7. Policy Alternatives for Future Rice Production Growth

Paul Heytens

Indonesia's efforts to increase rice production since the late _ig6os have been enormously successful. Between ₁₉₆₉ and _1988, rice production grew at an annual rate of _4.5 percent. As a result, Indonesia moved from being the world's largest importer of rice in the _1970s to self-sufficiency in _1985. Successful development has been the result of policies that emphasized new technology, infrastructure investments, and remunerative prices to farmers. The dissemination of new and shorter-duration seed varieties, larger applications of fertilizer, better advice to farmers, and improved water control were key factors in increasing output. Rates of growth in rice output were roughly the same on and off Java and came mainly from increased productivity per hectare. Yields grew at about 3.3 percent annually, whereas harvested area increased only _1.2 percent annually.

Despite the past production success, continued future growth of rice output will require close attention by policymakers. Tabor et al. ₍₁₉₈₈₎ project that growth in rice consumption will remain above ₂ percent per annum through the year _2000, although the rate of increase will decline from _2.6 percent a year in the early _1990s to _2.1 percent in the late _199os. These consumption projections imply that rice production will need to grow at about _2.5 percent per annum through the remainder of REPELITA V for Indonesia to maintain self-sufficiency on trend. Although this rate is only half the growth in production attained during the past two decades, achieving a steady growth in rice production of _2.5 percent is no simple task. Much of the easier gains in both yields and irrigated area planted to rice already have been made.

Evidence from the field gives some reason for optimism that self-sufficiency can be sustained. Large-scale shifts into rice on _sawah will not occur because rice already is being grown wherever feasible. But ample area could be transformed into higher-productivity systems through investments to improve water control. On Java, over 1 million hectares of rainfed and poor-control, irrigated _sawah could be transformed by investments in new irrigation systems or rehabilitation of existing systems. Field evidence and the historical record in Indonesia indicate that beneficiaries of better-controlled water supplies respond with increased input use to exploit new opportunities. Further, to the extent that agricultural research and extension services can make available new technologies such as integrated pest management and hybrid rice seeds Indonesia's rice farmers can be expected to adopt innovations that will allow them to attain higher profitability and productivity. Finally, Java's sugar producers would move out of cane into rice-dominated cropping patterns if the government allowed them to do so.

The desirability of making investments in irrigation infrastructure, maintaining incentives for farmers and a vital research and extension program, and deregulating sugar production depends partly on how efficiently Indonesia is able to produce rice and competing crops. Do increases in rice output, needed for greater food security, also contribute to efficient growth of national income, or is there a trade-off between food security and efficiency? Does Indonesia have a comparative advantage in producing rice for import substitution or for export? Are there competing crops that provide higher social profits than those for rice? The answers to these questions, addressed below, go a long way in defining an appropriate rice strategy for Indonesia.

Social Profitability of Rice and Competing Crops

To assess Indonesia's comparative advantage, the social profitability of rice and competing crops must be calculated. For that purpose, policy analysis matrices (PAMs) were constructed. This method of analysis simultaneously addresses the prospective changes in levels of rice income, the contribution of rice to efficient growth of national income, and the income transfers resulting from price and macroeconomic policies. The main task is to construct, for each rice production system defined in Chapter 4, a set of accounting matrices of revenues, costs, and profits, first including and then excluding the effects of policy and market failures. The approach is described in detail in Monke and Pearson (1989).

The first part of this task was accomplished by the construction of budgets from data gathered during the farmer surveys. The second part requires revaluing private costs and benefits at social or efficiency values-what costs and returns would be if there were no distorting policies or market failures. The world price of any tradable input or output, with appropriate adjustments for location and quality, represents the price at which inputs and outputs could be bought or sold without intervening policies. For example, in the absence of intervening policy, farmers could purchase an input traded on world markets for the world price (c.i.f. at the nearest port) plus the cost of transport to the farm gate. Similarly, traded output (which includes rice) could be sold for the world price (c. i. f. at a domestic port for an import substitute or f. o. b. for an export). Assessing the social values for inputs that are not traded (e.g., transport, capital, and labor) is more difficult, but the principle remains the same as with tradables-to discern costs in the absence of distorting policies and market failures.

