National Taxes and Local Inequality

ANDREW HAYASHI: Thank you, Warren, for that kind introduction and the long tour through my personal biography. Usually, I write about things that have nothing to do with my own experience like the corporate alternative minimum tax. But actually the inspiration for some of the work I'm going to show you today comes from my own experience moving from Northern California to New York City and then to Charlottesville and observing and experiencing the different nature of inequality in each of those places.

Before I begin though, Warren, I want to thank you. And I also want to thank the foundation for giving me the privilege of speaking to you today. As Warren mentioned, for the last three years, along with Ken I've been serving as faculty liaison to the alumni council.

And it's been really very encouraging. And to be honest, more than a little inspiring actually to attend these meetings with the foundation board and the alumni council and listening to the really good, heartfelt counsel that's given to the dean and to our leadership and to the trust that you have in her and in our leaders.

It's really a gift I think to work at a place where everybody is focused on the good of the institution. And it's not something I take for granted. And we owe you thanks for your role and the board's role in that. I also just briefly want to acknowledge the passing of Justice Souter yesterday.

I saw that this morning. Justice Souter had three UVA alumni clerked for him, including Dean Kendrick. And I want to extend my condolences to her and to other friends of Justice Souter, who may be here today. OK, for the next few minutes, I'm going to talk about measurement, which is probably the worst line to open a presentation with, the easiest way of losing an audience before you begin.

But please bear with me because I teach and write about taxation, which means I spent a lot of time thinking about how the law takes really the very complex reality of people's lives, their personal relationships, the business, and personal dealings they have, and abstracts from all of that complexity something called income, something that's quantifiable, something that can be written down on a tax return, and something that can serve as the basis for determining how much we contribute to government.

And as I teach my fed tax students, when we get that measurement wrong, things can really go awry. And they can go awry in two ways. So one of those ways is that people will take advantage of the mismeasurement of income to reduce their taxes. I would say that when tax law fails to accurately measure income, that is the fertile soil from which tax shelters grow.

My interest today though is in the question of measurement from a tax policy or from the government's perspective. So Congress is currently in the middle of negotiating a tax cut and spending package that among other things, will address a number of expiring tax provisions from the 2017 Tax Cuts and Jobs Act. And if provisions are extended, it'll cost about $4 trillion over the next 10 years.

And when that package is really ready for consideration, the Congressional Budget Office, the Joint Committee on Taxation, any number of nonprofit groups will score the legislation. And they'll estimate its revenue effects, as well as its distributional effects determining how it'll affect taxes and benefits for different parts of the income distribution. And these estimates are going to be crucial for how that legislation is evaluated.

So this is the truth behind the maxim I have on the board behind me. What gets measured, gets managed. A clever quote, which is often but incorrectly attributed to the management guru, Peter Drucker, but it is-- there is there's more than a little bit of truth to that.

The problem is that I think the way that we do these distributional analysis is missing something important. So today I want to talk about how we measure income inequality and in particular, at what level of granularity we should measure income inequality. So most of the media's representation of income inequality is driven by narratives, including specific people, Jeff Bezos, or Elon Musk, or archetypes.

The Connecticut financier or the Silicon Valley tech wizard who most of us will never meet, who live hundreds or thousands of miles away and who probably spend their things, certainly in the case of Elon Musk on very different things than most of the rest of us do. In most parts of the country, income inequality at the top end is driven by lawyers and doctors and the owners of auto dealerships.

So I'm interested in the question of in terms of the experience of inequality and its consequences, what's the right level to measure it? Is it at the level of a city like Charlottesville, Virginia, a state like Virginia, the country, or something else. And the level at which we measure inequality really matters.

So for example, if you lived in Blaine County Idaho, in 2020, you lived in one of the most unequal counties in the entire country. But as a resident of Idaho, you lived actually in one of the most equal states in the country.

So you would have lived in a very unequal county in a quite equal state, in a rather unequal country. What's the relevant-- what's the most important level of aggregation? But we typically almost always estimate, report, discuss income inequality at a national level. Federal tax policies are evaluated as either progressive or regressive, depending on their effects on national income inequality.

But you can think of national inequality as having two components. You can think of it as the sum of inequality within individual regions, plus inequality across regions. The United States income inequality is the sum of inequality within individual states, for example, plus inequality between the states, inequality within individual counties plus inequality across those different counties.

And the problem with our focus on national. Inequality is that it conflates the two. It conflates how inequality has evolved over time within regions and across regions, say, for example, between as between rural and urban areas treating them as equivalent and basically indistinguishable.

And that blind spot prevents us from identifying meaningful differences. I'm going to suggest created by these two different kinds of inequalities and their political consequences. And it also means that when we consider the effects of our tax policies. On inequality, we may be evaluating them at the wrong level, not thinking about their effects on the level at which they matter most.

