wholly in that field. This is the age of the multiple-cylinder engine, of conservation, and of the efficiency expert. The first rush when only the primary product and the primary effect were of importance has passed and an era has come in which the field must be more carefully gleaned.

In the field of insurance then the case stands as follows: insurance is an instrument with a primary effect, wholly good, of relieving uncertainty and preventing the effects of misfortune, and with a secondary effect, wholly bad so far as it goes, of decreasing individual responsibility and opening the door to possible fraud. What, under the urge of this increased modern sensibility to secondary effects, should be our reaction to this adverse quality in insurance?

I venture to assert the thesis that the public welfare demands that the operation of insurance shall not decrease the sum-total of individual responsibility. This is only asking insurance to clean up whatever bad conditions it has itself produced. It has produced a marvelous new order in which the complexities of modern life and business can be carried on in comparative security, but in doing so it has shaken the foundations of personal responsibility and opened the door to abuse. The public, even granting that what is good in insurance far outweighs what is bad, may nevertheless properly say to the insurance companies that their duty does not end when they have developed to the greatest possible extent the positive, helpful effects of insurance, but that they must also do all in their power to minimize its evil effects.

The implications of this public admonition to look after its secondary effects are significant. The way in which insurance solves the primary problem of relieving uncertainty is by mass-action, by grouping a large number of risks together and thus securing the working of the law of averages. The secondary problem must be solved in the same way, that is, by mass-action. It is impossible to replace individual responsibility on the basis of immediate self-interest; that is gone; the best that the insurance company can do is to substitute something else for personal responsibility and since it has already associated many risks together for the sake of securing an average it may very naturally take advantage of that fact to institute preventive actions of a collective nature whose sum-total will be equivalent to the sumtotal of the personal responsibility that has been lost.

There are various activities of a collective nature, their character depending upon the particular type of insurance concerned, that can be carried on by the insurance carrier to prevent loss; for instance, the excellent survey of conflagration conditions in the cities of the country by the National Board of Fire Underwriters, the standardization work of the National Fire Protection Association and of the

Underwriters Laboratories, the excellent work in factories of the New England Mill Mutuals, the nursing and tuberculosis work of the Metropolitan Life Insurance Company and of the German social insurance system. Each of these has either largely or completely made up for the effect of lessened personal responsibility among the assured; each of these is a piece of work that is possible only through collective action, and furthermore there is no agency which so naturally can undertake this work as the insurance carrier because the insurance carrier already has intimate relations with the assured and its interest is recognized to be direct and immediate.

Such activities as these are already in operation. All that I desire to do is to show that such activities are a logical and necessary corollary to the concept of insurance, that the institution of insurance is not a complete social instrument, not merely not adequate for work that needs to be done, but not complete in the sense of fulfilling its own intrinsic nature, until it makes use of mass-action not merely in averting the evil effects of misfortune but in preventing the misfortune itself. I believe the public may rightly insist for its own protection that the insurance carriers take this larger view of their responsibilities.

Insurance may thus become not merely one of the great fiduciary and distributive agencies of the country but one of the great conservators. Intermediate between the direct self-interest and personal responsibility of the individual and the action of the state for the benefit of its citizens will come this organized effort of the insurance companies, substituting mass-action for the individual efforts of its policyholders but yet more immediately controlled and inspired by self-interest than in the case of state action. With greater flexibility than state action and with greater power than individual action, with the theoretical effectiveness of socialism but without its dangers and practical inefficiency, the conservational activities of insurance may turn out to be as important an instrument in the organization of our life as its present distributive activities.

I have not undertaken to differentiate between the preventive possibilities in the various lines of insurance. This is not the place to do this in detail; in passing, however, I may acknowledge that the problem is very different in different fields. The intrinsic value of living and the distaste for pain are sufficient to make the secondary effects in life insurance and in workmen's compensation relatively much less important than in fire insurance, for instance, where the consequent loss of personal responsibility may be one of the largest contributory items in the make-up of the hazard, of such magnitude in fact that it will be a serious question whether preventive activities are feasible which will be a complete offset.

It is evident that in the last analysis insurance must be judged by its net effects; the good results of insurance in the elimination of uncertainty appearing on one side of the ledger and the results of lessened personal responsibility on the other. A more accurate formulation of my thesis would then be that the public welfare demands that this debit side be reduced to the greatest extent possible, and even where feasible turned into a credit. In some cases even with the best endeavor there may remain a debit item; in other cases this may be converted by the preventive work of the company into a credit so that in that case insurance both in its primary and secondary effects will be a positive public benefit.

One of the ear-marks of a correct solution or of a truly fundamental development is the fact that it clears up a larger field than that for which it was designed. That is the case here. The preventive activities of insurance are not only good theory but good business. The saving to be had is far greater than the expense involved. It is on this basis that the already very considerable developments in this field have been carried on.

