Eurostat Task Force on Subsoil Assets

B1 - National accounts methodology,

statistics for own resources

Luxembourg, September 1999

Eurostat Task Force on Subsoil Assets

Summary of Conclusions and Results

Revised version

Eurostat B1

Note: this revised version takes into account comments and additional inputs provided by Mr. Todsen (N) and Mr. Van de Ven (NL). Revisions affect section 2.2.4, section 2.4 and the table 3 in annex.

1

1Background and objective

1.1Background

In 1996 Eurostat started a Task Force in order to examine the issue of subsoil assets in the context of environmental accounts. The Task Force met four times:

March 1996

November 1996

January 1998

June 1999

Norway, France, the Netherlands and the UK participated in the four meetings; Austria joined the Task Force in January 1998.

The Task Force focused on valuation of reserves more than on broader environmental accounting issues: hence, the issues of sustainability of extraction at the world level, as well as emissions related to the use of fossil energy consumption and extraction were postponed or referred to other groups. The Task Force developed its work in four main directions:

definition of the reserves to be taken into account,

definition of the resource rent,

calculation of the value of reserves,

and ownership issues.

Characterisation and treatment of “depletion” in national accounts was not explicitly discussed, but left to the London Group and to the National Accounts Working Party.

At the January 1998 meeting a set of tables was agreed upon and some volunteer countries carried out test exercises.

1.2 Objective of the meeting

The purpose of the June 28th 1999 final Task Force meeting was

to present and discuss the results of the test exercises,

to review the work done and to draw final conclusions concerning the issues discussed,

to agree on the way results should be disseminated.

2Summary of meeting discussions and conclusions

2.1Volumes and physical tables

2.1.1Inclusion of reserves and their weightings

Questions raised

It had previously been decided by the Task Force to include in the calculations both proven reserves, as recommended by SNA 1993, probable and possible reserves, as well as undiscovered reserves. Reason was that due to the high costs for proving new reserves, particularly in the North Sea, oil firms only prove the volume necessary for a limited extraction period, typically 5 to 10 years. Therefore the volume of proven reserves is not representative of the overall volume of reserves present on the economic territory of a country.

Conventionally, hydrocarbon deposits are categorised, on the basis of the stage of exploration they have undergone, as proven, probable, possible and potential / undiscovered. This corresponds to the expected proportion of reserves that can be recovered. The SORP (Statement of Recommended Practices) of the OIAC (Oil Industry Accounting Committee) definitions are the followings:

Proven reserves: the estimated quantities of crude oil, ... which geological, geophysical and engineering data demonstrate that there is a 90 per cent statistical probability that the actual quantity of recoverable reserves will be more than the amount estimated as proven, and a 10 per cent statistical probability that it will be less.

Probable reserves: the estimated quantities of crude oil, ... which geological, geophysical and engineering data demonstrate that there is a 50 per cent statistical probability that the actual quantity of recoverable reserves will be more than the amount estimated as probable, and a 50 per cent statistical probability that it will be less.

In a discussion paper of the OIAC, when discussing the use of only proven reserves in calculating the value of discovered reserves of oil and gas, authors state that "thus instead of including only proven reserves, the calculation would be based on proved and probable and might even give some recognition to possible reserves. Due weight could be given in the calculation to the different probabilities associated with each of these categories, e.g. 50% of probable and 10% of possible reserves". [1]

Thus, on January 1998 the Task Force decided that when the volumes of reserves in the various categories are raw volumes they should be weighted by their probability of being recoverable, before being added and accounted for in balance sheets. The SORP probabilities could be used as default average values elsewhere, so that, for example, probable reserves are weighted with value 0.5.

Results of the test exercises and conclusions

Two countries reported on raw volumes. For weighting reserves a 0.9 weight was applied to proven reserves, which resulted in some discrepancy in the tables, due to the fact that extraction is known with certainty. The position of the Task Force was that proven reserves should be weighted with probability 1.0. Any revisions to proven reserves need to be explicitly noted.

In the test exercises, three countries did not report on the raw (non-weighted) quantities but, due to data availability, disclosed already weighted data (e.g. Norway, Netherlands). In this case there is of course no need for weighting. In order to ensure comparability of data, it was asked to these countries to investigate about the weights used for obtaining weighted reserves.

It was considered in earlier meetings that some deposits would be known, but would be sub-economic or non-economic with present prices and technology. However, it appeared from pilot exercises that, even for the physical tables, the countries were not able to report on the categories sub-economic and non-economic reserves; as a consequence, it was impossible to distinguish between the different categories of changes: all changes were grouped together under the item: "other changes in volume"[2]. Therefore the Task Force decided that these categories should be excluded from the tables.

