Chinese / English
Home >> Great Minds China Forum >>  2nd Great Minds China Forum >>  China's Economic Growth: A Historical and Empirical PerspectiveChina's Economic Growth: A Historical and Empirical Perspective
James Kai-sing Kung

The award ceremony of “China's Economics Prize 2016” (National Economics Foundation), together with the 2nd Great Minds China Forum, was successfully held at the Renaissance Beijing Wangfujing Hotel in Beijing on 4th December 2016. Professor James Kai-sing Kung, of the Hong Kong University of Science and Technology and a discussant of Debin Ma’s speech, opines that the study of Chinese economic history is still at a nascent stage, with much room for growth and development. He suggests that apart from the conventional topics in economic history, it would also be fruitful to examine historical issues by way of what are popularly known as “natural experiments”.


Below is a translation of Professor Kung’s speech:


I only received Professor Ma’s slides a couple of days ago, which gives me a greater degree of freedom to discuss his speech and indulge in what I would like to speak on.

Debin (Professor Ma) mentioned the comparative advantage of Chinese scholars, which is the first point I want to comment upon. I remember back in January of 2004, I met Professor Ma for the first time at a conference held at the Chinese University of Hong Kong. Actually before that meeting many well-known scholars of China’s economic history, with the likes of the late Ramon Myers, Thomas Rawski, and Loren Brandt, had all asked me if I had heard of the name Debin Ma. I said I hadn’t. They told me that I should remember this name, for he was a rising star. Indeed, it wasn’t long before I came across his name in many prominent journals of economic history and economic development. Ma is the first Chinese scholar whose rise in the academic field of China’s economic history testifies to the importance of his idea of realizing one’s “comparative advantage” (as an ethnic Chinese); prior to that research on China’s economic history was dominated almost exclusively by Western scholars.


By now I am quite familiar with Professor Ma’s work. Over the past decade or so he has compiled and constructed data series on both wages and prices, which are essential for historical economic analysis. Above all, the “Great Divergence” debate cannot be addressed at all in the absence of data on prices and wages, which are absolutely necessary for measuring the standard of living. But these numbers have other uses as well. For instance, comparative research on age heaping between China and Europe as conducted by Professor Ma reveals, if indirectly, the destructive role of peasant rebellions in Chinese history, for they had wreaked such havoc on the living standards of the common people as well as destroying human capital. It was only recently that Chinese scholars began working on the long-term consequences of the Taiping Rebellion, for example, the largest peasant rebellion in Chinese history. The Opium War represents another example demonstrating the importance of time series data on prices and wages. A question of overriding importance underlying this historic event of epic proportions is whether the opium trade had had any effect on the real economy. To address this question, we must first and foremost work with data on prices and wages, without which we cannot say much about the standard of living and real earnings as a consequence of such trade. It is from this vantage point that Professor Ma’s groundwork in data construction proves useful for other economic historians interested in exploring a wide gamut of important topics related to China’s economic history and testing the pertinent hypotheses empirically.


Debin also mentioned the importance of using natural experiments in economic history. Let me elaborate a bit on this. As economists we pay much attention to causal relationships. Unfortunately, unlike science in historical analysis we are unable to employ such powerful methods as controlled and replicated laboratory experiments, so we have to look for the “second-best” compromise. Jared Diamond and James Robinson coined the term “natural experiments of history” to describe a method that is an imperfect substitution for manipulative experiments (Diamond and Robinson, 2010). In my remaining speech I will introduce several studies based on the use of this method in studying the economic history of China. But before that I need to emphasize the importance of data and accordingly quantitative analysis in the context of natural experiments. Increasingly, I feel very fortunate because our ancestors have bequeathed upon us an invaluable trove of historical materials. The amount and variety of data available for exploitation is really beyond our imagination. What’s more, experience suggests that it is possible to construct unique data sets from various sources and combine them in a creative manner to address questions of unique historical interest.

In the past ten years or so I have had the pleasure and fortune of working with a number of graduate students at the Hong Kong University of Science and Technology. Let me briefly introduce to you some of these projects. The first is Sino-nomadic conflict. Throughout world history, many settled civilizations have been invaded and conquered by nomadic tribes, amongst which the bullying of Han Chinese by their nomadic neighbors was the longest and most enduring. While reading Sima Qian’s account of the nomads, which he characterized as “war loving” and “barbaric”, I was immediately struck by the idea that the underlying causes of the Sino-nomadic conflict can be empirically tested. Why? We know from the economics literature that economic (climate) shocks play a pivotal role in causing conflicts (see, e.g., Edward Miguel of the University of California Berkeley’s famous study, in which he and his coauthors exploited climatic variations to exogenously identify the effects of economic performance on conflict). Coming back to the historical Chinese context, we asked the question of whether the nomadic invasion was rooted in their alleged war-loving nature, or whether as a pastoral people and hence dependent on rainfall the nomads were forced to invade the Han Chinese in the face of weather adversity. To test this empirically, we used data on historical warfare and climatic variations in China—specifically droughts and flood—as exogenous predictors. The graph here provides a glimpse of the result: represented on the y-axis, nomadic invasions are positively correlated with an increasing incidence of droughts and negatively correlated with an increasing incidence of floods over time (which is represented on the x-axis).


The second project is about trade suppression and piracy. Interestingly, this natural experiment is inspired by Zhu Yuanzhang—the first emperor of the Ming dynasty.

