Betting with Bryan Caplan

Bryan Caplan responds to the data on US and EU-15 unemployment by offering a bet

Cry of the Owl psp .

The average European unemployment rate for 2009-2018 (i.e., the next decade) will be at least 1 percentage point higher than U.S. unemployment rate. The bet will be resolved when Eurostat releases its final numbers for 2018.

Betting is usually unwise, but nonetheless I’m willing to take Bryan on, with one amendment. I will take the bet provided that people in prison are counted as unemployed. By my estimate, that raises the US rate by about 1.5 percentage points and the the EU-15 rate by about 0.2 percentage points. That is, assuming current imprisonment rates remain unchanged, the bet is that the Eurostat measure of unemployment (which excludes prisoners) should be no more than 2.3 percentage points higher in the EU-15 than in the US.

A few points about the odds. I haven’t been able to download the time series data, but eyeballing this graph suggests Bryan would have won the bet narrowly if it had been run over the last 15 years.
US and EU-15 unemployment rates since 1993

I think the bet is fair, since unlike 1993, the EU-15 is starting ahead. Also, although EU geographical mobility is still much less than in the US, it has increased dramatically over the past fifteen years, and that is likely to continue, particularly if some countries recover from the current crisis faster than others.

Looking to the short term future, the big question is whether, as I argued recently, the EU system is characterized by lower variance than the US, which would suggest EU rates should be lower during the global recession. The alternative view is that the EU is just not as far into the cycle as the US and that the US will recover earlier and faster.

Thinking about the bet more generally, if you regard it as supporting the view that the proposition “in the long-term average US unemployment rates are about 1 percentage point lower than average EU-15 rates” is an even money bet, that has a number of implications.

First, since the EU-15 countries are quite disparate, this suggests that the US is likely to be, on average, around the middle of the pack of developed countries as regards unemployment rates (adding in non-EU countries like Japan and Australia wouldn’t change this much).

Second, although the US is middling on unemployment outcomes, it’s an outlier on a range of measures that have been presented as important in promoting high employment. In addition to higher geographical mobility, it has very low minimum wages (lower now in real terms than it was in the mid-1950s), very weak trade unions, almost no restrictions on hiring and firing, and very limited welfare benefits for unemployed workers*. It’s quite surprising, even to me, that all of these things should add up to a difference of only one percentage point in unemployment. In part, I suspect that these institutions create their own kind of dual labor force structure.

In political terms, it’s hard to see how the pressure to adopt “more flexible” * labour market institutions can be justified by reference to the US example. While lower unemployment is better, it’s hard to see why a country with a decent minimum wage, strong union movement and good social welfare systems would want to scrap those things to achieve a one percentage point reduction in unemployment.

* As far as I can determine, in most US states, childless adults who have exhausted their unemployment benefits (or were ineligible) don’t have any access to cash benefits, and aren’t, in general, eligible for Medicaid or even, in some cases, food stamps. Can someone confirm or correct my understanding?

Update Quite a bit of quibbling in comments as to whether prisoners should count as unemployed. To short-circuit this dispute, can we all agree that breaking rocks (or similar) under armed guard is a bad labor market outcome, as is being unemployed. As the desire to debate this point shows, the excess prevalence of the first of these outcomes in the US, relative to the EU, is similar in magnitude to the excess prevalence of the second in the EU. Combined, they pretty much cancel out.

36 thoughts on “Betting with Bryan Caplan

  1. AR, this is a pretty good measure for prime-aged males (25-54) since it can generally be taken for granted that being employed is a Good Thing for this group, under current social conditions.

    But, outside this group, alternatives like study, childraising, and retirement may be Better Things.

    And, there’s still the problem that the standard data on E/P ratios refer to the civilian non-institutional population, thereby excluding prisoners.

  2. JQ @22, I don’t really wish to defend the US criminal justice system, as I believe it is generally too harsh and punitive. Although Australia’s system is perhaps too lenient. So there is a happy medium somewhere.

    But I do agree that providing a good, efficient, and fair system of law and order is a more positive contribution to general societal wellbeing than anything else governments can do. Not that the US does a particularly good job at that.