For this analysis, the production budgets presented in Chapter 4 are generalized to represent wetland systems for Java, South Sulawesi, and West Sumatra, the project's survey regions. Private budgets estimated from survey results were checked for consistency against data from the triannual cost of production surveys done by Indonesia's Central Bureau of Statistics (BPS)-that is, a weighted average of the rice production budgets described in Chapter 4 was estimated to verify its similarity to the BPS average budget for Indonesia.

Policy Analysis Matrix Assumptions

The PAM approach requires evaluation of costs and benefits at social or efficiency values. In practice, this task is based on a revaluation of private costs and benefits and involves making judgments about key parameters. Therefore, if the PAM analysis is to be a useful guide to Indonesia's policymakers, assumptions for important parameters must be made explicit and sensitivity analysis done to assess the robustness of results to different assumptions.

For this analysis, social values for rice output, the foreign exchange rate, the interest rate, and labor costs are particularly important. Domestic rice prices were in line with world prices of comparable quality grains in 1987 (c.i.f Indonesian ports of about $200 per metric ton for Thai 35 percent brokens). Although prices on the world rice market are very difficult to project (as Chapter 3 makes clear), world rice prices in 1987 were judged to be below their expected long-run trend. A c.i.f. Indonesia price of $250 per metric ton consequently is used as the social value of Indonesian rice in the base case.

As described in Chapter 2, the history of Indonesian exchange rate policy in the 197os and ig8os is characterized by periods of overvaluation caused by positive differential inflation (i.e., domestic inflation higher than that of major trading partners) followed by large corrective devaluations. Although differential inflation was positive in 1987 and import restrictions created some unrealized demand for foreign exchange, any overvaluation of the rupiah then was not large. A major devaluation of the rupiah in September 1986 probably brought the Indonesian currency back into line with its opportunity cost at that time, and any subsequent overvaluation during 1987 was offset by the sharply rising value of the yen relative to the dollar (and thus the rupiah). Therefore, as a base assumption, no correction is made to the private Rp/U. S. $ exchange rate.

The private return to capital does require some adjustment for social valuation of capital costs. Credit policy differs from that in other developing countries because the private real interest rate (14 percent) is higher than its social counterpart. This supposition is corroborated by the 4 to 5 percent decline in nominal interest rates that occurred after the implementation of the financial sector deregulation package in late _1988. Other analysis indicates that a real social rate of return of 1o percent is a reasonable approximation for a country such as Indonesia (Monke and Pearson _1989).

In the calculation of social costs, the 1o percent rate is applied to purchases of large capital equipment such as tractors and large trucks for transport. It is not intended as the social opportunity cost of farmers' working capital. Evidence from the KUPEDES (rural credit) program indicates that the interest rate at the village level for small savers and borrowers is considerably higher than the rate in urban centers. Nominal KUPEDES savings rates were in the ₃₀ percent range in _1987. Given an inflation rate of ₈ to ₉ percent, the real social opportunity cost of farmers' cash outlays is considered to be zo percent per year in the base case.

Finally, a judgment must be made about the competitiveness of Indonesia's rural labor markets. Chapter 5 provides evidence that the labor market is operating with reasonable efficiency. Therefore, no divergence is assumed between private and social labor costs in the analysis. A complete listing of parameter values used and other assumptions made is found in the appendix at the end of the chapter.

Policy Analysis Matrix Results

The results of the social and private profitability calculations for rice and competing crops for the base year ₁₉₈₇ are reported in Table _7.1. The table shows that rice as an import-substitution activity (i.e., measured relative to a c.i.f. Indonesia world price) is socially profitable on all types of _sawah in the three regions. Further, social profits for rice are higher than for competing _palawija crops, and the social returns to land and management for soybeans and sugarcane are negative. A move toward world prices or a reduction in planted area targets for either of these commodities, but particularly for sugar, would free up additional cropland for the production of more socially profitable rice. This efficient change would increase farm incomes, food production, and national income.

Government policy effects a relatively small magnitude of income transfers to rice producers through tradable inputs, as indicated in Table _7.2. The source of policy transfers for tradable inputs is the direct subsidy on chemical fertilizers and pesticides, and transfers are largest in the most productive systems. But the magnitudes involved are small. The fertilizer subsidy varies by type but is largest for urea and triple super phosphate, the most heavily used tradable inputs.

