World 3.0

by Pankaj Ghemawat

  Current account balance: −5% of GDP

  Trade competitiveness: Competition versus countries with lower labor costs and other cost savings such as lower health care costs (and trade deficit with China elevated by its undervalued currency).

  Developed capital markets: Attraction of U.S. capital markets (until recently perceived as most advanced, secure, well regulated).

  Reserve currency: U.S. dollar status as reserve currency attracts capital inflows.

  Note: Potential causes shown in [brackets] in tables 8-1 and 8-2 reduce the size of current account imbalances or otherwise make them less problematic.

  Source: The percentage of GDP figures are 2007 data from World Bank, World Development Indicators, with savings distributed among sources based on the ratios (but not the exact amounts) shown in Goldman Sachs Global Economics Paper 191, “China's Savings Rate and Its Long-Term Outlook,” October 16, 2009. Potential causes drawn from a variety of sources.

  According to Lindsey, then, there is no market failure or fear to worry about here, and nothing to be gained economically, from G20 get-togethers. This emphasis on integration-without-intervention clearly aligns with World 2.0. The question here is whether that worldview provides a good perspective on imbalances.

  Models and Messages

  Since imbalances do not show up on the standard list of market failures, the case that market-based processes can lead to problems in this regard needs to be made. This section does so by considering a suite of simple theoretical models of growth that exclude all the standard sorts of market failures to make the point that it is nonetheless possible for market processes to lead to problematic imbalances. Note that this approach of using a theoretical model to illustrate a possibility that has been denied is very different from—and logically much more defen-sible than—the usual approach of using incomplete models to make predictions about what will actually happen in the real world (which is what chapter 4 critiqued).

  So there are no small numbers, externalities, or informational imperfections in the suite of models considered here, nor any currency problems. Instead, attention is reserved for the effects of two kinds of differences between two countries: in terms of initial capital stocks per capita, or how rich the two countries are, and in terms of time rates of preference, or how patient they are.10 In line with the discussion of Chimerica, the poorer country, denoted as C, is assumed to be more patient than the richer country, A, as elaborated below. But apart from these cross-country differences, the rest of the context is simplified (e.g., technological progress is ignored) so as to minimize the number of moving parts extraneous to the dynamic mechanism highlighted by the models.

  Begin by considering, as a simplifying baseline, a model in which A and C have different capital stocks per capita—A's are higher—but are otherwise identical, including in terms of access to technology and, for now, time rates of preference. Also assume provisionally that they operate side-by-side but in isolation from each other. In other words, this baseline allows for one of the two kinds of cross-country differences that we ultimately want to consider (the second kind will be considered next).

  Convergence in the long run is guaranteed under these assumptions because if you run time far enough out, initial conditions don't affect projected outcomes. Convergence in the long run implies that C is going to have higher rates of capital accumulation and growth than A in the short to medium run. The asymmetry in growth rates—in favor of C—will be higher the greater is the initial asymmetry in capital stocks (wealth endowments). And extending this basic model to include governmental investment and consumption suggests that C will also have a proportionately bigger public sector than A in terms of productive investments—but not in terms of consumption (e.g., environmental public goods). So this very simple setup suffices to generate some of the differences evident in tables 8-1 and 8-2.

  Now, add to this baseline setup an asymmetry in terms of discount rates, with C also being more patient or long-term-oriented than A—but continue, for now, to keep the two economies isolated from each other. This assumption of different discount rates is motivated by several aspects of Chimerica: the long decline in U.S. savings rates that began in the mid-1950s and accelerated in the 1980s, the much higher savings and investment rates in China, the Chinese government's emphasis on maximizing growth rather than welfare,11 and suggestions by culture gurus that the disparity in short-term versus long-term orientation is perhaps the key cultural difference between the two countries.12 While economists may not be convinced, especially by the cultural evidence, note that there is nothing in economic theory to rule out the possibility of different time rates of preference across countries. So this is a possibility that must be accounted for by those who assert that no need can arise for governments to intervene to fix imbalances.

  Allowing for different discount rates has such extreme implications in terms of tilting long-run outcomes in favor of C that the impacts of different discount rates usually go unanalyzed in most theoretical treatments. But one textbook provides a summary (for the case of multiple countries):

  If all countries have different discount rates, the country with the lowest discount rate determines up to which point capital is accumulated in the world economy. The country with the lowest discount rate provides capital until its discount rate and the real interest rate are equal. Then its consumption stays constant. For the other countries, consumption decreases in the course of time.13

  The country with the lower discount rate, C, eventually also becomes richer than A under this scenario (unlike the baseline scenario, which implied convergence in their incomes). One could imagine this outcome being at least mildly disturbing from the perspective of A's citizenry.

  Even more disturbing, though, is what happens when one connects two countries with different rates of time preference: the divergence implied is much more dramatic. The spendthrift economy, A, falls behind even more quickly as it borrows from C to finance current consumption. Exactly how this happens depends on the channels that connect A and C. Start by allowing capital mobility while continuing to disallow trade in output. With different discount rates, the less patient country, A, will import capital from a more patient country, C, in the short run and incur some level of ongoing debt service payments in the long run that effectively reduce its long-run consumption levels.

