The problem is the (lack of) points

The Writers Guild of America is on strike because everyone in Hollywood is unhappy. Nobody is making enough money or getting hired for long enough stretches. Complaints appear fairly universal as well, from showrunners and producers all the way down to the newest hire. In fact the only thing that doesn’t appear to be running a shortage is blame.

I know the complexity of trying to collectively bargain such an enormous industry means that any deal struck, even if it starts out relatively efficient, will be a bad deal for someone within 3-5 years. That’s just the nature of things when you’re trying to forecast both technology and the shape of multiple labor markets. It’s an impossible task. That said, I think we can boil down a lot of what has gone to hell to a single problem. Streaming is currently the dominant platform and streaming studios don’t pay points i.e. profit share on the back end.

Of course streaming services aren’t the totality of the television and movie business, but they are sufficiently dominant that the lack of points, and the problems downstream of it, have fundamentally altered both the labor market and the incentives facing the people creating entertainment. Writers, directors, producers, and even some of the bigger actors have routinely accepted a percentage share (i.e. percentage points, get it?) of the profits of an enterprise as a portion of their compensation. Blumhouse films famously got bigger name actors to join their small genre projects for scale (i.e. guild actor minimum wage) by contracting them for a sizable profit share. It’s a great way of mitigating risk for small studios while also attracting actors with the possibility of a big pay day.

Profit sharing is about far more than just mitigating risk, though. They are about aligning incentives. Everyone cares that much more about the quality of the product when their eventual payout is directly connected to success. Aligning incentives is both exceptionally important and difficult in contexts where quality is difficult to assess and forecast at intermediate stages of production. The only thing you need to know about Hollywood is a cliche that I choose to attribute to legendary screenwriter William Goldman who always asserted, with regards to the entertainment business, “Nobody knows anything.”

“Nobody knows anything” means that 1) forecasting success or failure of a project is difficult, but just as importantly, that all of the really important information is tacit. There’s somone who knows what the best bit of dialogue is, the perfect costume, who to cast as the best friend and the villain, how to frame the shot, how to light the third scene, how to make the write kind of fake snow, what kind of bagels to get, etc. Someone knows, but that someone probably couldn’t explain why they know what they know, couldn’t pass that information on, couldn’t identify the underlying factors in audience and historical film data. They just know. And if you want them to take the optimal action, and make the necessary committments and sacrifices to make the film the best it can be, they need to have incentives in place to ensure that they are rewarded for infusing their tacit knowledge within the final product.

Those incentives are currently broken in the film and television business.

As best I can tell, what broke them was that a) streaming is a lot less profitable than Netflix, Disney, et al. hoped, and b) out of a desire to keep internal data private, compensation contracts could no longer include an “outcome” based component i.e. profit sharing.

Streaming services make money from subscriptions. How an individual production contributes to those subscriptions has a relationship to data associated with that production, but streaming companies have thus far been unwilling to release data associated with individual productions. Yes, Netflix started releasing a little bit of data, but nothing compared to previous entertainment paradigms. Ticket sales, nielson ratings, and advertising always allowed for intuitive mechanisms for profit sharing via points and residual checks. In order to keep their data private, firms have opted instead to move to simple compensation structures, including before-after compensation where a writer, producer, or director is paid when a project is started and when it is deemed completed.

Receiving roughly half of your pay when a an amorphous product, like a film script, is “done” is a really terrible idea for the simple reason that the writer has far less incentive to make the script “good” and far more incentive to make it acceptable to person who must stamp it “done” so they can get paid. How do you get that stamp? By doing whatever they tell you to do. The problem is, they don’t know what the script needs to be good, that’s why they aren’t a writer. They don’t have the tacit knowledge. The writer has it, but fighting to infuse that knowledge in the script does nothing to improve their pay, in fact, it only delays it.

These breakdowns in incentives are all over Hollywood and you can feel it in the films and shows coming out. Scripts feel half-assed, as if written by disconnected employees scattered to the winds of Zoom and a California housing crisis. Performers seem unsure of what to do within a menagerie of green screen and tennis balls on sticks. Audio is famously terrible in movies now.

So how do they fix it? Well, I’m not even sure who “they” are one answer might be found in the Blumhouses, A24s, and every troupe of actors and filmmakers who keep working together. Shrink the number of people involved to the point where you solve incentives not with perfect contracts but with longer term relationships. Repeated interactions have always been a great way of aligning incentives. When credit can more naturally flow to smaller number of people, when working together over and over is the best way to create job security, all of the other problems become easier to solve.

