Thursday 5 May 2011

Youth unemployment and evidence-based policy

I'm probably the only one on my block who runs around the house shouting "The new Household Labour Force Survey is here! The new Household Labour Force Survey is here!". The neighbours likely already think I'm a jerk anyway.

And so it's time to update the estimate of how many kids aged 15-19 are currently unemployed because Labour, at the behest of the Greens, eliminated the differential lower youth minimum wage. And let's not forget that National, who opposed the abolition when Labour was doing it, decided that they liked pricing kids out of the labour market when they had a chance to change things.

The method is fully explained here. The .do and .dta files are also posted there. To update things for the current quarter, add in another line for March 2011, quarter number 101, with an adult unemployment rate of 5.63% (rate for all persons aged 20 and up), a youth unemployment rate of 27.5%, and a labour force population of 150.9 (thousands) for the group aged 15-19.

The model expects, given the current adult unemployment rate, that the youth unemployment rate would be 19.3% if the youth unemployment outcomes were no worse (relative to adult outcomes) than in the worst quarter from 1986 to 2008. As the actual youth unemployment rate is 27.5%, the rate is 8.2 percentage points higher than would have been expected under the prior trend. That translates to 12,350 kids who don't have work who we would have expected to be in work had the prior relationship between youth and adult unemployment rates continued.

Here's the table, updated with this quarter's results.

QuarterAdult unemployment rateExpected youth unemployment rateActual youth unemployment rateExcess youth unemployment, in thousands

I'm less confident about this quarter's figures because all of this quarter's stats will have been jostled about by the February earthquake. If the earthquake differentially affected firms employing youths (compared with those employing adults), the residual for this quarter will be picking up that effect. It's not implausible that kids employed in retail were hit harder than folks whose jobs shifted location.

But it seems fairly safe to attribute five to eight points of youth unemployment to forcing employers to pay sixteen year olds as much as older workers. Hit the "minimum wage" tag for further background and responses to counterarguments.

I wasn't surprised to see Green MP Catherine Delahunty exclaiming surprise about the current high unemployment rate. But I'd be interested to hear from Green MP Gareth Hughes on it. Hughes is one of the more promising MPs in Parliament; he was excellent on the recent copyright bill.

Gareth, and the rest of the Greens, if you want everybody else to take the general consensus of academics working in a field as sufficient basis for policy, as in climate change, you're going to someday have to face up to that the general consensus of economists is that minimum wages are a really poor way of helping the folks you want to help. The theory behind economists' beliefs - basic price theory - is at least as well established in our field as the basic theory behind greenhouse gas effects on the overall climate. Price theory is kinda the basis for everything we do. And, as in climate science, most debate is about the relative magnitude of the effects. Some studies find no effects of minimum wage changes, others find large negative effects. It's often hard to isolate effects because states usually only increase minimum wages by any non-trivial amount when the labour market's doing well anyway. But it's also often hard to isolate the effect of year on year carbon emission changes on global temperatures. The Canadian study, linked above, showing that increases in minimum wages increase poverty rates I find fairly damning.

To dream the impossible dream: that politics someday stops being about rallying around stupid symbolic policies that often hurt those they're intended to help and instead gives choice among a few policy platforms, each of which would be the best means of achieving its proponents' desired and stated ends. It'll never happen.


  1. Thanks Eric. Following our tweets the other day wanted to understand your "Green economic policies" comment better.

  2. "The neighbours likely already think I'm a jerk anyway."

    Not just the neighbours!!!

  3. It is a bit condescending to assume that an MP is not familiar with the price/demand model underpinning basic economics.

    "In 2000, only 46 percent of members of the American Economic Association agreed that minimum wages increase unemployment among young and unskilled workers. Another study published in 2006 showed that slightly less than half of all economists surveyed thought the minimum wage should be eliminated, while more than a third favored increasing it."

    Seems like there is some disagreement between experts about how simple the minimum wages <---> unemployment relationship is!

  4. Also your data covers only 4 years, and you use that to try to show a connection???!

    I did pretty bad at stats at uni so I'm not 100% on this but my gut doesn't feel like that is a long enough time to average out statistical noise. Why did you only use 4 years of data? Why not 40, or go back to when the minimum wage was introduced?

  5. @Rimu: I encourage you to go and read the prior post where I discuss the method. The data set goes back to 1986. I only report results here for the era post the abolition of the youth minimum wage. I start in 1986 because that's when the data series starts; going back prior to that would introduce some data series commensurability problems. And quarterly data back to 1986 is plenty sufficient for showing that results since 2009 have been very anomalous.