So to see why it matters when we consider the effects of-- and distinguishing between and across region inequality, you can imagine two different Americas with exactly the same amount of national inequality in America. Number one, people are perfectly segregated by income. We have high income cities, low income cities, and middle income cities.

In America, number two, every city in the country is perfectly representative of the country as a whole. They have their own Elon Musk. They have their class of professionals, academics, lawyers, doctors, a working class, and then a low income group. And I'm going to I want to suggest that inequality in these two Americas is very different with different problems, maybe different benefits and potentials, but also different solutions for how we address it.

So the level at which we measure inequality affects the stories we tell about the current state and history of inequality, the problems we see and that we can identify, and the policies that we choose to address those problems.

So this graph shows the evolution of inequality over time. So the top line in blue shows how inequality at the level of the entire world has evolved over the last 43 years. And you may have to squint a little bit to see it, but it's actually modestly declined since 1980.

Now, at the same time, in the largest economies in the world, the United States, India, and China, income inequality has increased dramatically. So what that means is this blue line is masking two trends going in opposite directions. So global income inequality has fallen mostly because there's been a convergence across countries in their incomes, which is mostly a good thing.

But the trend going in the other direction is the dramatic increase in within country inequality. And so there are really two importantly different things going on during this period, which need to be described and accounted for separately in telling the story of global income inequality in the last 50 years and in understanding its effects on people's lives and the political consequences as well. I suggest.

And if you were to zoom out over time say over the last two centuries, the graph looks like this. So now the dark blue line at the top shows total global inequality since in the last 200 years. And you can see a pretty steady increase over time, which is mostly reflecting that increase in inequality between countries.

Countries actually have been growing apart over this longer time horizon in how rich they are. If you look at inequality within countries, it was mostly steady until the early part of the 20th century and then dropped actually into the middle of the 20th century before rebounding in the last 25 years or so.

So what that means is the phenomenon I showed you on the last slide of growing within country inequality within say the United States, for example. And convergence between countries is really only a pretty recent phenomenon of the last 50 years. Now, we can do the same exercise in the United States.

For example, from 2010 to 2019, inequality within the United States as a whole grew, but it fell in a number of states like Illinois, Delaware, Minnesota, Georgia, and Michigan. These all became more equal places during the 2010s.

We could do the same analysis within a state. So inequality within Virginia also rose in the 2010s, but it fell in a number of places. There are over 130 I think counties and independent cities in Virginia. And 55 of them, inequality fell during this time period, including in Albemarle County and in a number of counties in the Western part of the state.

Inequality grew in Charlottesville, in Staunton, and in a number of counties in the Eastern part of the state, New Kent County, Charles City County, and Middlesex County, for example. So the story of the United States as well, in terms of the evolution of income inequality has multiple layers with different dynamics.

And you can't see it unless you measure and take account of inequality at those different levels. And I don't think and I've read a lot of the research, whether inequality is better or worse for-- at the low, it's better or worse for it to be localized or for it to be distributed across regions. Whether there are some benefits to economic segregation that counterbalance its costs.

I think there are pluses and minuses either way, but we can't even have that discussion unless we're keeping track of these two kinds of inequalities separately. And some of the ills of income inequality are definitely localized. So there's evidence that local income inequality is associated with a variety of negative outcomes, higher crime rates, and poor community health, social mistrust, and political polarization, and lower mobility across generations economically.

Another underappreciated effect of local inequality is how it increases the cost of living disproportionately for low income households. And the reason for this is it turns out rich people and poor people consume different bundles of goods. They buy different kinds and qualities of homes, different clothing, and even different kinds of groceries just as they donate to different charities.

So in the case of wealthy donors, it tends to be arts and culture organizations, private schools, and hospitals. And in the case of middle and lower income individuals, they tend to be local religious organizations, food banks, shelters, and community organizations.

And the local economy tends to reflect the tastes, the preferences, the spending patterns of higher income residents crowding out and driving up the cost of living for low income households. Stores, for example, tend to favor higher income consumers more in wealthy areas than in lower ones.

So consider just one example. So Flint Michigan has a median household income of about $36,000. Bridgeport, Connecticut has a median household income of about $56,000. And consider somebody earning $25,000. If they decide to move to Bridgeport, Connecticut and consume exactly the same groceries that they were buying in Flint, their grocery bill goes up by 9%.

Somebody's earning $200,000 and buying the kinds of groceries that somebody earning $200,000 buys who moves to Bridgeport, Connecticut, you actually see their grocery bill fall by 19%. Moreover, many of the things that the low income individual is used to buying in Flint are simply not available in Bridgeport.

You probably would not be surprised to learn there are $14 general stores in Flint, Michigan, and there is exactly one in Bridgeport, Connecticut. The fact that we don't measure or keep track of local income inequality means that we're blind to the fact that income that the incomes people report can vary dramatically across the country and the goods that they're able to buy. It has very different purchasing power depending on where you live in the country.

