The Government Petroleum Fund 1998
The composition of the Government Petroleum Fund’s portfolio
The Ministry of Finance is responsible for laying down the strategy
of the Government Petroleum Fund. This strategy is reflected in a
benchmark portfolio that forms the basis for Norges Bank’s management
of the Petroleum Fund. This article describes the Petroleum Fund’s
benchmark portfolio, addressing in particular issues relating to the
equity portion of the benchmark.
The benchmark portfolio
The return on a portfolio is essentially determined by the strategy and
framework laid down for its management. Empirical research in the US shows
that 90-95 per cent of the differences in return between large pension
funds are due to their different choice of strategy. The most important
strategic decisions concern the distribution of investments among various
asset classes, such as bonds, equities and property, and the distribution
by country. The remaining 5-10 per cent of the differences in return is
related to the managers’ selection of securities.
In the operational management of the Government Petroleum Fund, Norges
Bank may deviate slightly from the benchmark portfolio defining the
strategy. However, there are limits imposed on this deviation. The quality
of the operational management can be measured by comparing the return on
the actual portfolio with the return on the benchmark portfolio. One
important management objective is to achieve an excess return, measured in
this manner, within the stipulated limits and restrictions. Another
important goal is to implement the established investment strategy in a
cost-effective manner. Requirements concerning satisfactory risk
management systems and control routines still take precedence over the
objectives of returns and effectiveness. In building up the management,
Norges Bank has given higher priority to developing control systems and
expertise than to active management with a view to achieving higher excess
returns.
The benchmark portfolio consists of equities and fixed-income securities
(bonds). The equity portion constitutes 40 per cent of the market value of
the benchmark portfolio, while the fixed-income portion accounts for 60
per cent. The equity portion consists of securities from 21 different
countries, while the fixed-income portion includes paper from 18
countries. The shares for the individual countries are listed in the
documentation section of the annual report, together with a graphical
presentation of the benchmark portfolio.
The countries selected for the investment universe must have well
functioning, liquid securities markets and satisfactory company, stock
exchange and securities legislation. The Ministry of Finance decided, on
the advice of Norges Bank, to delimit this by looking at those markets
that are defined as developed in one of the most widely used international
market indices. This is further discussed in the National Budget for 1998.
The countries in the investment universe are geographically located in
three regions: Europe, Asia and Oceania, and North America. The Ministry
of Finance has decided that the benchmark portfolio is to consist of 50
per cent from Europe, 20 per cent from Asia and Oceania, and 30 per cent
from North America. These figures were arrived at partly on the basis of
GDP figures in each region, and partly on current Norwegian imports from
each region. The portions may thus roughly correspond to the weights in a
price index for future Norwegian imports. The idea is that changes in
import patterns will probably take the form of a shift towards regions
with high production.
The Ministry of Finance has chosen to construct the benchmark portfolio
for equities within each region in proportion to market capitalisation
weights. This implies looking at the market values of all companies listed
on the stock exchanges of the countries in each region, and determining
the weights accordingly. The benchmark portfolio benefits from economic
growth in countries outside the investment universe because companies
listed in approved markets may have production activities in several
countries.
Market capitalisation weights are constantly changed because price
movements in the equity markets in question are not synchronised. The
composition of the benchmark portfolio therefore has to be adjusted on a
regular basis. The Ministry of Finance has decided that this will take
place at the end of each quarter.
In each market, the benchmark portfolio consists of shares in the
companies included in the Financial Times/Standard & Poor’s
Actuaries World Index. This international equity index includes the
largest companies listed on each stock exchange. In the Petroleum Fund’s
investment universe, it includes some 2 040 companies, which accordingly
are included in the Fund’s benchmark portfolio. However, the Fund’s
actual investments may be made in all the listed companies in the
countries in question, which total about 20 000.
