Information about Europe Europa
Chapter

VII External Trade and the Aftermath of the 1992 Devaluation

Author(s):
Desmond Lachman, Ramana Ramaswamy, J. Green, Robert Hagemann, and Adam Bennett
Published Date:
October 1995
Share
  • ShareShare
Information about Europe Europa
Show Summary Details
Author(s)
Adam Bennett

The 1992 devaluation of the Swedish krona fundamentally altered the trading circumstances of the Swedish economy. It did so in much the same manner as did the devaluations of other European economies following the exchange rate mechanism crisis in September 1992. Since 1992, Sweden, like Italy, Spain, and the United Kingdom, has experienced a substantial increase in export volumes. Moreover, as in most of these other countries experiencing large devaluations, the pass-through of the Swedish devaluation to domestic prices has been relatively muted, in reflection of the large degree of excess capacity characterizing the economy in the depth of its recession. However, the distinguishing feature of Sweden’s recent trade performance has been the surprisingly large increase in import volumes despite the devaluation.

This section examines in more detail the various aspects of the Swedish devaluation since November 1992. First, it describes the impact of the devaluation on competitiveness and on export and import performance, including a comparison with the Swedish experience in 1981-82 and a comparison with the recent experience of Italy, Spain, and the United Kingdom. Second, it investigates empirically Sweden’s recent trade performance and the pass-through of the devaluation on import and export prices in local currency. A final section provides a brief summary of the conclusions.

Overview of External Trade Developments

Exports and Export Competitiveness

The abandonment of the fixed exchange rate in November 1992 resulted in a devaluation of the krona of some 25 percent in effective terms. Because of the modest pass-through from the exchange rate adjustment to export prices in domestic currency, the price competitiveness of exports improved markedly (Chart 7-1). Moreover, labor shedding and subdued wage pressure, combined with such rises in export prices as occurred, contributed to a large increase in export profitability and cost competitiveness.1 This improvement in competitiveness has led to a major recovery in Sweden’s share of foreign markets. Thus, despite the stagnation in the size of foreign markets, the volume of Swedish exports of goods rose by 9 percent in 1993 after several years of negative or negligible growth (Table 7-1). More dramatically, continued growth in market share has combined with a recovery in activity in overseas markets to produce a growth of 18 percent in industrial exports in the first half of 1994 relative to a year earlier.

Table 7-1.Exports by Commodity(Percent change)
VolumeUnit Price
First half of 1994First half of 1994
Value (In millions of kronor) 199219921993Change on year earlier1Change on previous period219921993Change on year earlier1Change on previous period2
Total exports326.01.29.017.86.9-3.29.23.52.2
Agriculture, forestry, and fish products1.9-25.1-44. 5118.822.111.227.81.1-5.7
Mineral products4.4-0.88.4-10.7-13.11.48.71.90.6
Of which: iron ore2.711.81.31.98.6
Industrial production, excluding ships315.31.38.817.97.7-3.58.83.62.2
Food5.6-6.623.822.77.63.27.15.22.6
Wood products11.120.221.77.61.2-9.45.021.010.2
Pulp8.40.33.1-3.4-9.6-9.10.67.411.1
Paper and cotton28.92.12.27.85.9-6.85.40.60.1
Petroleum products9.314.14.916.3-4.3-16.825.3-7.1-8.6
Iron and steel15.9-0.29.325.815.5-4.710.55.01.4
Nonferrous metals5.7-5.96.512.43.3-4.814.04.33.0
Metal products, excluding ships162.1-1.47.924.011.0-1.28.72.71.9
Other68.32.712.713.05.2-2.27.92.70.4
Electricity1.146.3-22.2-7.4-1.4-19.313.848.043.1
Ships3.342.9122.814.3-21.5-2.03.64.14.5
Memorandum items:
Manufactured goods275.30.18.619.49.8-2.38.32.91.2
Raw materials47.44.86.111.7-1.7-7.413.05.75.2
Source: Konjunkturinstitutet, Analysunderlag.

