International Monetary Policy Coordination and Financial Market Integration

Transcription

1

2 An important paper that opens an important conference. In my discussion I will attempt to: cast the paper within the broader context of the current literature and debate on coordination; suggest an interpretation of its results; evaluate its empirical implications.

3 For my generation, Analysis of monetary coordination = Mundell-Fleming meets game theory Canzoneri-Henderson [1991] book as key reference: analysis of monetary interdependence similar to analysis of tariff wars in the trade literature Essential message: there are significant gains from coordinated response to macroeconomic shocks Typical example: uncoordinated reaction to inflationary pressures (oil price shock) leads to excessive monetary tightening and output contraction in the world economy.

4 However: cooperation needs commitment (binding agreement) and imposes constraints on sovereignty: difficult to achieve and maintain reputation concerns in repeated games can make cooperation unnecessary: policymakers do the right thing even in the absence of coordination; cooperation may be counterproductive [Rogoff (JIE 1985)]. in practice: institutional attempts at cooperation in the mid 80s (Plaza, Louvre) problematic and short-lived;

5 By the mid 1990s, the literature on coordination was stagnant. Representative quote: Virtually all of the literature [on international monetary policy cooperation] is based on obsolete Keynesian models, which lack the microfoundations needed for proper welfare analysis... Because ad-hoc Keynesian analyses of cooperation can yield seriously misleading policy prescriptions, there is a compelling case for basing policy-coordination analysis on choicetheoretic models. Maurice Obstfeld and Kenneth Rogoff, Foundations of International Macroeconomics, pp , published 1996.

6 Second half of the 1990s: resurgence of interest in the analysis of interdependence: Early New Open Economy Macroeconomics [see survey by Lane (JIE 2001)]: microfounded analysis of international transmission of (monetary) shocks [Obstfeld and Rogoff (JPE 1995)] microfounded interdependent responses to shocks look very much like the familiar Canzoneri-Henderson reaction functions [Corsetti and Pesenti (QJE 2001)] monetary transmission can be beggar-thy-neighbor or prosper-thy-neighbor depending on structural parameters such as elasticities of substitution [Tille (JIE 2001)] and pass-through coefficients [Betts and Devereux (EER 1996)] there are gains from cooperation, but cooperation can be counterproductive [Benigno (JIE 2002)]

7 Newer Open Economy Macroeconomics (referred to as workhorse model by Canzoneri, Cumby and Diba [2002]) builds on previous results in a fully stochastic setting with forward-looking price (wage) setters; shift of emphasis from ex-post coordination (optimal response to shocks) to ex-ante coordination (optimal design of monetary rules among interdependent policymakers)

8 The key result: Obstfeld and Rogoff [QJE 2002]: everything considered, welfare gains from monetary coordination are very small; optimizing policymakers maximize global welfare even when they act in a non-coordinated way. Devereux and Engel [2000], Corsetti and Pesenti [2001]: gains from coordination are small even when exchange rate pass-through is low (leaving open possibility of significant gains for intermediate degrees of pass-through). Substantially the new literature confirms (provides theoretical foundations to) something that monetary policymakers and central bankers have long understood and recognized...

9 The reaction: these results depend on the specific parameterization of the model; for a more general specification, the gains from cooperation are (theoretically) large; This point is made in a form or another in several recent papers: Benigno and Benigno [2001] Clarida, Gali and Gertler [2001] Canzoneri, Cumby and Diba [2002] Sutherland [2002]...

10 Sutherland [2002]: the problems with the workhorse model are: the elasticity of substitution between home and foreign goods is restricted to unity; the structure of financial markets is not considered.

11 The unit elasticity problem This parameter determines the strength of the expenditure switching effect of monetary policy and is therefore an important determinant of the spillover effect of monetary policy. Everybody agrees that the elasticity of substitution between domestic goods and imports should be lower than the elasticity of substitution between brands (larger than one). Not everybody agrees that this elasticity should be exactly one. The model becomes extremely tractable if the elasticity is set equal to one (closed-form solution). But tractability comes at the price of generality. Sutherland [2002] shows how to circumvent the tractability problem: work with non-linear (quadratic) approximations. Major methodological achievement.

