The Future of Inequality - Branko Milanović

Now, this is an excellent question. Now, let me start like from my narrow perspective with somebody who works on inequality, on income inequality. The situation becomes even more difficult with unrealized capital gains. And you know, just let me give a comparison. Unrealized capital gain on stocks that you have or wealth that you hold in like, you know, mutual fund or something like that, is very similar to the painting that you have in your room whose value for some reason, you know, such a painter became now much more popular than before. The value has gone up. But has anything changed in reality? Yes, your wealth has changed in the case you wanted to sell that painting or you wanted to sell the stock that you had. But actually otherwise nothing has really changed. So, I'm really somewhat reluctant on sort of going big time on...

Well, thank you very much for inviting me. It was a little bit difficult to set it up all on without the work between the different time zones and technical problems, but I'm very glad that we are here now.

You know, this is really a difficult question and also an easy question. It's difficult because we try to actually understand something which will happen in the future and we make analogies with the past. It's easy in the sense that objectively nobody knows. So, we can be all equally wrong. Now, let me explain what is my current thinking on that. My previous thinking was that I saw it as a technological change like any other over the last 200 years. This is essentially you replace labor, sometimes skilled labor, sometimes unskilled labor, but you replace it with capital. And of course artificial intelligence is capital, you know, we are used to seeing capital as a machine, but artificial intelligence is also a machine that you might not see in a physical sense the way that you see like for example a construction engine, but it's the same idea. However, I think that there is now, and that's why I changed a little bit my mind, there is now something new. And that something new is the ability of the machine to invent itself. So, that's something that we did not have before. In other words, artificial intelligence, if it becomes, and it might become what people sort of think now, it's not simply going to do what, for example, I've used it a little bit on the internet, collect information and put it in a way that you can read and maybe summarize and so on. It would actually learn and become have agency. And actually when the machines have agency, I think then the situation changes. Then it really almost becomes very difficult to put it exactly where it belongs because it could become labor. So, let us suppose that artificial intelligence has this agency, that it starts doing things that you and I are doing now, and then you do have a big change in the sense that even if you look from the point of view of factors of production, it could become labor. It's no longer a tool because a tool is always sort of used by a laborer and the laborer says to the tool what to do. But in this case the tool will become the labor. So that's why I think it is new and I am not very much into big hypes, but when I think of this I do think that it is a significant change, a new change, something that we have never experienced before.

You know, this is of course one way of looking at it, but let me continue what I was saying before. Let us think of this as really transformative technology. And if you think of that like that, then you are no longer simply thinking, okay, well, this machine might replace, you know, labor that for example now is working in India and it's linked to the international labor. It might replace this particular people by machines. But you have one has to start to thinking that actually it is such a quantum leap that actually developing countries have an advantage in the sense that everybody starts more or less from the same position. So in other words, the technological changes we have seen now is really driven by maybe three countries, you know, US, China and India. And each of them is really basically at the very similar level of development. So in that sense I would see it as a big opportunity for well educated and also countries that can actually are able to apply that not only to catch up in the usual sense, but really do what for example China has clearly done. It's no longer catching up, it's actually pushing the frontier. So I am in that sense also optimistic. As I said, I don't think we have to think of this only simply within the current confines where it would replace somebody like where the taxi drivers would be replaced by autonomous cars. We have to really think of this as possible really big change. Even if you look at Africa, you know, Africa would always start also with a clean slate in that sense. And Africa would have an advantage that it has many, like India, many young people who have shown themselves to be much better, including myself who actually had to have support from my son in order to start this conversation. So, that's the advantage.

You know, I don't know, you know, obviously I mean there was always talk that it would replace people like, you know, I mentioned before taxi drivers. It can replace cleaners. We can replace waiters. I've seen of course robots that actually deliver food. It can replace professors. It can because you basically can have autonomous professor who is going to tell you more or less the same thing that I'm telling the students. People of course said that actually it would replace, that was rather obvious, legal offices, junior clerks who write things basically summarizing the previous cases. It might also replace people in medical professions. But I think like every other technology it would create also new jobs that we cannot imagine now. And I'm quoting for example many years ago Nassim Taleb wrote that very nicely. He says if you're like a stone age expert and your tribal chief asks you to tell him what would be the new technologies that would be developed. If you knew what were the new technologies you would have already developed them. So, you know, we cannot in advance know what would be new technologies and new jobs that would appear that this particular development would require. I remember when computers and PCs became prevalent we couldn't imagine like what kind of jobs they would create. And they did create quite a lot of other jobs of people including the people who are, you know, obviously like servicing the computer servers who are doing architectural design and so on. So, we don't know for the artificial intelligence either. But, just to finish this long answer, let me just say that the part which I think is very difficult to predict is that transformative nature of artificial intelligence itself when it becomes a new labor. And that's something that I think is totally new.