The FT/S&P Actuaries Index is used by many other international
equity investors. Nevertheless, the Petroleum Fund’s investment
universe does differ from what is customary, for example, for large
pension funds with international equity portfolios. Such funds normally
have a slightly larger spread of equity investments with significant
portions also outside markets defined as ‘developed’. In a
submission published in the National Budget for 1998, Norges Bank raised
the question of whether emerging markets should be included in the
investment universe, but the Ministry of Finance did not want this issue
to receive priority while Fund management was being developed. The
Ministry indicated, however, that it would revert to this issue at a later
stage, and announced in the Revised National Budget for 1998 that further
work would be done to expand the Petroleum Fund’s investment area.
The fixed-income portion of the benchmark portfolio has the same
regional weights as the equity portion, but the distribution principle
within each region varies. The use of market capitalisation weights might
lead to the Petroleum Fund purchasing most securities in countries with a
high debt level. Instead, the Ministry of Finance decided to use GDPs as
weights.
At the end of 1998, the Petroleum Fund had a market value of about USD
23 billion, and the equity portion of the benchmark portfolio was thus
worth about USD 9 billion, whereas the fixed-income portion was worth
around USD 14 billion. By way of comparison, the total market value of all
the companies included in the Petroleum Fund’s country indices for
equities was around USD 19 000 billion. This means that with an even
investment distribution among all companies in the index, the Petroleum
Fund would have an ownership interest of less than 0.1 per cent in each
company. This applies to all three regions in which the Petroleum Fund
invests. The market value of the bonds in the Petroleum Fund’s bond
indices was approximately USD 7 000 billion, so that the Fund’s
market share is around 0.2 per cent.
Given the assumptions about allocations to the Petroleum Fund made in
the National Budget for 1999, the market value of the Fund at the end of
2002 will be about NOK 600 billion, or USD 80 billion. However, this
figure is based on a substantially higher oil price than the current
level. Moreover, the markets will probably have risen in value, so that
the Fund’s shares in the market can be expected to remain very small.
Nevertheless, the Petroleum Fund is so large in relation to the liquidity
in many of the individual shares in which it has invested that the
possibility of influencing share prices must be taken into account in the
management of the Fund.
The basis for recommendations concerning equity
investments
When the first allocations to the Petroleum Fund were made in 1996, the
Fund’s entire capital was invested in bonds and other fixed-income
instruments, according to roughly the same guidelines that applied to the
foreign exchange reserves. In the course of 1997, however, considerable
capital was allocated to the Petroleum Fund, and in a submission to the
Ministry of Finance, Norges Bank raised the issue of investing some of the
Fund’s capital in equities. This submission was published as an annex
to the Revised National Budget for 1997, in which the Ministry of Finance
also proposed that investments in equities be permitted. The Storting
endorsed the proposal.
The point of departure for Norges Bank’s submission was that the
size of the Petroleum Fund made it improbable that its capital would be
used in the next decade. Consequently, when deciding how to invest the
capital, the return over a time horizon of at least ten years should be
considered. With such a long investment horizon, Norges Bank believed that
equity investments would, with a high degree of probability, provide a
higher return than investments in fixed-income instruments. This was
supported by a study carried out by the G-10 countries, with return
figures for the past two decades, and a study of equity and bond markets
in the US since 1871.
It was also pointed out that, calculated over five-year periods, the
return on portfolios consisting of both equities and fixed-income
instruments has varied less than the return on portfolios consisting only
of securities from one of these asset classes. Fixed-income instruments
have often provided a solid return when the return on equities was low,
and vice versa. Such diversification gains have been achieved for equity
portions of up to 50 per cent. The underlying data for these calculations
were from the UK, for the period 1926-1996.
The recommendations on equity investments were thus based on historical
material that shows a higher average return on equities than on
fixed-income instruments, and less fluctuation in five-year rates of
return on mixed portfolios. The former finding, in particular, has been a
controversial issue in the economic literature of recent years.
Will equities provide a higher return than bonds?
Jeremy Siegel’s book, "Stocks for the Long Run" (Irwin,
1994), which compares the return on equity investments with the return on
investments in fixed-income instruments in the period 1871-1992 in the US,
has made an important contribution to the recent debate. Siegel found
that, for the period as a whole, equities provided an average annual
return that was almost 7 percentage points higher than the return on
fixed-income instruments. Looking at individual decades, equities provided
the highest return in eight of ten cases. Siegel documents that for an
investor who is solely considering long-term returns, a diversified equity
portfolio has been the best choice as a rule.
