The issue concerns an OECD report on China’s R&D spending and the way that organization calculated the RMB exchange rate. In other words, a classic statistical argument. Disclaimer: I am not an economist or statistician, but I will give this a shot and try to keep the econ errors to a minimum.
WAIT – don’t go away. I promise there is an interesting part to this, sort of.
OK, some background. My original post on this is here. I received a comment from China Law Blog that there was actually some controversy on the way that OECD got its numbers. I have read the criticism and have two comments. One, the OECD did not "hype" the numbers, but quite a few journalists did not tell the whole story (I was intellectually lazy too, I suppose). Two, the basic point of my post is still valid, albeit slightly blunted.
Warning: some econ ahead. Those of you with frail constitutions should log off for a while.
The OECD report stated that China’s R&D spending would surge ahead of Japan this year ($136 billion as compared with $130 billion). Simple story so far. The details are a bit muddy, however. China’s 2005 spending was $30 billion, which would obviously make that $136 billion a bit of a jump, to say the least. Whenever you see a large statistical leap like this with these sorts of statistics, you can bet your bottom that there has been some sort of change in methodology. This happens quite frequently, although it can be jarring. When it involves a complete change in methodology of a statistic that is published on a regular basis (e.g. an unemployment figure or a GDP amount), then we get a series of debates from statisticians and economists, some of whom like the old way and others who support the new methodology.
In this case, it is not a change in methodology, but just the way that the authors of the report presented their findings. This is perfectly reasonable, and I do not therefore think that the OECD was irresponsible here.
So why are the 2005 numbers (from the Chinese government) so different from the 2006 OECD numbers? The answer is the exchange rate calculation. Now this is a hot topic in many circles, not just with R&D guys. The easiest way to do this kind of thing is just take the aggregate numbers in RMB and calculate a dollar amount based on the current nominal exchange rate. This is where the 2005 numbers come from, I would guess.
Apparently the OECD felt that using the current quasi-fixed value of the RMB (many economists feel the RMB is quite undervalued) does not yield a realistic number for total R&D spending. Fair enough. However, how does one go about determining what is a more accurate value of the RMB? The experts do not agree, unfortunately, with economists suggesting that the RMB is undervalued anywhere from 10 to 40 percent. A good read on the subject is Fred Bergsten’s testimony to the U.S. Senate Finance Committee last March. Bergsten, an international econ guru at the IIE, says 20 – 40 percent is a good range to consider. Read Nicholas Lardy, Morris Goldstein, Albert Keidel, etc. and so on. Everyone agrees that the RMB is undervalued, but there is no exact agreed-upon figure.
Back to the OECD report. So if the RMB is undervalued, then that would mean the 2005 figure of $30 billion is probably quite understated. How would one correct for this? One way would be to restate the value given the range discussed above. The $30 billion then becomes something like $33 to $42 billion, still quite far away from that $136 billion figure.
What the OECD did was use something called Purchasing Power Parity. If you remember back to Macroeconomics 101, this is a concept that essentially says that a nominal exchange rate is in equilibrium if you can buy a widget in either country A or country B with the same amount of money (using that exchange rate).Your purchasing power in both countries would be identical. This is what exchange rates are supposed to reflect in a perfect world, but of course the world is not perfect and exchange rates are also influenced by financial flows, speculation, interest rates, etc. PPP is based on the price of a representative basket of goods in each country to determine a differential that can be used like an exchange rate to make comparisons. The subject often comes up when comparing per capita GDP between countries, which is the most common measure used by growth economists.
The OECD thought that PPP was a better way to go than to use the nominal exchange rate, presumably for two reasons: one is that PPP might be a better comparison method, and two, China maintains a fixed exchange rate. Lots of possible distortions here. Please also note that the OECD does a heck of a lot of work publishing international stasticial data using PPP – they are not just pulling these numbers out of thin air.
I’m not sure if all this makes the story more or less murky, but using PPP, the OECD restated China’s R&D spending and ended up with $136 billion. This represents an almost four-fold increase in the reported dollar amount. This was criticized by some academics and journalists (see articles in Businessweek and SciDevNet for background). Certainly the way this was reported in the press was, if not misleading, not an accurate presentation of the entire issue.
I am left with two issues. The first is whether this PPP figure should be taken seriously. Well, I am not entirely convinced. I do agree that the current value of the RMB is way too low, but I have never supported the 40% crowd. But is the purchasing power of the RMB that far out of whack with even the admittedly low nominal exchange rate? This is inherently difficult to assess, since prices in China vary widely by region. Just take a look at the price differential between Shanghai and Xinjiang as an example, and you can see what I mean. Price surveys in China, as with many other statistical surveys here, are notoriously unreliable and incomplete, so that gives me pause as well. That’s a long-winded way of saying that I am not sure. On the other hand, all those headlines, including the one that I included in my post, were inherently "sexed up", to use a memorable phrase from the UK Iraq War debate. Mea culpa.
The second issue is the big picture item. Without the hyperbole, what is the underlying trend here? China does have a 20% growth rate in R&D spending and is training/graduating and employing a huge number of technical experts and researchers. This is significant, particularly when you hold this up against a lot of other policies implemented and emphasized by Beijing in recent years on promotion of home-grown innovation. As I have said in earlier posts, China is consciously trying to transition from a manufacturing-based economy to an innovation and knowledge-based economy, and they are doing it fairly quickly. Whether the dollars spent should be seen as $36 billion or $136 billion or, probably more accurately, somewhere in between, the trend is the important thing.