The Case for Compassionate Meritocracy
Technology is increasing inequality. What are we going to do about it?
Author’s note: since publishing, I have received a lot of great feedback on this piece. Some of my views have updated, and I have edited the post to reflect that.
In his recent Substack post, provocatively titled “Meritocracy is bad”, Matt Yglesias makes the case that although America is pretty good at identifying and elevating the “best” people, this is actually a bad thing. The problem with meritocracy, he says, is that smart people still do bad things—take, for example, the case of shrewd private equity firms gutting nursing homes for profit. According to Yglesias, what we need to do is move away from meritocracy entirely. He writes:
What you need to do is actually change the framework — have a society that’s less based on sorting and ranking, and more based on equality.
In his meritocracy book, the Harvard political theorist Michael Sandel suggests that the most exclusive colleges should move away from tournament-style admissions. Instead, he’d like to see them set a minimum competency bar and then accept everyone who clears the threshold. That seems like a fine idea to me. What I’d like to see even more is for Michael Sandel to teach at a community college.
I think Ygelsias and I agree on a lot of things here. We both think that America actually does a pretty good job of sorting smart people into high-paying positions, although it could be much better, especially for students from disadvantaged communities. We also agree that massive wealth inequality and smart people doing anti-social things to get ahead are unequivocally bad things.
The problem is that I just don’t buy Yglesias’ solution. Deconstructing meritocracy would require fighting against the arc of history.
There are two big economic forces pushing us towards meritocracy. The first is that technology is leading to increasingly heterogenous productive ability. If John von Neumann was born as a 15th century peasant, he probably would have done about as well for himself as most other 15th century peasants. Maybe he would have figured out a slightly more efficient way to plant his crops and ended up with 10% more wheat than his neighbor. In the best case scenario, he might have joined a monastery and gotten really good at illuminating manuscripts or something. Actual John von Neumann was born in 1903 and managed to revolutionize both computing and mathematics. Today, the leverage granted by existing technology and the instantaneous transfer of ideas over the internet mean that our best and brightest can write code or start companies that add billions of dollars to GDP.
The second trend is the move towards winner-take-all competition. As a result of greater global and regional integration, more and more rewards are filtering towards the very best in each field. In the past, an individual buyer might not even have been aware of firms beyond their local area; these days, it’s easy to pick from among hundreds of products online and have them shipped across the country. Digital products are even less location-bound than physical ones. The best software is capable of spreading across the globe almost instantaneously. Tech companies can give thousands of candidates online programming exams before deciding who to interview. With competition no longer limited by geographic area, the best ideas, products, and employees will ruthlessly outcompete everything else.
The rise of automation also contributes to winner-take-all dynamics. Factories that once employed thousands of people may now achieve higher productivity with only a few dozen. Again, the rewards filter upwards, towards the robotics experts and engineers. It’s not just traditional working class work, either: the results achieved by the text-generating GPT-3 neural network give me a sense that secretarial and administrative jobs might be on the chopping block in the next century. After they go, who knows how long bloggers have left.1
Taken together, increasing heterogeneity in productive ability and widespread winner-take-all dynamics mean that there will be intense competition for the very best positions and less left for everyone else. These are long-term trends that seem very hard to defeat. I’m just not convinced that breaking up companies or closing down Wall Street or opening up Harvard is going to effectively fight forces which are closely tied to technological progress.
Even if we could fight back against meritocracy, perhaps by stopping the concentration of talent at elite schools or making it harder for companies to assess the ability of potential employees, I’m not sure we would want to. Critically, I think the first-order effects of the pro-meritocracy economic trends are good: I would rather live in the world where John von Neumann realizes the full extent of his talents, the best ideas spread quickly, and a few engineers can produce enough cars for a whole city.
So what do we do? Do we just sit and wait while power accumulates upwards until we the masses become serfs in thrall to a select few roboticists and AI researchers? I would rather avoid this. Increasing concentration of wealth generation in a few very-high paying positions is already leading to increasing economic inequality. Widespread poverty prevents lower-income people from effectively competing for top positions; without restraints, one can imagine this vicious cycle leading to an entrenched upper class, a humanitarian crisis, and widespread social unrest.
It seems to me that what we need is compassionate meritocracy, a system where we acknowledge and even encourage the spread of good ideas and productivity of our John von Neumanns but also take strong government action to prevent runaway inequality.
One proposal I’m not particularly salient on long-term is the drive for increasing unionization and the minimum wage. This might well be the best action we can take now, but the ongoing rise of automation makes it pretty easy to foresee a future where Amazon doesn’t need fulfillment center workers at all. I worry that putting all of our hopes in the labor movement risks being left high and dry when Silicon Valley eliminates their need for labor entirely. If we want to keep rising inequality at bay, we must do something more radical.
Universal basic income is one such radical proposal. I won’t say that I’m completely sold yet, but the results of preliminary studies have piqued my interest. Some on the left prefer robust public options for health care, housing, and other essentials. At the moment I am less enthusiastic about this approach, but it’s definitely aiming at the right problem. Perhaps some combination of public option and cash transfer will be the optimal solution. These programs would be expensive to fund; luckily, meritocracy and technology seem to be pretty good at driving GDP growth.
There’s also the issue raised by Yglesias that smart people are still capable of doing bad things, and meritocracy rewards ethical and unethical ideas equally. I can imagine an argument that the U.S. needs some sort of secular moral revolution, especially in light of the void left by declining religious participation. I’m longer than Matt on the prospects of regulations to solve problems like private equity takeovers of nursing homes—for example, you could require nursing homes to be more up-front about patient outcomes with potential customers (this sounds like it would be a good idea independent of how evil and/or smart Wall Street is). I don’t think either of these solutions are mutually exclusive with meritocracy. In general, it seems to me that incentive alignment through some combination of regulation and informal social pressure are a better solution to this problem than trying to make radical economic changes.
All told, I agree with Ygelsias about the risk of increasing inequality but don’t think dismantling American meritocracy is a feasible solution. Instead we should embrace a compassionate meritocratic system that takes direct action to address its own deficiencies and ensure all citizens can thrive.
In fact, GPT-3 wrote parts of several paragraphs in this post.