Tuesday, July 20, 2004
A reporter friend of mine asked me about the quality of Chinese statistics. Here are my thoughts, which is basically the same as anyone who is using Chinese statistics. "Can't live with it; can't live without it." It is better than nothing. However, in some sense, the evolving Chinese statistical system makes it terrible for statistical analysis, especially usingtime-series data.
There is no doubt that the statistical system has come a long way, from Soviet-style command economy reporting to a hybrid system where statistics is reported by local governments and verified by random surveys conducted by the NSB. The problem is that because the statistics is improving in quality all the time, it makes data across time not really comparable. GDP figures from say 1992 were collected and calculated in a very different way thanGDP figures today. What is worse, the quality of data and the rate at which data improves vary across sector. For example, I think banking and fiscal data is quite accurate in China, since it is difficult to hide anunbalanced balance sheet (with the exception of grassroots level budget;that's a big mess). The government at least knows what is happening.They may choose to be ambiguous with the public.
Investment and output data is more tricky, since you need firms and local governments to report it. As you point out, "the numbers make a cadre;(thus) cadres make numbers." The central government has to randomly sample industries and regions to verify these figures. As an analyst, there is not much we can do but to take all of ourfindings with a grain of salt.
There is no doubt that the statistical system has come a long way, from Soviet-style command economy reporting to a hybrid system where statistics is reported by local governments and verified by random surveys conducted by the NSB. The problem is that because the statistics is improving in quality all the time, it makes data across time not really comparable. GDP figures from say 1992 were collected and calculated in a very different way thanGDP figures today. What is worse, the quality of data and the rate at which data improves vary across sector. For example, I think banking and fiscal data is quite accurate in China, since it is difficult to hide anunbalanced balance sheet (with the exception of grassroots level budget;that's a big mess). The government at least knows what is happening.They may choose to be ambiguous with the public.
Investment and output data is more tricky, since you need firms and local governments to report it. As you point out, "the numbers make a cadre;(thus) cadres make numbers." The central government has to randomly sample industries and regions to verify these figures. As an analyst, there is not much we can do but to take all of ourfindings with a grain of salt.
Comments:
Post a Comment
