The following argument against the military draft was written while I was at the Rand Corporation during the summer of 1957. It was never published, and I had lost any copies until recently, when Charles Wolf of Rand found it on microfiche in the files. I had submitted it as a working paper. It was rejected. Fifty years ago, the Air Force provided most of the financial support to Rand, and Air Force leaders were convinced that they benefited from the draft—that many high-quality potential draftees volunteered for the Air Force only because they preferred serving there to serving in the Army. As I recall, I was told informally that the Air Force’s attitude toward the draft was why Rand had turned down this paper.

I was discouraged by the Rand rejection. I also felt that the opposition to a fully voluntary military was so powerful that the United States would never abandon the draft. Naturally, I did not anticipate the Vietnam War and the enormous reaction against it, especially by young people subject to the draft. This reaction and opposition led then-president Nixon to set up the Gates Commission in 1969 to consider whether to keep the draft. Fortunately, Nixon asked Milton Friedman to serve on the commission, and he was a persuasive advocate of fully voluntary armed forces, turning an initially divided commission into one that unanimously supported ending the draft. The draft ended in 1973. Since then, other countries—France and Germany, for example—have been persuaded by the U.S. example to end conscription.

This experience taught me that ideas and policies that are politically impossible at one time can come to fruition later if events and circumstances change. Analyzing defects in public policies can create intellectual ammunition, if you will, for dramatic policy changes not yet on the horizon.

More than 30 years have passed since the abolition of the draft and the introduction of the voluntary armed forces. The performance of the volunteer military has exceeded all that I expected in my 1957 paper. The proportions of minorities and people from poorer backgrounds in the armed forces are larger than in the population at large, but these groups are far less than a majority of all enlistees as well as of officers. Moreover, opportunities for advancement in the military by African-Americans, such as Colin Powell, and other minorities are at least as great as in the civilian sector. Use of drugs by enlistees is less than in the civilian workforce of similar ages. Morale has been high, and the volunteer armed forces have performed well in all the wars since they were introduced, including Iraq, despite difficult circumstances.

Fifty years ago, I felt that opposition to a volunteer military was so powerful that America would never abandon the draft. I did not anticipate the Vietnam War and the enormous reaction against it, especially by young people.

Another measure of the volunteer Army’s success is the lack of any significant movement in the United States to bring back the draft, even during the current unpopular war as the armed forces are being stretched to their limit.

One important omission from my 1957 analysis is the social cost imposed on taxpayers by all-volunteer armed forces. The tax burden of a large voluntary Army during wartime could be substantial, because much higher pay might be needed to attract a large number of volunteers. Yet while the tax increase to finance a war that doubled or tripled the size of the armed forces would not be negligible, it would still be a small part of the federal budget. A voluntary Army might become too costly only if military size increased far beyond such levels, which is unlikely under present weapon conditions.

I am indebted to Charles Wolf for obtaining a copy from the microfiche files at Rand, and to Charles Lindsey and his staff at the Hoover Digest for preparing the text from a difficult-to-read copy. To preserve the original document I have not changed the text, even though it is obviously only an early version, and I am not satisfied with some of the statements and how they are expressed. (The references to drafting men are not really an anachronism, because women were exempt.) In the interest of saving space, I have omitted some sections, including a discussion of the close relationship between a voluntary army and one where, as during the Civil War, draftees could buy substitutes. That discussion is available on request.


(19 August 1957)

During the last election, [Adlai] Stevenson proposed that conscription be abolished; it received very little general support and was frequently viewed as a purely “political” proposal by a desperate candidate. The reaction to this proposal may seem surprising since conscription was never used in peacetime before 1940 and was really first used in wartime during World War I. Five of our seven major wars have been fought without effective conscription. Or, to look at the problem differently, while the United States is the major advocate of free choice of occupation and condemns the use of slave or forced labor elsewhere, use of the latter is tolerated in a sector that employs about 5 percent of the labor force. This paper examines in detail the arguments for using conscription to secure military personnel.

