Stumbling and Mumbling

Unemployment kills

Mass lay-offs are, literally, deadly according to this new paper (pdf):

We find that job displacement leads to a 15-20% increase in death rates during the following 20 years. If such increases were sustained beyond this period, they would imply a loss in life expectancy of about 1.5 years for a worker displaced at age 40. These results are robust to extensive controls for sorting and selection.

This isn't just because unemployment leads to worse mental health. It's because low and volatile wages do too. Nor is it a US peculiarity; the same thing seems true of Sweden.
This should dispose of the myth that class inequalities (pdf) in health are due simply to differences in education or IQ. Even in advanced capitalist economies, you pay a high price for being economically powerless. The macho management of men like "Chainsaw" Al Dunlap is murderous.
This suggests that one trade-off might be more acute than realized. On the one hand, corporate restructuring is sometimes necessary; without it, economies would stagnate. But on the other hand, the costs of that restructuring might be higher than supposed.

Brown on skills

Brown says we need to become more skilled:

As a result of changes in the global economy many of the jobs British workers do now are becoming redundant.
Of today's 6 million unskilled workers in Britain we will soon need only half a million - over 5 million fewer.
We have 9 million highly qualified workers in Britain - but the challenge of the next ten years is that we will need 14 million - five million more. Higher standards of living will depend on higher standards of learning.

There are four problems here:
1. The misplaced confidence  in futurology. How can Brown claim to know what  thousands of employers will need in years to come, when he and his colleagues don't even know what their own employees are up to today?
2. The fallacy of composition. For any one individual, getting skills is the key to getting better wages. But it doesn't follow that the same is true writ large. If we all become more skilled the effect might be merely to bid wages down.   
3. The Star Trek fallacy. It's a common error in forecasting to extrapolate present trends; Star Trek writers foresaw space travel but not the internet. Maybe Brown is doing the same. Let's grant - which might not be true - that globalization has in recent years raised demand for skilled workers relative to unskilled ones. It doesn't follow that this will remain the case. This paper (pdf) argues that even skilled workers will soon face competition from lower-paid Chinese and Indians. Meanwhile, demand for many workers deemed to be low skilled - cleaners, shelf-stackers, care home workers - is sheltered from foreign competition. And indeed, there might be a replacement demand for such workers, as the current generation of older unskilled workers retire.
4. Blaming the victim. What Brown doesn't say is that there is already an over-supply of skills. Table 3 of this paper estimates that over one-fifth of men and a quarter of women are over-qualified for their jobs. Table 5b of this this paper (pdf) reckons that over half of workers are moderately or severely over-qualifed.  And this paper (pdf) estimates that a third of graduates are over-qualified for their jobs.
One could equally well argue (pdf), therefore, that what the economy needs is not more skilled workers, but rather that employers use the skills workers already have.
However, one feature of Boss party ideology is that workers must adapt to the needs of capital, not vice versa.

Child abuse & earnings

One of the most controversial claims Richard Dawkins makes in The God Delusion is that indoctrinating children into religion might do them more damage than sexual abuse. Well, here's some evidence, at least about sexual abuse - it is associated with higher education and earnings in later life.
Researchers estimate that Canadian women who were  first sexually abused between the ages of 7 and 12 were more likely to get degrees, and more likely to earn good money, than women who weren't abused. 
Now, one might object that there's sampling bias here. The abuse is self reported. And it's possible that women who have faced up to the fact that they were abused as children are more resilient than those who are in denial, and this resilience might be associated with higher earnings.
This objection, however, runs into the problem that women whose self-reported first experience of abuse began before the age of 6 or after 12 do no better (or worse) in education and earnings than women who weren't abused.
The mechanism here might be straightforward.  Abused women might be deterred from forming social and sexual relationships, and so divert the effort they'd put into social life into education and work.
This does not of course  mean sexual abuse is a good thing, even leaving aside the obvious reasons.  This finding is consistent with the possibility that abuse reduces women's ability to convert educational or work effort into achievement. It might just be that the extra effort they divert into work effort compensates for this inward shift in their production function.
There might be a broader message here. Not all psychological damage is consistent with poor labour market outcomes. Outwardly successful people needn't be psychologically healthy.