The major policy transfers in the domestic factors category result from subsidies on transport and handling of fertilizers and pesticides to the village cooperative (KUD) and the free provision (generally) of irrigation water. The transport subsidy for fertilizer and pesticides is of roughly the same magnitude in all regions, but the irrigation subsidy is considerably larger on Java than in the outer islands. In addition, divergences in urban capital markets have the effect of increasing private tractor rental and transport costs for rice output above social costs. The magnitude of factor market transfers, however, is small and does not offset the overall positive effects on farmers' incomes of the tradable input subsidies.

On the output side, Table _7.2 indicates that government policy is taxing rice producers. As discussed above, world rice prices in ₁₉₈₇ were judged to be below long-run trend values. Since domestic rice prices in ₁₉₈₇ were broadly in line with actual world prices, there is an apparent tax on producers. This implicit tax arises because government policy forced rice producers to accept a domestic price that was below the trend world rice

price in _1987. The difference is about $5₀ per metric ton plus domestic transport costs, more than 25 percent of the social value of rice.

Sensitivity Analysis

Rice production's comparative advantage as an import-substitution activity is very robust to changes in parameter assumptions. For example, adjustments in the foreign exchange rate to reflect a possible overvaluation only strengthen Indonesia's comparative advantage because a higher Rp/U.S.$ exchange rate serves to raise the rupiah cost of competing rice imports and hence the rupiah value of domestic rice. In addition, doubling the real social rate of return to 20 percent and increasing the opportunity cost of farmers' working capital to 30 percent raises domestic factor costs by not more than RP 50,000 per hectare for any of the rice systems studied, hardly enough to challenge base case conclusions.

The more interesting parameters for sensitivity analysis are the break-even prices for import and export competitiveness, that is, the minimum c. i. f. and f. o. b. prices that would have to prevail for Indonesian rice to be competitive. Under the base case assumptions, the c.i.f. price for Thai 35 percent brokens could fall to roughly _$18o per metric ton before Indonesia's least efficient wet-rice systems would cease to be socially profitable. Furthermore, Indonesia might be able to export rice profitably when world prices are at or above long-run trend values (f.o.b. prices at or above $200 per metric ton). Because of problems associated with the thin world rice trade, the depressing impact of Indonesian sales on export prices, the quality of Indonesian rice, and the long gestation effects of investments related to market development, exports are more problematic and the results must be interpreted more cautiously. But the analysis indicates at least that occasional exports in times of domestic surplus could be done efficiently. The future choice of domestic price levels needs to consider the possibility of efficient exports because price support levels, not production costs relative to f.o.b. prices, will determine whether Indonesian exports can compete without subsidy.

Defining a Future Production Strategy

In _1987, Indonesia had a strong comparative advantage in producing rice as an import substitute. At that time, it appeared economically efficient to invest resources in rice agriculture to maintain trend self-sufficiency. Not doing so, and importing rice, would have been an inefficient use of resources. This result is strong evidence against the adoption of a near-term future rice production strategy that would imply a low rate of growth of output (about i percent annually) and lead to regular imports. In the remainder of this chapter, therefore, the low production target is

ignored, and attention is focused instead on the two other output targets and their associated strategies-z.5 percent annually to achieve self-sufficiency on trend and 4 percent annually to support absolute self-sufficiency and regular exports. The central issue is whether achievement of either of these two growth targets would be consistent with the objectives of food security and efficient income generation. Their likely impacts on equity (personal and regional income distribution) are examined in Chapter 8.

Public Investment in Irrigation

REPELITA V provides estimates of the public-sector investments that are likely to take place by early 1994.₁ Of particular importance for this study are the irrigation investments that will convert unused or low-productivity land into higher-yielding systems. Transformation of rainfed land involves the building of new irrigation works, whereas upgrading poor- and moderate-control sawah requires rehabilitation of already existing infrastructure. In addition, expenditures on operations and maintenance of the entire system are needed to prevent physical deterioration of past investments and the consequent loss of rice output.

The Fifth Plan allocates Rp 8.5 trillion to the irrigation subsector between 1989 and early 1994. The physical targets include about 5oo,ooo hectares of new irrigation development, 450,ooo hectares of upgraded swamp irrigation, and 450,ooo hectares of additional flood protection. REPELITA V also calls for various operation and maintenance expenditures on nearly 6 million hectares.