You know, I would be very much against taxing specifically things that have to do with artificial intelligence in the sense that if you make a lot of money out of artificial intelligence, you should be taxed differently than other people. I don't think it makes sense. I think it is actually done directly opposed to technological progress. And I don't think that we should be opposed to technological progress. But, if people become billionaires thanks to the artificial intelligence and those developments, they should be treated like other billionaires. And I do actually think that in many cases billionaire tax or tax on really extraordinary high wealth is something that should be done. So, whether that wealth is made because you did it in real estate or casinos or Miss Universe as Trump did or you did it artificial intelligence, I think you should be treated the same.

No, this is an excellent question. You know, let me start like from my narrow perspective of somebody who works on inequality, on income inequality. I don't work that much on wealth inequality, but on income. Unrealized, first we have a problem with realized capital gains because they're really not part of income. They're part of your increase of wealth. Your wealth is what is called like below the line. So, we look at the flows. Flow is income and wealth is below the line. And I mean, even realized capital gain is something which increases your wealth. It's not a flow. But, the situation becomes even more difficult with unrealized capital gains. And you know, just let me make a comparison. Unrealized capital gain on stocks that you have or wealth that you hold in like, you know, mutual fund or something like that, it's very similar to the painting that you have in your room whose value for some reason, you know, such a painter became much more popular than before, the value has gone up. But, has anything changed in reality? Yes, your wealth has changed in the case you wanted to sell that painting or you wanted to sell the stock that you have. But, actually otherwise, nothing has really changed. So, I'm really somewhat reluctant on sort of going big time on unrealized capital gains because they neither increase the consumption of people who are actually having them and they don't, and if you believe that there is some type of consumption that is ostentatious or ostentatory and that you would like actually to limit, this doesn't come into that play. They are actually, they're a difficult issue both from the point of view of accounting and from the point of view of taxation. But, I'm not, in some sense you can actually even see in a very Marxist or Ricardian way unrealized capital gains is a kind of, how should I say? It is a kind of homage paid by the rich to development. In other words, if they are always unrealized, they would be permanently fueling investments. And in that sense, this is something that also you don't want to discourage.

But you know, the problem becomes, I think actually I'm not a tax expert, but I think the problem starts and becomes a problem when either you want to liquidate, I mean to sort of basically take into cash to realize the shares that you owned or the painting that you own, or you want to actually give it to somebody, to your family, to your sons or others. At that point, I think you have to tax it because it becomes then realized and you have to tax it. But taxing it before coming to that point, I really do not see exactly the purpose of that because as I was saying before, it is not increasing consumption in an ostentatious way. It is essentially like the painting that you keep on your wall. Moreover, it is a productive capital so it is actually used primarily to grow up and to actually invest. So, we have the situation, you know, Keynes in the opening paragraphs of The Economic Consequences of the Peace wrote, I mean, somewhat idealized view of Europe before 1914 and he said the system was very unequal or inequitable, but it maintained itself because the assumption was that the rich people would actually mostly use the money that they have made to do investments. And in reality, if you have rich people who practically do that all the time, that it is not that much of a problem until it is withdrawn and it is inherited by somebody. At that point, yes, I think it needs to be taxed.

No, I do see it actually what you're saying exactly it is true and that's another technological change which was accelerated as you know by COVID. We had before COVID also the ability to use online basically jobs to have them, but they're relatively few and then COVID really pushed us in that direction. Technically, yes, it is true that actually if one were to move further and further in that direction, which can happen, you really don't need to be physically in one country and not another country. You can do the same job and pay it exactly the same and so forth. And actually, even if you can say you're in a poorer country and the price level in that country is lower, you can actually have in real terms a higher salary there than you would have it elsewhere. But my point was more generally, it was not only about the labor market and the point that you made, but it's also the very fact that there are still huge differences in income levels between the countries. Just to illustrate that, you can take the median income of a person in Tanzania and you have a median income of a person say in Norway and the ratio is probably, I don't know offhand, but it's probably more than 10. And this is really the citizenship premium because you could have exactly the same person driving a bus, you know, in Kampala and the person in Uppsala and they would have a different wage. So, that was the premium. And moreover, what I actually meant, which I think is an interesting point and maybe I should have elaborated it more, is that that premium of course depends on the bilateral comparison between the countries, but it also depends on where you expect to end up in the income distribution of the rich country. Because if you, for example, going back to the example of Tanzania, if you expect that you're going to be in a lower part of the distribution of a rich country to which you migrate, then for you it's better to go to a country like Sweden than to a country like the US. But if you expect the opposite, if you expect actually that you're fairly highly qualified, you actually like to go to a country with very high inequality because you expect to end up in the top of the income distribution there. So, the premium itself, and that's my last point, is not simply one number. It is a number depending on where you are in the income distribution of the emitting country and where you expect to be in the income distribution of the receiving country. And in function of that, you can actually, it varies along the income distribution.

