Three Pitfalls in Statistics in Economic History

Mataji Umemura


1. Production for Self-Consumption

In the system of national accounts, only self consumption for foods is included in the calculations. However, such an approach is appropriate only on the assumption that a market economy has fully developed in a society. In a society in the early stages of economic development, applying such a calculation method is problematic.

In such societies, much of the clothing and housing supply does not go through a market; therefore, these categories also are often not included in production statistics. The village and town officials responsible for statistics often conduct statistical surveys in addition to other responsibilities. Consequently, statistical surveys are not considered of great importance in local administrative and corporate organizations even today. Under such conditions, it is highly unlikely that local officials will have the time or resources to survey self-consumption in clothing and other categories.

Further, even within the category of food, with some items that are obtained in quantities whenever needed, such as vegetables, milk, or eggs, it is unlikely that housewives know how much was consumed per year, nor do they need to keep track. Naturally, these output statistics also are not included in national accounts for such situations. Thus, when consumption and production statistics are compared, significant gaps can be seen. Such foods are fundamentally different from rice and grains, which are harvested once a year and divided for sale and self-consumption.

2. The Importance of Subsidiary Occupations

In the early stages of economic development, labor is not very specialized, as most people live everyday hand to mouth. In such situations, in order to support the family, the head of the family must do anything that generates income. There are no differences between main and subsidiary occupations. In other words, everything is income. Statistical surveys come in from the outside and divide income into minute categories, classifying one thing as the main job, another as a subsidiary occupation. This may help us portray occupational characteristics of the population, but does not represent industry-specific labor inputs. We can only examine the extent occupational profiles represent industry-specific labor inputs through statistics on subsidiary occupations. However, if we skip over this step, we create fertile ground for gross misinterpretations.

3. The Issue of Measures

In Japan, weights and measures were standardized on March 24, 1891. Some scholars claim that the Tokugawa bakufu set official standards for weights and measures, but the bakufu could not regulate measures used within each domain (han). In the various domains, there were local practices and standards even before a market economy was established. Tributes were levies on the basis of these local practices, as were weights and measures. Thus, whenever rulers attempted to alter measures, among commoners, there was likely a fear of increased taxes. Hence, people resisted revisions of weights and measures. Revisions of measures are in effect a political issue. The fact that the metric system was implemented in the middle of the French Revolution highlights this point. The 1873 land tax reforms generated tax rates that decreased from 3.0% to 2.5% five years later, but contained no other changes. For around 20 years, rice prices rose, and agricultural production increased. As a result, the overall effect was a drastic decrease in the burden of the land tax in the early Meiji period. The standardization law for weights and measures in 1891 was enacted in response to this condition.

Before that, measures were likely based on local practices. There was a transition over many years from measures and standards based on local practice to standardized legally-binding measures. Economic statistics are influenced by such changes. To my mind, James Nakamura's criticisms of Japan's agricultural production statistics overlooks these gradual changes in measures.

(Professos Emeritas of Hitotsubashi University)