Owing to a multitude of reasons overseas trade in China was prohibited by a “sea ban” policy in the Ming dynasty. Curiously, piratical activities, much of which was highly organized and armed and resulted in massive casualties, rose in tandem (as a matter of fact piracy in Ming China has remained the most rampant to this day). By some historical account, many pirates were in fact merchants before 1550. To put things in perspective, I plot in this graph the frequency of piratical attacks on a yearly basis for the entire Ming dynasty. Ironically, if you ask historians they will tell you systematic data such as these are not readily available. But in fact they are available in fine detail in the Veritable Records of the Ming (e.g., the frequency of attacks on counties and prefectures). Moreover, we can combine these data with other historical sources to construct a unique data set for analysis. In a nutshell, we are able to show that coastal prefectures with higher trade potential—where trade potential is measured by whether these prefectures were silk centers or historical ports and enjoyed above average economic prosperity—were more likely raided by the pirates as a result of the sea ban (trade suppression).


The third project pertains to the possible role of culture in conflict mitigation. The literature on conflict draws heavily on Africa, which has had a shorter history than China. Because culture is long lasting and tends to change only very slowly, China presents an ideal situation for testing the possible effect of culture on conflict mitigation. We use information on crop failure to proxy for economic (weather) shocks and the number of Confucian temples and chaste women to proxy respectively for loyalty and filial piety, which are the very essence of Confucian norms. As with the other two studies mentioned earlier these data did not exist before, but they were all recorded in a historical source entitled the Encyclopedia of Shandong Province (Shandong Tongzhi), which we coded systematically. Why Shandong? On the one hand it was the birth place of Confucius, yet on the other hand it was also home to the many social bandits and popular unrests in historical China. By combining the data drawn from Encyclopedia with the Veritable Records of the Qing we were able to construct the variables on peasant rebellion, crop failure, the number of Confucian temples and the number of chaste woman—variables required for testing our hypothesis. In short, we found significantly fewer peasant revolts that were triggered by crop failure in counties with stronger Confucian norms.


Our fourth natural experiment concerns maize adoption and population growth in late imperial China (pretty much the Qing dynasty)—a study motivated by two rather different strands of literature. First, a recent study has found that crops (the potato) imported from the New World (the Americas) into the Old World (predominantly Europe) had caused not only population but also economic growth measured by the rates of urbanization (Nunn and Qian, 2012). Second, Qing China had experienced explosive population growth but, unlike Europe it did not result in an industrial revolution. Since China had adopted maize and (to a lesser extent) sweet potato in Qing times, it raises the question of whether these two New World crops had anything to do with the explosive population growth in China’s last dynasty, and accordingly their effects on the country’s urbanization (as a proxy for its economic prosperity). Again, local gazetteers contain information on whether a prefecture had ever adopted maize and if so the time of adoption. The graph you are seeing now shows the diffusion of maize across China from 1580 to 1900. While we found a significantly positive relationship between maize adoption and population growth, this New World crop failed to bring about urbanization, which corroborates well with the historical fact that China was unable to break the stranglehold of a Malthusian regime; i.e., in the absence of a technological breakthrough an increase in food supply would only induce an increase in population growth, which in turn would consume the increased food resources.


Last but not least, my fifth example may be regarded as an empirical test of Max Weber’s thesis regarding the relationship between the Protestant ethic and what he termed “the spirit of capitalism” in the historical Chinese context (circa the mid-19th to early 20th centuries), which according to Acemoglu, Johnson, and Robinson (2005) is the “most famous link between culture and economic development.” Specifically, we asked the question of whether Protestantism promoted economic prosperity in China, by drawing upon the analysis of a survey data set conducted during 1901-1920 by the missionary Milton Stauffer, later published as “The Christian Occupation of China: A General Survey of the Numerical Strength and Geographical Distribution of the Christian Forces in China.” We first digitized the data collected in this survey, and then combined it with other data sources—most notably Zheng Yufa’s valuable compilation entitled “The Private Industries in the Late Ch’ing and Early Republican China” and the microfilmed missionary diaries, among others. We found that Protestantism did promote economic prosperity, where prosperity is proxied by the degree of urbanization and modern industrial establishment across China. Moreover, we showed that it was the diffusion of Western knowledge, as measured by schools with a modern curriculum and hospitals, rather than culture that contributed to economic prosperity.

Most recently we have been examining the long-term effects of China’s imperial examination system (keju) on human capital outcomes. We begin with the intuition that, with a millennium-long history keju may have fostered a culture among the Chinese of valuing education (and hence a distinct willingness to invest in human capital). For example, in contemporary times we can see the importance the Chinese society has accorded to Gaokao (formally known as the National Higher Education Entrance Examination). Specifically, we find that prefectures that historically produced more jinshi (jinshi was the highest level of exam achievements under keju) have higher human capital outcomes today as measured by the years of schooling. The same can be said for GDP. Data on jinshi is systematically available in the Directory of the Ming-Qing Imperial Exam Graduates, so what we did—as with our other projects—was to digitize them.


To conclude, research on the economic history of China is still at a nascent stage, hence there is much room for development. For that reason I would strongly encourage students and young scholars to find the many hidden treasures in this field of inquiry. With a rich variety of empirical sources spanning a long historical period, data should not be an obstacle to quantifying economic history, as long as you are willing to raise questions. Earlier on Professor Qian has mentioned the importance of raising good questions in doing research. Here my advice is, one may adopt the strategy of “walking on two legs” (a famous saying of Chairman Mao in the context of strategizing China’s economic development); to wit, one can certainly work on such conventional topics in economic history as the impact of the opium trade on China’s real economy and standard of living, the relationship between population growth and market integration, social mobility under the keju system, and the likes, but we should go beyond that—as I hope I have successfully demonstrated—and also conduct a variety of natural experiments in history.




◆please indicate the source if authorized: National Economics Foundation

◆photo:National Economics Foundation