Abstract
Countries strategically use statistics to secure foreign aid and investment, burnish their international image and wage information wars against adversaries. The dogfight over the COVID-19 statistics that has rocked the fragile US–China relations is the latest reminder of the importance of statistics in international relations. The literature has, however, paid insufficient attention to the role of statistics in international affairs.
Making It Count examines scientific and state-building activities of the early Maoist China and highlights ‘the interplay between technical considerations and broader shifts in domestic and international politics’ (p. 7) that generated path-dependent institutional choices in the field of statistics. Drawing on archives in China, India and the United States and interviews in China, Ghosh also illuminates an interface between statistics and international relations by examining statistical exchanges among India, China and the USSR. Other questions raised in the book, which we will skip due to space constraints, relate to science in non-liberal, non-Western contexts and the assumption of universal acceptance of probabilistic thinking.
The book consists of 10 chapters divided into three parts. The first part discusses the use of numbers in pre-War China and explores the philosophical underpinnings and technical apparatus of Soviet-inspired socialist statistics in Maoist China. The second and third parts, respectively, examine socialist statistics in action, including from the perspective of statistical workers, and attempts in the late-1950s to resolve ‘challenges generated by socialist statistics’ (p. 20).
Ghosh’s account suggests that at least three things were at stake in early Maoist debates on statistics. First, the ‘nature of social reality’ and the role ‘of mathematical statistics… in ascertaining that reality’ (p. 5). Second, the appropriate means of erasing ‘national humiliation’ entailed by a ‘lack of factual self-knowledge’ and making China ‘legible’ in modern terms (pp. 58–59). Third, the distinction between the new regime and both the country’s feudal past as well as bourgeoisie imitators of ‘shameful’ Anglo-American methods (p. 112).
Ghosh points out that unlike Republican China, Maoist China’s search for answers was restricted by its political orientation. This constrained both the choice of statistical methods and design of the statistical system as Soviet-inspired statisticians maintained that the natural and social realms were not only analytically separable but also amenable to distinct statistical tools. ‘Bourgeois’ mathematical statistics was fit only for the natural realm governed by universal laws, while the social realm had to be studied using ‘socialist’ statistics (p. 48, 56). This separation meant that advances in sample surveys bypassed China resulting in heavier demands on the limited trained manpower, information overload for policy-makers, undue delays and even sustained inaccuracies. In practice, what counted as socialist statistics was a fragile combination of Soviet statistics and Maoist ethnography contingent upon a host of domestic and international factors. In other words, socialist statistics in China were an amalgam of exhaustive enumeration and typical sampling sanctified by socialist scientificity and Mao’s stamp.
Ghosh points out that once its epistemic dominance was institutionalised, socialist statistics was confronted by the reality of understaffed statistical departments and a poorly trained workforce. The choice of socialist methods, particularly, exhaustive enumeration, strained statistical departments, which in turn bred disillusionment, particularly, among the lower cadre, who were exhorted to ‘ardently love statistical work’. This was also the time when the Sino-Soviet relationship was flagging, which provided the much-needed opening for reforms that effectively meant incorporating sample surveys in the toolkit with assistance from outside the Soviet bloc. The rest of the review will focus on Ghosh’s account of the international linkages of the Chinese statistical system.
We have already noted that Maoist China’s turn to socialist statistics was governed by its Soviet orientation. Ghosh suggests that the ‘enthusiastic and comprehensive adoption’ (p. 74) of Soviet statistics was ‘voluntary’ (p. 75) and part of a broader fascination with the Soviet Union as the only role model for an industrialising socialist society. The Soviet statistical influence in Maoist China first emerged following the liberation of the Northeast region, when translation of Soviet treatises on statistics was attempted. The Northeast Statistics Bureau (NSB), headed by Wang Sihua, was ‘the first subnational statistical agency to fully incorporate Soviet/socialist approaches’ and served as the model for statistical offices elsewhere in China (p. 26).
The NSB began a systematic translation of Soviet statistics treatises on its own as early as 1950. Soviet experience and translations of Soviet works routinely featured in Chinese statistical journals during the 1950s. In 1956, the departments of the culture of the two countries began to formally collaborate resulting in a greater supply of Soviet literature. The Soviet influence was not limited to the realm of ideas though. Soviet consultants ‘offered key advice in 1953’ for ‘the first complete modern census in China’s history’ (p. 81). They also ‘spent time at designated factories and enterprises, helping set up or rationalise statistical work’ (p. 82) and inspected ‘local statistical work’ (p. 83).