What about the revenue problem though? Well, maybe ticket sales never recover, and streaming never prints money the way advertising did, but modern technology also means you can make a show or film with a smaller number of people than ever before, which means every single one of those people could feasibly participate in profits. Maybe the answer is to simply recreate the old model externally to the big studios, at the earliest stages of coordination, recruiting and writing, before distribution or marketing even enter the equation. Smaller teams, direct participation in profits. Smaller movies and shows. Smaller can be better, just ask anyone who’s ever tasted a beer in the last 20 years.

Maybe the WGA (and the directors and actors guilds too, for that matter), should be thinking a little bit less about what they can negotatiate from the studios, and more more about how much it can cut them out of pre-distribution entirely. Vertically integrating the film and television business is not the only possible state of the world. In fact, given the current state of technology and the broader labor market, it’s entirely feasible that one of the main reasons that industry has collapsed into such a small number of firms is that guild contracts grant cost advantages to being enormous, even in the case where much of the companies in question have next to zero knowledge about the creative side of their industry. There may very well come a time where a million micro-studios, teams of writers, actors, directors, and production staff simply work to make their products and then sell ex-post to the highest of the content-starved bidders. There was a time with the Shane Blacks of the world got to enjoy the benefits of bidding wars for their scripts. Maybe the WGA,DGA, and SAG can bring those back. Does anyone have Sotheby’s number?

Or maybe the world collapses into a hellscape of reality-influencer tik tok stars and I wholly divorce myself from entertainment entirely. Bird watchers seem happy. Maybe I could get into bird watching.

Is Your Head in the Arm Hole of Your Dress?

There is something morally instructive about watching a preschooler melt down. It was the morning of my __th birthday yesterday. Kids still had to be dressed and fed and shipped to school on time.

My daughter, who is almost 5, was screaming on the stairs instead of coming to breakfast. Upon inspection, I realized that her head was through the arm hole of the sleeveless dress she had chosen to wear to school. I offered to help her. She screamed louder and lurched away from me. Her pride was more hurt than her neck at the thought of accepting help. She was not yet really wearing the Anna (the character from Frozen) dress because of the snafu of the sleeves. She stomped around screaming for minutes, refusing all offers of help or comfort from me.

Adults do this kind of thing all the time, although it looks different. People do the stupidest things and then dig in instead of accepting help and reversing course.

My daughter is exceptionally brilliant and kind. She is loved by everyone she meets. Even she has these moments, because we all do. That is some behavioral economics for you.

Beware of Scatterplots

Scatterplots are a great investigatory tool. You can scatterplot raw data for two variables and, if the relationship is strong, then you can see the functional form that relates x and y (linear, polynomial, exponential, etc.). However, there are two data characteristics that are a scatterplots Achilles’ heel: large samples and discrete variables. And they create misleading scatterplots for the same reason.

Examine the below scatterplots for y vs the discrete variables x1, x2, & x3 on the interval [0,10]. What do you think slopes or correlations are?

Continue reading

Music Rights Are Surprisingly Cheap and Easy to Buy

When music rights make the news, it’s generally because a superstar’s entire catalog is selling for hundreds of millions of dollars. That may be why I always assumed that buying music rights would be difficult and expensive- that you’d both have to know the right people to even hear about potential deals, and have to be quite rich to afford them.

But this week I found out about Royalty Exchange, a site that currently lists hundreds of music rights for sale. They certainly appear to make the process of finding and buying rights, and collecting royalties, easy (I haven’t bought any yet so can’t say for sure). They currently list songs and partial catalogs from all sorts of artists you’ve heard of:

When I say I find many listings to be surprisingly cheap, I mean this relative to the hundred million dollar deals you hear about. Of those that offer a list price (as opposed to simply asking for offers), the vast majority are over $10,000, and many are over $100,000. Overall I’d put it in the “luxury car” bucket- expensive enough that its a bad idea for a normal middle-class person to buy one, but cheap enough that they could if they really wanted to. It’s a bit of a better idea than a luxury car, since its more investment than consumption. But if I actually bought the Flogging Molly catalog like I want to, I’d be taking an unnecessary risk by putting a large proportion of my net worth in a single investment. Their music is great and I think it will maintain its popularity, but if I’m wrong and people stop listening to it I’d lose out. So, for most people it’s a bad idea in the same way that putting half your retirement account into a single company’s stock is bad idea. But I’m surprised its even possible.