Irrigation projections for REPELITA V are based on the actual experiences of REPELITAS I through IV. Table 7.3 outlines financial and physical data for four key irrigation programs between 1969 and 1988. The data on plans versus actual realization show considerable variation by category and time period; however, these performances and experiences are consistent with those encountered by other countries in Asia. Without considerably more fieldwork to verify specific local investments, even the most carefully made projections are likely to provide only rough approximations of irrigation's contribution to rice growth during the next five years.

One such approximation is given in Table 7.4. Each of the investment categories is assessed relative to its REPELITA V target. Each category then is broken down into intensity, area, and yield components. The assumptions for each type of investment are made explicit in Table 7.4 so that analysts who have alternative views on irrigation investments during REPELITA V and the lagged effects that could accrue from investments completed during earlier plans can produce their own assessments.

_{1 This section draws heavily from Varley (1989).}

Four important conclusions follow from the assumptions and results of this projection. First, continued investments in irrigation should produce substantial further growth in rice output. Specifically, a 2 million ton increase in _gabah production seems plausible by the end of ₁₉₉₃ from irrigation improvements. This tonnage translates into about a 1 percent per year increase in rice output during REPELITA V Second, the projections in Table 7.4 imply no production increases from operation and maintenance expenditures on irrigation systems. Keeping irrigation systems at reasonable efficiency is a continuing problem. For irrigation, as well as for new seeds, Indonesia will require major expenditures and organizational efforts simply to maintain previous gains in rice output. Third, the gains from new irrigation investments will not be captured so easily as those from earlier investments. Many of the easiest systems have been improved in earlier plans, leaving the more remote and costly irrigation projects for the _19gos. Finally, the resurgence in actual irrigation investments since ₁₉₈₆ and the proposed investments in REPELITA V are good indicators of the importance the government attaches to rice self-sufficiency and of the key roles played by irrigation in such a strategy.

Regulatory Policy

Perhaps the quickest way to increase rice production during REPELITA V would be to change sugar policy. Roughly 150,ooo hectares of _sawah were in sugarcane on Java in _1988, mostly in the TRI program (Nelson _1988). If the government were to phase out the TRI program, perhaps over a five-year period, and current relative prices were to hold, about 150,ooo hectares of _sawah, mostly in moderate- and good-control irrigated systems in East and Central Java, could shift into rice production.

An informal phaseout of TRI sugarcane land began in Java in _1989, because program regulations were not enforced in some parts of Java. For purposes of estimation, a phaseout plan is modeled so that _30,ooo additional hectares of sugarcane land are planted with other crops each year for five years beginning in June _1989. In _1989, two dry-season crops per plot could be grown on the initial _30,ooo hectares. In the first dry season, rice is assumed to be grown on four-fifths and _palawija on one-fifth of the newly converted _sawah. In the second dry season, two-fifths of the available land is assumed to go into rice, two-fifths into _palawija crops, and one-fifth into fallow rainfed _sawah. In _1989, then, an additional _36,ooo hectares of rice (and _18,ooo hectares of _palawija) would come into production. In subsequent years, it is assumed that 40 percent of the converted sugar land could accommodate three rice crops, 40 percent two rice crops, and 20 percent only one rice crop. In _19go, for example, cultivated rice area would increase by 102,ooo hectares. Land changed over in ₁₉₈₉ would provide _66,ooo hectares of rice production (plus _i8,ooo hectares of _palawija production), and _36,ooo hectares would be converted anew to rice. The implications of such a program for rice production are summarized in Table _7.5.

The table indicates that the sugarcane phaseout under the above assumptions would raise cultivated rice area on _sawah by _{33o, ooo} hectares by ₁₉₉₄ because of multiple cropping. If yields on converted land average _5.25 tons of _gabah per hectare in ₁₉₈₉ and increase by _0.5 percent annually, the phaseout would result in about _1.8 million tons of additional _gabah production per year by _1994. In comparison with production of over ₄₁ million tons of unmilled rice in _1988, the production changes shown in Table _7.5 imply, _{ceteris paribus,} annual growth in rice production of 1 percent through the end of REPELITA V.

