The Massachusetts Resource Curse?

In the field of the economics of development, one of the most fundamental questions is about the so-called resource curse . Basically, the idea is that countries with high endowments of natural resources tend to have worse governance and lower rates of economic growth (this thesis is not uncontroversial empirically, I should point out). The list of reasons why this might be true is long, but one of the primary problems with natural resources is countries endowed with them are spared the hard work of developing a prosperous market economy and can simply fund their activities with the sale of oil or diamonds.

We tend to think of this as a phenomenon that afflicts developing countries, but there's no reason that it couldn't apply to the United States just as well. To wit, Massachusetts lawmakers are now proposing a tax on college endowments:

Massachusetts lawmakers desperate for additional revenue are eyeing the endowments of deep-pocketed private colleges to bolster the state's coffers by more than $1 billion a year, asserting that the schools' rising fortunes undercut their nonprofit status.
more stories like this

Legislators have asked state finance officials to study a plan that would impose a 2.5 percent annual assessment on colleges with endowments over $1 billion, an amount now exceeded by nine Massachusetts institutions.

Now, I should start by pointing out that Massachusetts is one of the richest states in the United States, a fact much beloved of left-leaning economics commentators trumpeting the superiority of their system. The fact that they are in this kind of fiscal crisis is a disgrace.

But if you're a Massachusetts lawmaker, this is perfect. Universities have huge fixed and relatively unmovable assets (buildings, stadia, etc.). Greg Mankiw (jokingly?) suggests that Harvard move to sunnier fiscal climes:

1. Instead of expanding the university into Alston, Harvard could create a second campus in another state. Call it Harvard South. (Put it in a better climate than Boston, and I would be one of the first faculty to volunteer for the move.)

2. Transfer much of the endowment to Harvard South. Support Harvard North by slowly selling off land in Massachusetts.

3. Eventually, make Harvard South the main campus, and Harvard North the satellite. If Massachusetts state lawmakers remain hostile, close Harvard North down entirely.

Now, this isn't terribly likely to happen, though frankly I'd love to see it. When resources are fixed geographically, it's very difficult for governments to commit not to plunder. (Maybe Patri should try to get Harvard to move to a seastead.)

You see this elsewhere in America. New York can get away with crummy government service for decades because the financial industry is stuck there (one of the dark sides of economies of aggregation). I had high hopes a few years ago that the internet would help to break this trend, but I turned out to be mostly wrong, at least so far.

But nevertheless, even with fixed endowments, there's only so far the government can go. Felix Salmon claims that "Right-wingers always say that if you tax people or institutions with money they won't pay more tax, they'll just move; they're rarely proved right."

Well, maybe. I haven't seen any academic economic work on internal migration in the United States, but something sure as hell is driving people away from high tax states. Housing is probably the major driving force, but at least anecdotally, taxes are a contributing factor :

The San Francisco Business Times noted recently that Bay Area millionaires are starting to take their gold out of the Golden State. The paper writes: "The Bay Area's wealth boom is producing an explosion of millionaires--in Nevada, Wyoming and perhaps Canada. Advisers to the well-heeled say 'wealth migration'--taking the money and running--is behind a surprising drop in the number of Bay Area millionaires."

The annual loss of millionaires from the Bay Area--home to Silicon Valley--knocks this extremely rich and fertile place down near the bottom of the new millionaire list, putting it in the company of Detroit, Pittsburgh and Cleveland. Of course, the middle class has been fleeing California for years. High house prices have been the main culprit. Long commutes and deteriorating public services play a part. But the flight of millionaires is very 21st century. The cause? You guessed it. Taxes.

Continues the San Francisco Business Times: "'I'm hearing from more California baby boomers, I need to get out,' said Diane Kennedy, a Phoenix accountant and financial adviser to the wealthy. 'You can still make a lot of money in California. The problem is, then you have to pay taxes on that money,' said Kennedy, who recently helped a California client with annual income of about $1 million save $96,000 annually by making his home in Jackson Hole, Wyo. his primary residence.

Felix Salmon may think "people respond to incentives" is a right wing concept, but I don't. And even Harvard may find that when pushed far enough, it might have to go slumming down in Charlotte or Houston, like the rest of America has been doing at a high rate for years now.

Share this

Right. Of course "people

Right. Of course "people respond to incentives" is not exclusively right-wing. If that weren't the case, the left would be without a reason to prefer to democracy to monarchy. Presumably monarchs have no incentive to respond to the will of the people.

