ANNEX III.1: Data Description
Brash, Donald T., “The New Zealand Dollar and the Recent Business Cycle,” address to the Canterbury Employers’ Chamber of Commerce, January 29, Reserve Bank of New Zealand Bulletin, 1999, Vol. 62, No. 1.
Burstein, A. Eichenbaum, M. and Rebelo, S., 2002, “Why are Rates of Inflation so Low After Large Devaluations?, unpublished manuscript.
Dwyer, Jacqueline and Leong, Kenneth, 2001, “Changes in the Determinants of Inflation in Australia,” Research Discussion Paper 02, Reserve Bank of Australia, May.
De Brouwer, Gordon and Ericsson, Neil, 1995, “Modeling Inflation in Australia,” International Finance Discussion Papers, No. 530, Board of Governors of the Federal Reserve System, November.
King, Alan and Douglas Steel, 1998, “Import Prices and Exchange Rate Pass-Through in New Zealand” Economic Discussion Papers, No. 9805, University of Otago, New Zealand.
McCarthy, Jonathan, 2000, “Pass-Through of Exchange Rates and Import Prices to Domestic Inflation in Some Industrialized Economies” Federal Reserve Bank of New York, Working Paper.
Orr. A., Alasdair Scott, and Bruce White, 1998, “The Exchange Rate and Inflation Targeting,” Reserve Bank of New Zealand, Bulletin, September, Vol. 61 No. 3, pp. 183–191.
Svensson, L., 2001, “Independent Review of the Operation of Monetary Policy in New Zealand: Report to the Minister of Finance,” Reserve Bank of New Zealand, February.
Prepared by Martin Cerisola (x38314) who is available to answer questions.
For example, the Reserve Bank of New Zealand (RBNZ, 2000) notes that “…while the first stage of pass-through (from the exchange rate to import prices) was very much dampened between 1994 and 1999, it seems to have returned in the third and fourth quarter of 2000…At the second stage of pass-through (from import prices to consumer prices), the pass-through relationship is even less clear….”
The sharp rise in import penetration experienced in New Zealand, Australia, and Canada, would suggest that the direct effect of the changes in the exchange rate to consumer prices should have increased proportionately with the share of imported goods in the consumer price index. For example, see the Bank of Canada, Monetary Policy Report, Technical Box 2, November 2000. However, increased import penetration may result in more competition vis-à-vis domestic producers, limiting their ability to pass higher costs onto prices.
Alternatively, proxying exchange rate volatility as deviations in the NEER from a trend estimated by using the Hodrick-Prescott filter shows that volatility rose sharply after 1991, and even rose further after 1994, only to decline from 1999 on. However, these deviations should be interpreted with caution as they do not appear to be stationary.
Arora and Jeanne (2000) note that the Canadian dollar has fluctuated by less against the U.S. dollar than have several other floating currencies, such as those of Australia and New Zealand. The low relative volatility seems to be consistent with the behavior of official interest rates and foreign reserves (an indicator of market intervention), both of which have fluctuated more in Canada.
Most of the variables are nonstationary, and the Augmented Dickey Fuller tests for stationarity suggest that most of them are difference stationary, with the exception of unit labor costs. Nevertheless, the VAR models were specified and estimated in first differences given that the models in levels did not meet the stability conditions specified by Lutkepohl (1991). While this specification may entail a potential significant loss of information on the long run relationships, it may preserve the efficiency of the estimates. However, differencing the data would imply imposing the restriction that all shocks tend to be permanent, which may not necessarily be the case for some of the variables used in the VAR. Moreover, the reduced form residuals from the VAR were orthogonalized using a Choleski decomposition to identify the shocks. Since the order of the variables affect the results, the order chosen was as follows: output gap, unit labor costs, foreign inflation, exchange rate, and prices. The variance decomposition results were relatively robust to different orderings in these variables. See Annex III.I for a detailed description of data.
The results also suggest that the estimated first stage short-term pass-through to-import price inflation does not differ when long-run PPP is not imposed in the model.
In addition, the increased reliance by firms on foreign exchange risk hedging may have also contributed to reduce the degree of first-stage exchange rate pass-through.
The estimated pass-through coefficient declined from about 0.65 (0.76 excluding oil) in early 1991 to 0.5 (0.43) in early 1999, and has risen to about 0.6 (0.52) in the third quarter of 2001.
In doing so, the estimated coefficient for the second-stage pass-through is not strictly comparable to more traditional estimates of the pass-through, which are based on the impact of import prices on consumer price inflation.
This model was extended and estimated to assess the long-term impact on consumer prices of terms of trade and of the upward trend in import penetration, and whether these variables affected the estimated exchange rate pass-through. Terms of trade shocks influence the intratemporal substitution of consumption between tradable and nontradable goods, as well as private savings and the current account. Cashin and McDermott (2001) find that terms of trade shocks induce large and significant intratemporal and intertemporal substitution effects in Australia, New Zealand, Canada, United Kingdom, and the United States. In addition, a higher share of imported goods in the CPI would tend to increase the direct effects of exchange rate changes on the CPI. Including proxies for both variables in the model support the hypothesis that the increased share of imports in the basket of goods comprising the consumer price index has positively affected the exchange rate pass-through in the long-term, but these proxies do not appear to significantly affect the short-term exchange rate pass-through. When including these proxies, the estimation is subject to multicollinearity with other long-term determinants, such as unit labor costs, making it difficult to interpret the long-run relationship between these determinants and the price level.
The estimation for Australia is based on the consumer price index and differs from that in Dwyer and Leong (2001), since theirs is based on the price index which excludes volatile components.
In the case of New Zealand, the estimated speed of adjustment does not appear to be statistically significant, suggesting the lack of a long-term relationship between these variables. However, this result does not appear to be robust to a decomposition of consumer prices into tradable and nontradable good prices, which suggests the existence of a long-term relationship.
In the case of Australia, when excluding the volatile components of the consumer price index, the exchange rate pass-through was significantly more stable during the 1990s.
The impulse response function traces the extent to which adjustment to a 1 percent shock to the exchange rate is estimated to have occurred by the period shown. For example, each point in time traces the effects of a shock that took place in the past four and six quarters.
The sensitivity of the long-term elasticity to unit labor costs may be influenced by the inclusion of the output gap. Also, the specification based on core inflation results in a more reasonable long-run elasticity for unit labor costs in Canada.
The countries included were Finland, Sweden, Mexico, Brazil, Korea, Indonesia, Philippines, Thailand, and Malaysia.
Therefore, the prices for these local goods tend to be determined mainly by domestic factors and may be largely immune to changes in the exchange rate.
The results for Canada are not presented as models for tradable (goods) and non-tradable (services) consumer price inflation were highly rejected.