This Background Paper examines the medium-term economic outlook (1997–99) for Norway. The central feature of Norges Bank’s reference case projection for the medium term is that the expansion of mainland output will slow from 3.3 percent in 1995 and 2.8 percent in 1996 to an annual average of 2 percent in 1997–99. Overall GDP growth will also slow from about 4 percent in each of 1995 and 1996 to 2 percent in 1997–99. Inflation is forecast to remain low, at 2 percent in 1996 and on average 2.3 percent per year in 1997–99.

Abstract

This Background Paper examines the medium-term economic outlook (1997–99) for Norway. The central feature of Norges Bank’s reference case projection for the medium term is that the expansion of mainland output will slow from 3.3 percent in 1995 and 2.8 percent in 1996 to an annual average of 2 percent in 1997–99. Overall GDP growth will also slow from about 4 percent in each of 1995 and 1996 to 2 percent in 1997–99. Inflation is forecast to remain low, at 2 percent in 1996 and on average 2.3 percent per year in 1997–99.

VII. Norway’s State Petroleum Fund 1/

The Norwegian authorities have established, and will soon begin to endow, the State Petroleum Fund (SPF) to preserve in the form of financial assets part of the nation’s petroleum wealth. The rationale for the SPF stems from the special problems that arise from reliance on oil income--a highly volatile source of revenue whose generation involves the dissipation of wealth and whose spending can adversely affect the competitiveness of other sectors. With these pitfalls in mind, the SPF was conceived as a device designed to impose discipline and transparency on the use of oil revenue, protect the mainland economy from loss of competitiveness and contribute to long-term fiscal sustainability.

The SPF will be activated with the transfer to it of the entire 1996 fiscal surplus, budgeted at Nkr 12.6 billion or 1 1/3 percent of GDP, which is to be invested in a portfolio of foreign financial assets managed by Norges Bank. The Government has announced that subsequent fiscal surpluses--which are expected to mount prodigiously over the medium term as the result of a final surge in petroleum revenue coupled with public expenditure restraint--should be similarly allocated. 2/ In so doing, the authorities aim to prepare for the long-term fiscal challenge of dwindling petroleum revenues and an aging population. Norges Bank estimates that the SPF will need to marshall assets in excess of 150 percent of GDP by 2030 for the SPF to become a self-sustaining fund capable of offsetting the expense of future social security obligations and the tapering off of oil revenue.

1. Rationale for the SPF

The idea for a facility like the SPF dates virtually from the initial exploitation of Norway’s North Sea reserves in the early 1970s. Although no such facility was established, substantial financial assets were accumulated, as reflected in the steady improvement in Norway’s net foreign asset position, especially prior to the recession that began in 1988. The SPF was established in 1990, only to lie dormant, as the government pursued an expansionary fiscal policy to counter the continuing recession. The delay in establishing a mechanism for setting aside part of the income generated from the petroleum sector is also attributable to the authorities’ concern that this wealth should benefit the entire population through increased public spending to meet social objectives, lower taxes, and a shorter work week. With petroleum revenue so allocated, there was little need for a special facility to garner savings. Moreover, in the early days of Norway’s oil development, the authorities pursued a policy of moderate petroleum extraction, on the expectation that oil prices would rise, making petroleum in the ground a good investment (Box 1). This extraction policy initially limited the revenues available for accumulation in a facility like the SPF.

Resource Depletion

The Norwegian authorities have sought to maximize the future revenue stream from petroleum resources. In theory, the optimal rate of extraction of a depletable resource is determined by equalizing the rate of return on the natural resource and financial assets--both domestic and foreign. The appropriate rate of extraction is thus a function of the expected evolution of petroleum prices and the return on financial assets. In practice, however, the latitude to fine tune the rate of extraction is extremely limited, and often restricted to the initial decision of whether or not to begin to develop a field--a process typically regulated through licensing. Once begun, the process of exploration, development, and extraction can take on a momentum of its own, reflecting the need to capitalize on costly and short-lived investment, independent from changes in expectations on petroleum prices and financial asset returns. The Norwegian authorities initially adopted a moderate annual rate of extraction of 90 million tons of oil equivalent (TOE) on the expectation that oil prices would rise, making oil in the ground a sound investment. In the event, this target was outstripped in 1988, and current output stands at about 160 million TOE. The rapid rate of petroleum extraction reflects efforts to capitalize on the heavy investment in the sector and the expectation that oil left in the ground will return less than financial assets.

