A large number of individual elements counted or assumed, and the uncertainty which accompanies the aggregation of catalog estimates can be quite large.

Imputation of dollar costs is a similarly complex and difficult task. On the surface it sounds simple: One simply estimates the cost of reproducing the item in the United States. This, however, soon proves more difficult than one might think. Where there are direct military analogues, for example, tanks of similar size, armor, and firepower, the cost of the Soviet system is taken directly as the cost of the similar U.S. system. Where there are no closely similar U.S. systems to Soviet systems, estimates must be made by U.S. cost estimators based on the observed or assumed characteristics of the Soviet system. The cost estimating relationships used in such estimates are usually those appropriate to U.S. industrial practice in the U.S. economy. They do not usually take account, for example, of Soviet propensities toward simpler and more rugged machinery, or toward standardization. On occasion, where some well-known Soviet practice has a clearly identifiable impact on costs, the estimates are adjusted to accommodate this.

Marshall, in his paper referred to above, pointed out a specific problem with the costing of Soviet military manpower. The CIA's method of costing manpower is to derive an average U.S. cost-perman by dividing the total manpower costs, including pay, allowances, and subsistence of a particular service by the total manpower in the service. This factor is, in turn, multiplied by the total manpower in the appropriate Soviet service to derive a total dollar cost. One of the problems with this, as Marshall pointed out, is that it assumes the grade and pay structure of the Soviet service is the same as that of its equivalent U.S. service-an unlikely assumption.

RUBLE COSTING

Ruble costing of Soviet defense programs involves much of the same kind of estimating process as the dollar costing. Ruble prices are available for many of the items in Soviet military programs, such as pay to individuals and items that are much the same as other items in the civilian economy, such as trucks. Ruble prices for Soviet weapons systems are not, however, available in the open literature. For such systems, what is commonly done is to make a dollar cost estimate, and then convert this to rubles by using a ruble-to-dollar conversion ratio based on a similar exchange ratio; e.g. for a tank, one might use what is known for tractors.

As the preceding brief discussion makes clear, there is a very large element of estimation in any of the cost comparison methodologies, and hence a large amount of uncertainty. Further, since the method used to make cost estimates frequently depends on available information, the basis of the estimates may change over time as more information becomes available or as sources of information are closed to us. There is a final theoretical point to be made about cost comparisons of activities in different economies. The method of constructing the

2 For example, it is reported that the diesel engine used in the Soviet T-54 tank is essentially the engine that has been in serial production in the U.S.S.R. since the mid-1930's.

aggregated defense costs resembles the method of construction of index numbers used for economic comparisons across time. It suffers from the same theoretical deficiency that index numbers do. That is, the evaluation of a collection of goods in one period by the prices prevailing in another does not exactly reflect its real value in either period, because there is no opportunity to change the mix of goods in response to changing relative prices between the periods. Given that relative prices do change between periods, the index number always overstates the value of one period's goods in another period's prices. Applied to spending comparisons, this means that dollar costing of Soviet programs, or ruble costing of U.S. programs tends to overstate the costs of both. A brief example may help clarify this. For the United States to attempt to duplicate a Soviet missile capability by making a "Chinese-copy" of a Soviet missile would be a very expensive undertaking, because U.S. industry and engineering just do not do things the way the Soviets do, and a lot of changing around and nonstandard practice would accompany such an undertaking. It would be cheaper to replicate the capability of the Soviet missile using the design and fabrication practices common in this country. But the point of this example is that Soviet practice is cheaper in the U.S.S.R. and U.S. practice cheaper in the United States because in both cases people have proceeded in ways to minimize their own costs, and these are different ways.

In summary, with respect to methodologies, there are theoretical reasons to think that cost comparisons between different economies are biased and there are practical reasons for thinking that there is considerable uncertainty in the CIA aggregated dollar cost and ruble cost estimates.

C. ARE THE ESTIMATES USEFUL?

Even if accurate, unbiased, and credible cost comparisons could be made between U.S. and Soviet military programs, would they be useful? To put the point another way, would it matter if one side were out spending the other on military forces? Clearly, the answer depends on what each side is obtaining from its resource inputs. What is important about military forces is certainly much more than their ability to absorb resources. Senator Proxmire has recently suggested that an explanation of estimated increased Soviet military spending is that the Soviets are learning to waste money on military programs at a much greater rate. What is meant by waste in this context should, perhaps, be the subject of another paper. But the important point is to be concerned with the relevant outputs.

One form of military output loosely linked to the input comparisons and size comparisons is the impact on the perceptions of adversaries and allies of one's military programs. It is how tough we appear; how much we reassure our friends; how much we deter our adversaries from testing us. This kind of military output has both political utility and drawbacks, and is certainly related to how much effort (money) we put into military programs, and how big our apparent forces are. It is perhaps less important in this regard that our apparent forces have a high degree of readiness, or have really competent, survivable weapons systems in real battle environments.

Another form of military output is generally termed "capability." It is the capacity to destroy enemy forces, defend friendly territory and forces, and generally to win military engagements. Calculating

such capability is a complex technical task which requires a host of assumptions about the engagement scenarios, characteristics of opposing systems, and many other things. However, it is this form of military output that many people think of in relation to money spent on military programs. But here the link between resource inputs and appropriate outputs is even more tenuous.