  But at least the physical capital formation that corresponds to this process of capital-based connection is expected to occur in A (although it is financed by C). With trade as a channel of connection as well, there need not be physical capital investment in A at all: trade opens up the possibility of C running a trade surplus based on productive capacity at home and accumulating IOUs from A rather than investing in productive capacity in A! If we focused on just the trade part of the relationship, we might characterize it as benignly as Lindsey: since trade embodies a voluntary exchange, it supposedly can't be bad. Yet the possibility of cross-border connections (further) shifting consumption from the future toward the present, in the spendthrift economy to the point of putting pressure on living standards over time, is surely very disturbing.

  To those potential problems with opening up, one can add currency issues—which were simplified away from the suite of models presented above, even though they occupied center stage in the G20 leaders' discussions in late 2010 of imbalances. Imagine that A issues IOUs to C in A's own currency, which it can print as much of as it likes. Obviously A has a huge incentive to devalue its currency over time. And C could anticipate this, taking us out of the reach of the “nonstrategic” models that we have looked at.14 But even without modeling, currency issues clearly create additional challenges for those of us intent on preserv- ing the benefits of international exchange while managing the side effects—that is, avoiding reversion to World 1.0 just because World 2.0 doesn't work.

  The broad message from the suite of models discussed in this section is that World 3.0 clearly requires some management of imbalances of the sort exe
mplified by Chimerica. While the precise approach that will be undertaken remains to be determined, the models do provide grounding for discussions of the proposals in the air in late 2010, as this issue moved to center stage. First, the G20 leaders' vague agreement to move toward more market-based determination of exchange rates might be a step in the right direction (if there were any specific follow-up) since the renminbi does seem to me, as it does to many others, to be significantly undervalued. But to leave matters there would be to peel off once again into the “markets know best” camp, problematic both because of the character of foreign exchange markets (they are considered very inefficient compared to other categories of financial markets and, relatedly, have a very high speculative component) and because there are reasons for imbalances that have nothing to do with exchange rate misalignments (for a recap, consult tables 8-1 and 8-2).

  Second, the idea of moving toward some kind of cap on imbalances seems attractive in light of the discussion in the last chapter, particularly around figure 7-4. So, at first blush, does the initial U.S. proposal in 2010 of specific limits on current account balances as a percentage of GDP that, once reached, trigger remediation. However, the proposal was rejected in summary fashion by countries whose current account balances have recently exceeded the proposed limits (e.g., China). Looking beyond Chimerica, there is also an issue of fairness in imposing the same percentage caps on large economies and small ones when they translate into very different contributions in absolute terms to global imbalances. The broader implication, encountered before (in chapter 6), is that one might want to have differentiated targets to account for vast differences across countries.

  Third, despite the differences across countries, this is a situation that illustrates how multilateral approaches are, in at least some situations, to be preferred to narrower ones. There is no presumption that, in a world without trade barriers and currency manipulation, trade would balance out between country pairs; rather, the presumption is that of a rough balance in a country's trade with the rest of the world. As a result, it is natural to think of capping imbalances in those terms—but that requires multilateral agreement rather than unilateral action.

  The final thought worth adding in regard to Chimerica is that while currency and trade adjustments have a role to play in reducing the problem of imbalances, the United States also needs to address the domestic factors that underlie its savings-investment gap (if not the overall low levels of both). At the first academic conference on international competition that I ever attended, in 1985, similar concerns were in the air in the United States about Japan. A young Larry Summers pointed out that while forcing adjusments on the Japanese might help alleviate the chronic U.S. current account deficit in the short to medium run, action was required on the domestic front if the problem was to be solved in the long run. Looking at recent numbers, I suspect the same idea can be applied to the United States and China in 2010. Which illustrates another broader World 3.0 theme: don't focus on the international to the exclusion of the domestic since many—or most—effects are still highly localized.

  Human Imbalances

  Demographers Wolfgang Lutz and Warren C. Sanderson pegged the dawn of the twenty-first century as the shift from the “century of population growth” to the “century of aging.”15 After rapid growth dating back to the advent of World 1.0, world population is expected to level off at some 9 billion-plus by the middle of the twenty-first century. The “demographic transition” that generates this pattern—rising life expectancies followed by falling birth rates—is well along in rich countries but at different stages and progressing at different rates across the developing world, creating widening demographic disparities.

  Consider the world's oldest and youngest large countries, Japan and Niger. In Japan, the median age is already forty-four years and by 2050 is expected to reach fifty-five! And from 2010 to 2050, Japan's population is projected to shrink by 20 percent to only 100 million. There's so much concern about how Japan's declining working-age population can support an unprecedented number of elderly that they're racing to develop robots to help them out. Much of Europe faces similar conditions, and the challenge is particularly difficult in eastern European countries with a fraction of Japan's per-capita income. The OECD has estimated that without large-scale immigration, Japan's living standards could drop by 23 percent, the EU's by 18 percent, and the United States' by 10 percent.16

  In Niger, in contrast, the median age is only fifteen and a full 68 percent of the population is under twenty-five years old. While it seems everyone in Japan is sprouting gray hairs, in Niger it's hard to find an adult, much less a senior citizen (Niger is also one of the world's least densely populated countries). And Niger's population is projected to more than triple over the next forty years. Figure 8-3 puts these examples in broader context. In the chart on the left, South Asia and Africa are shifted to the top to emphasize that most of the population growth through 2050 will take place in those regions.

  Thinking about the link between demographics and economics has come a long way since the old Malthusian caricatures of people as mouths to feed. More recently, the “neutralist” perspective has gained favor as research has indicated no clear positive or negative effect of population growth on per capita income.17 Age distributions, however, do seem to have a significant impact, with countries enjoying favorable “windows of opportunity” when a high proportion of their populations are of working age. A bulge of children like Niger's or a high proportion of retirees like in Japan both mean that each worker has to support more dependents (a high dependency ratio), which makes it harder to accumulate wealth. As figure 8-4 shows, the proportion of working-age people in East Asia (including Japan), Europe, and the United States will be declining while it rises in South Asia and Africa.

  More migration is an obvious remedy for demographic imbalances but it stirs up many fears. This section summarizes patterns of migration and the potential global gains from enhancing such cross-border flows of people, and the sections that follow look at migration's impact in sending and receiving countries.

  Figure 8-3: Long-term population growth and age distribution trends

  Source: United Nations, World Population Prospects, 2008.

  Figure 8-4: Ratio of working-age to dependent population, 1950–2050

  Source: David E. Bloom and David Canning, “Global Demographic Change: Dimensions and Economic Significance,” NBER Working Paper 10817, September 2004.

  As we saw in chapter 2, human flows across borders are much smaller than product or capital flows, with only about 3 percent of people living outside the country where they were born, roughly the same proportion as fifty years ago.18 And today's levels of migration pale in comparison to the late 1800s, when “total emigrants over a decade accounted for 14 percent of the Irish population, 1 in 10 Norwegians, and 7 percent of the populations of both Sweden and the United Kingdom.”19 With stark wage disparities, falling transportation costs, and the addition of more national borders, why hasn't the number of migrants surged upward like other kinds of cross-border flows?

  The simple answer is that while the benefits of migration net of the costs have gone up, barriers have been raised to block its growth. As Ng and Whalley explain:

  Prior to 1913, visas were not required for transit between most countries, and work permits were also not required for employment of foreigners. Passports were largely used as proof of identity and/or citizenship once inside national borders in case help were needed, typically from an embassy or ambassador abroad. Border formalities focused on revenue collection via tariffs from those crossing borders with goods in transit, not on documentary proof of identity.20

  In fact, up to 1924, visas weren't required to settle in the United States and “in 1905 only 1 percent of the one million people who made the transatlantic journey to Ellis Island were denied entry into the country.”21 Migration was even subsidized in many cases by both sending and receiving countries to address labor market imba
lances.

  In today's “borderless” world, visas and work permits are almost always required, and the processing of these documents alone has been estimated to cost $88 billion globally, or about 0.3 percent of world GDP.22 In one in ten countries a passport itself costs more than 10 percent of the country's per capita income. In the Democratic Republic of the Congo, a passport costs 125 percent of the per capita income—and does not, of course, guarantee visas to travel!23 Furthermore, there's no organization like the WTO dedicated to removing barriers to migration.

  Because of the fears I will address below, policies in most countries are aimed more at restricting than facilitating migration. And yet, with such strong incentives for movement, enforcement is highly uneven. Some 50 million people live and work abroad without proper authorization, and many endure unspeakable horrors trying to cross borders illegally, with thousands dying each year.

  With figures like that, one might assume that most migration is from developing countries to developed countries, but actually that is not the case. Sixty percent of migration from developing countries is to other developing countries—although almost always relatively more developed ones. Thirty-seven percent is from developing countries to developed countries, and the balance (3 percent) is from developed countries to developing countries. And we can see the usual distance effects here, too. Nearly half of migrants stay within their home regions, 60 percent stay in a country with the same major religion, 40 percent move to a place with the same major language, and so on.24

  Readers from developed countries might have a hard time believing that the stock of immigrants hasn't gone up in recent years. Intuitions about rising immigration in developed countries are, in fact, correct, but reflect a shift in the composition of migratory flows, with a higher proportion of migrants now going to developed countries, rather than larger aggregate flows. This change, combined with different population growth rates, has led the proportion of immigrants in the populations of developed countries to rise from 4.6 percent in 1960 to 12.1 percent in 2010. It's also true that a much larger proportion of new immigrants into developed countries come from developing countries. In the United States, this proportion rose from about half in the 1960s to about 90 percent in the 1990s and early 2000s.25