Friday, 30 January 2015

A paragraph to ponder from Schumpeter's review of Keynes

It is, however, vital to renounce communion with any attempt to revive
the Ricardian practice of offering, in the garb of general scientific truth, advice which - whether good or bad - carries meaning only with references to the practical exigencies of the unique historical situation of a given time and country. This sublimates practical issues into scientific ones, divides economists - as in fact we can see already from any discussion about this book - according to lines of political preference, produces popular successes at the moment, and reactions after - witness the fate of Ricardian economics - neither of which have anything to do with science. Economics will never have nor merit any authority until that unholy alliance is dissolved. There is happily some tendency towards such dissolution. But this book throws us back again. Once more, socialists as well as institutionalists are right in judging economic theory as they do.

Is he right? Hat tip to Alexia.

Thursday, 29 January 2015

Grade inflation

The media has noticed the problem of grade inflation in UK degrees (BBC, Telegraph). There are now twice as many first class degrees being given out as there were ten years ago (for non-UK readers, a “first” is the best degree grade).

The economics of this seems clear. A university that gives better grades to its students benefits them in the job market, and also looks better in league tables that count the number of grades students get. It also devalues that university’s degrees, but, since most UK employers cannot distinguish between universities except perhaps Oxbridge at the top, this devaluation is a “public bad” which is shared with all universities... and also with past and future students, neither of whom the short-term-focused administration cares about. Result: grade inflation.

This story is true as far as it goes, but it misses something important. Grades are given by the academic staff who do marking. None of us benefits directly from inflating our students’ grades. The benefit to the university is a public good for each individual academic: why should I care about my university’s score in the rankings?

The real cause of the change is not pure self-interest, but a combination of “bounded ethicality” and its exploitation. I feel loyal to my colleagues in my department, and to my university; whereas UK education as a whole is too abstract and remote to care about. And then, these feelings are played on by the administration. A memo comes round about “using the top end of the grading system more” so as “not to short-change our students”. Your colleagues knuckle under – after all, everyone else is doing it. If you rock the boat or grade too low, then you are told not to make trouble. In this way, a new norm is developed. We shift our ideas about what constitutes first class work, just a little at first....

This is typical. Selfishness is not just about the breakdown of norms; new norms are also created. As Thucydides put it:

Words had to change their ordinary meaning and to take that which
was now given them. Reckless audacity came to be considered the courage of a loyal ally; prudent hesitation, specious cowardice; moderation was held to be a cloak for unmanliness; ability to see all sides of a question, inaptness to act on any.... (3.82)
The promising side of this is that, in economists’ terms, there are multiple equilibria. Individuals will always be tempted by selfishness; but an organization can only act selfishly if the individuals in the organization tolerate this. When the greater society has a strong claim on people’s affections, it is possible to resist organizational selfishness. Let’s hope that UK academics recognize this, and try harder to uphold our standards in future.

Friday, 23 January 2015

Social science and the migration debate

Large majorities in the UK want less immigration.

What are the possible responses to such views? We could:
  1. Treat them as preferences. Preferences are legitimate inputs into the democratic political process, so we should reduce immigration accordingly.
  2. Treat them as preferences, but ones based on xenophobia. Preferences like that are ugly; they deserve no weight in our democracy. The political process should launder them out.
  3. Treat them as coming from rational beliefs. Perhaps people are concerned about the impact of immigration on wages of native workers, for example.
  4. Treat them as coming from irrational beliefs, due to psychological quirks such as overvaluation of the ingroup, misestimates of the number of immigrants in the country and so forth. These beliefs can safely be ignored or dealt with by public education.
None of these really seems satisfactory. Politicians who have to get reelected, tend to favour 1. But politicians ought to lead, not just follow, public opinion.

2 is appealingly resolute and uncompromising. But the future membership of any group, including a nation, is of fundamental importance for its current members. No democracy can put this issue out of the reach of public control.

3 is also appealing, since it treats the public as sensible rather than racist. The problem here is simple: there is really not much evidence that immigration harms even unskilled workers economically.

4 draws strength from a large psychological literature on ingroup bias, and from empirical evidence that people state mistaken beliefs in answering pollster's questions. It is less good at explaining why immigration has currently risen up the political agenda. Option 4 must be partly true, but it feels patronizing to dismiss people's concerns.

I have no neat solution, but a few comments.

