Creative Destruction
A Nobel paper I can write about!
I went through a few phases in grad school where I thought certain papers had the answer to everything — in particular, to the most critical question: “What should I write my thesis on?”
One of those phases sent me down this rabbit hole on Aghion and Howitt (1992) and the inverted U-shaped relationship between innovation and competition, creative destruction, Schumpeter, and all that.
Recently, Aghion and Howitt won the Nobel — largely because of that paper and research program.
So, we have two factors coming together:
I know the paper well because of the hours I spent thinking and writing about it — generating zero usable output — at many Madison-area coffee shops circa 2014.
I have a blog.
So, I have to write about it to cash in on those sweet Nobel clicks — maybe? We’ll see. In any case, it’s a great paper, and you should know about it.
The paper wrestles with the question of how innovation drives economic growth.
Economic growth is the most important thing in the entire world. And that’s understating it.
Throughout most of human history, economic growth measured by something like GDP per capita was anemic by modern standards. 2% in a year would have been considered insane progress. And it kind of is: 2% growth means the economy doubles every thirty-five years. Not doubles because of inflation — doubles in the sense that we really produce twice as much or stuff that’s twice as valuable. Nowadays, that’s the growth rate of even already very advanced economies like the United States.
Why? What changed? And what causes economic growth in any case? We only have one Earth and whatever resources God and the stars deigned to give us. And ourselves. So how do we produce more per person from the same materials and the same people?
At first, the answer seems obvious. Technological progress. We get better at making things. And, in fact, that answer is correct, more or less. There’s an old model that already won a Nobel Prize — before Aghion and Howitt even wrote their paper (well, maybe it was already a working paper, publication takes time) — by Mr. Solow.
Solow showed, among other things, that, in the long run, all economic growth comes via improved productivity.
But what is productivity?
Well, stuff we can’t measure and don’t know about. Factors of production besides capital and labor.
So, productivity is a great answer but a partial one. Economic growth can’t be driven by technological progress because progress is itself an outcome. It begs the question: why is there technological progress in the first place?
And so, economists started a literature on “endogenous growth theory,” which tried to — you guessed it — endogenize growth, i.e. make growth an outcome of an economy, not an exogenous technological improvement.
Aside: Economists as a species are horribly uninspired when it comes to naming things. Machine learning types would have written about Real Economic Learning or the Progress Function or something that would sell. There’s room for improvement in the Econ marketing department.
Aghion and Howitt (1992) is part of that literature. They propose a new avenue for economic growth: Creative Destruction.
Reading Schumpeter (1942) is the first step toward radicalization:
“This process of Creative Destruction is the essential fact about capitalism. It is what capitalism consists in and what every capitalist concern has got to live in… The problem that is usually being visualized is how capitalism administers existing structures, whereas the relevant problem is how it creates and destroys them. As long as this is not recognized, the investigator does a meaningless job. As soon as it is recognized, his outlook on capitalist practice and its social results changes considerably.” — Joseph Schumpeter (1942)
“The fundamental impulse that sets and keeps the capitalist engine in motion comes from the new consumers’ goods, the new methods of production or transportation, the new markets…. [This process] incessantly revolutionizes the economic structure from within, incessantly destroying the old one, incessantly creating a new one. This process of Creative Destruction is the essential fact about capitalism.” — Joseph Schumpeter (1942)
Model Setting
Aghion and Howitt’s model has three kinds of goods in it:
Labor
A Good People Want To Consume
An intermediate good used to make the consumption good
The world consists of a continuum of people who live forever (nice!), and don’t mind working (no disutility from labor).
There are three kinds of labor folks can do:
Unskilled labor, which can only make the Consumption Good.
Skilled labor, which can do “research” or produce the Intermediate Good.
Specialized labor, which does “research”.
A key simplifying assumption here is that the mass of each of these three labor types is fixed at (M, N, R). Note that this still allows for different research intensities because the Skilled Labor can choose to produce research or intermediate goods.
Because unskilled labor is fixed and supplied inelastically, we can write the production function for the consumption good without including it directly, i.e. y = A F(x), where x is the flow of the intermediate input and A is current productivity.
The flow of intermediate input is determined linearly by the amount of Skilled Labor working on the intermediate good (L): x = L.
Research produces innovations arriving randomly via a Poisson parameter that depends on the amount of labor allocated to work on research, i.e. q x g(n, R) where n is the skilled labor working on research.
This is the key part. When a new innovation arrives, it increases productivity by a constant factor A’ = v A, but the new innovation requires a new type of intermediate input. To be more productive, we need to do different things than we did before.
Property rights are perfect so that the innovative firm can capture the full rewards from the more productive technology, and the lower-cost product immediately monopolizes the industry.
There is only one firm active at any given point in time, and it’s a monopolist — but only until a new firm comes along with a better production process to replace it.
The monopolist, however, is not the only firm. There are other firms, waiting in the wings, deciding how much to invest in research to have a chance of being the firm to produce the next innovation.
The beauty of this model is that it’s trying to capture a very complex problem: how economies evolve with an endogenous innovation and, yet, at any point in time, there’s really only one decision for the economy to figure out: how much skilled labor to allocate to research and how much to allocate to production.
The model had such a great impact exactly because of this setup, I think. It’s an elegant way to model out and think seriously about Creative Destruction (discussed in economics since at least Marx, who viewed the process as eventually leading to the end of capitalism, etc).
Okay, so that’s the setup. What do Aghion and Howitt do with it?
Implications
So, first, we have some direct comparative statics about how much research skilled labor does in equilibrium.
Research increases when interest rates fall (because the marginal benefit of research rises as current consumption can be delayed), as the size of the innovation (v above) increases (again, because the benefit of research increases), and if innovations arrive more rapidly. That more rapid innovation leads to increased research is not obvious because more rapid innovation cycles decrease future profits from monopolizing the industry, so the marginal benefit of innovation falls. However, it turns out, the marginal cost of innovation falls more.
But the most interesting result is how the equilibrium research intensity differs from the socially optimal one.
When the size of innovations (v) are large, firms will underinvest relative to the social optimum because the discount rate of private firms is larger than the discount rate of the social planner, who understands that the innovation will have permanent effects on economic growth. The private firms understand they have only a finite lifespan until they are destroyed via competition.
When the size of innovations is relatively small compared to the size of monopoly power, then firms will overinvest because the rewards to winning the research race are greater and longer-lasting. In other words, less competition drives increased innovation!
The paper has a great slew of other interesting extensions to this core idea — you should check it out!
Conclusion
If you haven’t read it, or if you haven’t read it carefully, the paper and the entire growth research program really, is well worth it. Most of Modern Economics is a lot of statistics, demand models, details. This literature steps back and asks the big questions — which can be a nice break from figuring out how to calculate standard errors
Thanks for reading!
Zach
Connect at: https://linkedin.com/in/zlflynn
Take my Udemy course at: https://www.udemy.com/course/identifying-causal-effects-for-data-scientists/?couponCode=1DD1B90E5C29DD496967