Seasonally unadjusted, change on first half of 1992.

Seasonally adjusted, change on second half of 1993.

Source: Konjunkturinstitutet, Analysunderlag.

Seasonally unadjusted, change on first half of 1992.

Seasonally adjusted, change on second half of 1993.

Chart 7-1.Export Competitiveness, Demand, and Capacity

Sources: International Monetary Fund, International Financial Statistics; Konjunkturinstitutet; and IMF staff estimates.

1Nominal effective exchange rate; cost competitiveness defined as relative normalized unit labor costs; price competitiveness defined as the ratio of Sweden’s export price to the average export price of industrial countries; 1992: Q3 = 100. An increase in the exchange rate index denotes a depreciation.

2Average percentage capacity utilization of reporting firms.

3Manufactured exports; OECD demand measured by total imports of Sweden’s 14 largest OECD trading partners; indices rebased second half of 1992 = 100.

4Constant 1985 prices, in billions of kronor.

The greatest improvement in Sweden’s market share was in the smaller, non-European OECD countries and in the non-OECD area, suggesting that the beneficial effect of the devaluation was felt primarily in improvements in competitiveness in third markets (Table 7-2). The Far East figures prominently in Sweden’s recent strong export performance, with rapid growth in exports to Japan and China. The rise in the market share of Swedish exports was broadly equal in the smaller OECD countries and the non-OECD area, but the higher rate of growth of the non-OECD countries combined with the improvement in market share to generate an especially rapid recovery.

Table 7-2.Exports by Destination1(Percent change)
19891990199119921993First Half 1994
Total
Market growth6.94.81.63.8-0.2
Change in market share-4.8-4.0-4.4-4.18.6
Swedish export volume growth1.80.6-2.9-0.58.4
OECD-142
Market growth6.04.80.52.7-1.86.8
Change in market share-4.8-3.7-5.1-2.06.410.4
Swedish export volume growth0.90.9-4.60.64.517.9
Other OECD
Market growth16.57.83.68.11.2
Change in market share-4.5-6.1-9.9-11.513.0
Swedish export volume growth17.31.2-5.8-4.314.4
Non-OECD
Market growth7.73.410.08.17.8
Change in market share-4.5-4.4-1.3-11.918.3
Swedish export volume growth2.8-1.28.6-4.827.5
Source: Konjunkturinstitutet, Analysunderlag.

Change on previous year.

Sweden’s 14 largest OECD trading partners are Austria, Belgium, Canada, Denmark, Finland, France, Germany, Italy, Japan, the Netherlands, Norway, Switzerland, the United Kingdom, and the United States.

Source: Konjunkturinstitutet, Analysunderlag.

Change on previous year.

Sweden’s 14 largest OECD trading partners are Austria, Belgium, Canada, Denmark, Finland, France, Germany, Italy, Japan, the Netherlands, Norway, Switzerland, the United Kingdom, and the United States.

In 1993, the largest pass-through to prices from the devaluation occurred for commodities such as agricultural products, petroleum products, iron, steel, and nonferrous metals, where prices are strongly influenced by related prices in world markets. The response of export volumes to the devaluation for these goods was generally rather muted. The largest gains for exports were recorded in food and wood products, engineering goods, and other manufactured goods, particularly pharmaceuticals.

The growth in exports has led to a significant recovery in the rate of capacity utilization of Swedish industry. In the last quarter of 1991, capacity utilization was 80 percent. By the second quarter of 1994, this had risen to 88 percent, significantly narrowing the margin with the peak rate of utilization of 90 percent reached during the last boom. As capacity utilization increases, it is possible that export prices might begin to rise more sharply over the remainder of the year, and export growth could slow. However, planned investment in the key export industries should result in expanded capacity, potentially de-laying the slowdown in exports. It is notable that the sharpest increases in capacity utilization have been in industries with more differentiated products, such as wood products, metal products, and engineering, in comparison with industries whose products are more homogeneous on the world market, such as wood pulp, chemicals, and petroleum products.

