Why the World Should Care About America's Middle Class

Tim Worstall, in his Forbes blog, attacks my series (here and here) on whether globalization is good for America's middle class. Not on the basis that he disagrees with my conclusion (though he does), but because, he argues, there are much more important facts about globalization than a decline in the economic well-being of the middle class in America and Europe. In particular, he points to the great decline in poverty among developing nations that have embraced globalization:

This growth in incomes, in wealth, has been uneven, this is true. Largely speaking those places which have been taking part in globalisation, Indonesia, China, India, have been getting richer. Those that have not been, Somalia perhaps as an example, have not been.

 Let's leave aside the fact that these successful countries are hardly poster children for the kinds of so-called "free-market" policies that Worstall espouses, a point made particularly well by Dani Rodrik. And in the spirit in which Worstall granted my claims for the sake of argument, let's grant his as well. (But if you want to get down into the weeds on the extent to which poverty reduction claims may be overstated, take a look at Robert Wade's work.)

Here is the crux of Worstall's argument:

So I would actually posit that whether the American, or European, or rich world, middle class benefits from globalisation is actually an incomplete question. Incomplete enough to be the wrong question. Almost to the point that the answer is “who cares?”.

The correct question is what is the distribution of all of the costs and all of the benefits of globalisation? To which my answer would be that a generation, perhaps even two generations, of stagnating lifestyles for the already rich, those middle classes, looks like a reasonable enough cost to pay for the other thing that is happening: the abolition of absolute human poverty in the rest of the world.

First, I think we should certainly care when hundreds of millions of people are suffering unnecessarily. Yes, unnecessarily, because contrary to Worstall's claim, we are not trading off reduced economic well-being for hundreds of millions of middle class people for the lessened poverty of billions of other people. Indeed, the two are happening simultaneously, but as Ronald Rogowki pointed out in Commerce and Coalitions, it is perfectly feasible to have rich country winners compensate rich-country losers and still have all of them be better off from trade. Politically, it is a hard row to how, as Rogowki pointed out: the winners from expanding trade increase their political power as a result of their increased income, making compensatory policies less likely. But ending globalization's harm to the middle class in rich nations does not require us to take anything away from poorer people, not if you accept the theory of comparative advantage and the Stolper-Samuelson Theorem. It does require us to figure out a political solution to the problems faced by the losers, which as we can see in the United States is made more difficult by the decline of unions and by the Citizens United Supreme Court decision.

And second, we should care about the U.S. middle class (and Europe's, for that matter) because how they react to their situation politically will have enormous consequences for the world economy and world politics. If the U.S. comes up with a "Smoot-Hawley" response to its economic problems, that would undo a lot of the gains Worstall sees as flowing from globalization, a point made recently by Dani Rodrik (via Mark Thoma). Even more ominously, in both the U.S. and Europe, we see increasing political polarization and the rise of nationalist political parties and movements, as noted by Paul Krugman. Economic decline is a scary thing, and people's reactions to it can get downright ugly, to put it mildly.

For both of these reasons, then, what happens to the middle class in the U.S. and Europe will have repercussions far beyond those acknowledged by Worstall.

Is Globalization Good for America's Middle Class? Part 2

In Part 1, I examined what economic theory has to say about the winners and losers from trade. The main conclusion is based on the Stolper-Samuelson Theorem: Because the United States is labor scarce in a global perspective, an expansion of trade will reduce the real wages of labor. As we have seen, this theoretical prediction has been borne out as real wages remain below their peak level for the 39th year running.

In this post, I analyze what I consider to be the other main element of globalization, the expansion of the mobility of capital. Just as transportation innovation and cost declines made trade easier, they also make it easier for owners of capital to locate it in a broader range of places than 30 or 40 years ago. Similarly, the decline in communication costs make it easier for owners of capital to coordinate production on a global scale as well as offering additional ways of moving financial capital (think tax havens).

Note that I have said nothing about actual movements of capital. Simply the ability to move capital strengthens capital owners in their negotiations with business and labor, because it makes the threat of moving credible and thereby gives companies greater bargaining power. Kate Bronfenbrenner showed clearly that after the passage of the North American Free Trade Agreement (NAFTA) in 1993, companies more frequently resorted to threats in their bargaining with workers, even to the point of violating the National Labor Relations Act by threatening to move during union organizing drives. In this blog, I have previously discussed the case of Boeing's establishment of a Dreamliner plant in South Carolina and admitting it was due to workers in Washington state exercising their right to strike, a form of retaliation that was a prima facie violation of the Act.

