AI Can Solve Wealth Inequality
A few months ago, I stumbled upon a research paper teeming with fascinating discoveries. It was a rigorous economic study, but came decorated with a colourful animation reminiscent of an Animal Crossing game. It demonstrated an AI simulation. Little agents would gather resources, build houses and trade with one-another. All the while, a different agent came up with new tax policies that would redistribute the wealth of these virtual citizens.
This experiment taught me a lot. It got me thinking about the dominant opinions people hold about how their wealth should be allocated (or retained). It seems like even prominent public figures, whose thoughts inspire and mould the minds of their many millions of listeners, have a very weak grasp of basic economic principles and arguments. One of these, I believe, is the common opinion that the free market economy is the best of all.
The free market principle is the heart of Capitalism. The concept centers around the power of supply and demand. In order to achieve the optimal price, where both consumers and producers are satisfied, these two forces must eternally perform a balancing act. Another description for this driving force towards optimal prices (also called price equilibrium) is the ‘guiding of the invisible hand’.
Living under the free market definitely sounds better than say, living under the thumb of a monarch, oligarch or fascist ruler. Certainly, in the time of Adam Smith (who founded the theory), this way of living sounded more Utopic than it does in today’s context. But think about what it would be like to have a corporation control all the means of production- doesn’t that sound terrifying (and plausible, given the hegemony of Big Tech over the world today, especially Amazon)?
For a country to be truly said to have a free market, you would need to take out any external forces on the marketplace. For example, government intervention, such as taxes, subsidies, and other forms of legislation. After all, these forces may push against the direction of invisible hand of the free market. If there were no governement intervention, how exactly would a country build infrastructure? How would it build education systems? How does a government even come to exist, if it has no agency to take some of the wealth from individuals who live within its borders? Unless you want to live in a city straight from the Golden Age of Piracy, it’s unlikely you’d ever find yourself living in a free market society. Even Smith made the argument that there are certain functions a government needs to perform, and he explicitly states that the rich should be taxed in “something more than in proportion” to their wealth.
When most people talk about the free market, they implicitly concede a few things like the need for taxation (at least, I assume they do this). So, given some basic concessions like a tax on people’s property and revenue, what other problems does the free market principle run into? The most obvious problem is how do we decide who to tax and how much do we tax them? It’s one of those questions that can turn a conversation turbulent, with sparks of passion flaring on both sides of the debate. This controversy can be distilled to two fundamental elements: productivity and equality. Both are desirable, but it can easily be a zero-sum game.
It turns out that in at least a few economic models, a free market tax policy system maximizes productivity. But having large amounts of production doesn’t mean that a sufficient number of people will reap the rewards of it. This is the essential problem of wealth inequality, which is so rampant in the USA that the three richest men own more wealth than the poorest 50% of the country. Chances are that both you and I are in the top 5% of the world in terms of wealthy living. Our little 5% club, a modicum, when compared to remaining 7 billion people, own 80% of the world’s wealth.
Is wealth inequality always bad? When you look at the staggering salaries of prestigious brain-surgeons, you may finding yourself nodding with approbation. After all, a brain surgeon is far more useful to society than say, a rural farmer or janitor. Unfortunately, this is not the current state of wealth inequality, at least not in the USA. Most Nobel Peace price winners, scientists, and brain surgeons fall sharply outside the top 1% of richest Americans.
The real wealth holders tend to be people who inherited it, and 50% of wealth in the country is inherited or gifted. Of the mostly highly payed workers, financial executives and CEOs, there seems to be a massive disparity between what they actually contribute and how much they get paid. Daniel Kahneman analyzed 25 wealth advisers, and their performance over 8 consecutive years. Looking at whether these advisors improved their clients returns year after year, as well as determining what skills these advisers displayed, Kahneman ultimately found a near 0 correlation between the work of the advisors and the return on their clients investments. CEOs, in a similar vein, have a near 0 correlation with the performance of the companies they work at.
This is just a bit of buttressing to highlight possible problems with valuing productivity over equality. I’m not anti-capitalist any more than I am anti-communist. As Martin Luther King Jr says, “Both represent a partial truth”. If we combine capitalist and socialist ideas, perhaps we can find the best economic model for everybody. And most countries actually do this, including the USA. Generally, it’s called a ‘mixed-market economy’ or just ‘mixed economy’. People are very quick to declare that the USA is a capitalist economy- it’s totally not. It has a taxation system, subsidies, a federal bank, a medicare programme, and many more policies that are strikingly socialist in nature. It’s closer to being a Capitalist country than say, most Scandavian countries, and there are Conservatives and Libertarians trying to crack down on some of these policies. But for the most part, people are in favour of these largely socialist policies. What the US tax policy today is attempting to do, and rightly so, is balance the trade-off between productivity and equality.
Coming up with a good policy to do this is one of the greatest challenges for any country to rise up to. You can tax those who produce the most value and wealth, but then, they may end up demotivated, deflated, and ultimately, you end up losing out on many of the gains they once provided. You can leave them with most of their wealth, but then how will you supply resources and opportunities to the underprivileged classes? It’s a problem that has existed since the neolithic era and inception of agricultural societies. Thankfully, that was a long time ago, and now AI is here to show us how it’s done.
In the experiment I discussed in the beginning of this post, there were two sets of AI agents. One was the citizens, gathering resources and trading, and the other created tax policies that governed how these resources were ultimately redistributed. The researchers created simulations for how the society functioned under three different kinds of philosophies: the free-market policy, the US federal policy and the Saez formula. It also allowed the AI model to come up with its own tax policy, in an attempt to maximize on both productivity and equality. And the result was astounding: it came up with a policy that had relatively high levels of production while retaining high levels of equality, compared to any of the other tax policies tested in the simulation. The agent was able to make a formula that outperformed any of the major approaches that have been dominant in economic theory for many decades.
People are scared and skeptical to let AI make decisions that shape our everyday lives. But at the end of the day, economics is a mathematical and empirical field, which more often than not suffers due to the irrational decision making of human beings. If we know and agree on our objectives (maximizing productivity and equaulity), perhaps it shouldn’t be up to us to figure out the exact method to get there. I for one, think it’s time we take some of the AI’s tax policies to Congress, and test them out on a real-world marketplace.