Why are these rights so affordable? Sometimes, of course, its because the artist isn’t that popular. But why are the rights to songs and musicians that are household names affordable? It seems to mainly be because the rights have been sliced and diced so that you’re only buying a small piece of them. Consider Miley Cyrus above. First of all its only the rights to one of her songs (admittedly a hit song). Second, you’re only buying the rights for ten years (lifetime rights are sometimes available but naturally they cost more). Finally, you’re only buying some of the rights, in this case the right to get paid when someone publicly performs the song (but not when someone streams it or buys a copy):

Even given all that though, I’m surprised how cheap the rights are. I expected that people would overpay for them because they like an artist, or for the bragging rights. But the yields seem pretty reasonable, often over 10%. Yields could rise or fall over time as an artist becomes more or less popular, or as the economics of the music industry change, but current prices generally seem justified by the income stream. I look forward to having enough money that this could make sense as an investment for me; I expect I might in 10 or 20 years, but maybe some of you are already there.

The cheapest listing from an artist I’ve heard of, Busta Rhymes (only performance rights, only certain tracks)

“Studies Show”: Marijuana Legalization and Opioid Deaths

In his NY Times column today, Ross Douthat argues that legalizing marijuana is a big mistake. Douthat makes a number of arguments, but let me focus on one point he makes in the column: that recent research suggests legalizing marijuana increases opioid deaths. This point is made in just one sentence of the essay, so let me quote it in full:

There was hope, and some early evidence, that legal pot might substitute for opioid use, but some of the more recent data cuts the other way: A new paper published in the Journal of Health Economics finds that “legal medical marijuana, particularly when available through retail dispensaries, is associated with higher opioid mortality.”

Kudos to Mr. Douthat for actually linking to the paper. That’s what the internet is for! Yet so many writers in traditional news sources fail to do this.

Now, on to the paper itself. There is nothing untrue in what Douthat writes. First, there was plenty of “early evidence” that legalizing marijuana reduced opioid deaths. More on this in a moment. And the study he cites by Mathur and Ruhm is particularly well done. It is published in the top health economics journal. But the main point of the paper is to say “we think the rest of the literature is wrong, and we’re going to try really hard to convince you that we are right.”

What does the rest of this literature say? Here’s a brief tour (all of these are cited in Mathur and Ruhm). The variable in question is opioid deaths.

Continue reading

Save $$$, Easily Change Your Car Cabin Air Filter Yourself

I have done various maintenance and repairs on my cars over the decades. Usually, they turn out to be harder and more time-consuming than I thought. Changing the engine oil and oil filter has become genuinely harder since the oil filters have migrated deep up under the engine, where it is hard to access them without putting the car on a lift, and disposing of a milk jug of used oil has gotten more difficult.  I used to be able to easily change out a light bulb in the headlight, but the last car where that needed doing required you to take apart much of the front end of the car to get at the headlight. However, I recently found that changing the cabin air filters in my two vehicles (van and sedan) is so easy, I wish I had started doing it years ago.

Why Change the Cabin Air Filter?

The cabin air filter filters the air coming into the passenger section of the car. It knocks out road dust and pollen, and other bits of whatever that might get sucked into your air system as you are going down the road. So, it protects your and your family’s lungs as well as the components of the air handling system. Typical recommendations are to change out the filter about once a year or every 15,000-20,000 miles.

The photo below shows the cabin air filter I just pulled out of my van after maybe 2 years and 25,000 miles, next to a relatively clean filter. Obviously, I let this one go a bit too long: it is grey with dust/dirt, and partly blocked with plant debris.

I have not been quick to change out these filters because garages or dealers often charge something like $80-$100 for this. And until recently, I never considered doing it myself, because for some reason I thought it was a hard job. I had read of people having to contort in unnatural positions with heads inserted under dashboards as they disassemble layers of car to get at the filter.

It Is (Often) Super Easy to Change a Cabin Air Filter

It all depends on where the filter is located. For most models of cars, you can find guidelines on line, including YouTube videos. There are some models where you indeed may have to unscrew a cover plate somewhere below the dashboard to expose the filter. But in most cars, you remove the glove box to expose the filter. That may involve undoing come screws or a snap or strut, and squeezing the edge of the glove box inward. For my Hondas, all I had to do was empty the glovebox, (authoritatively) squeeze in the edges, and the glove box pivoted down, and behold, there was the filter in its little holder. Then slide out the holder, pull out the old filter and put in the new filter (purchased at AutoZone for $20 each), slide the holder back in place, and finally tilt the glovebox back up until it snapped in place.