The draftee is performing “slave” labor and the enlisted man “free” labor, and there is evidence that slave labor is less efficient than free labor, particularly in the skilled jobs. Slave labor has traditionally been condemned on economic as well as on moral grounds.

It is often claimed—particularly by military men but also by economists—that the military couldn’t obtain (say) three million men without conscription. This assumes either that the present military pay scale is the only possible scale or that the supply curve is completely vertical. In this naive form the argument is clearly erroneous.

A more sophisticated version of this argument is that it would be too costly to raise a voluntary armed forces. By “too costly” is not meant too costly to society but too costly to the general taxpayer, since conscription merely shifts some of the costs to draftees from the general taxpayer. Draftees are, as it were, hidden taxpayers. A voluntary program may increase costs to the general taxpayer, but the increase is apt to be overestimated precisely because conscription hides these costs. Therefore, if the cost to the military rose—i.e., if military costs became more equal to social costs—there would be a decline in the demand for men. The change in personnel costs does not decline, however, since the military can hire fewer men at a higher price partly because they use more hardware, and the increased expenditure on hardware must also be included in the estimate of the change in costs. The cost of getting a given amount of military “output” would rise, but perhaps by not very much; for example, if men and hardware are perfect substitutes, the cost would be unchanged.

There is evidence that hardware and manpower are good substitutes; e.g., the greater the number of men available the smaller the inventory of hardware that would be desired; or to look at a broader substitution, the orga-nization of the Chinese army is surely based on more manpower relative to hardware than ours is.

The increase in costs to the military depends also on the elasticity of supply; for example, if the demand for men were completely inelastic, the increase would depend only on the elasticity of supply. Supply is probably most elastic to those military activities that pay a wage most comparable to the nearest civilian alternative. Direct knowledge of these elasticities is very limited; all the evidence indicates that the elasticity of supply to a civilian occupation or industry is extremely high. Moreover, in 1947–1948 the military was able to raise about 11/2 million men with low wages relative to the civilian sector. My guess is that the military would have to increase wage rates by no more than 30 percent on the average to raise a force of 21/2 million to 3 million men.

So far only the supply and demand for men has been considered, although, of course, the military demands effort. If the effort supplied per person was independent of whether he was a draftee or a volunteer, the supply of men could be used to represent the supply of effort. But there is an important dependency: the draftee is doing something against his will and has no incentive to perform his tasks above the minimum required to avoid trouble; the enlisted man has voluntarily accepted his job and has the usual incentives to keep his job and improve his lot. To put this in other more colorful language, the draftee is performing “slave” labor and the enlisted man “free” labor, and there is evidence that slave labor is less efficient than free labor, particularly in the skilled jobs.

This latter point is particularly relevant today, since the fraction of skilled personnel in the armed forces has risen strikingly. The difficulty in quantifying this difference should not cause an underestimate of its importance: slave labor has traditionally been condemned on economic as well as on moral grounds. The more important this difference is, the less likely that a voluntary armed forces would be more expensive to the military than a conscripted one; it may even be cheaper.

If “duty” requires conscription to provide men for the armed forces, why doesn’t it also require conscription to provide men for aircraft factories, Civil Defense units, or the State Department?

There is another reason why a voluntary armed forces might not be much more expensive and perhaps even less expensive than a conscripted one. The military provides and pays for much training for its personnel. The cost of this training can be recouped only if the trainee enlists. It is more likely that a volunteer will reenlist than that a draftee will, since the former has shown an interest in military life. Hence the net costs of training volunteers would be less than the net cost of training draftees.

After all these factors are considered, it no longer is clear that a voluntary system would be more expensive to the military than one using conscription; the wage rate would be higher with a voluntary system, but the effort expended by the average person would also be higher, and the number of men and the amount spent on training would be lower. It is not at all unlikely—to judge from the little evidence available—that the latter changes would more than offset the higher wage rate.

