Note No. 181 April 1999 The 1996­97 Gas Price Review in Argentina Andres Gomez- Argentina's natural gas industry was privatized at The regulatory framework Lobo and the end of 1992. Prior to divestiture, the state- Vivien Foster owned monopoly Gas del Estado was divided The tariff paid by final gas consumers in Ar- into two transport and eight distribution compa- gentina is composed of three parts: nies, all of which were sold through international bidding. Final price = gas wellhead purchase price + transport margin An independent government body, Ente Naal + distribution margin Regulador de Gas (Enargas), was established to regulate the transport and distribution seg- Gas wellhead purchase prices are not regu- ments of the industry. As part of its mandate, lated, but are determined by the contracts Enargas is in charge of price reviews. These negotiated between gas suppliers and pro- reviews, which occur every five years, deter- ducers. Purchase costs are passed through to mine the allowed tariffs for each transport and final consumers, subject to Enargas's approval distribution company. The first such review of the prices as reasonable. There is, however, took place in 1996­97, and the new tariffs no formal mechanism to promote efficient went into effect in January 1998. This Note ex- purchases. amines the methodology and outcome of this experience. Price margins for transport and distribution are set by Enargas for five-year periods. The price The Enargas price review is of interest for sev- control system is similar to the price cap regu- eral reasons. First, it is the first of its kind in lation used in the United Kingdom. But unlike Argentina, and one of the first in a developing U.K. firms, Argentine firms do not have flexibil- country. The outcome of the process provides ity in setting individual prices subject to an a test of the regulatory framework adopted by aggregate revenue or tariff basket constraint. the Argentine government, and may influence Enargas sets the maximum tariff for all individ- regulatory reform in other parts of the world. ual services and customer categories. These tar- Second, as the first such event, the 1996­97 iffs must be sufficient to: review set a precedent for methodologies and Cover operating costs, taxes, and depreciation. approaches to be used in future utility price Provide a reasonable rate of return on invested reviews in Argentina. Finally, the approach capital. used to calculate the cost of capital--as well as Guarantee a secure supply by providing the other parameters used to set prices-- resources to fund system maintenance and provide an interesting illustration of how theo- expansion. retical and practical methods from regulatory practice in industrial countries can be adapted Tariffs are automatically adjusted every six to developing countries, where data availability months according to: and other restrictions prevent a direct transfer of techniques. (1) g = PPI ­ X + K, The World Bank Group Finance, Private Sector, and Infrastructure Network 2 The 1996­97 Gas Price Review in Argentina where g is the percentage change of the tariff, each tariff, and different types of outputs. Another PPI is the producer price index in the United simplification in the above formula is the omis- States,1 X is the efficiency factor, and K is the sion of taxes. Cash flows should be net of taxes, investment factor, both of which may differ since it is the posttax income that is relevant for across firms. The last two parameters may also the valuation of companies. differ for each six-month semester of the five- year period. To give investors a fair rate of return on their invested capital, the regulator should set g in The efficiency factor (X ) reflects the cost reduc- equation 1 (and therefore tariffs Tt) in such a way tions that the regulator estimates can be achieved that the NPV is equal to the capital invested in in the next five-year period, which are thus the firm at the beginning of the review period. passed directly on to customers. The investment This initial investment was set equal to the price factor (K ) is an adjustment to allow revenues to paid for the companies at the time of divestiture cover expected investments in improving and rolled forward to the review date by adding the expanding network infrastructure. Company new investments made during the interim period investment plans are first screened by Enargas. and subtracting depreciation. Once approved, they are assigned K factors. The K factors are project-specific and are contingent There is a further issue in determining the time on the investment being undertaken. They come profile of the tariff changes: changes could be into effect only after a project has come on stream gradual--through small changes each semester-- and is delivering benefits to customers. or made all at once at the beginning of each five- year review period. Enargas decided that the full Setting the new tariffs price reductions would occur on January 1, 1998. Thus the first X factor was set to achieve this tar- Estimates of future cash flows for each firm are iff reduction, while the X factor for each subse- the starting point for the setting of new tariffs. quent biannual price adjustment was set to zero. The cash flow analysis should extend until the This is in contrast to K factors, which are imple- licenses expire, since this is the relevant time mented gradually, as noted above. horizon over which owners can recoup their investment. At the time of the review, current Information requirements licenses were set to expire in thirty-five years. Estimates of the growth in demand (that is, pre- The net present value (NPV) of