Lane Kenworthy’s New Book Takes on a Controversial Question: ‘Is Inequality the Problem?’

For nearly two decades, scholars have argued that reducing income inequality would have beneficial effects on living standards, health, economic opportunity, the functioning of democracy, and even subjective measures of happiness. Yet during that same period, inequality scholar and Stone Center Affiliated Scholar Lane Kenworthy grew increasingly skeptical of that conclusion. In his latest book, Is Inequality the Problem?, Kenworthy offers answers that are neither simple or expected — while at the same time making the case that ethical arguments in favor of reducing inequality are stronger than ever.

Kenworthy, who is a professor of sociology and the Yankelovich Chair in Social Thought at the University of California, San Diego, spoke to the Stone Center about his book earlier this year. This interview is edited for length and clarity.

What would you say is the key finding of your book?

Kenworthy: The question I address is whether economic inequality has bad consequences for living standards, democracy, economic opportunity, health, and happiness. I look at the experience across the rich, longstanding-democratic countries since the late 1970s — a period in which income inequality has grown in almost all of those countries, but at different rates — to see if nations with either higher levels or faster increases in income inequality have tended to see worse outcomes. I find very little evidence that that’s the case.

I then go on to say there are still very compelling ethical reasons why we might want less economic inequality, especially than what currently exists in the United States. So I’m in favor of reducing inequality, but I don’t think it should be at the top of policymakers’ priority list.

What led you to investigate this question of whether inequality is, in fact, the problem?

Kenworthy: It was the growing number of very thoughtful, very smart people who had come to believe that income inequality has these sorts of harmful consequences and therefore that it ought to rise either to, or very close to, the top of policy priorities. This wasn’t the case prior to around 2005 or 2010. That’s when we began to get more research looking into this question of income inequality’s possible harmful effects. Some of that research concluded that yes, inequality is bad for health or democracy or economic growth or happiness. And so more policymakers, more op-ed writers, and also scholars embraced this conclusion — which I should say I leaned toward embracing, too, in those early years.

I was particularly motivated by a book called The Spirit Level, which came out in 2009. It was published by two British epidemiologists, Kate Pickett and Richard Wilkinson. It did a lot of cross-country comparisons and very strongly endorsed the conclusion that income inequality has harmful effects on all sorts of things that we might care about. They focused on health, but they concluded that there were lots of other bad effects. When that book came out, I read it and mostly nodded my head in agreement. But then I decided to poke around in the data. Once I began doing that, I grew much more skeptical.

I then thought about writing a book, more or less the kind of book that I’ve now just written. Around 2010, I discovered that Christopher Jencks, a very careful empirical social scientist, was thinking along similar lines. He, too, was interested in this question of inequality’s impact. We thought for a short time about combining our efforts and writing a book together. And then we both concluded after a couple of years that it was not a good moment to do it, because research was starting to come out in droves on this question. We thought, “Well, whatever we conclude might turn out to be wrong, or we ourselves might disagree with it within just a couple of years. So we ought to wait and see what this outpouring of research concludes.” And then I decided, I guess about two years ago, to come back to this. It seemed like a good moment to do it.

When you came back to this, did you have any expectations about the conclusions you might reach? Did your results surprise you?

Kenworthy: My rough conclusion as of the mid-2010s was that income inequality probably hasn’t, in the rich democratic countries, had these harmful effects — or, at least, they’re not nearly as strong as a lot of people at the time were suggesting. And that’s essentially what I find in the current book. In that sense, what I find isn’t surprising to me.

But I was open to the possibility that with more years, and to some degree better measures, the data might suggest something different than I previously thought was the case.

What do you see as the potential policy implications of the book?

Kenworthy: We should be very cautious before telling policymakers that if we want to improve health — let’s say that we want to improve life expectancy or reduce infant mortality here in the United States — we can achieve that by reducing income inequality. Or that if we want to improve our democracy, or if we want to help equalize opportunity or increase happiness, an effective way to do that would be through reducing income inequality. I don’t think the evidence supports that.

I don’t want policymakers to think, “Well, there’s no point in reducing income inequality at all. We don’t need to do it because it’s not going to have these good effects.” Again, I think the moral argument is just as compelling as it’s always been. In fact, it’s arguably more compelling nowadays, because income inequality has gone up, especially in the United States but also in many of these other countries, too.

My policy message is to still care about getting to or maintaining moderate or low levels of income inequality, but not to think that that’s going to get us all kinds of other good downstream effects. If you want to improve life expectancy, try to fix the healthcare system and people’s health behaviors. Reducing economic inequality shouldn’t be our top priority. There are other things that are more important, in my view.