To move away from economic arguments, it is often alleged that the draft is desirable because it is the duty of every citizen to serve his country. It is difficult to see the meaning to attach to this statement. It may be the community’s duty to provide resources for the military, but this occurs under both schemes. Why is there any argument based on “duty” which implies that draftees should pay the costs of having an armed forces rather than the general taxpayer? Surely the age, sex, etc. distribution of draftees is much more arbitrary than that of the general taxpayer. Once it is seen that much of the difference between a voluntary and a conscripted force is in who provides the revenue for the military, arguments supporting the draft in terms of “duty” disappear. Or to look at this from another viewpoint, if “duty” requires conscription to provide men for the armed forces, why doesn’t it also require conscription to provide men for aircraft factories, Civil Defense units, the State Department, etc? Surely they are all equally important and equally in the country’s interest, yet most persons would be horrified at the thought of conscripting men for the State Department or factories.

Those obtained under a voluntary system would be more, not less, “patriotic” than those obtained with conscription, if “patriotism” is defined as the willingness to give up money to serve one’s country.

It is often said that it would be bad to obtain men by paying them since a good armed forces depends on patriotic spirit. I do not see why a voluntary force is inconsistent with patriotic spirit or why a conscripted force is a particularly patriotic one. If patriotism was a sufficient motivation for entering the military, all the men required could be voluntarily obtained even at low wages. Since this is not true, most people must be unwilling to forfeit much income for a feeling of serving one’s country. Enough men could be obtained either by drafting them or by raising the real income in military service. The draft is prima facie evidence that patriotism is insufficient; real income could be raised either by raising money income, or by using propaganda to raise the amount of “patriotism.” Even if real income were raised by raising money income, those obtained under a voluntary system would be more, not less, “patriotic” than those obtained with conscription, if “patriotism” is defined as the willingness to give up money income to serve one’s country. A voluntary force always obtains those men most willing to forfeit income, while a draft, because of its random nature, obtains men with an average amount of patriotism. If real income were raised by raising the amount of patriotism, a fortiori a voluntary armed forces would be more patriotic than a conscripted one.

Another argument often made is that a voluntary force is a professional one and that a professional force is a constant threat to a democracy, while the heavy turnover with a draft would prevent the cohesion necessary for military subversion. Several comments are relevant. The first is that the present system encourages professionalism in the officer corps, the sector posing the biggest military threat to civilian control. The second is that turnover could be obtained with a voluntary armed forces by limiting the number of reenlistments. This presumably would raise the cost of securing personnel but would be quite feasible. The third is that knowledge on this subject is limited. Countries like Germany, in which the military did obtain an excessive amount of political power, were also countries in which the military traditionally had a prestige and importance way beyond that in this country. Furthermore, the German army had a professional officer corps with conscripted enlisted men, so it was as similar to our present system as to the proposed system. Civilian control of the military can be maintained only with constant vigilance, but there is no evidence that this would be more difficult under a completely voluntary system.

It is often said that the desirability of having a reserve of trained manpower strongly argues for a continuous draft of men to be trained for a fixed period and then released. But a draft is not necessary since volunteers could be obtained for short periods by an appropriate wage offer. Again the preference for a draft implies a strange preference for taxing young, physically able persons rather than the average income recipient.

However, the alleged advantage of a reserve force can itself be brought into question. In discussing this we distinguish three kinds of wars: all-out thermonuclear war, limited war of the Korean type, and large-scale war with thermonuclear and other extremely destructive weapons excluded.

A reserve force would be almost useless in all-out thermonuclear war as the outcome of such a war would depend on the quality of the fighting force in being; the quality of this force would be better if it were an experienced professional force than if it were an inexperienced drafted one.

Mobilization of a large reserve force would also be of little use for a limited war as again the outcome would closely depend on the existence of an efficient force that could bear the brunt of any fighting. Here, too, the voluntary professional force is to be preferred.

