Macroeconomic Policy Coordination in a Competitive Real Exchange Rate Strategy for Development

¹I want to thank the comments of two anonymous referees, Robert Blecker, Arslan Razmi, Roberto Frenkel and Daniel Aromi. I am specially grateful to Peter Skott for his encouragement and insightful suggestions. The usual disclaimers apply.

²See, among others, Hausmann et al. (2005), Prasad et al. (2007), Gala (2008), Rodrik (2008), Rapetti et al. (2012) and Bereau et al. (2012).

³I follow the definition of nominal exchange rate as the domestic price of a foreign currency. Consequently, a higher RER implies a more competitive, depreciated or undervalued domestic currency in real terms.

⁴This idea resembles the view of classical development economics that conceived manufactures as the engine of economic development. Agriculture and services were seen as the backward sectors. These days, there is a recognition that modern activities are comprised not only by manufactures but also by some highly productive tradable services (Rodrik 2010).

⁵Another common objection regards the global implications of this strategy, which are not discussed here.

⁶Chile’s experience fulfills the of-cited definition of growth acceleration episode developed by Hausmann et al. (2005). Argentina’s also fits this definition for the period 2003–2011.

⁷For an analysis of these experiences, see Frenkel and Rapetti (2012).

⁸An increase in tradable profitability can be achieved or strengthened with sectoral policies such as subsidized credits, tax exceptions, direct subsidies, infrastructure building and other conventional industrial and trade policies. Most successful development experiences used these instruments extensively (Rodrik 1995). Since the 1980s, however, they gradually lost mainstream appeal as the critics pointed to information problems, rent-seeking and government failures. Their use has also been threatened by WTO regulations. Without implying any judgment, in this paper I ignore the role of these policies and focus exclusively on the RER.

⁹A devaluation can certainly have contractionary effects in the short-run (Frankel 2005). In the case they exist, once digested, the expansive effects described above are assumed to prevail.

¹⁰Investment is mostly dominated by expected profitability (although actual profitability may also play a relevant role when firms face credit constraints). Consequently, expectations about the future are crucial for investment decisions. The view I adopt about the way agents form expectations is rather “behavioralist.” Future is by its own nature unknown. Besides, agents face costs at gathering information and have limited cognitive abilities to process the information at their dispose. In such contexts, past and present conditions could be reasonable guides to form expectations about the future. Thus, if the RER competitiveness incentive has been stable and lasted long enough, it would likely induce tradable firms to invest.

¹¹Although a key input, labor is not the only one influencing supply conditions of non-tradables. Infrastructure, energy and land may be other relevant factors. For simplicity, I will ignore the possibility that these factors may constrain the supply of non-tradables.

¹²Given the assumptions that non-tradable productivity is constant and tradable productivity is catching up due to the accumulation of capital, technology and knowledge, this tendency towards RER appreciation resembles the Balassa-Samuelson effect.

¹³For an analysis of the conditions under which sterilized FX interventions are sustainable see Frenkel (2007). See Ocampo (2003) on the effectiveness of capital controls in developing countries.

¹⁴For simplicity I use the product wage as a proxy of unit costs, which would also include non-tradable goods. Since the determination of nominal wages are arguably influenced by the evolution of non-tradable prices, as modeled in the next section, the product wage captures the set of relative prices that are relevant to determine tradable sector’s profitability.

¹⁵The product wage is calculated as the nominal wage rate divided the exchange rate; i.e., domestic wages measured in US dollars. For tradable labor productivity, I use real GDP per capita as a proxy.

¹⁶This, in turn, is achieved by setting θ=1, θ<1 and θ>1, respectively.

¹⁷The domestic interest rate, i, could influence the behavior of private savings. However, since it is given by the international interest rate, s appears in the model as a parameter. As discussed in Section 2, the government could use sterilized interventions and capital management techniques to regain some control over monetary policy. In that case, it could manage i within a certain range and thus influence the evolution of the price of non-tradables through s. In other words, demand management policies would also include monetary policy. Although governments targeting competitive RER levels actually use a combination of policies to have some autonomy in conducting monetary policy, I will ignore this possibility to keep the algebra simple. However, s could be thought as a variable potentially influenced by monetary policy and other saving incentives offered by the government.

¹⁸The price of non-tradables also falls when private saving rate, s, rises; i.e., X_s<0. Thus, in cases in which governments have some autonomy to conduct monetary policy, a rise in the domestic interest rate would increase s and thus reduce the price of non-tradables (i.e., ↓ X).

¹⁹Two other alternatives are possible. In one case the nullclines do not intersect and in the other they are tangential to each other at a single point. None of these cases are of economic interest.

²⁰See, for instance, Galindo and Ros (2008) for the case of Mexico and Barbosa-Filho (2008) for the case of Brazil.