Digitized by the Internet Archive in 2011 with funding from Boston Library Consortium Member Libraries http://www.archive.org/details/speculativegrowtOOcaba YQ 1 15 \.o2- Massachusetts Institute of Technology Department ot Economics Working Paper Series SPECULATIVE GROWTH Ricardo Caballero J. MASSACHUSETTS INSTITUTE Mohammad Hammour OFTECHNOLOGY L. JAN 2 1 2003 WORKING PAPER 02-45 LIBRARIES November 27, 2002 RoomE52-251 50 Memorial Drive MA Cambridge, 02142 This paper can be downloaded without charge from the Social Science Research Network Paper Collection at http://papers.ssrn.com/abstract=355901 Speculative Growth Ricardo J. Caballero Mohamad L. Hammour* This draft: November 27, 2002 Abstract We propose a framework for understanding recurrent historical episodes ofvigorous economic expansion accompanied by extreme asset valuations, as exhibited by Japan in the 1980s and the U.S. in the 1990s. We interpret this phenomenon as a high- valuation equilibrium with a low effective cost of capital based on optimism about the future availability of funds for investment. The key to the sustainability of such an equilibrium is feedback from increased growth to an increase in the supply of funding. We show that such feedback arises naturally when the expansion is concentrated in a — "new economy" sector and when it is supported by sustained fiscal surpluses both ofwhich would constitute an integral part, as cause and consequence, ofa "speculative growth" equilibrium. The high-valuation equilibrium we analyze may take the form of a stock market bubble. In contrast to classic bubbles on non-productive assets, the bubblesinourmodel encourage realinvestment, boostlong-runsavings, andmayappear in dynamically efficient economies. 'Respectively: MIT and NBER; DELTA and CEPR (DELTA is a joint CNRS-EHESS-ENS research unit). E-mail: [email protected]; [email protected]. We are gratefultoBjorn Brugemann foroutstanding research assistance and to the European Central Bank Research Department and the International Institute forEconomic Studiesfortheir hospitality. WethankFranklin Allen, Marios Angeletos, Gadi Barlevy, Olivier Blanchard, Thomas Chaney, PeterDiamond, Thomas Philippon, Frank Portier, PietroReichlin, Jean Tirole, Jaume Ventura and seminar participants at CORE, DELTA, ESSIM, IIES, MIT, the NBER EFG meetings, Northwestern, Toulouse, and Wharton foruseful comments. Caballerothanks the NSF for financial support. First draft: May 2001. 1 Introduction The last two decades have witnessed stark "speculative growth" episodes of extreme stock market valuations accompanied by brisk economic growth. Most notable was the experience ofJapan in the 1980s and the United States in the 1990s. Figure la illustrates the sharp rise in the Nikkei in the 1980s, followed by the two-year slide of 1990-1992; and the sharp rise in the NASDAQ in the 1990s, followed by the collapse of 2000-2001. The extremes reached by valuations shown in figure lb, and their collapse in the absence of obvious changes in fundamentals, are suggestive of a stock market bubble. Figures lc and Id illustrate the growth and investment boom-and-bust that accompanied the market's gyrations. La: The Nikkei 225 and the NASDAQ 1 b: Price/EarningsRatio 50 Indices 85 AA« 45 10 65 *' f\ \l \>\ 35o3n 830000 30 g20000 45 X7 y '"• *«*" 25 10000 20 95 - , , 1 1 , 1 j 1 1 1. 15 -5-4-3-2-1012 -5 -4 -3 -2 -1 1 2 3 4 5 3 4 5 YearsfromPeak YearfromPeak l.c: Real GDPGrowth (%) Id: Investment/GDP (%) -3-2-1012 -5-4-3-2-1012345 -5 -4 3 4 5 YearfromPeak YearfromPeak Note Panelb:numeratorisreal(inflation-adjusted)DaiastreamTotalMarketIndex;denominatorismovingaverageoverpreceding tenyearsofrealearningscorrespondingtotheindex. Sources:Panela: NasdaqCompositeIndexfromTheNasdaqStockMarket.Inc;Nikkei225fromDaiastream.Panelb:Dataslieam TotalMarketIndexforUSandJapan.CPIsfromOECDMainEconomicIndicators(allitems).Panelc:BureauofEconomicAnalysis CBEA),NIPATable1.2;EconomicandSocialResearchInsitute(ESRJ).Paneld:BEA,N7PATable1.1;Annualreportonnational accounts(CD-ROM)1998 Figure 1: Speculative Growth Episodes The nature and policy dilemmas of speculative expansions have attracted much atten- tion (e.g., International Monetary Fund 2000, Shiller 2000, Cecchetti et al. 2000), but our formal understanding ofthe macroeconomic mechanisms that underlie the relation between stock market speculation and real economic activity remains quite limited. In fact, the clas- sic literature on bubbles in general equilibrium bears little relation to speculative growth — episodes. In this literature, bubbles understood as asset prices that deviate from funda- — mental values can only appear in a "dynamically inefficient" economy, that is one whose structure is conducive to the over-accumulation ofcapital. Their emergence helps to absorb saving away from investment and to alleviate the over-accumulation problem. However, the notion that the emergence ofa bubble would crowd out investment, while its bursting would increase investment, is contrary to the patterns depicted in figure 1. Moreover, empirical testing of "dynamic inefficiency" has been negative. Abel et al. (1989) tested an implication of dynamic inefficiency, whether the aggregate market for assets acts as a long-term "cash sink" for investors, but found no evidence of it in OECD economies. This paper proposes an alternative framework for understanding speculative expansions that is better able to address the facts. Our interpretation of this phenomenon rests on two premises: (i) the rise in market valuations in speculative growth episodes results from a low effective long-term cost of capital; and (ii) the boom in investment and growth during such episodes is a response to increased asset prices along the standard g-theory line. The central question that arises under this interpretation is: How are expectations of a low long-term cost of capital consistent with the high demand for savings needed to fund a high-investment equilibrium? A key component of our equilibrium account is a growth- saving feedback by which the future supply ofsavings increases as a result of the conditions created by a speculative expansion, and ends up lowering the effective cost of capital. The kind of stock market bubbles that can arise in the presence of growth-saving feed- back have very different implications from standard bubbles. First, rather than crowding out capital accumulation, they naturally encourage investment and growth. Conversely, their bursting leads to an investment slump as the economy rids itself of what appears ex post as widespread overcapacity. Second, because they induce greater saving, such bubbles are expected to reduce the long-run cost of capital; this, one can show, allows them to arise in a dynamically efficient context. Finally, such bubbles and the resulting specula- tive expansions are manifestations of a more general class of low cost-of-capital equilibria that are the result of the feedback from growth to saving. Such equilibria are qualitatively similar to bubbly equilibria, in that they exhibit high asset valuations and can involve a collapse in asset prices in the absence of any news. However, these equilibria may or may not incorporate a bubble in asset markets, depending on how extreme valuation levels are. In other words, although consistent with market participants' perception of a stock market bubble, our account of speculative expansions does not hinge on ascertaining the presence of a bubble. In the particular case ofthe U.S. in the 1990s, we argue, at least two factors created the conditions for a speculative growth episode: the emerging information technology sector and conservative fiscal policy. Both factors created favorable conditions for growth-saving feedback and for the possibility of a speculative equilibrium characterized by extreme stock market valuations and a potential crash. How does the information technology sector generate feedback from growth to saving? This is an emerging sector ofthe economy, subject to a lesser degree ofdiminishing returns than more mature sectors. As the economy expands, such a "neweconomy" sector naturally will expand proportionally more than the rest of the economy and will have a greater ability to generate income for saving. Moreover, in an endogenous growth version of this model, we show that a bubble not only might be necessary for productivity growth to accelerate, but that the resulting expansion in productivity, income, and saving also might be critical to creating the growth-saving feedback necessary to support the bubble. Thus, a fundamental expansion in productivity and stock market speculation are not necessarily — competing explanations. A technological revolution might form an integral part both as — cause and consequence of a speculative growth equilibrium. What about fiscal policy? We note that a significant share of increased saving in the recent U.S. speculative expansion is attributable to the public sector. To the extent that pro-cyclical government revenues increase public saving, they reinforce the feedback from growth to saving. In the short run, fiscal surpluses can arise as a consequence of the stock market boom and can facilitate the initial rise in investment. More importantly, fiscal surpluses can play a central role in making the speculative equilibrium feasible by providing the funding necessary to sustain high investment in the long run. This consideration has a particularly stark implication for fiscal policy during a speculative growth episode. The possibility ofusingthe fiscal surplusesthat result fromsuchexpansionstocut taxesand raise spending might be illusory, as these surpluses might be necessary to sustain the speculative growth equilibrium that generated them. Thespeculativegrowth episodes to which ourtheoryrelates arearecurrent phenomenon. Earlier episodes of vigorous economic expansion under speculative asset valuations have been documented by economic historians. In the case of the U.S., this phenomenon also can be observed in the expansions ofthe turn ofthe 20th century, the 1920s, and the 1960s (see, e.g., Shiller 2000). Equally important, our theory can be brought to bear on sustained lowcost-of-capital equilibria, such astheprolonged expansionsexhibited in anumberofEast Asian economies in the post-war period. In fact, the key feedback from growth to saving in these economies has been documented. Examining the aggregate relationship between income growth and saving in a cross-country panel of 64 countries over the period 1958- 1987, Caroll and Weil (1994) find that growth Granger causes saving with a positive sign, but that saving does not Granger cause growth. The pattern of an acceleration in growth followed by strong increases in saving rates is particularly clear in the high-growth, high- saving East Asian economies of Japan, South Korea, Singapore, and Hong Kong. Gavin et al. (1997) elaborate on this evidence and show that the estimated impact of growth on saving is not only statistically significant but also very large in economic terms. Relation to the literature In its economic theme, this paper is part of a long tradition of studies of speculative growth episodes (e.g., Kindleberger 1989). In terms of the recent U.S. experience, this literature is divided between those who see a case of speculative behavior (e.g., Shiller 2000), and those who, based on the significant acceleration in underlying U.S. productivity growth, conclude that the expansion was driven by a technological revolution that affected real fundamentals (e.g., Greenwood and Jovanovic, 1999; Hobijn and Jovanovic, 2000). From this perspective, our contribution is to provide a unified perspective under which these two views need not be mutually exclusive. On the contrary, an expansion of technological opportunities and stock market speculation may be inseparable complements.1 On the methodological side, our paper belongs to the literature on bubbles in general equilibrium. As we mentioned earlier, the classic theory of bubbles in general equilibrium (see Tirole 1985) is not designed well for addressing speculative growth episodes. It implies that a bubble crowds out investment and only can arise in a dynamically inefficient econ- omy. Neither of these elements is observed during these episodes. From this perspective, the main distinctive feature ofour framework is the centrality of the growth-saving mecha- nism. The latter overturns standard conclusions by breaking the crowding-out mechanism of equilibrium bubbles. Moreover, bubbly equilibria in our setting are a particular case of low cost-of-capital equilibria, that exhibit the same qualitative features but do not hinge 'The standard "technological revolution" interpretation faces a double challenge. First, in order to explain the collapseofthe market, it must assume a reversal oftechnological market expectations that finds no counterpart in the commentary of market participants and their broad consensus that a bubble had burst. Second, this interpretation does not match well with the decline during the expansion of long-term interest rates and other measures ofthe cost ofcapital. The interpretation we offer in this paper shows that these two aspects of the US expansion are not only compatible with a technological revolution, but might be essential factors in its emergence. on ascertaining the existence of a bubble. More recently, several papers have demonstrated that, in the presence of externalities that create a wedge between private and social returns on investment, bubbles can arise even if the bubbleless economy is dynamically efficient (e.g., Saint-Paul, 1992; Grossman and Yanagawa, 1993; King and Ferguson, 1993). That is because dynamic efficiency requires that the social rate of return exceed the rate ofgrowth ofthe economy; the possibility of a bubble instead requires that the private rate of return be below the rate of growth. These two conditions are simultaneously feasible in the presence of a wedge between private and social returns. However, the cash-flow test for bubbles developed by Abel et al. (1989) still applies in the case of externalities because it concerns private returns, and rejects the possibility of standard bubbles. In contrast, our mechanism does not depend on a wedge between private and social returns, and abubble can arise even ifthe speculative path starts with a private return that exceeds the rate of growth of the economy. In particular, our mechanism passes Abel et al.'s test. In standard models, bubbles reduce investment and growth by crowding out investment. Olivier (2000) contributed to this literature by developing a model of bubbles on firm creation which, like ours, leads to more rather than less investment.2 However, because he does not incorporate a growth-saving feedback, bubbles in his model raise rather than lower the long-term cost of capital. This makes it challenging for a bubble to account simultaneously for high stock market valuations and high investment. In fact, for the bubble in Olivier's model to be price-increasing (see Weil 1990) -- i.e., for the negative effect on asset prices caused by the rise in the interest rate to be dominated by the positive — effect of the bubble a large degree of ex-ante dynamic inefficiency is required. The rest of this paper is organized as follows. In section 2, we present a prototypical model of speculative growth. Our analysis is based on the Diamond (1965) overlapping- generations model, to which we add adjustment costs and a special assumption about the saving function. Adjustment costs allow us to characterize stock market booms and their key role in facilitating the transition from low- to high-investment equilibria. Our special saving function generically captures the growth-saving feedback that is central to the existence of a high-valuation equilibrium. Section 3 presents a specific instance of the generic mechanisms ofsection 2 in a model of the "new economy." In section 4, we extend our analysis to encompass fiscal surpluses and current account deficits as short- and long- run funding mechanisms. Section 5 concludes and is followed by an appendix that presents 2See also Cozzi (1998) for a related model applied to education decisions (culture as a bubble). the proofs of the main propositions. 2 Speculative Growth: The Mechanics In this section we present a prototypical model ofspeculative growth. Our analysis is based on the Diamond (1965) overlapping-generations model, to which we add adjustment coststo capital and aspecial assumption about the saving function. Adjustment costs are important in any model in which changes in stock market valuations are central; otherwise the price ofcapital would be constant. The special assumption about the saving function is intended to capture the growth-saving feedback that is central to our interpretation of speculative expansions. This section introduces the feedback generically through the functional form of the saving function. The next two sections pay closer attention to specific macroeconomic — — mechanisms the "new economy" and fiscal conservatism that can create feedback from growth to saving and help explain recent speculative growth episodes. 2.1 Model Setup Consider a standard Diamond (1965) overlapping-generations structure with no population growth and a unit mass ofyoung and old agents who coexist at any datei. Each generation is born with a unit of labor, Lt = L = 1, to be used when young, for which it receives a total wage Wt determined in a competitive, full-employment labor market. The economy's single consumption good is used as a numeraire. Technology. Consumption goods are produced with capital, Kt, and labor, Lt- The pro- duction function at any time t is determined by the level of technology, At, which grows at an exogenous rate 7: At+i = (1+7)^*- Production is given by a constant returns technology: Yt = F(KuAtLt) = AtLf(kt), kt = Kt/AtLu where /' > 0, /" < 0, and lim^of'(k) = +00. Purely technical conditions on this and other functions are stated in appendix A. The labor market is competitive, and the wage wt = Wt/AtL per unit of effective labor is wt = w(kt) = f(kt)-ktf'(kt), w'>Q,w"<0. (1)