Was there a Marxian bias in Austrian manufacturing? Evidence on the direction of technical change, 1978–1994

Abstract

This paper presents a study of the bias of technical change for 20 industries of the Austrian manufacturing sector for the period 1978 to 1994 using empirical methods derived from both classical and neoclassical economics. Overall, the results suggest that the technical bias in Austrian manufacturing did not follow a Marxian pattern. Estimates based on the classical framework suggest that 5 out of the 20 industries show a Marxian bias over the period 1978–1994. Evidence based on a neoclassical CES production function is provided and critically evaluated.

Keywords:

• capital productivity
• bias of technical change
• aggregate production function
• Austrian manufacturing industries

JEL Classifications:

• B51
• O30

Acknowledgements

I am grateful to two anonymous referees for their comments and suggestions. I thank Jesus Felipe, Thomas Grandner, Dieter Gstach and Andreas Reinstaller for useful comments on earlier versions, and Robert Leisch for his excellent research assistance. This paper has been presented at research seminars at the Vienna University of Economics, the Vienna University of Technology, as well as at the EAEPE 2006 conference in Istanbul. I would like to thank the participants for their comments and suggestions. The usual disclaimer applies. The research was supported by a research grant (Project no. 9800) from the Jubilaeumsfonds der Oesterreichische Nationalbank (OeNB).

Notes

1. The periodization is dictated by data availability constraints; a change in industry classification makes it impossible to extend the period.

2. For a discussion of the differences between modern classical and neoclassical economics which are sympathetic to classical economics, see Kurz and Salvadori (1995) or Foley and Michl (1999). For a critical point of view, see Hahn (1982).

3. There is no capital productivity in classical theory; however, I will abuse terminology for convenience.

4. Only the technique of production is depicted in the figure. The determination of one of the distribution parameter (wages or profits) would allow determination of the distribution of income. However, this is not needed for the investigation of the direction of technical change. The interest is in changes of technique, not in changes of distribution parameters.

5. Note that this methodology is not immune to the issues raised by the Cambridge capital controversies (Harcourt 1972; Kurz and Salvadori 1995).

6. The use of the Hodrick–Prescott filter to adjust annual productivity growth rates for business cycle movements led to slightly different result, but did not change the substance of the results.

7. Different estimates for different specifications are reported also by Nerlove (1967), Antras (2004), Jalava et al. (2005). Also Arrow et al. (1961) in the first estimation of the CES production function report different values of a across specifications.

8. One objection to these results might be that the functional form imposed on the factor augmentation term does not fit the data. Technical change may not be a smooth process. Experimenting with other functional forms did not lead to more efficient estimates.

9. The proof is as follows: from differentiating (9) with respect to time and expressing it in terms of proportional growth rates results in , where a_i,t is the factor share of capital in value added. By assuming that income shares remain constant a_i,t = a_i and setting follows . Taking anti‐logarithms gives . This expression is mathematically equivalent to the Cobb–Douglas production function with constant returns to scale and exogenous technical progress given by v_i,t .

10. The Fachverbandsgliederung (see also table A2) is an institutional industry classification by the Austrian chambers of commerce. It is similar to the two‐digit SIC code.

11. Fraumeni (1997) presents the estimates of retirement and associated depreciation weights used by the United States Bureau of Economic Analysis (BEA) for a variety of assets. Katz and Herman (1997) provide estimates of the US capital stock covering the period 1929–95.

12. The estimates are available from the authors upon request.