by Ian Fletcher
One consequence of this is that economic growth is path-dependent. To grow, an economy must continually break into new industries. But to do this, it needs strong existing positions in the right industries. So a national economy that doesn’t get onto the right path (and stay on it) risks being sidelined into industries which lead nowhere in the long run. We noted this problem before in Chapter Five: 18th-century Portugal derived no other industries from winemaking, while Britain derived many from textiles because the construction of textile machinery spawned a machine-tool industry that could produce innovative machinery for other industries. Similarly, electric cars may be the wave of the future today, but without a strong position in conventional cars, a nation is unlikely to have the know-how or supplier industries to build them.
Path dependence applies to economies at all levels of development, not just those starting to industrialize. Infant-industry protection is, of course, one of the best-known cases for protectionism and industrial policy. (It is often the one case grudgingly conceded even by free traders.) But it is, in fact, only the most obvious case of the more general phenomenon of the path dependence of economic growth. Infant industries are merely the first rungs of the ladder.
A key concept here is the driver technology, which enables progress in multiple other technologies. As Clyde Prestowitz of the Economic Strategy Institute writes of Japan’s makers of industrial policy:
They knew that the RAM [random-access memory chip] is the lynchpin of the semiconductor industry because, as the best-selling device, it generates not only revenue but also the long production runs plant managers use to test, stabilize, and refine the production and quality-control processes. Compared with many other chips, it is a relatively simple product, which makes it a more attractive vehicle for developing new techniques. The latest technology has always been incorporated first in RAMs, which have always been the first product to appear as a new generation. Once RAMs are refined, new generations of other products follow...The Japanese knew that if they could grow faster than the Americans in the RAM segment of the market, they could become the low-cost producer of RAMs. And if they controlled RAMs, they would have taken a long step toward dominance in other semiconductors. And if they had semiconductors, semiconductor equipment, materials, and everything that semiconductors went into, such as computers, would be next.655
Free market, free-trade economics systematically maintains the opposite of all this. It maintains that any industry can drive an economy upwards, just so long as it enjoys comparative advantage right now. And because free trade economics holds that free trade automatically steers an economy into those industries where it enjoys comparative advantage, it holds that free trade will therefore maximize economic growth.656 For free-trade economics, there is, in fact, no important distinction between the long and the short term: comparative advantage is always right, period. Free-trade economics holds that it is profoundly impossible for one industry to be “better” than another. This is the cause of an infamous (subsequently denied) comment by Michael J. Boskin, George H.W. Bush’s chairman of the Council of Economic Advisers:
It doesn’t matter whether America exports computer chips, potato chips, or poker chips! They’re all just chips!657
Why would Boskin make a statement so brazenly contrary to common sense with such confidence? Because free-trade economics holds that markets are so efficient that no industry can be special. In its view, there can be no ladder externalities because there can be no industry externalities at all—certainly none that are big enough and evident enough to understand and manipulate. Every industry’s profits today must accurately reflect its value in both the short and the long term. Why? Because if any industry did have superior value for future growth, its expected profitability today would reflect this, that superior profitability would draw new firms into the industry, and the superior profits would be competed away.
If every industry’s short-term profitability were indeed a correct measure of its long-term value, this would indeed be the case. But when long-term returns may well accrue to another company, even another industry, and someone else may capture them, short-term profitability is not a reliable metric of long-term value. So any strategy that relies on short-term profitability alone to steer an economy will necessarily underperform. (As noted in Chapter Two, short-termism is a crucial hidden part of America’s trade and industrial problems.)