Price Policy

Price policy is the residual lever that Indonesia can use to affect rice production during REPELITA V. If a production target needed to realize a chosen strategy cannot be achieved with changes in investment and regulatory policy, price policy is the last option available to provide additional incentives to rice farmers. Altemeier et al. ₍₁₉₈₈₎ estimate that the price elasticity of rice yields _{is o.2} in the first year and _0.3 after three years if the price change persists. Hence if the real rice price increases _1o percent and is held at that level, rice yields can be expected to increase about ₂ percent after one year and a total of ₃ percent over a three-year period. Price policy influences cultivated area as well. Altemeier et al. ₍₁₉₈₈₎ estimate the one-period lagged own price elasticity for rice area to be _0.17.

The rapid increases in real prices during ₁₉₈₇ and ₁₉₈₈ have encouraged area expansion and intensification of input usage in rice, which led to a strong rebound in production early in REPELITA V. The roughly _2o percent real price rise from ₁₉₈₆ to ₁₉₈₈ should generate yield growth of

about i percent in ₁₉₈₉ and about _0.5 percent in _19go. Harvested rice area probably increased by nearly z percent in ₁₉₈₉ in response to the earlier price increases. If real prices were held constant throughout the REPELITA V period, the effects of the earlier price increase would stop contributing to rice production growth by the end of _19go.

Figure _-7.1 shows the movements of nominal rice prices, by month, between ₁₉₈₆ and _1989. The rapid increase in prices occurred between August _198-7 and August _1988, reflecting drought and a decision not to import rice. This abnormally large increase moved Indonesian domestic prices from only _8o percent of expected long-run world price levels to approximate equality.

Between August ₁₉₈₈ and late _1989, nominal domestic prices moved very little, and real prices fell somewhat (the annual rate of inflation was about ₆ percent during this period). The stability in prices resulted from excellent weather, repayments of rice "loans" from the Philippines and Vietnam, changes in the enforcement of the TRI sugar policy (discussed above), and price responsiveness of rice farmers. This combination of good fortune and policy permitted BULOG to rebuild its food reserve stock and the country to regain aggregate supply-demand balance in rice. By the close of _1989, then, the Indonesian rice economy had attained

trend self-sufficiency. But policymakers had very little unused efficient policy space within which to raise real domestic prices.

Determining a Policy Mix for REPELITA V

Policymakers are concerned with the changes in rice policies required to pursue a rice strategy, such as trend self-sufficiency or regular exports. The first step in this assessment is to quantify the increases in rice production during the Fifth Plan period that are already in the pipeline. The sources of growth include carryover effects from price changes in earlier years, the impact of public investments (both lagged and future) in irrigation infrastructure, the impact of reductions in sugarcane area already made on Java, and trend effects attributable to improved practices apart from the irrigation program. The potential contribution of these factors to rice production during REPELITA V is summarized in the first column of Table 7.6.

Table 7.6 incorporates the carryover effects from the rice price increases of 1986-88 as well as the phase-in of the incremental _gabah production resulting from irrigation investments, shown in Table 7_.4. In addition, the first year of the sugarcane phaseout plan described above is assumed to have been implemented informally. Moreover, a o.5 percent per annum growth in yields, additional to the impact of irrigation investments, is assumed to incorporate the effects of wider implementation of integrated pest management, the spread of other intensification programs, and local innovations. No trend effect outside of irrigation investment is assumed for rice area. The trend assumptions are low in comparison with past achievements, reflecting the existing high rates of HYV adoption and intensive use of modern inputs, especially chemical fertilizers.

The results in the first column of Table 7.6 show that, without additional policy incentives, a sharp downturn in production growth would occur after 1989. Annual growth during the final four years of REPELITA V averages somewhat over i percent per year. To avoid consistent rice imports during REPELITA V, then, Indonesia likely would have to implement additional incentive policies. The policy analysis matrix results (Table 7.1) indicate that Indonesia has a strong comparative disadvantage in sugarcane production. Irrespective of the rice strategy Indonesia ultimately chooses, it is clearly in the country's economic best interest to continue with the sugarcane phaseout summarized in Table 7.5. The implications of following through with the phaseout of the TRI sugarcane policy are summarized in the middle column of Table 7.6. That column shows that growth in rice production increases to over 2 percent per year

during the final three years of the Fifth Plan, but not enough to reach z. 5 percent, the likely rate needed to attain self-sufficiency on trend. Some increase in incentives to farmers seems necessary as well, although budgetary pressures make high fertilizer subsidy levels difficult to maintain. As a result, administered fertilizer prices were raised considerably for iggo (see Table 2.1). The recent series of fertilizer price increases should not have a large disincentive impact on rice production because Indonesia's wetland rice farmers have had long experience using chemical fertilizers and application rates are already very high (Altemeier et al 1988). Further, there is strong agronomic evidence in Indonesia that TSP and KCl applications could be reduced substantially with virtually no impact on rice yields (Manwan and Fagi 1989; Adiningsih et al 1989) but with considerable savings for the government budget.

If output price policy is used, a 5 percent real price increase in i9go or 1991 would be required to achieve rice self-sufficiency on trend during REPELITA V The impact of a 5 percent price increase in iggi is summarized in the final column of Table 7.6. The projected production response would place Indonesia somewhat above 2.5 percent annual growth through REPELITA V and would allow policymakers the possibility to continue phasing out fertilizer subsidies.

Higher annual growth rates in rice production than these, of course, would require even higher real rice prices (in the assumed absence of even greater investment or faster technical change). For example, an average annual growth rate of 4 percent after 1989 would require an additional real increase in rice prices of at least io percent during REPELITA V Because real rice prices rose sharply during the 1986-88 period, potential political difficulties are likely to impede further large increases in the near term. Higher domestic rice prices also make Indonesian exports less competitive. To generate exportable surpluses, policymakers would have to raise domestic rice prices considerably, but this high price policy would cause the exported quantities to require subsidies in most years. Therefore, pursuit of the high-growth rice strategy probably is not desirable, unless it could be achieved with public investments and technical changes additional to those considered here.

A final caveat is in order. This chapter has focused on trends and longer-run orders of magnitude. Short-term variations in international rice prices, domestic rice production, or Indonesian oil revenues could create opportunities or constraints for rice policy that planners cannot foresee in advance. A successful rice strategy for the 19gos needs to be based on sound long-run planning and investments. It also requires short-run flexibility and the capacity to adjust policy to changing year-to-year circumstances.

Conclusion

In comparison with either the actual 1987 world rice prices or the expected, long-run trend prices, Indonesia produces rice efficiently to substitute for imports. Comparative advantage in import substitution is strongest in the good-control, irrigated systems, but exists in all wetland rice production (about go percent of total production). Because real domestic rice prices rose 20 percent between mid-i987 and mid-i988, however, policymakers cannot afford to allow domestic levels to rise much more if they wish to maintain efficiency. World rice prices are expected on average to fluctuate around a long-run trend price of about $250 per metric ton (c.i.f. Indonesian ports, for 35 percent brokens). At the end of ig8g, Indonesian prices were approximately aligned with this expected trend of c.i.f. import prices. Consequently, policymakers cannot use large price policy incentives to induce efficient increases in rice output and farmers' incomes; production achieved at domestic prices in excess of comparable world prices is inefficient and wastes scarce resources. In this circumstance, Indonesia faces an emerging trade-off between the objectives of improved food security and the efficient generation of income.

The analysis of likely sources of growth for rice production shows that this trade-off could emerge as soon as iggi. By then, the residual incentive effects of the price rise in 1987-88 will have disappeared. Moreover, the projected increases in rice output from public investments in irrigation, research, and other infrastructure and from phasing out the TRI sugarcane program over five years amount to just over 2 percent annually. By iggi, therefore, price policy incentives could be needed to permit yearly growth in rice production of 2.5 percent, the projected rate of increase of rice consumption.

The implications of the analysis are clear. A strategy of regular imports (based on a 1 percent annual growth target) is inappropriate; Indonesia can compete efficiently with rice imports at expected c.i.f. prices. At the other extreme, a strategy of planning for regular rice exports with a high production target (about 4 percent annually) is also likely to be ill-advised. Given the recent (and expected long-run) relationships between Indonesian domestic prices and export prices, the country cannot sell regularly on world markets without export subsidies; this lack of export competitiveness would be exacerbated if Indonesia chose to raise real domestic rice prices substantially in an effort to generate exportable supplies. The most desirable strategy, from the viewpoints of both food security and efficiency, is self-sufficiency on trend. But even this strategy faces potential efficiency problems if substantially higher domestic prices are required to balance domestic supply and demand.

Appendix 7.1. Policy Analysis Matrix Assumptions

Following are the assumptions used for the policy analysis matrix analysis as well as tables of private and social values for general and regional prices and system input use and output data. Important assumptions are justified in the text.