Imports and Import-Weighted Domestic Demand

Reflecting the decline in import-weighted domestic demand during 1989-92, imports fell during the recession, although there was evidence of continued market penetration of foreign goods. This market penetration no doubt owed much to the relative price competitiveness of imports prior to the devaluation, but other factors appear to have been at work too. Following the devaluation, prices of manufactured imports rose by about 9 percent in 1993, appreciably less than the depreciation of the exchange rate. Broadly speaking, the strongest rises in domestic prices for imports were recorded for raw materials, where the pass-through was more or less full, by contrast to semimanufactured and manufactured goods where the pass-through was much more modest (Table 7-3).

Table 7-3.Imports by Commodity(Percent change)
VolumeUnit Price
First half of 1994First half of 1994
Value (In millions of kronor) 199219921993Change on year earlier1Change on previous period219921993Change on year earlier1Change on previous period2
Total imports290.50.13.4-13.57.6-3.710.93.41.1
Agriculture, forestry, and fish products9.85.0-8.80.45.9-6.817.812.212.5
Mineral products16.52.16.17.02.1-9.516.9-9.4-7.1
Of which: iron ore13.63.16.37.52.8-10.118.8-11.8-7.0
Industrial production, excluding ships260.90.34.114.27.7-2.810.43.30.6
Food16.17.13.56.24.40.415.86.84.4
Wood, pulp, and paper2.82.71.733.819.2-5.65.35.0
Petroleum products10.410.1-6.520.714.0-12.523.7-4.3-4.5
Iron and steel6.58.912.737.415.7-6.18.611.03.4
Nonferrous metals5.4-7.17.112.45.5-5.814.84.33.2
Metal products, excluding ships127.6-2.35.617.18.9-1.79.52.6-0.4
Other92.70.73.77.94.3-2.28.24.60.8
Electricity0.642.1-9.9-13.5-23.0-22.7-18.878.158.4
Ships2.5-44.3-48.0189.4180.6-4.713.39.3-3.4
Memorandum items:
Manufactured goods228.4-0.85.114.27.4-2.18.93.60.1
Raw materials60.04.80.69.45.6-7.116.90.73.1
Source: Konjunkturinstitutet, Analysunderlag.

Seasonally unadjusted, change on first half of 1993.

Seasonally adjusted, change on second half of 1993.

Source: Konjunkturinstitutet, Analysunderlag.

Seasonally unadjusted, change on first half of 1993.

Seasonally adjusted, change on second half of 1993.

Importers of manufactured goods chose to preserve market share at the expense of their profits (Chart 7-2). Nonetheless, there was some erosion of import competitiveness. For this reason, it was surprising that manufactured imports rose by 5 percent in volume terms in 1993, despite the fact that overall import-related demand in total demand continued to fall. The only component of import-weighted demand that increased during 1993 was related to exports: this normally generates derived demand for imports of semimanufactured goods equal to one third of the export value. It is possible that since 1992 the derived demand for imports from exports has been higher than usual because of closures of Swedish subcontractors resulting from the low level of demand from the export industry and the severity of the recession. Certainly, the strongest growth recorded from imports was in components and other manufactured inputs to production. Imports of consumer goods were relatively flat.

Chart 7-2.Import Competitiveness, Demand, and End-Use

Sources: International Monetary Fund, International Financial Statistics; Riksbank; Konjunkturinstitutet; and IMF staff estimates.

1Nominal effective exchange rate (domestic value of foreign exchange); relative price of imports defined as the manufactured import deflator divided by the GDP deflator; relative profitability defined as the ratio of import price to index of industrial country export prices; 1992:Q3 = 100. An increase in the exchange rate index denotes a depreciation.

2Seasonally adjusted constant price indices rebased to 1992:Q3 = 100.

3Seasonally adjusted constant price indices rebased to second half of 1992 = 100.

