rest of the world’s savers and labourers, whose incomes are devalued. Wealth accrued by such means is no more a reflection of free market forces and returns to value added than is wealth accrued through taxation.