Unpaid internships are good, actually

On decoupling hiring from payment


I. Introduction

Two things happened recently. First, I screwed something up at work. And by work, I mean my unpaid internship. It wasn’t the end of the world, but it wasn’t trivial either. I don’t think I’m at liberty to discuss much (it’s incredibly banal, but it is the government) but I basically forgot to filter out observations from a data set that I should have so dozens of corresponding graphs and charts had to be remade by yours truly. I’m glad I caught it, and after about two days of work things seem to be cleaned up.

Second, I listened to the podcast version of Astral Codex Ten’s recent review of How Asia Works. One of the book’s key points is this: protectionism (things like tariffs and import subsidies) is, contrary to traditional economic wisdom (or dogma, if you prefer),important for economic development because it allows first-time entrepreneurs to put together shitty businesses while still making some money. More importantly, it allows those shitty businesses to get less shitty over time until that Salvadorian smartphone company can compete on a level playing field with Samsung and Apple.

So why is the Econ 101 “free trade is good” logic incorrect? Because it doesn’t account for the positive externalities associated with learning. Firms learn just like people do, and the benefit of firms’ learning how to make cars and pharmaceuticals and stuff is captured largely by domestic employees and consumers rather than by the business owners themselves. And, like any positive externality, this too must be subsidized.

How Aaron is like a Haitian pharmaceutical company

While hearing this, I couldn’t help but think that I, an intern who screwed up the data, am a lot like a Haitian1 pharmaceutical company that keeps mixing up the Prozac and the Percocet. That is, both of us need a little subsidization to become competent.

In the case of HaitiPharma, subsidization comes in the form of tariffs raising the price of American/European/Indian pharmaceuticals and perhaps direct subsidies from the Haitian government (hey, this is a hypothetical). In the case of me, it comes in the form of my parents giving me food and shelter while I give away my labor for free.

This is where the analogy breaks down, though. People dislike economic protectionism for a variety of reasons, but mostly because it reduces Econ 101 “gains from trade” and helps domestic producers at the (short term) expense of everyone else. People dislike unpaid internships because they’re exploitative and perpetuate inequality; those like me with generous, affluent parents get to put fancy knowledge work action verbs (“analyze,” “generate”) on their resumes, whereas the only things my less-privileged peers get to analyze and generate are rude customers and Macchiato lattes with extra whipped cream.

Nonetheless, a fundamental similarity remains; both unpaid interns and uncompetitive developing world companies need to be subsidized because they can’t yet compete on their own.

Forcing the analogy

One fundamental issue with unpaid internships is that they disadvantage even those who don’t take one. Implicitly, my ability to take such an internship puts those who cannot at a disadvantage. In the case of international trade, this would be analogous to only some developing world companies getting subsidized. If HaitiPharma is getting tax breaks left and right, they can undercut the prices of Port-Au-Prince BioTech and run them out of business. Obviously, this isn’t very fair for PAPBT.

The wrong path

Now, one option to level the playing field—which seems to be quite popular—is to denounce unpaid internships and call for their abolition. Yesterday, a friend and fellow government intern told me about the Twitter account @payourinterns which—if you can believe it—advocates for paying interns, especially those who work for the government.

From the organization’s website:

Our Mission

Pay Our Interns is the leading organization fighting for an increase in the amount of paid internships across all work sectors. 

We believe a person’s socioeconomic status should not be a barrier to getting real-world work experience. Pay Our Interns is committed to reducing the gap between education and work experience through research, advocacy, and storytelling. 

As written, I wholeheartedly agree! If I could snap my fingers and have all internships become paid, I’d do it in an instant. But…

A few weeks ago, I wrote A progressive-ish case against the minimum wage. My point was quite similar to the one I’m making here: our goal (well, mine anyway) is to help low and middle income people, and raising the minimum wage is a pretty darn bad way of doing this relative to other options such as a negative income tax.

Likewise, our goal is to help less privileged students and young professionals, and demanding that interns be paid is a bad way of doing this. Why? Because just like forbidding employers from paying “starvation wages” at least might result in them eliminating the position instead of increasing pay, forbidding unpaid internships will almost certainly result in some internships being eliminated instead of becoming paid.

