RDP 8713: Estimating the Inflationary Effects of Depreciation Appendix 2: International Import Price Experience

This Appendix looks more closely at international import price experience. The table below (from OECD and country sources) shows movements in import prices and effective exchange rates for 16 OECD countries. As can be seen, there was not a single country to experience rising import prices in their own currency in the period in question. This was despite the fact that a number of the countries experienced substantial depreciations over the period.

In an attempt to quantify the various factors (including currency movements) influencing import prices over the period, a simple cross-section regression was conducted. The estimated equation was as follows:

View MathML

where

The second exchange rate term is included to capture possible lagged effects from exchange rates changes. The domestic price variable is included to capture the possibility that foreign suppliers' pricing behaviour is partly determined by the objective of market share preservation. The oil price variable is included to capture direct oil price effects, since the import price data shown above are for all imports, including petroleum imports. The expected signs of α₂, α₃ and α₄ are positive. The parameter α₅ should be equal to the percentage change in oil prices over the period. Finally, α₁ should represent that movement in import prices which cannot be explained by exchange rate changes, differential domestic inflation rates or direct oil price effects. It should therefore be the underlying movement in the price of non-oil imports over the period.

In the initial estimation, the first exchange rate term was statistically very significant and showed the expected sign. The second exchange rate term and the domestic inflation show the expected sign but were not statistically significant. The oil price effect was highly significant, though somewhat larger (−0.8) than one might expect given the fall in oil prices over the period of around 50 per cent.

In the preferred equation, the oil price effect was constrained to −0.5 (the approximate fall in oil prices over the period) and the second exchange rate and domestic inflation terms were omitted. The equation is shown below.

Because heteroskedasticity is often a problem in cross-sectional regressions, the Breusch-Pagan test was employed and could not reject the hypothesis of uniform variance of the errors. The results do not seen sensitive to the exclusion of either Japan (which shows the largest fall in import prices) or the United States (which would be most subject to “large country” effects).

The constant term in the above equation can be interpreted as an estimate of the average movement in non-oil import prices, apart from exchange rate effects over the period. That is, the “underlying” movement in non-oil import prices for these 16 countries is estimated to have been significantly negative. According to these estimates, the underlying movement was a fall of around 7 per cent.

Footnote

The exchange rates here are defined as units of domestic currency per unit of foreign currency. [1]