Critics immediately pointed out that this result only applies to the US.
The US and the UK are the only two major stock markets in the world that
have functioned without breaks of any length during this century. In other
stock markets, shareholders have in periods incurred major losses as a
result of war, nationalisation or a collapse of the financial system. The
same applies to some extent to owners of bonds and other fixed-income
instruments. Looking at one of the two markets that have avoided such
interruptions may nevertheless provide a distorted picture of what can be
expected to happen in stock markets in the future, in both the US and
other countries. When Siegel’s book was published, there were no
comparable studies on other countries’ markets.
Chart 1: Annual returns on stocks, bonds and bills in the
US: Moving 30 year averages. 
Source: Siegel, "Stocks for the Long Run"
Chart 2: Average real returns on Global Stock Markets 
Source: Goetzmann and Jorion, "Global Stock Markets in the Twentieth
Century"
However, William N. Goetzmann and Philippe Jorion recently made a study
of stock markets in 39 countries in the period 1921-1996 ("Global
Stock Markets in the Twentieth Century", to be published in the
Journal of Finance in 1999). The study covers virtually all the stock
markets for which data exist for at least parts of the period 1929-1970.
The countries covered include Japan and Germany, whose stock markets were
interrupted for long periods in the 1940s. The data also cover a number of
countries in Central and Eastern Europe which had Communist regimes after
the Second World War.
Goetzmann and Jorion look at the real rise in share prices, over and
above the rise in world market prices, that would have been achieved by
buying a diversified equity portfolio in each of these markets in 1921 (or
later in cases where there are no data for 1921). They have no figures on
dividend payments, and can therefore not calculate return figures. The
median for the annual real rise in prices up to 1996 is only 0.8 per cent,
and in many countries share prices have risen less than prices in the
world market. For the US they find that real share prices in this period
rose at an annual rate of 4.3 per cent, which is higher than in any other
country. This illustrates the fact that experience in the US is not
representative of the world’s stock markets.
However, if an investor in 1921 had spread equity investments in all the
countries included in the data, using GDP as the weight for each country,
the real price rise of the portfolio would have been 4 per cent a year.
The stock markets in the countries that were wealthiest in 1921 were the
same markets that provided the best return later. In this respect the
study provides some support for the guidelines concerning the distribution
of equity investments that have been established for the Petroleum Fund.
There are two possible reasons for regarding both Siegel’s and
Goetzmann and Jorion’s studies with scepticism. First, history is
only a good guide to the future if we can assume that the processes that
have generated the historical data are the same processes that will be
important in the future. This is a general problem with empirical studies
that can never be entirely ignored. But if no major changes take place in
the manner in which the return on equities and fixed-income instruments is
generated, historical data form the best available basis for arriving at
expectations of what will happen in the future.
Second, the available return series are fairly short considering that
the objective is to provide some indication of the expected return over
ten-year periods. Return figures are not available for so many independent
ten-year periods, and the results we obtain thus have little statistical
significance. This is a real objection, which underscores the uncertainty
of expectations.
However, there are sufficient empirical data from stock and bond markets
to enable us to conclude with a fair degree of certainty that the annual
return on equity investments fluctuates more than the annual return on
investments in fixed-income instruments. This is also because the return
to shareholders is a residual, ie the income remaining after an enterprise
has serviced its debt and met other obligations. It can furthermore be
assumed that investors will generally demand a risk premium in the form of
a higher expected annual return if they are to be willing to buy a
security with sharply fluctuating rates of return.
It follows that most investors will only buy equities if they expect
investments in equities to provide a higher rate of return than
investments in fixed-income instruments. If this condition is not met,
there will be few buyers in the stock markets, and prices will inevitably
fall. Naturally, buyers’ expectations may later prove to have been
incorrect, but systematically wrong expectations cannot last; few
investors will continue to believe in something they see is not happening.
In a normal situation, equities must therefore have a higher long-term
return than fixed-income instruments. This follows from undisputed
empirical results showing greater fluctuations, coupled with generally
accepted assumptions about rational investor behaviour. This thereby
enhances confidence in the above empirical data, which are consistent with
the conclusions arrived at by theoretical reasoning. But there is reason
to stress that the long term in this line of reasoning may well be more
than the 10-20 years that have been examined in the empirical data.
The equity index for the Petroleum Fund - The FT/S&P Actuaries World Indices were established in 1987,
and are constructed by FTSE International, Goldman Sachs & Co
and Standard & Poor’s, in collaboration with the Institute
of Actuaries and the Faculty of Actuaries. There are currently 29
countries, and around 2 400 shares, included in this world index,
and of these, 21 countries and around 2 040 shares are included in
the Petroleum Fund’s benchmark index.
- The FT/S&P indices aim to cover around 80 per cent of the
value of all listed shares in every country, excluding equities with
inadequate liquidity, shares which foreigners may not own and
long-term government shareholdings in companies. The choice of
companies to be included in the index is made by a committee,
according to relatively clear guidelines. The object is that the
index should be representative of the market with respect to
sectoral composition, so that the index does not necessarily consist
only of the largest companies in each country.
- The daily price return is calculated by dividing the value of
all the shares in the index at the close of the stock exchange by
the value at the close of the previous day, after adjusting for
changes in the composition of the index. This is done for each
country and each sector within each country, for regions and for the
index as a whole. In addition, the total return is calculated by
adjusting the return for an estimated annual dividend payment
distributed evenly over the year. Since dividend payments show
seasonal variations, the total return on an actual portfolio
composed according to the index will follow a different course from
the reported total return on the index. Taxes on dividends are
excluded when the total return is calculated.
- There are also other world indices. The MSCI World Index,
administered by Morgan Stanley Capital International, has excluded a
number of smaller companies, so that the index only represents 60
per cent of the market value. The Salomon Brothers Broad Market
Index covers a full 95 per cent. The MSCI index is used most by
American investors, while the FT&S&P index is most popular
with European investors.
- The composition of the FT&S&P World Index at 31 December
1998, including those countries that are not included in the
Petroleum Fund’s investment universe:
| | Country | No. of equities | Share |
|---|
| The Americas | | * | Brazil | 32 | 0.28 % |
| Canada | 117 | 1.87 % | | * | Mexico | 29 | 0.28 % |
| US | 619 | 53.19 % | | Europe and Africa |
| Belgium | 22 | 0.91 % |
| Denmark | 34 | 0.38 % |
| Finland | 29 | 0.57 % |
| France | 75 | 3.73 % | | * | Greece | 36 | 0.22 % |
| Ireland | 16 | 0.24 % |
| Italy | 53 | 2.30 % |
| Netherlands | 26 | 2.65 % | | * | Norway | 37 | 0.13 % |
| Portugal | 18 | 0.22 % |
| Spain | 30 | 1.34 % |
| UK | 207 | 10.00 % |
| Switzerland | 30 | 3.10 % |
| Sweden | 44 | 1.10 % | | * | South Africa | 38 | 0.35 % |
| Germany | 55 | 4.26 % |
| Austria | 21 | 0.11 % | | Asia and Oceania |
| Australia | 76 | 1.16 % | | * | Philippines | 22 | 0.04 % |
| Hong Kong | 67 | 1.07 % | | * | Indonesia | 26 | 0.04 % |
| Japan | 445 | 10.13 % |
| New Zealand | 18 | 0.09 % |
| Singapore | 41 | 0.20 % | | * | Thailand | 35 | 0.03 % |
* Not inluded in the investment universe of the Petroleum Fund. • The total market value of all companies included in the
index is NOK 145 000 billion. |
Building up the equity portion of the Petroleum Fund
At the beginning of 1998, the Petroleum Fund’s capital was invested
in its entirety in fixed-income securities, but during the next five
months equities were purchased. The equity portion thus increased from
zero to about 40 per cent, which is the intended level as long as the
current guidelines apply.
Equity prices fluctuate considerably over time, and it is difficult to
know whether they are unreasonably expensive or especially cheap at any
given time. When purchasing equities, there is always a risk of paying
such a high price that the return is low at a later time. The risk can be
reduced by spreading equity purchases over time, with purchases at
different price levels. This reasoning may suggest that the process of
acquiring the equity portion of the Petroleum Fund should have extended
over a longer period than five months.
Such a conclusion would be reasonable if the size of the Petroleum Fund
could be expected to remain fairly constant. But the projections presented
by the Ministry of Finance in its Long-Term Programme, and which applied
in 1997 when the management strategy was changed, show new transfers of
capital to the Petroleum Fund for at least 15-20 years ahead. The Fund
will then be considerably larger than it was at the beginning of 1998. If
the equity portion is to remain at around 40 per cent, further equities
must be purchased every year for a long period. The equity purchases made
in the first half of 1998 were not much larger than the quantities that
will probably be purchased later within equally short periods of time.
This means that a rapid increase in the equity portion while the
Petroleum Fund is relatively small will help to provide precisely the
spread of equity purchases over time that can reduce the risk that the
total equity portfolio is purchased at too high a price. It would have
been considerably more risky to postpone the building up of the equity
portion until the Fund had grown larger, because it would then have been
necessary to concentrate the equity purchases over a shorter period (cf
the National Budget for 1998).
The bond index for the Petroleum Fund - Salomon Smith Barney’s World Government Bond Indices were
established in 1981 and now contain 19 countries. Indices are
published both for each country and for groups of countries (eg the
global index, the world excluding the US and the euro area).
- Government bonds in markets that satisfy certain minimum
requirements regarding size and capital mobility are represented in
the index. In addition, each issue must be of a certain minimum
size. Each bond and each market is represented in the index
according to its market value (the market capitalisation principle).
- The index is reconstructed every month. New issues are registered
daily, and must be dated no later than the 25th of the month prior
to implementation. They must also be publicly announced, ie
announced with all relevant information, before the 25th.
Participation in the index by an individual bond is contingent on a
minimum issue (USD 1bn for the US). The intention is that there
should be bonds available to investors under normal market
conditions.
- The index contains only government bonds, and may not contain
bonds that are difficult for potential investors to buy. Special
costs associated with individual issues, such as withholding tax,
pledging of collateral etc. must be announced and easy for investors
to predict. The index must not be changed often, and changes must be
easily predictable.
- The index contains only bonds with a maturity of more than a
year.
- The world index and the various currency indices are priced
daily. Even more detailed return figures for sub-segments of the
country indices are also provided on a monthly basis. Bonds are
priced at offer price, and mainly at the close of each market.
- The total return is calculated on the assumption that bonds are
bought at the beginning of a period and sold at the end. The basis
for calculating the total return is price changes, principal
payments, coupon payments, accrued interest and reinvestment of
coupon income. Adjustments for exchange rate effects are made using
prices quoted in London at 5 pm each day (Reuters closing spot
rates, Reuters page WMRH).
- Daily, weekly, monthly and annual calculations of returns are
published on Salomon Smith Barney’s web pages, and in monthly
publications from their index group. Figures for the various
maturities are also presented.
- Salomon Smith Barney has appointed special teams in London and
New York to maintain the indices, monitor all relevant instruments
with respect to criteria for inclusion, solicit prices, publish
return figures on the world index and its various components, and be
responsible for providing service to users of the index.
- Salomon Smith Barney is not alone in offering a recognised global
bond index, but is one of the leading suppliers of indices. The
indices are used today by a large number of international managers
to measure their own return against a benchmark. Since the Salomon
Smith Barney World Government Index includes so many countries, it
is very suitable for global management.
|
The portfolio composition of other large
international funds
The distribution 40/60 between equities and fixed-income securities in
the benchmark portfolio for the Petroleum Fund was chosen as a compromise
between higher expected returns on the one hand, and wider fluctuations in
the annual rate of return on the other. Both effects accompany an increase
in the equity portion: wider annual fluctuations are the price of a higher
expected return. Other investors face the same need to strike a balance
when they decide on the composition of assets in their portfolios.
The Petroleum Fund is an institutional investor, like investment funds,
pension funds and life insurance companies. The OECD publishes annual
statistics on institutional investors, including asset distribution
figures. The most recent figures are from 1995, and show substantial
differences among OECD countries. The equity portion of investors’
portfolios varies from an average of 12 per cent in Germany to 69 per cent
in the UK, with the bulk in the interval from 20 to 40 per cent equities.
In making comparisons with the Petroleum Fund, it should be remembered
that most other institutional investors invest some of their capital in
property, which shares many features with equity instruments.
It is also important to remember that most institutional investors have
definite commitments, which influence what kind of portfolios they want.
For example, pension funds have to pay pensions to their customers
according to rules laid down in pension contracts. A pension fund can
therefore calculate in advance how much capital must be available at a
given point in time, and can develop an investment strategy which entails
little risk that sufficient capital will not be available. In principle,
pension funds can completely eliminate this risk by buying fixed-income
instruments that mature at the time the capital must be available. This
requirement suggests that pension funds should have a substantial
portfolio of fixed income instruments, at the expense of other asset
classes with a more uncertain return.
Chart 3: Asset composition at institutional investors in
1995
Source: OECD, Institutional Investors Statistical Yearbook, 1997
Taking into account the above considerations as well as economic policy
goals, the governments of some countries have issued rules that restrict
the freedom of pension funds and life insurance companies to choose the
composition of their portfolios (see E.P. Davis, "Regulation of
Pension Fund Assets", in Institutional Investors in the New Financial
Landscape, OECD, 1998.) German pension funds, for example, may not invest
more than 36 per cent in equities, and the limit in Japan and Switzerland
is 30 per cent. Norway’s Nordic neighbours, Denmark and Sweden, have
limits of 40 per cent and 50 per cent respectively for asset classes that
include more than equities. The limit for Norwegian life insurance
companies and pension funds is 35 per cent.
The nature of commitments can also influence the composition of
portfolios in countries without such rules. For example, pension funds may
have commitments based on either defined-contribution or defined-benefit
schemes. In the first case, disbursements to pension customers depend on
the realised return on the capital in the funds. In the second case, the
disbursements follow rules that are independent of the realised return. A
study of US pension funds shows what can be expected: pension funds with
defined-contribution-based commitments have higher equity portions than
those with defined-benefit-based commitments (Eckbo in Annex 3 of NOU
1998:10)*. However, both types of funds had on average more than 50 per
cent in equities in 1996.
The Petroleum Fund is an investment fund without commitments as clearly
defined as those of pension funds or life insurance companies. The capital
in the Fund is to be used to finance deficits on future government and
social security budgets, but it is uncertain when such needs will arise.
Discussions of the purpose of the Petroleum Fund have often focused on the
needs that will arise when the large post-war generation reaches
retirement age in 15-20 years, but there are also many other factors that
may influence government budget balances. The need to draw on the
Petroleum Fund for the first time may arise either earlier or later than
around the year 2015.
Chart 4: Composition of two North American petroleum
funds, taken from their annual reports. The left-hand column shows the
benchmark composition of Alaska Permanent Fund Corporation, and the
right-hand column shows the benchmark for the asset composition of the
Alberta Heritage Savings Trust Fund
Source: Annual Reports
It is reasonable to maintain that the objective of a high return on the
Petroleum Fund, to a greater degree than for pension funds and life
insurance companies, is to apply primarily with a time horizon of more
than 10 years. Consequently, the Petroleum Fund should have a higher
proportion of asset classes with a high expected return than these other
funds, because less importance should be attached to the annual
fluctuations in the return in the case of the Petroleum Fund.
There are few other funds with the same type of purpose as the Petroleum
Fund. Closest are probably two petroleum funds in North American: Alaska
Permanent Fund in the US, and Alberta Heritage Savings Trust Fund in
Canada. The return on these funds is intended to benefit the inhabitants
of these states in various ways, but there are no specific obligations
attached.
The equity portion of the Petroleum Fund is thus smaller than those of
two comparable North American funds. It is also smaller than the average
US or British pension fund, but a little larger than the average pension
fund in continental Europe. This latter, however, is partly because
European pension funds are subject to restrictions that are not always
based on the interests of pension customers.