In the first half of 1994, imports of manufactured goods rose by over 14 percent in real terms. Whereas most components of domestic demand were generally flat over this period, strong export demand resulted in an increase of 6 ¼ percent in import-weighted demand. Moreover, as discussed above, the recovery in investment, possibly related to the emerging capacity constraints in the export industry, contributed to the surge in imports in 1994. Reflecting the high import content of investment in machinery, the value of imports of investment goods rose some 20 percent in the first half of 1994.

Comparison with the 1981-82 Devaluation

The last major realignment of Sweden’s exchange rate took place during 1981-82, likewise in the context of a recession (Chart 7-3). The cumulative depreciation of 28 percent (10 percent in 1981, and 16 percent in 1982) resulted, initially, in a sharp fall in the relative price of Sweden’s exports (Chart 7-4). In consequence, Swedish exporters were able to improve their market share, which rose by 1.7 percent in 1982 and 7.7 percent in 1983. With world markets recovering from recession during 1982-84, this translated into real export growth of 5.3 percent in 1982 and 13.6 percent in 1983. By the end of 1983, however, Sweden’s export prices in domestic currency had more or less fully adjusted for the exchange rate devaluation, and market share accordingly began to retreat steadily. This decline in market share accelerated in 1989, by which time Swedish export prices had become extremely uncompetitive, but showed signs of bottoming out at historically very low levels in 1992.

Chart 7-3.Comparison with the 1981-82 Devaluation

Sources: International Monetary Fund, International Financial Statistics; and IMF staff estimates.

11992:Q3 = 100.

21981:Q3 = 100.

3Nominal effective exchange rate.

Chart 7-4.External Trade Performance

Source: Konjunkturinstitutet.

1Indices rebased to first half of 1980 = 100; market share defined as the ratio of Sweden’s exports to total imports of Sweden’s 14 largest OECD trading partners; relative price defined as the ratio of Sweden’s manufactured export price to the average industrial country export price.

2Indices rebased to first half of 1992 = 100; manufactured imports and import-weighted demand measured in constant prices.

As noted above, the surprise of the present recovery has been the unusual strength of imports. Chart 7-4 shows the evolution of import-weighted demand in Sweden since 1980, and compares it with the total volume of manufactured and semimanufactured goods imports. Apart from a slightly different uptrend, both imports and import-weighted demand moved fairly closely after the 1981-82 devaluation, and continued to do so in the downturn during the 1990-92 recession. However, as can be seen, since the end of 1992 there has been a sharp deviation in the relationship between imports and import-weighted demand despite the large devaluation of the currency. Thus, imports rose sharply following the devaluation despite only modest growth in import-weighted demand.

International Comparisons

The limited impact on the domestic currency prices of traded goods observed in Sweden after the 1992 devaluation is a feature shared with Italy and Spain but not with the United Kingdom, after these countries’ most recent currency depreciations. Chart 7-5 shows the path of export and import unit values on a monthly basis for Sweden during 1991-94 and compares their evolution with the path of the effective exchange rate. As can be seen, only about half of the exchange rate adjustment was passed through to export prices by the end of 1993. The pass-through to import prices was equally modest initially, although by the end of the year it exceeded that for exports. This behavior is similar to the response observed in Italy, and Spain, where the pass-through was even more muted, but dissimilar to that of the United Kingdom, where the pass-through was more or less full.

Chart 7-5.Sweden, Italy, Spain, and United Kingdom: Export Prices, Import Prices, and Exchange Rates

Sources: International Monetary Fund, International Financial Statistics; and IMF staff estimates.

1Export and import unit values and the effective exchange rate; August 1992 = 100 (except for Sweden where October 1992 = 100). An increase in the exchange rate index denotes a depreciation.

Chart 7-6 shows that export volumes in Sweden picked up four to five months after the devaluation and were rising strongly by the fall of 1993. Imports did not react significantly one way or the other until the fall, when they also rose quite sharply, consistent with the supposition that a higher-than-average proportion of imported components may have been employed in the production of exports. In Italy, export volumes rose more steadily following the 1992 devaluation. The more pronounced movement was in imports, which declined. Exports from Spain responded after about six months, and rose strongly thereafter, while import volumes were little changed. In the United Kingdom, on the other hand, the main effect of the devaluation was initially to set both exports and imports on a higher trend, although exports have since accelerated.