“Just chips” economics is wrong because industries are very much not alike in their long term consequences. In the words of Laura D’Andrea Tyson, Bill Clinton’s chairperson of the Council of Economic Advisors (who never got to apply her valuable theoretical insights in office):
The composition of our production and trade does influence our economic well-being. Technology-intensive industries, in particular, make special contributions to the long-term health of the American economy. A dollar’s worth of shoes may have the same effect on the trade balance as a dollar’s worth of computers. But...the two do not have the same effect on employment, wages, labor skills, productivity, and research—all major determinants of our economic health.658
NOT ALL COMPARATIVE ADVANTAGE IS CREATED EQUAL
Free market, free-trade thinking can’t comprehend the above realities. But it still has a contribution to make to our understanding here. In light of these realities, comparative advantage doesn’t disappear from the picture entirely. But a crucial insight is added: not all comparative advantage is created equal.659 It is better to have comparative advantage in some industries than in others, and what a nation has comparative advantage in determines its standing in the global economic pecking order. In the words of Paul Krugman:
Each country has a “niche” in the scale of goods; the higher a country is on the technological ladder, the further upscale is the range of goods in which it has comparative advantage.660
This may sound obvious, but this reality is relentlessly obscured by free-trade thinking, which defines away the possibility of some kinds of comparative advantage being better than others by its insistence that it is always best to act according to the comparative advantage that one has today.661 As the Norwegian economist Eric Reinert puts it:
The very idea of a nation lifting itself to higher levels of living standards through competitiveness—being engaged in activities that raise the national living standard more than other activities—goes directly against the assumptions and beliefs which form the foundations of the neoclassical economic edifice. This is not the way economic growth is supposed to take place in the neoclassical model.662
One implication of all this is that national economies tend to rise (or decline! Read “Argentina”) over time to the level of income embodied in their exports. Dani Rodrik has summarized this fact as “you become what you export.”663 This is a fact with vast significance for industrial policy—especially for developing countries, which are desperately trying to become something other than what they already are. And this fact profoundly contradicts Ricardian economics. As Rodrik observes:
Under received theory, a country with an export package that is significantly more sophisticated than that indicated by its current income level is one that has misallocated resources (by pushing them into areas where the country does not have a comparative advantage). Such a country should perform badly relative to countries whose export packages are more in line with current capabilities.664
That climbing a ladder of industry externalities can lift an economy upward shows up in the data in the fact that economies with more-sophisticated exports are not only richer today (which one would expect) but also grow faster over time.665 The latter fact, although not terribly shocking to common sense, is not obvious at all to free-trade economics. But in reality, having a foothold in industries which intrinsically lead somewhere is a big part of what makes economies grow.
All of the above was, of course, well known to mercantilist governments for centuries. In the words of economist John Culbertson:
This view...had been
well understood by governments and writers on economic subjects centuries before Adam Smith, that industries are not homogenous. Some lead to cumulative advances in knowledge and technology, some bring new skills and capabilities to people and firms, some permit high incomes to be earned in foreign markets because of the absence of competing producers—especially of competing low-wage producers. Other lines of production have none of these favorable characteristics, and are dead ends. The nation that specializes in them will be economically second rate, at best.666
WHAT ARE GOOD INDUSTRIES?
If the industries a nation needs in order to grow economically are those whose intrinsic nature it is to lead onward and upward, and free trade won’t automatically nourish them, which lucky industries might these be?
Let’s start with the fact that sustained economic growth only really occurs in industries which exhibit increasing returns. This means that for a given increase in inputs, output goes up by more than the size of that increase. For example, because the cost of baking bread consists in a one-time investment in an oven plus a per-loaf cost for ingredients, the cost per loaf will go down with each additional loaf baked, as the cost of the oven is spread out over more loaves. So 10 percent more money will deliver 11 percent more bread and so on. The opposite of increasing returns is diminishing returns: after a certain point, 10 percent more money delivers nine percent more output, then eight percent and so on.
Increasing returns is a simple concept, but it ramifies endlessly, forming the ultimate basis of a long list of the opposite characteristics of “good” (increasing returns) and “bad” (diminishing returns) industries.667 Historically, manufacturing is the quintessential increasing-returns industry and agriculture the quintessential diminishing-returns one. But some types of each behave like the other, and since the mid-1970s the line between manufacturing and services has blurred, with a small segment of high-end services acquiring some of the desirable characteristics traditionally associated with manufacturing. And low-end manufacturing has increasingly come to resemble agriculture. But the underlying characteristics of increasing and diminishing returns industries have remained stable, even as which industries exhibit these characteristics has changed.
Having a lot of increasing returns industries is really the only way to be a developed economy. This is, in fact, the fundamental purely economic difference between the First and Third Worlds: the former is full of such industries, the latter is not. As a result, examining why some industries exhibit increasing returns and some do not can tell us a lot about why some economies grow and some do not. And how free trade can easily lead an economy astray.
WHY DO SOME INDUSTRIES HAVE INCREASING RETURNS?
Industries which exhibit increasing returns do so mainly because they can absorb endlessly rising capital investment. Not all industries can: buying another $1,000,000 worth of tractors for a coffee plantation that already has them won’t increase the plantation’s productivity very much.668 Neither will buying every lawyer at a law firm another desk. But putting another $1,000,000 into production machinery in an automobile or semiconductor plant will do a lot. And capital doesn’t just mean factory floor hardware. It also means human capital or skills accumulation, and research and development.
