NBER WORKING PAPER SERIES SEGMENTED ASSET MARKETS AND OPTIMAL EXCHANGE RATE REGIMES Amartya Lahiri Rajesh Singh Carlos A. Vegh Working Paper 13154 http://www.nber.org/papers/w13154 NATIONAL BUREAU OF ECONOMIC RESEARCH 1050 Massachusetts Avenue Cambridge, MA 02138 June 2007 This paper was published in the Journal of International Economics (May 2007; pp. 1-21) and is reprinted here with permission from the Journal of International Economics and Elsevier. We would like to thank Andy Atkeson, Mick Devereaux, Huberto Ennis, Andy Neumeyer, Mark Spiegel, and seminar participants at Duke, FRB Cleveland, FRB NY, Penn State, UBC, UCLA, UC Santa Cruz, USC, Warwick, CMSG 2003, ITAM-FBBVA Summer Camp 2003, SED 2003, and NBER IFM meeting Fall 2003 for helpful comments and suggestions. The usual disclaimer applies. Végh would like to thank the UCLA Academic Senate for research support. The views expressed herein are those of the author(s) and do not necessarily reflect the views of the National Bureau of Economic Research. © 2007 by Amartya Lahiri, Rajesh Singh, and Carlos A. Vegh. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is given to the source. Segmented Asset Markets and Optimal Exchange Rate Regimes Amartya Lahiri, Rajesh Singh, and Carlos A. Vegh NBER Working Paper No. 13154 June 2007 JEL No. F3,F40,F41 ABSTRACT This paper revisits the issue of the optimal exchange rate regime in a flexible price environment. The key innovation is that we analyze this question in the context of environments where only a fraction of agents participate in asset market transactions (i.e., asset markets are segmented). Under this friction, alternative exchange rate regimes have different implications for real allocations in the economy. In particular -- and contrary to standard results under sticky prices -- we show that flexible exchange rates are optimal under monetary shocks and fixed exchange rates are optimal under real shocks. Amartya Lahiri Carlos A. Vegh Department of Economics Department of Economics University of British Columbia Tydings Hall, Office 4118G Vancouver, BC V6T 1Z1 University of Maryland [email protected] College Park, MD 20742-7211 and NBER Rajesh Singh [email protected] Department of Economics Iowa State University 280D Heady Hall Ames, IA 50011 [email protected] 1 Introduction Fifty years after Milton Friedman(cid:146)s (1953) celebrated case for (cid:135)exible exchange rates, the debate on the optimal choice of exchange rate regimes rages on as (cid:133)ercely as ever. Friedman argued that, in the presence of sticky prices, (cid:135)oating rates would provide better insulation from foreign shocks by allowing relative prices to adjust faster. In a world of capital mobility, Mundell(cid:146)s (1963) work implies that the optimal choice of exchange rate regime should depend on the type of shocks hitting an economy: real shocks would call for a (cid:135)oating exchange rate, whereas monetary shocks would call for a (cid:133)xed exchange rate. Ultimately, however, an explicit cost/bene(cid:133)t comparison of exchange rate regimes requires a utility-maximizing framework, as argued by Helpman (1981) and Helpman and Razin (1979). In such a framework, Devereux and Engel and (2003) reexamine this question in a sticky prices model and show how results are sensitive to whether prices are denominated in the producer(cid:146)s or consumer(cid:146)s currency. On the other hand, Cespedes, Chang, and Velasco (2000) incorporate liability dollarization and balance sheets e⁄ects and conclude that the standard prescription in favor of (cid:135)exible exchange rates in response to real shocks is not essentially a⁄ected. An implicit assumption in most, if not all, of the literature is that economic agents have un- restricted and permanent access to asset markets.1 This, of course, implies that in the absence of nominal rigidities, the choice of (cid:133)xed versus (cid:135)exible exchange rates is irrelevant. In practice, however, access to asset markets is limited to some fraction of the population (due to, for example, (cid:133)xed costs of entry). This is likely to be particularly true in developing countries where asset mar- kets are much smaller in size than in industrial countries. Table 1 shows that even for the United States, the degree of segmentation in asset markets is remarkably high. The table reveals that, as of 1989, 59 percent of U.S. households did not hold any interest bearing assets (de(cid:133)ned as money market accounts, certi(cid:133)cates of deposit, bonds, mutual funds, and equities). More strikingly, 25 1There are some exceptions when it comes to the related issue of the costs and bene(cid:133)ts of a common currency area (see, for example, Neumeyer (1998) and Ching and Devereux (2003), who analyze this issue in the presence of incomplete asset markets). 1 percentofhouseholdsdidnotevenhaveacheckingaccountaslateasin1989. Giventhesefactsfor a developed country like the United States, it is easy to anticipate that the degree of asset market segmentation in emerging economies must be considerably higher. Since asset markets are at the heart of the adjustment process to di⁄erent shocks in an open economy, it would seem natural to analyze how asset market segmentation a⁄ects the choice of exchange rate regime.2 Table 1 here Thispaperabstractsfromanynominalrigidityandfocusesonastandardmonetarymodelofan economy subject to stochastic real and monetary (i.e., velocity) shocks in which the only friction is that an exogenously-given fraction of the population can access asset markets. The analysis makes clear that asset market segmentation introduces a fundamental asymmetry in the choice of (cid:133)xed versus (cid:135)exible exchange rates. To see this, consider (cid:133)rst the e⁄ects of a positive velocity shock in a standard one-good open economy model in the absence of asset market segmentation. Under (cid:135)exible exchange rates, the velocity shock gets re(cid:135)ected in an excess demand for goods, which leads to an increase in the price level (i.e., the exchange rate). Under (cid:133)xed exchange rates, the adjustment must take place through an asset market operation whereby agents exchange their excess money balances for foreign bonds at the central bank. In either case, the adjustment takes place instantaneously with no real e⁄ects. How does asset market segmentation a⁄ect this adjustment? Under (cid:135)exible rates, the same adjustment takes place. Under (cid:133)xed exchange rates, however, only those agents who have access to asset markets (called (cid:147)traders(cid:148)) may get rid of their excess money balances. Non-traders (cid:150)who are shut o⁄from assets markets (cid:150)cannot do this. Non-traders are therefore forced to buy excess goods. The resultant volatility of consumption is costly from a welfare point of view. Hence, under asset market segmentation and in the presence 2In closed economy macroeconomics, asset market segmentation has received widespread attention ever since the pioneering work of Grossman and Weiss (1983) and Rotemberg (1984) (see also Chatterjee and Corbae (1992) and Alvarez, Lucas, and Weber (2001)). The key implication of these models is that open market operations reduce the nominalinterestrateandtherebygeneratetheso-called(cid:147)liquiditye⁄ect(cid:148). Inanopeneconomycontext,Alvarezand Atkeson (1997)and Alvarez,Atkeson,and Kehoe (2002)have argued thatassetmarketsegmentation modelshelp in resolving outstanding puzzles in international (cid:133)nance such as volatile and persistent real exchange rate movements as well as excess volatility of nominal exchange rates. 2 of monetary shocks, (cid:135)exible exchange rates are superior to (cid:133)xed exchange rates. Asset market segmentation also has crucial implications for the optimal exchange rate regime when shocks come from the goods market. We show that when output is stochastic, non-traders in the economy unambiguously prefer (cid:133)xed exchange rates to (cid:135)exible exchange rates because pegs provide a form of risk pooling. Under a peg, household consumption is a weighted average of current period and last period(cid:146)s output which implies that the consumption risk of non-trading households is pooled across periods. Under (cid:135)exible rates, however, the real value of consumption is always current output which implies no intertemporal risk sharing. Trading households, on the other hand, prefer (cid:135)exible exchange rates to (cid:133)xed exchange rates since maintaining an exchange rate peg involves injecting or withdrawing money from traders which makes their consumption more volatile under a peg. However, trading households(cid:146)access to asset markets ensures a much smaller increase in their consumption volatility relative to the reduction in consumption volatility of non-trading households. Using a population share weighted average of the welfare of the two types, we show that under fairly general conditions, the non-traders(cid:146)preferences dominate the social welfare function. Hence, when output is stochastic, an exchange rate peg welfare dominates a (cid:135)exible exchange rate regime. In sum, the paper shows that asset market segmentation may be a critical friction in determining the optimal exchange rate regime. Our paper is related to an older literature on exchange rate regimes. Perhaps the closest paper is Fischer (1977) who showed that in an economy with no capital mobility, (cid:133)xed exchange rates produced better outcomes than (cid:135)exible exchange rates when shocks are real while (cid:135)exible exchange rates are better when shocks originate in the money market. There are two main di⁄erences between Fischer (1977) and our paper. First, we solve a micro-founded optimizing model while Fischer obtained his results in the context of a reduced-form model. The reduced-form nature of the model made his analysis unsuitable for a choice-theoretic welfare analysis. Second, we analyze an economy with heterogenous agents whereas Fischer did not. In our model, agent heterogeneity is key to understanding the role of monetary policy in a⁄ecting real outcomes. In our framework, monetary policy can react to output disturbances by redistributing resources from one agent to the 3 other. This channel is critical in achieving the (cid:133)rst-best in our model and is missing completely in Fischer(cid:146)s model. In fact, it is this heterogeneity of agents which also di⁄erentiates our work from the work of Helpman and Razin (1982). Helpman-Razin studied an environment with uncertainty and incomplete markets to show that (cid:135)exible exchange rates can produce higher welfare than (cid:133)xed exchange rates. However, the key feature of our model is incomplete market participation (cid:150)some agents are absent from asset markets. In a previous version of this paper we have shown that our results carry over to the complete asset markets case. Hence, what is central for our results is incomplete market participation, not incomplete asset markets. The paper proceeds as follows. Section 2 presents the model and the equilibrium conditions while Section 3 describes the allocations under alternative exchange rate regimes and derives the optimal regime under monetary and output shocks. Section 4 studies the optimal, (cid:133)rst-best monetary policy rule. Finally, Section 5 concludes. Algebraically tedious proofs are consigned to an appendix. 2 Model The basic model is an open economy variant of the model outlined in Alvarez, Lucas, and Weber (2001). Consider a small open economy perfectly integrated with world goods markets. There is a unit measure of households who consume an internationally-traded good. The world currency price of the consumption good is (cid:133)xed at one. The households(cid:146)intertemporal utility function is 1 W = E (cid:12)s tu(c ) ; (1) t t (cid:0) s ( ) s=t X where (cid:12) is the households(cid:146)time discount factor, c is consumption in period s, while E denotes s t the expectation conditional on information available at time t. The households face a cash-in-advance constraint. As is standard in these models, the house- holds are prohibited from consuming their own endowment. We assume that a household consists of a seller-shopper pair. While the seller sells the household(cid:146)s own endowment, the shopper goes 4 out with money to purchase consumption goods from other households. We assume that house- holds are heterogenous. In particular, only a fraction (cid:21) of the population, called traders, have access to the asset markets, where 0 < (cid:21) 1. The rest, 1 (cid:21), called non-traders, can only hold (cid:20) (cid:0) domestic money as an asset. We restrict (cid:21) to be strictly positive. As will become clearer below, the model has a discontinuity at (cid:21) = 0. Since all money injections occur in asset markets, the monetary authority has no way of introducing money into the economy if there are no traders at all. There are two potential sources of uncertainty in the economy. First, each household receives a random endowment y of the consumption good in each period. We assume that y is an t t independently and identically distributed random variable with mean y(cid:22)and variance (cid:27)2.3 Second, y following Alvarez et al., we assume that the shopper can access a proportion v of the household(cid:146)s t current period (t) sales receipts, in addition to the cash carried over from the last period (M ), to t purchase consumption. We assume that v is an independently and identically distributed random t variable with mean v(cid:22) (0;1) and variance (cid:27)2. In the following we shall refer to these v shocks as v 2 velocity shocks.4 The timing runs as follows. First, both the endowment and velocity shocks are realized at the beginning of every period. Second, the household splits. Sellers of both households stay at home and sell their endowment for local currency. Shoppers of the non-trading households are excluded from the asset market and, hence, go directly to the goods market with their overnight cash to buy consumption goods. Shoppers of trading households (cid:133)rst carry the cash held overnight to the asset market where they trade in bonds and receive any money injections for the period. They then proceed to the goods market with whatever money balances are left after their portfolio 3Wecouldallowfordi⁄erentmeansandvariancesfortheendowmentsoftradersandnon-traderswithoutchanging our basic results. 4There are alternative ways in which one can think about these velocity shocks. Following Alvarez, Lucas, and Weber(2001)onecan(cid:145)thinkoftheshopperasvisitingtheseller(cid:146)sstoreatsometimeduringthetradingday,emptying thecashregister,andreturningtoshopsomemore(cid:146). Theuncertaintyregardingvcanbethoughtofastheuncertainty regarding the total volume of sales at the time that the shopper accesses the cash register. Alternatively, one can think of this as representing an environment where the shopper can purchase goods either through cash or credit. However, the mix of cash and credit transactions is uncertain and (cid:135)uctuates across periods. 