For a large-scale war with limited use of weapons the ultimate size of the armed forces would be much above 3 million; the argument for a reserve for this war is that it would enable the military to increase its size rapidly. The value of this depends on the gains from the more rapid increase and the costs of doing it. It is relatively easy to obtain some idea of the costs. Suppose a million men were to be taken each year for a one-year term and then released. If they could have earned an average of $3,000 during that year as civilians, the total opportunity cost of their service would be $3 billion a year. To this must be added the cost of the personnel and equipment used in training, which (say) comes to about $1 billion per year; adding together these costs gives a yearly cost of $4 billion. The total cost of this program depends on the date at which war begins; if 10 years elapse the total cost (ignoring discounting) is $40 billion; if 20 years elapse the total cost rises to $80 billion.

Conscription is inequitable, causes bad feeling, and unnecessarily increases the uncertainty surrounding decisions. Fewer than 150,000 men are drafted each year, although about 1.3 million males reach age 20 each year.

The gains depend on the rapidity of the rise and the advantages of a more rapid rise. The former depends largely on how well training is retained; e.g., if more reservists “forget” their training, or if technological advance makes their training outmoded, the force could not be increased much more quickly by having reservists available. The latter depends on the extent of mobilization of the enemy, and other aspects of the war. Since costs of a reserve program would be substantial, the gains would have to be quite large to overcome them, and in light of present knowledge I would guess that they are not. Note also that the $4 billion which would be used to train reservists could be used to increase the size of a professional army, thereby decreasing the need for a rapid build-up. Note finally that only in this third type of war would the reservists be of any use.

Arguments in favor of conscription appear very weak when subjected to close examination. The case against conscription is strengthened still further by some additional arguments.

Involuntary servitude is immoral. This is the basis for our whole economic system and is a strong, although not conclusive, argument against any proposal to make use of it.

Conscription involves a misallocation of resources. It results from the different skill structure in civilian and in military activities. A draft obtains men with a more or less representative distribution of civilian skills; there would be too many men in this group with skills little used by the military and too few with skills greatly used by the military. This loss may be extremely large even though there is no a priori way to estimate it.

Conscription is inequitable, causes bad feeling, and unnecessarily increases the uncertainty surrounding decisions. Fewer than 150,000 men are drafted each year, although about 1.3 million males reach age 20 each year; if the military only drafted 20-year-olds the chances that anyone in this age group would be taken would be less than one in eight. Most of them would not be, and this would cause much ill-feeling, regardless of whether the criterion to exclude some was physical condition, occupation, education, marital status, etc. Even when the draft is confined to 20-year-olds there would be some uncertainty surrounding one’s chances of being taken. When the draft takes men of different ages (as at present) the uncertainty is multiplied; those eligible do not know if they will be taken this year, next year, etc.

These arguments against the draft apply not only to the military as a whole but also to individual services, even to those like the Air Force which do not take draftees. It is often said that the Air Force is better off if the Army uses the draft, since this encourages some to avoid the draft by enlisting in the Air Force. However, these “volunteers” to the Air Force are really draftees since they would not “volunteer” if the draft were not used by the Army. All the arguments against conscription apply equally to these “volunteers,” and thus are equally relevant in formulating Air Force policy.

These arguments strongly argue against the draft on economic, moral, equity and military preparation grounds. Next we briefly consider the kind of income that should be given to volunteers. We have already ruled out a capital transfer to the volunteers from individuals selected at random, that is, a system of buying substitutes. Capital transfers from the government to the volunteers should also be excluded since this encourages desertion and insincere enlistments.

The alternative is to increase the real wage rate. This can be done by increasing money wages, providing rent-free housing, better pension plans, free education, more prestige for military men, increasing the variety of goods in the PX, etc. Military men in their testimony before Congress usually emphasize housing, prestige, pensions, etc. in addition to money income. I suspect that a major reason for doing so is the feeling that Congress would be unwilling to raise money income very much. The ideal solution would be to raise income primarily through raising money income and not through housing or pension plans, because of the preference for generalized purchasing power. There may be some exceptions to this in isolated areas where government provision of housing and other goods would be a way to avoid a private monopoly position.

The military should be a respected occupation, but I would not raise real income primarily through increases in prestige. I want to avoid a situation like that in Germany where the prestige of the military was so great as to set the military above the civilian sector.

overlay image