Do you have any concerns that some readers might miss this nuance — in particular, that policymakers, politicians, journalists, or even inequality scholars might come away with the idea that the book argues that reducing inequality shouldn’t be a goal at all, and isn’t important?

Kenworthy: Yes. I try as hard as I can in the book to be clear. But you’re right that it’s a nuanced case. I’m saying, “Look, the evidence doesn’t support this conclusion that reducing income inequality will have all kinds of beneficial effects on other outcomes, but at the same time, you should still care about income inequality.”

The media, ordinary citizens, voters, policymakers, academics can always interpret conclusions of social science research in their own way, and in a way the author didn’t intend. I try to make the argument as clear as possible and reiterate it several times throughout the book. Beyond that, it’s in some ways out of my hands.

We’re at a political moment where many policies in the U.S. are changing that will impact people at the lower end of the income distribution. As a scholar, do you feel you have a responsibility for the public perception of your work?

Kenworthy: It’s definitely something I worry about. Nevertheless, if the evidence doesn’t support the notion that reducing inequality will have beneficial spillover effects, I don’t think it’s helpful to continue to encourage people to embrace that argument or that conclusion. And for those on the conservative side of the political spectrum who don’t care so much about inequality anyway, I doubt this is going to encourage them to work harder to not reduce inequality. They’re already pretty committed to that.

One way I think about this is that since the early 1980s, among conservatives — especially but not only in the United States — it’s been widely believed that tax cuts are good for economic growth and, in fact, good for a variety of other things. That is to say: the benefits of tax cuts or lower taxes will spill over into things like family and community and have a bunch of beneficial impacts.

We know that people respond to financial incentives on average in certain ways, so the hypothesis that tax cuts will yield better economic growth and other good outcomes is quite plausible. Yet the country-level evidence doesn’t support it.

I don’t think the notion that inequality reduction is a cure-all is quite as strongly entrenched on the left side of the political spectrum. But I worry that we’re moving in that direction and have been for about 20 years now. I don’t think it’s helpful to be quiet about that.

Poverty doesn’t seem to appear very often in the book. I was wondering why you didn’t focus on poverty more.

Kenworthy: I do very much focus on poverty, in the sense that I make an argument for policymakers to emphasize lifting the living standards of those at the bottom of the distribution. I just don’t use the term “poverty” a lot. I’m not opposed to it. In fact, in my research I’ve used it countless times. But a big part of the policy approach that I try to make the case for involves increasing government services, and that doesn’t show up in our standard income-based measures of poverty.

For example, if we provide health insurance or childcare or eldercare to everybody, that doesn’t raise their income. It can in the long term, because it allows more people to be in employment, but in a direct sense these services don’t boost income.

Yet these and other services are hugely beneficial in terms of increasing living standards and economic opportunity, especially for people at the bottom. I’m making an argument to focus on these kinds of things. I also think that here in the United States, we should increase the benefit levels for a variety of social programs, and that would have a direct impact on poverty. But I try to use the language of “lifting living standards” rather than “reducing poverty” in order to emphasize the helpfulness of public services.

There’s a second reason, which is: I worry about the fact that we have a variety of poverty measures. Here in the United States, we typically use an absolute measure of poverty, whereas Europeans tend to favor a relative measure of poverty. This is an interesting and important debate, but it’s easy to get sidetracked by definitional disagreements when you tell policymakers to try to reduce poverty. Over the last 10 or 15 years, when I’ve given direct policy advice, I’ve tended to move away from focusing on poverty reduction or using the language of poverty reduction because it’s possible for policymakers to take a different measure of poverty and say, “Oh, look, the poverty rate’s not really that high or it hasn’t gone up very much when we use this other measure.”

As you say, giving benefits such as free childcare and healthcare, which would increase the well-being of ordinary people, wouldn’t show up in measures of reducing poverty. They also wouldn’t show up in measures of inequality. Do you think we should be measuring inequality differently?

Kenworthy: I have mixed feelings. The Congressional Budget Office, which every year puts out a report on income inequality in the United States, includes in their income measure what they consider to be the value of public health insurance (Medicare and Medicaid primarily) and employer contributions to health insurance. The reasoning is that this should be counted as though it were income, because otherwise people would have to pay for health insurance. That makes sense. But imputing to everyone an average value of what we spend on Medicare or Medicaid probably overestimates the true monetary value to some people, especially people with lower incomes who may use healthcare services less frequently.