Why are some industries so good at absorbing capital? One big reason is that they are susceptible to innovation, and R&D is a big capital absorber. This activates a virtuous cycle in which innovation absorbs capital then repays it by raising profitability, generating more capital and repeating the cycle. It is no accident that manufacturing and related fields generate over 70 percent of research & development in the U.S.669 And within manufacturing, high technology accounts for roughly 20 percent of output, but 60 percent of R&D.670
This susceptibility to innovation derives largely from the fact that good industries tend to produce goods capable of infinite improvement, like laptops or airplanes, while bad industries produce goods whose character is fixed, like fruit or t-shirts. The products of good industries are also susceptible to meaningful variety, so firms don’t end up selling the exact same product in pure head-to-head competition. This spares firms the raw price competition that drives down profits, wages, and funds available for further investment. Instead, firms compete on quality, reliability, reputation, marketing, service, product differentiation, special understanding of buyer needs, rapid innovation, and managerial sophistication. This enables them to accumulate strongly entrenched competitive positions where vulnerability to competition—crucially by cheap foreign labor—is not a big issue.
This lack of perfect competition in good industries activates something free-market economics despises: market power, also known as monopoly or quasi-monopoly power. From the point of view of free markets, this is inefficient on first principles, because industries with monopoly power earn higher profits than the free market would allow. They are parasitic. The confusing term economists use for this excess profit is “rent” (which has nothing to do with rent in the normal sense),671 so in the words of Eric Reinert:
In the static system of neoclassical economics, rent-seeking is seen as a negative term. In a world where increasing returns to scale, imperfect information, and huge barriers to entry dominate all industries of any importance, dynamic rent-seeking seems to be a key factor for economic growth and competitiveness. 672
This dynamic rent-seeking generates a number of virtuous spirals. One is that rising worker incomes provide the purchasing power to sustain industrial growth. And as incomes rise, what economists call “quality of demand” also rises: people demand not just more but better products, driving the industries of their home nation to upgrade and reinforcing the ladder externalities discussed earlier.673
Good industries also readily absorb rising human capital or skill. The fact that capital accumulates in the workers themselves tends to encourage well-cared-for labor for the same reason factory owners do not let valuable machinery rust away. It generates corporate and state paternalism and the “countervailing powers,” like bargaining leverage by workers, that spread the profits of industry beyond its owners.674 This is reinforced by the fact that good industries tend to produce products for which income elasticity is high, i.e., people buy more as their incomes go up. As a result, productivity gains don’t just drive down the price of the product, and output can rise along with productivity, enabling wages to stay steady despite productivity gains which require fewer and fewer workers per unit of output.
BAD INDUSTRIES AND DEAD-END ECONOMIES
The opposite of good industries is, of course, bad ones. These are dead-end jobs writ large. For centuries, this has meant agriculture and raw materials extraction, but since the mid-1970s, unskilled manufacturing has been inexorably joining this category. In these industries, diminishing rather than increasing returns apply, so all the previously discussed benevolent dynamics are absent—or run in reverse.
These industries are hobbled by their very nature. For a start, demand for agricultural products is intrinsically less elastic than demand for manufactured goods simply because of the finite size of the human stomach.675 As a result, productivity growth in agriculture tends to translate into lower prices for consumers, not higher wages for farmers. Because productivity growth in agriculture tends to go into lower prices, while productivity growth in manufacturing does not, agricultural prices generally decline over time relative to the price of manufactured goods.676 This problem has been around for a very long time: according to one British estimate of 1938, the same quantity of primary products bought only 63 percent as many manufactured goods as it had in 1860.677 Thus nations whose main exports are agricultural or raw material products have slipped further behind the industrial nations, decade by decade.
Agriculture and raw materials also tend to be bad industries because they are too easy for competitors to break into and thus attract too many rivals. When Vietnam, on the advice of the World Bank, started exporting coffee and rapidly became the number two producer after Brazil,
this flooded the market and drove the price down from 70 cents a pound to around 40 cents.678 Economies dominated by bad industries are subject to volatile income swings due to distant commodities markets, swings exacerbated by undiversified exports and impossible to hedge against. The dependence of agriculture on the weather only makes this worse.
Most agriculture simply can’t absorb technological innovations that upgrade productivity and wages on anything like the scale manufacturing can. For example, as Eric Reinert notes:
Mexico specializes in unmechanizable production (harvesting strawberries, citrus fruit, cucumbers and tomatoes), which reduces Mexico’s opportunities for innovation, locking the country into technological dead-ends and/or activities that retain labor-intensive processes.679
Because agriculture can’t absorb technology, it can’t absorb capital, either, as there’s nothing to spend the money on that will pay back a return. In any case, without a strong manufacturing sector, it’s hard to raise even agricultural productivity because increased productivity means fewer workers are needed, and there’s nowhere for the workers released from agriculture to go. So fear of mass unemployment tends to lock society in place.680 For most of the people they employ, agriculture and other bad industries also tend to hit a fairly low ceiling in the amount of skill they can usefully absorb, so human capital doesn’t accumulate any more than capital invested in technology. As a result, these industries remain undercapitalized and the societies that host them do not accumulate wealth in these industries. Whatever money is made is siphoned off elsewhere: into castles in Medieval Europe, into Europe in colonial Africa.