A common question among aspiring screenwriters is: how much does a screenwriter make?
Every year, the WGAw releases an annual report, listing various statistics about its membership. It provides an interesting way of answering that question.
Acccording to the latest report, only 55% of all WGAw members had any income from film or TV. Among those who did have some income, median income was $106,756. So… if you have a 55% chance of working in a given year, and you get $106,756 that year, your annual expected income is $58,715. And there you have one answer: “A screenwriter earns $58,715 a year.”
Of course, the problem with this calculation is that WGAw statistics only cover writers who are already in the WGAw–and you can’t join the WGAw until you’ve already gotten one or more film or TV writing gigs. If you factor in a couple of years in a low-paying job as an assistant (or a waiter), the average income goes down further.
Furthermore, if you don’t work for long enough, you can lose your active Guild membership–at which point, you vanish from the statistics.
I suspect that if you really factored everything into account, the median lifetime salary of a writer would be about $40-45k per year… which happens to be exactly the same as the median income for all Americans.
In short: screenwriting is a middle-class job. A hugely lucky screenwriter can become a millionaire–just like a hugely lucky office worker can become a stock-option millionaire–but most of us are just happy to be earning a solid, middle-class living doing something we love.
That’s a very nice lesson in the difference between conditional and marginal expectations!
Can you explain the difference for us non-statisticians?
Well, if you let Z be a random variable where Z=1 if you have a writing job and Z=0 if you don’t, then Z could be modelled with a Bernoulli distribution where Pr(Z = 1) is 0.55. Then let X be the salary you make as a writer. We know that the expected value of X given Z = 1 is equal to 106,756—that’s the conditional expectation (conditional on Z = 1). But what is the expected value of X by itself? That is, what is the marginal expectation of X. Well, you need to integrate over the sample space of Z, which in this case is discrete, so it’s a sum. Doing that gives you 106,756 * 0.55 + 0 * (1 – 0.55).