a firm's future diction of the parameter V in equation 2) are cru- cash flows is: cial for determining appropriate tariff levels. Enargas used information provided by other gov- 35 ernment departments (such as the Secretaría de (2) NPV = T t t V - Ct - It, Energía y Puertos) and the gas companies them- t=1 (1 ) t + r selves. The available information was used to con- struct probable demand scenarios for each firm. where Tt is the tariff for transporting (or distrib- uting) gas in period t, Vt is the volume of gas Detailed information on operating costs and transported (or distributed), Ct is the operations revenues, as well as planned investments, was and maintenance cost, It is additional investment, submitted to Enargas by each regulated firm ac- and r is the firm's cost of capital. cording to a format provided by the regulator (Enargas 1996b). In the actual modeling of cash flows, T is a vec- tor of tariffs and V is a vector of outputs, since Productivity increases were considered by the there are different tariff zones, different parts to regulator when forecasting costs and, therefore, The World Bank Group 3 TABLE 1 DEBT AND EQUITY IN THE ARGENTINE GAS INDUSTRY, 1996­97 US$ millions setting the efficiency (X ) factor. The exact Activity Debt Equity Debt to equity ratio approach used by Enargas is described below. Investments that expand the network will Transport 874 1,512 0.58 increase the volume of gas sold and will also Distribution 1,015 2,403 0.42 affect future cost estimates. These activities are considered when setting the investment (K ) fac- Source: Enargas 1996a, table 7. tor and are also described below. TABLE 2 FINAL COST OF CAPITAL FOR THE ARGENTINE The cost of capital GAS INDUSTRY A fundamental parameter that must be estimated Percent before tariffs can be set is the cost of capital-- that is, the cost to a firm of raising an additional Nominal Expected Real unit of capital. Enargas estimated separate rates Activity capital cost inflation capital cost for the transport and distribution sectors but did not attempt to differentiate these rates across Transport 13.4 1.9 11.3 individual companies. Distribution 15.2 1.9 13.1 Source: Enargas 1996a. The aggregate cost of capital is a weighted aver- age of the cost of debt capital and the cost of equity capital. The weights are the portion of debt and equity relative to company assets. return investors will demand to hold the asset. Formally, Calculating this risk premium for a regulated util- ity is a necessary but contentious aspect of any price review. A widely used tool in this endeavor E (3) rc = rd 1- t ( ) DV + re , is the capital asset pricing model. V Use of the capital asset pricing model is problem- where rc is the cost of capital, rd is the nominal atic in countries where stock markets are under- interest charged on the firm's debt, t is the profit developed or where the industry under analysis tax rate faced by the firm (0.3 in Argentina), D is has not historically been quoted on the stock the debt of the firm, re is the opportunity cost of exchange. This was the case in Argentina. Only equity capital, E is the value of the firm's equity, two of the ten gas firms had been quoted on the and V = D + E. Debt and equity are measured by stock exchange at the time of the review, and even their market value. Where a firm's debt or shares for these companies, the time series for stock mar- were not traded in the market, information from ket data was rather short. Recognizing these obsta- financial statements was used. The equity and cles, Enargas used the following adjusted capital debt figures for the Argentine gas industry at the asset pricing model: time of the review are presented in table 1. The final nominal and real capital costs for transport (4) re = rf + be (rm ­ rf) + riskARG, and distribution, as calculated by Enargas, are presented in table 2. where re is the cost of equity for an Argentine firm, rf is the return provided by a risk-free asset The cost of equity capital. The cost of equity cap- in a reference industrial country, rm is the return ital must include a risk premium to compensate on a well-diversified portfolio in the reference investors for the nondiversifiable volatility of the country, b is a parameter proportional to the financial returns on a firm's equity. The higher covariance between the return on the equity of the volatility, the higher the expected rate of the gas firm and the return on the diversified 4 The 1996­97 Gas Price Review in Argentina portfolio in the reference country, and riskARG cient to the gearing ratios of firms. Details on is a premium reflecting Argentina's sovereignty this adjustment can be found in Enargas risk. Except for the last term, the above formula (1996a). The adjustment for the differences in is the standard capital asset pricing model. regulatory regimes is undertaken by examin- ing the beta coefficients for firms regulated The risk-free interest rate (rf) was taken to be the under different regimes in Britain and the yield to maturity on U.S. Treasury bonds of the United States. same average life as the Argentine gas compa- nies. The sovereign risk premium (riskARG) was The final result was an average beta coefficient obtained by comparing the rate of return of a for Argentine gas distribution companies of 0.78. foreign currency­denominated Argentine bond For gas transport companies there was no equiv- with that of a U.S. Treasury bond. Enargas used alent information on beta coefficients from the Argentine euronote bonds as the basis for com- United States. Instead, the parameter for this sec- parison. These bonds are denominated in U.S. tor was obtained by rescaling the beta coeffi- dollars and in deutsche marks and were issued cient for the distribution sector by the relative by the Argentine government in European finan- standard deviation of the returns to each type of cial markets. The risk premium of a well- activity in Argentina. (Details can be found in diversified portfolio (rm ­ rf) was estimated as Enargas 1996a.) The result is an estimated beta the difference between the return on a basket of coefficient for gas transport companies of 0.58. stocks in the United States and the rates on U.S. This accords well with prior expectations. Gas Treasury bonds with long maturities. transport is less risky than the more competitive distribution sector, and so should have a lower The final parameter needed is the beta coeffi- risk premium. cient (be). This parameter is proportional to the correlation between the returns on the firm's The cost of debt capital. The cost of debt capital equity and those on the market portfolio. was estimated as rf + riskARG, the sum of the Estimated betas were available in the United risk-free interest rate (measured by the rate of States only for gas distribution companies. The return on U.S. Treasury bonds of similar average average beta coefficient for these companies was life as the Argentine gas firms) and the sovereign 0.58 (Enargas 1996a). The estimated beta coeffi- risk premium for Argentina. The cost of debt cap- cients for the United States were adjusted for two ital amounted to 12.56 percent for transport and factors before being applied to Argentina. Both 13.02 percent for distribution. adjustments relate to differences in the risk char- acteristics of the companies in each country: Determining the efficiency factor First, the financial gearing of a firm will influ- ence the beta coefficient. Unless firms in Enargas analyzed three sources of information Argentina have the same financial gearing as on potential efficiency gains to forecast costs and firms in the United States, the beta coefficients set efficiency (X ) factors: for U.S. firms are not applicable to Argentine Efficiency-enhancing project and restructuring firms. plans submitted by the firms. Second, gas distribution firms in the United Global productivity trends in the industry. States are regulated by a rate-of-return system. Financial models to check the consistency of This regulatory regime may be considered results. inherently less risky for investors than the price cap regime practiced in Argentina. Legislation requires Enargas to identify and quantify the impacts of specific efficiency pro- An adjustment for the first factor requires the jects as a basis for setting the X factors. To that use of a formula that relates the beta coeffi- end Enargas, with the help of independent con- The World Bank Group 5 sultants, analyzed detailed programs that The transport and distribution companies pre- allowed for reliable estimates of efficiency gains. sented investment projects worth 1,774 million Examples included inventory control programs, pesos, of which just 192 million pesos were changes to firms' input purchasing strategies, approved by Enargas and qualified for a K fac- and changes in billing systems. tor. Because Enargas was still evaluating many projects at the time of the final tariff determina- In this respect, the legislation requires Enargas to tion, it retained the power to approve further adopt a method that requires detailed knowledge projects up to ninety days after this date. of the management of firms, and thus contradicts the spirit of arm's-length regulatory control. The The K factor is broken down by project and problem with this approach is that asymmetric semester. It is activated only when an investment information prevents the regulator from identify- project has been completed according to its orig- ing all the efficiency improvements that a com- inal specification and it meets the objectives for pany could introduce and that, moreover, not all which it was proposed. The K factor for each pro- efficiency gains can be linked to specific pro- ject is estimated as the percentage increase in tar- grams. For these reasons, Enargas also analyzed iffs that would be required so that the value of historical total factor productivity in setting the X the firm is the same with and without the project. factors. However, since license conditions re- quired that the X factors be based on clearly iden- Tariff increases for system expansion apply only tified and quantified projects, Enargas had to to customers who benefit from the investment. expend some legal effort in justifying the appli- Consequently, K factors are specific to each cation of total factor productivity analysis in set- ting the factors. The final X factors for each company are shown TABLE 3 FINAL EFFICIENCY (X ) FACTORS FOR in table 3. These factors are applied once at the TRANSPORT AND DISTRIBUTION FIRMS beginning of the five-year period. Thus on Percent January 1, 1998, tariffs were reduced by the full amount of the efficiency factor. Firm X factor Determining the investment factor Transportadora de Gas del Sur 6.5 Investment (K ) factors, if positive, increase tar- Transportadora de Gas del Norte 5.2 iffs each semester. Their purpose is to stimulate Distribuidora de Gas Cuyana 4.8 investment in improving and expanding the gas Gas Natural BAN 4.8 system. Investment projects are approved by Litoral Gas 4.7 Enargas if they: Metrogas 4.7 Have reasonable costs and schedules. Cannot be funded with the original tariffs and Distribuidora de Gas del Centro 4.7 so require additional investment. Camuzzi Gas del Sur 4.6 Expand the system--maintenance investment Camuzzi Gas Pampeana 4.5 is considered when setting the efficiency (X ) GASNOR 4.4 factor--and improve the quality and security of supply beyond the requirements stipulated Note: For transport firms the X factors apply to all interruptible supply and firm supply in the license conditions. tariffs. For distribution firms the X factors affect residential, general small, general, Benefit the majority of the firm's customers. compressed natural gas, and subdistribution tariffs. Are structured so that companies assume all Source: Enargas 1997. construction cost risks. 