The BLS’s Import/Export price release, from May 13th, might seem like old news. And some aspects are. But I think it is useful to think about what the trends in these price indices mean for general inflation and the adjustment process (this is in some sense an update on this post).
First, consider import prices.
Figure 1: Log dollar exchange rate (up is weaker) against broad basket of currencies (blue), log goods import price (red), log goods import ex-oil price (green), all normalized to 2002M02=0. Source: Federal Reserve Board via St. Louis Fed FREDII, BLS Export/Import price release (May 13), and author’s calculations.
Note that while total goods import prices (including commodities such as oil) have surged nearly 40% since 2002M02 — a remarkable 9/10 of the 45% depreciation of the broad trade weighted dollar over the corresponding period — the non-oil goods import price has only moved by a third of the dollar’s change. (This is why measured United States exchange rate pass-through is higher for prices including dollar-denominated commodities , ).
What does this mean? In particular, will these import price changes feed into a generalized change in the CPI? In a mechanical sense, they must since imported goods comprise some portion of the CPI. However, in an economic sense, they might not, to the extent that non-import prices fall.
Note that export prices have also risen. Figure 2 depicts the corresponding picture. Both the goods export and goods export ex-agriculture price indices have risen, accounting for between 4/5 and a half of the dollar’s depreciation. And while it is typical to focus on import prices as the key driver of domestic inflation, it is important to understand that as long as goods used in domestic consumption and exportables are to some degree substitutable, then rising export prices can also exert upward pressure on inflation. So while there is some appropriate focus on ethanol demand driving up food prices , to a certain extent the fact that US agricultural exports are demanded globally pulls upward domestic food prices.
Figure 2: Log dollar exchange rate (up is weaker) against broad basket of currencies (blue), log goods export price (red), log goods export ex-agricultural commodities price (green), all normalized to 2002M02=0. Source: Federal Reserve Board via St. Louis Fed FREDII, BLS Export/Import price release (May 13), and author’s calculations.
Returning to import prices, note that prices of imported goods from China have started rising substantially. Figure 3 depicts the time series for the USD/CNY nominal exchange rate and the price index for imports from China, normalized to 2005M06=0 (in logs). So, the long awaited adjustment of dollar prices to exchange rate changes appears to be underway.
Figure 3: Log USD/CNY exchange rate (up is weaker dollar) (blue), and log Chinese goods import price (red), all normalized to 2005M06=0. Source: Federal Reserve Board via St. Louis Fed FREDII, BLS Export/Import price release (May 13), and author’s calculations.
It is interesting that import prices kept on declining so long after the change in the Chinese exchange rate regime. The continued decline in the Chinese goods import price index suggests that productivity in Chinese export industries was increasing at such a substantial pace as to outweigh the combined effects of rising wages and appreciating yuan (although, without detailed knowledge on unit labor costs, one can’t be sure).
How to interpret the implications of these trends in tradable goods prices? First, it is important to observe that higher import and export prices need not necessarily lead to faster inflation. It could be that the exchange rate change and inflation are moving as a consequence of a common shock (namely monetary policy). In this interpretation, the correlation of exchange rate changes, import and export prices, and general price indices such as the CPI, are not structural. Estimation of pass-through equations with the CPI inflation rate on the left hand side then do not make sense.
On the other hand, if exchange rate changes are to some extent exogenous, then estimation of pass through equations might still make sense, especially to the extent that one can find valid instruments for those exchange rate changes.
I don’t want to focus on this particular issue (I’ve discussed this elsewhere ). Rather, I want to highlight the fact that the relative price of non-oil imports, as compared to the full basket in the CPI or the Core CPI, has essentially reversed trend.
Figure 4: Log ex-oil goods imports price relative to CPI (blue), and relative to Core CPI (red). Source: Federal Reserve Board via St. Louis Fed FREDII, BLS Export/Import price release (May 13), and author’s calculations.
In Figure 4, one will note that from 1990 to the beginning of 2008, the price of non-oil goods imports was declining relative to the total bundle in the CPI. Relative to the Core CPI, the price was declining until about the beginning of 2002. The relative price trend was essentially flat until the beginning of 2008.
The longer horizon movements are shown in Figure 5, which depicts the NIPA measure of non-oil goods imports prices relative to the GDP deflator. While the slope of the trend is somewhat different over the 1990-2008 period, the same essential result shows up — that recent months have a seen a radical break in the relative price.
Figure 5: Log ex-oil goods import price relative to GDP deflator (blue). NBER defined recessions shaded gray. Source: BEA GDP release of April 30, 2008, NBER, and author’s calculations.
What remains to be seen is whether this recent trend is extended, so that something like the mid-1980’s episode is replicated; at that time, the non-oil net exports deficit shrank by about 2 percentage points of GDP before the recession of 1990 set in.