Labour market mismatch and immigration

If you want to know why immigration is so high, today's labour market figures (pdf) give you one reason.
They show that unemployment and vacancies have both increased in the last two years. This is evidence of increasing mismatch in the labour market. The skills employers want are, increasingly, not the same as the skills workers are offering. The upshot is that jobs lie vacant whilst workers stay unemployed. And the gap gets filled by migrant workers.
In this sense, Brown's slogan "British jobs for British workers" is not just morally obscene, but economically moronic too.
In a dynamic market economy, jobs are continuously destroyed and created; today's figures remind us that in a typical month, almost one-in-four of the unemployed leave the claimant count, with another one-in-four joining it. These changes will inevitably result in differences between what employers want and what workers offer. Sometimes these differences will be small. Sometimes (as now) they'll be high.
To claim that the government can somehow ensure British jobs for British workers is to suppose that government can somehow anticipate these mismatches. That's just not possible.

How much discrimination is there?

Samara at the F-word laments this report which shows that men earn £1000 more than women within three years of graduating. This, she says, is because of either outright discrimination, or because men are more likely to study well-paying subjects.
There are, though, (at least) two other possibilities:
1. Men are fussier about which jobs they take. This means men are more likely to be in jobs with which their skills are well matched and so earnings higher. Women, by contrast, take any old job which might be a bad match. The full report (pdf) says that men are more likely than women (20% vs 13%) to have had a spell of unemployment since graduating.
2. Women, more than men, are more likely to choose job satisfaction over pay; the report finds that 40% of women are very satisfied with their careers compared to 34% of men.
Both these theories are consistent with a form of discrimination - that women are socialized differently, to be grateful for any job and not to be so obsessed with money.
I should confess, though, to a Bayesian prior here. Intuitively, I find explanation 2 much more plausible than the idea that women suffer outright pure discrimination in the sense of being paid less than men for doing the same job. Is this because I'm overly influenced by Adam Smith's theory of compensating advantages which says good pay is only a reward for hassle? Is it because I generalize from personal experience which vindicates Smith's theory? Or is it because I'm naive?

Migrant numbers: do they matter?

Immigrant workers impose less cost upon society than previously thought. That's the obvious message from the fact that the number of migrant workers is 300,000 higher than the government previously estimated.
The thing is, the social costs of  migrant workers  - crime, loss of social cohesion, pressure on public services, over-crowding - haven't changed. It follows, then, that the social costs per migrant worker are lower than thought*.
Why, then, is there so much fuss about this revision to the figures? Partly, of course, it's just normal party politics. But there might be other things happening.
Lib Dem spokey Danny Alexander  says it "undermines the credibility of the government's claims to be able to deliver a well-managed system for foreign workers." Maybe. But this strengthens the case for free migration - because the state, whoever runs it, just cannot control migration.
Another reason, I suspect, lies in fact fetishism.  People think that aggregate national data - be it the number of migrant workers, GDP, average house prices or whatever - must somehow be important. But it ain't necessarily so. Why should the number of non-native workers in Britain be any more important than, say, the number of people working in Hinckley who weren't born there?
There might, however, be a more sinister reason why the political class is so het up about this. It's that, as Foucault pointed out, facts and knowledge matter not (just) because they explain and describe the world, but because they are a means of controlling people. That the government can't be sure how many foreigners work in the UK is therefore a symptom that perhaps government cannot control us, that there's a crisis of governmentality. And for politicians of all parties, this is a terrifying thought.
* Of course, the benefits are lower too.