First, with regard to the economic effects of migration, technical research is necessary but not  enough. The stream of history and politics has been flowing against the poor in rich countries for many years in many ways - both ideologically, and through globalization in many forms, from mobility of skilled workers and capital, to offshoring. Exactly which forces have contributed how much to wage stagnation is very hard to disentangle. But the overall outcome is not much in doubt. Immigration is not so much a policy that can be neatly exogenously varied, and more part of a package which people in wealthy democracies may quite reasonably fear. (One may wonder whether poor immigrants should bear quite so much of the brunt of these feelings, compared to, say, financial geniuses or clever-clever right-wing academics.) And migration itself doesn't just have short-term economic impacts. It changes communities and transforms politics; it has reshaped Britain profoundly and will keep doing so. You can welcome that or not, but you cannot limit it. Of course, this is not a call to ignore the evidence, but for humility about the limitations of our research. How would greater or less migration change Britain? We don't know.

Second, we need to understood more about the social and psychological effects of diversity. We still know startlingly little. Socially, diverse communities seem to find it harder to provide public goods. But we don't know when or how they happen - though there are some great starts at finding out.  Social networks, norms, habits and political coalitions are all likely to change with the makeup of society; and there are probably many mechanisms which could come into play.

Psychologically, existing work is good at explaining how prejudice works: "symbolic racism", "implicit attitudes" and so forth. I think we are less good at understanding how individuals experience ethnic outgroups in the context of their daily lives. Paul Celan, himself an immigrant to Paris and a victim of racial persecution, writes of an elderly relative in London in the 1960s:
Die dir zugewinkte
Stille von hinterm
Schritt einer Schwarzen....
The stillness that waved at you 
from behind the gait of a black woman... 
Understanding how people actually experience ethnic strangers might help us understand preferences. Perhaps it might also help us live together better - since that is certainly our fate. Celan's poem finishes with a humanistic flourish: Vertag dich nicht, du. "Don't adjourn yourself."

Wednesday, 21 January 2015

How Open Access works

1. Before open access


You write a paper. It is available for free from your website, SSRN, IDEAS etc. It can easily be found at Google Scholar and read for free.

The paper gets accepted at a journal.

The publishers make you take it down from your website. They publish it on the journal website. They charge large fees to universities for access. People outside universities cannot read it. (Luckily, early drafts are available at the university websites where you have presented the work.)

2. Open access arrives


Information wants to be free! Activists demand that journals provide articles on open access so that anyone can read them -- especially if the government has paid for the research.

The government agrees. From now on, government-funded research must be open access.

3. Now with open access

You write a paper. It is available for free from your website, et cetera. It gets accepted at a journal. The publishers offer you an option: publish it as normal, (i.e. stop people reading it) or publish it open access so that everyone will be able to read it. The fee is a snip at £2000. Your research is government-funded, so you (your university) have to pay the fee. Or, if you want it to be included in the next Research Excellence Framework assessment, again you have to pay the open access fee.

Your research is available for free on the journal website. Anybody can read it. It is easy to find at Google Scholar.

 Everybody is happy!

Tuesday, 13 January 2015

More on Churchill's Mahommedanism

Lady Gwendoline Bertie's letter to Winston Churchill has the line:

“If you come into contact with Islam your conversion might be effected with greater ease than you might have supposed, call of the blood, don’t you know what I mean, do fight against it.”

 The "call of the blood" phrase was puzzling... it seemed like an in-joke. By chance tonight I read that Henry Stanley, third Baron of Alderley, converted to Islam.  Now Lord Stanley's sister was Henrietta Blanche, the grandmother of Churchill's wife.

Unfortunately the timing doesn't quite work: Churchill courted Clementine in 1908 and the letter was written the previous year.

Footnote: according to Nancy Mitford, at the third Baron's Islamic funeral, his brother Algernon turned to the new Lord Stanley, who had removed his hat as a mark of respect, and snapped "Not your hat, you fool, your boots".

Thursday, 8 January 2015

On Charlie-Hebdo: group identity, pride and fear

Humans favour their own group – even when that group is an arbitrary label given them by experimenters. They like members of their group more, see them as superior, and behave more altruistically and forgivingly to them. The dominant explanation among social psychologists for this "group identity" is pride. People's "social identity" as a group member is part of their personal identity, and they support their group to shore up their own self-esteem.

I prefer the theory that group identity comes from fear. People stay close to their own group and support it, because they fear the other group. That in turn makes members of the other group afraid and pushes them closer together.

So far Western politicians have been responsible in differentiating between terrorists and peaceful Muslims. George Bush was adamant that the US was waging a war on terror, not on Islam.

Let us hope that this will last. But there is a lot of fear in France now, and a big potential audience for leaders who promise to keep the majority safe. The decline into mutual fear and hatred can be fast: when the situation gets dangerous, everyone is forced to seek protection from their own group's warlords and thugs.