5 rebalancing. After acquiring goods in exchange for cash, the non-trading-shopper returns straight home while the trading-shopper can re-enter the asset market to exchange goods for foreign bonds. Afteralltradesforthedayarecompletedandmarketsclose, theshopperandthesellerarereunited at home. 2.1 Households(cid:146)problem 2.1.1 Non-traders The non-trader(cid:146)s cash-in-advance constraint is given by: MNT +v S y = S cNT; (2) t t t t t t where MNT is the beginning of period t nominal money balances while S denotes the domestic t t currency price of consumption goods at time t. Perfect goods mobility then implies that S is t also the period t exchange rate (i.e., the domestic currency price of foreign currency). Equation (2) shows that for consumption purposes, the non-traders can augment the beginning of period cash balances by withdrawals from current period sales receipts v (the velocity shocks). Notice t that while writing (2) we have assumed that the cash-in-advance constraint binds in equilibrium.5 Appendix 6.1 provides su¢ cient conditions to ensure that (2) indeed holds for all t. Money balances at the beginning of period t+1 are given by sales receipts net of withdrawals for period t consumption: MNT = S y (1 v ): (3) t+1 t t t (cid:0) The usual (cid:135)ow constraint follows from combining (2) and (3): MNT = MNT +S y S cNT: (4) t+1 t t t t t (cid:0) Given the cash-in-advance (2), it follows that: MNT +v S y cNT = t t t t: (5) t S t 5Relaxingthisassumptionwilladdcomplexitytoouranalysiswithoutqualitativelychangingoraddinganyfurther insights to the results we obtain. Note further that it is a standard assumption in the literature; see for example Alvarez, Atkeson, and Kehoe (2002), and Alvarez, Lucas, and Weber (2001). 6 2.1.2 Traders The traders begin any period with assets in the form of money balances and bond holdings carried over from the previous period. Armed with these assets the shopper of the trader household visits the asset market where she rebalances the household(cid:146)s asset position and also receives the lump sum asset market transfers from the government. Thus, for any period t; the accounting identity for the asset market transactions of a trader household is given by B B T M^T = MT +(1+i ) t t+1 +S (1+r)f S f + t; (6) t t t(cid:0)1 (cid:21) (cid:0) (cid:21) t t(cid:0) t t+1 (cid:21) where M^T denotes the money balances with which the trader leaves the asset market and MT t t denotes the money balances with which the trader entered the asset market. Also, B denotes aggregate one-period nominal government bonds, i is the interest rate on these bonds, f are foreign bonds (denominated in terms of the consumption good), r is the exogenous and constant world real interest rate, and T are aggregate (nominal) lump-sum transfers (i.e., negative taxes) from the government.6;7 Note that nominal bonds maturing at date t pay an interest rate i since this t 1 (cid:0) rate was contracted in t 1.8 (cid:0) Afterassetmarketsclose,theshopperproceedstothegoodsmarketwithM^T innominalmoney balancestopurchaseconsumptiongoods. Likenon-traders,traderscanalsoaugmentthesestarting money balances with random withdrawals from current sales receipts to carry out goods purchases. 6We assume that these transfers are made in the asset markets, where only the traders are present. Note that since B and T denote aggregate bonds and aggegate transfers, their corresponding per trader values are B=(cid:21) and T=(cid:21) since traders comprise a fraction (cid:21) of the population. 7The assumption of endogenous lump-sum transfers will ensure that any monetary policy may be consistent with the intertemporal (cid:133)scal constraint. This becomes particularly important in this stochastic environment where these endogenous transfers will have to adjust to ensure intertemporal solvency for any history of shocks. To make our life easier, these transfers are assumed to go only to traders. If these transfers also went to non-traders, then (5) would be a⁄ected. 8InanunpublishedAppendixweshowthatthekeyresultsobtainedherecarryovertothecompletemarketscase. The appendix is available from the authors upon request. 7 Thus, the cash-in-advance constraint for a trader is given by9 S cT = M^T +v S y : (7) t t t t t t Combining equations (6) and (7) gives T B B MT + t +v S y = S cT + t+1 (1+i ) t +S f S (1+r)f : (8) t (cid:21) t t t t t (cid:21) (cid:0) t(cid:0)1 (cid:21) t t+1(cid:0) t t In this set-up the only reason that traders hold money overnight is the separation between markets. In particular, if the seller could access the asset market at the end of the day, then the trading household would use all their remaining sales receipts from the period to buy interest bearing bonds. Thus, period-t sales receipts net of withdrawals become beginning of next period(cid:146)s money balances: MT = S y (1 v ): (9) t+1 t t t (cid:0) Note that since v, S; and y are all exogenous, the traders(cid:146)money holdings evolve exogenously over time. Atraderchoosesc ,B andf tomaximize(1)subjecttothe(cid:135)owconstraint(8). Combining t t+1 t+1 (cid:133)rst-order conditions, we obtain: u(cT) = (cid:12)(1+r)E u(cT ) ; (10a) 0 t t 0 t+1 u0(cTt ) = (cid:12)(1+i )E(cid:8) u0(cTt+(cid:9)1) : (10b) t t S S t ( t+1 ) Equation (10a) is the standard Euler equation for the trader which relates the expected marginal rate of consumption substitution between today and tomorrow to the return on savings (given by 1+r) discounted to today. Equation (10b), on the other hand, determines the optimal holdings of nominal bonds. Equations(10a)and (10b)jointlydetermine the modi(cid:133)ed interestparitycondition for this economy which re(cid:135)ects the standard portfolio choice between safe and risky assets. 9In equilibrium i > 0, which implies that the cash-in-advance constraint for traders always binds. It can be t shownthatthesu¢ cientconditionsfornon-traders(cid:146)cash-in-advanceconstraintstobind,derivedinappendix6.1,are su¢ cient to ensure i >0 in equilibrium. t 8