Friday, September 26, 2008

Fred Thompson on the Minimum Wage

Editor's Note: A thousand thanks to Fred Thompson and his contributions to this blog during a time when I have been unable to provide much content. Fred promises to take a breather for a while and, based on this contribution especially, I hope he doesn't really mean it. For what it is worth, I think this post pretty accurately describes mine and many other economists preference for extra-market transfers in lieu of market distorting taxes and rigidities (see: my rants abut the prohibition on self-service gas). Take it away Fred...


Last week, at Blue Oregon, Chuck Shekatoff of the Oregon Center for Public Policy (OCPP) accused the communications and lobbying firm Conkling Fiskum & McCormick (CFM) of slipping into “into Political BS (Bogus Statement) mode. The CFM article asserted ‘[h]owever, for the majority of that time [since 2002] the state unemployment rate has remained higher than the national average.’ The suggestion by CFM and business groups that a relatively high minimum wage and higher-than-average unemployment levels are related is Political BS.”

Shekatoff asserts that Oregon’s typically higher-than-national-average unemployment is due to population growth and the structure of our economy and cites an OCPP report and economists at the Oregon Employment Department in support of this claim. When, however, one checks the OCPP report (Who’s Getting Ahead) the evidence cited is short discussion of the issues by an Oregon Employment Department economist, Art Ayre, “Why Does Oregon Have a High Unemployment Rate?” published April 27, 2005. This is, indeed, a very nice discussion of the issues. It correctly observes that “economists usually speak of unemployment as having three components: frictional, cyclical, and structural. All three contribute to Oregon's higher-than-national rate.” However, while it is clear that Ayre has a keen grasp of the issues and the Oregon labor market. The evidentiary basis for his claims is simply not reported.

Indeed, it is hard to see how two of the factors he cites, the structure of the economy and population growth could explain Oregon’s RELATIVE unemployment when compared with the rest of the United States. The structure of Oregon’s economy is essentially a constant. It is axiomatic that a variable that doesn’t vary cannot explain any variance, although it might explain the intercept in an empirical model (which is how I understand Ayre’s claim, although I would stress that ‘might’ leaves a powerful lot of wiggle room). But if you take population growth as the independent variable in an empirical model and relative unemployment (or employment growth) as the dependent variable, one actually obtains a negative relationship. Now I don’t believe for a moment that population growth really causes unemployment to fall (or employment to grow). My hunch is that the causal arrow goes in the opposite direction.

The main driver of unemployment in Oregon is America’s business cycle. America’s booms and busts are also Oregon’s. But what explains Oregon’s RELATIVE unemployment. As a classroom exercise, I have had my students look at this issue several times over the past decade or so, using monthly data, the difference in unemployment (or employment growth) between Oregon and the US as the dependent variable and anything they could think of as independent (or causal) variables. The two variables that seem to explain Oregon’s relative unemployment best are the dollar’s exchange rate with other currencies and the state government's stop and start spending behavior.

The fact is that Oregon industry and agriculture are highly affected by foreign trade. When the dollar is low, they do well, at least so long as the rest of the world isn't in the tank. When it is high, they don't. This is a structural factor that varies measurably over time. Moreover, the state relies on a highly progressive tax structure. Progressive taxes are necessarily volatile revenue sources. Revenue volatility encourages spending volatility, making our booms and busts bigger than elsewhere. This is a cyclical factor unique to Oregon.

Early this year, I took a spreadsheet from the work of one of my better students and tossed a dummy variable representing Oregon’s adoption of a premium minimum wage into her model. The effect was statistically significant, as was the increase in the adjusted coefficient of determination. However, from the standpoint of relative unemployment (or the relative change in total employment), the negative effect was very small. This isn’t a very good test. It’s not a good econometric model. Moreover, using the change in Oregon’s minimum wage relative to the national average would have been better. But when I did the analysis, I lacked that data. Nevertheless, I think its results are likely to be correct. Oregon’s high minimum wage almost certainly makes relative unemployment worse, but the effect is probably not very big, at least not compared to other significant factors.

The reason I believe this is because it is consistent with what most other economists, who have looked closely and carefully at this issue, have found. My reading of contemporary research that the effect of high minimum wages on low-income employment is far more likely to be negative than neutral. Of course, employers can increase prices. But, other things equal, increased prices mean lower sales volume and fewer employees (otherwise, presumably, those employers would have already raised prices). High minimum wages also lead to rationing inefficiencies, which are probably more important than the job losses, but that is another story. One can recognize that high minimum wages have adverse effects and still support minimum wages, even high minimum wages. The consensus among labor economists is that minimum wages help many more low-wage workers than they hurt.

But its supporters shouldn't fool themselves that its effects are entirely benign. If anything is political BS, that is.

Obama proposes to raise the national minimum wage to $9.50 per hour in 2011 and index it to inflation. He also wants to increase the Earned Income Tax Credit (EITC) for working Americans with no children and for those with three or more children and a tax credit of up to $500 per person or $1,000 per couple. This would be a rebate of the worker’s Social Security contribution on the first $8,100 of earnings. Like the EITC, this helps low-income working families without creating employment disincentives.

Most economists believe that we could get more equality at a lower cost by focusing on bottom-end personal income tax brackets and expanding the EITC than by raising minimum wages. Consequently, I’d prefer it if Obama were less enthusiastic about increasing minimum wages and more committed to across-the-board increases in the EITC. Here in Oregon, one thing that we could do is to eliminate the first two brackets of the state personal income tax. Failing the better, however, we should probably, as Chuck Shekatoff proposes, celebrate the pretty good.