Martin Jol and hiring problems

Martin Jol's sacking by Tottenham, just months after  he had been given a huge transfer budget, raises the question: why do football directors make so many bad hiring and firing decisions?
In theory, you'd expect the opposite to be the case. Football directors are usually very successful businessmen with presumably lots of experience of hiring people. And football coaches' track record is publicly available, in a way that employees' records in other businesses are not; most us in a job interview could exaggerate our achievements without being found out, whereas Gary Megson could not claim to have done a great job at Nottingham Forest when he applied for the Bolton job.
You'd therefore expect better matches between coach and club than in most businesses. But this is not so. Last season saw 48 management changes at League clubs, implying that the average boss lasts less than two years - much less than job tenure in other businesses.
Here, then are some theories:
1. There are threshold effects. It's moderately easy to hire someone to do a basically competent job - though even here, bad hiring decisions cost economies around 1% of GDP a year. But it's vastly harder to spot someone who can do really well. Spotting real talent is hard.
2. Success in football coaching is not easily replicable. What works at one club can fail at another - the classic instance being Brian Clough at Leeds. So a coach's track record gives less clue to his chances of success than one might think.
3. Football directors' success in business amplifies their over-confidence and egomania. And just as overconfident share traders take too many decisions, so overconfident bosses are too quick to hire and fire, believing that they can get it right this time.
4. Criteria for success are too high, as directors are not just overconfident, but over-optimistic. Worse still for coaches, success raises expectations which increases the likelihood they'll fall short. This contributed to Jol's problems. His record at Tottenham (45.3% of games won, 28.3% lost) was in fact better than Tottenham's long-term average. But his bosses demanded even better than better-than-average.

Is Coulter right?

The ridiculous Ann Coulter says:

If we took away women's right to vote, we'd never have to worry about another Democrat president...women are voting so stupidly, at least single women.

There's some academic evidence to support at least the first half of this assertion: this paper (pdf) is called "Women prefer larger governments":

We show, for a wide cross-section of developed and developing countries, that higher female participation rates in the labor market are positively associated with larger governments...Female labor force participation is found to cause an increase in government size, with a 10 percent rise in the former leading to a 6.5 to 9 percent rise in the latter.

Insofar as Democrats traditionally offer bigger government, this corroborates Coulter.
But it doesn't follow that women vote stupidly.  Quite the opposite. As they work outside the home, they want more government services such as education and healthcare - because government provision of childcare helps them work and earn more. This is rational enough, even for single women as these often anticipate becoming mothers later.
The thing is, only the most bigoted partisan could possibly see this as an argument for taking away women's right to vote. Surely, a more reasonable position is that it merely shows the truth of Isaiah Berlin's saying, that political goods cannot all coincide. There's a tradeoff between the good of limited government (assuming this to be a good) and the good of empowering women.

Costly discrimination

Sex discrimination costs billions of pounds around the world. That's the message of this paper (pdf).

We find that for several countries a large fraction of the actual difference in output per capita between the U.S. and the different economies is due to gender inequality. For countries such as Ireland and Saudi Arabia, wage discrimination actually explains all of the output difference with the U.S.

It can explain around two-thirds of the difference in per capital income between Iran or Egypt and the US.
There are two reasons why gender discrimination is so expensive. One is simple maths. If women stay at home, they're not earning money. That depresses GDP per head. The other is that when women don't work, they have more children. And this spreads GDP over more people, again depressing per capital incomes.
This suggests that gender equality and economic development go hand-in-hand.

Migration as brain gain

One of the apparently strongest arguments against immigration is wrong, according to this new paper.
The argument is that migration depresses the economic prospects of poor countries, as their brightest people leave.
But this isn’t true.  The authors studied the determinants of education in Cape Verde, a country notable for having a large proportion of its natives living overseas - and, not uncoincidentally, a better economic performance than most African countries.
They found that emigration increased the human capital, and hence aggregate economic prospects, of the remaining population.
This is simply because youngsters get educated in the hope of emigrating. And as some of these either don‘t subsequently emigrate, or come back after they have (with valuable experience)  the result is a higher stock of human capital. The authors estimate that banning emigration from Cape Verde would reduce the proportion of graduates in the labour force by almost 40 per cent. Unless you’re an extreme advocate of education as signalling, this implies a big hit to growth.
This suggests that free migration of educated people can be a form of efficient aid - it helps both donor and host country.