Wednesday, 30 September 2015


George Borjas has a new paper on the Mariel boatlift (an exogenous shock that allowed many Cubans to migrate to Miami). This looks at a famous paper by David Card, and reverses its conclusions. But note, it also admits that the econometrics are tricky:
The analysis also documents the sensitivity of the estimated wage impact to the choice of a placebo. The measured impact is much smaller when the placebo consists of cities where pre-Mariel employment growth was weak relative to Miami.

A novelist's take on the demand for bad social science.

Kenan Malik on multiculturalism - very thoughtprovoking and particularly interesting is the idea that class in the 19th century was in some way close to race today. (Remember George Orwell:

Here you come to the real secret of class distinctions in the West — the real reason why a European of bourgeois upbringing, even when he calls himself a Communist, cannot without a hard effort think of a working man as his equal. It is summed up in four frightful words which people nowadays are chary of uttering, but which were bandied about quite freely in my childhood. The words were: The lower classes smell.
That was what we were taught — the lower classes smell. And here, obviously, you are at an impassable barrier.

Glossy science versus messy science

TL;DR: the editor said a bad thing, and now I am sad about incentives for researchers.

 Just got a desk reject for my honesty paper. The journal editor's comments were brief, but very thoughtful and useful, and I'd much rather have a quick decision than wait 3 months for the same one.

There was one point I disagreed with. The editor commented that my two measures of honesty didn't correlate highly at country level, and this left the reader unsure of how to interpret the results.

Well, sure. Here below is honesty measured by a coin flip. (You ask people to flip a coin and report the result, and you give them a reward for getting heads. If everyone was honest you'd see 50% reporting heads; the proportion actually reporting heads gives you an estimate of the proportion of honesty in the group – in this case, each of the 15 countries.)
 And here is honesty measured by answers to quiz questions. (The questions were so hard that you could only answer, I claim, by cheating and looking up the answers on the internet. See the paper for more details.)

The two measures indeed have some big differences. For example, the Japanese sample goes from really dishonest in the coin flip to really honest in the quiz. (Why? Well, gambling is illegal in Japan. Maybe Japanese people saw the coin flip as already a bit naughty, and felt less inhibited about unethical behaviour in that context. Or maybe [insert your preferred post hoc explanation here].) So, yes, these differences make it harder to interpret the results.

But suppose I only had the coin flip measure – like, ahem, several other papers on this topic. Then my results would be really easy to interpret! Of course, they would also be potentially misleading. No single measure is guaranteed to capture exactly what we want, all have potential confounds like attitudes to gambling, or attitudes to quizzes. That's why we need to use multiple measures and check robustness of our results.

I think we should be careful about the incentives we create. If we only demand data that tells a simple clear story, that is what we will get: glossy science that sells well, instead of messy science, which is the reality.

Maybe what I should have done is published one paper with the coin flip results. Then I could publish a  second paper saying "Ah, but results are different in the quiz experiment!" Ha ha. Of course I would never....

Sunday, 27 September 2015

Re Volkswagen: corporate culture is not enough

Well, this is absolutely a problem of neoliberalism (an idea I basically support, by the way). The whole thing with European regulators handing over emissions tests to the companies themselves, then not bothering to check them? Oosh. Classic abdication to market forces.

It is also absolutely a problem of corporate culture, and I expect to see a lot of talk from VW about culture, principles and values... probably a "value change" programme, et cetera.

The idea of corporate culture is that an organization can have its own inspiring values. This idea has a weakness: what happened at VW was probably not about employees being selfish. It was probably about people doing what they thought would benefit VW. (Mistakenly and short-sightedly, for certain, but that is hindsight.) Now it is possible to create a culture where cheating the company is frowned upon, but it seems hard to create a company culture which will consistently frown upon doing things that benefit the company. For, what incentive do the company and its employees have to maintain that culture?

A second weakness is that employees have already been socialized in ways that are hard to undo. Just as it is hard to learn a new language as an adult, it is also hard for adults to learn that cheating and lying are wrong, if they have not internalized that already. (Almost everyone knows by that age to say that cheating and lying are wrong, but that is not the same.)

So, a better way to improve corporate culture is to start outside the company, in schools and in families, and to ensure that we are bringing young people up right.

Which leads me back to my research on honesty in different countries. Here's the per cent of nationals from 15 countries who reported flipping a coin and getting heads. They were paid $3 to $5 for heads, but nothing for tails. The excess over 50% gives you an estimate of how many people in each country are lying.

GB Britain; ZA South Africa; PorTugal; GReece; CH Switzerland; DenmarK; TuRkey; US; ARgentina; RUssia; BRazil; INdia; K: South Korea; JaPan; ChiNa
Ordinary people in Britain – this is not a student sample, by the way – appear strikingly honest on average. I am not sure the same is true at higher echelons of society, but let's hope so. (No Germany, sorry! But this paper suggests that the Germans too are highly honest.)

Jeremy and Dave

Jeremy Corbyn is planning to apologize for the Iraq war and say that Britain must never use armed force for regime change again.

I used to think that Corbyn was great news for Conservatives. This is the Michael Foot analogy. But Foot was under Thatcher, who went on in the 83-87 period to basically refound the British economy. Now we have Cameron, who will go down in history as a gap between more interesting people. Cameron enrages me the way Tony Blair once enraged the hard Left.

The defence of this government is that (i) it is putting our house in order and (ii) the economy is growing. (i) is false. Austerity stopped in 2012. The government is still in deficit, seven years after the 2008 crisis. The date for breaking even has repeatedly been put back. I doubt that it will ever come: there will be some new headwinds and we will return to living beyond our means. This is a long-term threat to our political system and I see no will or courage to fix it among our politicians.

Point (ii) is true but it has nothing to do with the bold reforms undertaken by our government. What bold reforms? Cameronism means handouts for the middle classes, much of it based on the house speculation economy: free help to buy a house; no inheritance tax if you buy a house; free chunks of the Post Office.... Economic growth is more to do with cheap money finally coming into the economy, and the resulting increase in consumption. That is not a hard trick to pull off.

With a leader like this, you actually want a decent opposition who can hold the government to account. Unfortunately we have Corbyn, who hopes to appeal to the British people by apologizing for the events of a decade ago and promising to tie our hands behind our backs. I marched against the Iraq war, but come on. (Here is Corbyn on the Russian propaganda channel RT, by the way, looking high minded and preaching about US imperialism while not mentioning the Ukraine.)

Saturday, 26 September 2015


The incredible proof machine (nerdy).

Mokyr on technological change, and Brad Delong responds (via Economist's View).

Another take on behavioural economics: "Please, not another bias!"
There aren’t 165 human biases. There are 165 deviations from the wrong model.
This reminds me of something Werner Güth said to me once: "I saw so many papers coming out with stories of inequality, betrayal aversion etc. that I decided to cultivate aversion aversion."

He then added a typically profound and provocative aside, that much later, he had realised that risk aversion was equally problematic. His point, I think, was that the concept of risk aversion, which is absolutely mainstream microeconomics, is itself a sticking plaster over our failure to come up with good descriptions of human decision-making.

Wednesday, 23 September 2015

Psychology from Francis Bacon

It is a point of cunning, to wait upon him with whom you speak, with your eye; as the Jesuits give it in precept: for there be many wise men, that have secret hearts, and transparent countenances....
Another is, that when you have anything to obtain, of present despatch, you entertain and amuse the party, with whom you deal, with some other discourse; that he be not too much awake to make objections. I knew a counsellor and secretary, that never came to Queen Elizabeth of England, with bills to sign, but he would always first put her into some discourse of estate, that she mought the less mind the bills....

I knew one that, when he wrote a letter, he would put that, which was most material, in the postscript, as if it had been a by-matter.

I knew another that, when he came to have speech, he would pass over that, that he intended most; and go forth, and come back again, and speak of it as of a thing, that he had almost forgot....
It is strange how long some men will lie in wait to speak somewhat they desire to say; and how far about they will fetch; and how many other matters they will beat over, to come near it. It is a thing of great patience, but yet of much use. 

Bacon, Of Cunning

Radicalization redux

A while back I wrote:
Politically, governments want social science to support their narratives.... Right now, fear of Islamists means there is a demand for studies of "radicalization". Maybe radicalization is a good object of study – maybe the concept cuts reality at the joints. More likely, it is a concept du jour which adds nothing to the idea of persuasion. 
At a recent workshop preparing for a big government grant application, a distinguished economist explained to me how work in his (thoroughly abstract and theoretical) field could be usefully applied to radicalization.
Noone really understands the true nature of fawning servility until he has seen an academic who has glimpsed the prospect of money.
– Yes, Minister

Tuesday, 22 September 2015

Experimental economics after the demolition job on theory

The front cover of Richard Thaler's autobiography Misbehaving shows a single bird, looking quizzically at a huge flock of birds who are all flying in the same direction. It's a metaphor for behavioural economics. The single bird is the lone behaviourist; the flock is the economic theorists who all think the same way. And indeed, the book is a story of conflicts in which standard economic theory and presuppositions turned out to be wrong.

Economic experiments started as critique. A typical experiment takes real humans and puts them into a situation designed to be as much as possible like an economic model. Payoffs are well-defined. Subjects know all about the experiment in advance. Play is anonymous - so there's no chance to "change the game" by, say, taking your partner out for a pint afterwards. None of this is anything like the real world. But, if economic theory doesn't even get it right in these situations, what hope does it have in the real world?

I happen to think that the critique has been successful. Sensible economists no longer take (e.g.) rational maximization of monetary payoffs, or Bayesian updating, as Gospel which somehow must be true. They accept that other things might happen.

Having performed this negative task, experimentalists have gone further. They want to build their own "behavioural" theories of how people do act. But now there is a problem.

Standard theory is a function from game forms to behaviour. Given such-and-such a set of players, payoffs, information sets and action spaces, Nash equilibrium predicts such and such behaviour from each player. If standard theory is right, then people in a lab experiment which implements a "game" in this way will behave exactly as the theory predicts. They don't, so standard theory is falsified. Fine.

But it does not follow that there is another theory, which also takes games forms as the input, and which will always correctly spit out people's behaviour. Indeed, we know that this does not work. For example, people behave very differently if you call a public goods game "the community game" or "the Wall Street game". But the game-theoretic description is unaltered by this label. There is no correct theory of behaviour which depends only on game forms.

Unfortunately, lab experimenters have blithely continued to write down game forms that correspond to real world situations; implement them in the lab; observe behaviour, perhaps building a model to predict it; and assume that this is what will happen in the real world. There is now a huge accumulation of such experiments. You can see them at any experimental conference.

There is no reason to believe this approach will work. Other things than actions, information and incentives matter. So mimicking actions, information and incentives in the lab does not guarantee people will behave the same way as in the real world.

This whole research programme has to stop. Then we can start designing experiments that will tell us about the real world. In particular, a first step is to think more carefully about the theory that makes any given laboratory experiment informative about the social phenomenon of interest. Such a theory is always needed. Psychology can be helpful here, because it opens the black box of the mind and tries to find how the subject perceives a given situation.

But my main point is: stop the mindless empiricism in economic lab experiments.

Saturday, 12 September 2015

A bit more on ffplot

This week I spent a few evenings and train journeys developing ffplot. The ffplot README gives technical details, here I'll talk about the motivation.

I spend a lot of time plotting my data, either to explore it or to create output for papers and presentations. I'm pretty experienced with basic R plotting, and I also use the excellent ggplot2. But I kept finding myself looking up solutions on Stackoverflow and mailing lists, and thinking "but what I want is so simple! Why can't I understand how to do it?"

ffplot is a simple frontend to ggplot - basically, a hack, born out of my own stupidity. It doesn't do anything new, it just makes it easier for the user to think about.

For example, for my honesty paper, I had people from 15 countries who reported either heads or tails in a coin flip experiment. So, I wanted a bar chart of the proportions of people reporting heads in each country, with some confidence intervals.

In ffplot, this looks like:

ffplot(prop(heads) + ci(heads) ~ country, data)
This gets you:

I like this way of thinking, because what you say is what you get.

Another example: my data includes a test of "citizen integrity", as well as a quiz which respondents could cheat on. I'd like to know if there's a relationship between answers to the test, and scores on the quiz. I'd also like to spot patterns in the data.

So, let's plot integrity score and quiz score, and add a smooth mean. Again, what you say to ffplot corresponds to what you want:
ffplot(quizperf + smooth(mean(quizperf)) ~ integ, data)

Hmm, no obvious linear relationship, but people with very low scores either got 0 or full marks in the quiz - probably because they either cheated, or didn't bother with it at all.

Lastly, let's look at the relationship between reporting heads on the coin flip – which gave respondents a cash reward – and scoring high on hard quiz questions. I hope this is positive, as that will support the idea that both the quiz and the coin flip are tapping the same underlying dimension of willingness to lie. In fact, I want to make sure it's positive in every age group.

So, I'll plot a bar chart of proportions reporting heads, with confidence intervals, split up by age.  Again, the command corresponds to what you want:

ffplot(bar(prop(heads), fill = "red") + ci(heads) ~ quizp.hard | age, data)

This looks pretty reasonable - top scores on the hard questions are generally associated with more reporting of heads.

There's lots still to do, and this is definitely a hack, based on my own ignorance of ggplot2 – if I was smarter I could get it to work more easily. But I like it when things are simple.

Many more examples, and installation instructions, can be found at the github page for ffplot.