Exceptional Argentina Di Tella, Glaeser and Llach I. Introduction Argentina began the 20th century as one of the richest places on the planet. In 1913, it was richer than France or Germany, almost twice as wealthy as Spain, and its per capita GDP was almost as high as that of Canada. Until the 1930’s, people in France used the phrase “riche comme un Argentin” to describe the foolishly rich. Over the last 100 years, Argentina’s place in the hierarchy of nations dropped precipitously, falling behind not only Europe but many of the growing countries in Asia. Governments enacted questionable policies, like very high trade barriers and widespread nationalization, long after they were economically and politically attractive. How did a nation that was doing so well end up doing so poorly? This introductory essay summarizes the ten papers in a volume dedicated to exploring the puzzle of Argentina in the twentieth century. Section II begins by summarizing the facts. We rely, like much of the research community, on the data of Angus Maddison, but we supplement this with other sources. While there is some controversy about whether Argentina was really wealthier than Western Europe, there is little doubt that Argentina was, in fact, quite prosperous by world standards. The country certainly had much inequality, but so did much of the world. Between 1870 and 1930, Argentina experienced robust growth and remained a prosperous nation. Between 1930 and 1970, Argentina continued growing, but at a slower pace than the world as whole. By 1975, Argentina’s income had slipped to being 60 percent of incomes in France. Then after 1970, Argentina stagnated and during some years even declined. As such, Argentina certainly serves as a cautionary tale about how a wealthy country can lose its way. In Section III of this paper, we outline four basic hypotheses about why Argentina fared so poorly in the twentieth century. These hypotheses are not mutually contradictory; in some cases, they reinforce one another. They are advanced and explored by different papers within this volume. The first hypothesis is Argentina’s early success is somewhat illusory. According to this view, Argentina did have natural resources that briefly made it rich when those resources were in high demand, but it did not share the other attributes of advanced countries before World War I. In particular, its human capital, physical capital and access to cutting edge technologies were far below those in many, poorer countries. According to this view, the decades around 1910 should be seen as a brief outlier, and Argentina post-1945 has just returned to the level of wealth implied by its core assets. The second hypothesis emphasizes bad policies and politics. During its heyday, Argentina had a pro-growth, liberal democracy. That regime collapsed in 1929 and was replaced by a stream of different governments, some of which were highly protectionist and economically intrusive, in quite harmful ways. Others took a less interventionist approach, but were also short lived. Political instability fostered economic short-termism and policy reversals became the norm. Failures of policies and politics thus led to Argentine economic stagnation. A third explanation of Argentina’s economic malaise emphasizes the role of terms of trade shocks that were external to the economy. In the 1920s and earlier, the world placed a high premium on Argentina’s agricultural output. Over the rest of the 20th century, technological advances in agriculture reduced the value of Argentina’s fertile agricultural land and this led to an understandable impoverishment of a country that primarily depended on the fruits of its soil. The final hypothesis emphasizes the lack of innovation and economic development in Argentina. This hypothesis, like the third one, also de-emphasizes politics, but it lays the blame for Argentina’s woes not on external factors, like the terms of trade, but on a domestic failure to produce new ideas and new technologies. Taylor, for example, has stressed the low savings rate in Argentina as a cause of its slow growth rates. This hypothesis is linked to the first hypothesis because, presumably, Argentina’s lack of economic growth reflects a lack of the core inputs into that growth during the early years of the 20th century. Section IV of this essay reviews the ten papers in the volume. We summarize their individual contributions, but also tie them to the four hypotheses. For example, the Llach paper and the Campante and Glaeser paper both relate to the first hypothesis, arguing that Argentina may have been rich but was not yet developed. The Di Tella and Dubra paper on Perónism analyzes the persistent political beliefs that were associated with Argentina’s dominant political movement during the late 20th century. Such beliefs were directly connected to disappointing performance through fiscal deficits, macro instability and low investment due to political volatility. And indirectly, by converting the low legitimacy of business into a set of commercial institutions that fostered rent-seeking instead of innovation. The Brambilla, Galiani and Porto essay documents a particular expression of