6 The 1996­97 Gas Price Review in Argentina TABLE 4 INVESTMENT (K ) FACTORS APPROVED FOR GAS DEL NORTE, 1998­2002 Percent 1998 1999 2000 2001 2002 First Second First Second First Second First Second First Second Tariff zone semester semester semester semester semester semester semester semester semester semester Salta 0 0.84 0.52 0.74 0.46 0 0 0 0 0 Tucuman 0 1.74 0.61 0.82 0.55 0 0 0 0 0 Central 0 1.85 0.65 0.87 0.58 0 0 0 0 0 Litoral 0 1.83 0.64 0.86 0.57 0 0 0 0 0 Aldea Brasilera 0 1.81 0.63 0.85 0.57 0 0 0 0 0 Gran Buenos Aires 0 2.56 1.38 1.55 1.24 0 0 0 0 0 Note: These factors are for a project to reinforce gas mains in some areas with high population densities. Gas del Norte proposed two other projects that were still pending approval by Enargas at the time the new tariffs were formally announced. Source: Enargas 1997. TABLE 5 INVESTMENT (K ) FACTORS APPROVED FOR METROGAS, 1998­2002 Percent 1998 1999 2000 2001 2002 First Second First Second First Second First Second First Second Tariff affected semester semester semester semester semester semester semester semester semester semester Residential 0 0.57 0.54 0.51 0.48 0.45 0.42 0.40 0.37 0.35 Commercial and industrial 0 0.41 0.39 0.37 0.35 0.33 0.31 0.29 0.27 0.26 Note: These data are provisional K factors for two of the three projects presented by Metrogas. The factors are provisional pending the evaluation and approval of the third project. Source: Enargas 1997. geographic tariff zone. In addition, only tariffs for Conclusion services that benefit from an investment project are affected. The 1996­97 Enargas price review offers a sophis- ticated approach to the price regulation of natural Final K factors for the northern transport firm monopolies. An extremely complex procedure, (Transportadora de Gas del Norte) and for the review was based to the extent possible on Argentina's biggest distribution company (Metro- objective information and rules. Its well-defined gas) are presented in tables 4 and 5. procedures and methodologies turned out to be very helpful in resisting pressure from lobbies and It must be borne in mind that K factors are preventing regulatory capture. applied to tariffs in the semesters when the cor- responding investment projects are under way. The review also set important precedents for Thus the dates shown in tables 4 and 5 may not future price reviews in the Argentine gas indus- be the effective ex post dates if projects fall try, as well as subsequent utility price reviews in behind schedule. other countries and sectors. The World Bank Group 7 A point of interest in the Enargas review is the adaptation of methods developed in the United States or Europe to overcome the lack of data in Argentina. In particular, the methodology adopted for estimating the cost of capital may be relevant to other developing countries. Gas transport and distribution licenses require Enargas to identify projects that will lead to effi- ciency gains in order to set the X factors. This approach is unlikely to detect the full range of efficiency improvements that could potentially be made, and furthermore may lead to excessive micromanagement of firms by the regulator. To overcome these problems, Enargas used a more aggregate method (total factor productivity analysis) to determine companies' potential effi- ciency gains. The use of this methodology was contested by the industry, however. Viewpoint is an open 1 Legal restrictions prevent domestic price indexation in Argentina. forum intended to But because of the country's currency board system, domestic infla- encourage tion is not expected to differ from international inflation. dissemination of and debate on ideas, innovations, and best References practices for expanding the private sector. The Enargas (Ente Nacional Regulador del Gas). 1996a. "El costo de cap- views published are ital en la revisión quinquenal de tarifas." Working Paper. Gerencia those of the authors and de Desempeño y Economía. should not be attributed ------. 1996b. "Metodología revisión tarifaria solicitud de informa- to the World Bank or any ción." Letter and annexes sent to licensed companies. of its affiliated organiza- ------. 1997. "Informe de aprobación revision tarifaria quinquenal tions. Nor do any of the periodo 1998­2002." Informes GDyE/GAL/GT/GD/GR N. 58­67. conclusions represent Hetherington, B. 1992. "Estimating the Rate of Return for Gas official policy of the Transportation." U.K. Office of Gas Supply. World Bank or of its National Economic Research Associates. 1996. "Contrato de consul- Executive Directors or toría N° 006 metodología de cálculo del factor de eficencia (X )." the countries they Report prepared for Enargas. represent. To order additional Andres Gomez-Lobo (andres_gomez-lobo@ copies please call oxera.co.uk) and Vivien Foster (vivien 202 458 1111 or contact Suzanne Smith, editor, foster@oxera.sprint.com), Oxford Economic Room F11K-208, Research Associates Ltd., United Kingdom The World Bank, 1818 H Street, NW, Washington, D.C. 20433, or Internet address ssmith7@worldbank.org. The series is also available on-line (www.worldbank.org/ html/fpd/notes/). Printed on recycled paper.