Bryan Caplan  

Greg Clark at the Movies

Cancer Screening Costs and Ben... Bryan's Challenge...
Over at Cato Unbound, I gave Greg Clark a challenge:
Name the most credible measure of idea production that isn't at least moderately positively correlated with population.
In his reply, he pointed to Iceland:
The 300,000 people of Iceland produced 70 films (features, documentaries etc) between 2000 and 2010, in Icelandic! I recently saw one, Jar City, which was very well done.
I'm not satisfied.  My reply to Greg's reply:
Greg wants to know how I would account for Iceland. The obvious answer is that Iceland is an extreme outlier. My challenge for Greg was, "Name the most credible measure of idea production that isn't at least moderately positively correlated with population" - not "Name a single counter-example."

Greg says that Iceland's 318,000 inhabitants produced 70 films between 2000 and 2010. It's great to be such an over-achiever. But what about the overall correlation between annual movie production and population? Using IMDB's numbers for 2003's top fifty film-producing countries, I calculate movie-population correlation of +.67. If you drop underperforming mainland China from the sample, the correlation jumps to +.88. (Raw numbers here, Microsoft Excel sheet.)

Question: If you were Greg, how would you respond to my challenge?  What is the most credible measure of idea production that isn't at least moderately positively correlated with population?

Comments and Sharing

COMMENTS (20 to date)
Steve writes:

Patent data by state may be a goldmine of information. Perhaps there are trends over time. It would be interesting to run a regression with a dependent variable of patents produced in a state in a year with independent variables including a state's population that year, its per capita income that year, other demographics such as education, etc. If you could find a variable that has a higher regression coefficient than population then the problem is solved (if not, I suppose you could always say that it's out there waiting to be found). Or, how would you interpret a variable with a lower regression coefficient by a higher t-statistic?

david (not henderson) writes:

Nobel prizes?

ajb writes:

What about Murray's Human Achievement indices? If there is a correlation, it's gotta be very weak and would suggest how nonlinear the relationship is between population and ideas. Also the idea leaders in the last few hundred years weren't the most populous since China and India did little during the period of rapid econ growth. Moreover, you can't weight all time periods equally since MOST of economic growth has come in the last three centuries.

Steve Roth writes:

I really don't understand this discussion. Are you suggesting that large countries (by population) produce more ideas (pretty obvious, though one would not expect perfect correlation), or more ideas per capita (which would not be obvious)?

There's no indication here that you're correlating ideas per capita to population *density* (which would be interesting), but...?

Eric Falkenstein writes:

Since population has grown steadily since 1750, I think you have a large econometric problem. I could just as well explain the increase in ideas on the decrease in the number of pirates.

But looking in Europe, how about the Black Death and the Renaissance. Lower population, increased wages, the building blocks for our current prosperity. The transitory population boom from 1000 to 1250 created very little.

Evan writes:


Also the idea leaders in the last few hundred years weren't the most populous since China and India did little during the period of rapid econ growth.
You can't forget though that those were both extremely socialist until recently, and still are to some extent. The USA was the third most populous, and most productive nation for most of the 20th century.

I am stumped. I cannot think of any idea that is not correlated with population.

@Steve Roth

or more ideas per capita (which would not be obvious)?

The reason why it's likely that increased population density would increase the amount of ideas per capita is simply that it would be easier for people who specialize in thinking about an idea to meet and expose each other to criticism. If I live in a single village and think up ways to make bread, I would probably be less productive than if I was able to meet people from other villages and learned what their techniques were and listen to their breadmaking advice.

I do think the Internet will probably make it harder to measure ideas production by country, since it's now easy to talk about ideas with people on the other side of the planet. That might make measuring this harder for Bryan.

Steve Roth writes:

@ajb: so you're saying that higher population density makes for more ideas per capita. Makes all kinds of sense, and there's empirical evidence (and research) to demonstrate it. (Though I'm not very familiar with it.)

I'm still not clear what Bryan's suggesting, unless it's kind of obvious.

Various writes:

I'm not exactly sure what you are asking. Are you asking what the most significant causes of idea generation are? Or, by "measure", do you mean an outcome of idea generation? For now, I'm just going to assume you mean what variable, other than population, has the highest positive correlation with idea generation, irrespective of whether this variable is a cause, an outcome or neither.

If that is what you are asking, my votes are:
1. GDP
2. Property rights
3. Average life expectancy
4. Quality of higher educational institutions

frankcross writes:

The best counter example would probably be Finland, but . . .
there's so much noise here, it would be hard to prove empirically. Better education, more money, more R&D spending, etc. are all factors. But just as a matter of theory, how could you deny this? More people will produce more ideas. Interacting with more other different people of choice surely will further more ideas. Not to mention that more people create a bigger market for those ideas, increasing the incentive.

Various writes:

I apologize. I did not read your question properly. You said what factor that ISN'T at least mildly positively correlated with population. My vote would therefore be property rights.

David O writes:

Nobels in Literature are a strong candidate.

But it depends on how you weight these things, surely. And we need to be mindful of results that are limited to one time period. China and India had 42% of the world's population in 1985, but had very poor technological progress credentials (find me an index). Now that's changing, and historically it wouldn't have been the case.

If you weight by country, then we're evaluating whether or not individual countries can improve their idea generation by increasing their population (probably yes). If you weight by population, you're evaluating the impact of national population on idea creation (very possibly negative).