Incentives

"People don't respond to incentives" is a left-wing concept, so it's understandable that they might think that the opposite is a right-wing concept.

For the record...

...I don't think that "people respond to incentives" is a right-wing concept. I think it's very true. But at the same time it's worth looking empirically at what happens in reality, rather than just assuming ex ante what people will do. The minimum wage is a good example: right-wing economists tend to simply assume that raising the minimum wage, even a little bit, will increase the unemployment rate. While real-world economists look at the data, and find that it doesn't.

The problem is the reductio ad absurdum. If you raised the minimum wage to $50/hr, yes unemployment would rise. If you taxed the Harvard endowment at 50% of its wealth per year, yes it would move. But relatively small changes don't necessarily work that way.

Minimum wage hikes probably bad

The minimum wage is a good example: right-wing economists tend to simply assume that raising the minimum wage, even a little bit, will increase the unemployment rate. While real-world economists look at the data, and find that it doesn't.

Oh really. How about checking this at the Wikipedia article. That seems flatly to contradict your summary of the situation regarding the real-world evidence.

Anyway, the "assumption" is not "right wing". It's what any economics student can predict practically from day one, when he learns how to use the supply and demand curves to make predictions. The model is an idealization of course, but I don't know of any economist who does not recognize this.

It is also true that simple logic forces pretty much the same conclusion even if one abandons the assumptions of continuity, differentiability, and monotonicity of supply and demand (though the conclusion must be worded carefully). Specifically, everyone seems to agree that if the minimum wage were increased by fifty dollars, employment (especially of those whose lack of skills commands low wages) would go down (barring general inflation) - this is regardless of whether the minimum wage is increased all at once, or fifty times one dollar at a time. The conclusion is logically inescapable that if the minimum were increased by one dollar at a time, then on average, the effect of a dollar increase would be a reduction in employment. This probabilistic prediction can be expressed as follows: the expected result of an increase in the minimum wage, however small, is to reduce employment.

Now, it could well be that a particular increase in the minimum wage will not reduce employment. However, this is not the way to bet. This prediction is entirely compatible with data that show that some particular increase in the minimum wage does not (noticeably) reduce employment. That does not mean it's unfalsifiable: on the contrary, were it the case that a fifty-dollar increase in the minimum wage did not reduce employment, then the prediction would be falsified. The thing is, few would expect it to be falsified.

In brief, while it is possible that a minuscule increase in the minimum wage will not reduce unemployment, that is not the way to bet.

As far as data gathering, it's notoriously difficult reliably to come to any conclusion when small effects are being investigated. Anyone who has followed the news is aware of the unreliability of data gathering for drawing conclusions when it comes to elections. Polls have in recent memory badly failed not only to get the spread right, but even on occasion to predict the winner of an election. While data gathering is the only way to do empirical science, its limits need to be acknowledged. Specifically, we should not in general expect the data to reliably reveal small effects. One solution is to focus our efforts on gathering data to check predictions of large effects. The unreliability of the data gathering matters less the larger the effect is being checked. Interestingly, most people, even advocates of a minimum wage increase, are so sure that a fifty-dollar increase in the minimum wage will reduce employment, and therefore that the model is absolutely and inescapably right at this increment, that they don't suggest checking it. In short, at the fifty-dollar level, even the fans of the minimum wage hike don't insist that the evidence contradicts the theory.

It's tough to figure out the

It's tough to figure out the consensus within a discipline from outside. The people most likely to claim the consensus is on their side are the partisans, who are also most likely to be the outsiders' sources--whereas neutral parties are more likely to stay quiet.

I tended to doubt that modest increases have any effect on unemployment after Steven Landsburg admitted it in a Slate column (after stubbornly resisting the Card and Krueger findings in his first book). That's a weak data point but I've nothing better.

Then again, the first paper I found after googling around was this. There may be other papers reaching the opposite conclusion--I stopped after I found one. Quotes follow.

I also question the notion that belief in the minimum wage affecting employment is meaningfully a right-wing economists' concept. The implication is that empirical economics is the domain of leftist economists and theoretical economics is the domain of the right. Neglecting the Austrians, as everyone does, I've no reason to believe that true.

Quote:

• Almost three-fourths of labor economists (73%) believe that a mandated minimum wage increase set at 150% of the current wage would result in employment losses. Similarly, more than two-thirds of labor economists (68%) believe a mandated minimum wage would result in employers hiring more applicants with greater skills, and nearly one-third (31%) believe there would be no change in hiring practices. Figure 1 and 2

• Nearly half of labor economists (49%) believe a mandated minimum wage set at 150% of the current wage would lead to no change in poverty rates, 32% believe it will reduce poverty rates and 19% believe it will increase poverty rates. Figure 3

• Labor economists were asked to rate the efficiency of three proposed policies which address the income needs of poor families: a higher minimum wage, the Earned Income Tax Credit, and general welfare supports. Of these three options, the Earned Income Tax Credit is rated most efficient followed by general welfare supports. A higher minimum wage is judged least efficient. Economists’ ratings of the efficiency of welfare and the EITC did not change between 2000 and 2007; the minimum wage question was not asked in 2000. Figure 4

• More than half of labor economists (53%) rated the Earned Income Tax Credit as very efficient, another 42% believe it is somewhat efficient, and only 5% think it is not at all efficient.

• General welfare grants are rated very efficient by 12% of labor economists, 67% believe they are somewhat efficient, and 21% think they are not at all efficient.

• Only 6% of labor economists believe a higher minimum wage is a very efficient way to address the income needs of poor families, 39% think it is somewhat efficient, and 55% think it is not at all efficient.

• Not surprisingly, when asked which of these three best addresses the income needs of poor families, 70% said an expanded EITC, 21% said general welfare supports, and only 9% said a higher minimum wage. Figure 5

• Labor economists were asked to estimate the effect a higher minimum wage (150% of the current level) and an expanded EITC would have on employment. Of these two options, economists believe an expanded EITC would lead to employment gains, and that a higher minimum wage would result in employment losses. Figure 6

• Nearly two-thirds of labor economists (64%) say an EITC would lead to employment gains, another 34% believe it would lead to no change in employment, and only 2% believe there would be employment losses.

• In contrast, only 6% of labor economists believe an increased minimum wage would lead to employment gains, 29% believe it would lead to no change in employment, and 65% believe there would be employment losses.

• Labor economists were asked what type of effect minimum wage mandates would have if they were set at the local or state level rather than at the national level. Of these two scenarios, economists predict larger minimum wage effects when set at the state or local level rather than at the national level. Figure 7

• Nearly half of labor economists (49%) believe minimum wage effects to be larger if set at the state level rather than the national level, 29% believe there is no difference, and 22% believe the effect will be smaller.

• The majority of labor economists (61%) believe minimum wage effects will be larger if set at the local level rather than the national level, 18% believe there is no difference, and 21% believe the effect will be smaller.

• More than eight in ten labor economists strongly oppose using a family of three as the standard for setting hourly minimum wage levels. Economists are also strongly opposed to using a family of four as the standard for setting minimum wage levels. There was no change in economists’ opinion on this issue between 2000 and 2007. Figure 8

The dieter's delusion

I tended to doubt that modest increases have any effect on unemployment after Steven Landsburg admitted it in a Slate column (after stubbornly resisting the Card and Krueger findings in his first book).

As a general statement, this cannot be true. A large increase is nothing other than an aggregation of small increases, so if it were generally true that a small increase has no effect, then it would follow by mathematical induction that a large increase has no effect. Dieters would love this to be the case, because then they could simply chop up their food into sufficiently small bits that each bit has no effect on their weight.

And it's not what Landsburg wrote, if I have the article right. He wrote:

Now that we've re-evaluated the evidence with all this in mind, here's what most labor economists believe: The minimum wage kills very few jobs, and the jobs it kills were lousy jobs anyway. It is almost impossible to maintain the old argument that minimum wages are bad for minimum-wage workers.

He's saying that a modest increase has a modest effect on unemployment, and he's giving an evaluation of it ("lousy jobs anyway").

As a general statement, this

As a general statement, this cannot be true. A large increase is nothing other than an aggregation of small increases, so if it were generally true that a small increase has no effect, then it would follow by mathematical induction that a large increase has no effect.

I'm not sure what you mean by "as a general statement," but certainly your statement's not true in all cases. Trace amounts of arsenic over time can have a different effect than the sum total amount of those trace amounts given in one dose. Or consider (perhaps more appropriately, as it deals with an irrational agent's response) the frog in the boiling pot. A sudden large increase has the effect of making the frog jump out, but the same increase, spread over time has no such effect (save for boiling the frog). It's perfectly possible that small lifts in the minimum wage spread over time could have no effect on employment in the same fashion.

Perhaps I'm not grasping your meaning here.

He's saying that a modest increase has a modest effect on unemployment, and he's giving an evaluation of it ("lousy jobs anyway").