Similarly, we have seen how companies have used the threat of relocation to extract subsidies from state and local governments. Sears, with its $275 million (nominal) retention package from Illinois, is just the most egregious in recent years. That package alone could support 550 state jobs at $50,000 a year for 10 years (assuming no raises, something pretty common for state workers lately though unlikely to last 10 years). And remember, Sears did this in 1989 as well, when it got $178 million not to move out of state.

More generally, who should win and who should lose from the growth of capital mobility? One possibility is that it would simply speed up the effects of trade. If Mexico had needed to wait for the growth of domestic entrepreneurs, it could not have expanded its exports to the U.S. nearly as rapidly as in the actual situation where U.S. companies could provide the money. In that case, we would simply expect the effect of heightened capital mobility to be the same as the Stolper-Samuelson Theorem.

But this would not explain why European labor appears as opposed to globalization as U.S. unions. Western Europe is labor abundant, so we would expect western European worker to benefit from the expansion of trade. Yet one does not have to look hard at all to see that European unions are not in love with globalization. The right answer now might be that those who are mobile win in the global economy, while those who are immobile lose. While capital is mobile geographically, governments are bound to their location. Workers, even where they have significant legal opportunities to move, as in the European Union, are still restricted in their mobility by their language abilities or lack thereof, and by the common desire to live near their families (another way in which corporations are not people, by the way). And it is not as if European capital can only be invested in the EU.

This is consistent with studies of the effect on home country labor of foreign investment (see Richard Caves' Multinational Enterprise and Economic Analysis): reduced employment because exports are replaced with foreign production, some possible increased employment due to supplying goods and services to foreign subsidiaries, but at best the result is a wash and more likely the net effect is negative.

If this is right, U.S. workers may have the worst of both worlds: they are harmed by expanding trade, and they are harmed by being less mobile than capital. While this does not explain the political changes that have happened in the U.S. since 1970 (though it is certainly relevant), it gives us a pretty good handle on the economic market pressures that the middle class needs to address politically. I will have more to say about these issues in future posts.

Is Globalization Good for America's Middle Class? Part 1

In this blog, I have frequently documented economic trends that have been bad for the middle class: Declining real wages, steadily falling bang for the healthcare buck, stagnant educational attainment, the gigantic cost of tax havens, etc. With this post, I want to begin exploring one possible reason for the economic insecurity of the middle class, namely globalization. Today, we will look at who wins and who loses from international trade, one of the key elements of globalization.

In some circles, one is likely to see a variant of the claim that "everybody" is better off because of freer trade. Even according to the most mainstream economic theory, this is simply false. The workhorse theory for determining the distributional effects of trade (i.e., who wins and who loses) is called the Stolper-Samuelson Theorem, first enunciated in an article by Wolfgang Stolper and Paul Samuelson in 1941.

To understand this theory, you need to know that economists think about national economies in terms of the amount of land, labor, and capital they have compared to all other countries in the world. These "factors of production" can be in relatively high supply compared to the rest of the world, in which case they are referred to as "abundant," or in relatively low supply compared to the rest of the world, in which case we call them "scarce."

The theorem can be stated in quite simple terms, but its consequences are not at all simple: As trade expands, owners of abundant factors of production benefit, and owners of scarce factors of production are harmed. Here, "benefit" means their real income increases, while "harmed" means their real income decreases.

Remember, trade can expand for two main reasons. First technological innovations can reduce the cost of transportation, making it first possible, then cheaper, to send goods long distances. For example, political scientist Ronald Rogowski, in his great book Commerce and Coalitions shows how the introduction of the steamboat made it possible to export North American wheat to Western Europe, displacing wheat from Eastern Europe. Second, policy changes like the North American Free Trade Agreement (NAFTA) or the trade agreements embodying the World Trade Organization (WTO) reduce or eliminate costly barriers to trade and lead to its expansion.