Intergenerational fairness and the desirability of capital preservation implicitly underlie the establishment of the SPF, whose ultimate objective is to allocate oil revenue over time, thereby preserving for future generations some of the benefits of Norway’s oil wealth. However, the current impetus behind the SPF is more concrete and pragmatic: social security expenditures are expected to rise as oil revenues decline, leading to difficult fiscal challenges over the long term.

The SPF is also conceived as a means of dampening the inflation and loss of competitiveness that has been an unwanted consequence of the expansion of the petroleum sector. Since the commencement of oil exploitation in the early 1970s, export-oriented industries and those exposed to competition from imports declined sharply as relative unit labor costs in Norwegian manufacturing rose by almost 40 percent. 1/ Manufacturing was effectively crowded out by the exchange rate appreciation, increased domestic expenditure, and wage and price pressure generated by the growth of the petroleum sector, while public sector employment grew to 31 percent of total and subsidies and transfers rose to one quarter of GDP.

The decision to invest the SPF’s resources in foreign financial assets is designed to protect the competitiveness of the mainland economy by sterilizing oil revenue and thereby relieving possible upward pressure on the real effective exchange rate. Domestic investment, in contrast, would be inflationary, undercut competitiveness, and result in the allocation of resources to low-yielding ventures. Moreover, the investment in foreign financial assets is aimed at diversifying Norway’s wealth to dampen the volatility of government revenue and shield the economy from the disruption of oil prices fluctuations.

2. Proposed structure of the SPF

In addition to serving as a store of wealth for future generations, the SPF is conceived as an accounting device designed to discipline and render transparent the spending of petroleum revenue. To this end, all petroleum revenue will first flow through the SPF. Any budgetary call on petroleum revenue would then represent a general transfer from the SPF account to the budget. The level of assets accumulated in the SPF would represent the difference between these flows. Accordingly, the SPF will not accumulate funds unless a budgetary surplus is achieved. Moreover, any non-oil fiscal deficit not covered by petroleum revenue will be financed through previously accumulated SPF assets so long as there are assets available to tap. As a result, changes in the SPF’s asset position will provide an indication of changes in the central government’s financial wealth.

Norges Bank will manage SPF assets in a segregated account. Although detailed investment guidelines have not yet been established, it is now anticipated that the SPF’s assets would be managed in a manner similar to the way in which Norges Bank allocates its investment portfolio of foreign exchange reserves, which is designed to provide transactions balances and income. 1/ The investment portfolio, which is allocated to foreign government bonds, has an average duration 2/ of about three years. As discussed in greater detail in below, the investment guidelines eventually established to guide the composition of the SPF portfolio should reflect the risk tolerance and return objectives deemed appropriate to fulfill the SPF’s funding objectives.

3. Rate of accumulation

The authorities have yet to determine the targeted level and timing of flows to the SPF, its ultimate size, and whether the rate of asset accumulation would be determined according to a fixed rule or on an ad hoc basis. Although the government is assessing the level of savings from oil revenue and budgetary adjustment required to meet the long-term fiscal challenge, only the mechanics of the SPF and the principle that its assets should represent a net increase in the government’s financial wealth have been established.

The pace of asset accumulation in the SPF will be determined by a three-part procedure: (i) broad targets are to be established in successive long-term programs; (ii) within that framework, the targeted annual transfer will be assessed in the revised national budget; and (iii) Parliament will approve the targeted transfer as part of the budget. The actual transfer to the SPF will be determined by the budget surplus achieved, so that the accumulation of assets in the SPF will reflect the actual level of government savings. In practice, the annual transfer to the SPF will be made once the fiscal outturn is known; the transfer of the 1996 budget surplus (projected at NKr 12.6 billion) will be made in early 1997. From the authorities’ perspective, the decision to calculate the accumulation of SPF assets as a residual based on actual government savings was designed to ensure that assets accumulated in the SPF were not indirectly funded (and effectively offset) by government borrowing.