And this neglect of local inequality also means that our policy attention is overwhelmingly focused on the people and the parts of the country that drive national income inequality. So this is a map showing the counties in the United States that contribute the most to what the media will report as our source of national income inequality measures. And they're represented in dark blue.

And you may recognize some of them. A lot of them are in coastal areas Los Angeles County, Miami-Dade County, New York County, Cook County, and Harris County, for example. And if you're focused on national inequality, these are the places and it's the industries in these places that you're going to focus your attention.

But if you focus on local inequality, you get a very different map with a different much more diverse set of counties. And this slide shows, again, in dark blue, which counties are the most unequal in the country. Blaine County Idaho, which I mentioned a few minutes ago, is on this list.

It's not only the Fourth most unequal County in the country, but also has undergone one of the biggest increases in inequality over the last 20 years. You can see it in the data. You can see it in the local reporting where there's a lot of anxiety about this influx of high income migrants from California and from South America.

Our own Charlottesville, I did not rig this list actually appears. I don't if that's number maybe 13 or so among the most unequal counties or independent cities in the country. So this map looks very different than the last map. And given the differences between those two maps, it shouldn't be surprising that when you adopt policies that have one effect on national inequality, they could have very different effects at the local level in your community, which one you focus on really matters.

If we were to think, try and analyze, for example, the Tax Cuts and Jobs Act and these expiring provisions that may be extended over the next 10 years, the story doesn't look very different at a local level than it does at a national level. That legislation was regressive nationally increased inequality locally and across regions.

But other federal policies have much more nuanced effects. So to illustrate this, we can consider the moratorium on student loan interest that was implemented as part of the CARES Act in 2020. This graph here shows the effect of the student interest moratorium on national inequality.

So on the horizontal axis here, you have the effect of the policy on inequality. And the fact that vertical line is to the left of 0 means it was negative. It reduced national inequality. And this was pitched as a progressive policy and its effect indeed on national income inequality represented by this black line was in fact progressive.

But there are many parts of the country-- Charlottesville probably being one of them-- where the people who benefit the most from student debt relief are doctors and lawyers who are right near the top of the local income distribution. In these places, the policy increased income inequality.

So this slide shows the wide variety of consequences that the student loan interest moratorium had across the country in different counties. And you can see there's a wide range of effects. And there are a number of counties, all the ones in blue where the interest rate moratorium reduced local inequality.

But all the counties in the red are places that became more unequal as a result of a policy that was pitched as progressive on a national level. So it really matters that we distinguish between local and national inequality. Some of the provisions, the expiring provisions affected by this year's negotiations probably have similar features as the student interest moratorium having different effects on the national and local level.

So for example, the mortgage interest deduction, which benefits lower, higher income households who tend to itemize their deductions, and are likely to have large mortgages on valuable homes. But probably doesn't benefit the very richest, who are more likely to own their homes outright.

There's also a deduction scheduled to expire for income from services in the fields of law, accounting, financial services, and investment management. That phases out, as a lot of these benefits do for incomes above $400,000.

Incidentally, it's a pretty frequent refrain in Washington over the last several administrations that taxes should never go up for the middle class. And the cut point, apparently is $400,000. I hope I don't need to show you the data to convince you that in many parts of the country, the middle class is not earning $400,000.

All right, so what to do? I've complicated the picture of national inequality. And there's no easy fix for local income inequality. You might wonder, well, maybe there's an increased role for state and local tax policy. And maybe there is one well-known concern with strongly redistributive policies at a local level is it's actually quite easy for people and businesses to get up and move across county boundaries or even across state boundaries in some areas to avoid higher rates of tax.

One interesting proposal involves indexing federal taxes for the local cost of living to reflect the real purchasing power of income in various parts of the country. But if you think about what that means, practically it's that there would be lower taxes in higher cost of living parts of the country, which is one tax observer put it, means that such a law might as well be called the Bicoastal Elite Tax Relief Act. And I'll let you speculate about the political prospects of legislation with that name.

But to do anything about local income inequality, we need to start measuring it. And it would be helpful for it to become common practice in assessing the distributional effects of federal legislation to track its effects both within and across regions.

In tax policy, I do believe that what gets measured, gets managed. And developing the capacity to do that, I think, would be a really important service to lawmakers and is something that I'm currently working on. So most of the numbers that I've shown you today come out of analysis that I did using publicly available data from the IRS website.

And the goal is to make that data available through a tool that can be used by anybody to estimate the effects of tax or other spending policy and its effects on local inequality really to complement the national analysis, not entirely to displace it, which I think will provide a richer picture of the effects of our policies and hopefully, spark more research on the evolving differences between economic conditions in our communities and across them. So that's it. Thank you.