With the reservation that your "modest effect" strikes me as having a connotation of a larger effect than Landsburg's "kills very few jobs," you're right and I am wrong, and have misremembered. NB: he also describes the effect as "at most a tiny impact on employment." Landsburg said tiny, and I rounded it down to zero in my brain.

Feel free to read my prior comment replacing "no effect" with "kills very few jobs."

I did use "modest increases," which is as as ambiguous as your usage. So feel free to replace "modest increases" with "increases of the size described by Landsburg in his column," as that's all I meant to convey.

Gradual increase included

I had written (in the comment you were replying to):

Specifically, everyone seems to agree that if the minimum wage were increased by fifty dollars, employment (especially of those whose lack of skills commands low wages) would go down (barring general inflation) - this is regardless of whether the minimum wage is increased all at once, or fifty times one dollar at a time

which covers your scenario of gradual increase. To apply an analogy you mention, the loss of jobs is more like the cooking of the frog than like it jumping out. A minimum wage hike is a price control, and price controls do harm whether or not they are imposed suddenly or gradually. The standard explanation of the harm they cause does not depend on their sudden imposition.

No linear response curve?

A large increase is nothing other than an aggregation of small increases, so if it were generally true that a small increase has no effect, then it would follow by mathematical induction that a large increase has no effect.

This assumes a linear response curve; unemployment and minimum wage could very well follow something akin to an advertising response cuve (famous for its S-shape), where each incremental advertising dollar spent yields little in sales until the ad budget size reaches a certain threshold (at which point it becomes exponential--past the 2nd threshold, you start seeing diminishing returns). Similarly, minimum wage increases might have little affect on unemployment until they pass some threshold.

It doesn't assume that

This assumes a linear response curve

Actually, it doesn't. Start with the minimum wage at $7, and consider every one-dollar increment, i.e., from 7 to 8, from 8 to 9. To say that it is generally true that a modest increase has no effect on unemployment is to say that for each of $7, $8, $9, etc., a modest increase has no effect on unemployment. If there is no effect from 7 to 8, no effect from 8 to 9, no effect from 9 to 10, and so on, then there is no effect from 7 to 57.

Similarly, minimum wage increases might have little affect on unemployment until they pass some threshold.

Indeed, but if there is a threshold past which a statement is false, then it is not generally true.

Keep in mind what I have acknowledged already. I have acknowledged:

Now, it could well be that a particular increase in the minimum wage will not reduce employment.

and

This prediction is entirely compatible with data that show that some particular increase in the minimum wage does not (noticeably) reduce employment.

And with your mention of thresholds this is all that you are saying: that for some particular range (below the threshold), there might be little or no effect. My point is that it's not the way to bet.

Similarly, someone might leave Las Vegas richer than he came. My point isn't that he definitely won't. My point is that it's not the way to bet, given that on average, people leave poorer than they came.

Adding to the discussion

The minimum wage is a good example: right-wing economists tend to simply assume that raising the minimum wage, even a little bit, will increase the unemployment rate. While real-world economists look at the data, and find that it doesn't.

I'm not an economist, but IMO, the data in these experiments is flawed by one clause that's missing from the statement above: "all else being equal".

All else is never equal in the real world, and thus, no accurate controls are available. Regions have immigration, emigration, sprouting and demise of industries, better or worse roads, change of politicians, economic growth, etc. So these minor tweaks in the minimum wage aren't really measuring much of anything. There's no isolated independent variable. It's no surprise that the result is "a tweak in the minimum wage has little effect on employment". The marginal effect is too small to measure.

I'm not a labor economist

I'm not a labor economist (and haven't ever studied labor seriously, even at the undergrad level) so I can't claim to be any sort of expert about the minimum wage. But my understanding is that Jonathan is right: The majority of the literature supports the proposition that higher minimum wages raise unemployment, but the effect is small and usually swamped by other factors.

Generally speaking, everyone arguing the converse tends to cite Card and Krueger almost exclusively. Those two are top notch economists without a doubt, but at a minimum their study is controversial in the economics profession.

Neither am I...

...but you can get a bit more precise and have empirical support. The minimum wage raises unemployment among those who typically make the minimum wage, which is a category teenagers disproportionately fall into. There used to be a nifty chart on some government site from the debate over Clinton's increase to the minimum wage which plotted the real minimum wage and the unemployment rate among teenagers on the same graph. The correlation was unmistakable. However, google has failed me, so instead I give you this. Notice the second bullet point.