Norges Bank has published calculations based on what it terms the “Oslo criteria”, an accumulation rule that requires both fiscal consolidation and savings from the expected future stream of petroleum revenue. The Oslo criteria would allocate to the SPF (i) all further nominal increases in oil revenue and (ii) the resources released by reducing the non-oil budget deficit by an additional 3-4 percent of GDP. 1/ This rate of accumulation would result in SPF assets equivalent to 150 percent of GDP by 2030 that would ultimately peak at 170-180 percent of GDP. 2/ Norges Bank calculates that a portfolio of this magnitude generating a 4 percent annual real return would be sufficient to meet the projected long-term non-oil budget deficit and maintain the SPF at a constant percent of GDP. 1/ Another possible criterion that has been examined--based on permanent income--(Box 2) was considered by the authorities to be insufficiently restrictive given the uncertainty of its calculation, its tendency to fall as a percent of GDP over time, and the size of the financing requirement.

Petroleum Wealth, Permanent Income, Rents, and Net Cash Flow

Petroleum fields are not traded: there is no observable market-determined price for such assets, and the risk of their ownership cannot easily be diversified. Petroleum wealth is typically calculated as the present value of the expected income stream from its exploitation, which in turn depends on price estimates (which are highly variable), extraction costs and techniques (which are both also constantly changing) and exploitable reserves (whose extent is uncertain and influenced by price and technological developments and new finds). In Norway, the pure resource wealth from petroleum is calculated as the net present value of petroleum rents, or the expected excess return to the sector, plus fixed capital investment in the sector. Because of the variability of both petroleum prices and reserve estimates, great uncertainty surrounds the extent of Norway’s pure resource wealth from petroleum, which has ranged between Nkr 500 billion and Nkr 2,000 billion. It is now put at Nkr 1,000 billion (150 percent of mainland output), of which Nkr 700 billion represents the present value of the economic rent from the petroleum sector with the balance attributable to investment in the sector. About 80 percent of this wealth accrues to the state.

Permanent income is calculated as the average real rate of return on the estimated petroleum wealth; it is an estimate of how much petroleum wealth may in principle be consumed without reducing its real value. Using a real discount rate of 7 percent--the hurdle rate used for public sector projects in Norway--results in a permanent income of Nkr 70 billion, of which Nkr 56 billion accrues to the state. Like the wealth estimate upon which it is based, permanent income is volatile and its risk is not easily diversifiable, factors that complicate budgetary planning and consumption and savings decisions. Moreover, the choice of discount rate is controversial--the rate used in Norway seems quite high compared with the return likely from other assets.

Rents represent the excess return--i.e. returns above a normal profit margin--to an activity. Petroleum rents are defined as the difference between petroleum revenues and the cost of production, including a normal rate of return on investment in the sector. During 1976-91, estimates of these rents have ranged between a negligible level to Nkr 60 billion in 1985.

The net cash flow generated by the petroleum sector represents the annual contribution of the sector to the budget. It is the sum of tax revenues, share dividends from Statoil, and revenues from the state’s direct financial interest (SDFI) in the petroleum sector less the government’s share of current investment. Net cash flow is also highly variable, fluctuating with petroleum price movements, extraction rates, and investment The net cash flow from petroleum activities of 5 percent of GDP in 1996 is projected to peak at 8.6 percent of GDP in 2001 and to decline thereafter, reaching 1.3 percent at 2030.

4. Use of SPF resources

The SPF is structured such that its resources may be used to finance the non-oil budget deficit, invest in foreign financial assets, and underwrite up to half of the net lending of state investment banks. As already discussed, the SPF is designed to ensure that its resources represent an accumulation of net financial assets by the government, rather than an account offset by liabilities accrued elsewhere. For this reason, the SPF’s resources are to be used to meet the general financing requirement of the budget, rather than a particular expenditure. Thus, in the authorities’ view, earmarking the SPF to future social security expenditures would not be consistent with sound budgeting practice, could complicate the reduction of the generous benefits that exacerbate the budgetary impact of an aging population, and would in any case account for only a portion of future budgetary requirements.

The decision to use the SPF to underwrite up to half the net lending by state banks was similarly motivated by the concern that the SPF’s resources should represent a net accumulation of financial assets by the government. Indeed, the government had originally proposed to Parliament that the entire net lending of the state banks be funded through the SPF, thereby effectively transforming a below-the-line transfer to state banks into a more transparent current expenditure accounted for above the line in a way that would preclude the accumulation of offsetting liabilities.