I feel this change even in myself. I struggle to remain open-hearted. But a little voice within me starts whispering that perhaps the bigots are right, that Le Pen has a point, shouldn't we ban the headscarf, on feminist grounds of course... How many people in France are going through the same mental process?

Wednesday, 7 January 2015

A smackdown for the APA on smacking

"Spare the rod", says the Economist magazine, with the subtitle "spanking makes your children stupid". An American Psychological Association task force has reported that corporal punishment goes together with anxiety, depression and behavioural problems.

Six of the best for the Economist, and a sound flogging for the APA, for committing the fundamental statistical sin: confusing correlation with causation.

Sure, corporal punishment goes together with behavioural problems. But there are several ways to interpret that. Maybe corporal punishment causes long run behavioural problems: i.e., if as a parent you decide to use corporal punishment, you make your children's behaviour worse in the long run. But it is also possible that people who use corporal punishment may be less smart, happy, liberal, urban, educated and/or wealthy than those who talk to little Tim about moral values; and that such people have more anxious, depressed, or naughty children, irrespective of how they punish. And isn't it also possible – nay, likely – that children with behavioural problems induce their parents to smack them more?

A dissenting member of the task force made some of these points:
 In the few studies that have compared spanking with other forms of punishment, such as restriction of privileges, grounding and time-outs, all the punitive measures examined resulted in similarly negative outcomes in children....  Scientific American
And a recent meta-review of the evidence [gated content] says:
Concerns with [these analyses] include:
1) Conflation of spanking with more severe forms of corporal punishment...
2) The temporal order of spanking and negative outcomes is not well documented. From cross-sectional correlational studies, it is not possible to determine whether spanking and CP [Corporal Punishment] lead to negative outcomes, or whether children with greater problem behaviors are more likely to be spanked....
3) Controlling third variables... effect sizes based on bivariate correlations run the risk of inflating effect size estimate due to failure to control for other relevant variables.
4) Overreliance on retrospective recall... Retrospective reports of past CP may ultimately say more about the respondent's current psychological status rather than accurate memories of disciplinary acts, which took place years or decades earlier....
It is surprising that such an important topic has been little studied using longitudinal methods, to figure out which came first, the spanking or the naughtiness. But even that would not deal with point (3). Correlation plus being earlier in time is still not causality (e.g. taking one's umbrella to work does not cause rain).

This is a classic case where nothing but a proper experiment will do; but random assignment is very hard to do ethically. "I'm going to give you [flips coin] 20 strokes of the birch!" One approach might be similar to that used by economists studying divorce: look at whether child outcomes are improved in states which ban corporal punishment. This would answer a useful policy question, especially regarding corporal punishment in schools. But the most important "policy" audience here is the parent who wants to know how to discipline his or her children. Perhaps an encouragement design, in which 50% of a group of parents who use corporal punishment are encouraged to give it up, would help.

Meanwhile, there is just very little evidence to go on.

I read my mother's diary from when I was one year old. In one place she writes of getting angry and smacking me: "something (the Devil, I suppose) made me..." My mother smacked me very rarely, that I can remember - perhaps on one or two occasions. Still, her words made me think how one-sided the power relationships are between parent and child. I hope I never feel I have to spank my children.

At the same time, this is a disciplinary method that has been used by parents throughout human history. The idea that it is wrong is very, very recent. Sometimes, traditional ways of doing things have advantages which scientists take a long time to discover (sleeping without a light on, turning the radiator down). Experts should be wary of issuing broad recommendations to the world's parents – especially on such a thin and inadequate evidence base.

(Here is the other point of view.)

Monday, 5 January 2015

Using refset in your R analyses

refset is an R package that creates subsets which refer to an original dataset. Here is a short example of how to use refset to simplify your data analysis.

I have a dataset from a laboratory experiment. In total, 532 subjects participated. Each subject was in a group of 4 and made decisions in 20 rounds. There were several treatments.

I typically create two scripts: R-create-data.R makes a data frame from my raw files, including any computed variables, and saves it to a .RData file. R-analysis.R loads that file and runs any statistics or graphs. That way, I can rerun my analyses without recreating the data.