  3. JQ set the conditions of the bet re the prison populations. If people don’t like the conditions then don’t take the bet. The proposer of the bet proposes the odds and the conditions.

    Mind you, I agree that prison populations should be added to the unemployed count. In addition, I would the extra prison guards the US needs over the EU model (on a per capita basis), as these guards are not productive. The EU justice system model shows that these extra prisons and guards are not necessary in a more socially enlightened system.

  4. My understanding is that in the US companies close their production lines then re open them in prisons because the prisoners are paid $1 a day. So I suppose the former workers steal, get sent to gaol then continue in the former job for $1 per day plus food and accommodation. I saw a program on private gaols in the US operated by Wackenhut who operate the Australian Detention Centres.

    So I disagree with Ikonoclast’s ascertain that the prison population always be added to the unemployed numbers. Prisoners have laboured in prison farms and stamp out our number plates

  5. “The proposer of the bet proposes the odds and the conditions.”

    The proposer was Caplan. It was Quiggin who wanted to Quibble.

    “I agree that prison populations should be added to the unemployed count.”

    Fine, but it’s a weird thing to do. If the War on Drugs was scaled down and prisoner numbers declined, would every single one of them be unemployable? I think not. Obviously a proportion of the prison population could be included, but not all.

  6. Prison work is not productive. Most of what is called “work” in prisons is merely a punishment detail.

  7. As well as those in prison there must be a lot more “self employed” in the US – those who are not eligible for support and therefore need to turn to crime. Also, there are probably a larger percentage who are not counted in surveys – the homeless, those living in ‘no go’ areas and so on. The existence of an army of illegals must also have some sort of impact on the numbers. It could be worthwhile to find out how the numbers are estimated in the US and what difficulties they have when estimating unemployment.

  8. I think only a proportion of prison populations (and likely very small – so it seems a red herring anyway) should be added. What if unemployment drove them to crime or drug dealing or other criminal activities which caused their imprisonment them in the first place….I do9nt see $1 a day production from prisons as productive…just the use of almost slave labour

    and then as someone suggested why dont we count the unemployed illegals living with relatives..

  9. I wouldn’t have thought you have to believe every prisoner would be unemployed if let out to endorse JQ’s point. If someone who is employed gets sent to prison most of the time someone will get their job, and probably be doing it pretty much as well, but reducing the official unemployment figures.

    If you add together all the prisoners who would be unemployed if on the outside, and all who are easily replaced in the workforce it would come close to the prison total. Maybe Professor Q should knock 0.1% off for the proportion of prisoners who are genuinely hard to replace.

  10. “If someone who is employed gets sent to prison most of the time someone will get their job, and probably be doing it pretty much as well, but reducing the official unemployment figures.”

    This kind of thinking underpins much discussion of the labour market, but is fairly spurious. That is, the notion that there is somehow a fixed amount of work available and therefore if one section of the population works less or stops working it will automatically make more work available elsewhere. This is a kind of flat-earth economics.

    If one section of the population stops working it has a negative impact on the economy and therefore the demand for labour elsewhere. Also, if people stop working their income usually goes down which means they have less to spend on goods and services and thereby generate jobs elsewhere.

    It doesn’t necessarily follow that every time someone stops working a job is freed up for someone else. Sometimes it is. But sometimes a company that loses many good workers goes out of business, thereby creating further job losses. Or sometimes a lost worker is not replaced if there aren’t applicants with similar skills available.

  11. Pr Q says:

    Update Quite a bit of quibbling in comments as to whether prisoners should count as unemployed. To short-circuit this dispute, can we all agree that breaking rocks (or similar) under armed guard is a bad labor market outcome, as is being unemployed. As the desire to debate this point shows, the excess prevalence of the first of these outcomes in the US, relative to the EU, is similar in magnitude to the excess prevalence of the second in the EU. Combined, they pretty much cancel out.

    Commenters are quite right to quibble. Pr Q’s formulation of the incarceration problem is fatally flawed. Any analysis of US labour market exceptionalism which focuses on economic, and ignores ethnic, factors is not up to scientific par.

    In particular, the explanation of the US’s “bad labour market outcomes” relative to the EU needs to take into account extraordinary rates of African-American incarceration and Mexican-American immigration. Otherwise it would not be an apples-to-apples comparison.

