Investment cycle of the Brazilian economy: a panel cointegration analysis of industrial firms based on Minsky’s financial instability hypothesis—2007–2017

Abstract

The aim of this paper is to argue that one of the impacts of the 2008 international financial crisis and the European crisis in the Brazilian economy was in the investment rate. Our interpretation is that the impact of the international crises made the balance sheet of non-financial firms more fragile, mainly the industrial firms, and the economic measures implemented to sustain profit margins from 2011 onwards were not enough to counterbalance the deceleration in the aggregate demand and the rise in financial costs. To investigate the evolution of the financial fragility of industrial firms , following Minsky’s financial fragility hypothesis , we propose an econometric model using a cointegration panel. We develop the analysis in two stages using instrumental variables to estimate the rate of investment and the share of financial receipts in total receipts. Among the instrumental variables, the share of wages in value-added, the domestic interest rate, the exchange rate, and the world income were statistically significant. We conclude that the deterioration of the capacity of industrial firms to generate enough internal funds to face debt commitments can be explained by the deceleration of the aggregate demand, expressed by investment rate and the deceleration of the world income, and the increase in the weight of the wage share in total value-added.

Keywords:

• Brazilian economy
• investment cycle
• margins of safety

JEL CLASSIFICATIONS:

• E12
• E22
• O16

Notes

1 In brief, Minsky (1992, 7) says: “The financial instability hypothesis, therefore, is a theory of the impact of debt on system behavior and also incorporates the manner in which debt is validated.”

2 The role of banks and other financial institutions is discussed, for instance, in Carvalho (2016).

3 On the asymmetrical hierarchy of currencies, see de Paula, Fritz, and Prates (2017).

4 Kregel (2004, 580) when analysing the position of developing economies with net international lending concludes that “by definition and by design, [developing economies will be] operating a “Ponzi” financial profile since the interest and amortization on the borrowing can only be met by additional foreign capital inflows.” This implies that for a country to be successful in engaging in foreign indebtedness to finance its growth in a financially integrated world means that “its success depends on the willingness of foreign lenders to continue to lend.”

5 See, for instance, Karwowski and Stockhammer (2017).

6 The concept of ‘balance of payments dominance’, as proposed by Ocampo (2013), offers an explanation for the external fragility of these economies, highlighting the importance of capital account in causing cyclical shocks. Financial markets, under the assumption of non-probabilistic uncertainty, adopt practices and procedures leading to risk evaluation, which tend to stress the financial cycle derived from the capital account of financially integrated developing economies. In this context, macroeconomic policies in developing countries tend to be procyclical. There are three main consequences to the macroeconomic policy management stemming from the dependence of financially integrated developing economies.One is that nominal and real interest rates tend to be relatively higher; the second is that real exchange rates tend to be appreciated over time. The third is that since both trends can lead these economies to reduce their long-term capacity to grow, fiscal policy is procyclical, limiting the space for growth policies. For a discussion on the constraints to grow with foreign savings in the Brazilian economy, see, for instance, Bresser-Pereira and Nakano (2003); Feijo, Lamônica, and Bastos (2016) and Feijo and Lamônica (2019).

7 Rey (2015, abstract), in influential research, concludes that “There is a global financial cycle in capital flows, asset prices and in credit growth…. Assets markets in countries with more credit inflows are more sensitive to the global cycle.” The author shows that no matter the exchange rate regime, economies financially integrated face a ‘dilemma’ of political economy and not a ‘trilemma’ as in the Mundell-Fleming model. In the absence of capital controls, monetary policy loses autonomy, mainly in economies dependent on foreign savings.

8 de Paula, Modenesi, and Pires (2015) refer to the Brazilian growth in 2004–2008 as a miniboom.

9 Domestic demand was fueled by several stimuli, such as: the improvement in income distribution (through the increase of the real minimum wage and income policies to fight inequality); an increase in employment rates and an increase in the supply of consumer credit.

10 For a list of the measures, see de Paula, Modenesi, and Pires (2015).

11 The liberal economic reforms that have been implemented since 2016, as a way out of the deep recession, have not yet delivered the resumption of growth, as they have not reversed expectations in favor of a sustained recovery. It should also be remarked that since 2016, the fight against corruption in the country, through the operation named ‘lava-jato’ has also contributed to the increased uncertainty in the economy.

12 Tax reductions were mainly designed to reduce corporate costs such as cuts in payroll.

13 Ferrari-Filho, Cunha, and Bichara (2014, 536) point out that here was an inconsistency between the Keynesian approach of New Macroeconomic Matrix and the macroeconomia tripod based on the New Macroeconomic Consensus.