The initial fascination notwithstanding, Maoist ethnography along with its decentralising tendencies was always in tension with Soviet statistical methods that entailed centralisation. Ghosh’s account suggests that several other factors fuelled disillusionment with Soviet statistical methods. First, the ‘initial idolization of Soviet models was confronted by the reality of their execution or their sheer inapplicability to Chinese conditions’ (p. 217). Zhou Enlai himself noted that ‘undue haste, arbitrary learning’ led to the ‘mechanical application’ of Soviet ideas (p. 218). Second, the Chinese complained that Moscow did not send the best of its experts. Third, Soviet experts tried to block access to alternatives. Early on, the Chinese realised that the Soviets were ‘not overly enthusiastic about… surveys’ (p. 44). Soviet experts, in fact, went out of their way to discourage the Chinese from even reading Russian works on sample surveys (p. 78). Fourth, the Chinese seemed to be wary of over-reliance on the Soviets (p. 273).
Nudged by a combination of domestic and international political factors and operational problems, the Chinese eventually turned to ‘bourgeois countries that are relatively good, especially [in] the methods of sample surveys and the problems related to statistical technology’ with the goal being ‘to provide better services in the construction of a socialist society’ (p. 241). The ‘bourgeois’ country being referred to here is India, which itself was not satisfied with Soviet statistical methods (p. 232, see also Engerman, 2018).
In the early 1950s, China had ‘rebuffed Mahalanobis’s invitation to join the International Statistical Association for Asia and the Far East’ describing it as an imperialist misadventure (p. 214). However, by mid-1950s India, under the leadership of its most prominent statistician P.C. Mahalanobis, had successfully experimented with large scale random sampling as well as contributed to the formulation of several international statistical standards. Mahalanobis, who had established independent India as a statistical powerhouse in a world hitherto dominated by western nations, was invited to China after Zhou Enlai visited the Indian Statistical Institute (ISI) in Calcutta in 1956. Zhou Enlai was followed within 2 days by a delegation led by Wang Sihua. Mahalanobis, who was grappling with the difficulties of drafting India’s Second Five Year Plan, was keen to interact with the Chinese to understand their planning experience and led a delegation to extensively tour China in mid-1957. However, the Chinese ‘still considered [public data] secret… and, as a result, many of Mahalanobis’ requests were eventually denied’ (p. 233).
The Chinese sent a delegation of their own on a year-long study tour through India in 1958, providing important exposure to international statistical developments for the otherwise isolated Maoist statisticians. In October 1958, Mahalanobis invited Chinese ‘planners with expertise in heavy industry to help with the drafting of India’s Third Five Year Plan’ (p. 247). However, much had changed in Beijing since Mahalanobis’ visit. China was hurtling towards the Great Leap Forward and the Sino-Indian border dispute was becoming increasingly intractable. ‘The archival trail turns cold’ after March 1959, signalling the end of the ‘much-touted [India–China] statistical cooperation’ (pp. 246–247). Nonetheless, two decades later when post-Mao China revisited its statistical foundations the debate was triggered by one of the delegates who visited India in 1956.
A few key issues are left unaddressed in Ghosh’s study. First, there is hardly any reflection on the impact, if any, of Taiwan’s experience on China’s choices even though during the period of interest the boundaries separating it from the mainland were still fluid. Second, the book is silent about the Soviet reaction to China’s turn to India at the time when its relations with China had begun to cool. Third, the book offers an inside out account from the perspective of official statisticians without hinting how the common people might have experienced the transition to ‘New China’ (p. 4) driven by socialist statistics. Fourth, the narrative ends in the early 1980s without offering any clues about how some of the contemporary concerns about China’s statistics might be linked to the path-dependent evolution of the socialist statistical system.
These lacunae notwithstanding, Making It Count makes three key contributions. First, it offers an insight into the formative years of Maoist China from a novel perspective. The book puts together an account of how non-linear changes and uneven ruptures generated by the intersection of ideological commitments, domestic political imperatives and material constraints and international stimuli, shaped the growth and development of a complex and decentralised statistical system in a presumably centralised polity. What Ghosh tells us about the statistical wing of the government is quite likely to be true of other wings. So, beneath the façade of centralisation, a lot of decentralisation would have thrived by default due to Beijing’s infrastructural capacity constraints. Second, it provides a good starting point to launch a historically informed debate about the quality of China’s official statistics. Popular and even academic commentaries often invoke ex post facto manipulation as the key driver of data quality in China. Ghosh, on the other hand, begins with the implicit assumption that China’s statistical system would have been put in place to provide reliable data to the government. Indeed, the early 1950s seems to be a period ‘when considerable data were available’ that ‘were widely considered to be reliable’ including by external observers such as India’s P.C. Mahalanobis (p. 1, 45, 47). Third, it opens up new avenues to understand the possibilities of South–South cooperation and explore international relations away from the shadow of diplomatic history. In conclusion, despite its historical orientation, Making It Count is a valuable source for understanding the workings of China’s official use of statistics and building a more nuanced understanding of the country’s government at large.