T C A Srinivasa-Raghavan: Inflation with Chinese characteristics

OKONOMOS

Why is such high growth in China not accompanied by rising inflation?

A couple of days before the latest monetary policy tweaking by the RBI was due, a colleague testily asked how China could grow at 10 per cent and have inflation of only 2-3 per cent.

One obvious answer was the old text-book one: high levels imports to keep supply and demand in balance. This is possible on a scale that is impossible in India because in China the government doesn't have to face elections. This means that it is unbothered if cheap imports hurt local producers.

Another colleague said it was because rural productivity was very high. He may be right, but my own view is the unfashionable one, namely, that China fiddles its numbers in the way it collects the inflation data. Rural prices enter the index tangentially or even asymptotically.

Be that as it may, a quick browse of the Internet revealed two very useful papers. The first* is by three economists from the Bank of Japan (BoJ), Ryota Kojima, Shinya Nakamura and Shinsuke Ohyama. Basically, say the three, taking the underlying dodginess of Chinese data as a given, when commodity prices start to rise, Chinese employers simply refuse to raise wages. After all, they don't have to worry about trade unions nor the government about opposition protest in parliament. This is what really keeps inflation low.

For the benefit of the Indian communists, let me quote in full: "Before the systematic reform of state-owned enterprises began in 1997, the real wage exceeded the marginal productivity of labour and led to higher inflation. By contrast, the real wage has been below the marginal productivity of labour since 1997. This means that wages have put downward pressure on inflation and upward pressure on wage growth... wages are controlled by two regulations... that the growth rate of total wages must be lower than that of after-tax productivity... and that per capita wage growth must be lower than the growth rate of labour productivity."

Lack of space prevents me from going into the fascinating details. Suffice it to say that everyone who is interested in the problem should read the paper, especially Mr Quixote and Mr Panza of the CPM.

As for monetary policy, this is what they say, "... quantitative monetary control, through administrative restrictions on bank lending... prevents overheating." The lesson here is for those who scream when the RBI does the same thing.

The second paper ** is by Stefan Gerlach and Wensheng Peng of the Bank for International Settlements. It looks at the relationship between inflation and the output gap. They reach broadly the same conclusion as the BoJ economists, namely, that the data is dodgy and that "a simple application of the Phillips curve does not fit data well and that this may reflect an omission of some important variables."

To allow for this, they have constructed a modified Philips Curve, one, I daresay, with Chinese characteristics. Since so much of the data is unreliable, they say "it is difficult to capture or measure the influence of these forces empirically, we model them as an unobserved variable." Unobserved? What a nice new term. So, "once a serially correlated omitted variable is allowed for, the model fits the data much better."

And what does that mean? It means that prices go up when aggregate demand goes up because of the output gap but "the omitted variable also explains what at first glance appear to be surprising features of the simple Phillips curve estimates."

The short point: high growth shows itself in higher aggregate demand and rising inflation. But China simply uses administrative measures to keep the overall inflation rate down. Remember, China is not a normal country and the rules it imposes on the economy are not normal either.

*Inflation Dynamics in China, No 05-E-9, July 2005, www.boj.or.jp/ en/type/ronbun/ ron/wps/data/ wp05e09.pdf

** Output gaps and inflation in Mainland China , BIS Working Papers, No 194, February 2006, www.bis.org/publ/ work194.htm