Ten minutes max, easy-peasy. Obviously, this saved money, but it also felt empowering. I highly recommend trying it.

Video of Joy Buchanan on Tech Jobs and Who Will Program

Here are some show notes for a keynote lecture to a general audience in Indiana. This was recorded in April 2023.

Minute Topic
2:00“SMET” vs STEM Education – Does Messaging Matter?  
(Previous blog post on SMET)
5:00Is Computer Programming a “Dirty Job”? Air conditioning, compensating differentials, and the nap pods of Silicon Valley  
(post on the 1958 BLS report)
7:50Wages and employment outlook for computer occupations
10:00Presenting my experimental research paper “Willingness to be Paid: Who Trains for Tech Jobs?” in 23 minutes  

Motivation and Background 10:00 – 15:30
Experimental Design         15:30 – 22:00
Results                    22:00 – 30:00
Discussion                 30:00 – 33:30
33:50Drawbacks to tech jobs  

See also my policy paper published by the CGO on tech jobs and employee satisfaction
35:30The 2022 wave of layoffs in Big Tech and vibing TikTok Product Managers  

I borrowed a graph on Tech-cession from Joey Politano and a blog point from Matt Yglesias, and of course reference the BLS.
39:00Should You Learn to Code? (and the new implications of ChatGPT)  

Ethan Mollick brought this Nature article to my attention. 
Tweet credits to @karpathy and @emollick
48:00Q&A with audience

Video: Joy Presents Two Experimental Papers to a Macro Class

Here are some show notes to a talk I gave in April 2023. I had the opportunity to talk to an undergraduate macroeconomics class at Indiana University East.

Minute Topic
2:00Research on Behavioral economics and Macroeconomics
4:25Labor Market Equilibrium Concepts and Incomplete Labor Contracts
6:50The Gift Exchange Game and the Fair Wage-Effort Theory
13:00Recessions and Downward Wage Rigidity
19:00Presenting my Experimental Study “If Wages Fell During a Recession” in 13 minutes
32:00-33:00How question raised in “If Wages Fell During a Recession” pointed the way to the Reference Point paper
33:00 – 41:00Presenting my Experimental Study “My Reference Point, Not Yours” in 8 minutes
41:00-44:00Conclusion of “My Reference Point, Not Yours” and tying it back to macroeconomics

The “If Wages Fell…” paper directly inspired the “My Reference…” experiment. But I don’t cite “If Wages Fell…” in “My Reference…,” so you would never know how closely they are connected unless you listen to this talk.

Improved University Retention By Selecting for Mission Fit

The president of my university said that he wants the following strategy publicized.

The purpose of an admission application is to find good matches between students and the university. We want the application to be easy for people to complete, but to filter out those with low conscientiousness and those who aren’t a good mission fit. If the application is arbitrarily difficult or convoluted, then we’ll lose great applicants. But, if it’s not costly enough, we’ll attract students who are closer to indifferent about attending. Those are the freshmen who don’t return for their sophomore year.

Continue reading

The End of Erdogan? Forecasting the Turkish Election

Predicting elections is hard. Poll aggregators and prediction markets can help. Many of the usual suspects like FiveThiryEight and PredictIt aren’t covering Sunday’s election in Turkey, partly due to their own issues, and partly because US organizations often ignore foreign elections. But we do have several good predictors to consider, and they all list opposition candidate Kiliçdaroglu as a slight favorite.

Polymarket is most optimistic for the opposition, giving them a 67% chance. British betting site Smarkets gives them a 61% chance. Play-money site Manifold Markets gives them 56%. Finally, no-money prediction site Metaculus gives a 60% chance that the opposition wins, and a 79% chance that Erdogan leaves office if he loses the election. I’m not sure how the count the Swift Centre, a small closed panel of forecasters, but they are the exception in seeing Erdogan as a slight favorite.

My economist’s instinct is to trust the real-money markets more here, although Manifold and Metaculus outperformed them in the 2022 US midterms. The usual bias is to predict a win for the candidate you like more (which for Westerners on these markets means betting against Erdogan), and have real money on the line can help counteract this. On the other hand, some might use betting markets as a hedge and bet on the outcome they don’t want. In this case the betting markets are slightly more favorable to the opposition, but the gap is small.

Of course, the biggest real-money markets are those that don’t ask directly about the election: the markets for Turkish stocks and bonds. These have generally performed well in the past year as the opposition’s chances have risen, which may indicate that markets think a new Prime Minister with more conventional economic views will get inflation under control.