Can you trust Chinese government statistics?

by on May 16, 2014 at 3:06 am in Data Source, Economics, Political Science | Permalink

Political scientist Jeremy Wallace has a recent paper on this topic:

Economic statistics dominate policy analyses, political discussions, and the study of political economy. Such statistics inform citizens on general conditions while central leaders also use them to evaluate local officials. Are economic data systematically manipulated? After establishing discrepancies in economic data series across regime types cross-nationally, I dive into sub-national growth data in China. This paper leverages variation in the likelihood of manipulation over two dimensions, arguing that politically sensitive data are more likely to be manipulated at politically sensitive times. GDP releases generate headlines, while highly correlated electricity production and consumption data are less closely watched. At the sub-national level in China, the difference between GDP and electricity growth increases in years with leadership turnover, consistent with juking the stats for political reasons. The analysis points to the political role of information and the limits of non-electoral accountability mechanisms in authoritarian regimes as well as suggesting caution in the use of politically sensitive official economic statistics.

All good points.  I would stress, however, that Chinese statistics have many problems in them and so they are not simple overestimates of how the economy is doing, at least not over the last thirty years as a whole.  In some ways Chinese growth statistics have been, until 2008-2009, probably underestimating the actual progress on the ground.  In general, growth figures underestimate progress when changes are large, and overestimate progress when changes are small.  (One reason for this is that extreme progress brings a lot of new goods to the market and their marginal value is underestimated by their price ex post, since it is hard to adjust for the fact that the price ex ante was infinite or very high.)  In Western history for instance, our most significant period of growth was probably the late 19th through early 20th century, when the foundations for the modern world were laid, yet estimated growth rates for this period are not astonishingly high.  We’re missing out on the values of the new goods, for one thing.

For the pointer I thank Henry Farrell.

Dr.D May 16, 2014 at 3:50 am

I hear your argument, but please enlighten us what genuinely new goods have come out of China during that period?

dan1111 May 16, 2014 at 5:30 am

I think he means new goods coming to market in China. e.g. how many Chinese owned cars 30 years ago?

prior_approval May 16, 2014 at 3:54 am

‘In Western history for instance, our most significant period of growth was probably the late 19th through early 20th century’

Depends on where you place mass industrial production of things like cars and trucks, airplanes, refrigerators, both private and commerical, radios and televisions, and pretty much the entire field of modern medicine from things like x-rays or antibiotics. And definitely places things like the Green Revolution, computer technology, and global trade facilitated through containerization completely beyond this time point.

dan1111 May 16, 2014 at 5:44 am

Some of the things you mention, like cars and X-Rays, actually were adopted widely in the early 20th century.

I do think your larger point is true, though. For example, if you look at life expectancy as a rough measure of medical advances, rapid progress continued at least until about 1950.

Adrian Ratnapala May 16, 2014 at 6:11 am

So “early 20th century” is a bad description for things which, in a subset of fields fields, continued through the first half of the 20th century?

dan1111 May 16, 2014 at 6:13 am

I just don’t think it’s clear that the mid to late 20th century was worse. Air travel, computers, etc. are a pretty big deal.

prior_approval May 16, 2014 at 11:27 am

‘like cars and X-Rays, actually were adopted widely in the early 20th century.’

Depends on what you mean by widely adopted. Here was the state of transcontinental road transportation in the U.S. in 1919 – ‘In the course of its journey, the convoy broke and repaired[21] wooden bridges[2]:10 (14 in Wyoming),[22] and “practically” all roadways were unpaved from Illinois through Nevada.[1]:4 The convoy travelled up to 32 mph (51 km/h),[10] and the schedule was for 18 mph (29 km/h)[23]:111 to average 15 mph (24 km/h).[24] The actual average for the 3,250 mi (5,230 km) covered in 573.5 hours[25] was 5.65 mph (9.09 km/h) over the 56 travel days for an average of 10.24 hours per travel day. Six rest days without convoy travel were at East Palestine, Ohio; Chicago Heights, Illinois; Denison, Iowa; North Platte, Nebraska; Laramie, Wyoming; and Carson City, Nevada. The shortest driving periods between control points were from Council Bluffs, Iowa, to Omaha, Nebraska (2 hrs for 5 mi) and Delphos, Ohio, to Fort Wayne, Indiana (6 hrs for 51 mi), while 4 days had average speeds over 9 mph (14 km/h): E Palestine OH to Wooster OH (9 hr for 83 mi), South Bend IN to Chicago Heights Il (8¾ hr for 80 mi), Jefferson IA to Denison IA (7½ hr for 68 mi), and Anderson’s Ranch NV to Ely NV (8 hr for 77 mi).’ http://en.wikipedia.org/wiki/1919_Motor_Transport_Corps_convoy

And of course, when one looks at the logistics of WWI, you don’t find much in the way of motor transport at all. Trains and horses, certainly. Vehicles with internal combustion motors? Not so much.