CONCERNS WITH REAL MILITARY CAPABILITIES AND REQUIREMENTS OR PERCEPTIONS OF POLITICAL USES OF MILITARY FORCES

Whether or not spending comparisons are relevant to discussions about how much to spend probably depends in the short run upon whether one is more concerned with perceptions and the political use of military forces, or with issues of real military capability. In the longer run these issues converge and thus credible spending comparisons can. be useful in assessing trends in real military capability as well as in resolve and determination not to be second best. If one side puts more resources than the other into military programs consistently over a period of years, it is difficult to escape the conclusion that the first side will eventually accumulate a preponderant military capability. Such a military capability may still not give the first side a free hand in imposing its will on the other. The usefulness of military preponderance can have severe limitations. The significance of military superiority or inferiority is presumably an issue which is capable of being analyzed and settled on its own merits. To put the point another way, spending comparisons are not likely to be useful in predicting what happens as the result of one side or another achieving a lead in military capability.

There is considerable uncertainty about whether the costing methodologies presently in use to compare military spending in the U.S. and U.S.S.R. will support conclusions, at least in the short run, about the status and trends in real military capabilities between the two countries. There is difference of opinion about whether spending levels would be relevant measures even if the costing methodologies were credible. This leads to the conclusion that cost comparisons must be used with great care and discrimination in arguing appropriate policy options. At the least, such comparisons should be given context by reference to appropriate military outputs, and by reference to observable forces and activities of the U.S.S.R. As Marshall observes, ". . . the policy issue to which sizing analyses are directed would be better served if several complementary estimates of Soviet activity were available and appropriately documented."

NEED TO ASSESS U.S. NATIONAL SECURITY REQUIREMENTS

Finally, the foregoing discussion prompts the observation that the tendency to compare inputs to military programs between the U.S. and U.S.S.R. in addressing policy questions may beg a more fundamental question. That is, what sort of capabilities does the United States require to defend itself and its allies, and to carry out its foreign policy? Should we assume that required capabilities must be attained by matching inputs with our adversaries, or should we tailor our programs to provide specific outputs where we perceive the need for them?

IV. U.S./U.S.S.R. Quantitative Force Comparisons: Cost Implications for Fiscal Year 1977 Budget

Over the past 15 to 17 years, the U.S.S. R. has been committed to a policy of upgrading its military forces and firepower to achieve parity with the United States in commonly accepted measures of global power. The Soviet Union, through concentration and steady growth in defense expenditures, has advanced its technology for military purposes. Consequently, U.S. superiority in strategic forces which was unquestioned during the 1960's has been challenged by the U.S.S.R. in the 1970's.

UNACCEPTABLE DAMAGE

The overall strategic balance remains in favor of the United States. Because of America's capacity to inflict unacceptable damage on the U.S.S.R., and the uncertainty of the outcome of even limited military provocations, the Soviet Union is deterred from initiating hostile military action.

The proposed programs in the FY 1977 defense budget indicate the preferred direction of the Administration for U.S. military posture through the end of the 20th century. The recommended strategic and general purpose force strengths, when added to existing United States and allied force power, are intended to solidify NATO capability to defend against the Soviet Union and its allies.

SALT

Many welcomed the SALT agreements as a positive if not conclusive means of constraining the nuclear arms competition. Yet, since June 1972, both sides have improved the quality of their strategic systems, and the U.S.S.R. has continued to develop and deploy (a) a wide variety of ballistic missiles, (b) the Backfire bomber, and (c) assorted ships for an enhanced naval capability.

In the general purpose forces realm, the Soviet Union has expanded its force levels and added new equipment, especially tactical aircraft, artillery, and navy ships. Much of the new land and tactical air force strength has been deployed to the Sino-Soviet border to face the Chinese. At the same time, the Soviet Union's western-oriented forces have improved in terms of size and sophistication. The Soviet Navy is now capable of operating worldwide.

MILITARY BALANCE

The following force assessment of the military balance between the U.S. and U.S.S. R. focuses on the quantitative comparisons of men and equipment and the positioning of present NATO and Warsaw Pact forces. Major U.S. programs and costs proposed in the FY 1977 budget are identified within each mission area of strategic and general purpose forces. Further, the outyear implications resulting from FY 1977 budget proposals are also identified. No attempt is made to address the quality of forces. Not only is there difficulty in addressing such values, but the variables of conflict could render such compari

sons invalid. The amounts indicated herein reflect the January 1976 budget request and do not include figures for subsequent budget amendments.

1 Includes deployed, strike-configured operational aircraft only.

2 The total throw-weight of the delivery vehicles is approximately the same for both nations. The U.S.S.R. holds a sizeable lead in missile throw-weight which is offset by the U.S. lead in bomber payload capability. Throw-weight is defined as weight of missile above last boost stage with projected or programed configurations 1,000's of pounds. Bomber payload is 1,000's of pounds representative aircraft loading rather than maximum or worst case.

INTERCONTINENTAL BALLISTIC MISSILES (ICBMs)

Soviet ICBM programs center around the older SS-7, SS-8, SS-9, SS-11, and SS-13 systems, many of which will probably be phased out during the next decade and replaced by the SS-X-16, SS-17, SS-18, and SS-19. The new ICBM family offers significant growth in target accuracy and multiple independently targetable reentry vehicle (MIRV) capability. There are indications that testing has been conducted on a land-based mobile capability for the SS-X-16; however, deployment in the near future appears unlikely.

U.S. ICBM programs are the Titan II (54), Minuteman II (450) and Minuteman III (550) missiles. Major development efforts in the DoD FY 1977 budget associated with the ICBM program are: (a) Minuteman (silo hardening, guidance, and target accuracy improvement), (b) M-X missile (advanced development of a new ICBM and mobile basing options), and (c) advanced ballistic reentry systems capability (ABRES). The costs for these efforts are: