Around the turn of the decade, low inflationary pressures in many emerging market and industrial countries prompted some to argue that China was bringing down prices in the rest of the world by exporting cheap goods to more mature markets. More recently, as inflationary pressures have started to build up around the globe, these arguments have been turned on their head, with some asserting that demand from China is now pushing up prices in the rest of the world. What’s going on? A new IMF Working Paper looks at the empirical evidence and finds some intriguing, and unexpected, linkages.
How can a country export inflation or deflation? After all, isn’t inflation in any country a monetary phenomenon determined by domestic policies? The short answer is that it is theoretically possible for countries to export and import inflation or deflation, and there is empirical evidence that they do so. In fact, when there was a global system of fixed exchange rates (such as during the gold standard era), it was not uncommon to see periods of synchronized deflation in a number of countries, as falling prices in the reserve country led to falling prices in others. And, more generally, for any relatively small open economy with a pegged exchange rate, domestic inflation will tend to be determined to a large extent by inflation in its trading partners.
There is no strong evidence that inflation in China caused inflation in the United States or Japan, but there is some evidence of the reverse.
This phenomenon of exporting inflation or deflation can also occur for a country with a flexible exchange rate regime, provided the exporting country is large. For example, deflationary pressures could be exported directly when exports of final goods, whose prices in the domestic currency of the exporter are falling, directly lower the import price indexes of other countries. Deflation could also be exported indirectly if cheaper final goods lower prices in other countries because they adversely affect demand in those countries when producers lose markets and profits. Also, cheaper exports to third countries could affect the cost structure of enterprises in those countries, enabling them to reduce the prices of their exports. Similarly, inflation brought on by demand pressures in a country could increase imports from its trading partners, reducing domestic supply and causing inflation in the exporting economies.
A number of these arguments have been used to claim that China has exported deflation or inflation. Some said that during the boom-bust cycle of the early 1990s, China built huge excess capacity in its manufacturing sector that later caused the price of manufactured goods to decline, causing deflation in China. With China’s share in global trade increasing rapidly, this deflation spread to the rest of the world through cheap Chinese exports. As deflation in China ended in 2003 and China’s import demand for various goods surged, the arguments changed somewhat. Several claimed that China was, at this time, exporting inflation because it was sucking in goods at such high rates that consumers in other countries had to pay higher prices.
What the data show
To examine the validity of these claims, the Working Paper uses a number of progressively more complicated econometric models to investigate what happened. As a first step, a simple autoregressive model is used to run the Granger causality test. This model and test impose minimal economic assumptions on the estimates and potentially capture both direct and indirect effects of inflation in one country on inflation in another.
When applied to China, Japan, and the United States, the results are intriguing: there is no strong evidence that inflation in China caused inflation in the United States or Japan, but there is some evidence of the reverse. Although these reverse effects are not emphasized in the media, they should not be surprising. China’s imports from the United States have been growing strongly for more than a decade. This finding is also consistent with the notion that in regimes with limited flexibility, inflation tends to be transmitted from the reserve-currency country (the United States) to other countries (such as China).
For the next step, the Working Paper uses a more sophisticated model to capture a broader picture of these economies and to explain how inflation is transmitted. Specifically, linkages among consumer price inflation, commodity and import prices, the exchange rate, and output growth are estimated jointly in a vector autoregression setting. This allows supply and demand shocks in China to have contemporaneous effects on inflation in the United States through import prices. To circumvent a typical weakness of such models—the ordering of the variables affects the results—two separate models are estimated, one of them biased toward finding a positive effect of inflation in China on inflation in other countries.
The results of this model confirm the earlier results. There is no significant effect of inflation in China on U.S. or Japanese inflation. In fact, whereas inflation in the United States responds positively to higher import prices and higher output growth, and its import prices respond positively to increases in world commodity prices, these import prices do not respond to inflation in China. Similarly, inflation in China explains only about 5 percent of the variation in U.S. import and consumer prices.
There do appear to be stronger and possibly growing sector-specific linkages between prices in China and those in the United States and Japan.
On the other hand, there is some evidence that U.S. economic activity affects China: shocks to output in the United States are found to affect inflation in China, and variance decomposition suggests that more than 10 percent of the variation in Chinese inflation is explained by changes in U.S. output growth. The results for Japan are broadly similar.
These models have one potentially important drawback, however. They may underestimate the current effect of inflation in China on inflation in the United States and Japan because the data span two decades, including the period when trade between China and these two countries was relatively small. To address this potential problem, the Working Paper also used a variable coefficient model, which allows the coefficient that captures the effects of Chinese inflation on other countries to change across time as the trade volume between these countries changes. An increase in this coefficient over time, and as trade volumes grow, would be expected, reflecting strengthening links between these countries’ inflation rates.
Contrary to expectations, however, taking into account the increasing trade between China and its trading partners does not change the conclusions of the simpler models. There are only a few instances of inflation in China having statistically significant effects on the United States and Japan, and they are short-lived (see chart). Again, inflation in the United States appears to have a stronger impact on inflation in China.
There is no clear evidence to date that Chinese inflation is influencing inflation rates elsewhere.
Data: IMF, World Economic Outlook.
What’s going on
Why is there scant evidence of a growing link between inflation rates as China’s trade volume increases? One possibility is that the effects of the various subcomponents of this relationship are canceling each other out. For example, over 2003-04, Chinese prices of household appliances declined, possibly depressing global prices for these goods, while food prices in China—which were affected by drought as well as by domestic policies—increased.
To test this hypothesis, a stochastic coefficient model was run using the components of the consumer price inflation (CPI). One component examined is food inflation, and the other—an important export category for China—is household furnishings, including appliances. The results suggest that changes in the prices of food and household furnishings in China do indeed explain a part of the corresponding U.S. prices. The results on household furnishings are less clear, possibly owing to the shorter sample period. Nevertheless, they do point to an increase in the link between price changes in China and the United States. These results also provide some support for the hypothesis that components of the CPI have been negating each other. Specifically, in 2003, as food prices pushed up the CPI, prices of household furnishings worked in the opposite direction. Results for Japan are again broadly similar.
In sum, there is only limited evidence at the aggregate level that inflation in China has led to price changes in the United States and Japan. And there is some evidence that U.S. inflation has had an impact on Chinese inflation—consistent with the literature that argues that inflation is propagated from the reserve-currency economy to other economies. In either case, the effect is short-lived. At a more disaggregated level, though, there do appear to be stronger and possibly growing sector-specific linkages between prices in China and those in the United States and Japan.
More broadly, however, it is other factors—notably common global shocks (for example, oil price developments) and the similar behavior of central banks (in particular, their shared commitment to keep inflation low)—that offer more likely explanations for similarities in inflation patterns among the world’s major economies.
IMF Asia and Pacific Department
Copies of IMF Working Paper No. 06/36, “Does Inflation in China Affect the United States and Japan?” by Tarhan Feyzioglu, are available for $15.00 each; see page 224 for ordering information. The full text is also available on the IMF’s website (www.imf.org).