Physical exertion and stature in the Habsburg monarchy, 1730-1800. (response to John Komlos and Albrecht Ritschl, Journal of Interdisciplinary History, vol. 26)

Heights were falling precipitously in the Habsburg empire during the second half of the eighteenth century. Two alternative interpretations have been advanced to explain this peculiar trend. According to the "Austrian model," proposed by Komlos, rapid population growth caused increasing strains in the empire's economy, and the deterioration in the ratio between resources and population - portending an imminent Malthusian crisis - caused drastic reductions in human stature. According to Komlos, only enlightened government intervention saved Empress Maria Theresa's subjects from starvation. In a contribution to this journal, I challenged Komlos' view, on the grounds that he ignored the large number of holy days that were abolished as part of the reform effort. Additional labor, I argued, reduced the amount of nutrients available for growth. Falling heights were not a sign of impeding Malthusian doom, but of a more productive socioeconomic regime. I demonstrated that the decline in the stature of army recruits, identified by Komlos, coincided with the two major reductions in the number of holy days within Habsburg lands. If the effect of additional working days is discounted, there is no trend toward smaller stature.(1)

In their reply to my article, Komlos and Ritschl provided new evidence supporting the Austrian model. They conceded that Komlos' initial contribution overlooked the major changes in the festive calendar instituted under Maria Theresa. Their refutation of my earlier argument consisted of two elements, the first theoretical and the other based on Komlos' data set of Austrian heights. Additional labor, they argued, not only reduced the amount of nutrients available for such purposes as growth; it also led to additional output, thus increasing the total amount of calories available. Further, their reestimation of my regressions on the basis of Komlos' full data set (instead of the published data used in my earlier article) showed that additional work did not affect heights significantly. Contrary to my contention, they found that the additional output generated from the new days of work actually boosted stature, even if the effect could not be estimated with great accuracy.(2)

In this rejoinder to their comment, I demonstrate that the negative impact of physical exertion postulated in my earlier contribution (and demonstrated by third world studies) holds for Komlos' larger data set as well. The technique used by Komlos and Ritschl is unsatisfactory from an econometric point of view; the insignificant coefficients that they found are due to misspecification and error. After controlling for additional labor input, my evidence demonstrates no trend toward diminishing stature in the second half of the eighteenth century. I also refute the theoretical arguments against physical exertion as the prime cause of diminished stature, and discuss implications of the present debate for the historiography of the Habsburg empire: Conceptual differences have been overstated. What is subject to debate is the context in which enlightened government intervention took place. Finally, these points lead me to some remarks about the anthropometric research project in general.

On a purely empirical level, Komlos and Ritschl argue that the timing of reductions in the height of Austrian recruits shows no significant association with the abolition of holy days. Instead of using the aggregated data on which my earlier contribution was based, they deploy the full data set collected by Komlos, containing a total of 145,873 observations. As my critics advise, the larger number of observations overcomes some of the statistical problems associated with a limited number of degrees of freedom. In contrast to their assertions, however, both aggregated and disaggregated data give the same result: When we control for the effect of additional physical exertion, there is little evidence of an impending Malthusian crisis in eighteenth-century Austria.

Komlos and Ritschl's reestimation of the regression equations from my original article, using the full complement of observations was intended to relate the height of recruits to their occupational status (skilled or unskilled), their location (urban or rural), their age, their province of origin, and the number of working days in the year (which serves as a proxy for work effort). In addition, they included a set of decade dummies to test for trends over time,

H = [Alpha]O + [Beta]L + [Gamma]A + [Delta]P + [Phi]W + [Psi]D, (I)

where H is height, O is a dummy variable for occupational status, L is a dummy for urban background, A is a vector of dummies capturing age, P is a vector for …