Stubborn Attachments

by Tyler Cowen

  When we are examining a policy change or an act of redistribution, it is important to know whether it involves an upfront, once-and-for-all benefit (or cost) or a systematic boost (or decline) in the growth rate over time. This somewhat arcane distinction, drawn from the economic literature on growth, is of great importance for adjudicating potential social changes.

  To consider a simple example, many scientists believe that global warming will increase the number of virulent and persistent storms on our planet. Many of these storms may come only after some time, but a greater concern for the future means that we must pay heed to these consequences. More generally, many environmental problems hurt our prospects for long-run sustained growth. I am suggesting that such problems are especially important.

  At the same time, a concern for the distant future can, counterintuitively, dissuade us against certain environmental investments. In contrast to the threat of severe and ongoing storms, some of the costs of climate change take the form of one-time adjustments, such as the cost of relocating coastal settlements. The induced relocation might count as a rights violation of some sort, but from a consequentialist point of view, maximizing the growth rate takes priority over avoiding one-time expenditures and one-time adjustments. Even if those expenditures are large, we will earn back that value over time, due to the logic of compounding growth. So, contrary to what other frameworks might suggest, we should pay greater heed to the storms and less heed to the relocation costs.

  It’s very important to take note of whether or not a specified policy choice will affect the overall growth rate. The details of this distinction are complicated, however, because economic models provide differing accounts of which changes alter growth rates, as opposed to bringing one-off changes or improvements.

  The most prominent economic approach to growth, the Solow model, is named after MIT economist and Nobel Laureate Robert Solow, who laid out the basics of the model in the 1950s. The Solow model postulates a stripped-down economy-wide production function based on constant returns to scale. National output is the result of capital inputs, labor inputs, and technological progress, which renders both capital and labor more effective.10 In this model, the primary way to increase ongoing growth is to induce a higher rate of technological innovation. Indeed, much empirical research has shown embodied technological progress to be a major factor behind U.S. economic growth.

  The Solow model helps us to understand the phenomenon of catch-up growth, which has been so significant in East Asia. In the model, the rate of return on capital diminishes as the capital stock increases. So when there is not much capital, the returns on investment and thus the incentives to invest are high, at least provided that the labor force is of good quality and institutions will protect the property rights of investors. Poorer countries should therefore be expected to catch up to richer countries as they borrow new technologies and increase their capital stocks to implement new and lucrative opportunities. The economic growth in China that began in the 1980s had this catch-up flavor.

  The Solow model also implies that economies should recover quickly from one-time negative shocks such as earthquakes or destructive wars. Although the capital stock has fallen from the destruction, the rate of return on capital is now higher precisely because the capital stock has fallen. Additional savings should make up the gap and, over time, restore the economy to its previous growth path. The very rapid recovery of Japan and Germany after World War II demonstrates this mechanism in action. So according to this theory, the rate of growth will remain lower only if the negative shock somehow permanently reduces the rate of technological progress.

  For similar reasons, a boost in savings and investment is seen as contributing to transition growth paths but not to steady state growth in the long run. In the model, more savings means a given amount of catch-up growth happens more rapidly, but the savings won’t raise long-run growth rates.11

  That’s the Solow model, but I’m not arguing that it is always the best model for understanding our world or for judging redistribution. It is simply one possibility.

  In contrast to the Solow model, the increasing returns model suggests that growth begets more growth. In this view, larger economies should grow more rapidly than smaller economies, and growth rates should continue to increase over time. Improvements beget further improvements and negative events are likewise cumulative, thus the moniker “increasing returns.”

  Ideas, and their non-rival nature, are often cited as the fundamental source of increasing returns. Once an idea has been generated, it can be used many times by many different people at very low marginal cost. The first idea spreads, begets subsequent ideas, and so growth increases. Or there’s another way to look at increasing returns. Larger markets generate stronger incentives for idea production because innovators can sell their product to a larger market (e.g., it would not be worth inventing the iPhone for customers in New Zealand alone). That means large economies can grow more rapidly than small economies. The more the economy grows, the greater the incentive for subsequent new ideas, which in turn reinforces the incentive for growth. New ideas will lead to more growth, which encourages more new ideas, and so on.

  The increasing returns model is most commonly associated with the economist Paul Romer, but it can be traced back to Adam Smith and the very beginning of economics as a systematic object of study. In Smith’s implicit model, a larger market size supports a greater division of labor, which in turn makes the economy more productive. In other models, greater openness to trade, or a common market area such as the United States, can drive an increasing returns to scale process.

  To put these growth theories into my own terminology, the increasing returns model suggests that there are lots and lots of Crusonia plants out there, whereas the Solow model indicates that we can find Crusonia plants only in policies that very directly and very specifically raise the rate of idea generation. The increasing returns model holds that virtually any gain in resources can be translated into higher growth in the long run, rather than washing out in the adjustment process.12

  Under the increasing returns model, a one-time negative shock harms the long-run rate of growth, which implies that we must take great care to avoid or limit each and every possible negative shock. The Solow model suggests a picture of greater resilience, since catch-up effects prevent each and every mistake from compounding over time into a larger collapse.

