Does the currency board matter? US news and Argentine financial market reaction

Abstract

Using a Generalized Autoregressive Conditional Heteroscedasticity (GARCH) model, we study the effects of the US monetary policy and macroeconomic announcements on Argentine money, stock and foreign exchange markets over the period January 1998 to July 2007. We show, first, that both types of news have a significant impact on all markets. Second, there are noticeable differences in reaction for different subsamples: Argentine money markets were more dependent on US news under the currency board than after it was abandoned as the floating exchange rate partly absorbs spillover effects from the US. Finally, we find that the US-dollar-denominated assets react less to US news than peso-denominated assets, which suggests that the currency board was not completely credible during its final years.

Keywords:

• Argentina
• central bank communication
• financial markets
• federal reserve
• macroeconomic announcements

JEL Classification::

• E52
• F33
• G14
• G15

Acknowledgements

We thank an anonymous referee, Roland Füss, participants of the Money, Macro and Finance Group Annual Conference in Limassol, 8th Workshop on Money, Banking, and Financial Markets in Halle, and MAGKS research seminar in Rauischholzhausen for their helpful comments on earlier versions of the article. Of course, all remaining errors are our own.

Notes

¹ The trade share of Argentina with the US is between 10% and 16% and declines during the sample period.

² Wolf et al. (2008) provide a comprehensive overview of the technical functioning of currency boards and the macroeconomic challenges faced by countries that peg their exchange rate.

³ Further information on the sample period is provided in Section III.

⁴ The Argentine currency board and its breakdown are discussed, for instance, in Bleaney (2004) and Gurtner (2004).

⁵ In designing these categories of news, we carefully read each speech twice, with a considerable time lag between each reading, and then coded them independently into the respective dummy categories. In the case of a conflict between the two codings, we checked the relevant speeches yet another time and adjusted our indicators accordingly. We employed extensive robustness checks to ensure that our results do not depend on the particular coding of ambiguous individual observations. Data are obtained from the official website of the Board of the Governors of the Federal Reserve System. An alternative to subjective coding is using content analysis software (for instance, Lucca and Trebbi, 2009). However, communications other than post-meeting statements are not standardized and, thus, content analysis programmes fail to detect systematic patterns in these more complex texts.

⁶ For example, this type of asymmetry can be found in the impact of International Monetary Fund (IMF) statements on financial returns in emerging markets (Hayo and Kutan, 2005).

⁷ We explore several specifications for US macroeconomic news. In a first step, we test whether the actual values, the standardized shocks, or both have an impact on our financial market indicators. The shocks are significant, whereas the actual values remain insignificant. The same outcome occurs when including actual values and news dummies instead of shocks. In a second step, we discover that the results using news dummies weakly dominate the ones using standardized shocks in terms of significance and, therefore, we employ news dummies in the analysis presented below.

⁸ Data sources: US bond market and foreign exchange market series – Federal Reserve's statistical releases H10 and H15; stock market series – Yahoo! Finance database; Argentine interbank lending rates – Central Bank of Argentina statistical database; Surveys of macroeconomic announcements – Bloomberg newswire.

⁹ Exchange rates are defined in price notation, which implies that an increase in the exchange rate indicates a depreciation of the peso against the dollar. The exchange rate is used as an endogenous variable only in the second subsample, as it was pegged until February 2002.

¹⁰ In July 2001, the Argentine economic and financial crisis hit the money market as returns increased sharply. Markets displayed the lack of confidence in the government's plan to balance the budget and subsequently the country rating worsened.

¹¹ In December 2002, the financial crisis effects on money markets decreased substantially after restrictions on peso-denominated bank withdrawals were relaxed and the government showed its ‘good faith’ by symbolically repaying some debt to several international organizations.

¹² Bloomberg surveys are used to identify surprises occurring at scheduled meetings. Intermeeting moves are naturally classified as surprises. For instance, a ‘surprise hike’ can be (i) an unexpected rise in the target rate or (ii) an unchanged target rate when a rate cut was expected.

¹³ The coefficients can be interpreted as follows. All target rate variables are coded as multiples of 25 bps. For instance, the coefficient 0.198 (−0.801) implies an increase of 19.8 bps (decrease of 80.1 bps) in the 3-month peso rates (the Merval index) after a surprise hike in the US target rate. In the case of macroeconomic surprises and central bank communication, we rely on news dummies. For instance, the coefficient 0.055 (−2.262) denotes an increase of 5.5 bps (decrease of 2.26 pp) in the 3-month peso rates (the Merval index) after announcement of a higher than expected US CPI.

¹⁴ The influence is statistically larger than for target rate hike surprises (Chi²(1) = 4.2*).

¹⁵ All unexpected interest rate hikes took place at scheduled meetings.

¹⁶ The influence of positive and negative CPI news is statistically equal (Chi²(1) = 0.43).

¹⁷ To test for statistical differences between the currency board era and the floating exchange rate subsample, we estimate an unrestricted version of Equation 1 in both periods for the peso-denominated money market and the equity market, respectively. Tests of differences in means across the two samples are mostly insignificant, except for two cases. The influence of MPR/Testimony EO– on the money market is larger during the second subsample (z = 2.47*) and CPI + shocks have a larger impact on the equity market during the currency board era (z = −3.07**).

¹⁸ Only two significant coefficients from Tables 1 and 2 are found to be marginally insignificant when controlling for Argentine macroeconomic news: (i) surprise interest rate cuts in the peso-denominated money market during the currency board era and (ii) positive ISM news in the equity market during the second subsample. To conserve space, we do not report these regressions. All omitted results are available upon request.