The heart of the matter

Underneath the empirical fact that some positions will get eliminated is the fact that interns like myself aren’t often contributing much net value to an organization. This is pretty counterintuitive. I’d like to think that I (generally) produce good work, and I’m no workaholic or savant. So, common sense goes, aren’t my peers and I contributing value to the organizations we work for—value we could use to bargain for a nonzero wage?

I don’t think so. Consider me in particular. It is not because my work sucks (though who am I to say?), but because I’m taking up a lot of very well-paid people’s time. For example, one person I regularly have meetings with is a PhD economist who is very busy and (hopefully) pretty well paid. Let’s say he’s getting paid $150k to generate $200k of value for 2000 hours of work per year. That’s $100 of value an hour! Even if I only take up 15 minutes of his time every day, this one tiny meeting alone is effectively costing the organization $25!

My story in particular isn’t important here, but I think the anecdote is illustrative; interns might be contributing a lot of value, but we’re the vocational equivalent of human infants—constantly in need of direction (if not assistance) from our elders. In an efficient market, interns wouldn’t get hired if their net contribution falls below zero. I have no doubt, then, that a lot of unpaid interns who are very smart and hardworking are contributing only a little bit of net value to their employer. If they had to get paid anything more than a pittance, the calculus would change and they wouldn’t get hired.

I’d like to emphasize that this isn’t any more an indictment of interns than it is of infants. Even the hyper-competent, Ivy-minted, ex-idealist McKinsey intern is going to have to learn the ropes—and might even defect over to Deloitte the following year. Despite what his academic accolades would suggest, Mr. Perfect is absorbing a lot of McKinsey time in the course of his diligent work.

A contrarian proposal

So, I propose a different rallying cry. Instead of forbidding unpaid internships, I say we have universal access to unpaid internships! Before you rage-tweet this (though honestly I wouldn’t mind the exposure) hear me out…

I mean that we should decouple intern hiring from intern payment. A basic program would go something like this: all Americans aged 18-24 get a stipend of $25,000 annually, no strings attached. Some might choose to seek paid work anyway to supplement their income, but everyone would have the opportunity to give away their labor for free, whether to a McDonalds or Human Rights Watch. In other words, a UBI for young adults.

We can haggle over the precise ages of eligibility and pay-fors. Some people will quibble that the program means giving money to rich kids, but I tend to fall in the camp that says that programs should be universal and made progressive through the tax code (i.e. couple the law with more progressive income tax code to compensate).

The economics of it all

For some reason, society has decided that from ages 0-18 people should generally not have to work so they can pursue an education. In neoliberal shill language, “education” can be read “human capital accumulation.” And, we all recognize, many of those who can afford it will spend another 2-9 years in college and grad school.

The implicit economic reasoning goes something like this: education gives people human capital and/or signals competence, which are important for the economic wellbeing of individual and society alike. Therefore, we will ban child labor and make sure that kids can at least get a high school diploma if they want to and are capable.

And then, all of a sudden, we decide that people have to sell their labor in the labor market with no sort of assistance.

As this chart shows, though, human capital accumulation doesn’t stop in one’s 20s. It slows down after that, sure, but eyeballing the line suggests that 50 year olds make about $12,000 more per year than 30 year olds, the age at which almost everyone is done with formal education.

Some of this human capital might come from general maturity/life experience, but I’m pretty certain work experience itself is the main thing that allows people to negotiate for higher wages. Part of the compensation given to working young adults, then, comes not in the form of income or benefits or stock options but training and education, which promise to elevate their future bargaining power.

It’s no great insight that young adults often take jobs to pad their resumes, gain experience and otherwise accumulate “career capital.” The more interesting corollary to this, though, is that firms often hire young adults because they can get away with paying them less, and they can get away with this because some of the compensation comes in the form of training and experience—just like an internship.

In other words, the difference between unpaid internships and real jobs for people below, say, 40 is a matter of degree, not of kind. In the former case, 100% of compensation comes in the form of experience. In the latter case, somewhere between 1% and 99% does.