At the end of R-create-data.R, I have made a data frame called subj. A row of subj records one decision by one subject, so it contains 532 * 20 = 10640 rows. The script then create subsets of subj to represent, e.g. the data from particular treatments:
subjhist <- subj[subj$treatment1=="H",]
subjPX <- subj[subj$treatment1=="PX",]
subjn <- subj[subj$treatment1=="N",]
subjst <- subj[subj$treatment1=="S",]
save(subj, subjhist, subjPX, subjn, subjst,    file="data.RData")
In theory, the script has created all the variables I need, and now I can just run regressions using whichever subset is appropriate. In practice, I will look at the data and think of a new analysis I should do, which requires a new variable.* Hmm... did my subjects play differently in earlier than later rounds?

When that happens, I need to create a new variable in all my datasets:
subj$lateround <- subj$Period > 10
subjPX$lateround <- subjPX$Period > 10 subjhist$lateround <- subjhist$Period > 10
## et cetera...
Eventually, I will get round to including lateround in R-create-data.R, rerunning that file, and removing the code above from my analysis script. But in the meantime, doing this is a hassle, and creates repetitive, verbose, unreadable code.

Let's replace those subsets by refsets.
subjhist %r% subj[subj$treatment1=="H",]
subjPX %r% subj[subj$treatment1=="PX",]
subjn %r% subj[subj$treatment1=="N",]
subjst %r% subj[subj$treatment1=="S",]
Each of these variables now refers to the main data frame subj, rather than being a copy of it.

We'll put this in R-analysis.R, so that it happens after our data is loaded. Otherwise, saving and loading breaks the connection between refset and data frame, and simply creates subsets as before. The data will look just the same, and I can run my existing analyses unchanged.

Now, when I create a new variable, I just do it in the main data frame:
subj$lateround <- subj$Period > 10
The variable is then visible in all of my refsets:
head(subjPX$lateround, 20)
## same for subjhist, subjn...

Problem solved: I can create new variables once, not several times, and immediately use them in any analysis.

* Yes, I know... one should have prepared all my analyses ex ante and simply run them at the press of a button. Otherwise one is just doing post hoc exploration and curve fitting. I applaud the ideal and strive to live up to it, but the real world isn't always like that -- and in any case, sometimes post hoc exploration is worth doing.

Political economy of the UK and Europe, January 2015: Doom and Gloom edition

This week I started a new job in the Department of Economics at the University of East Anglia. To celebrate, I will share my thoughts about the UK economy. As I am not a macroeconomist, these ideas are opinionated ramblings, not expert knowledge.

Florence, once at the heart of early capitalism, is now a tourist town, lying in state among its architectural heritage. Europe is making the same slow transition from a place where wealth is created to a place where it is spent. Better economic policy could slow but not stop this change, since the markets of Asia contain the majority of the world's population, including two massive polities: that is the customer base to start your business, if you want to conquer the world. Since the gains from better economic policy would be limited, we can predict that it will not be forthcoming. Europe will see some economic reform, perhaps even Thatcherite revolution in places, but not enough to clear the huge overhang of debt.

The UK is no exception to this story. It now has an economy driven by house prices, like a tropical island whose inhabitants give up fishing so as to sell their shacks to millionaires. The Conservative party's constituency is the homeowner, its policies are designed to shore up house prices, and the economic recovery to date is not based on real economic reform but on low taxes (unsustainably low given our debt) and a house price boom with the attending rise in consumption. The Labour party's constituency is the public sector: its policies may be more competent and responsible, but it is equally incapable of economic reform. Both parties have second rate leaders.

Institutions – the police, judiciary, local government and so on – need to be coordinated: this can happen either via political leadership, or by coordination through the informal networks we call the establishment. Unfortunately, we have a vacuum of political leadership, and the traditional establishment has collapsed. There are plenty of people in power, and they certainly know each other, but they have no shared values or mutual trust on which to act in concert, and they are equally untrusted by the man in the street. When institutions are not coordinated, they continue to function, but begin to work in a way that serves their own interests rather than that of the whole. This will increasingly happen here.

On the “high politics” side, the UK's decision to remain outside the Euro, while economically prudent, has cemented the Franco-German axis. I do not know whether the Eurozone will survive the next year in its current form. If it doesn't, then the UK is on the outside of a failing continent, which, like Italy in the 16th century, will be pulled apart by the surrounding larger powers (think the Cold War, but more chaotic). If it does, then the UK is on the outside of a strong regional player which will understandably make the rules to favour French and German capital, not UK financial capital. Neither is appealing, though the second is preferable.

I see I've ended up writing about politics not economics, but as politics determines economics, that's fine. Clearly this is not an optimistic view. I wish I could suggest policy solutions, but in political economy, notoriously, there is no actor to appeal to outside the system. The role of social science is sometimes just to help people understand and cope with ineluctable realities.