The drama of Japan's population history continues to fascinate. Sandwiched between two periods of population increase—the 1600s and the century and a half after 1800—was a period of plateau-like stagnation. For instance in central Honshū, Japan's main island, bristling with great metropolitan centers like Edo, Osaka, and Kyoto—an urban graveyard effect checked population increase. In Hokuriku on the Japan Sea coast, emigration kept a lid on expansion in human numbers. In Southwestern Japan expansion prevailed, albeit modest. Amongst the various subregions of Japan, the most intriguing case is Northeastern Japan (Tōgoku). Here rampant resort infanticide appears to have suppressed the birth rate, ultimately precipitating collapse in human numbers. Indeed some villages became virtually bereft of humanity, wild animals threatening to take over the abandoned fields and houses at their perimeter.
Systematic infanticide: is this possible? Part I of Fabian Drixler's carefully researched volume screams “yes.” Drawing upon a Ten (Tokugawa era) Province data set containing estimates of total fertility rates for over a thousand villages (780 thousand observations; 5.5 million reconstructed person years) between 1660 and 1790, Drixler establishes that killing newborns—referred to in Japanese by the euphemism mabiki (“winnowing out” as in plucking out rice plants from the fields so the remainder can benefit from sunlight and grow to a healthy maturity)—was common in the Northeastern region. How common?
Drawing upon the low total fertility rate village averages he extracts from his database, Drixler leads us along a systematic quantitative path in search of an answer. Distorted sex ratios provide a lower bound baseline percentage to the infanticide rate. Injecting the impact of late marriage, labor migration, separated spouses, sterility, chronic malnutrition, and the impact of damaging physical work upon expectant mothers, into Monte Carlo simulations, yields an upper bound to the rate. Deciding exactly where to locate a statistically plausible range for the infanticide rate between the two bounds poses additional problems. To what degree did abortion suppress births? What about contraception? What about stillbirths? Arguing from plausibility—for instance maintaining contraception was largely restricted to the world of prostitutes —Drixler concludes mabiki reduced the number of births by a whopping 40 percent.
To this analysis, Drixler adds qualitative evidence. Drawing upon motive tablets (ema), scrolls, posters, and literary documents, he demonstrates mabiki was widely discussed, its demerits vigorously debated. Indeed after 1790 feudal domains adversely impacted by declining human numbers, introduced childbearing grants, pregnancy surveillance systems, and—eliciting the assistance of Buddhist priests—moral suasion campaigns to combat mabiki. Why mabiki? Drixler argues that the practice became acceptable as peasant households adopted the stem family system, one and one child inheriting the farming assets amassed over generations by the household. Since in-adopting a fictive son was an option, keeping the overall level of reproduction low offered a low cost strategy for coping with non-inheriting offspring, simultaneously guaranteeing the family line was perpetuated. To be sure, in districts where mabiki was prevalent, a standard rationalization evoked crushing poverty. In the second part of the volume, Drixler advances a thesis about what infanticide slowly but surely disappeared. He argues that extinction of household lines due to low fertility; distorted sex ratios negatively affecting marriage opportunities; and Buddhist inspired moral revulsion played roles during the early nineteenth century. Subsequently, opening the country up to Western eyes in the 1850s (believing Westerners viewed infanticide with disdain) and nationalistic ardor aimed at insuring a sufficient stock of soldiers seemed to have loomed large. As a result, marital fertility rates in Tōgoku soared. During the period between the world wars, Northeastern Japan boasted the highest levels. The swing from low to high is truly remarkable.
The economist in me wanted more. I cannot deny that I think the spread of best practice technique in rice production from Southwestern Japan to the Northeast increased yields in Tōgoku. In doing so, it swelled the demand for child labor encouraging larger families. Marital fertility responded.
As well I cannot agree with Drixler that Japanese infanticide invalidates the theory of the demographic transition. Let's be clear about this: a freshly delivered baby smothered unto lifelessness constitutes a package of birth and death events. Add these mostly unrecorded births terminated by infanticide to the births recorded. What this yields is a regime of moderate birth and death rates, not a regime of low fertility. Add to this the absence of evidence that stopping behavior (parity-specific control) occurred in Northeastern Japan, hence no evidence of planned birth control, per se, further encourages skepticism about some of Drixler's conclusions.
These quibbles aside, I found Mabiki informative, innovative, and impressive.