Chart 7-6.Sweden, Italy, Spain, and United Kingdom: Export and Import Volumes

Sources: International Monetary Fund, International Financial Statistics; and IMF staff estimates.

1Seasonally adjusted; 1992:Q3 = 100.

Recent Econometric Relationships for Trade Volumes

While Swedish export and import volumes have both increased sharply since 1992, the question arises as to whether this is entirely explicable in terms of the normal econometric relationships for trade volumes or whether there has been a structural break with the past.

Following the standard models of international trade, exports may be expected to be influenced by both demand and supply considerations. The demand for Sweden’s exports may then be posited as a function of world demand yw and the relative price of those exports px by comparison to other goods on the world market pw:

where e is the exchange rate (domestic price of foreign exchange). The supply of exports will normally be a function of the overall capacity of the economy yc, perhaps measured by a trend output term or an index of capacity, and the profitability of exports versus producing for the home market, measured by the export prices relative to domestic prices ph:

If the price level in the world and home markets, world income, domestic capacity, and the exchange rate can all be considered exogenous, then the equilibrating variable between demand and supply becomes the price of exports. However, because firms may wish to avoid undue fluctuations in their output price, they may choose to respond to cyclical changes in demand by adjusting capacity utilization in the first instance rather than prices. The greater the amount of unused capacity, the more costly it becomes, and prices may then assume a larger part of the burden of adjusting supply to demand. As the less than full pass-through from exchange rate movements to export prices suggests, Swedish manufactured exports are not perfect substitutes for other goods, and firms thereby have some leeway in setting their prices.

The less than full pass-through of exchange rate movements to domestic import prices suggests that the Swedish market is differentiated from others, and that importers have discretion in setting their prices. In these circumstances, the appropriate model for imports would be analogous to that for exports. Thus import demand would be related to a measure of Swedish demand yh, or income, and to the relative price of imports versus domestically produced goods:

The supply of imports to the Swedish market would be a function of overall foreign supply, perhaps measured by trend OECD GDP yw and the relative profitability of exporting to Sweden by comparison to world markets generally:

A simplifying assumption often made in the work on Sweden’s external trade is to treat px and pm as exogenous in the short run, reflecting stickiness, enabling the demand equations to be estimated directly as a function of exogenous variables. More generally, the focus of attention tends to be on the properties of demand equations rather than supply. In this regard, most of the work has concerned trade in manufactured goods. Demand elasticities are variously freely determined, or imposed at unity. Estimated price elasticities of demand are usually much lower in the short run than in the long run.2

In general, apart from questions surrounding the size of price elasticities, export equations for Sweden are reasonably well behaved and indicate that export behavior since 1992 has been fairly normal. Import equations for Sweden, on the other hand, have proved to be considerably less satisfactory or stable than for exports. Accordingly, the results of these equations need to be interpreted with caution.

Export Volumes

Alexius (1993) discusses a number of empirical estimates for export price elasticities for Sweden from various studies, mostly conducted during the 1970s or early 1980s. These estimates ranged from 1.75 to 2.16, although short-run price elasticities were generally much lower, of the order of 0.5 to 0.6. The specification employed by the National Institute, which covers exports of manufactured goods to Sweden’s 14 largest OECD partner countries, and which is used in their forecasts (as of fall 1994), is a semireduced form of the model presented above.3 Re-estimated over the period 1977 to 1993 their specification results in the following equation:

SE = 0.0193; R2 = 0.93; DW(1) = 1.82; first half 1977 to first half 1993.

where t-values are in parentheses. This equation, estimated over semiannual data, has a long-run price elasticity of demand of 2.15, and an even higher long-run price elasticity of supply of 3.45. The unit elasticity with respect to the demand variable (overall OECD imports) is imposed, which is an acceptable restriction for the variable employed.

The equation is generally quite well behaved. The restriction that real export demand be independent of the absolute export price level is acceptable. It is statistically stable over time, although it is notable that the estimated price elasticity of demand is higher for the 1985-93 period than for the 1977-84 period. For the full period, however, inclusion of a time trend has the effect of significantly lowering the estimated long-run price elasticity, from 2.15 to 1.53. Although the time trend is not significant at the 5 percent level, this suggests that some missing factor, such as the increasing integration of Sweden with partner countries in terms of trade, may be giving rise to a spuriously high price elasticity. The magnitude of the price elasticities of this equation is also called into doubt when the equation is estimated in long run format.4 The effect of regressing exports on the levels of the explanatory variables is to halve the long-run price elasticities:5

SE = 0.0277; R2 = 0.988; DW(1) = 1.81; first half 1977 to first half 1993.

Import Volumes

Long-run import price elasticities tend to be substantially lower for total imports than for imports of manufactured goods alone. Imports of fuel and raw materials are generally insensitive to price changes, whereas imports of food products are somewhat more sensitive. Imports of manufactured goods, on the other hand, are usually estimated to have a rather high price elasticity of -2, as with exports.

The National Institute’s semiannual estimated equation for manufactured imports m treats import supply as fully elastic and thereby represents a structural equation for the demand for imports, as in equation (3). The domestic demand variable employed yh is “import-weighted demand,” which comprises the sum of the components of total final expenditure, weighted according to the average import content.6 The estimated equation is

SE = 0.0246; R2 = 0.775; DW(1) = 1.90; first half 1973 to first half 1993.

This equation includes a term measuring the 15-year moving average trend ratio of exports to GDP, intended to capture the process of trade integration and import penetration referred to above. It is not clear, however, that this term performs this task adequately, since a simple time trend proves to be superior and knocks it out of the equation. Furthermore, such a simple time trend also knocks out the already rather weak estimated relative price elasticity. While the equation is technically stable over time, the relative price elasticity is not, and for the subperiod from 1984 to 1994 the estimated coefficient attracts the “wrong” sign. The equation also has difficulty forecasting 1993. It is notable that the equivalent equation estimated in long-run form also indicates that the term representing the relative price of imports is insignificant in this specification:7

SE = 0.0283; R2 = 0.990; DW(1) = 1.97; first half 1973 to first half 1993.

The Determination of Export and Import Prices

The pass-through of the depreciation of the currency since November 1992 to both export and import prices observed to date has been relatively subdued. In general, both export and import prices appear to be related in a plausible way to the expected determinants. Nonetheless, the empirical evidence reported in this section suggests that the response to the devaluation has been more muted than past relationships would have indicated.

Some limited econometric work on the relationship between the exchange rate and the price of traded goods has been done in the Riksbank. The Riksbank looked at the import prices of four categories of import: (1) crude oil, (2) pulp and paper, (3) automobiles, and (4) pharmaceuticals. This study employed monthly data, estimated over the years 1980-93, and regressed import prices on (1) export prices, (2) world prices, (3) the exchange rate, (4) domestic prices, and (5) GDP (as a proxy for the tightness of the domestic market). It found that the pass-through from the exchange rate was very rapid for crude oil (about one month); much slower for pulp and paper (about two years); and between 50 percent and 70 percent after 18 months for automobiles and pharmaceuticals. That the pass-through was rapid for homogeneous goods whose price is set in the world market is not surprising. Less clear, however, is the extent to which prices of manufactured imports have behaved abnormally.8

The behavior of manufactured export and import prices have been re-examined in this section using quarterly data. The specification selected was similar to that chosen by the Riksbank, except that an index of capacity utilization was found to give markedly better results than GDP in proxying the tightness of the domestic market.9 Both export px and import pm prices were related to the same set of explanatory variables. This approach yielded the following unrestricted equation for manufactured export prices:

SE = 0.0086; R2 = 0.998; DW(1) = 1.90; 1980:Q2 to 1993:Q4.