5. Asset allocation issues

Institutional investors typically minimize risk by matching the variability of the assets under management with that of the liabilities being funded using a mean-variance optimizer (Box 3). The case of the SPF is somewhat complicated by the decision to use its assets to fund the general government financing requirement, rather than a more clearly defined liability stream. However, estimates of the likely evolution of the general government financing requirement and Norges Bank’s estimate of the target rate of return for the SPF portfolio provide some guidance for determining an appropriate asset allocation policy. First, the SPF’s time horizon is long and its early liquidity needs low. Second, a significant proportion of the assets being funded are indexed to wages, suggesting the need to invest in assets that provide a degree of inflation protection. Third, since part of the rationale for the SPF is to shield the budget and the economy from the ebb and flow of petroleum revenue, a premium might be placed on assets that are negatively correlated with petroleum income. Finally, the 4 percent real return targeted by Norges Bank suggests the need to construct a portfolio tilted toward equities and long duration bonds, and away from cash equivalents. 1/

Objectives and Pitfalls of Portfolio Optimization

The standard approach for identifying portfolios with the best tradeoff between risk and return is mean-variance optimization, the main inputs for which are the expected return, standard deviation and correlation coefficients of the various investable assets. The central insight of portfolio construction is based on Markowitz’s observation that if individual assets are imperfectly correlated, the total risk of the portfolio is less than the weighted average of the risks of its constituents. The objective of the optimization exercise is to construct portfolios such that, for any level of risk, there is no other combination of assets that results in a higher expected return.

By varying the level of portfolio volatility, a continuum of such “efficient” portfolios can be constructed to form a frontier that defines an optimal balance of risk and return. These efficient portfolios will be characterized by different weights to the various investable assets and a different maximum level of portfolio return for each level of portfolio risk. A rational investor should choose that portfolio on the efficient frontier that best reflects the investor’s willingness to trade risk for return.

In practice the actual portfolio chosen by investors also reflects institutional factors and is influenced by the portfolio preferences of other, like, investors. These institutional factors are typically expressed in the form of constraints on the weights that may be assigned to any asset. The most common such constraint is one that precludes the possibility of short selling; others include limits to avoid a concentration in a particular asset that is outside the norm followed by similar investors, or otherwise considered imprudent.

Mean-variance optimization suffers from a number of inadequacies. First, the optimization process tends to emphasize forecasting errors, a phenomenon referred to as error maximization. Assets whose expected returns are overestimated or whose risks are underestimated will be inordinately favored in the optimization results. Since optimization models are highly sensitive to the underlying assumptions, the resultant portfolios can significantly underweight or overweight assets as a result of those errors. Second, the capital market assumptions underlying the optimization are either subjective or, if based on historical returns, period specific. Third, optimal portfolios do not generally result in portfolios that are “diversified” in the sense of placing relatively small weights on any individual asset. Finally, the results of the optimization are unstable: small changes in the underlying capital market assumptions result in wide swings in the weights assigned to the various assets.

It is instructive to consider the portfolios of U.S.-based pension and endowment plans, which have a similar time horizon, if somewhat more modest real return objective. At end-1992 the portfolios of these institutions were invested in U.S. equities (44 percent), U.S. bonds (34 percent), U.S. cash equivalents (8 percent), non-U.S. equities (5 percent), real estate (4 percent), venture capital (4 percent), and non-U.S. bonds (1 percent). Using recent historical performance data as a guide, this aggregate portfolio had both an expected annual return and a standard deviation of about 10 percent.

A main task still facing the authorities is to determine the investment guidelines and benchmark portfolio that will be used to direct investment decisions for the SPF portfolio and measure its performance. 2/ The benchmark sets forth the acceptable exposure of the portfolio to various market risks--mainly stock market and interest rate risk. Thus, the benchmark embodies the target return sought for the portfolio and the variability of returns that may be countenanced by establishing a range of acceptable weights for the assets that may be included in the portfolio. In addition to determining the weight to attach to each asset class (mainly the stocks, bonds and bills of various countries), the investment guidelines eventually adopted for the SPF will need to indicate the degree of exposure to currency movements that may be maintained and provide guidelines on the acceptable level of credit risk. The authorities will also need to decide whether to manage the SPF’s portfolio “in house” or to entrust this task to external money managers, and whether to pursue an active or passive approach to security selection (Box 4).

