هراس از سیاست های پولی معمول ویا شناور ؟ تجزیه و تحلیل ساختاری از سیاست پولی مکزیک

This paper uses Bayesian methods to estimate a small open economy dynamic stochastic general equilibrium (DSGE) model for the period in Mexico after the 1994 crisis. I consider a Taylor rule as the expression of the evolution of monetary policy to gauge its response to the exchange rate in the post-crisis period. The estimation results favor a consistent response of the nominal interest rate to the short-run nominal exchange rate after 1994. Although fear of floating is present, Mexico's monetary policy has taken steps toward a credible free-floating exchange rate that targets inflation.

Fear of floating1 has been an endemic problem in emerging market countries for the past several decades. This issue developed after the contagion of financial and balance of payments crises in emerging markets (e.g., Mexico in 1994). Before the crises such countries had as a common characteristic pegged or managed exchange rate policies. As part of their recovery, several countries allowed the exchange rate to float.2 In Mexico's case, low international reserves, a large current account deficit, and lack of credibility drove policymakers to adopt a floating exchange rate regime. However, Calvo and Reinhart (2002) conclude that most emerging markets are not true floaters. Instead, they are characterized by fear of floating. This paper examines the Mexican economy experience after its 1994 financial and balance of payments crisis. I estimate a structural model that includes a Taylor rule as the expression of the evolution of monetary policy. The monetary policy rule evaluates the extent to which the interest rate responds to inflation or the exchange rate in the post-crisis period. This approach allows me to evaluate whether the Mexican central bank's post-1994 monetary policy targeted inflation or reflected fear of floating. Announced central bank policy was characterized by a significant regime change after the 1994 crisis. Policy shifted from a predetermined exchange rate to a floating exchange rate thereby making monetary policy the nominal anchor of the economy after the crisis. Aside from the desire to stabilize inflation through the exchange rate, several reasons explain why emerging markets—in this case, Mexico—have sought to establish a predetermined exchange rate. Emerging markets have adopted predetermined exchange rates in part because of the association of currency devaluations with recessions. In addition, a large portion of these countries’ debt is denominated in dollars, which make defaults and general debt servicing more pervasive if the exchange rate varies widely. If large depreciations occur, adjustments in the current account are more pronounced in emerging markets. Moreover, large swings in the exchange rate have a larger impact on trade mostly because trade is invoiced in U.S. dollars in these countries (Calvo & Reinhart, 2002). Calvo and Reinhart (2002) compare the behavior of the volatility of the interest rate, the exchange rate, and the international reserves from different countries, including Mexico, in which fear of floating is suspected with the behavior of such variables in true floater countries such as Japan and the United States. Ball and Reyes (2004) use the same analytical approach as Calvo and Reinhart (2002) in addition to a vector autoregressive (VAR) analysis to determine (i) whether Mexico exhibits fear of floating or targets inflation and (ii) if the two are empirically distinguishable.3Ball and Reyes (2004) find that Mexico indeed exhibits fear of floating by standard measures. Their results also suggest that a rise in the exchange rate today signals future higher inflation. Therefore, they find evidence of both inflation targeting and fear of floating with more weight on inflation targeting. The contribution of the present study is that it estimates a small open economy model under two different regimes: one that is known for having a managed exchange rate and the second that targets inflation but in which fear of floating might be present. To attain my goal I estimate a New Keynesian small open economy model using likelihood-based Bayesian methods. The small open economy model framework, derived from micro-foundations, offers a structural channel for the study of monetary policy. The New Keynesian DSGE model has been a widely used to study monetary policy. Large New Keynesian DSGE models have been estimated for the United States and the euro area, and small open economy models have been estimated for countries such as Australia, Canada, Chile, and New Zealand, among others. In this paper, I estimate a small open economy DSGE model for Mexico's emerging market economy. The estimated model is based on a simplification of the Galí and Monacelli (2005) small open economy model by Lubik and Schorfheide (2007). The estimation of a DSGE model for Mexico is performed for two subsamples: the period before and after the 1994 crisis. Within this framework my paper builds on the work of Kamin and Rogers (1996) and Martinez, Sanchez, and Werner (2000), who estimate a Taylor-type policy rule for the periods before and after the crisis, respectively. Furthermore, multiple equation models have been estimated in Santaella and Vela (1999) before the crisis, in Martinez et al. (2000) for the period after the crisis, and Del Negro and Obiols-Homs (2001) for the pre-and post-crisis that portray the characteristics of a small open economy. However, the goal of Kamin and Rogers (1996) and Santaella and Vela (1999), is to pin down whether the monetary authorities used an exchange rate stabilization program or a money-based stabilization program to contain inflation before the crisis, whereas my the purpose is to evaluate whether Mexico's monetary policy shows fear of floating after the crisis. The estimation procedure used by Martinez et al. (2000) and Del Negro and Obiols-Homs (2001) is a VAR designed with the objective to study the mechanisms of monetary policy transmission in Mexico. My estimation is closer to the model structure approximated by the VAR used by Martinez et al. (2000); however, I estimate a small open economy structural model. Del Negro and Obiols-Homs (2001) use a money supply equation as a monetary policy reaction function; my estimation uses Taylor-type rules. The empirical analysis shows that the central bank of Mexico (El Banco de Mexico) favors a consistent response to movements in the exchange rate during the floating exchange rate regime. Although the estimated positive response to exchange rate movements can be attributed to interventions by the central bank to relieve inflationary pressures, the response to the exchange rates in Mexico is more pronounced than in other small open economies such as New Zealand, the United Kingdom, and Canada. Moreover, when the exchange rate and inflation policy parameters are normalized using their respective standard deviations (SDs), a comparable response to both variables by the central bank is still observed. Thus, there is some evidence that monetary policy in Mexico is conducted in accordance with fear of floating. In addition, I find that responding to inflation has become important since the 1994 crisis. In the monetary policy rule analysis, the response of the interest rate to inflation in the floating exchange rate regime is more than one to one, which compares with the reaction behavior addressed by the Taylor principle. I find much less concern by the monetary policymakers regarding inflation control during the managed exchange rate regime. One finding of this paper is that policy behavior changed from the pre- to the post-crisis period; however, the estimation results also provide evidence of changes in the structural parameters in the model. The estimation of impulse response functions for the pre- crisis and post-crisis periods also suggests that since the crisis the Mexican central bank has improved its ability to insulate the economy against real shocks. This is a relevant feature of the dynamics experienced by a country with a flexible exchange rate regime. Therefore, Mexico's central bank does target inflation but it also exhibits some degree of fear of floating. 1.1. Brief history of Mexico's monetary policy (1976–2006) From September 1976 to February 1982, Mexico had a fixed exchange rate regime, with the peso pegged to the U.S. dollar at almost a constant rate. By February 1982, the Mexican annual inflation rate was growing at 31% and reserves were dangerously low. In addition, the country experienced a 28% overvaluation of the peso. All these factors led to a large devaluation in February 1982. On February 17, 1982, the peso was devalued by 85%, after which Mexico adopted a flexible exchange rate regime that lasted until August of the same year. The economy experienced inflationary pressures caused by (i) a large deficit financed by money creation and foreign borrowing and (ii) an emergency wage increase. Moreover, Mexico was unable to meet its foreign obligations due to a critical currency shortage. The shortage was caused by the outflow of foreign capital, the maturing of loan obligations acquired in 1981, and the fall in oil prices. These events unleashed the August 1982 debt crisis that resulted in the nationalization of the banking system and the adoption of tighter capital controls. In addition, monetary policymakers adopted a pegged rate of depreciation of the peso to the dollar that lasted until 1988. By the end of this regime, annual inflation was at an all-time high—150%—caused in part by constant wage revision. Mexico's currency was devalued in September and October 1988. The devaluations of September and October 1988 marked the beginning of a new monetary policy regime that would gradually liberalize the exchange rate. The purpose of this regime was to abandon the exchange rate as the nominal anchor of the economy. In December 1987 Mexico adopted an anti-inflationary program that ended the constant wage revisions, privatized the banking system, and fixed the exchange rate to the dollar. The fixed exchange rate regime from March to December 1988 changed into a “Tablita”-type regime that prevailed until October 1991. The Tablita regime allowed the peso's exchange rate to adjust at a slow fixed pace that moved from a daily devaluation of 1 peso per dollar to 0.40 peso per dollar. The exchange rate was further liberalized into a limited flexibility regime that used a widening exchange rate band to control the depreciation of the currency. From October 1992 to the balance of payments crisis of December 1994, the band width reached 15% of the peso exchange rate with the dollar. The predetermined exchange rate bands helped to anchor inflation expectations in the economy during this regime. On December 22, 1994, the beginning of the Tequila crisis, the monetary authorities adopted a floating exchange rate regime. After a flexible exchange rate was adopted, the economy experienced two regimes that intended to target inflation in the economy. The tool used to anchor the economy shifted from a controlled exchange rate regime to a regime that controls monetary policy. In recent years, monetary policy oriented to target inflation has used the interest rate as the policy instrument. Policy evolved from a borrowed reserves regime to one that used “discretionary actions of monetary policy,” as Martinez et al. (2000) concluded. Monetary policy authorities in Mexico have taken into consideration the output gap, expected inflation, and shocks originating in the foreign sector to formulate a discretionary monetary rule. The evolution of monetary policy led to a rule that uses explicit inflation targeting as a preemptive source of stability.

I estimate a small open economy model for Mexico that uses a Taylor-type rule as an expression of the evolution of monetary policy. In particular, I address the question of whether there is evidence that the period after the financial and balance of payments crisis shows fear of floating. I conclude that fear of floating is present after the crisis. Furthermore, through the estimation of a DSGE model I find that the exchange rate was the nominal anchor of the economy before the crisis, while the inflation rate has become the nominal anchor since the crisis. Additionally, the monetary policy transmission observed in the model implies that a contractionary monetary policy shock produces a decline in inflation, the output gap, and an appreciation of the currency. However, the shocks to monetary policy do not persist in the model for longer than five quarters. In contrast, shocks to the U.S. business cycle cause dynamics in the variables of the model that persist in the economy over 25 quarters. In addition, after estimating impulse responses to shocks to the terms of trade and the exchange rate, I can conclude that a floating exchange rate regime by the central bank of Mexico has helped to insulate the economy against real shocks. The responses in the model for the pre-crisis and post-crisis periods are consistent with the responses associated with a fixed exchange rate regime and a floating exchange rate regime, respectively. The estimated structural parameters show that the model can fit the high degree of trade openness observed in Mexico and could be extended to study the behavior of this characteristic in other emerging markets. In addition, when compared with estimates for other small open economies, the value for the sacrifice ratio between inflation and the output gap in Mexico is similar. Even though fear of floating is present in the data, I still observe a low degree of exchange rate response in the period after the 1994 crisis compared with the period before the crisis. The inclusion and study of pass-through in Mexico and its link to fear of floating is a topic that requires further research.