12 Consider the case of financial autarchy with high elasticity of substitution: Social loss increases (welfare falls) when: Volatility of domestic output gap increases (with price staggering we would have domestic inflation as well) Volatility of foreign output gap increases. So far nothing controversial (also in the workhorse model ).

13 Social loss falls (welfare increases) when: Volatility of exchange rate increases! The latter result represents the big difference wrt the standard model. All the results of the paper stem from the positive effects of exchange rate volatility on social welfare: Policymakers care about the variance of the exchange rate (they want more!). Under coordination the exchange rate is more volatile, and this underlies substantial gains from coordination. It is very counterintuitive!

14 Explanations (the Sutherland effect): the CPI is concave in the prices of the goods: exchange rate volatility reduces expected CPI and increases welfare when home and foreign goods are highly substitutable, agents can reduce the average cost of their consumption basket by switching expenditure towards whichever set of goods are cheapest ex post. Relative price variability is therefore a utility benefit. Other possible explanation: exchange rate volatility can be good or bad for firms profits according to parameter values [Bacchetta and Van Wincoop (2002)].

15 A conjecture (and analogy): In international trade, when two large countries trade with each other, each country has monopoly power on trade (large country externality: a reduction in supply of exports improves the term of trade). Taken the behavior of the other country as given, each country has an incentive to impose an optimal tariff, tilting the terms of trade in its favor. But the other country retaliates and there is a trade war. The final outcome (Nash equilibrium) is inefficient: trade shrinks worldwide and the relative price is unchanged. In this case there are large gains from tariff coordination: actually, under coordination the two countries refrain from imposing tariffs and achieve free trade.

16 Is the same mechanism at work in this paper? The two economies are large, each has monopoly power. Taken the behavior of the other country as given, each country has an incentive to tilt the terms of trade in its favor and pass the burden of labor effort onto the other country. But when the elasticity of substitution is sufficiently high, this comes at the price of the Sutherland effect. In a coordinated regime the two countries avoid manipulating the terms of trade, and consumers gain from exchange rate volatility to the extent that the Sutherland effect is at work. The gains are particularly large when consumption fluctuations are highly synchronized across countries (risk-sharing case)

17 Should we care? From an empirical vantage point, how bad is the assumption of unit elasticity? It is occasionally argued that the value of the elasticity of substitution between home and foreign goods (θ in the paper) is quite high, almost as high as the elasticity of substitution among differentiated brands (φ in the paper). In fact, there is a link between φ and the steadystate price markup on marginal costs: φ φ 1 = 1 + MARKUP Martins, Scarpetta and Pilat [1996] estimate the average markup for manufacturing sectors at around 1.2 in most OECD countries over the period

18 Other estimates (Morrison [1990], Domowitz, Hubbard and Petersen [1988]) range between 1.2 and 1.7. Benigno and Thoenissen [2001] estimate φ at 6.9 (for non-tradables) and 6.5 (for tradables) in the UK, and 7.9 in the euro area. The GEM (new simulation model of the IMF) sets φ/ (φ 1) = 1.2 or φ = 6.

19 But φ is not θ! The elasticity of input substitution in final good production θ can be estimated as the price elasticity of import demand. The benchmark estimates for θ are in the range [1,2] [Chari, Kehoe and McGrattan (1998)]. Many studies set the elasticity at 1.5 [Gali and Monacelli (2002), Smets and Wouters (2001)]. More recent estimates suggest that the price elasticities of imports and exports have declined over time. According to Hooper and Marquez [1995] the median value of the estimates for Japan, Germany and UK is 0.6.

20 Bottom line: A very plausible range for the elasticity of substitution θ is [0.5, 1.5]. So the workhorse model assumption θ = 1 falls right in the middle. Empirically, the assumption of unit elasticity seems warranted. But if this is the case, all the large gains from cooperation discussed in the paper disappear... A result corroborated by recent empirical research [Pappa (2002), Kollmann (2002)] stressing that joint maximization of world welfare delivers very limited gains compared to the Nash outcome.

21 Conclusion A key contribution to the new literature on coordination, especially from a methodological vantage point. Theoretically, the explanation for gains from cooperation relies on the positive contribution of exchange rate volatility to social welfare. Tough to sell... Empirically, the gains from coordination may still be small even accounting for a more flexible model. Perhaps central bankers are doing a good job after all...