In the last decade, China’s serious gender imbalance has made headlines: millions of Chinese men are doomed to bachelorhood due to a shortage of women, with awful social consequences. The conventional wisdom is that this skewing — a sex ratio at birth far higher than the natural ratio of 105 males to 100 females — is caused simply and solely by China’s one-child policy. Given Chinese parents’ supposedly ancient cultural preference for sons, the argument goes, if they can only have one child, it had better be a boy. That the sex ratio started skewing around 1985, about five years into the new birth planning policy, seems proof enough. The claim’s logical conclusion is that abolishing the policy will eliminate China’s sex ratio imbalance.

But some basic comparisons show us how the conventional wisdom is faulty. The claim that the skewed sex ratio has occurred due to the birth planning policy usually rests on two premises: that it alone caused fertility decline and that this fertility decline led to a skewed sex ratio at birth. But, in fact, China’s fertility decline began in the early 1970s, years before the one-child policy. And, China’s neighbours had similar fertility declines in the 1970, even without such draconian policies.

Fertility declines also do not inevitably lead to a skewed sex ratio. Japan’s fertility rate has been among the world’s lowest for nearly two decades, but its sex ratio at birth has remained in the natural range. While China’s skewed sex ratio of the past three decades does coincide with the start of the birth planning policy, this coincidence is somewhat misleading. A longer historical view reveals that China’s sex ratio was terribly skewed over much of the last two centuries.

The one-child policy itself is a bit of a misnomer: three distinct policy variations are in place across China. Rural, majority-Han areas practice a ‘1.5-child’ policy, in which families whose first child is a girl are allowed a second in hopes of having a boy. Urban areas have a strict one-child limit, while poor ethnic-minority areas have a two-child limit.

Sex ratio skewing is higher in rural 1.5-child policy areas (about 119:100 at birth) than in urban one-child areas (about 115:100), and is lowest in two-child policy areas (about 112:100). These numbers indicate that switching to a universal two-child policy would reduce but not eliminate the problem: China’s sex ratio at birth would still be higher than almost anywhere else in the world. So there is some truth to the conventional wisdom, but the birth planning policy is not the only important driver of sex ratio skewing.

The problem with the conventional wisdom is that it treats son preference as a cultural given: it says Chinese people just prefer sons. But son preference is not a constant. Incentives for Chinese families to have sons have changed considerably over time, rising and falling in tandem with a skewed sex ratio at birth. So efforts to normalise China’s sex ratio at birth ought to attack existing incentives for families to have sons.

Son preference incentives appear in four realms: labour, property ownership and inheritance, ritual life, and old-age security. Societies with strong incentives in these areas tend to have a skewed sex ratio. Indeed, differences in the sex ratio at birth parallel differences in levels of son preference incentives through time, across regions of China and across countries. Old-age security seems to be the most important driver of son preference, while ritual-related incentives matter less.

China’s sex ratio at birth was skewed before 1960, normal during 1960–85 and skewed again after 1985. In both periods of high skewing, sons were highly necessary on all four measures: for farm labour, property inheritance, ancestor worship and old-age care. In contrast, during the commune era (approximately 1958–83), production was socialised, property was collectivised, ancestor worship was suppressed and pensions for the elderly were provided by the commune. Families didn’t require sons, and so they had little incentive to practice female infanticide or abandonment.

Since the mid-1980s, son preference incentives have differed starkly between urban and rural areas of China. In urban areas, educated women make important economic contributions to their birth families and are thus able to provide old-age care for their parents. Ancestor worship is also less relevant to urban life compared to rural areas, while urban women also share equally in property and inheritance. Altogether, these factors mean that urban families have less incentive than rural ones to prefer sons.

The sex ratio at birth has also changed along with son preference incentives in Japan and South Korea. Japan scores low on our measures of son preference throughout the 20th century. Women contribute meaningfully to family income and inherit property equally, and Japan has excellent old-age pensions. While sons are strongly preferred for family rituals, Japan’s sex ratio at birth is not skewed. As a second example, South Korea saw a sharp increase in sex ratio skewing in the mid-1980s, followed by a decrease after 1995. This normalisation coincided with changes in South Korean family law specifying that women did not have to marry into their husbands’ families, had equal rights and responsibilities in ancestor worship and equal inheritance rights.

Normalising China’s skewed sex ratio will require a concerted effort to reduce son preference, targeting policies and institutions that create these incentives. Simply haranguing Chinese citizens to change their ‘backward’ ways of thinking and culture will not suffice.

Elizabeth Remick is an associate professor of political science at Tufts University. Charis Loh is an independent scholar. This article was originally published in the East Asia Forum on 25 August 2015. It is republished with permission.