Yes, it does. But let me, I mean, sort of divide your question in two parts. First question, is it true that really people at the top of the income distribution, I'm talking about income distribution, not wealth. Wealth distribution is even more unequal and I think is more inequitable also in terms of taxation, but let's stay with income. Is it true that they are paying less than they used to pay? First, yes, generally true because of course the top marginal rates were much higher before than they are now. Although it has to be said that actually this top marginal rates that used to be 90% really covered a very small number of people. But I think they do pay now less. That was the argument that Zucman guys made. I think that they are right. They probably pay less. But what is even more sort of striking, and that's my second point, is that they argue, and they show that the average tax rate in that particular range, which is lately like top 1% or 2% at most, is lower than the average tax rate that is paid, for example, by people who are at the top decile. So also rich people, but not as rich as the top 1%. So then we have really paradoxical situation that the progressivity of the system, like the American system, is really working well up to 98% because the average rate is higher for richer people. Not only marginal, but the average. But then you come to the top and then suddenly for the top you have a reversal that the average rate goes down. And then clearly the system does not maintain progressivity throughout. So I believe, you know, in their findings. I have to say that in the household surveys you do not find that, but you do not find that decrease in progressivity. You find basically flat progress, like flat rates towards the end, but in part because taxation, part of the direct taxation in household surveys, is imputed, not actually taxation that is paid or which IRS has the data. In household surveys it is imputed. So obviously, if these people at the top have really shenanigans to hide their income or to pretend to put them in a different trust or actually to use corporate taxes instead of personal, I think there is an under and overestimation of how much they are paying. I'm not sure if I made this correct, but what I want to say is that I think that towards the end of the income distribution, on the top of the income distribution in the US, you do not have progressivity anymore, but you have regressivity because the average tax rate at the top is smaller than the average tax rate for the previous percentiles.

Yes, yes, I'm sorry I forgot the second part. Yes, they are talking a lot and actually as you know in the United States there is also struggle between state taxations and people move for example from California, they move to Texas, they move to, you know, Kentucky where, you know, state tax is zero and it's not negligible because basically you have in high tax states you pay something, you know, two digit number maybe it is 10%, 12% and in no tax states you pay zero. So there is a movement within the United States. Now whether that movement is significant worldwide, well we have seen like for example now with Dubai, there was lots of talk about of course people from England in particular for the US and so on, Russia, China, they go to Dubai and that was really tax haven, but we have seen sort of problems now with the war. I'm not totally convinced that that problem is huge, but maybe I'm wrong on that because if we combine the tax jurisdictional incentives that we talk about with what you said before, ability to do your job from wherever you are in the world, yes, I could see the advantage that some people would really sort of see by moving to places that are really tax havens like Dubai.

Yes, I think actually when I wrote the book, you know, obviously the book is couched in the way of liberal meritocratic capitalism. By meritocratic, I didn't mean that there's something meritocratic in the sense that people get what they deserve. Meritocratic is self-perception. It does not mean necessarily reality. So, liberal capitalism and political capitalism. I think the term political capitalism, as I said, actually comes from Max Weber, and I think that China fitted that model quite well, and it continues to fit that model. However, in the meantime, as you said, I think that we have seen other examples of political capitalism, which are not maybe as clear-cut as China is, but India definitely has moved in that direction. Russia is obviously there. Turkey has moved. And I don't want to use that in a pejorative way. And obviously, United States, you can say under Trump has moved there. I just see it as a different way of, how should I say? Different way of managing and regulating a polity. So, in the book, I try to avoid that value judgment one is good or another is bad. There are certain advantages, and if you look at actually at all these countries that I mentioned, well, not all, but most of them, people forget that, for example, for Turkey, Turkey had the highest growth rate among the rich countries. It has now, for the first time in history, overtaken in GDP per capita Greece. That has never happened before. And has really created a very significant middle class. So, when we speak of Erdoğan, we tend to forget that there are actually, indeed, you know, many people who vote for him. And likewise for Modi in India. So, you know, we should not simply say, you know, this is a system that somehow is flawed. You know, it has certain features that we might not like, but on the other hand, it is able to deliver economic growth.

No, I totally agree. We should be doing that and actually we are doing that. The issue there is the following. You know, I worked on economic history actually on that particular topic by looking at the 19th century data. Now, for the 19th century, and we are trying not to really do much more than before. For 19th century, we have reasonably good estimates of the GDP per capita. You can actually disagree with the Maddison data and Maddison project is revising them. What we do not have is we really have very, very poor data on income distribution by large countries. So, this is where there is I think a push now with I think at most, I think we might have some estimates for about a dozen countries and maybe only two or three points in time. As we get more data, I think that the comparisons will become more common. But there is no doubt, even if you take only the GDP per capita as the mean and then impose any kind of distribution to that mean. Let's suppose you take some reason seems reasonable distribution. You would still find that person who are the median at that point, let's suppose the person who is a median in India in 1850 and a person at the median today, obviously they would differ in real terms enormously. So, yes, we have had a huge increase in mean income in the world. You know, there is this chart which is a sort of like a hockey stick that it is true actually when you look at the mean, for example, GDP per capita in the world then and now you obtain exactly that result. So, yes, we have had, we should actually do much more of that, but we are limited by data.