The grain example helps show why trade creates winners and losers. The Midwest U.S. and Canadian Prairie provinces are a gigantic breadbasket made possible by low population density, which implies abundant land and scarce labor. Expanding trade gave these farmers new markets and higher incomes. In much more densely populated Europe, the reverse is true: labor is abundant and land is scarce. As a result, expanding trade in grains meant more import competition and lower income for European farmers..

Fast forward to today and we can ask what U.S. factor endowments are currently. As a rich country internationally, the United States is necessarily a capital abundant country. As a comparatively low population density country, it is land abundant but labor scarce. The answer is to our initial question is then quite clear: expanding trade is harmful to U.S. workers because imports of labor-intensive products and services from abroad create competition for American workers, reducing their real wages. As I have discussed before, U.S. real wages have remained below their peak for 39 straight years, just as the Stolper-Samuelson Theorem would predict.

What about all the cheap goods we now buy at Wal-Mart? It doesn't change this story at all, because the lower price of imported goods is already reflected in the inflation rate we use to calculate real wages.

Rogowski's book also argues that we can expect certain pattens of political coalitions to form, with the winners from trade on one side and the losers on the other. NAFTA illustrated this well, with capital and agriculture generally in favor of the agreement (minus a few small specialty agricultural products like oranges), while labor was strongly opposed. And of course, this only helps us understand economic reasons for support or opposition to trade agreements; for non-economic reasons such as the environment, we have to look elsewhere. Although beyond the scope of this post, Rogowski's analysis of the entire world through phases of rising and falling trade (i.e., the Great Depression) lends strong credence to his claims. You should definitely read his book sometime.

Economists are divided over how big this effect is. In the 1990s, when I first started teaching, the most common view of economists was that technological change was the driver increasing the premium for high skilled labor while reducing wages for low-skilled labor. Adrian Wood's 1994 book, North-South Trade, Employment, and Inequality, argued that trade was in fact the main culprit, (a good, ungated analysis is  Richard Freeman's "Are Your Wages Set in Beijing?"). Although this met with a lot of resistance at the time, Wood's view has gained a lot of traction among economists based on developments over the last 15 or so years. Paul Krugman, a particularly noteworthy example due to his Nobel prize, has gone from being a fanatic adherent of free trade to someone who sees trade as a big problem, though even today he is not quite willing to pull the plug on free trade.

One important point Rogowski makes (and Stolper and Samuelson did before him) is that the theory of comparative advantage tells us that the winners from trade gain more than the losers lose, which makes it possible in principle to compensate the losers and have everyone be better off. But he also argued that those who benefit economically from trade will see their political power increase, something that has certainly been borne out in the United States in the more than 20 years since his book was published. This makes it less likely that such compensation will occur, and we certainly haven't seen any policy in the U.S. that comes close to making everyone better off as a result of trade.

One small bit of comfort comes from Paul Krugman's book The Conscience of a Liberal (pp. 262-3). He provides us some reason to think that the Stolper-Samuelson Theorem isn't necessarily destiny, as he shows that the United State and Canada, two countries with the same factor endowments as each other, have distinctive differences in political outcomes, particularly with regard to unionization rates.

Overall, unfortunately, it looks like the answer to today's question is clear: freer trade has harmed, and is harming, the American middle class. But globalization is more than trade, and I will continue to analyze other elements of globalization in my next few posts.

New Report Highlights Flaws of North Carolina Mega-Incentives

My new report for the North Carolina Budget and Tax Center, Special Deals, Special Problems--An Analysis of North Carolina's Legislature-Approved Economic Development Incentives, has just been published. It covers a range of issues I've emphasized here before as well as some basic considerations reporters really need to pay more attention to.

North Carolina has some of the best economic development practices in the country, in terms of online transparency, performance requirements, use of clawbacks for non-performance by companies, sunset clauses for tax expenditures, hard caps for many tax credit programs (see my report on these points), etc. The state publishes an economic development inventory I consider to be of very high quality and consistent with international definitions of a subsidy. The most recent edition shows that in the 2008-9 fiscal year the state spent about $1.2 billion on economic development, enough to hire 24,000 people at $50,000 a year in wages and benefits.

