Timely Economics Nobel – Creative Destruction and genAI
The recent Economics Nobel Prize focuses on economic and technological growth
Thank you for reading our work! Nominal News is an email newsletter read by over 4,000 readers that focuses on the application of economic research on current issues. Subscribe for free to stay-up-to-date with Nominal News directly in your inbox:
If you would like to support us in reaching our subscriber goal of 10,000 subscribers, please consider sharing this article and pressing the like❤️ button at top or bottom of this article!
On October 13, 2025, the Nobel Prize in economics was awarded to Joel Mokyr1 – “for having identified the prerequisites for sustained growth through technological progress” – and Philippe Aghion and Peter Howitt – “for the theory of sustained growth through creative destruction”.
The work by Aghion and Howitt is very timely, as it may help explain some of the rise in genAI investment spending. Today, we will focus on this research.
Growth
The main motivation behind the work of Aghion and Howitt is economic growth – the increase in the quantity and/or quality of goods and services produced. Commonly mentioned causes of economic growth are institutions (forms of government and laws) and overall knowledge accumulation.
Aghion and Howitt (1992) (“AH”) looked at knowledge accumulation through the lens of technological adoption and obsolescence. Many new technologies often fully replace former technologies – like email replacing the fax machine, or online streaming replacing DVDs.
This type of technological development leads to economic gains – higher efficiency – but the process itself results in certain losers (fax makers and DVD producers no longer exist). The discovery of new technologies that replace old ones is known as “creative destruction”, a term popularized by Joseph Schumpeter:
“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”
Creative Destruction
To model creative destruction, AH propose the following approach.
Suppose the final good – for example, a car – is produced using a worker and a machine (a factory robot). The machine, on the other hand, is produced by a single firm that owns the intellectual property (IP) behind it. This “IP firm” produces this machine using skilled workers. Moreover, this IP firm is a monopoly, as no other firm knows how to produce the machine. This allows the IP firm to earn monopoly profits (i.e. higher profits than a firm in a competitive market).
Lastly, there are ‘research’ firms. These ‘research firms’ aim to develop a machine that’s better than the machine the current IP firm produces. Since this better machine would be preferred by the car manufacturer, the ‘research’ firm that develops this new machine would usurp the monopoly status of the IP firm, becoming the new “IP firm” and earning the monopoly profits.
‘Research’ Firms – the Source of Growth
It shouldn’t be surprising that all the economic growth in this economy will come from the ‘research’ firms that are trying to develop a new machine.
To conduct this research, ‘research’ firms employ the same type of skilled worker as the IP firm. The ‘research’ firms and IP firm are competing for the same type of worker. Moreover, the more skilled workers working at the ‘research’ firm, the higher the chance of discovering a new technology.
Quickly summarizing:
Final good (car) is produced by a worker and a machine;
The machine is produced by an IP firm that is a monopoly and uses skilled workers to make the machine;
‘Research’ firms try to develop a new machine to become the monopoly IP firm. To conduct this research, they employ skilled workers.
Making Profits
From the above structure of the economy (side note: in economics, we call this the ‘model environment’), it can be seen that what motivates ‘research’ firms to develop a new technology is the hope that they will get to be a monopoly IP firm one day and earn monopoly profits.
So how much do these ‘research’ firms invest in research, i.e. in this particular instance, how many skilled workers should they hire?
Choosing Research Level
The amount a ‘research’ firm will choose to invest really depends on how much profit the ‘research’ firm can make in the future as an IP firm, once it discovers the new machine. This future profit, just like the current profits of the current IP firm, depends on two crucial elements:
How much the IP firm must pay skilled workers to produce the machine;
How long the IP firm remains a monopoly, i.e., how long before a new ‘research’ firm discovers a new technology.
It can be seen from these two elements that the more a ‘research’ firm invests in trying to find a new technology, the less it will earn as an IP firm. Why? This is because in this model of the economy, everything is symmetric – any decision made by one ‘research’ firm, will be made by all ‘research’ firms. So the following two things occur:
Firstly, to increase the likelihood of developing a new technology, a ‘research’ firm hires more skilled workers, pushing their wages up. This in turn means that wages for skilled workers are higher, resulting in lower IP firm profits. Once the ‘research’ firm discovers the new machine and becomes the IP firm, all other ‘research’ firms will continue to try to dethrone it. This means skilled worker wages will remain high, keeping monopoly profits lower.
Secondly, as the ‘research’ firm hires more skilled workers to increase the probability of discovering the new machine, it means that the rate of discovery of new machines is permanently higher. But that implies that once you become an IP firm, you will be an IP firm for a shorter time, as other ‘research’ firms now also have a high probability of discovering the new machine. Since you will be a monopoly IP firm for a shorter time, you will earn lower profits.
Both of the above elements imply that as you increase spending on research, your potential profit from this research falls.
Outcomes for the Economy
So what is the rate of growth/innovation in this economy? It depends on the parameters of the economy – things like the number of workers, general likelihood of discovering a new technology, costs of capital.2
AH find that the level of research increases when:
Interest rates are lower (as this increases the value of all future profits one will earn as a monopoly IP firm);
The discoveries are more likely to be more substantive, i.e. the new machine is much more productive in making cars (as this increases future IP firm profits);
There are more skilled workers (as these workers are cheaper, resulting again in higher IP firm profits);
The probability of a discovery is higher, all else constant, i.e. if we keep the number of researchers fixed, the probability of discovery goes up.
The last point – higher discovery probability – is not obvious. This is because with a higher discovery probability, the ‘research’ firm is more likely to make a discovery (increasing profits), but once it becomes a monopoly IP firm, it is also more likely to be usurped by another ‘research’ firm (decreasing profits). AH show that the former effect dominates, resulting in ‘research’ firms conducting more research when discovery probabilities are higher.
Socially Optimal Research Levels
Economists also like to look at what the socially optimal outcome is. A socially optimal outcome is modeled by assuming that there is a ‘social planner’ that can pick and choose what every person does – i.e. the ‘social planner’ chooses how many people work in ‘research’ firms and how many work in the monopoly IP firm. This is in contrast to a ‘market’ world, where each individual chooses to do what is best for them without taking into account how their decision impacts others. Do you think a social planner would want more research, and thus more innovation or less?
The answer to this question is surprisingly unclear. That is because the social planner realizes that a new innovation makes the old technology obsolete, destroying the value of the old technology. To protect the old technology, the social planner may reduce the amount of research compared to a ‘market’ world.3
Moreover, the more research that is conducted, the more expensive it is to produce the machine since both research work and machine production use the same skilled workers. More research work implies higher wages paid to skilled workers, making machine production more costly.
On the other hand, the ‘social planner’ values the future benefits of new technological discoveries more than the ‘research’ firms. That’s because better machines allow for cheaper final goods (cars) in the future for everyone. Thus, the ‘social planner’ may increase the amount of research. Overall, whether the level of research in a ‘market’ world is optimal is unclear in an AH model, as it depends on the parameters.
Application of AH to genAI
The beauty of the AH model (and probably the reason for the Nobel Prize) is that with the relative simplicity of the model, many insights can be had. The AH model shows the importance of benefitting from discoveries (the duration of monopoly status) and how it can impact research levels. Moreover, the tension between how easy it is to discover new technologies and how much research is conducted is an important factor to consider when designing research incentives. Some of these insights appear to be quite pertinent to our current focus on genAI technologies.
As AH mentioned, if the probability of discovery of a new technology increases (point 4), more research is conducted. Some have posited that the probability of new discoveries within a field of research can go up after a particular breakthrough. We may have seen some of this happen with genAI – after the release of ChatGPT in December 2022, investment in genAI research has skyrocketed. Additionally, another implication from AH’s model based on point 4 is that if the probability of discovery changes over time, it is unclear whether this will lead to more research being conducted. In the extreme, if new discoveries are too easy find, no research will be conducted! This is because you will never benefit from your discovery, since your technology will immediately be replaced by a better one.
The simple AH model may reflect the world of genAI quite well – the maker of the best genAI model may become a monopolist, as everyone will only use their technology. Knowing this, many firms have invested significant capital into research hoping to become the genAI monopoly. Once some of these discoveries occur, the amount of investment in genAI research might start to fall, as new discoveries become less likely. This helps explain the meteoric rise in genAI expenditures.
I have not read Joel Mokyr’s work before, but I’m thrilled to see an economic historian win the Economics prize, as it demonstrates (mainly to other economists) that the field of economics is much broader. I am looking forward to reading his work – “A Culture of Growth: The Origins of the Modern Economy”.
In theory, the economy could end up in a ‘no-growth/no-innovation’ world. This can occur if all ‘research’ firms believe that even if they become an IP monopoly firm, they will quickly be usurped by another ‘research’ firm. Thus, no ‘research’ firms undertake any research at all.
The old technology still generates positive value, while research is costly. Replacing old technology very often may be welfare decreasing, since we may end up using a lot of resources just on research, which, on its own, does not generate value.
Have you seen the new paper on AGI by Restrepo? It's more of a thought experiment, but it's interesting.
Thank you, I just learned something! Knowledge is power and you just empowered me!