Furthermore, although the oil price, and hence the resource rent, experienced large fluctuations during the period under review, in particular with the sharp decrease in oil price from 1985 to 1988, which should have resulted in the decrease of economic reserves, the test exercises showed no significant changes in the volume of reserves classified as economic[3]. See graphs 1 to 6 in annex.

As a conclusion, the participants agreed that the yearly tables proposed for the physical description of reserves should be simplified, no reference being made to sub-economic and non-economic reserves. Flows should be restricted to extraction and other changes in volume. This last category would group together: discoveries, revisions of previous estimates, and changes due to changes in price and classifications; when available, discoveries should be distinguished.

It was also agreed that National Accounts departments would not normally calculate the volumes and that these would have to come from energy statistics departments, petroleum directorates, or directly from oil companies.

2.1.2Expected lifetime of reserves.

Questions raised

The conventional measure of the expected lifetime of reserves is both a widely used indicator of the status of reserves (how many years the reserves would allow the present extraction to continue) and an important element in the calculation of the value of reserves.

Because future extraction is discounted over time, the path or profile of extraction has a potentially considerable impact on the present value of reserves. This is more important for the longer life reservoirs, thus for gas. The expected lifetime of existing reserves is generally based on the extraction of the current year. There are three ways to estimate the lifetime of reserves on the basis of current year information:

(a) stock at the beginning of the year / extraction of the year

(b) stock at the beginning of the year plus appearances / extraction of the year

(c) stock at the end of the year/extraction of the year

Results of pilot exercises and conclusions

a) Incorporation of weighted non-proven reserves

Incorporation of the weighted non-proven reserves was shown to extend the lifetime of reserves without changing its profile (see graph 4).

The Task Force considered the graph depicting the ratio of reserves to current extraction may be misleading, as this ratio results of two different factors: the rate of extraction and the additions to reserves. Possible alternative representation were considered that show the absolute values of extraction and "additions to reserves" (see graphs 5, 6).

b) Calculation of the lenght of life

The Task Force considered that, as the objective is to value the closing stock of reserves, method (c) is consistent in its use of the information available when balance sheets are compiled. However, the choice between (b) and (c) issue was not considered to be of primary importance.

The Task Force also discussed an alternative method where the stock is divided not by the extraction of the year but by [extraction less additions]. With this method the expected lifetime of reserves increases considerably. The Task Force rejected this method, not just for the practical problems that may appear (need for some averaging, situations where additions are higher than extraction …) but also because it would mean to value volumes that are larger than those reported in physical balances – above that part of undiscovered reserves which is already included through the weighting procedure. However this method could be used for the assessment of depletion (see 2.3 below). See also graph 7 in annex.

Finally the Task Force decided that:

Where production forecasts exist this data should be incorporated to give a more realistic future extraction profile.

When no additional information is available on the future path of extraction, the use of a constant extraction profile is recommended.

2.2Resource rent

2.2.1Definition of the resource rent

Following the recommendations in the SNA, the Task Force previously chose the present value method as the basic method for the valuation of reserves This requires the series of future receipts corresponding to the extraction of the subsoil asset, as well as the choice of a discount rate.

The Task Force had previously also agreed on the following definition of the net resource rent: "net operating surplus plus specific taxes less the return to fixed capital"[4]. When this calculation gives a negative value, the net resource rent has to be set equal to zero.

Government can appropriate part of the rent in the form of royalties or specific profits taxes (i.e. after NOS) or as other taxes on production (i.e. before NOS of the extractor). In the above definition “specific taxes” refer to those "other taxes on production" which are specific to the extracting industry. The OIAC defined these taxes as "allocable taxes", i.e. "taxes which can be specifically identified with oil and gas production activities"[5].

Addition of specific “other taxes on production” provides a measure of the full resource rent derived from national reserves, whatever the units/institutional sectors (non-financial corporations or general government) that appropriate the rent.

Value of reserves by present value of the net resource rent

PV(NOS* -K)

NOS* = net operating surplus plus specific other taxes on production

K = net stock of fixed capital (inc. exploration expenditure)

 = rate of return to fixed capital

Two questions emerge as concerns the definition of the resource rent:

a) the rate of return to fixed capital to be used

b) an alternative approach by the Netherlands

2.2.2Rate of return to fixed capital

Questions raised

In order to calculate the net resource rent as defined above, the return to capital has to be estimated. Therefore an assumption as concerns the rate of return is necessary. The January 1998 Task Force had agreed that the indicative range for the real rate of return to fixed capital would be 8 to 10%.

During the September 1998 OECD meeting on capital stock measurement it was stated that the level of 8% for the rate of return to fixed capital was "unrealistically high".

Results of pilot exercises and conclusions

After having examined empirical data on the ratio that the net operating surplus bears to the net capital stock for manufacturing industry as a whole (see table 1 in annex), as well as national standards, the Task Force concluded that, for European Economic Area countries, an 8% rate of return on fixed capital should be taken as the default value in the absence of more detailed information. It was also suggested that a sensitivity analysis should be conducted around the 8% value.

2.2.3Alternative for the assessment of the resource rent

The Netherlands considers that the auction for exploitation rights is fully competitive and that it therefore precludes any significant part of the resource rent being appropriated by extracting firms. Under this view, extracting companies receive a “normal” rate of return to capital and the General Government appropriates the whole resource rent.

For European countries, where North Sea oil and gas are owned by governments, the Netherlands’ hypothesis would permit the valuation of oil and gas reserves directly as the present value of the Governments’ oil and gas related revenues. Empirically, this valuation could appear to be sufficiently accurate owing to the considerable uncertainties that affect other methods and would also result in some implicit "smoothing" of the resource rent.

In particular, this method would avoid having to make an assumption as concerns the rate of return to fixed capital. It would also avoid a negative resource rent. Further, this hypothesis permits a valuation to be arrived at if data on net capital stock and consumption of fixed capital are not available but the receipts of general government from specific taxes and/or royalties are known.

Results of pilot exercises and conclusions

Pilot exercises show that for Norway[6], over the period 1984-1996, the proportion of resource rent appropriated by the extractor was an average of 12% of the total resource rent (see graph 8 in annex). However, due to changes in the oil price, this proportion is very variable. Resource rent for the extractor was near zero over the period 1988-1995[7]. Resource rent for the extractor was, in the contrary, highly positive in 1983-85. For the year 1996, with oil prices increasing relative to previous years, the proportion of net rent appropriated by the extractor was 38% in Norway (but only 10% in UK for the same year).

Components of the resource rent

(1996 – Mn Ecu)

A counter-argument for the Dutch method was that some governments have followed policies of permitting extraction companies to retain a part of the resource rent, e.g. because they are public and thus there was no competition to eliminate the extractor’s part of the rent, which fully accrues to government in any case; because of a policy decision to overcome uncertainty in exploitation; or because of a policy decision to ensure energy security.

Moreover, the CBS method can only be applied for these countries where the general government is the legal owner of the subsoil assets, and therefore entitled and in position to appropriate the whole resource rent. Also, in many countries outside of Europe, the existence of a positive resource rent for the extractor is highly probable.

It was suggested that the Netherlands undertake a test of the hypothesis on their own data. It was also suggested that the Netherlands hypothesis be tested by the simulations around the 8% return to capital.

The suggestion remains that, should the Netherlands’ hypothesis be verified by these tests, it could be recommended that this method is considered a second best method of calculation of the resource rent in appropriate cases, when data on net capital stock and consumption of fixed capital are not available.

2.2.4Distribution of total resource rent between oil and gas

Question raised

In National Accounts figures, production costs of the oil industry are generally not divided between oil and gas. In some cases the oil and gas wells are physically distinct and separate data can be compiled but often this is not possible. As a consequence, except for the UK and to some degree for France, the data supplied by the countries only permitted calculation of the total resource rent for oil and gas.

For the calculation of the oil resource rent Norway has divided the costs (and thus the resource rent) in proportion to the output value. In Norway, some platforms produce both oil and gas, others are pure oil or gas producers. The industry also has onshore establishments. For each platform data are available as concerns intermediate consumption, wages and investments. These may be used to distribute the production costs of the extraction industry roughly between oil and gas - distribution of taxes seems more doubtful.. The Netherlands would also be able to provide the level of information required.

It was noted that, as government receipts from oil and gas extraction are generally separately identified, the Dutch method would allow to directly know the resource rent for oil and gas separately.

2.2.5Volatility of resource rent

Question raised

As previously noted, the resource rent per unit is highly volatile, amplifying the movements of oil price (see graph 9 in annex). It is generally recognised that there is no identifiable cycle in prices of oil or gas and by extension in the value of the rent.

Results of pilot exercises and conclusions

In order to provide a consistent basis for the valuation of the future stream of resource rent, the Task Force accepted the need to develop an expected resource rent per unit, imposing some "smoothing" or averaging technique. However consistency with yearly National Accounts has to be carefully examined.

Along the same lines, when data show that the per unit resource rent drastically changed from one year to the following, it may be considered to use a "sliding" per unit resource rent (see graph 12 in annex)

2.3Calculation of the value of reserves

2.3.1The rate of discount

In discounting the value of the future stream of resource rent, after a discussion as concerns the meaning of the rate of discount for government's receipts, and considering that the rate of discount should be considered as a "social" rate of discount, a central value of 4% was judged acceptable by the Task Force: in their calculations Norway used 3.5%, the UK 5.0%, the Netherlands 4% and France 5%.