If import prices and domestic prices adjust in full proportion to a change in the exchange rate, then the freely estimated pass-through to export prices is 92 percent, and can be acceptably constrained to unity. With a fixed exchange rate, export prices are close to bearing a unit overall elasticity to world, import, and domestic prices combined (the freely estimated total elasticity is 0.83), but the restriction to unity is not quite acceptable by the usual criteria.10

For prices of manufactured imports, the same specification yielded insignificant terms for domestic prices ph. Removing this latter variable resulted in the following general unrestricted equation:

SE = 0.0098; R2 = 0.997; DW(1) = 2.11; 1980:Q2 to 1993:Q4.

With a proportionate change in export prices in response to a change in the exchange rate, the freely estimated pass-through to import prices is 91 percent, and this can easily be constrained to unity. The freely estimated total elasticity with respect to world and export prices with a fixed exchange rate, however, is only 0.55, and is well below the level that would permit a restriction to unity. Like the export price equation, this equation substantially overpredicts import prices during 1993. Also, as in the export price equation, capacity utilization is significant, indicating that importers set prices in a manner that depends on the tightness of the domestic supply of goods. Since capacity utilization is trending upward again, this implies that a widening of both export and import margins relative to other prices may be in prospect.

Since the import and export price equations interact, an analysis of the adjustment lags requires them to be solved together. To estimate the response to a change in the exchange rate, however, also requires an assumption about the behavior of home prices. These will depend, in turn, on the state of domestic demand, the price of import-competing products and imported inputs, as well as the wage determination process. Other things being equal, the evidence tends to suggest that home prices follow traded goods prices with a lag of about one year.11 Factoring in this lagged effect from changes in export and import prices to home prices, these equations imply that 50 percent of the exchange rate pass-through occurs within 9 months for import prices and within 15 months for export prices. Three quarters of the pass-through occurs after 2 ½ years for import prices and after 3 ½ years for export prices, other things being equal.

One argument that has been put forward to ex-plain the tendency of the pass-through equations to overestimate the reaction of export and import prices to the 1992 devaluation relates to the change in regime when the fixed exchange rate system was abandoned at this time. It has been suggested that following the 1981-82 devaluation, when those engaged in trade knew that the exchange rate adjustment was a deliberate realignment and would not be reversed, prices were adjusted relatively quickly. By contrast, during the float in 1993 there was a feeling that the exchange rate may have overshot and this may have caused decisions about export and import price adjustments to be delayed. To test this latter hypothesis, the estimation period for both export and import prices was shortened to exclude the 1981-82 period and the equations were rerun. For import prices there was indeed some tendency for the estimated rate of pass-through to be diminished, and the forecasting record of the equation improved. For export prices, on the other hand, there was little difference in the estimated rate of pass-through, and the forecasting record was not improved.

A second hypothesis concerning the delayed pass-through to export prices relates to exporters’ unit labor costs. These fell sharply during 1992 and 1993, and may have allowed exporters to use some of their increased profitability to moderate their prices when the exchange rate depreciated. Including a term representing relative normalized unit labor costs in the export price equation, however, did not significantly alter the equation’s performance. While the estimated coefficient was of the expected positive sign, it was not significant. Nor did the inclusion of such a term appreciably improve the fore-casting performance of the equation over the 1993 period.

Concluding Remarks

The large depreciation of the krona that followed the transition to a floating exchange rate regime has had important, although not fully anticipated, consequences for Sweden’s external trade. Export volumes have risen sharply, as would be expected, but imports have risen almost as strongly. More surprising has been the limited pass-through, so far, of the depreciation to export and import prices, and thereby ultimately to domestic prices. The determination of export and import prices seems to be generally explicable in terms of the usual variables, but the slowness in the rate of pass-through to date is still more modest than earlier relationships would have indicated. Investment by Sweden’s exporters has been vigorous, and this will no doubt have expanded capacity somewhat. However, for the future, tightening domestic supply and the continuation of the pass-through will likely have more noticeable effects on price inflation in 1995 than has been the case so far.

