Nishant Yonzan earned his Ph.D. in Economics from the Graduate Center in September and is now working at the World Bank in Washington, where he analyzes the effects of Covid-19 and other international crises on global poverty and inequality. Yonzan is also an author of a recently published study in a special issue of The Journal of Economic Inequality, on which he collaborated with Stone Center scholars Branko Milanovic, Salvatore Morelli, and Janet Gornick. (For more from the special issue, which includes contributions by Stone Center postdoctoral scholar Ignacio Flores and affiliated scholar Stephen P. Jenkins and was edited by affiliated scholar Nora Lustig, see the Read More links below.)

Your study compares the estimation of top incomes from tax data and survey data. What are your main findings?

Yonzan: The study aims to understand the difference in incomes between two data sources: household surveys, and tax data. First, we focus on where — i.e., in which part of the distribution — these differences start. To do this, we rank everyone from the poorest to the richest in the respective distributions, and compare the differences in total, average, and share of income held by each income percentile. We find that the difference, for the three countries we study — the U.S., France, and Germany — is limited to the richest 1 percent.

Second, we try to see why these differences exist. The differences are primarily driven by non-labor components of income, such as income from capital, self-owned business, and rent. Tax data capture more non-labor income than survey data and hence the top percentile in tax data, whose total income consists of a large component of non-labor income, captures more total income than its counterpart in survey data. We additionally suspect that tax policies influence this gap between tax and survey data. In the U.S., for example, the 1986 Tax Reform Act created incentives to shift corporate income to personal income. Consequently, for the top 1 percent, we see an increased discrepancy between what is reported in tax data versus what is reported in surveys post-1986 compared to before.

How did you get involved in the research that led to this paper? And are you working on any other projects related to this topic?

Yonzan: I’d started to work with some LIS data, and was investigating various aspects of the income distribution without a clear idea about what exactly I was looking for. That is when I chatted with Branko, who had started examining the difference between survey and tax data sources to see how the Stone Center could contribute to this discussion in this area. (Editor’s note: for more on this topic, see “Measuring Inequality and the ‘Missing Rich,’” a Scholar Interview with Nora Lustig.) Janet and Salvatore, who were also thinking along similar lines, soon became involved in the research and later the writing of the paper.

Since then, I’ve started another study along the same lines. I’m working on a paper with Laxman Timilsina, a colleague from the Graduate Center, in which we’re trying to further the why question. We suspect tax policies of driving some of the changes in the gap between survey and tax data, and we would like to quantify the impact of tax policies on this gap.

I’m also working on the question of what can we do about the gap. We know that survey data, especially for the very top earners, doesn’t capture all income. Can we adjust the incomes for the top earners in surveys to “correct” for this shortfall? It is important to keep using survey data for the rest of the distribution because it not only covers the whole population (unlike tax data, which cover only a segment of the population in most countries) but also includes very detailed information about household and demographic characteristics: people’s ages, incomes, education level, and so forth. Thus, survey data are generally more suited for welfare studies. One way would be to adjust the top with tax data — that is, to use the tax data for the very top of the distribution to make adjustments to survey data. This is certainly feasible in higher income countries where fiscal capacities are more developed and tax data do a fair job of capturing incomes of the top. In most countries in the developing world, however, fiscal capacities aren’t as well developed. Could we, in those countries, use other data sources, such as data of loans from financial institutions, etc., to adjust the top of surveys?

What was it like to collaborate with three professors at the Stone Center?

Yonzan: We started the research at a time when I was just finishing up my classwork and getting into research. By that time, you generally know what direction you want to go, what you’re interested in broadly, but finding very specific questions to work on can be difficult. Talking with Branko and Salvatore and Janet helped me to be more focused and ask more precise questions. These discussions helped me to find a direction.

And, obviously, working with senior colleagues has a lot of benefits. If I’d done this on my own, the structure of the paper would be entirely different. There would be perhaps more discussion than needed; it wouldn’t be as precise as it is now.

Another helpful thing was that I could go constantly to them. If I got stuck, I would talk to them, and they all have different specialties, and so their thinking is also very different. I could get instant feedback from these different perspectives.

What are you currently working on at the World Bank? I saw that some of your research is on the impact of Covid-19 on poverty and inequality.

Yonzan: My team is focused on monitoring global poverty and inequality, and we do analytical and data work related to this. For about the last year and a half, I’ve been working on the impact of Covid as part of the bigger picture, such as: Are people entering poverty? How has that affected poverty and inequality worldwide? We found that global extreme poverty — i.e., those who live below $1.90 a day (adjusted for differences in prices across countries) — increased for the first time since 1997 (during the Asian financial crisis). Extreme poverty had been decreasing for the last two decades or so, but now it’s going up. And the magnitude of the increase is also very large. The increase in poverty due to the pandemic will be the largest increase in at least three decades, and a lot larger than what we saw back in 1997.

Regarding income inequality, within countries, at least initially in 2020, on average everyone’s incomes were negatively affected, at least compared to what would have been the case without the pandemic. But then the pandemic had differential effects across countries. The differences between countries looks to have increased global inequality for the first time in three decades. And that was mostly driven by the differences in income shock between countries rather than within them.

Now we’re shifting our focus from Covid to other crises, like the war in Ukraine and how inflation is affecting poverty and inequality. So I’m thinking about that as well.

What initially led you to study inequality? Was that your primary interest when you came to the Graduate Center?

Yonzan: I’m from Nepal, and growing up, I saw poverty all around me. I was one of the lucky ones who got to go to school and get an education and so forth. But if you went on trips outside the capital city, you’d see extreme poverty, really extreme poverty. That was something I was very interested in trying to understand — and trying to help in any way I could — and that was sort of the starting point for why I wanted to go to graduate school.

But then along the way sometime, I realized how important the distribution was — not only the bottom of the distribution, but how the bottom of the distribution is influenced quite a bit by how things are distributed among everyone. So that led me to study more about the whole distribution, focusing on inequality as well as poverty. And this paper shaped my agenda as well, because it focused me on studying what the different data sources are, what the differences between them are, and how to use them.

Read More: