Inflation, growth, and distribution: The Brazilian economy after the post war

Abstract

Following Marglin and Bhaduri (1990), the purpose of this paper is to investigate empirically the interaction between income distribution and growth of aggregate demand during the 1951–89 period in Brazil. Applying Hein and Vogel’s (2008) methodology we conclude that the Brazilian economy showed a profit-led demand regime. In a context of high inflation, high concentration of markets, and wage control, retained profits were the main source to finance new capital. In this sense, we found a large sensitivity of investment relative to the wage share, a result that is compatible with a consumption pattern based on high income, which supported the growth trend with low wages observed during the period.

Key words:

• Brazilian economy
• demand-led regime
• industrialization via import substitution

Notes

¹For a reference of the high inflation regime in Brazil in the 1980s and 1990s, see Feijó and Carvalho (1992).

²We limit our period of analysis until 1989, as in March 1990, the new president elected after the military coup in 1964, Fernando Collor de Melo, gave a new orientation to economic policy, following neoliberal recommendations of the Washington Consensus. This implied deep changes in the institutional framework with impacts on income distribution and capital accumulation.

³We should note that in Marxists models there is no role for aggregate demand in determining long-term growth, which is explained exclusively by supply forces.

⁴Assuming that variations in the income distribution are not offset by variations in the markup.

⁵Changes in labor cost levels affect the relative prices of domestic goods, which affects import levels. However, the Brazilian economy in 1951–89 was extremely closed and import control mechanisms were widely used given the import substitution strategy. Hence, following Naastepad (2006), we consider that imports are not directly affected by changes in income distribution.

⁶We use data from Penn World Table 6.3 (PWT 6.3), because in more recent versions of this database government expenditures with education and health are included in private consumption.

⁷The main public enterprises created in the 1940s and 1950s were: Companhia Siderúrgica Nacional, in 1941, in the metallurgic branch; Companhia Vale do Rio Doce, in 1942, in the mining branch; and Petrobras, in 1953, in the exploration and processing of oil.

⁸Trebat (1983) lists factors such as operational autonomy, high dependency on financing with retained profits (and relative independency for public funding), and management aligned to “market rules,” which allow him to conclude that faced with distribution shocks, Brazilian state-owned enterprises made investment decisions in a way similar to those of private companies.

⁹Brazilian Statistical Office (IBGE), Estatísticas do Século XX, http://seculoxx.ibge.gov.br/economicas/contas-nacionais/.

¹⁰Using the series in logarithmic form implies that we are assuming multiplier effects of the wage share on the components of aggregate demand. The estimation using variables in level, which assumes additive effects, is not suited to the characteristics of the series used in this analysis, since growth is exponential over time.

¹¹Notes: (1) one lead and one lag were added to the estimated equation (additional leads and lags did not show significance); (2) the covariance matrix was estimated using quadratic spectral kernel and Andrews automatic bandwidth, and the residues were prewhitened with an AR (1); (3) the Ljung Box test up to the sixteenth lag did not reject the null hypothesis of no serial autocorrelation at 5 percent significance; (4) R² = 0.9996 and DW = 2.53.

¹²Notes: (1) one lead and one lag were added in both equations; (2) the covariance matrix was estimated using quadratic spectral kernel and Andrews automatic bandwidth, and the residues were prewhitened with an AR (1); (3) the Ljung Box test up to the sixteenth lag did not reject the null hypothesis of no serial autocorrelation at 5 percent significance; (4) Eq IPPE: R² = 0.992 and DW = 1.57 (5) Eq IP: R² = 0.982 and DW = 1.54.

¹³Notes:(1) two leads and two lags were added; (2) the covariance matrix is estimated using quadratic spectral kernel and Andrews automatic bandwidth, and the residues were prewhitened with an AR (1); (3) the Ljung Box test up to the sixteenth lag did not reject the null hypothesis of no serial autocorrelation at 5 percent significance; (4) R² = 0.987 and DW = 1.60.

¹⁴This growth experience was observed in most Latin American countries that followed the import substitution strategy toward industrialization (Bertola and Ocampo, 2012).

¹⁵“Investment … is likely to respond more cautiously to a change in profit margin/share compared to consumption. Thus, the depressing effect of a lower real wage rate on consumption may be felt within the short period without its stimulating effect on investment materializing within the same period … . To the extent that exports and imports have faster speeds of adjustment (to price changes) compared to investment” (Bhaduri and Marglin [1990], p. 385).

¹⁶After the war, three development plans were launched: the Targets Plan (Plano de Metas) in the 1950s and two National Development Plans (PND I and II), in the 1970s. For a review of Brazilian development, see Baer (2008).

¹⁷For a discussion of the development of the financial system in postwar Brazil up to the 1980s, see, for instance, Tavares and Serra (1971), Bresser-Pereira (1987), among others.

¹⁸See also Moreira and Puga (2001, especially Figure 5), for a discussion of how industrial firms financed their investment after price stabilization in Brazil in the second half of the 1990s.

¹⁹It should be mentioned that in the 1960s important reforms were introduced in the financial market aimed at developing a long-term credit market.

²⁰See Kapeller and Schütz (2015) for a recent debate on the profit-led and wage-led demand regimes integrating the Veblenian concept of conspicuous consumption.

²¹Underemployment and informal work is a common feature of Latin American economies with high structural heterogeneity (Cimoli et al. 2006).

²²It is worth mentioning that Bhaduri and Marglin (1990) point out that the expansion of the wage bill and increasing employment are important factors to socially sustain a profit-led regime, and to reduce the possibility of an overaccumulation crisis in long run.

Additional information

Notes on contributors

Carmem Aparecida Feijó

Carmem Aparecida Feijó is a full professor at Fluminense Federal University and a CNPq Researcher.

Felipe Figueiredo Câmara

Felipe Figueiredo Câmara is an economist at the Brazilian Statistical Office.

Luiz Fernando Cerqueira

Luiz Fernando Cerqueira is an associate professor at Fluminense Federal University.