1

It is possible that this exceptional increase in productivity allowed Swedish enterprises to limit the pass-through of the devaluation to export prices more effectively than usual. It was not, however, possible to isolate this effect econometrically (see below).

2

This is a normal econometric finding for trade equations for many industrial countries. Less typical, however, are the relatively high long-run price elasticities of close to - 2 that have been estimated. Such estimates are at the upper end of the range of estimated elasticities surveyed by Goldstein and Khan (1985).

3

World demand yw14 is measured by total imports of Sweden’s 14 largest OECD trading partners and world prices pw14 by the corresponding deflator. Export prices px14 are represented by the deflator corresponding to manufactured exports to the 14 countries and home prices php by the domestic producer price index.

4

The logarithm of both exports and OECD demand are expressible as 1(1), while the logarithm of the relative price terms are borderline 1(0), but with very strong persistence.

5

These findings are similar to those obtained from quarterly data. Regressing manufactured exports on an interpolated series for OECD import demand, relative export profitability and export competitiveness generated an apparently well-behaved equation, but where both long-run supply and demand price elasticities were below unity. Furthermore, a forecasting test for the period 1993:Q1 to 1993:Q4 did not indicate aberrant behavior during this postdevaluation period. Experimentation with instrumentation of export prices did not reveal strong evidence of endogeneity in the current quarter. This finding confirms the supposition that, in the short run (that is, in one quarter), prices are sticky, and it is capacity utilization that adjusts to changes in demand. Under this scenario, prices are predetermined in the short run, changing only after capacity utilization has made the initial adjustment. As discussed more fully below, direct estimation of export prices shows that they react with a lag to changing degrees of capacity utilization.

6

The weights are private consumption of durables, 40 percent; public consumption, 9 percent; investment (equipment), 58 percent; construction, 10 percent; inventories, 40 percent; and ex-ports, 32 percent.

7

Estimation on quarterly data failed to improve on these rather unsatisfactory results. An equation estimated with both import profitability and price competitiveness, as well as import-weighted demand, showed both profitability and competitiveness to attract the “wrong” signs. Furthermore, a time trend was highly significant, suggesting that there was some other important missing variable at play. The best long-run equation that could be estimated, excluding this time trend, was for an equation that was un-stable and that did not comprise a cointegrating vector either. Using dynamic versions of this equation to forecast over the four quarters of 1993 results in substantial underestimation. Ordinarily, this might be taken to suggest that import demand was un-usual during this period, but the poor properties of the equation preclude much interpretation.

8

See Riksbank (1994) for a summary of these findings. The Riksbank study did not conduct stability tests for the postdevaluation period. Their equations were freely estimated, with a large number of regressors.

9

World prices (pw) were measured by an index of export prices for industrial countries (valued in foreign prices in terms of Sweden’s effective exchange rate), the exchange rate (e) by the effective exchange rate, domestic prices (ph) by the GDP deflator and capacity utilization (caput) was defined as unity divided by the proportion of spare capacity. Export, import prices, and capacity utilization were treated as endogenous. The exchange rate, being fixed though subject to adjustments, was treated as exogenous, despite the fact that it floated during the last year of the 13-year period examined.

10

The rejection of full long-run homogeneity may reflect the inappropriateness of one of the price indices used to represent world or domestic prices. Imposition of homogeneity does, however, improve the forecasting performance of the equation. Freely estimated, the equation results in a significant overestimation of the pass-through to export prices of the 1992 devaluation, while the restricted version does not suffer from this forecasting bias.

11

A study of the determination of domestic prices is beyond the scope of this section, but Nilson and Nilson (1993) report that changes in export prices precede changes in the consumer price index, in a statistical sense, with a lag of one year. The lag from changes in import prices is estimated to range from one quarter to two years, depending on the category of import.

    Other Resources Citing This Publication