on geographical and other conditions. Since the slacks and shortages that develop under Scenario I are relatively small, we assume, probably over-optimistically, that they do not induce secondary cutbacks in other sectors. The general problem of bottlenecks becomes more serious under Scenario II, where it is discussed more fully.

Unemployment estimates in Table I are quite conservative since they (i) ignore the threat of bottlenecks, (ii) assume that final deliveries in excess of initial final demand can be sold to final users, and (iii) ignore the negative feedback effects of direct unemployment on consumer demand. It is unlikely that these second-order reductions in consumer purchases would neatly cancel the bottlenecks due to the petroleum cutbacks, the two phenomena would affect different products in different proportions.

SCENARIO II

For the second scenario we increased the cutbacks implied by the President's guidelines by a factor of 1.67 to cover a 20 percent (rather than the original 12 percent) petroleum shortage. Sectoral petroleum cutbacks are 41 percent for services and trade, 16.7 for process energy in all others. First-round effects of these cutbacks on employment and on computed final deliveries are shown in the second column of Table 1 and the second and fifth columns of Table 2, respectively. The energy shortage now reaches beyond all the cushion factors and every sector is forced to reduce employment. Layoffs vary from around 10 percent in heavily oil-dependent industries to as little as 2 percent in sectors that rely on other energy sources. Average unemployment due to the oil shortage now amounts to 8.3 percent.

Even this gloomy estimate is unrealistically low because the shortages and surpluses that result from this allocation are serious. For example, shortages of transportation (sector 65) and of plastics (sector 28), shown in Table 2, amount to 8 or 9 percent of those sectors' initial outputs and constitute 22 and 52 percent of final deliveries, respectively. They are too large to be absorbed by unplanned final demand adjustments. Bottlenecks induce successive reductions in production and employment. When petroleum feedstocks are inadequate, chemicals are cut; when chemicals are cut, plastics are cut; when plastics are cut, automobiles are cut; and so on.1 Transportation constraints rapidly limit the activities of all sectors.

One can only speculate about total unemployment induced by bottlenecks because it depends on how the scarce chemicals, transportation, etc., are themselves allocated among their potential users. If shortages are borne proportionally among intermediate users, all sectors will approach the largest percentage cutback sustained by the most restricted sector-in our example, close to 10 percent! When petroleum is not properly allocated in the first place, shortages may warrant controlled allocation of many other intermediate products as well. Reductions in consumer demand induced by direct unemployment are also likely but we have not estimated them. The picture is alarming enough to recommend that we seek an alternative approach.

SCENARIO III

The third scenario assumes the same direct cutbacks of petroleum to households and hence the same initial shortfalls as Scenario II. In this case, however, bottlenecks are avoided through a two-pronged approach. First, since petroleum is insufficient to satisfy pre-shortage final demand, some cutbacks in final deliveries cannot be avoided. Rather than reducing final delivery changes from a predetermined industrial allocation, we plan a program of reductions that nurposely discriminates against products that are energy-intensive rather than labor-intensive. Second, by an input-output computation, we estimate the levels of sectoral outputs required to deliver the revised final bill of goods and deduce the demand for petroleum from the new output levels. Because output levels are just sufficient to satisfy final demand, bottlenecks are avoided and no fuels are wasted in the production of goods that will not be bought.

Figure 2 illustrates an informational basis for deciding what final items to cut back. For selected final goods it shows the ratio of total (direct plus indirect) energy to total labor in units of 105 BTU's per man year. Other things equal,

1 Hegeman. George and Vince Ficcaglia. "United States Petrochemical Industry Impact Analysis." Arthur D. Little, Inc., Cambridge, 1973.

there will be less unemployment for a given energy deficit if items with higher total energy-to-labor requirements are cut back. Of course, reductions in some areas like food production might impose greater hardship to consumers than reductions in others. Furthermore, severe curtailment of a single industry might concentrate the burden too heavily on a particular group. The choice of a reasonable set of reductions in final deliveries involves judgment about social priorities and should not be formed mechanically.

A hypothetical program of final delivery curtailment is cited in Table 2, column 6. It is made up of the following reductions:

Percent

25

25

40

30

25

State and local government construction was singled out because it is very energy-intensive, because it includes a lot of highway construction that should be reconsidered now, and because it can be reduced by direct administrative action. Gross capital formation contains many energy-intensive products. Certainly 25 percent of private investment can be postponed in time of national emergency. The business climate will tend to discourage new investment in many areas anyway. If necessary, Federal action can discourage it further by emergency regulation. Reductions in automobile demand also seem to be taking place spontaneously. Cutbacks in defense expenditures are worth considering on many grounds. Finally, the last two items on the list result automatically from announced restrictions on private automobile and air travel. This set of cutbacks is only one of many possible energy-saving programs. Alternative programs should certainly be considered.

Using our input-output model we computed the reductions in sectoral outputs and employment that would result from the planned curtailment of final demand. The percentage cutbacks in employment are shown in column 3 of Table 1 and in Figure 3. The overall employment rate due to the oil shortage is well below

On the basis of computed reductions in output, petroleum cutbacks by individual sectors were computed from equations of the type pictured in Figure 1. Each sector's allotment depends on its output level, its cushion factor, its initial use of substitute fuels. Estimated reductions in petroleum consumption under Scenarios II and III are compared. Total energy cutbacks are the same under the two scenarios but individual sectors are affected differently. Under Scenario III, "negative cutbacks" occur in the food and utilities sectors. They must receive more than their initial consumption in order to make up for the six percent shortage of natural gas.

An input-output computation can provide guidelines for the industrial allocation of scarce petroleum. To implement a program requires a great deal of wise technical and administrative judgment as well. Reducing final demand will automatically curtail consumption in many sectors but administrative measures will still be necessary to reduce industrial consumption for space heating and lighting and to force all sectors to eliminate waste wherever possible.

utilities

7 percent, as compared with something between 8 and 10 percent under Scenario II. Unemployment is particularly high in construction as compared with other sectors; there may be some solace in its wide geographical dispersion. Although new construction is reduced, maintenance construction is sustained at a high level. Since outputs are sufficient to cover intermediate and final demands, serious bottlenecks are averted.

FIGURE 3

Petroleum Cutbacks urder Scenarios I and III

The cost per barrel of oil and per gallon of gasoline to intermediate and final consumers is likely to double within the next few months. Steep increases in petroleum taxes have been proposed. Alternatively, prices of energy may be allowed to rise to double or more with increased profits to the oil industry. Indeed, they have already risen sharply in the past few months.

How will other prices be affected? Using a standard input-output price computation, we estimated the effects on all sectors' prices of a 100 percent tax on refined petroleum products. In a second computation we estimated the effects of 100 percent taxes on all basic energy sources: on coal and on natural gas as well as on petroleum products. The estimates assume that wages and profits in each sector remain unchanged. Alternative assumptions could be built into the computations, but they probably would not significantly alter the thrust of the conclusions.

Computed price changes are listed in detail in Table 3. In column 2, they range from 14 percent for basic chemicals (27) which has a very high direct and indirect energy requirement to 1.3 percent for radio and tv broadcasting (67). Figure 5 summarizes the price changes over broad industrial groups. As might be expected, prices of construction and maufactured products rise more than those of service sectors, which are labor-intensive. For many years labor saving has been the key to successful enterprise in the American economy. These price changes may well shift the focus to energy saving.