Charlie writes:

Having a top 100 restaurant seems to fit the bill. If you separate Hong Kong from China (which seems reasonable) China, India and Indonesia are way under represented and Singapore and Hong Kong are way underrepresented. I didn't run a regression though.

rovesciato writes:

I assume this is related to the discussion you just had on EconTalk a couple weeks ago, which rather got my mind turning and even led me to download some articles for Julian Simon; so you got someone who had never heard of Julian Simon to download and read at least one article (paradoxically, growth may end war) which is in and of itself a success story.

This would seem to be an exception proves the rule kind of question; i'm sure there are way more ideas coming out of India today than when there were just 700,000 people there. The rate of ideas per person would vary between cultures, governments, regions, etc because of local factors, but another hundred thousand in India only would still increase the number of ideas worldwide, unless all that hundred thousand talked like John Cleese in suspenders.

I think the issue with more ideas = more population good is with the positive value attached to the idea of ideas. Communism is certainly an idea that gained a lot of traction as populations were rising in the nineteenth century, millions of people starving to death on account of it. Though it could be argued that communism is in eclipse there is no reason that new communist ideas coupled with new rhetoric ideas, new propaganda ideas, new genetic ideas, new behavioral science ideas,and new means of production ideas could lead to a new resurgence of the cult of personality, leading to all the old problems but potentially on a greater scale since there are now way more people who's subsistence needs are not met by a failure to the same baseline as in the 1950's. And, of course, the idea of the nuclear bomb has the potential to set human advancement back behind that of stone tools.

There is also the biological, or rather, bacterial problem. Genetic mutations in bacteria are analogous to ideas as far as correlation to population is concerned, and as every body is a host to 'illions of bacteria the increase of the former necessarily entail an increase of the latter. This sets up a race between human ideas in biology, sanitation, medicine, etc. and bacterial ideas, and a bacterial idea that eliminates on third of the population cannot be called a positive by any standard relevant to people living today regardless of the inevitable recovery.

Without any numbers, or ability to use them, I would argue that the real relationship between population and ideas (in that they multiply growth or wealth) is that rapid increases in population lead to increasingly rapid increases in ideas. I would argue that this is a creative destructive process however, and that population knockoffs via plague, war, starvation are inherent and necessary to the forward propulsion of ideas.

rovescaito writes:

I would also say that "aesthetic" ideas such as movies, music, literature in the sense of story telling, etc are not really relevant to the concept of an increase of ideas, where an increase in ideas is a justification of population growth. Ideas like movies or songs have a template nature such that it is entirely possible that 100% of new ideas create zero advancement in that there is nothing new under the sun, or even particularly well done, despite the present time enjoyment derived from them. It is like the pleasure (utility) derived from running; it can be redone every day but never improves upon its basic nature (ie. someone can get new running shoes, but no one learns to run without ever touching the ground).

So 70 films in Iceland is conceivably like 70 new McDonald's in Iceland, keeping the increased population just as fat as they were before their new bounty.

Where movies and songs, ect. are particularly well done and have cultural significance there is a chicken and egg quality: did the increased population generate a quality idea or is the idea a reflection of the static population of the time, such that every static population may produce up to a certain number of reflective ideas and after that they are all just copies, or franchises, if you will. I guess this isn't really chicken and egg, but it represents the possibility that aesthetic ideas at least may not work forward the way technological ideas do.

Charlie writes:

It's interesting that the correlation goes all the way down to .15 when you take the twenty countries by population (starting with the fifth highest pop) from Brazil to Poland.

If you take out the top 4, India, U.S., Indonesia and China, the correlation is only .5 for the other 45 countries. It seems more reasonable to do that than taking out China and leave the U.S. and India in.

Mike writes:

It seems to me to be obvious. Population is what academics would call a necessary but not sufficient condition for idea production. The untrained mind is, for the most part, not going to be able to contribute substantive, 21st century ideas that rise above the noise.

I would say idea production would best correlate with years of higher education a society invests in their population. That higher education can either be undergraduate and above or graduate and above.

A bi-product of this is you can then control for other factors to see if eliminating some sub-populations from the sample improves the correlation. For example, you could quickly test the implied 800 pound gorilla in the room of Nobels coming from Israel versus Muslim countries to see if cultural and/or non-secular variables partially explain the data.

Troy Camplin writes:

property rights protections? the presence of good institutions -- meaning, in this case, those that promote the production and realization of ideas -- rather than bad ones?

David C writes:

It doesn't really matter because this is double counting. The value of idea production is already a part of GDP growth and wage growth statistics. As Greg Clark mentioned, prior to the 17th century, consumption equaled production for the average individual even after incorporating the value of idea production. Nowadays, production exceeds consumption after incorporating the value of idea production. What's important is whether this is a permanent or tempory state of affairs, and if temporary, for how long.

tenkev writes:

Is everyone being willfully obtuse here?

The obvious answer is average IQ.

Zac Gochenour writes:

IQ isn't a measure of idea production. Part of the point of this exercise is to realize that measures of idea production are more correlated with population than some measure like average IQ.

My suggestion is high-tech entertainment eg video games. The video game industry has now surpassed music movies in revenue (there is some argument about this, suffice to say that video games are a really big thing in the entertainment business). China and India have no video game industry to speak of whatsoever. Almost all video games are developed and published in the US, Japan, Canada, UK, and France. Germany, Sweden, and the Netherlands have a few notable studios. As Greg pointed out, Iceland has a successful one.

Of the countries that make video games at all, number of video games is definitely correlated with population. But it seems this is a cultural product that only comes out of pretty rich countries with higher average IQ. Intellectual property rights may have something to do with this.

Comments for this entry have been closed
Return to top