such institutional distrust of the rich, namely policies dealing with the agricultural sector and more broadly the relative closing of the argentine economy. The Taylor essay explains the role of changing terms of trade. Again, this is directly related to the fourth class of explanations. Finally, we conclude in Section V with a synthesis of the different views that summarizes the views of these authors—although not necessarily the views of the other contributors to this volume. Argentina was different at the start of the 20th century and had less education, technology and probably also weaker political institutions. Those factors then made Argentina particularly vulnerable to economic shocks, and that vulnerability led to dire political consequences. The lack of human capital also made it particularly hard to find new ways of growing throughout most the past one hundred years. Argentina’s bad 20th century is surprising, but it is not inexplicable. It is the outcome of adverse shocks, and policies that responded to those shocks, impacting a country that had only natural-resource driven prosperity. II. The Basic Puzzle Was Argentina's growth experience during the 20th century an exceptional one? In terms of its rate of growth it certainly was, at least after the first couple of decades. Figure 1 is a good starting point for the questions posed in this book. It shows Argentina's per capita GDP expressed as a percentage of twelve rich nations' income per capita. The rise-and-fall pattern is clear beyond shorter-run fluctuations. Starting at two thirds in 1870, the share rose to around 90% at the beginning of the 20th century. After a brief crisis during World War I and recovery afterwards, a long relative decline ensued, save for a short spell in the late forties (the early Perón years) and stability −relative to the sample− in the 1960s. Relative decline accelerated after 1975, and by 1990 Argentina had reached a level of around one third of these (relatively rich) countries' per capita GDP. The two final decades suggest that in spite intense instability the downwards trends seems to have subsided, though there isn't still recovery in the income ratio with the rich. Figure 1. Argentina's GDP as a percentage of twelve rich countries 100% 90% 80% 70% 10-year moving average 60% 50% 40% 30% 20% 10% 0% 187018761882188818941900190619121918192419301936194219481954196019661972197819841990199620022008 No other country of some economic significance in 1928 −probably the final year of Argentina's own belle époque− took so long in doubling its per capita GDP, a feat which Argentina completed only in 2000. During the same period, the richest members of the "convergence club" (Britain, US, Canada, Belgium, Netherlands, Australia) multiplied their income by around 4 and other currently developed countries by somewhat more: Germany, France and Sweden by around 5, Italy and Spain by 6, Finland by 7 and Norway by 8 − not to mention Japan, Taiwan or Korea which grew more than tenfold. Even among the not too successful Latin Americans Argentina lagged behind: between 1928 and 2000 Brazil multiplied its per capita income by close to 5, Mexico by 4, Chile, Colombia and Venezuela by around 3. Only Uruguay (2.02) and Peru (2.08) are close to the ratio between Argentina's 2000 and 1928 per capita GDP, which is actually 1.9995. Argentina's divergence story by the end of the twentieth century is particularly sad for a country that looked so promising at the century's dawn. In the decade of 1900, Argentina had the highest immigration to population ratio in the world −an unequivocal sign of high hopes− which didn't prevent the country from enjoying the world's highest per capita growth rate in that decade, along with Canada. As documented in Llach's paper in this book, between the eve of World War I and the twenties Argentina was very close to being a rich country if it wasn't already there; it's per capita GDP lagged behind the average of the core Western European countries and Maddison's "Western Offshoots" by less than 20% −and sometimes by as little as 7%− between 1905 and 1929. Massive population growth and economic success combined to multiply by 4 Argentina's share in world GDP between 1870 and 1930.1 Disheartening and disappointing as Argentina's post-1930 performance was, does it really involve an economic puzzle? Is there any reason to believe that a country that has reached the league of the rich should remain there, or even approach the richest among the rich? The question has an empirical and a theoretical side to it. Let's formulate the empirical question with some precision. For example: are there many other examples of countries which, after reaching at some point at least 80% of the per capita GDP of the twelve originally rich2, subsequently fell to a level consistently below 50%, as Argentina did? The list is short, and has the peculiarity that all five cases hovered around 40% of that sample by the year 2000: Uruguay (around 100% in the 1870s), plus four oil exporters: Venezuela (more than 100% in 1945-1960), Saudi Arabia (90% during the first oil shock), Kuwait and Qatar (both of them, more than 400% in the early 1950s). Three of the five are tiny, with a population less than a tenth of Argentina's. In all five cases, the high per capita GDP at their summit is definitely due to an exceptionally high availability of natural resources per person, including not only the oil exporters but also the 213,000 Uruguayans who held, each, an average of 24 cattle and 12 sheep in 18603. Was that also the case for Argentina? In the next section we discuss that possibility as part of one of the hypotheses for Argentina's decline, namely, that Argentina's per capita GDP during the initial decades of the 20th century was a misleading indicator of its real wealth − in particular, that GDP was more related to land and less to physical and human capital than was the case for countries of similar income per head. We leave the full discussion of that hypothesis for the following section and for some of the articles in this volume, but it is interesting to pursue at this point the theoretical aspects of economic growth as applied to countries in which the stream of income springs, to a larger extent than the norm, from natural resources. Is there, in the neoclassical theory of economic growth, anything special going on in countries with a high ratio of natural resources to population? If we define modern economies as those where accumulation of capital −physical and human− and technology account for all of the growth in per capita GDP, a corollary of Solow-type models is the prediction of "conditional convergence": countries with lower per capita stocks of human and physical capital are poorer; other things equal, in countries with lower stocks of both types of capital marginal productivity is higher, due to decreasing returns; then, if "institutional incentives" and other traits −such as propensities to save− are similar, poorer countries should grow faster than richer ones4. What are the implications of neoclassical growth theory for 1 All data from Maddison, in http://www.ggdc.net/maddison/, retrieved March 2010. 2 Ie., the richest in 1900, a group including England, France, Germany, their economic hinterland (Denmark, Belgium, Netherlands, Switzerland, Austria) plus the four "Western Offshoots" (USA, Canada, Australia and New Zealand). The thirteenth was Argentina (77,7% of the richest twelve), and, after a gap, Uruguay (62,5%), Sweden (62,3%) and Chile (61,7%) followed. 3 Uruguayan population in 1860: extrapolation from Maddison's data for 1850 and 1870 using a constant rate of growth. Cattle and sheep from Gran Enciclopedia Rialp, 1991. 4 In addition, there could be some "technological catch up" as improving the backward technology of poorer countries requires the cheap expedient (of copy-pasting foreign techniques, rather than the more expensive process of developing new technologies. countries with high ratios of natural resources per capita? Should we expect convergence anyway? Does conditional convergence apply independently of the factorial combination behind a certain level of per capita GDP? Imagine two countries with a similar per capita GDP in 1910, one of them with more natural resources and less physical capital than the other, in per capita terms. Call them Argentina and the Netherlands (actually, in 1910, their per capita incomes were the same and they both had a population between 6 and 7 million). By 1910, Argentina had already reached its agricultural frontier so it's perhaps appropriate to consider land a fixed factor −though not an exhaustible one− by that time. Imagine, moreover, a similar technology and savings rate5 and let's leave aside for the moment the question of human capital. Think of a per capita production function of the form y = A.f(k,t) (1) where y is income per worker, A is the level of technology, k is capital per capita and t is land per capita. Using subscripts a and n for Argentina and the Netherlands, we are saying that y = y A = a n, a A , k <k and t >t . At this very basic level, and considering that t depends only on population n a n a n growth (because total land is fixed) should we expect Argentina to grow less than the Netherlands, more or the same? If both countries have the same savings rate and the same per capita income, and assuming they invest what they save (i.e., the closed economy case) then they would both be adding the same amount of capital to their current stock. If both have the same rate of population growth, then capital per worker will be increasing at the same rate in both countries. However, because physical capital is initially lower in Argentina than the Netherlands, if there are decreasing returns to capital then Argentina's income per worker should be growing more than that of the Netherlands6. Put differently: if natural resources play a relevant role, a resource-rich country (say, Argentina in 1910) could have higher capital productivity and thus grow more than another country which is poorer but more capital-intensive (say, Austria in 1910) − i.e., there could be divergence favoring the resource-rich. 5 Taylor (1998), though, argues forcefully that Argentina had a very low savings rate. 6 For example, with a Cobb-Douglass production function for equation (1), y=A.kα.tβ the rate of growth is g = g + α.g + β.g (2) y A k t where g is the growth rate of variable i. The growth of land per capita is −n, where n is population growth. If s is the i savings rate, the growth rate of capital is g = s.y/k− (n+δ) (3) k where δ is depreciation, so that (2) becomes g = g + α.s.y/k − α δ− (α+β).n. y A The higher the output-capital ratio, the faster economic growth is. In our example, Argentina has the same output than the Netherlands but a lower level of capital, so −ceteris paribus− it should be growing more. A final point on this mechanistic approach to growth: with a fixed amount of natural resources in the production function, it could appear that population growth would dilute "land" more rapidly than capital, leading to a lower rate of economic growth. That is not necessarily the case. If suddenly both Argentina and the Netherlands double their population, per capita GDP would fall by half in both cases if there are constant returns to scale, no matter what the factorial combination behind that income. Both could compensate for that increase in population by increasing capital − actually, as pointed before, Argentina could have an advantage here as it would need less investment to attain the original per capita income, as capital would be more productive there if there are decreasing returns. Higher population growth reduces economic growth with or without natural resources in the production function7. The message here is that there aren't obvious reasons in mainstream growth theory telling us that Argentina should have diverged from the rich as soon as incorporation of new land −a key to its earlier success− came to an end. The motives behind Argentina's decline need to be more subtle in trying to explain the dynamics of factor and technology accumulation. That doesn't exclude, of course, models in which natural resources can be a curse, in any of the many ways surveyed, for example, by Sachs and Warner (1997). Two- and three-sector models have been central to the debate on Argentina's growth difficulties8. Most of them touch upon the question of whether Argentina could have grown by persisting in its bet on its natural resources or if, rather, capital accumulation and technological advances necessarily required a structural transformation towards a more diversified economy − and the related question of whether that transformation would result from a market process or could only take place with government's assistance. Models in the endogenous growth tradition do probably make a difference between Argentina and Netherlands under the conditions described above. For example: Campante and Glaeser's paper in this volume show that Buenos Aires had lower levels of physical and human capital than Chicago. In models such as Lucas (1988), the level of human capital is a significant determinant of economic growth, as the rate of increase in human capital depends on its level, through externalities. A similar story could be made of technology or physical capital. Would Argentina fit in such a model or would it still be an outlier? Can a model along such lines explain the unstable timing of Argentina's decline, with periods of accelerated divergence (the 1980s) and some of moderate convergence (the 1960s)? The answer is far from obvious. The general point here is that even if arguments relating Argentina's subsequent development to its conditions at some point in its prosperous may be true, in any case there's nothing evident about them. In other words: Argentina's twentieth century economic performance is in fact a puzzle. There are no straightforward reasons why, contemplating Argentina in 1910 or 1928, one could predict Argentina's unfortunate divergence. It should come as no surprise that contemporary observers tended to be optimistic about Argentina's future, in 1900, the twenties or even as late as the immediate postwar.9 An almost-rich country turned almost-poor, Argentina is 7 And it's probably a factor of some significance to understanding Argentina's comparative decline. In 1910, Argentina's population was 2.4% of the population of the "richest twelve" (footnote 2); in 2000, it was 6.4%. 8 The list of explanations in this vein are too numerous to be listed here. Di Tella and Zymmelman (1967) and Díaz Alejandro (1970) are two examples. 9 One of them was Paul Samuelson: "In 1945 I was a young talented economist. I was at the height of my abilities. If someone had asked me what part of the earth would develop the fastest in the next 39 years, I would have said: Latin a likely outlier for many theories of economic growth. In what follows we present and briefly discuss some of the hypotheses that can be advanced to account for Argentina's exceptionalism. III. Major Hypotheses In this section, we present some of the historical context and summarize four over-arching explanations of Argentina’s painful 20th century. These explanations are not mutually exclusive, and indeed, many are closely connected. They do, however, map out much of the intellectual terrain associated with explaining Argentine exceptionalism. Just Say No to Exceptionalism: Not Rich then; Not Poor Now The first hypothesis is that, once we properly measure things, Argentina is not truly exceptional in any interesting economic sense. The hypothesis comes in two versions. The first is that Argentina is not particularly poor now. And the second version is that it was not truly rich at the turn of the 20th century. The first version of the hypothesis is that Argentina is in fact richer now than what GDP figures indicate. Corrections of GDP measures are not uncommon in developing countries and Argentina is no exception, with a large upward correction implemented in the national accounts implemented in 1993. A standard rationale behind such changes is a desire to incorporate the large informal sector that arises when regulations and market limitations proliferate under a relatively weak State. An adjustment of approximately 30% of GDP for Argentina is not unusual using the “monetary method” (this number comes from Ahumada, et al, 2003).10 In this spirit, some have argued that the usual approach to measure GDP has to be adjusted when the economy undergoes big changes in the number and/or quality of available products or when the tendency of consumers to substitute away from expensive products bias the price index. In this volume, Gluzman and Sturzenegger explore a related approach exploiting the change in trade regime that allowed for increased product variety during the 1990’s to derive large upward estimates of current levels of GDP. The second version of the hypothesis denying “exceptional” status to Argentina is that its initial position was far less promising than it seemed. While Argentina was relatively rich, it may not have been as rich as some studies have found and it was sorely lacking in many of the fundamental attributes of more developed nations. Its wealth was more of a temporary nature, a shock which has more in common with the booms in oil-producing nations during the 1970s than with the more permanent prosperity associated with well developed stocks of human and physical capital. Perhaps the most controversial variant of this hypothesis is the literal questioning of standard income numbers, such as those from Maddison (1983). Some have argued that Maddison overstates Argentine prosperity at the beginning of the twentieth century, in part because he underestimates the role of the expanding informal economy since the 1960’s. The implied corrections, however, are significant but not dramatic: while Maddison puts Total America − Argentina or Chile. There is a moderate climate there and a population with European roots... I was completely off the mark." Interview for Der Spiegel, 28:2005. 10 The idea, broadly, is to use an estimate of the amount of money held by agents in excess of what they need to finance official transactions (making assumptions about the relative velocity in the shadow and official economy).Another revision in the late 1990’s, updated the prices used and led to a downward correction. GDP in 1900 at 12,100 (in constant pesos; basis: 300,000 in the year 2000), the revised estimate of Gerchunoff and Llach (1998) is 10,800.11 Far more common is the view that while Argentina was relatively rich (see for example, Míguez, 2005), those riches didn't extend widely throughout the population and they were not accompanied by other common correlates of development. For example, Adelman (1995) and Engerman and Sokoloff (2000) have discussed the high level of inequality in Argentina at the turn of the century, particularly in the agricultural sector. According to this view, the United States managed to share land and prosperity to a much greater degree than Argentina, where large estates where far more common. As such, Argentina should be seen as a much poorer nation that managed only to enrich a tiny slice of its population. There is little doubt that Argentina had significantly less education than many other wealthy nations a century ago. For example, the primary school attendance rate in 1910 Argentina was 48% percent of that in France and 57% percent of that in Germany, despite the fact that Argentina was 29% and 14% percent richer, respectively, in terms of per capita GDP12. As the Llach essay in this volume illustrates, Argentina was catching up in terms of primary school enrollment, but it remained significantly below Western Europe and far below western offshoots, like the U.S., Australia and Canada, throughout the pre-World War II period. Just as pre-World War II Argentina seems to have less human capital than other wealthy nations, it also seems to have had less physical capital, at least if one excludes the great value of its land and livestock. The Campante and Glaeser paper in this volume compare industrial output and capital stocks in Buenos Aires and Chicago at the beginning of the 20th century. They find that there is a wide gap between the two cities. Value added per worker is far lower in Buenos Aires, and capital per worker is too. In some cases, capital per worker is more than 75 percent lower in Buenos Aires at then-contemporary exchange rates. The lower level of human and physical capital also seems linked to a technology gap between Argentina and many other western countries, at the turn of the last century. The Campante and Glaeser paper documents that Chicago was the home of many cutting edge industries, and the site of many significant inventions (e.g. the skyscraper). The same thing could be said of Detroit (mass produced cars), New York (alternating current), Paris (radioactivity), London (subways, vacuum cleaners) and Berlin (electric streetcars and elevators) at the same time. By contrast, Argentina was primarily an importer of technologies developed elsewhere. This hypothesis suggests that Argentina in 1910 should not be compared to other rich countries, because it lacked the key ingredients that make development durable. According to this view, Argentina was essentially an undeveloped economy made temporarily rich by an abundance of high quality land and better transportation technologies (which were again developed elsewhere). 11 The discrepancy occurs because early figures are estimated working backwards from current estimates (using growth rates) and there where upward corrections to GDP numbers in the 1990’s. Note that the revised estimates of Ferreres (2005) are very similar to those of Maddison. For the period before 1900 things are even sketchier; the best available estimates come from Della Paolera (1988) and Cortés Conde and Harriague (1994). 12 Data from Peter Lindert's database: www.econ.ucdavis.edu/faculty/fzlinder/Lindert%20data%20CUP%20book/App._T._A1__primary_enrol.xls As such, it isn’t surprising that Argentina had a bad 20th century—it just reverted to form. In a sense, the other hypotheses can be understood as explaining the channels through which Argentina’s lack of early 20th century resources resulted in less economic development. Bad Politics, Bad Policies There is some dispute as to whether Argentina was behind other advanced countries in 1900 in the state of its political development. Its policies, which seem to have preserved a reasonable amount of rule of law and which allowed free trade in goods, capital and people, seem to have been quite benign. Politically, Argentina was a republic, albeit one with a limited franchise until 1912, and strongly empowered local landowners. But there is no question that Argentina had significant political instability over the 20th century, and enacted many policies that seem to have been harmful for growth. The second hypothesis is that these problems are responsible for Argentina’s economic malaise. It is perhaps useful to briefly describe the historical context for such political instability. An important event is the 1916 election of Hipolito Yrigoyen, which replaced the conservative regime by the Radical party, and also brought into the country’s political life a large portion of the middle class. This was the beginning of a dramatic change in the way the country would be governed. The Radicales would wield power until 1930. While it is true that Yrigoyen enacted some policies, like minimum wage laws, that some economists would argue are detrimental to growth, there is little evidence that Argentina suffered during this period. The real watershed seems to have been in 1930, when a military coup brought down Yrigoyen’s second government. Argentina’s “returned to democracy” with the 1931 election, where the radicals were banned from participating. As the great depression impacted world trade, a more conservative regime was put in place by an elite-dominated coalition known as the “Concordancia”. It soon began Argentina’s turn inward by implementing more interventionist policies: public works were started, import duties were increased and a system of multiple exchange rates favored industrial activities (over agriculture). The resulting industrial growth led to some migration of rural workers to urban centers and to a changing composition of labor organizations. Real wages remained stagnant while the perception of concessions to foreign trading partners, principally Britain, irked nationalist sentiment. A 1943 military coup named Juan Domingo Perón to the hitherto harmless post of secretary of labor and social welfare. From there, Perón enacted a comprehensive set of pro-labor laws that included a scheme to establish and periodically adjust minimum wages, often leading to increases in real terms; yearly paid vacations; retirement and health insurance benefits; and an annual mandatory bonus equal to an additional month’s salary. He also instituted the Agricultural Worker Statute (Estatuto del Peon) in late 1944, which outlined the specific rights and obligations of both rural workers and employers and was perceived as a defiance of the landed elite. In 1945, he enacted the Law on Professional Associations, which gave his Labor Secretariat veto power on the formation of new unions. By the end of his tenure, Argentina had advanced to a world leader in labor legislation (see for example, the description in D’Abate, 1983). The nine years of Perón’s presidency starting in 1946 witnessed intense political polarization. Perón enacted policies that eventually antagonized the rich. He continued a set of pro-union policies: between 1946 and 1954, union membership increased from 880,000 to 2.5 million, which represented 42.5% of all workers (see for example, Smith, 1991). He also