    About 70% of the US prison population are colored minorities, in a population where 30% fit that classification. Those disparities in prison demographics cannot be explained soley by socio-economic disadvantage.

    If a substantial fraction of the US population have trouble abiding by common sense laws (“say no to drugs”) then how are they going to hold down a regular job on the outside? The evidence so far is not looking good.

    From the late fifties through the early noughties black male labour force participation in the key prison demographic (16-24) has plummeted to rates 30% to 50% below comparable white rates. Andrew Sum crunched the numbers:

    Employment rates among black male teens and young adults ages 16 to 19 have dropped considerably over the past 50 years, the study found. In 1954, a slight majority – 52 percent – of black male teens worked, a rate slightly in excess of their white peers. By 2003, however, only one of five black male teens was employed in a typical month – just 20 percent – only half the employment rate of white male teens.

    Among 20 to 24 year old black males, employment rates also have declined considerably from their peak values of 77 to 83 percent in the mid to late 1960s to dramatic 50-year lows more recently. During 2003, for example, just 56 percent of such young black men ages 20 to 24 was employed.

    The prison demographics suggest that, in the post-segregation era, black male labour morale, at least for the bottom half of their Bell Curve, has more or less collapsed. That would imply that even if the prisons were emptied those liberated prisoners would not be willing or able to enter the regular workforce.

    They would not become unemployed as they are, apparently, unemployable under current conditions. They would most likely economically fall off the BLS radar, just like vast numbers of their un-incarcerated brothers. It is therefore a fallacy to treat, as Pr Q does, the US prison population as “hidden unemployed”.

    I dont have all the stats handy to do accurate algebra. BOTE calculation, blacks are 14% of the population therefore are, pro-rata, likely to be 7% of the eligible workforce. About one-quarter of them would have done jail time say 2% of the total workforce. More than half of that lot would likely fall into the non-participating half of the black male labour force. That pretty much eliminates Pr Q’s 1-2% of potentially employable incarcerated persons.

    Its implausible to count most incarcerated black males, given their typical employment participation rates, as “willing and able” for gainful employment. Therefore Bryan Caplan is quite right to be wary of Pr Q’s wagering stipulation.

    But there is one form of economic liberalism that is actually wrecking the labour market outcomes of the bottom tier of American workers. And that is mass illegal immigration, mainly from Hispanic countries.

    These workers take jobs that might be done by native born and drive down wages for those natives in jobs. Particularly native born black Americans. George Borjas surveyed African American employment trends in the post-desegregation era and found that mass immigration had been a direct cause for the disaster for black male job opportunities and had indirectly caused a massive spike in black male incarceration:

    The employment rate of black men, and particularly of low-skill black men, fell precipitously from 1960 to 2000. At the same time, the incarceration rate of black men rose markedly. …

    As immigrants disproportionately increased the supply of workers in a particular skill group, the wage of black workers in that group fell, the employment rate declined, and the incarceration rate rose.

    Our analysis suggests that a 10-percent immigrant-induced increase in the supply of a particular skill group reduced the black wage by 3.6 percent, lowered the employment rate of black men by 2.4 percentage points, and increased the incarceration rate of blacks by almost a full percentage point.

    A moments thought and this conclusion is exactly what anyone with any real knowledge of the US economic system would draw. (I have worked in the US for lengthy periods of time, less than a year all up.)

    In fact this conclusion has been drawn by an economist of impeccable Left-liberal persuasion: Heres Paul Krugman arguing for much tighter restrictions on mass immigration, to protect the worst-off native workers:

    many of the worst-off native-born Americans are hurt by immigration — especially immigration from Mexico. Because Mexican immigrants have much less education than the average U.S. worker, they increase the supply of less-skilled labor, driving down the wages of the worst-paid Americans. The most authoritative recent study of this effect, by George Borjas and Lawrence Katz of Harvard, estimates that U.S. high school dropouts would earn as much as 8 percent more if it weren’t for Mexican immigration.

    Hats off to Pr K for at least mentioning the unskilled immigration gorilla in the labour market institution living room. I wonder, idly to myself, whether Pr Q will follow suit.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s