  6. Check the wording of the survey you're citing. I would agree with the 46% cited that minor changes in the US federal minimum wage, which is already low, make little difference in overall employment. American minimum wages tend to rise very slowly - they've deteriorated over time in real value. So they have little employment effect in aggregate statistics. But they do have real effects on poor urban blacks.

    Suppose that for the last 50 years we've had carbon taxes. And the carbon tax tends only to rise with inflation. If you polled climate scientists after that period, most of them too would likely say that carbon taxes have little effect on overall emissions, because the context of the question is around the very small changes that really don't make much difference. But that doesn't mean that the overall effect of the system as a whole isn't large.

  7. Rimu: In a 1996 survey of labour economists, in answer to the question “A minimum wage increases unemployment Among young and unskilled workers”, 13% generally disagreed, 37% agreed with provisions and 50% generally agreed. That looks like majority support for the idea that increases in the minimum wage increases unemployment among young workers.

  8. The empirical evidence and academic consensus are brutally clear:

    to remove unemployment NZ needs to only three things
    - eliminate minimum wage rates (the youth min wage is only a subset of the real problem, the adult min wage)
    - eliminiate the unions and other restrictive work practice legislation (notable the ERA & Holidays Act)
    - eliminiate the dole and other benefits

    Make those three changes and unemployment will disappear - literally overnight. The good news is that all these three commonsense policies are advocated in the 2025 taskforce, are now official ACT policy, and will be implemented in Don Brash's emergency budget this November

  9. @Anon. I'd put really good money against each part of that. First, I don't buy that Key would do it. Second, the claims of policy efficacy are way too strong.
    - Even if you abolished the youth minimum wage entirely tomorrow, there's no way youth unemployment would disappear. It would instead slowly retreat to a level somewhat below the historical trend prevailing through the 90s relative to the adult unemployment rate. For starters, wages are downward sticky - many employers would choose to shed staff rather than cut wages even if they would have hired folks on at a lower rate had the lower rate been available at time of hiring - motivational effects of pay cuts aren't good. So the action would be relatively slow and come as new jobs were created.
    - Existing workplace regulations in NZ can't have that huge of effect - we did manage to get unemployment down to very low levels around 2005/6 with those regulations in place. A lot of nonsense comes of them, but I can't imagine the aggregate effects are massive.
    - Making use of some of the policy evidence from America's experience with welfare reform would help, but there are lots of folks for whom employment will never work out. It would be pretty surprising if any government were to cut things that radically. I'd expect civil society to work through charity to help folks were it to happen, but I doubt we'd ever see those kinds of cuts.

    I'm really hoping that you're a leftie trying for a caricature of pro-market economist beliefs about the effects of policy. It's borderline crazy to expect that any policy changes could eliminate unemployment overnight.

  10. On the plus side, additional money in wages was presumably paid out to the youths who remain employed at the new higher minimum wage. Is it possible to figure out something like how many dollars were paid to youth workers per year before and after the rate change? Might be an interesting comparison.

    I think comparing climate change research to economic research is pretty cheeky. Your linked 'no effect' vs. 'large effect' studies show that the data does not always fit the theory due to unknown confounding factors. Whereas it is entirely possible to explain climate change data with fundamental theories (thermodynamics etc.), and there are no equivalent peer-reviewed papers finding that carbon emissions have zero effect. Also climate change and other 'green issues' (long-term energy prices, resource shortages etc.) might end up being far more consequential than the current lack of $6/hr jobs.

  11. @gileswrites: That's why the Canadian study is telling: increases in the minimum wage correlated with increased numbers of families below the poverty line. The gains to the folks who kept their jobs are nice, but the folks who lost their jobs were the second earners who kept their families above the poverty line.

    I'm pretty sure there are papers in climate change arguing that changes in human emissions are too small to account for observed changes in temperatures, just like plenty of papers find that observed changes in minimum wages account for very little of observed changes in employment. The consensus of the literature in climate science seems that carbon emissions matter, but there are dissenters. Just as there are dissenting papers on the effects of minimum wages.

  12. I fear you may be fighting an uphill battle here Eric. There seem to be some within the Green camp who are entirely idealogically driven and who cannot, or will not, accept that well-intentioned policy with a social conscience may have unintended negative impacts. Being a leftie myself I tend to be a supporter of many Green policies, but one thing I have noticed is that some Greens have a distrust of science, to the point of refusing to acknowledge the outcomes of research when it disagrees with their preconceived notions. Mind you, this isn't restricted to the left of the political spectrum, there are probably just as many scientifically retarded folk to the right. ;)

  13. Rimu, you clearly have not seen any of Meteria Turia's comments on economics. I cringe when I read them.