AI and the research production function

I've been a bit quiet on the blog lately, due to travelling, and attending the North American Regional Science Conference in San Diego. Regional science is a multidisciplinary field that overlaps economics, geography, planning, and many other disciplines. There were many interesting sessions at the conference, although apparently not the session that I was presenting in, which had an audience of two (one of which was another speaker in the session, with the other three speakers in my session no-showing).

Anyway, one of the most interesting sessions was the day before the conference proper started, on "The Potential of AI for Regional Science", hosted by The Regional Science Academy. Speakers included many of the contemporary great minds in regional science such as Rick Church (University of California, Santa Barbara), Tomaz Dentinho (University of Azores), Peter Nijkamp (Open University), and John Östh (Oslo Metropolitan University), among others. They noted (with examples) the great opportunities that artificial intelligence offers, especially as an aid for research. For example, Östh demonstrated a really interesting application of AI to generating data on neighbourhoods, while Patricio Aroca (Universidad Andres Bello) showed how AI could help researchers whose first language is not English to improve their chances of publication in top-ranked English language journals.

I came away from this session in equal parts excited and concerned (which might be an appropriate mix of reactions to any sufficiently disruptive technology). There is great potential in using AI as a research tool, and I've already seen many examples (and blogged about some ideas here and here). However, my concern comes from how AI will change the research production function.

Before we get that far, let's go back to an earlier technological revolution that greatly changed how research in economics was conducted. This story is not mine, but I forget where I got it from. Consider a simple research production function in economics, with two inputs: (1) econometric modelling; and (2) economic explanations. As computers increased in computational power, the relative price of econometric modelling decreased. So, researchers reallocated their scarce research resources from economic explanations (which had become relatively more expensive as an input) towards econometric modelling (which had become relatively less expensive as an input). The result was a rapid increase in econometric modelling, and a corresponding decline in the quality of economic explanations to go along with the econometric modelling. It is not clear that this was a positive change for the overall quality of economics research (but there certainly was an increase in the quantity of research).

Now consider a similar research production function in economics, where the inputs are: (1) human intuition and explanations; and (2) artificial intelligence. New AI models have made AI radically less expensive to use. We can probably expect research in economics (and in other fields) to adjust to much more use of AI, and much less use of human intuition and explanations. It remains to be seen whether that leads to an increase or a decrease in the quality of research overall. My money is on an overall increase in the quantity of research, but a reduction in the average quality and an increase in the variance (with some very high quality research resulting from AI, as well as a lot of very low quality research).

Unlike the science fiction and fantasy magazine Clarkesworld, which shut off submissions earlier this year due to a flood of AI-generated stories were submitted, I'm not aware of any academic publishers that are feeling the strain, yet. It will be coming though, and may intersect with the increasing volume of predatory publishers (see here and here). As the Managing Editor of a journal myself (the Australasian Journal of Regional Studies), I'm really not looking forward to policing the submission of AI-generated articles of low quality.

Taken altogether, it is clear that there is a trade-off inherent in the impact of AI on research (in regional science, in economics, and in other fields). We will need to accept (and act to mitigate) the negative impacts, while endeavouring to maximise the good.

Jibbitz trading bans offer a missed opportunity to introduce some basic economics to children

I read this article from the New Zealand Herald from earlier this week with some interest:

Jibbitz - accessories that clip on to Crocs - are being banned in schools in Northland due to escalating arguments between youngsters over the sought-after items.

Kamo Primary School principal Sally Wilson was forced to take action after students became upset over Jibbitz trades and some resorted to stealing...

Wilson said attempts to create a safe environment for trades were a learning curve for tamariki and sometimes “ended in tears”.

Often, tamariki trade an item in the hopes of getting it back, and when they realise that isn’t going to happen, they “emotionally can’t cope”, she said.

Eventually, Wilson banned Jibbitz from the school because they had become “disruptive”.

“They were getting stashes and holding on to them, and there was an uneven trade for a certain one that they were after.”

While some kids have their “eye on the prize” and trade cheap Jibbitz for more expensive ones, Wilson said there have also been cases of stealing.

“It’s a learning curve about possessions.”

My children are well beyond the age of adding accessories to Crocs (or going to school, for that matter), but I can remember past crazes for trading Pokemon or Yu-Gi-Oh cards, where trading for sought-after cards got a bit out of hand. So, I can understand the attractiveness of a ban, to protect vulnerable, younger traders from being taken advantage of by older, more savvy traders, who understand that trades are 'for keeps' and better recognise the real value of what is being traded.