At the same time, however, the state has persistently had problems in overvaluing potential investments and consequently offering wildly excessive subsidies for them. The best known case is Dell in 2004, when Virginia offered the company a $37 million incentive package, while the state and local bid from North Carolina came to almost $300 million on a nominal basis ($174 million present value). Other deals discussed in the report are Google ($260 million nominal value, $140 million present value), Apple ($321 million over 30 years nominal value, no present value calculation available), and a provision in a 2011 special incentives bill to allow Alex Lee Inc. to keep $2 million it should have forfeited for not keeping job promises. This last case illustrates how special legislative deals weaken the state's performance requirements; this case will make future companies think that there may be no penalty for non-performance.

Reporters take note! This publication describes useful techniques for comparing the size of incentive packages regardless of project size or payout period of the incentive. From the European Union I borrow the term "aid intensity," which measures the size of the incentive relative to the amount of the investment or the number of jobs created. The idea is that a $1 million incentive would be large for a call center but a rounding error for an automobile assembly plant. As a result, we need a standardized way of comparing incentives.

While in this country one can sometimes find cost per job analyzed for some subsidy packages, the EU actually uses the subsidy/investment metric as its primary measure of aid intensity. In my last post I discussed a mall redevelopment which could conceivably have an aid intensity of 96%. For comparison purposes, we should note that the highest aid intensity allowed for large firms anywhere in the European Union, is 50%, and that is only allowed in the poorest regions of the EU, mainly in eastern Europe. (Richer regions have lower allowable maxima.) A region's maximum is cut by half for large projects over 50 million euro, and by 66% for spending over 100 million euro.

The other important concept is present value, a familiar one to accountants and economists, but not widely understood among the general public. The basic idea is simple: receiving a dollar today is worth more than receiving a dollar next year, which is worth more than receiving a dollar in two years, etc. Since incentive packages can pay out immediately (with a cash grant) or over a period of 30 or more years, we need to use present value to properly compare the size of incentives with different payout periods. This requires finding a a "discount rate" by which to reduce future payments. We then use the present value as the numerator in calculating aid intensity to be able to compare across different sizes of projects.

Using Google as an example, this $600 million project will receive $260 million over 30 years and create 210 jobs. As mentioned above, this is its nominal cost, before discounting the future dollars. Following the practice of a 1990s study by the Organization for Economic Cooperation and Development to compare subsidies among its then 23 members, I used a discount rate equal to the 10-year Treasury bond yield to come up with a present value of $140.6 million. Then the aid intensity is $140.6 million/$600 million, or 23%, and the cost per job at present value is $669,489. We can then use these two measures of aid intensity to compare the incentive to that given for other projects and inform our judgment of whether it was a better or worse deal than other states have made, in the current context where states make such deals all the time. Of course, I believe there should be limits placed on state and local governments so we can sharply reduce net incentive spending, which has few national benefits--but that is a long time in the future.

North Carolina provides an intriguing case study because it does so much right in economic development, but it makes special deals outside its statutory incentive programs. The result is high costs and weakened bargaining position in the future. It's a case we can learn a lot from.

How bad can "corporate socialism" get? David Cay Johnston gives us a glimpse

David Cay Johnston has a new column up today showing us some of the worst outcomes from corporate subsidies: incentives for retail development.

Johnston analyzes the case of a proposed redevelopment of the nearly-shuttered Medley Centre Mall in Irondequoit, Monroe County, New York, where developer Scott Congel is seeking a $250 million sales tax TIF. Originally a $260 million project, Congel now says he will invest $750 million to build a hotel and condos as well.

Johnston points out that retail is practically the worst thing government can subsidize, because it is almost entirely derivative of a region's population and income. If income falls (actually, even if it stays the same), the new mall can only succeed if it takes sales away from other existing outlets. And it's even worse than Johnston says, because retail jobs tend to have relatively poor pay and benefits.

The best study on this subject was conducted by the East-West Gateway Council of Governments, which is the regional planning agency for the St. Louis metropolitan area. This report found that from 1990 to 2007, local governments in the region had spent $2 billion in retail subsidies, repeatedly shifting the location of sales but generating no tax growth beyond the area's income growth. This comes to an astonishing $370,000 per net job if we believe that the incentives created the jobs, which is unlikely since sales growth did not exceed regional income growth. The cost per job is actually infinite.

Johnston points out more outrageous aspects to the Medley Centre proposal. Instead of conducting its own analysis of the economic impact of mall redevelopment, the developer commissioned and paid for a report, which "found" that the subsidized mall would see its sales grow from $30 million a year to $420 million per year. Johnston, by contrast, found that real income had fallen by $2.5 billion (13%) from 2000 to 2008 in Monroe County, and rightly argues that it makes the 14-fold increase in sales predicted "unlikely." Similarly, Johnston found that hotel demand in Monroe County had been flat for two decades.

Johnston also reviewed building costs for condos and hotels, finding that the $750 million price tag was "wildly inflated." As he points out, even if the figure was right, the subsidy has an aid intensity (subsidy divided by investment) of 1/3, whereas if the investment is only $260 million as earlier promised, the aid intensity comes to 96%. This would rival the 98% aid intensity on the Electrolux manufacturing plant being built in Memphis, itself an absurd level of subsidization, but at least for manufacturing rather than retail jobs.

The Medley Centre project has not yet received final approval, but it is an excellent example of all that is wrong with subsidized retail development. It won't create any new net jobs, the jobs at the mall will be low quality, the economic "analysis" was bought and paid for by the beneficiary, and the cost will be outrageously high. Let's hope there is some way it may yet be stopped.

Basics: U.S. One of Worst Rich Countries for Child Poverty

Pat Garofolo of Think Progress reports on new data published by UNICEF on child poverty among rich countries. The study covered all 27 European Union members, plus Australia, Canada, Iceland, Japan, New Zealand, Norway, Switzerland, and the United States. Of the 35 countries studied, the U.S. ranks second-worst in the percentage of children living in relative poverty (half of median GDP per capita; see further below), with 23.1% below this poverty level, above only Romania at 25.5%. Romania is the second poorest member of the EU and not an OECD member country. The table below sums up the results (most data are for 2009):


Source: UNICEF via Think Progress

As the report says:

Previous reports in this series have shown that failure to protect children from poverty is one of the most costly mistakes a society can make....The economic argument, in anything but the shortest term, is therefore heavily on the side of protecting children from poverty. Even more important is the argument in principle. Because children have only one opportunity to develop normally in mind and body, the commitment to protection from poverty must be upheld in good times and in bad. A society that fails to maintain that commitment, even in difficult economic times, is a society that is failing its most vulnerable citizens and storing up intractable social and economic problems for the years immediately ahead.

Some people object to the use of this relative poverty measure, which is the OECD standard. Certainly, for the poorer EU Member States, their absolute levels of child deprivation are worse than that in the U.S. because their income per capita is so much lower. For example, Romania's GDP per capita is $12,300 at purchasing power parity (PPP) compared to $48,100 for the United States. The UNICEF report actually has an absolute measure of child deprivation based on lack of access to two or more of 14 resources it estimates are essential for children in an industrialized society (including everything from three meals a day to a quiet place to do homework to an Internet connection). 72.6% of Romania's children and 56.6% of Bulgaria's are deprived by this standard, but only four more EU members exceed 20% by this measure. (Comparable data were not available for the U.S.)

But we in America should not get too excited by this fact. If we compare poverty across the major OECD economies, we find that the rankings change very little whether we use the OECD's relative measure or an absolute measure. We know this because for a time the UN Development Programme's Human Development Report listed data for an absolute poverty threshold of $11/day in 1994-5, which comes to $16,060 per year for four people, little different than the Census Bureau's figure of $15,569 for a family of four in 1995. So, measuring major European economies plus Australia and Canada, which generally have a lower GDP per capita at PPP than the U.S. does, against the U.S. poverty line, what do we find? From the 2006 Human Development Report, , page 295, here are all the "high human development" countries with poverty data using both the relative and absolute scales (listed in order of their Human Development Index score):

Country          50% of Median Income Rate          $11 a Day Poverty Rate
                                   1994-2002                                1994-95

Norway                          6.4%                                         4.3%
Australia                        14.3%                                       17.6%
Sweden                           6.5%                                         6.3%
Canada                          11.4%                                         7.4%
United States                  17.0%                                      13.6%
Netherlands                      7.3%                                        7.1%
Finland                             5.4%                                         4.8%
Luxembourg                     6.0%                                         0.3%
France                             8.0%                                         9.9%
United Kingdom             12.4%                                       15.7%
Germany                          8.3%                                         7.3%

As we can see, the shift to the absolute rate improves the U.S. rank from only 11th (last) to 9th. As long as restrict ourselves to the richest of rich countries, it makes little difference whether we use a relative or absolute measure of poverty. Either way, the U.S. does very poorly. And because it does poorly in overall poverty, it does poorly in child poverty as well.

The U.S., with high levels of child poverty, is therefore setting itself up for permanently lower economic productivity, higher costs in social services and incarceration, and so forth. This is a powerful argument against cuts to safety net programs and to education. The only question is whether we can overcome the forces currently promoting such policies.

Basics: How Overrepresented Are Rural and Low-Population States?

We all kinda sorta know it: rural and small states are overrepresented in the Senate and, to a lesser extent, the Electoral College.This has deep roots in American history, of course: when the United States Constitution was drafted, small states demanded the Senate, with two votes for every state, to guarantee they would not be overwhelmed by the larger states politically. But today, when we have much greater population differences among states than in 1787, this takes on much more anti-democratic significance than it did then. Because each state has two Senators, political changes favoring the middle class are much harder to achieve than if everyone in the country were equally represented, in a mathematical sense, in Congress. Moreover, with the existence of the filibuster (recently challenged in court by Common Cause), the effect of this overrepresentation is substantially magnified. But how big is the effect after the 2010 Census?

Under the Senate's filibuster rules, 41 Senators can block debate on Senate bills and nomination confirmations. So the first question is what percentage of the 50 states' population do the 21 smallest states have. The 2010 Census showed the states to have 308.1 million (all quoted figures are subject to slight rounding error) population, with the smallest 21, from Wyoming's 564,000 to Iowa's 3 million, having a total of 34.8 million, or just 11.3% of the 50-state population. In theory, Senators representing those states could mount a successful filibuster. Of course, this is unrealistic, since some small states are heavily Democratic, such as Vermont, Rhode Island, Hawaii, and Delaware. Even Montana currently has two Democratic Senators.

Another way to look at the filibuster is to ask what percentage of the 50-state population is represented by the 41 Republican Senators from the least populous states. The answer takes the actual population of states with any Republican Senators, except Texas (Cornyn and Hutchison), Florida (Rubio), Illinois (Kirk), Pennsylvania (Toomey), and Ohio (Portman).  The population of the states represented by the other 41 Republican Senators is 104.7 million, or 34.0% of the population of the 50 states. Thus, states with just a third of the country's population can block legislation or Presidential nominations. With the recent skyrocketing use of the filibuster in the Senate, this is profoundly undemocratic.

Turning to the Electoral College, we can again see the effect of having a minimum of two Senators regardless of population, which means that each state (and the District of Columbia) has a minimum of three electors in the Electoral College. For example, the Real Clear Politics Electoral College map lists just 11 states and the District of Columbia as likely Obama, whereas 17 states are likely Romney. Even though the likely Obama states have more electoral votes than the likely Romney states (161 to 131), 6 of the Democratic states have double-digit  electoral votes whereas only two of the Republican states do, underlining how Romney benefits from the overrepresentation of rural states.

Finally, remembering the 2000 election, where President Bush was awarded more electoral votes despite losing the popular vote nationally, we can ask what the minimum percentage of population for the 50 states plus DC is needed to win the Electoral College. To answer this question, I tallied from the bottom to see how many states were required to top 270 electoral votes. According to Wikipedia (as I tell my students, only a potentially reliable source for non-controversial information, like this), you have to have New Jersey to top 270, but it actually takes you to 282. So I subtracted three Democratic states (DE, VT, and DC) with 3 electoral votes as well as Montana's 3 electoral votes (since it's the most competitive of the remaining states with 3 EVs) to get down to 270. The 37 remaining states have only 45% of the nation's population eligible to elect the President. Yet theoretically they could do just that.

This post has merely scratched the surface of the deep historical and constitutional questions that have led to Wyoming's 564,000 people having as many Senators as California's 37.7 million. The rural bias of the Senate and Electoral College make major political changes difficult to achieve, yet it is even more difficult to imagine that they could possibly be fundamentally altered, especially the Senate. Still, it is worth reflecting on these imbalances in order to understand the shortcomings that exist in American democracy.