Professional Money Managers

Many institutional investors rely on investment professionals to manage their portfolios. There are two main types of professional money manager those offering index funds, or some other passive investment approach; and those attempting to beat the market through active security selection. The former group competes on the basis of reliability and low cost arising from economies of scale and limited asset turnover. The latter competes on the basis of the (difficult-to-substantiate) ability to generate risk-adjusted returns superior to those of the market.

Institutional investors typically encounter three main problems in constructing portfolios of active managers. The first revolves around manager selection. Very few managers, if any, consistently perform better than the market. Identifying this select few is a difficult and time-consuming task. If, against the odds, some consistent superior performers are identified, their results must be strong enough to compensate for the shortcomings of managers hired in error. In practice, the narrow margin of superior performance, when evident at all, is unlikely to be sufficient to make up for the poor performance of others. Thus, for a portfolio of active managers to beat the market, those selecting managers must be unerring in their judgment.

Firing managers, the other side of the coin, also poses problems. There is a strong belief among institutional investors who hire external money managers that the performance of managers will revert to the mean: stellar performance will be followed by indifferent returns, and vice versa. This suggests that managers should be fired only after a period of superior performance, rather than immediately following poor results. However, such an approach only serves to highlight the difficulties of identifying managers: if even bad managers can be expected to have periods of strong performance, how can they be distinguished from the good? In practice, firing managers is difficult, especially following a period of good returns, and costly, generating unnecessary portfolio turnover.

The third major pitfall centers on the inefficiencies of portfolio composition using a number of active managers. The active stock selection of one manager favoring a certain sector of the market may offset those of another oriented toward a different market segment, with the result that the aggregate portfolio is nothing but an overpriced index fund. Moreover, the tendency of managers to minimize relative risk by straying only slightly from the market benchmark contributes to quasi-indexation and inefficient portfolio construction.

As portfolios increase in size, these problems are exacerbated. Additional assets often entail the need for more managers for prudential reasons, uncertainty about manager abilities, or a desire to experiment with new techniques. However, a multiplicity of managers only compounds the problem of manager selection, dissipates the performance of good managers, and increases the likelihood of closet indexation at high cost.

1/

Prepared by David J. Ordoobadi.

2/

The State Petroleum Fund’s guidelines envisage that the SPF may be used to finance the non-oil budget deficit and one half of the increase in the net lending of the state banks.

1/

A more detailed discussion on the evolution of Norway’s competitiveness is provided in Appendix IV of SM/95/17, January 1995.

1/

Norway’s foreign reserves are split into two pools: a hedging and an investment portfolio. Norges Bank has designed the hedging portfolio to immunize Norway’s foreign debt from currency and interest rate fluctuations by holding bonds in the same size, currency, and maturity as the foreign debt. The hedging portfolio will be used to pay off the foreign debt as it matures. The longest outstanding foreign bond is scheduled to mature by end-1998.

2/

Duration is a measure of the sensitivity of the price of a bond to interest rate fluctuations. It is calculated as the time-weighted average of the bond payments as a percentage of the bond price, and represents the weighted average time to full recovery of principal and interest.

1/

Further details on the Oslo criteria and other estimates of the resource requirements to smooth the transition to the post-oil era are provided in the background paper on Norway’s long-term fiscal challenge.

2/

Five possible paths of fiscal consolidation have been considered: a one time tightening of 2.9 percent of GDP; a consolidation of 3.2 percent over ten years; 4 percent over 25 years; 5.1 percent of GDP implemented pari passu with the expected increase in social security expenditures; and 7.6 percent of GDP over 80 years.

1/

Assuming a long-run GDP growth rate of 2.2 percent and a 4 percent real return on its portfolio, the SPF could finance a non-oil deficit of up to 3.2 percent of GDP and remain at a constant 180 percent of GDP. This (somewhat ambitious) target rate of return has obvious implications for the type of portfolio the SPF would need to hold.

1/

Since 1900, real return have averaged 6 percent for U.S. stocks 2 percent for U.S. bonds, and 1 percent for U.S. Treasury bills. Since 1919, the average annual real returns on U.K. equities, gilts and Treasury bills have been 7 percent, 1.4 percent and 1 percent respectively.

2/

The benchmark is a relative--rather than an absolute--performance measure. An investor that weights each asset in a portfolio at the midpoint of the ranges for that asset would take no risk vis-à-vis the benchmark, since both it and the portfolio would move together with the market. The portfolio would, of course, still face an absolute loss as a result of market movements.