However, I can't help but feel that there is a missed opportunity here. The gains from trade is a cornerstone of economic principles, and can be taught very easily. And Jibbitz offer the opportunity to teach the gains from trade in a way that young children can readily understand. Jibbitz trading also offers an opportunity for young children to understand that there are two sides to every trade, and that trading is a voluntary activity. So, thinking about Jibbitz trading, whenever there is a trade of Jibbitz from one child to another (there are two traders), the trade will only happen if both children agree to the trade (either trader can say "no" to a trade), and each child will only agree to the trade if they think that what they are receiving is better than what they are giving away (there are gains from trade for both traders).

Instead of a ban, putting some simple rules for trading Jibbitz in place could really help. Here's a few. First. all trades are voluntary. Both children have to agree. Forced trades can be cancelled by a teacher. Second, all trades are 'for keeps'. There are no take-backs. As an extra rule perhaps a 'current price list' could be maintained, where the number of common Jibbitz expected to be traded for particularly rare or valuable Jibbitz are recorded. This may help avoid problems of a thin market for rare Jibbitz.

There are also opportunities for young children to better understand demand (some Jibbitz are more sought after than others), relative prices (the most sought-after Jibbitz may be traded for several less-sought-after Jibbitz), and scarcity (the rarest Jibbitz are likely to be the most valuable). All of these economic principles can be taught simply, and without recourse to economic jargon, and would help children to better understand some simple economics.

Of course, then there is this objection to Jibbitz trading:

Dargaville mother Taiāwhio Wati-Kaipo was first annoyed when Jibbitz were recently banned at Dargaville Primary School, worrying for her children’s ability to express their “individuality”.

But Wati-Kaipo soon considered the issue and realised the ban had a “deeper meaning”.

She believed the ownership of Jibbitz is a “social indication” of where someone is “sitting on the financial bracket”.

Wati-Kaipo said the price hike in Crocs themselves has created a “has and has not” situation among students.

“Without the Jibbitz, the Crocs were already speaking volumes about someone’s identity,” she said.

Wilson said the craze created a social comparison, as it was about who had the coolest ones and who had the most.

I know that at least some people really believe that inequality can be addressed by banning markets, but it's not correct. In this case, banning Jibbitz will simply shift the outward expression and indicators of social status to some other margin. There are many ways that social status is conveyed. Why stop at banning Jibbitz? Why not ban Crocs altogether? Or premium school bags? Or premium stationery? Or phone games or apps where players can buy special skins or other in-game virtual merchandise? Or phones entirely? Anyway, I'm getting off topic. The Jibbitz ban is a missed opportunity to help young children to better understand some key economic concepts that will be helpful in their development as economic citizens. A ban isn't necessary, and there are better responses to protect children from exploiting each other in the Jibbitz market.

Who benefits the most from free speech?

Earlier this year, I read a most interesting article by Diana Voerman-Tam, Arthur Grimes (both Victoria University of Wellington), and Nicholas Watson (Motu), published in the Journal of Economic Behavior and Organization (open access, with less technical summaries here and here). We also discussed it in the Waikato Economics Discussion Group earlier this year. The article looks at the economics of free speech. Specifically:

Can we measure the impact of free speech on people’s wellbeing (rather than its impact on economic growth per se), and can we determine which groups value free speech the most? This paper takes an empirical approach to answer these questions. We test whether free speech is valued differently by different groups in society, according to their level of re- sources as proxied by income or education. These relative valuations are analysed both using surveyed stated preferences and using estimates of the realized relationships between individuals’ subjective wellbeing (SWB), their country’s degree of freedom of speech, and individuals’ income or education levels, controlling also for other influences.

Voerman-Tam et al. used individual subjective wellbeing data from the World Values Survey and the Latino Barometer, and looked at the correlation with data on free speech and other human rights drawn from the CIRIGHTS database and the Varieties of Democracy database. They test two hypotheses:

The first reflects a view that free speech is a ‘luxury good’... The second reflects a view that free speech has an ‘empowerment effect’ for people with lower socio-economic status and who are therefore more likely to be marginalized in society...

If free speech is a luxury good, then people with higher income (or education) would have a more positive relationship between free speech and subjective wellbeing. That is, higher income people would have the greatest wellbeing gains from more free speech. On the other hand, if free speech has an empowerment effect, then people with lower income (or education) would have a more positive relationship between free speech and subjective wellbeing. That is, lower income people would have the greatest wellbeing gains from more free speech.

Interestingly, the paper starts with an analysis of stated preferences over the importance of free speech, based on responses to a World Values Survey question that reads:

If you had to choose, which one of the things on this card would you say is most important?: 1. Maintaining order in the nation. 2. Giving people more say in important government decisions. 3. Fighting rising prices. 4. Protecting freedom of speech.

Respondents were then asked which of the four choices was the next most important. Voerman-Tam et al. create two measures of the priority attached to free speech. The first is equal to one only if the survey respondent ranked freedom of speech as the most important (and zero otherwise). The second is equal to one if the survey respondent ranked freedom of speech either first or second in importance (and zero otherwise). Then, looking at the relationship between these variables and personal characteristics (including income and education), they find that:

A positive gradient is observed across both income and education for each of the free speech prioritization variables.

In other words:

...people with higher incomes or education place higher priority on free speech (relative to other alternatives that they are asked to rank).

I don't think too many people would find it surprising that people with higher income (or education) are more likely to state that free speech is more important to them than people with lower income (or education. This bit was also interesting, and mostly unsurprising:

Several other associations stand out: free speech is prioritized more by people who are young, students, and/or have no children, and by people to the left of the political spectrum. These characteristics are more in keeping with the hypothesis that people who are more marginalized favour free speech.

So, people with higher income (and education) say that free speech is more important. That is suggestive that free speech is a luxury good. However, it is based on stated preferences for free speech. People with low income (or education) likely have much more important things to worry about in their daily lives than whether they are able to exercise free speech.

A better question to ask, then, is who gains the most (in terms of wellbeing) from free speech? In the second part of their analysis, Voerman-Tam et al. find that overall, there is no correlation between free speech and subjective wellbeing (after controlling for other variables). However, that isn't what they were really interested in. When they interact free speech with income (or education), they find that:

...people with lower income (relative to others within their own country) benefit more with free speech than those with higher incomes, especially in countries with full free speech.

The results are similar for education. Overall, these results support the second hypothesis. So, despite the stated preference results suggesting that free speech is a luxury good, it turns out that people with lower income (or education) benefit the most from free speech (in terms of its contribution to subjective wellbeing).

In my view, the combination of these two results is important. People with higher incomes (or education) make up the elites in most societies, often holding positions of political (or if not political, then at least bureaucratic) power. If those groups believe that free speech is important (which according to these results, they do), then they are more likely to argue for more free speech. That, in turn, creates the greatest benefits (in terms of wellbeing) for people with lower incomes (or education), who are likely to be more disenfranchised. These results suggest to me that we might be optimistic for increases in free speech and somewhat of an equalising of wellbeing between the richer and poorer segments of society as a result.

However, before we get carried away, there are some important caveats. This research is based on correlations. It doesn't demonstrate that greater free speech causes increases in subjective wellbeing. We'd want to establish that more definitively. And, as with all research that involves subjective wellbeing data, we must be aware of its limitations and the criticisms it faces (for example, see here and here, but also see this post by Arthur Grimes as well).

Nevertheless, in the meantime high-income people should feel good about continuing to believe that free speech is of high importance.

Antimicrobial resistance vs. climate change

Careful readers of yesterday's post on antimicrobial resistance might wonder whether I also prefer regulation as a solution for climate change. After all, the problems are superficially similar. Both antimicrobial resistance and climate change are listed among WHO's top ten global public health threats, and both involve negative externalities (where one person's actions make others worse off). Both problems will require concerted international action to properly address.

However, there is an important distinction between the two problems, which means that taxes or tradeable permits (such as the Emissions Trading Scheme) are likely to be effective solutions to climate change, but are less effective for antimicrobial resistance. That distinction may have gotten a little bit lost in yesterday's post.

In the case of climate change, all reductions in carbon emissions are good, in terms of reducing future climate change. So, any policy instrument that reduces carbons emissions is moving us towards the 'optimal' level of carbon emissions (which is not zero gross emissions - all of us emit carbon dioxide when we breathe, for example). The question then becomes, how do we reduce those emissions at the lowest cost to society? Taxes and tradeable permits are credible options for reducing emissions at the lowest cost, while regulation is not (a point that Eric Crampton has made many times, such as here).

Unlike carbon emissions, reducing all antibiotic use is not a good thing. Antibiotics are still needed to deal with infections. So, taxes and tradeable permits are not good instruments for reducing the problem of antimicrobial resistance, because we shouldn't want to reduce all antibiotic use, only inappropriate antibiotic use. So, as I noted in yesterday's post, antibiotic use in agriculture is an appropriate target for a tax to reduce use, but taxing all antibiotics used in medical care would likely make us all worse off.

Sometimes, regulation may actually be the best available option. But that sure doesn't mean that regulation is always the best available option.

Read more:

• Antimicrobial resistance, and health care as a negative externality