In this connection I may point out, however, one more practical consideration which is not obvious at first sight. If the rates for insurance become too high an adverse selection takes place; there will be a tendency for the careful not to insure and the average character of the insured risk will deteriorate. This will immediately show itself in a bad experience and it will drive the rates still higher, and so on in a vicious circle. There is no way of breaking this circle from the inside except in a limited field where a compensatory selection is possible; that is, in general, insurance in its purely fiduciary and distributive capacity is helpless. The circle must be broken from the outside, that is, by bringing the preventive forces of insurance to bear upon the problem. Insurance for its own best interests and for the continuance of its existence must adopt preventive measures.

I may now connect this with the subject in hand. There is so little in Dr. Downey's paper that I have to disagree with, and that I can do other than commend, that my function is rather to pick up the loose ends of the program. One of these is prevention; it is not enough to consider workmen's compensation merely as compensation; workmen's compensation must bring about prevention or it is a failure. In the last analysis the fundamental question is what workmen's compensation has done to decrease industrial accidents.

What I have had to say with regard to the place of prevention in insurance applies to a very considerable degree to workmen's compensation even though there be no insurance system underlying it. For the shifting of the financial responsibility for accidents from the employee to the employer has the essential characteristics of insurance.

It measurably relieves the workman of uncertainty and of the evil effects of accidents. It places this responsibility upon the employer, and the very fact that he has a number of employees provides to a degree for the working of the law of averages and makes it feasible for him to institute action of a preventive nature which will be collective so far as his employees are concerned.

In some cases the particular plant is so large that for practical purposes it may be safely taken as an insurance unit; this is then self-insurance. In most cases, however, the number of employees is not large enough to yield a wholly dependable average and the employer seeks outside insurance, which has the effect of increasing the field over which the losses are spread.

In both cases, however, the general tendencies are the same, and the line of reasoning that has been developed can be applied. The very fact, for instance, that the law now places the responsibility for accidents upon the employer may to a certain limited extent develop a carelessness among the employees as well as a tendency to malinger. The compensation law looked at as an assurance of indemnity by the employer to the employee may well as a corollary, therefore, carry with it a compulsion to be exerted by the state upon employers to institute preventive action in their plants. As a matter of fact, exactly this has happened in a number of states. The minute that the obligation of the employer is fixed to pay compensation that minute he also becomes obligated, if not under the law then by equity, to prevent so far as he is able the accidents that the law requires him to compensate.

The situation is not fundamentally changed when there is insurance. To the extent to which the employer through insurance is relieved of responsibility to that extent the insurance company is obligated on the grounds of public equity to prevent so far as possible the accidents that it insures against.

I have purposely emphasized the fundamental equities of the situation rather than the hard business sense which underlies preventive action. But in passing I may say that the whole development of the rating feature of workmen's compensation insurance has given an enormous impetus to the prevention of accidents as a sound business policy. The nub of this lies first in the development of rates that so far as is humanly possible will exactly fit the hazard of each individual risk, and secondly in correlating these rates with the conditions and experience in the plants, first in as great detail as possible and secondly with the greatest possible clarity. When this has been done it means that the employer will be shown exactly where and how his accidents occur, exactly what effect certain accident-producing causes have upon his rate and exactly how much he can reduce his rate by

removing these malign conditions. Such rating as this acts as an incessant economic stimulus upon the employer to improve his plant.

As you have realized, the general tendency of my remarks has been in the direction of prevention because I have felt that the program was deficient in this respect; but at this point it is evident that what I have to say connects immediately with the subject of Mr. Verrill's paper, for the development of the statistics of accidents both as to nature and number and severity and cause provides the material for the type of rating system that I have described and, therefore, acts strongly in the direction of prevention.

In passing also may I refer to the very great accomplishments in the workmen's compensation field during the last ten years. I can thoroughly agree with Dr. Downey as to the things that still need to be done, and it is well on such an occasion as this to stress these; but, nevertheless, the progress during these first ten years of compensation has been remarkable. And this has been particularly so in the field of rating. Workmen's compensation probably presents greater difficulties in rating than any other ordinary type of insurance; it has all the complications of life insurance and of fire insurance and then some more of its own. There is much more to be done in making rates right, but, nevertheless, the main features of the problem have been solved and solved in such a way as to be already a most valuable instrument for accident prevention.

It is not an easy matter to make a statistical estimate of what has been accomplished by workmen's compensation in the prevention of industrial accidents, for the conditions themselves have been changing. It is much simpler to state what is possible and what has actually been accomplished in individual cases. It can safely be said that it is entirely feasible without departing from a sound economic basis to cut the cost of accidents (including the lowering of industrial efficiency) in the average industrial plant in two. I admit that this is an ideal that has not been generally attained and yet more than this has been accomplished in hundreds of plants. The net annual saving of the United States Steel Corporation in accident cost, after the cost of its safety work has been taken out, is stated by it to be a million dollars a year. It is not an uncommon thing for a plant to get a credit of 40 to 50 per cent in its insurance rate on the basis of its actual experience. Most of what has been accomplished can be credited to workmen's compensation either directly or indirectly. The problem of the future is how to extend this preventive action into all corners of the field and into all plants however small, and how to push it still further and make it permanent without at any time getting away from a sound economic basis. And, if I may again emphasize the thesis of this paper, this is largely a problem of insurance and very distinctly a problem of workmen's compensation.