And there are two additional problems. One is: Why stop at health insurance? There are all kinds of public expenditures, from the military to police to parks to roads to childcare and more. Granted, health insurance is a particularly expensive type of service. Still, if we’re going to do that, we should be doing it for every type of public service the government provides and maybe some other things that charities or employers or families provide.

The other problem is that what tends to happen, when the Congressional Budget Office or others try to include the value of services, is that they usually continue to call it income inequality. But it’s not really income any longer; it’s now something closer to living standards inequality.

My tentative preference is to have two separate measures — to keep our existing measures of income inequality, which focus only on income that people have, and then have a separate measure of something like living standards inequality or material well-being inequality.

The graphs in the book tend to focus on the bottom 99 percent. And that’s an extremely broad group that includes people with very high incomes, and the poor, and the working class. And there’s a point in the book where you talk about measuring the links between inequality and measurements of living standards within the top 50 percent and the bottom 50 percent, but then you decide not to do that. I was wondering if you could explain that decision.

Kenworthy: I use two measures of income inequality for almost all of the analyses. One is the share of income going to the top 1 percent. That’s the measure that Thomas Piketty, his collaborators, and now many analysts tend to use. The other is the Luxembourg Income Study-type measure of inequality within the bottom 99 percent. (The surveys that are the source of most LIS data aren’t good at capturing income in the top 1 percent, due to survey non-response, item non-response, under-reporting of income at the top, and/or top coding.)

I made the choice to use these two measures partly to avoid the analyses becoming too messy and complicated, partly because these are now the two most common ways of measuring income inequality, and, most importantly, because the literature that hypothesizes that income inequality is going to have bad spillover effects on things like health and so on tends to focus on one or the other or occasionally both of these two types of measures.

To me, the most interesting alternative would be income inequality within the bottom half. That is pretty close to a relative measure of poverty, which is the share of people with an income below 50 percent or 60 percent of the median income. (The relative poverty rate tends to correlate strongly with a measure of income inequality within the bottom 50 percent of the distribution. They’re basically the same thing.) But in research on the consequences of income inequality, there’s very little emphasis on inequality within the bottom half. Hardly anyone hypothesizes that this is the real problem area. It’s much more about income inequality throughout the whole distribution.

Also, a significant amount of the analytical leverage that I use in doing these analyses comes from the fact that income inequality has increased at different rates across the affluent democratic countries. But there’s been fairly little increase in income inequality within the bottom 50 percent of the distribution in most of these countries. Most of the rise in income inequality has occurred between the top 1 percent and everyone else or within the top half.

There were a few points in the book where you conclude there isn’t enough evidence to show a causal link between inequality and a particular measure of well-being. Is this just a data or measurement problem, or do you think this suggests that there aren’t causal links, in some cases?

Kenworthy: I look at outcomes in five areas: living standards, democracy, opportunity, health, and happiness. For some outcomes, such as economic growth and life expectancy, we have enough data for me to feel fairly confident that in rich democratic countries there’s been no adverse effect of income inequality, or at worst a very small one. But there are others where I feel less confident because of limited data.

One is opportunity. There, we don’t have data at the country level for enough countries to do a good test. I spend much of the chapter on opportunity looking at indirect tests of the hypothesis that income inequality is bad for equality of opportunity. We have a good bit of data here in the United States that allows comparison across states, across commuting zones, or over time. I also try to look into the particular mechanisms or causal pathways through which the hypothesized adverse effect on opportunity is purportedly working, like neighborhoods and family developments and money that parents spend on their children. I do a lot of indirect analysis and convey the conclusions of others who have done this sort of thing, but I can’t do the kind of analysis where I say, “Let’s compare changes over time in income inequality across these 21 rich democratic countries to changes in equality of opportunity.”

Another is democracy, where the measure is a complicated one. The best measure we can use is the correlation between preferences among people at different income levels and policy changes. And if policy changes correlate more strongly with the preferences of people with higher incomes than of people with middle or lower incomes, we say there’s probably unequal political influence and maybe income inequality is the problem. We’ve now got pretty good data on this for the United States, and there are emerging similar data for other rich democratic countries. But we still have data only for a handful of countries, not enough for me to be really confident about cross-country differences or changes over time.

Read More:

• As the Nordic Model Gains Surprising Popularity in the U.S., Lane Kenworthy Assesses Its Potential Advantages for Families
• Lane Kenworthy’s presentation for Inequality by the Numbers 2020: Cross-National Inequality: Trends, Causes, and Consequences