  The increasing returns growth model will therefore make us more wary of non-growth-enhancing wealth redistribution than will the Solow growth model. In the Solow growth model, the costs of redistribution might be “once-and-for-all,” rather than lowering the long-term rate of growth. We can make up for our temporary losses and eventually get back to where we ought to be. In contrast, under the increasing returns model, any setback will make the economy smaller and thus limit future rates of growth, with significant implications for the standard of living in the distant future.

  This distinction points to yet another way traditional political debates should be redrawn. Individuals who believe in the increasing returns model should be much more skeptical of non-growth-enhancing redistribution than individuals who believe in the Solow catch-up model.

  A central question here is whether the logic of the Solow model or the increasing returns model holds. Even if you don’t buy into all of the details of these models, the two core options they present are either that one-time costs do matter a lot in the long run, or they don’t. I’m treating the models as stand-ins for these two broader views. The key question is whether gains and losses compound over time or dwindle into longer-run insignificance.13

  Whatever your exact view of the Solow and increasing returns models, the logic of the increasing returns model will likely carry significant weight in our final evaluation. In many cases our best answer, given current knowledge, is that a given cost brings some probability of an ongoing growth effect (as in the increasing returns model
) and some probability of a once-and-for-all adjustment cost, followed by catch-up (as in the Solow model). In our expected value calculations, this will operate as an expected impact on the long-term rate of economic growth. Therefore we should incorporate the logic of the increasing returns model into how we evaluate social changes, even if the increasing returns model is not our single best current theory of economic growth. In expected value terms, most of our social choices have an impact upon future rates of economic growth. Crusonia plants are everywhere, in expected value terms. We are making decisions about Crusonia plants all the time.

  Finally, both the Solow and the increasing returns models emphasize ideas as the wellspring of economic growth. New ideas are the product of human reason; it was Aristotle who defined man as the rational animal. A preoccupation with pursuing growth—or some modified version of the growth ideal—therefore means a preoccupation with ideas, a preoccupation with cultivating human reason, and a preoccupation with the notion that man should realize, perfect, and extend his nature as a generator of powerful ideas that can change the world. Cringe all you wish, but on this point I’ll send another credit along to Ayn Rand, who stresses this point even more than most philosophic rationalists. If we are pursuing self-sustaining and self-generating bundles of plural values, we are in one way or another paying homage to the power of human reason.

  1. See for instance Alesina and Rodrik (1994) and Persson and Tabellini (1994). For a survey of the growing literature on how income distribution can affect growth, see Greiner, Semmler, and Gong (2005, 132–133).

  2. See, for instance, Barro (1991). Goodin, Headey, Muffels, and Dirven (1999) argue that a democratic social welfare state does not lower the rate of economic growth, but they use only two data points, the Netherlands and the United States. See also Lindert (2004). He argues that higher welfare spending tends to be packaged with other growth-enhancing policies, such as low taxation on capital income. He does not show that higher spending at Western European levels is itself good for economic growth.

  3. On utilitarian obligations, see Scarre (1996). In addition to Smart and Williams (1973), Rand (1967), Scheffler (1982), Wolf (1982), Railton (1984), and Nagel (1986) also criticize the notion of extreme utilitarian obligations.

  4. Hurley (2009) raises the question of whether consequentialism, even if it stipulates that some set of actions is good, can generate a strong obligation to perform those actions. In the synthetic view outlined here, that sense of obligation can come from common sense morality, if need be.

  5. See Goodin (1995).

  6. For a left-wing view of discounting, see, for instance, Solow (1974). Beckerman (1996) offers a market-oriented view critical of zero discounting.

  7. Alternatively, a sacrificer could be specified by lot. Or we could look to game theory. We could think of morality as prescribing that individuals should play what economists call randomized Nash strategies, which will mean some probability of selfishness and some probability of sacrifice for each individual, as if we were rolling a die. A correctly specified randomization strategy will bring about the right amount of sacrifice—on average, at least—without requiring everybody to sacrifice. The proper randomized set of strategies will be those that maximize sustainable, expected global economic growth.

  8. On our obligations to save, see Rawls (1999, 252).

  9. See Harding, McGregor, and Muller (2013).

  10. See Solow (1956, 1957); Romer (2000) provides a more recent summary.

  11. Some later modifications to the Solow model allow for the rates of savings and investment to be correlated with economic growth in a more general manner; see Temple (1999, 139–140). Extensions by Uzawa (1965), Lucas (1988), and others stress the role of human capital in boosting or maintaining the growth rate.

  12. On increasing returns models, see Romer (1986, 1990). On the Solow model vs. the increasing returns model, see the 1994 symposium in Journal of Economic Perspectives.

  13. Neo-institutional approaches are less formal than either the Solow or the increasing returns models. They point to the importance of property rights, well-functioning institutions, trust, the rule of law, and properly aligned microeconomic incentives. Nonetheless, these views do not typically specify which policy changes cause permanent boosts in the growth rate, as opposed to once-and-for-all changes. The neo-institutional models have very real merits in explanatory terms, but for normative purposes they do not mitigate the importance of taking a stance on the long-run growth effects of a policy change. On neo-institutionalist approaches, see Douglass North’s work on American and European economic history (North 1981, North and Thomas 1976); see also Olson (1984), Bates et al. (1998), and Acemoglu and Johnson (2004).

  6—Must uncertainty paralyze us?

  When I was a kid, I loved science fiction stories. I loved to think about how things could be totally different from the way they were. One story in particular intrigued me, and I encountered its premise in a few different books. It still appears in popular culture, and if you don’t like science fiction, try the 1998 movie Sliding Doors, starring Gwyneth Paltrow, which asks how much a single life can be altered by the simple act of missing a train.

  The premise is pretty straightforward: if you could somehow manage to go back in time and alter one small event, the entire history of the world might change. One extra sneeze from one caveman millennia ago might overturn life as we know it today. Ray Bradbury’s short story “A Sound of Thunder,” published in 1952, is one of the early sources of this idea. It seems a little crazy, but the more you think about it, the more it seems to hold true.

  The key point is that small changes can very easily turn into big changes. What if Joseph pauses to pick up a penny on the sidewalk? This slight change in the timing of his life will almost certainly alter the particular identities of his future offspring, for example by delaying the occurrence of sexual intercourse with his wife by a tiny increment, or by slightly shifting the position of his testes, thus affecting which sperm will end up fertilizing the eggs of his wife. Further into the future, if a different set of individuals is born, the world will likely take a different path—sometimes a very different path. What if Hitler’s great-grandfather had seduced his wife just a moment later in time, changing the combination of sperm and egg that met as a result? Hitler as we know him would never have been born. Even if most people don’t much matter for broader aggregate outcomes, it sure seems like some of them do, for instance Jesus or Mohammed or Buddha, not mention Hitler or Lenin. Without Hitler, Nazism probably would not have succeeded; that would mean no World War II and no Holocaust, at least not in the forms we witnessed. Virtually every country’s subsequent history would have been different and humanity would be on a very different path for the rest of its time on Earth. Are we really so sure that the United States would still have been the first to build usable nuclear weapons?

  You might assume that aggregate global outcomes are mostly stable with respect to small perturbations in the basic events of daily life. You might think, for example, that the logic of positive-sum trade and the power of human reason to advance technological progress will win out in the long run, with or without Hitler on Earth. Other offshoots of these butterfly effects may cancel out or offset each other in the longer run; Tolstoy, in his novel War and Peace, argued that the “great men” of history had little impact, as their acts would be reversed by their successors (Napoleon being one case in point).

  While part of me wishes that this logic were true, the brute reality is that contingency is real and disturbances to the flow of temporal events need not dwindle into insignificance, given that even a small act can reshape the entire future genetic history of humanity. Napoleon’s actions changed the course of life in Germany, which underwent a liberal intellectual revolution as a result of the French invasion and subsequently modernized, as well as Egypt, which received the printing press and a large dose of liberal ideology before s
tarting to resent European interference in local affairs, a feeling that persists to this day. The histories of these regions were changed irrevocably, as were the histories of the Jews who were liberated under Napoleon’s rule.

  The key point is this: even if you’re not convinced that Napoleon really mattered, you don’t and indeed can’t really know this. There is a real chance that Napoleon being born, rather than a different child from a different act of conception, fundamentally changed world history. So in terms of expected value, it remains the case that small acts can have a big impact on the future, even if they do not always do so. It only has to be the case that some small acts steer the future along a very different path.

  Following the philosophers, I refer to this as the epistemic problem.1 The epistemic problem isn’t really about time travel at all, even though science fiction constructs allow us to visualize the problem in an especially vivid manner. The real issue is that we don’t know whether our actions today will in fact give rise to a better future, even when it appears that they will. If you ponder these time travel conundrums enough, you’ll realize that the effects of our current actions are very hard to predict, and that has nothing to do with whether or not time travel ever becomes possible.

  The epistemic critique suggests that the philosophic doctrine of consequentialism cannot be a useful guide to action because we hardly know anything about long-run consequences. While we can try to calculate expected values, such calculations are typically based on a very limited range of information about present consequences or consequences in the near future. As David Schmidtz once put it to me: can we have the correct moral theory if we cannot know ninety percent, or perhaps 99.9 percent, of what counts toward a good outcome?