14 The authors also quote the CEMEC (2016) study that shows that among the joint-stock firms, excluding Petrobras, the share of debt in foreign currency increased from 24.2% in 2010 to 46.3% in the second quarter of 2016. Including Petrobras, the situation is more alarming, since foreign currency debt increased from 32.7% in 2010 to 60.0% in the second half of 2016.

15 As for the foreign indebtedness of Latin American firms, Pérez-Caldentey, Favreau-Negront, and Lobos (2018) show that the composition of the external debt in Latin America and Caribbean has changed significantly during the 2007–2015 period, although the level of the external debt in relation to GDP had been rather stable, around 18% of GDP of the region. According to the authors, the debt of non-financial firms raised from 14.8% in total debt in 2007 to 32.9% in 2017. Based on a sample of 2,241 non-financial firms in Latin America (1,281 Brazilian firms) with open capital, the authors also find that during the 2007–2015 the financial fragility of firms with debts in foreign currency had increased significantly.

16 The index of real exchange rate (average of the year) was 89.89 in 2013 and 111.58 in 2015 (serie 11752 from Brazilian Central Bank).

17 See Serrano and Summa (2015); Arestis, Baltar, and Prates (2015) and Corrêa, Lemos, and Feijo (2017).

18 For an alternative proxy for financial fragility after the 2008 international financial crisis, with a sample of large firms, see Almeida, Novais, and Rocha (2016).

19 The Annual Industrial Survey is produced by the Brazilian Statistical Office (IBGE). For information on the methodology and sampling procedures, see: https://www.ibge.gov.br/estatisticas-novoportal/economicas/industria/9042-pesquisa-industrial-anual.html?=&t=conceitos-e-metodos; Accessed March 10, 2020, January 29, 2019.

20 Industrial investment accounts for most of the investment in machinery and equipment in the Gross Capital Formation.

21 This indicator compares cash flow after of-pocket- costs and taxes with current receipts. It is built as the ratio between total receipts minus total costs and expenditures, excluding taxes and duties, depreciation and non-operational expenditures, divided by the sum of taxes and duties, depreciation end non-operational expenditures. Data are from the Annual Industrial Survey, IBGE.

22 The authors, based on an econometric study about the degree of indebtedness of non-financial firms (op.cit, 14), conclude that: “One insight of our paper is that a wage-led demand regime would hardly imply an increase in capital accumulation in a scenario of persistent real exchange rate overvaluation. On the opposite, exchange rate misalignment increased wage costs, reducing industrial competitiveness, negatively affecting the growth rate of the economy. Also, it contributed to substitute domestic for imported goods. Therefore, we can say the Brazilian economy is in a ‘weak' wage-led demand regime, in the sense that it has not been able to sustain high growth rates, mostly after 2010.”

23 The tests performed to evaluate the statistical relevance of instrumental variables are of three types: subidentification, overidentification and extraneous exogeneity. Sub-identification tests are performed on first- and second-stage equations with the null hypothesis that endogenous regressors are not relevant or are poorly identified. The over-identification test is performed only in the second stage of the equation, and has as null hypothesis the restriction of excessive identification. The strict exogeneity hypothesis is a test of the second stage equation, this test is determinant in the choice of an endogenous modeling.

24 The tests of the first stage equation show the statistical relevance of the instruments, as can be seen from agreements with the tests listed in the statistical summary of the first stage equations. The Sanderson–Windmeijer (SW) statistics reject the hypothesis of underidentification and weak identification of individual endogenous regressors, with a 1% probability. Likewise, the Anderson–Rubin and Stock–Wright tests, robust to the presence of weak instruments, reject the joint hypothesis of endogenous regressors equal to zero, also at 1% significance. The second stage tests reinforce the results of the first stage on the relevance of the instruments. The endogeneity test rejects the hypothesis of extreme exogeneity of the endogenous regressors at 1% siginificance. The Kleibergen–Paap underidentification tests in the second stage, reject the hypothesis that the excluded instruments are not relevant to 1%, accepting that the model is identified, and the Hansen overidentification test accepts the hypothesis that the instruments identify the model.Therefore, there are no restrictions on the instruments used.

Additional information

Notes on contributors

Carmem Feijo

Carmem Feijo is at Economics, Fluminense Federal University, Niteroi, Rio de Janeiro, Brazil

Marcos Tostes Lamônica

Marcos Tostes Lamônica is at Economics, Fluminense Federal University, Campos, Rio de Janeiro, Brazil

Sergiany da Silva Lima

Sergiany da Silva Lima is at Economics, Rural Federal University of Pernambuco, Serra Talhada, Pernambuco, Brazil