Pages in this blog

Monday, December 17, 2018

Stagnation 1.1: In what way is the drug business like the movie business?

A couple days ago I uploaded a long post about a decline in productivity stretching back into the previous century. It was based on a recent paper by Bloom, Jones, Van Reenen, and Webb [1] that includes a case study of the pharmaceutical industry. Back in 2004 Art De Vany published a study of the movie industry which I reviewed in 2005 on Amazon [2]. On Nov. 9, 2004 De Vany gave a presentation at Harvard Business School, “Hollywood Economics: Dealing with ‘Wild’ Uncertainty in the Movies and Pharmaceuticals”. I didn’t hear the talk, but I managed to download the notes on which he based the talk (which are no longer on the web). The object of this post is to outline the similarities between these two businesses.

Superficially they are very different. The movie industry is marketing entertainment while the drug industry is marketing better health. The drug industry is thick with applied biological and medical science while the movie industry is based on various artistic disciplines, acting, directing, screen writing, cinematography, and so forth. Yet the underlying similarities are profound: many ‘ideas’ are proposed, a few make it into production, of those some break even and then become profitable and among those a very few become blockbusters (and keep the corporate boat afloat). Success is as much a matter of luck as of positive knowledge and deliberate intent. Among De Vany’s wry observations: “It’s not a Gaussian world out there.”

The movie business

In the case of the movies, an idea is just that, an idea for a movie. In it’s simplest form an idea takes the form of a “pitch”, a quick statement that can be delivered in a matter of minutes (e.g. on an elevator ride, hence the term “elevator pitch”). A pitch can be developed into a treatment, a written document of 10 to 30 type-written pages. And then we have a script. A screen writer might produce a script ‘on spec’ without any production prospect in hand or may start by pitching the idea to interested parties. The details of this process vary from case to case and, in any event, are irrelevant.

What’s important is that relatively few ideas are purchased. Moreover, the fact that a script has been purchased doesn’t mean it will go into production. Maybe it will and maybe it won’t. But some of them do and a fair percentage of those in turn actually get made and distributed.

This is where De Vany’s research kicks in. Of those movies that are produced, only a minority break-even or become profitable, much less become blockbusters. Can we predict which ones? Well, if we could, then those are the only ones that would get made. We can’t.

De Vany took a 10-year run of box office statistics from late in the previous century and asked whether we could predict success based on knowing the producer, director, or actors involved with the project. No we can’t, nor does it help to know the opening weekend box office. The deep and ineradicable condition of the business is that there is no reliable way to estimate the market appeal of a movie short of putting it on screens across the country and seeing if people come to watch.

The drug business

The market for drugs isn’t as fickle as that for films, though surprises do happen. Here the problem is simply finding a drug that’s worth the expense of clinical trials, where success is not guaranteed. Lots of compounds must be tested, with most being discarded early in the process. I queried my friend Rich Fritzson about this. He worked in pharmaceutical IT for a number of years ending about a decade ago, so his knowledge is a bit old. With that qualification, this is what he told me:
So, on the biology side you hypothesize that you want to have an impact on a biological pathway, which means affecting a protein-protein interaction. You want to find a chemical compound which will interfere with that interaction and you have a library of a million compounds and you don't know which ones are good candidates.

You devise a simple test which can be performed in a very small amount of liquid, contained in a small indentation in a pain of glass, known as a well, and detected visually, say by measuring light reflected from or refracted by this small amount of liquid after you've run the experiment. (This is done by adding the compound to the protein mixture or more likely, by adding the second protein to the well which already contains one protein and the target compound).

To test a million compounds, you have e.g. a thousand different wells in a sheet of glass, and a thousand sheets of glass, and you have machinery to automate the whole process.

Then you start analyzing the results. [...] The compounds you identify are rarely “good enough” to be a drug. They spend months or years tweaking the chemical structure of the compound trying to get a stronger effect.
And then, when it’s all tweaked up, we go for those expensive clinical trials.

De Vany indicates in his notes for the Harvard presentation: “A lot of success is just luck, e.g. antibiotics and ‘cross-over’ drugs for heart disease.”

Looking for a needle in a haystack

In both cases we’re looking for the proverbial needle in a haystack. In the movie business the haystack consists of a population of movie goes whose preferences are only partially known and can change on a dime. A needle is a film that satisfies an imaginative craving in that audience. This business is about human culture.

In the drug business the haystack consists of the molecular pathways in a population of people. A needle is a drug that interacts with a pathway in a way that promotes a person’s good health. This business is about human biology.

Both biology and culture involve extremely complex objects embedded in complex and unpredictable processes. Small differences can have major consequences.

What about the productivity of the movie business?

I don’t know. De Vany didn’t address that issue. He was only interested in the statistical properties of the uncertainty driving the business.

But that’s not quite what I have in mind. What’s the issue? In looking at the pharmaceutical industry Bloom, Jones, Van Reenen, and Webb were interested in the productivity of R&D. The defined productivity in terms of lowered mortality from disease (cancer and heart disease). That’s what drugs are supposed to do, allow you to live longer, right? The measured input in terms of volume of publications.

What would the appropriate analogs be for the movie industry? I’m thinking that productivity would be measured, not in terms of titles produced, but in terms of the variety of niches or taste profiles served. That’s what all this concern over appropriation and Asian actors playing Asian roles is about. It’s also what independent studios and Netflix & Amazon getting into production is about, not to mention YouTube, and so forth. What’s important is not the number of distinct title by the diversity of them. How would you measure that?

What about inputs to movie-industry R&D? Is there even such a thing? I suppose one might say that all those pitches, treatments, and scripts that never get made into films, that’s R&D. But that seems a bit artificial to me. Without having thought much about it I’m inclined to think that the movie business all R&D up to the point that prints are made and distributed to theaters.

But then that’s not so different from the drug business, is it? The cost of coming up with a viable drug is high while the production costs are marginal. It’s the same with movies. Getting that first print is costly. Running off duplicates is cheap.

References

[1] Nicholas Bloom, Charles I. Jones, John Van Reenen, and Michael Webb, Are Ideas Getting Harder to Find? March 5, 2018, https://web.stanford.edu/~chadj/IdeaPF.pdf.
.
[2] The Amazon review is here, https://tinyurl.com/y7ys9egm. I’ve republished the review on New Savanna, Chaos in the Movie Biz: A Review of Hollywood Economics, https://new-savanna.blogspot.com/2012/05/chaos-in-movie-biz-review-of-hollywood.html.

No comments:

Post a Comment