So universal-UBI-for-young-adults is, in economic terms, quite similar to normal old education: a way to level-up their human capital and productivity (or signals thereof) before having to produce net value for an organization sufficient to command a decent salary.

Some ambitious speculation

A final bit of speculation I’d like to make is that this internship subsidization program I’m proposing might be more efficient, dollar for dollar, than normal public education. We currently spend about $12,000 per K-12 student per year, and for what? No, education is not just about labor market optimization, but how much intrinsic value do we think society gets out of each student’s school year? Remember, a lot of kids hate school. Something tells me it’s way under $12,000 a year per kid.

So, I think, much of this $12,000 are effectively wasted or worse. Remember, money spent corresponds to real resources being deployed in the world. Paying a teacher $50,000 to teach means that this smart, socially minded person is not being paid to do something else.

Consider such a UBI program, on the other hand. Say we give Jess $25,000, no strings attached, and she uses it to intern at an cybersecurity firm. In the counterfactual world where she didn’t get that money, Jess would be working at Subway (and driving down the wages of other Subway employees). So, in reality, these $25,000 are buying us transfer of Jess’s labor from Subway to the engineering firm.

With teaching, there’s no guarantee that any actual value is being created. We hope that kids enjoy learning, gain skills, and more, but the burden of proof is on those who assert such value is being created. The engineering firm, on the other hand, is being paid to do work by people who want work done. Unless we think people go around hiring engineering firms by accident, this is pretty strong evidence that the engineers are producing value for society.

A metaphor might be illustrative. Consider two government programs. The first pays poor people to dig holes and fill them back in again, and the second gives poor people Amazon gift cards. Suppose both programs cost $1 billion per year, and a similar amount goes to overhead. Even though both programs are helping poor people at the cost of $1 billion, we can make a good guess that the latter program is better since the first is wasting a ton of actual, real life resources (people’s labor) and the second is not.

I don’t think all of public education is like digging holes and filling them back in, but some of it is. And, in the metaphor, UBI-for-young-adults is a lot like the Amazon gift cards—basically guaranteed not to waste real resources in the same way


Tying it all back to international development, one of the interesting lessons to me is that default, neoliberal, Econ 101 consensus is sometimes pretty misleading. Yes, free trade is generally good, but protectionism has an important role to play. Yes, subsidies are often inefficient, but subsidizing internships might be a good thing.

As Scott Alexander writes,

It really is striking how the countries that did the best were the ones that gave the world establishment the middle finger (unless of course this is cherry-picking and there are lots of big countries that followed IMF advice and did great). To whatever degree this is true, it belongs on the list of science failures that should keep us up at night, alongside all those people saying not to wear masks for COVID. Except that the COVID mistake lasted a few months and killed a 5-6 digit number of people, and the IMF advice lasted decades and kept hundreds of millions in poverty.

Fundamentally, I don’t know what the solution to this is. Is there any set of norms or rules that would have avoided such failures at a reasonable cost, besides just “be better at science”?

My best guess of the lesson is this: we shouldn’t assume that a theory’s loudest proponents espouse the best evidence in its favor (full explanation at this older post). Centralized planning’s biggest proponents might be hardcore leftists, and hardcore leftists don’t always talk in terms of “Pigouvian subsidization of positive externalities” and “firm learning.”

So 1960s economics professors or whoever was developing what Scott calls the “IMF advice” shouldn’t have listened to Troskyists discuss why centralized planning is good for class consciousness or whatever. They should have tried to come up with the best reasons why centralized planning makes sense under a neoclassical framework. If more people had done this, maybe fewer would be living in poverty right now.

In the same way, those of us who have a knee-jerk “unpaid internships are bad!” reaction should interrogate whether cynical Chamber of Commerce shills and naive libertarians are really offering the best arguments in their favor. Admittedly, I don’t think the status quo is fine—quite the opposite. But the approach to rectifying injustice matters, and banning unpaid internships is the wrong one.

So, I’ll repeat, let’s make sure that all young adults have the ability to give away their labor for free—not just those who work hard enough to have been born to the right parents.

Leave a comment


This was written before the assassination of Haiti’s president and my use of Haiti as an example has nothing to do with these current events.