We asked you, the viewers, how you think taxes should be collected, and here are your answers. So, we asked, what is worth taxing in the age of artificial intelligence? We have 35% for consumption, 31% for billionaires, 19% for computing power, nine for robots, and only five for land. Zanni, what do you think about this? Well, I think, as always, we have an extremely informed, thoughtful, and wonderful audience. I completely agree with them. You know, consumption is what economists generally consider to be an extremely efficient way to collect tax revenue. You can tax consumption and make it progressive, for example through social benefits or introducing large benefits before you start paying such a tax. So, the consumption tax can
be made progressive. However, I don't think that's politically possible in the world we're talking about, where there will be huge profits. We are not entirely sure who will receive them, but capital and its owners should be taxed not only on consumption. So while I think " billionaires" is hopefully shorthand for capital gains taxes. I think there will also be taxes on capital. Again, a tax on billionaires is very attractive from a political perspective. But you won't really collect much, plus they are very mobile and can find a place to move to, and so on.
So I'm not necessarily in favor of a billionaire tax, but I think overall we need a system that's more focused on a consumption tax with some capital taxation. And I would put the land on a higher level. You know, I still believe that land, as a scarce resource, can be usefully and effectively taxed, that would be my option. It's worth remembering that this means the biggest changes for the US, where consumption taxes are lower than in Europe. Europe already collects a lot of revenue from consumption taxes. So these changes will be more massive where I am now than anywhere else. Okay, I want to go into more detail on each of these elements later, but before that, Tom, let's start at the most theoretical level.
What is a good tax system? What are its characteristics? Yes, of course. I think the first thing we need to think about is why we need a tax system at all, and I think there are two main reasons. The first is revenue collection, and we must not forget that this is needed to fund public services. You must be able to raise enough funds. And the second, depending on the political color of the government, is redistribution. That is, helping the poorest and, perhaps, taking some of the income from the richest. How to do this-there are certain principles to consider. Simplicity. You don't want to spend hours of your life, like I did when I lived in America for several years, filling out tax returns, trying to figure out whether I'm eligible for this or that
credit, that's not good. You want to have a simple tax system. You want to have an efficient system. Well, that's the word. What does it mean? I think this means that you want to tax activities in a way that minimizes distortions, and tax different activities equally. So you don't want to encourage people to do one thing and not do another. And then there is progressivity. If you think billionaires will make a lot of money from AI, they should probably pay more taxes. Sandy, how does AI change this theoretical optimum for you in any way? How does this distort that, you know, old picture of how it should be done? I, you know, those three main goals. And it is worth noting that these three goals, as Thomas described them
, are somewhat contradictory. But the basis, you know, you want to get, how was it there? Income is like plucking feathers from a goose with minimal hissing Colbert, yes. No, Colbert was right about that. And you want minimal hiss, so you want it to be efficient. You don't want everything to be too concentrated. You don't want the activity to be distorted. Now, if we find ourselves in a world with zero returns to labor, you optimize where the profits are, you know, you have to go where the money is. Um, but you still want to do it, I suppose, with minimal distortion and maximum simplicity. I suggest that AI makes the focus on simplicity less important for the reasons we just discussed. If you can, you know, if you can use your AI to optimize
or minimize your tax bill, then you probably don't have that cost because of the complexity. However, I would still advocate for a very simple tax system. Um, can I add to that? One of the things I think people get confused about when they talk about taxes is that they want all these principles to be followed by every single tax, when in fact you need to think about the system as a whole. So you can use a consumption tax to raise a lot of money, but you can use a billionaire tax or something like that as a progressive element, and then redistribute some of that money to the poorest if they lose out on the consumption tax. So you don't have to achieve every goal
with every single tax, and that's one of the big problems that got us into this mess in the first place. Although I do believe that consumption taxes are quite unfairly criticized on the issue of progressivity. Many people have this model in their heads. Consumption taxes are flat, not progressive, so you need a more redistributive tax on top. You know, if you're worried about billionaires having a lot of money, if I raise the consumption tax by five percentage points, I'm going to reduce the economic value of a billionaire's wealth by five percent. For what is the use of wealth, what is
wealth for? They are needed for expenses. So, from a purely economic perspective, increasing the consumption tax is a bit like a wealth tax. But if you tell that to someone like Gabriel Zuckman, who I interviewed for Inside Economics, a big proponent of wealth taxes, he will say that the reason a wealth tax is needed is because of the political power that comes with wealth. It's not just about the economy. Where do you actually get wealth through consumption taxation, because if you never spend your wealth, you know what good it does? He says it's useful for influencing politics, buying out competitors, and so on. And that's why you need a tax system, which is, you know, a valid argument, but not a purely economic one.
Is that all? Yes, I think both Tom and Henry made important points, but Tom's point is that it's important to look not just at individual taxes, but at the tax system as a whole. And, frankly, it's worth looking not only at the tax system, but also at the system of taxation and redistribution. Because if you look at taxes alone, you know, the income tax system in the US is actually extremely progressive. And I don't remember the exact share of income tax that the richest 1% pays, but it's extremely high. Overall, redistribution in the US is less than in many European countries, although countries there rely on a less progressive consumption tax. And this is because the state there is larger and carries out more redistribution
. So we're making life a little more complicated for ourselves, but it's actually too simplistic to just ask, "What tax system is optimal?" We need to think about what system of taxation and spending is optimal to achieve the desired balance between progressivity, efficiency, and transparency. But the test question that we have posed to ourselves-and I think rightly so-is that there will be pressure for tax reform because governments simply will not be able to raise the necessary revenues under the current system. In addition, it inhibits economic growth and distorts the market. But when we take a blank sheet of paper and say, "Here's tax reform from The Economist." Which, by the way, guarantees that it won't go anywhere, but if we do that, I think
we should ask ourselves: are we assuming that the size of the state remains the same? So, if you are in France, what would be the optimal tax system for France compared to the United States? What do we assume is happening on the expenditure side? And only by answering these questions, I think we will be able to truly decide what kind of tax system we want. This also shows that what Danny said earlier was true: progressivity and simplicity can be in tension, as it is possible to make significant redistribution through spending while maintaining equal taxation. But people really want the tax system to be
progressive. And every time you make changes to taxes, people want them to become more progressive. So they end up coming up with increasingly complex ways to do it. As a result, you get a complex system. So this compromise is difficult. Has Zuckman convinced you that wealth taxes are a good idea? No, I wouldn't say that. No. I mean, the history of their use shows that countries that try them eventually Yes, definitely. And, as I think you said earlier, billionaires are usually very mobile. Personally, I am not convinced that the political influence of billionaires has gotten out of control. If I ever supported a wealth tax, it would only be on the grounds that billionaires have excessive political power, and that a instead of a
meritocratic society. I think if you're pushed to the edge, you have to be willing to go for it. But I believe that the situation should become truly critical. And I'm not sure that society has reached that point yet. And from a pure economic perspective, you know, I really see a problem with redistribution through the tax system: the principle that you need to redistribute funds through taxes has no natural limits. People will always strive to push it as far as possible. And wealth taxes, at least those under consideration in California, are in some respects very similar to ordinary property expropriation. And I find that quite disturbing. So no, he didn't convince me.
I also think that we already have certain taxes related to wealth, like the focusing on new ones that won't work, we should focus on improving the ones we already have. Capital gains taxes are very complex, but they often undertax the huge profits that AI billionaires can make and overtax the ordinary income that the average investor makes. So reforming this, making it fairer, might be a better starting point than introducing a completely new tax.