Scott Sumner on behavioural economics in introductory economics

In a recent blog post, Scott Sumner argues against a large role for behavioural economics in introductory economics:

The Atlantic has an article decrying the fact that economists are refusing to give behavioral economics a bigger role in introductory economics courses. I’m going to argue that this oversight is actually appropriate, even if behavioral economics provides many true observations about behavior...

Most people find the key ideas of behavioral economics to be more accessible than classical economic theory. If you tell students that some people have addictive personalities and buy things that are bad for them, they’ll nod their heads.  And it’s certainly not difficult to explain procrastination to college students. Ditto for the claim that investors might be driven by emotion, and that asset prices might soar on waves of “irrational exuberance.”  Thus my first objection to the Atlantic piece is that it focuses too much on the number of pages in a principles textbook that are devoted to behavioral economics.  That’s a misleading metric.  One should spend more time on subjects that need more time, not things that people already believe.

The whole post is worth reading. Although I don't agree with all of it, as I think behavioural economics does have a role to play in introductory economics. However, in my ECONS102 class I use it to illustrate the fragility of the rationality assumption, while pointing out that many of the key intuitions of economics (such as that people respond to incentives, or even the workhorse model of supply and demand) don't require that all decision-makers be acting with pure rationality. At the introductory level, I think it's much more important that students take away some economic intuition than a collection of mostly ad hoc anecdotes, which is essentially what behavioural economics is currently. That point was driven home by this article by Koen Smets (which I blogged about earlier this year).

Sumner argues in his blog post that we should focus on discouraging people from believing in eight popular myths. I don't agree with all of his choices there. For Myth #2 (Imported goods, immigrant labor, and automation all tend to increase the unemployment rate), I'd say it is arguable. For Myth #3 (Most companies have a lot of control over prices.  (I.e. oil companies set prices, not “the market”), it depends what you mean by "a lot". For Myth #7 (Price gouging hurts consumers), Exhibit A is the consumer surplus.

However, one popular myth that should be discouraged is the idea that behavioural economics will (or should) entirely supplant traditional economics. There is space for both, at least until there is a durable core of theory in behavioural economics.

[HT: Marginal Revolution]

Book review: Prediction Machines

Artificial intelligence is fast becoming one of the dominant features of narratives of the future. What does it mean for business, how can businesses take advantage of AI, and what are the risks? These are all important questions that business owners and managers need to get their heads around. So, Prediction Machines - The Simple Economics of Artificial Intelligence, by Ajay Agrawal, Joshua Gans, and Avi Goldfarb, is a well-timed, well-written and important read for business owners and managers, and not just those in 'technology firms'.

The title of the book invokes the art of prediction, which the books defines as:

[p]rediction takes information you have, often called "data", and uses it to generate information you don't have.

Students of economics will immediately recognise and appreciate the underlying message in the book, which is that:

[c]heaper prediction will mean more predictions. This is simple economics: when the cost of something falls, we do more of it.

So, if we (or to be more precise, prediction machines) are doing more predictions, then complementary skills become more valuable. The book highlights the increased value of judgment, which is "the skill used to determine a payoff, utility, reward, or profit".

The book does an excellent job of showing how AI can be embedded within and contribute to improved decision-making through better prediction. If you want to know how AI is already being used in business, and will likely be used in the future, then this book is a good place to start.

However, there were a couple of aspects where I was disappointed. I really enjoyed Cathy O'Neil's book Weapons of Math Destruction (which I reviewed last year), so it would have been nice if this book had engaged more with O'Neil's important critique. Chapter 18 did touch on it, but I was left wanting more:

A challenge with AI is that such unintentional discrimination can happen without anyone in the organization noticing. Predictions generated by deep learning and many other AI technologies appear to be created from a black box. It isn't feasible to look at the algorithm or formula underlying the prediction and identify what causes what. To figure out if AI is discriminating, you have to look at the output. Do men get different results than women? Do Hispanics get different results than others? What about the elderly or the disabled? Do these different results limit their opportunities?

Similarly, judgment is not the only complement that will increase in value. Data is a key input to the prediction machines, and will also increase in value. The book does acknowledge this, but is relatively silent on the idea of data sovereignty. There is an underlying assumption that businesses are the owners of data, and not the consumers or users of products who unwittingly give up valuable data on themselves or their choices. Given the recent furore over the actions of Facebook, some wider consideration of who owns data and how they should be compensated for their sharing of the data (or at least, how businesses should mitigate the risks associated with their reliance on user data) would have been timely.

The book was heavily focused on business, but Chapter 19 did pose some interesting questions with application to AI's role in wider society. These questions do need further consideration, but it was entirely appropriate that this book highlighted them while leaving the substantive answers to some other authors to address. These questions included, "Is this the end of jobs?", "Will inequality get worse?", "Will a few huge companies control everything?", and "Will some countries have an advantage?".