Disclosure, Gary Conkling of FSM is my colleague at the Atkinson Graduate School of Management.


Stevelle said...

A few quick notes:

What exactly does it mean to "eliminate the first two brackets of the state personal income tax"?

1)Reduce to 0% for those who fall into those brackets? (a tax cut for the low-income)

2)Expand the third bracket downward to include those currently in the bottom two brackets (a tax hike for the low-income)

3)Something else?

I presume the second, but want to be sure since I have stumbled on this phrasing more than once now.

Fred says:
"My reading of contemporary research that the effect of high minimum wages on low-income employment is far more likely to be negative than neutral."

While the effect on low-income employment may be more likely negative than neutral, the net effect is more likely to be positive than neutral. That is an important footnote.

Fred says:
"The structure of Oregon’s economy is essentially a constant."

then later says: "The fact is that Oregon industry and agriculture are highly affected by foreign trade. [...] This is a structural factor that varies measurably over time."

Well which is it? :)
Clearly the fact that Oregon's economy is driven by a little-varying set of market activity in a specific set of business sectors. That doesn't mean that the structure is fixed, so I'm not sure what was meant in the first point as the way I read this it seemed that I missed something.

Fred concludes:
"Most economists believe that we could get more equality at a lower cost by focusing on bottom-end personal income tax brackets and expanding the EITC than by raising minimum wages."

Were that efficiency were the only consideration, we would just expand the EITC and be done with it, as I understand it. Sadly, our world is messier and more complex than our models.

Thanks for sharing.

Fred Thompson said...

Stevelle askes "What exactly does it mean to "eliminate the first two brackets of the state personal income tax"?" The statutory rate for the first bracket is 5% and the second 7%. I am suggesting changing them both to zero, leaving only the 3rd bracket. That is like your option 1, except that it would apply to all taxpayers.

As for your second point, the late Ned Gramlich (no conservative) argued that the gains to low income workers from higher minimum wages were 2-4 times larger than losses to low income workers from higher minimum wages, which is good, but that its leaky-bucket ratio is quite high -- it costs society about $8-$10 bucks to increase the income of low-income families a buck using the minimum wage. If that were the only way do so, so be it. But it's not, we have alternatives with leaky-bucket ratios of close to 1, like the EITC.

Third, Oregon is a big exporter in international markets. That is a structural characteristic that doesn't change much from one decade to the next. But currency exchange rates change daily. Since most of the things we export are sold at world prices, this has a significant effect on the relative performance of the economy from one period to the next. The economy's structure doesn't change, but its response does. I could have been clearer when I wrote what I wrote, but I had Ayre's discussion in mind and implicitly presumed all my readers would too.

Fred Thompson said...

Since writing this I have read two articles that strengthen my conclusions:

Minimum wages and the economic well-being of single mothers

Joseph J. Sabia, Journal of Policy Analysis and Management, Volume 27 Issue 4, Pages 848 - 866

Using pooled cross-sectional data from the 1992 to 2005 March Current Population Survey (CPS), this study examines the relationship between minimum wage increases and the economic well-being of single mothers. Estimation results show that minimum wage increases were ineffective at reducing poverty among single mothers. Most working single mothers were not affected by minimum wage hikes because they already earned
wages above state and federal minimum wages. And less-educated single mothers who were affected did not see a rise in net income because of negative employment and hours effects. For this low-skilled population, a 10 percent increase in the minimum wage was associated with an 8.8 percent reduction in employment and an 11.8 percent reduction in annual hours worked.

Does the minimum wage affect welfare caseloads?

Marianne E. Page, Joanne Spetz, and Jane Millar, Journal of Policy Analysis and Management, Volume 24 Issue 2, Pages 273 - 295

Although minimum wages are advocated as a policy that will help the poor, few studies have examined their effect on poor families. This paper uses variation in minimum wages across states and over time to estimate the impact of minimum wage legislation on welfare caseloads. We find that the elasticity of the welfare caseload with respect
to the minimum wage is between 0.1 and 0.2, but this estimate is sensitive to the sample period and assumptions about state trends. We conclude that
higher minimum wages increase welfare dependence.

Torrid said...

doesn't the structural nature of the industries in OR have an impact? I'm thinking of the forests--actual and Silicon. Twice in the past 20 years, significant sectors of the state economy saw drastic downturns. We can argue about the reasons, but doesn't it suggest the idea that structure might indeed impact relative unemployment, in the same way an oil boom of bust changes thing in Texas?

Fred Thompson said...

To Torrid:

I don't think so. Hi-tech jobs have been increasing as a proportion of Oregon's jobs, but that is a national phenomenon. Fluctuations up and down have been mostly cyclical, largely driven by the business cycle and exchange rates.

The decline in forest jobs is another matter. There is evidence to suggest that the labor force adjustment caused by this change was both hard and protracted. But forest jobs have been in decline for a long time and their disappearance mostly predated the period in question.

Torrid said...

certainly the timber industry has had more impact than tech, yes. But I'd argue that the residual inability to rehire tech workers with white collar degrees, is less than for those doing skilled but manual work, typically without a fancy degree. Perhaps most are no longer in the labor force to drag down the numbers, but I think any look at Oregon's employment has to factor in the loss of the industry that sustained it for so long.