The exchange rate cycle in Argentina

ABSTRACT

This article discusses the secular volatility of output, inflation, the exchange rate, and poverty in Argentina. Inflation, currency devaluation and a wide gross domestic product (GDP) cycle have been recurrent problems in Argentina for several decades. The literature has extensively discussed those issues from different viewpoints. This study focuses on a relatively unexplored theme that may contribute to a partial explanation. It deals with the continuous tendency to equalize different profitabilities resulting, in turn, from remarkably different sectoral purchasing power parities. Thus, for any given exchange rate, an incessant tendency toward the equalization of profitabilities generates opposing inflationary and devaluatory pressures. The resulting inflation-devaluation cycle feeds income redistribution, GDP fluctuation, real exchange rate instability, and high levels of uncertainty.

KEYWORDS:

• Economy-wide country studies
• macroeconomic analysis of economic development
• open economy macroeconomics

JEL CLASSIFICATIONS:

• F41
• O11
• O54

Notes

¹Data available at www.economia.gob.ar and www.indec.gov.ar. From 2008 onward, official data on the economy began to diverge from private data collection. Inconsistencies within the official data would suggest that there are some grounds for concern. For example, Figure 1 shows the real exchange rate calculated with official CPI data, and the real exchange rate calculated with official GDP deflator data. These two series increasingly diverge from 2005 onward, whereas they were always pegged to each other before 2005. After 2011, the differences became too wide and a heated political debate broke out. To avoid unnecessary controversy we focus on the 1970–2011 period. Quarterly data for 1970–2011 are sufficient to illustrate our argument. Quoted econometric works are for 1970–2008. There is no indication, however, that events after 2011 respond to a different pattern from the one we are discussing for 1970–2011.

²In 1991–2011, the composition of exports in nominal U.S. dollars was as follows: agricultural products—56 percent; manufactures—29 percent; and fuels and minerals—15 percent. Abstracting from the rise in commodity prices in 2007–11 (i.e., in constant 1993 prices) the share of manufactures rose from 26 percent in 1991–95 to 39 percent in 2005–11. In all our calculations, we removed gold and copper (1.4 percent of total exports in 2007–11) from manufactures and added them to minerals. Argentina started exporting gold and copper in 2007.

³In his import demand function, Nicolini-Llosa (2011) detected a structural change in the first quarter of 1992 corresponding to the aggressive trade liberalization policy. Thus, income elasticity rose from 1.72 in 1970–91 to 2.95 in 1992–2008. Nicolini-Llosa (2011) regressed quarterly nonstationary unit-root time series into Engle and Granger’s ordinary least squares (OLS) long-run equation. The corresponding augmented Dickey–Fuller test of residuals of the OLS equation rejected the null hypothesis of no cointegration. The short-term values converged to long-term values in the error correction model. Specification errors, serial correlation, and heteroskedasticity were not detected. Income elasticity was 2.42 in 1980–97 (Catao and Falcetti, 2002) and 2.76 in 1993–2008 (Berretoni and Castresana, 2009).

⁴This results from multiplying (a) the 0.142 ratio of imported manufactured consumer goods to total imports by (b) the 0.239 ratio of total imports to total private consumption, that is, 0.142 × 0.239 = 0.034. The trade liberalization period 1993–2001 yields 0.173 × 0.161 = 0.028. This is a higher than average ratio of manufactured consumer goods imports to total imports of 0.173, but a lower than average total imports to total private consumption ratio of 0.161. The latter is because GPD grew 2 percent p.a. in 1991–2001, but 5.5 percent p.a. in 2002–2012, given the above-mentioned high income elasticity of demand for imports.

⁵An additional necessary condition for a single PPP to hold is that the quantity theory of money would apply. Evidence, however, is that causality does not go from money to prices and then to the exchange rate, but the other way around (Harvey, 2006). The PPP presents additional problems besides the necessity of the quantity theory (Flassbeck and La Marca, 2009; La Marca, 2004). However, this goes beyond the scope of this study.

⁶Obstfeld and Rogoff (1996, ch. 4) show that in a tradable/nontradable economy with capital mobility and equal profit rates, the real exchange rate need not converge to the PPP. In Argentina, PPP does not hold for tradables, even if nontradables are dismissed—Equation (2).

⁷The ratio of agricultural average export prices in 1991–2001 to their average price in 1986–2012 was 0.90. This same ratio for manufacturing exports was 0.98, indicating that in 1991–2001 agricultural export prices were below their long-term average and below the manufacturing export prices in 1991–2001.

⁸π = (α₂/α₁)^1/(2*3) − 1 = 2^1/6 − 1 = 0.122 p.a.

⁹GDP_t/GDP_t–4 − 1 at constant 1993 AR$ prices.

¹⁰De Gregorio et al. (1998) test a sharp rise in the consumption of manufactures at the end of large devaluations in Argentina and other developing countries with high exchange rate instability.

Additional information

Notes on contributors

Jose Luis Nicolini-Llosa

Jose Luis Nicolini-Llosa is a professor on the Faculty of Economics, Buenos Aires University and CONICET.