The Ballet Job Market Needs a Market (Re)Design

The pandemic has contributed to many shifts in the dance world, as the community has outspokenly criticized longstanding practices and cultural norms. However, no one has called into question the structure (or, rather, the lack thereof) of the job market. As a now-retired ballet dancer of nearly a decade and a PhD student in economics at Harvard, I know that incorporating basic lessons from economics could change the game for both dancers and directors.

The ballet job market is what an economist would refer to as a “matching market”—you cannot simply choose where to go, but you must also be chosen. What makes the ballet market peculiar is that, unlike most professional athletic markets, directors have vastly different preferences for dancers and they mostly do not (and cannot) compete for hires with salaries. Rather, dancers are first and foremost committed to finding their best artistic fits and are often willing to work for less than their worth.

This phenomenon would not be quite as problematic if dancers and directors were nonetheless matched efficiently. Unfortunately, there are two major failures that plague the current system.

First, although many, but not all, major ballet companies in the U.S. operate under the dancers’ union AGMA, there is virtually no regulation in terms of hiring. Deadlines to hold auditions, renew or cancel contracts are company-specific and are not standardized industry-wide. This is problematic because when streams of dancers are released into the audition market at different times, both companies and dancers can end up with undesirable results.

Here’s an example: A dancer knows she will be let go early in the season and begins a job search right away. Other companies, ones that hold auditions later or whose dancers have several months to return their contracts, do not know how many places they will have available and so they tell her they might—or might not—make her an offer down the road. She may then feel pressured to accept an offer that expires in a matter of weeks from another, less preferred company.

Still, she’s better off than the dancer who’s let go after auditions have already passed. I’ve also known directors to rescind verbal offers to dancers very late in the season, instigating a chain reaction that not only leaves the dancer worse off, but also the other companies that would’ve preferred to hire her but were unable to do so.

This coordination failure is made worse by the fact that many dancers wishing to leave their current jobs typically do not announce their departures until they’ve secured other positions—and for good reason. But directors at saturated companies cannot make additional offers until they know who is definitively leaving. This is what I refer to as “holdup.” There are many favorable “trades” available across companies, but either someone has to first give up their job without a guarantee of another, or a director needs to seek out additional funds in order to start the trading process.

This is where market design comes in. This field, which seeks to find economic engineering solutions to practical problems, has studied similar failures in markets like the medical residency matching program, public school choice and vaccine allocation.

One of the central tenets of good market design is creating market “thickness,” or bringing together as many people as possible at once so the best outcomes can be achieved. By centralizing industry-wide contract renewal dates and audition time-frames, not only would companies avoid coordination failures, but this would also eliminate the unnecessary anxiety that dancers face by not knowing when they might receive an offer and when they should accept one.

This practice has become commonplace in other highly competitive settings in the U.S., such as legal clerkships and PhD programs, which do not require students to accept any offers before a standardized date. It is also the reason why most markets that bring together various traders, including the New York Stock Exchange, open and close at the same time each day.

In other settings, centralized clearinghouses have been enormously effective in eliminating similar market failures. Specifically, what I have in mind is a variant that I’ve designed of the well-known top trading cycles algorithm. It would work something like this: After all company departures have been announced and auditions held, dancers and directors would simply submit their preferences to a centralized algorithm that would quickly determine final assignments based on those preferences. While this may sound radical, variants of top trading cycles have been used in school assignment settings and most notably in the kidney exchange system in the U.S., an innovation for which economist Alvin Roth was awarded the Nobel Prize in 2012.

The assignments resulting from this process are guaranteed to be efficient. The process also encourages all dancers to submit their preferences over companies truthfully, eliminating another major element of anxiety. By streamlining hiring in this way, both parties get their best shots at their most preferred counterparts, while trades across companies at capacity can happen swiftly and without the need for additional expense.

Of course, centralized clearinghouses are most effective when the majority of the market agrees to partake in them. While leaders may fear that this would require them to relinquish some control, they would only make offers to the dancers who they would under the best possible scenario, and the gains they would achieve by thickening and coordinating the market would far outweigh any perceived losses.

As new leaders begin to take the reins at companies around the globe, time will tell whether they will be brave enough to challenge the status quo and reshape the marketplace in a way that truly works for both dancers and directors.