So Much For Subtlety May 16, 2014 at 6:23 am

In Western history for instance, our most significant period of growth was probably the late 19th through early 20th century, when the foundations for the modern world were laid, yet estimated growth rates for this period are not astonishingly high. We’re missing out on the values of the new goods, for one thing.

If I read that right, that is a technical comment on the way that we measure economic growth. This, on the other hand, isn’t:

This paper leverages variation in the likelihood of manipulation over two dimensions, arguing that politically sensitive data are more likely to be manipulated at politically sensitive times. … At the sub-national level in China, the difference between GDP and electricity growth increases in years with leadership turnover, consistent with juking the stats for political reasons.

That is a complaint that Chinese statistics are still carrying on a fine tradition from Communist days of lying their little hearts out. For political reasons.

Those are two very different issues. Look, I know people who have to work with data from China. Their rule is that every figure that comes from China, whether from the government, a business or a university, is fabricated. That is a very different problem. Do we know what Chinese economic growth is? Does the government? Do they care? In the end, it all seems to depend on a bunch of guys who make stuff up rather than take a break from their heavy duties reading the newspaper and drinking tea.

sansfoy May 16, 2014 at 7:07 am

It’s hard to tell from our point of view, but I wonder if a hundred years from now it will be obvious that the internet was the final disruptive piece of technology to be introduced. There doesn’t seem to be much coming down the pipeline. Self-driving cars, maybe, but that’s an iterative improvement. Maybe some big medical breakthrough, but that seems a lot like fusion energy, always 30 years in the future.

dan1111 May 16, 2014 at 7:41 am

Big technological breakthroughs are hard to foresee, kind of by definition.

mofo. May 16, 2014 at 9:29 am

Self-driving everything. Cars, planes, buses, trains, tractors, construction equipment. Also, what Dan1111 said.

dan1111 May 16, 2014 at 10:41 am

Also, don’t forget quadrocopters bringing us every single thing we need, so we don’t ever have to get off the couch again!

Larry Siegel May 16, 2014 at 5:41 pm

In the late 1800s a very prominent physicist told his PhD students they wouldn’t discover anything important because the universe was already fully understood, and that they should find another field. Hmmph.

Owen May 17, 2014 at 3:12 am

Have you forgotten about the ROBOTS TAKING OUR JOBS?

Brenton May 17, 2014 at 3:12 am

“doesn’t seem to be much” <<< try reading more about what's going on… you may be surprised! I'm fascinated and inspired by the technological breakthroughs currently occurring. Go to your library and read MIT Technology Review to start :) or just do a google or youtube search.

100 years from now aging might be cured. We might all have exoskeletons, like those currently giving paraplegics the ability to walk. Robots might do all of our household tasks. China may be paying couples to have genetically engineered genius children, while the general population receives genetic engineering to make themselves fitter, happier, and smarter. And in general, premature death and suffering may be greatly reduced worldwide compared to today.

Regardless of future technology, the miraculous economic growth occurring worldwide should be enough for people to feel very happy about the direction the world is going. The fact that millions of people each year leave behind extreme poverty and suffering should be enough to consider the world to be currently changing drastically.

jon May 16, 2014 at 8:03 am

China is certainly understating their growth. This way they don’t have to go the way of Japan, who gave in to USA pressure. If you as a country appear to strong, you suffer the consequences so you have to appear weak.

Cahokia May 16, 2014 at 8:38 am

“This way they don’t have to go the way of Japan, who gave in to USA pressure.”

Did Japan give in to the U.S.?

The Japanese auto industry is essentially as closed as ever. Consider for how many decades U.S. leaders have been proclaiming that Japan’s markets are really this time about to open and that Prime Minister X’s government is about to initiate painful reforms that will make their country more like America.

It’s funny, people talk about the downfall of Detroit on here in isolation from the fact that Japan and Korean automakers were able to dump cars in the U.S. that were subsidized by their highly protected auto markets (Korea’s somewhat less protected than Japan, yet closed by American standards).

dan1111 May 16, 2014 at 9:08 am

Most outward signs seem to indicate that China wants to appear strong (unless they are intentionally trying to look like a weaker country that is trying to look strong, which would be pure genius. Like my friend who once dressed up for Halloween as a different person who was dressed up for Halloween).

And, yeah, I agree with Cahokia. How about some evidence on Japan?

Nathan W May 16, 2014 at 1:18 pm

It may be difficult to pinpoint the source of inaccuracy due to opaque aggregation processes, even if a broader analysis can be used to identify that most likely there is manipulation/massaging of data which is most politically sensitive.

Brett May 17, 2014 at 3:12 am

Didn’t Xi Jinping just straight up tell a US official in one of the Wikileaks cables before he became the top leader in China that he himself didn’t use the GDP data, and relied on electricity and railroad cargo data instead as proxies?

Comments on this entry are closed.

Previous post:

Next post: