Financialisation as structural change: measuring the financial content of things

Abstract

In this article, we present a multi-sectoral treatment of financialisation based on input–output analysis. Our main innovation introduces financialisation as an increase in financial content per unit of output produced. In this way, we may investigate changes in relative importance of financial activities, taking into account direct and indirect interactions among sectors. Although methods focusing on the disaggregation of input–output tables have been largely explored in past decades, they have received limited attention in the literature on financialisation. We aim to refocus on multi-sectoral issues by offering a simple structure of analysis to assess the interconnections between the real and financial sides of the economy. Using a 15 and 14-sector level of aggregation, we study the experiences of the United States and Brazil for the period 1947–2015 and 1995–2011, respectively.

Keywords:

• Financialisation
• FIRE
• input–output
• United States
• Brazil

JEL CLASSIFICATIONS:

• G20
• O11
• O14
• O33

Acknowledgments

An earlier version of this paper was presented at the Analytical Political Economy Workshop, University of Massachusetts, United States; the Pontignano Annual Meeting, University of Siena, Italy; and the 22nd FMM Conference, Berlin, Germany. We are grateful to the participants and in particular to Peter Skott, Gerald Epstein, Mauro Caminati, Jose Oreiro, and Daniele Girardi for helpful comments and suggestions. In addition, we thank the editors and two anonymous referees for their suggestions for improvement. The usual caveats apply.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 A well-known definition by Epstein (2005, p. 3) makes reference to the ‘increasing role of financial motives, financial markets, financial actors and financial institutions in the operations of the domestic and international economies’.

2 Early discussions on the relation between finance and the productive structure go back to classical economists such as Adam Smith and Karl Marx. Key subsequent contributions include Hilferding's (1910) view of an emerging fusion of financial and industrial motives, further extended by Hymer (1960) to an international set-up. More recently, several studies have addressed the correspondence of the financial sphere with the macroeconomic environment (Keen, 1995; Skott and Ryoo, 2008; Sordi and Vercelli, 2014), investment dynamics (Stockhammer, 2004; Orhangazi, 2008; Davis, 2018), income distribution (Lin and Tomaskovic-Devey, 2013; Jaumotte et al., 2013; Dünhaupt, 2017), and innovation (Mazzucato, 2003; Mazzucato and Tancioni, 2012).

3 A comprehensive review of the fundamental structure of the IO model and its main applications can be found in Miller and Blair (2009) and Miyazawa (1976). IO models were expanded to a SAM framework in order to incorporate households, financial and nonfinancial corporations, government, and the rest of the world. Chen et al. (2005) provide a discussion about the inclusion of financial assets in the original IO model. A review of countries for which SAMs including financial institutions is available in Aray et al. (2017).

4 Different indicators have been proposed to estimate backward and forward linkages. A comprehensive discussion is provided by Miller and Blair (2009, pp. 555–565). For a recent application see Borghi (2017).

5 Notice that $\mathbf{D}$ does not depend on relative prices. Denoting with $\overline{\mathbf{x}}$, $\overline{\mathbf{A}}$, and $\overline{\mathbf{f}}$ values expressed in physical quantities while $\hat{\mathbf{p}}$ is a diagonal matrix of prices, we can rewrite it as: $\begin{array}{rl}\mathbf{D}& ={\hat{\mathbf{x}}}^{-1}(\mathbf{I}-\mathbf{A}{)}^{-1}{\hat{\mathbf{f}}}^{-1}\\ & =\left({\widehat{\overline{\mathbf{x}}}}^{-1}{\hat{\mathbf{p}}}^{-1}\right)(\mathbf{I}-\hat{\mathbf{p}}\overline{\mathbf{A}}{\hat{\mathbf{p}}}^{-1}{)}^{-1}\left(\hat{\mathbf{p}}\widehat{\overline{\mathbf{f}}}\right)\\ & =\left({\widehat{\overline{\mathbf{x}}}}^{-1}{\hat{\mathbf{p}}}^{-1}\right)\hat{\mathbf{p}}(\mathbf{I}-\overline{\mathbf{A}}{)}^{-1}{\hat{\mathbf{p}}}^{-1}\left(\hat{\mathbf{p}}\widehat{\overline{\mathbf{f}}}\right)\\ & ={\widehat{\overline{\mathbf{x}}}}^{-1}{(\mathbf{I}-\overline{\mathbf{A}})}^{-1}\widehat{\overline{\mathbf{f}}}.\end{array}$ Therefore, $\mathbf{D}$ is invariant to relative prices.

6 The alternative would be to focus on the effects of inter-sectoral linkages on employment and output shares (for references in the deindustrialisation-tertiarisation literature, see Peneder et al., 2003; Montresor and Marzetti, 2011; Ciriaci and Palma, 2016; Peneder and Streicher, 2018). However, they are themselves the result of changes in technical coefficients and final demand. Therefore, we take a different route. For a very recent assessment of the role of financial flows and contagion using IO analysis, see Román et al. (2018) and Schumacher (2018).

7 Our exercise has similarities with the so-called structural decomposition analysis (SDA) defined in the nineties as ‘the analysis of economic change by means of a set of comparative static changes in key parameters in an IO table’ (Rose and Chen, 1991, p. 3). In most of SDA formulations, changes in the Leontief matrix are described as some ‘technological change’, often interpreted to include any factor that causes a change in a technical coefficient (Rose and Casler, 1996). That is, true technological change, response to input price changes, scale effects, etc. For a recent investigation on the persistent statistical structure of the IO coefficient matrices, see Torres-Gonzáles and Yang (2019).

8 The alternative would be to use OECD tables for Brazil and US. However, this implies a massive loss of data since the OECD series are available only after 1995. We should mention that even though they do not share straightforward, comparable sectoring, cautious comparisons of trends could be useful. This is particularly true for matrix $\mathbf{D}$, which is independent of relative prices.

9 Minsky's theory rests on the bifurcation of an economy's price system. On the one hand, there is the price system for goods and services. On the other, there is a wholly separate price system for assets. It is here that stability leads to asset price inflation, followed by a build-up in debt and, eventually, a crisis (see, for example, Wray, 2015).

10 At the time, several banks were specialised in using the institutional mechanism of monetary correction and price indexation to make profits. Hyperinflation was used to cover illegal operations and balance sheet issues. Once prices were under control, some of them faced severe problems and were about to break. In 1995 the Brazilian government released the ‘Programa de Estímulo à Reestruturação e ao Fortalecimento do Sistema Financeiro Nacional’ (PROER) to avoid the collapse of the financial system resulting from the control of inflation. PROER divided those institutions between ‘good’ and ‘bad’ allowing the former to be purchased by other banks, while the Central Bank would liquidate the latter.

11 A reduction in inflation and the expansion of credit and consumption help to explain the increase in financial content reported for ‘Wholesale and retail trade’. It happened in two waves. The first started in 1995 and lasted until 2000 with financial content doubling its figures. A second wave started at that moment with a recovery of financial content per unit of output.

Additional information

Funding

Lionello Punzo gratefully acknowledges financial support from the National Institute of Science and Technology (INCT/PPED).