d-cycloserine The rest of the paper is organized as follow S

The rest of the paper is organized as follow. Section 2 presents some stylized facts about the relationship between import penetration and labor productivity and real exchange rate. Section 3 presents the micro-funded model that supports the theoretical view about the importance of productivity over exchange rate. Sections 4 and 5 present, respectively, the empirical model and results. Finally Section 6 concludes this article.
Stylized facts: exchange rate, productivity and import penetration A preliminary and basic test of this relationship shows that the correlation is negative, as expected, but very small. Fig. 1(a) presents the relationship between the log difference of the share of imports in the domestic market (measured by import penetration coefficient calculated at constant prices) and the log difference of sectorial exchange rate for 17 manufacturing sectors, during the d-cycloserine between 1997 and 2011. Note that the linear relationship is weak and the correlation coefficient is only −0.06. This characteristic is mainly a consequence of the low correlation in the group of intermediate and capital goods sectors. Fig. 1(c) shows that in these sectors the relationship between the two variables is very close to zero, whereas in the sector of consumption good industries, as shown in Fig. 1(b), the relationship is stronger and the correlation coefficient is −0.21. According to Lisboa and Pessoa (2013), the effect of the exchange rate is ambiguous. Devaluations may protect the national industry, but, at the same time, raise the costs of imported inputs. Because of this ambiguity, Broz and Frieden (2006) pointed out that there is no clear economic policy that could determine one appropriate level of exchange rate. Thus, it is possible that the use of manufactured imported goods as input by the national industry decreases the sensitivity of import penetration to the exchange rate. In fact, intermediate and capital goods sectors are, in general, more dependent on imported inputs than the consumption goods sectors. According to the databases of input–output matrix estimated by different authors for the period between 2000 and 2009 (Freitas et al., 2012; Guilhoto and Sesso Filho, 2010; Martinez, 2015), the sectors present in these groups has, on average, a participation of around 16% of imported inputs on total inputs demanded in the final production against a percentage of 8% in the consumption goods sectors. Several economists, on the other hand, have attributed the recent raise of import penetration to the performance of labor productivity in recent years (Bonelli and Pessoa, 2010; Ferreira and Fragelli, 2011). According to Lisboa and Pessoa (2013), while certain sectors in the decade of 2000 were favored by labor productivity gains, e.g. agriculture, most of the manufacturing industry sectors showed a poor performance. Using data from IBGE manufacturing survey (PIM-IBGE), between the years of 2003 and 2011, a period of intense appreciation of the real exchange rate, the industry labor productivity increased only 21%. In particular, labor productivity performance in intermediate and capital goods sector presented a cumulative growth of 18%, while the sectorial exchange rate appreciated 57%. Whether these segments of economists are correct, we expect that the relationship between import penetration and productivity would be negative and high correlated. Therefore Fig. 2(a) presents the dispersion between the log difference of import penetration and the log difference of labor productivity for all sectors highlighted here. In fact, a preliminary analysis of the data in Fig. 2 reveals that the relationship between these variables is strongly negative. The correlation coefficient is −0.57 for the group of 17 sectors. The same occurs with the two groups of sectors as seen by Fig. 2(b) and (c). In the case of intermediate and capital goods sector the correlation coefficient is −0.65 and for consumption good sector its value is −0.40.