Notwithstanding my two gripes above, the book has an excellent section on risk. I particularly liked this bit on systemic risk (which could be read in conjunction with the book The Butterfly Defect, which I reviewed earlier this year):

If one prediction machine system proves itself particularly useful, then you might apply that system everywhere in your organization or even the world. All cars might adopt whatever prediction machine appears safest. That reduces individual-level risk and increases safety; however, it also expands the chance of a massive failure, whether purposeful or not. If all cars have the same prediction algorithm, an attacker might be able to exploit that algorithm, manipulate the data or model in some way, and have all cars fail at the same time. Just as in agriculture, homogeneity improves results at the individual level at the expense of multiplying the likelihood of system-wide failure.

Overall, this was an excellent book, and surprisingly free of the technical jargon that infest many books on machine learning or AI. That allows the authors to focus on the business and economics of AI, and the result is a very readable introduction to the topic. Recommended!

The leaning tower that is PISA rankings

Many governments are fixated on measurement and rankings. However, as William Bruce Cameron wrote (and which has wrongly been attributed to Albert Einstein), "Not everything that counts can be counted, and not everything that can be counted counts". And even things that can be measured and are important might not be measured in a way that is meaningful.

Let's take as an example the PISA rankings. Every three years, the OECD tests 15-year-old students around the world in reading, maths, science, and in some countries, financial literacy. They then use the results from those tests to create rankings in each subject. Here are New Zealand's 2015 results. In all three subjects (reading, maths, and science), New Zealand ranks better than the OECD average, but shows a decline since 2006. To be more specific, in 2015 New Zealand ranked 10th in reading (down from 5th in 2006), 21st in maths (down from 10th in 2006), and 12th in science (down from 7th in 2006).

How seriously should we take these rankings? It really depends on how seriously the students take the PISA tests. They are low-stakes tests, which means that the students don't gain anything from doing them. And that means that there might be little reason to believe that the results are reflective of actual student learning. Of course, students in New Zealand are not the only students who might not take these tests seriously. New Zealand's ranking would only be adversely affected if students here are more likely not to take the test seriously, or where the students who don't take the test seriously are better students on average than those who don't take the test seriously in other countries.

So, are New Zealand students less serious about PISA than students in other countries? In a recent NBER Working Paper (ungated version here), Pelin Akyol (Bilkent University, Turkey), Kala Krishna and Jinwen Wang (both Penn State) crunch the numbers for us. They identify non-serious students as those where there were several questions left unanswered (by skipping them or not finishing the test) despite time remaining, or where students spent too little time on several questions (relative to their peers). They found that:

[t]he math score of the student is negatively correlated with the probability of skipping and the probability of spending too little time. Female students... are less likely to skip or to spend too little time. Ambitious students are less likely to skip and more likely to spend too little time... students from richer countries are more likely to skip and spent too little time, though the shape is that of an inverted U with a turning point at about $43,000 for per capita GDP.

They then adjust for these non-serious students, by imputing the number of correct answers they would have gotten had they taken the test seriously. You can see the more complete results in the paper. Focusing on New Zealand, our ranking would increase from 17th to 13th if all students in all countries took the test seriously, which suggests to me that the low-stakes PISA test is under-estimating New Zealand students' results. We need to be more careful about how we interpret these international education rankings based on low-stakes tests.

[HT: Eric Crampton at Offsetting Behaviour, back in August; also Marginal Revolution]

The beauty premium in the LPGA

Daniel Hamermesh's 2011 book Beauty Pays: Why Attractive People are More Successful (which I reviewed here) makes the case that more attractive people earn more (or alternatively, that less attractive people earn less). However, the mechanism that drives this beauty premium (or even its existence) is still open to debate. It could arise because of discrimination - perhaps employers prefer working with more attractive people, or perhaps customers prefer to deal with more attractive workers. Alternatively, perhaps more attractive workers are more productive - for example, maybe they are able to sell more products.

However, working out whether either of these two effects, or some combination of both, is driving the beauty premium is very tricky. A 2014 article by Seung Chan Anh and Young Hoon Lee (both Sogang University in Korea), published in the journal Contemporary Economic Policy (sorry I don't see an ungated version) provides some evidence on the second effect. Anh and Lee use data from the Ladies Professional Golf Association (LPGA), specifically data from 132 players who played in at least one of the four majors between 1992 and 2010. They argue that:

Physically attractive athletes are rewarded more than unattractive athletes for one unit of effort. Being rewarded more, physically attractive athletes devote more effort to improving their productivity. Consequently they become more productive than less attractive athletes with comparable natural athletic talents.

In other words, more attractive golfers have an incentive to work harder on improving, because they can leverage their success through higher earnings in terms of endorsements, etc. However, Anh and Lee focus their analysis on tournament earnings, which reflect the golfers' productivity. They find that:

...average performances of attractive players are better than those of average looking players with the same levels of experience and natural talent. As a consequence, attractive players earn higher prize money.

However, in order to get to those results they have to torture the data quite severely, applying spline functions to allow them to estimate the effects for those above the median level of attractiveness. The main effect of beauty in their vanilla analysis is statistically insignificant. When you have to resort to fairly extravagant methods to extract a particular result, and you don't provide some sense of robustness by showing that your results don't arise solely as a result of your choice of method, they will always be a little questionable.

So, the take-away from this paper is that more attractive golfers might work harder and be more productive. Just like porn actresses.