Journal Issue

Norway: Selected Issues

International Monetary Fund
Published Date:
April 1999
  • ShareShare
Show Summary Details

IV. A Forecasting Model of Norway’s Non-Oil Current Account Balance19

A. Introduction

80. This note discusses a model for projecting the medium-term trends of the non-oil current account balance in Norway. The forecast is based on the September 1998 WEO growth and inflation projections for Norway and its partner countries, and assumes a constant nominal effective exchange rate (in terms of the ECU/euro) at the average level of 105 for 1998 through 2002. The model results suggest that with the economy moving back to a more neutral point in the cycle and with global demand strengthening in the medium term, the non-oil current account balance will improve by more than 3 percentage points of GDP, with the deficit declining from 9.8 percent of GDP in 1998 to 6.5 percent of GDP in 2002. This level of the non-oil current account deficit is sustainable in the long run: given an estimated net present value of oil wealth of 170 percent of GDP and assuming a 4 percent real rate of return on the State Petroleum Fund, Norway can finance a non-oil current account deficit of about 7 percent of GDP indefinitely.20

81. The medium-term projection of the non-oil trade balance is based primarily on estimated equations for trade in non-oil goods and services. The uncertainty about the changes in the other components of the non-oil current account balance (investment and other factor income and net transfers) adds uncertainty to the medium-term current account projections. Owing to the lack of information on the determinants of these flows, as well as their small size and lack of historical volatility, the staff has relied upon the projections of the Norwegian Ministry of Finance for these items.

82. The note is organized in the following way: Section B describes the staff’s model and discusses the estimation results; Section C summarizes the assumptions and compares the projections of the staff and of the authorities; and Section D presents the conclusions.

B. Determinants of Trade in Non-oil Goods and Services

83. The staff’s model was designed to estimate the non-oil exports and imports of goods and services (hereafter referred to as exports and imports). The data base was extracted from the Norwegian annual national accounts for 1978-97, with the non-oil trade flows comprising primarily exports and imports of goods and services produced by the mainland economy.

84. Non-oil trade flows were calculated as the difference between total and oil-related trade flows; the latter were defined as including oil and natural gas exports, oil platforms and modules, trade in other goods and services directly related to oil activities, pipeline transportation services, and oil drilling. In line with the classification of the Ministry of Finance, refined petroleum products were included in non-oil exports, making the non-oil current account somewhat dependent on petroleum production.21

85. The staffs model comprises export and import price and volume equations. All variables are expressed in logarithms. The tabulation below summarizes the names of the variables:

Variables of the model:

erwreal effective exchange rate based on wages
ernominal U.S. dollar-NKr exchange rate
gdpNorway’s real GDP
gdptpreal GDP of the trading partners
nmpinon-oil import price index for Norway (in domestic currency)
nmrnon-oil real imports of goods and services
nxpinon-oil export price index for Norway
nxrnon-oil real exports of goods and services
pdomfinal consumption deflator
PgdpGDP deflator
xdfeexport-weighted average foreign trade price22
xdfiimport-weighted average foreign trade price

86. Table 1 presents unit root tests for original variables in logarithms and for their changes. The null hypothesis of the unit root cannot be rejected for any of the original variables, but can be rejected at the 10 percent level for all the changes of the variables, with the exception of real effective exchange rates, import prices, foreign and domestic prices. The test on cointegration indicates the existence of a long-run relationship between the variables entering export and import price equations, and the import volume equation, which makes it possible to express these equations in the log-linear form. The equation for export volume is expressed in a similar form, although the cointegration test suggests no significant long-run relationship between the variables. The method of ordinary least squares (OLS) is used to estimate the model.

Table 1.Norway: Unit Root Tests1
VariableWeighted-Symmetric τ Test
Δ gdp-Δ gdp*-3.22*[0.046]
Δgdptp -Δgdptp*-2.96*[0.094]
Δnmpi- Δnmpi*-2.64[0.216]
Δnmr- Δnmr*-2.86*[0.124]
Δnxpi- Δnxpi*-3.95*[0.005]
pdom -pdom*0.93[0.999]
Δpdom - Δpdom*-2.60[0.236]
Δpgdp- Δpgdp*-2.82*[0.138]
xdfe -xdfe*-2.50[0.292]
Δxdfe - Δxdfe*-2.81*[0.141]

See text for data definitions. An asterisk denoted a test statistic that is significant at the 10 percent level The Weighted Symmetric τ test involves a weighted double-length regression in which the dependent variable is regressed on leads and lags of its own changes. P-values are shown in brackets.

See text for data definitions. An asterisk denoted a test statistic that is significant at the 10 percent level The Weighted Symmetric τ test involves a weighted double-length regression in which the dependent variable is regressed on leads and lags of its own changes. P-values are shown in brackets.

87. The estimation results suggest that export prices are dependent on their own lagged values, domestic prices, and on foreign prices of exported goods denominated in domestic currency;23t-statistics are reported in parentheses (Figure 8, Panel 1):

Equation 1:


Sources: Staff calculations.

All explanatory variables are found to be significant. The equation suggests that in the long run, Norwegian exporters are price setters. In the short run, however, the response of export prices to changes in domestic costs and foreign prices is almost equally strong.

88. Import prices are also found to be dependent on their own lagged values and on foreign prices, denominated in Norwegian kroner (Figure 8, Panel 2):24

Equation 2:

The equation suggests that Norwegian import prices are surprisingly insensitive to changes in foreign prices. The long-run effect of changes in foreign prices would not fully pass through to Norwegian import prices.

89. The estimation of Norwegian non-oil exports is based on a standard trade equation of the form (Figure 9, Panel 1):

Equation 3:


Sources: Staff calculations.

The estimated equation indicates a price elasticity of Norwegian non-oil exports that is below unity, but has greater sensitivity to contemporaneous partners’ income.

90. The equation for non-oil imports has the following specification (Figure 9, Panel 2):

Equation 4:

The estimated equation suggests that Norwegian non-oil import volumes have a very weak import price elasticity but respond strongly to domestic output. This is in line with the authorities’ claim that imports constitute a significant share of inputs for production of certain goods, thus making imports fairly insensitive to changes in their prices. In the long run import volumes appear to be driven by movements in domestic output.

C. Medium-Term Forecasting of the Non-oil Current Account

91. The staff’s forecast is based on the medium-term WEO assumptions, outlined in the tabulation below:

Underlying Assumptions(Change in percent, unless otherwise specified)
Real GDP2.
GDP deflator0.
REER based on wages (index)2595.097.898.898.898.8
Nominal ECU/NKr
exchange rate index26105105105105105
Nominal U.S. dollar/NKr
exchange rate (index)270.1320.1320.1320.1320.132
Partners’ real GDP2.
Export-price deflator (U.S. dollars)-
Import-price deflator (U.S. dollars)-

92. The forecasting is complicated by the two interrelated exogenous factors—the abrupt weakening of international demand for Norwegian exports, and the fall in oil prices. The projections of the Ministry of Finance suggest that the non-oil current account deficit would widen in 1998 to 9.3 percent of GDP—a 1.4 percentage point of GDP deterioration from 1997. The weakening of global demand and of international export prices is expected to have a stronger effect on exports than an estimated 7 percent nominal depreciation of the krone against the U.S. dollar in 1998. Cyclical factors—the economy has been operating above potential for the past two years—also have a strong impact on import growth.

93. While the staffs model predicted a deterioration in the non-oil current account in 1998, it did not capture the full impact of the shocks on the trade balance. Accordingly, the forecast for 1998 was replaced with the authorities’ projections of the non-oil current account balance, and the forecasting period set to start in 1999. The tabulation below summarizes the forecast:

Forecast of the Medium-Term Non-oil Current Account(In percent of GDP)
Investment and other factor income and transfers28-0.9-1.3-1.4-1.4-1.5-1.4
Current account balance-7.9-9.8-8.8-8.0-7.3-6.5
Memorandum item:
Ministry of Finance projection of the current account,-7.9-9.3-8.4-7.0-6.3-4.9
Of which
Trade balance-7.0-8.5-7.5-6.3-5.8-4.7
Investment and other factor income and transfers-0.9-0.8-0.9-0.7-0.5-0.2

94. As suggested by the tabulation, the non-oil current account deficit is projected to narrow by 3.3 percentage points of GDP in 1998-2002, which is a smaller improvement than projected by the Ministry. The discrepancy between these two projections is due to at least two factors:

  • The Ministry included the projected returns on the State Petroleum Fund in non-oil investment income, reducing the non-oil current account deficit, whereas the staff netted out these returns in order to be able to compare the results transparently to the income stream from Norway’s oil wealth; and
  • The authorities’ forecast incorporates some factors that are not captured by the staffs model, such as capacity constraints faced by industries dependent on hydroelectric power, that are expected to depress exports in 1999-2002.
  • Adjusting for the difference in treatment of SPF income, the Ministry and staff projections are roughly the same for 2001-2002.

95. The improvement of the non-oil current account in the medium term is expected to be brought about primarily by the slowdown of the domestic economy, which would have a dampening effect on imports, and by the assumed recovery in global demand, which would positively affect exports. A loss of competitiveness is projected to depress export performance in 1999-2000, but competitiveness would stabilize thereafter. This forecast, however, is subject to a large margin of error owing to the uncertainty of international economic developments, the evolution of oil prices, and the ability of the government to moderate the excess demand pressures in the domestic economy. The tabulation below shows the changes in non-oil exports and imports over time:

Evolution of the Non-Oil Current Account(Change in percent)
Nominal non-oil exports0.
Nominal imports5.
Changes in prices:
Changes in volume:

D. Conclusion

96. The staffs medium-term current account forecast shows a reduction in the non-oil current account deficit from 9.8 percent of GDP in 1998 to 6.5 percent of GDP in 2002, The staffs projections for 2001-2002 suggest that Norway would achieve a sustainable non-oil current account balance, assessed in relation to the permanent income stream on its oil wealth.


Prepared by Natalia Koliadina.


At current levels of oil production, a sustained 10 percent decline in oil prices would reduce the overall current account surplus by 1 percentage point of GDP. For the longer run, official calculations of the net present value of oil wealth were adjusted downward by about 10 percent in response to the decline in oil prices that took place in the first eight months of 1998, because the authorities considered that much of the previous increase in oil prices and some of the subsequent decline were temporary, and also because some oil production and exports was postponed to future years.


Exports of refined petroleum products constituted almost 2 percent of GDP in 1997, and were found to be highly correlated with oil and natural gas production.


Average foreign trade prices are based on countries’ 1987-89 composition of trade in manufactured goods and commodities, using world price indicators.


The result of the F-test shows that the null hypothesis of the same coefficient on foreign prices and dollar-kroner exchange rates cannot be rejected.


The null hypothesis of foreign prices and the exchange rate having the same coefficient cannot be rejected at the 5 percent level.


The projections of the real effective exchange rate based on relative wages assume a constant nominal exchange rate on an inverted ECU index, and slightly faster wage growth in Norway than in trading partners over the next two years.


Based on an inverted ECU index.


The dollar-NKr nominal exchange rate is fixed for 1999-2002 to avoid the effects of its changes on the Norway’s current account balance, since trade with the United States is not significant—less than 7 percent on both export and import sides.


This component includes non-oil investment income net of returns on the State Petroleum Fund, current transfers and net wages.

Other Resources Citing This Publication