4.3 Price stickiness
As noted above, perfect price stickiness generates a mechanical link between invoice currency and exchange rate pass through. To give an indication of the degree of price stickiness in the New Zealand data, Table 4 reports the share of short-run unit value changes falling below an absolute value threshold of 0.1 percent.[17] This measure of stickiness is calculated in three currencies - producer, local, and vehicle (where applicable) - and is reported separately by invoice currency group. The key point to observe from the table is that price stickiness is a phenomenon observed primarily in the invoice currency (the bold values). From the first column, 7.5 percent of producer-currency priced unit values are unchanged across consecutive trades, when the change is measured in NZD, compared to 1.7 percent when measured in the currency of the importer. For trades invoiced in local currency, 5.5 percent of unit values are sticky when changes are measured in the local currency, but only 0.7 percent are sticky in NZD. Similarly, vehicle currency-invoiced unit values are sticky 7.0 percent of the time, when calculated in the vehicle currency, but two percent or less when measured in the producer or local currency.
Our subsequent approach to sticky prices differs from that observed in several recent papers (including Gopinath et al. 2010), in that we do not restrict analysis to pairs of unit values in which we observe a change in the unit value. This decision largely reflects the nature of the data we use - sticky unit values make up only a small proportion of observations, and exploratory estimates suggest little impact on estimated exchange rate impacts from including them. Meanwhile, identifying nominally sticky unit values is complicated by the use of aggregate monthly prices, as prices may differ across trades within, as well as across, months.
| Currency of calculation | Invoice currency | ||
|---|---|---|---|
| Producer | Local | Vehicle | |
| Producer | 0.075 | 0.007 | 0.007 |
| Local | 0.017 | 0.055 | 0.021 |
| Vehicle | N/A | N/A | 0.070 |
Unit value changes are calculated in the "currency ofcalculation" and are unadjusted for the gap between observedtrades (M). A threshold of 0.1 percent is used to define a stickyunit value.
Price stickiness, in the form of "take it or leave it" offers by trading partners (perhaps driven by currency movements), could induce entry or exit by New Zealand exporters. Alternatively, fixed NZD price offers from New Zealand exporters could result in variable foreign demand. If a firm does not trade, we do not observe whether a product changes price. This analysis makes no adjustment for attrition and compares unit values across consecutive trading months. That is, we show that where firms continue to trade, the price set in the invoice currency is an important anchor for short-run price changes, suggesting that it will be useful to control for invoice currency when considering estimates of ERPT.
Notes
- [17] To be consistent with other analyses of price stickiness, the observed changes inunit values are not normalised by M in this table. A threshold value approach isadopted because perfect unit value stickiness cannot be consistently observed across currencies due to rounding issues induced by currency conversion, and the division of valueby volume. The patterns described in Table 4 are maintained for other choices of threshold.
