How Reliable Are China's Economic Statistics?
With so much attention on developments in the Chinese economy, the reliability of China's economic indicators has been the subject of some controversy. In the popular imagination, the production of China's economic data is regarded as a crude political farce: the controlling hand of the Communist Party intervening arbitrarily to direct the level of key indicators before they are published. In the past, that image was not too far from the reality.
In the Great Leap Forward, Chairman Mao's disastrous attempt to shift a backward agrarian economy to a modern industrial economy in a few short years, the failure of the statistical system contributed to catastrophe on a grand scale. Mao's plan to accelerate the development of the economy required producing an agricultural surplus that could be sold to fund investment in a modern industrial base. Whipped into a patriotic frenzy, and knowing that their future depended on meeting unrealistic targets for the production of grain, local officials engaged in rampant falsification of data. At the height of the insanity, in summer 1958, some provinces were boasting annual output of 10,000 pounds of grain per mu. Those numbers were already crazy. A mu is 1/15 of a hectare, and in the 1950s that area of land could be expected to produce around 1,000 pounds of grain a year. But even these bloated reports were soon surpassed by other provinces claiming output of 20,000 or even 30,000 pounds per mu. The point of the production data was not to reflect reality, but to generate the good news required to support morale among comrades and meet the unrealistic expectations of the higher-ups. But reality is distorted at a cost. The higher the production figures, the greater the tax owed to the central government. In some provinces, the exaggerated claims were so great that the entire harvest had to be handed over as tax, used to fund investments and extravagancies that China could ill afford. In some parts of the country, the only crops left behind were grown by villagers in secret locations, away from the acquisitive eye of the local production teams. But such success stories were few and far between. Tens of millions died in history's greatest man-made famine.
That was 50 years ago. Some things have stayed the same in the last 50 years, but a lot has changed. At its heart, the cause of over-reporting output during the Great Leap Forward was the divided loyalties of local officials, torn between the reality of stubbornly unchanging grain yield and the dream of career progression that depended on meeting unrealistic targets for output. That conflict of interest was slow to be resolved. The biggest reform-era controversy over China's economic data, a GDP growth figure for 1998 that many experts regard as grossly inflated, has been laid at the door of the exaggerated claims made by local officials. But NBS headquarters in Beijing is no longer reliant on the unreliable inputs it receives from local bureaus. Across the key industrial output, fixed asset investment, and retail sales data, the largest enterprises in the country report directly to the NBS in Beijing. Where there is a conflict between local and national data, the NBS typically resolves it in favor of the reliable national figures. The GDP data is a case in point, with the national total announced by the NBS consistently below the sum of the GDP reported by the provinces.
The second problem that bedeviled the grain data during the Great Leap Forward was the belief that boosting morale through exaggerated claims was more important than reporting reality. The audience for China's economic data might have expanded beyond the agricultural workers of the 1950s, but the numbers continue to play a role at home and abroad in buoying confidence in the China growth story. The magic 8% figure for GDP growth, in particular, has an almost talismanic significance. But if the government is ever tempted to play fast and loose with the statistical reality, there are also forces pulling in the other direction. The Information Age has reduced the scope for the use of economic data as an instrument of propaganda. Official data is available instantly around the world over the Internet. A horde of sophisticated and cynical journalists, spreadsheet-wielding economists, and hard-nosed investors are following developments in the Chinese economy. The extent of the 1950s famine was revealed only by demographers poring over the population data decades later. If an inconsistency arises today, a lot of people will know about it very quickly. The fact that the NBS was happy to announce GDP growth of 6.1% year on year in the first quarter of 2009, the lowest level in a decade, suggests that the powers-that-be realize that the benefits of supplying credible official data are greater than the costs of reporting an unwelcome reality.
Measuring a rapidly changing economy remains a challenge. One of the problems of the Great Leap Forward was that China's leaders were blinded by a belief in their own hocus-pocus technology. Mao genuinely believed that revolutionary fervor plus new planting techniques could result in massive increases in grain output. Changes in production techniques made it more difficult to measure output, or at least obscured for a time the fact that output was little changed. The dislocations of reform-era China might be less wrenching than those of the 1950s, but the mainland is still changing fast. The economy is many times larger today than it was in 1978, new sectors are driving increases in output and employment, new products are entering consumers' shopping baskets, and new property is coming online in the housing market—keeping track of developments is an enormous challenge. In some areas, the NBS has made an effort to keep its measuring tools sharp and clean. To keep track of GDP, the NBS has expanded its survey from a primitive 16 sectors to a more respectable 94 and has significantly improved its coverage of the services. But in other areas, surveying tools and techniques have been slow to adapt to a changing reality. Consider China's creaking system for measuring developments in labor markets. In 1978, 100% of the workforce was employed in the state sector, and a survey based on wages in state-owned enterprises worked well enough. In 2010, that number was less than 10%, and a wage data survey that continues to focus on a privileged subset of state-sector workers makes little sense. Survey tools that lag behind the reality of a changing China are a more serious problem for China's economic data than political interference.
A recalcitrant population continues to add to the problems. The NBS is not the State Administration of Taxation or the Public Security Bureau. But a culture of deceit among Chinese people when it comes to dealing with officials of any kind makes it difficult for the NBS to collect solid baseline information. In the Great Leap Forward, peasants growing crops outside the greedy gaze of the local production team distorted the data. Fifty years later, the problem is small businesses that keep three sets of books—one for the taxman, one for investors, and one for themselves—or rich households that refuse to disclose the income they receive from graft. The problem of a sample set that is incapable of telling the truth to anyone in an official badge remains, and that adds to the difficulties the NBS faces.
Finally, the NBS and other arms of the government charged with the production of China's economic data do themselves no favors by treating straightforward information on methodology with a degree of secrecy more suited to guarding the location of nuclear weapon silos. Transparency on the methodology underpinning key data points has improved considerably from the situation a few years ago. In 2010, for example, the results of a new effort to measure wages in the private sector were published alongside details of the survey approach and a discussion of some of its limitations. But crucial details of the methodology on key indicators are still kept hidden. The weights of different components in the consumer price index (CPI), for example, are public information in most countries and enable an easy check on the reliability of the official data. In China, the NBS does not disclose the weights, making it impossible to verify if movements in the index represent changes in prices or changes to the calculation method. Partly as a result, confidence in the CPI data is so low that, in some months, the rumor in the market before the release of the data comes in two parts—one on the real level and one on the level the NBS will announce. The statisticians do not have an enormous fund of good will and faith in the official data to draw on. By withholding key details of how the data is calculated, the NBS and other institutions raise doubts, perhaps unnecessarily, about its reliability.
The reality of China's economic data today is not the crude controlling hand of the Politburo dictating the GDP growth figure. It is an increasingly reliable and comprehensive set of economic indicators that remain compromised in some areas by the difficulty of measuring a rapidly changing economy, imperfect surveying methods, a recalcitrant sample set, and continued political sensitivity surrounding some numbers. The system is not perfect. Some data points are more reliable than others. But neither is it a farce. As shown by the mad scramble for the GDP data in the State Council Information Office and the billions of dollars that are traded instantly on its release, the shortcomings in the data are no impediment to the market reaction.