Statistics interest about 10% of the population. The processes that determine the underlying accuracy of statistics probably interest 1% of the population. And the number of people willing to spend more money to improve the accuracy of statistics approaches zero. But, unfortunately, without good statistics, those interested in economics are flying blind without a map.
As a young man working as a development economist in Pakistan, I quickly learned one central reality: never believe the reported numbers until you understand where they come from and have double-checked them against other numbers. It is seldom the case that anyone is deliberately lying, but it is often the case that they don’t know what is happening or simply write down numbers they think their bosses want to hear.
Collecting good data is expensive, complicated, and requires a central statistical agency with complete political independence. Few under-developed countries have the money or the expertise to generate good data. And even fewer give their statistical authorities the necessary political independence.
China as a case in point
Take China, for example. The country is much talked about on the economic pages of our newspapers – often due to reported growth rates. For the entire decade of the 1990s and again in the first quarter of 2004, China has claimed an annual growth rate of 9.7%.
First, how were the numbers generated? China has a bottom-up reporting system and a top-down promotion system. Local leaders send reports to Beijing detailing how fast their regions are growing and these are added together to generate national data.
But Beijing also determines who among those local leaders gets promoted. Do a good job as a mayor in a small town and they will make you a mayor in a middle-sized town. Do a good job in a middle-sized town and they will make you a mayor in a big town. Do a good job in a big town and they will make you mayor of Shanghai. From Shanghai, one can jump to the central committee and become one of the rulers of China. Would you mail in a report that said that your region wasn’t growing as fast as the rest of the country? It could end your political career. Economic exaggeration is the route to personal success.
In 2001, the central statistical agency reported that there were more than 60,000 cases of over-reporting. Every province but one reported a growth rate in excess of 8.5%, but they were judged down to 7.3%. Why 7.3% and not 6.3% was not reported.
Consistency
Second, what about statistical consistency? China admits that there has been little or no growth in the countryside in the last 15 years. A total of 70% of Chinese citizens live in the countryside. Do a little simple algebra. If a country grows at 10% per year and 70% of its citizens do not enjoy any growth in their incomes, how fast do the incomes of the other 30% have to grow?
The answer is 32%. Shanghai and other Chinese cities are growing rapidly, but they are not growing 32% a year. The same doubts arise when growth rates are double-checked against electricity consumption. In every country in the world, electricity use grows faster than GDP for some very simple reasons. To produce anything, motors have to be turned on. Consumer buys products that use electricity. Something gets plugged in. For the 12 most rapidly-growing countries in the world, it takes a 2% increase electricity consumption to generate a 1% increase in GDP. Not surprisingly, what was true for the rest of the world was true for China in the 1980s. Electricity usage grew about twice as fast as output.
In the most energy efficient country in the world – Japan – it takes a 1.4% increase in electricity usage to generate a 1% increase in GDP. To get to this result, Japan has had to shut down all electricity-intensive industries such as aluminum production.
But if we are to believe that 9.7% Chinese growth rate in the 1990s, China got a 1% increase in GDP for every 0.8% increase in electricity usage. China knows something about energy efficiency that no-one else knows. And whatever they know, they did not know it in the 1980s when they had patterns of GDP and electricity usage much like everyone else.
Commodity prices
Today, the financial press reports that commodity prices are rising because there is a boom in China, but if we are to believe Chinese statistics, there is no boom in China. The growth rate in early 2004 is exactly what it was for the entire decade of the 1990s.
Moreover, China is too small economically to determine world commodity price levels. In global macroeconomics, international purchasing power is what counts and China accounts for just 4% of the global GDP using market exchange rates.
Even a big boom in China isn’t going to have a big effect on world demand when China is that small. More likely, the run-up in commodity prices can be traced to a global inventory rebuilding process that feeds on itself as buyers get worried about rising commodity prices. It does not take much of an inventory swing to generate a huge increase in global demand for commodities.
Everyone talks about China as if it is much larger than it is economically because of its huge population and the fact that the World Bank and the IMF always publish GDP statistics using purchasing power parity (PPP) measures. These are the correct numbers if one is making welfare comparisons about relative standards of living across countries. In welfare terms, China’s per capita GDP is closer to $4000 than it is to the $1000 one gets using official exchange rates. But PPP numbers are not the right numbers when doing global macroeconomics.
Suddenly, in 2003, India reported a big acceleration in its growth rate to 8%. Should we believe it? None of the inputs into the economy dramatically changed. Neither local nor foreign investments went up. The economy was not dramatically deregulated. India’s reported exports of software are three times America’s reported imports of Indian services, a statistic that includes American tourists in India and American imports of Indian movies.
If you cannot find the causes of the Indian acceleration in the inputs into the economy, the chances are that the acceleration did not occur. The more probable answer is national pride. India cannot let China consistently report growth rates much better than it has. There are unacceptable political and military consequences if those numbers are believed and the China-India gap gets too big for too long.
But let us not throw stones at others. Let us turn the spotlight of statistical accuracy on the US.
When our statistical data collection systems were set up in the early 1930s, they focused on careful collection of data and definitions of output in agriculture, mining, construction, and manufacturing.
Together, the goods sector accounted for 65% of the US economy. Everything else was lumped into a category called services, with very little effort being devoted to either data collection or careful definitions of outputs. Since the heart of the growth process was to be found in the goods sectors of the economy, the effort to collect good statistics was concentrated there.
Today, we have neither good data collection nor good definitions of output in the service sector, yet services account for 78% of US output.
Maybe in early 2004 the US economy is growing at the announced 4%-plus rate of growth and maybe it is not. It depends upon the quality of the service data being collected and that isn’t very good.
Good measures
The issue is a simple one. Congress has never been willing to appropriate the money necessary to develop good output measures in areas such as healthcare and undertake the surveys necessary to collect good statistics in the services sector.
It is not easy to design good measures of output in some service areas, but that is also true in the goods sector. Today’s office buildings or cars are not yesterday’s office buildings or cars.
What is the extra value of a car that has to be repaired less often? Output measures have to have quality improvements built into them and that is not easy in any sector of the economy.
Bad data in the service sector is not a matter of intrinsic difficulty, it is simply a question of money. And in the meantime, those who are at the controls of both the Chinese and American economies are flying blind without a satellite navigation system.
Lester Thurow is a professor of economics and management at the Massachusetts Institute of Technology
