should be safeguarded to the fullest possible extent if the insurance carrier is to serve a useful purpose in such a field. Under present conditions this is not being done in the case of reciprocal exchanges. Their primary purpose appears to be to afford a remunerative activity for the promoter or attorney-in-fact. In short, so many objections can be raised against reciprocal exchanges as at present conducted, that one fails to see how they may find a place of usefulness in relation to a branch of insurance so peculiarly social in its character as that of workmen's compensation.

CONCLUSION

Through the maze of controversial claims made for the various forms of insurance of workmen's compensation, one sees clearly the single fact that no particular form has reached such a point of excellence that it stands out as superior. The various experiments now undergoing the test of practical experience afford the opportunity for future appraisal. All these forms of insurance must be judged upon the basis of security, cost and service. Self-insurance and inter-insurance seem fairly indictable and doubtless will be the first to pass away. The mutual principle, whether applied through private or through public management, bids fair to survive, even if only as a competitive force inhibiting monopoly by the joint-stock companies. There is no apparent reason why mutual insurance carriers worthy of public confidence and patronage should not develop. The effect of competition thus far indicates a gradual drift in the direction of the strongest and most efficiently managed stock and mutual companies. The largest question involved in the general problem is whether the insurance of workmen's compensation should be conducted competitively or as a monopoly.

Against competition the argument is a compelling one that lost motion results from the agency system which system unnecessarily absorbs a large proportion of the premium contribution. Further waste is introduced at the point where state supervision is invoked to promote equal competitive opportunity among the carriers. If a monopolistic system is created which form shall it assume? The probability is strong that society will not legislate in favor of a proprietary monopoly however efficient. The mutual association and the state fund will be the most likely candidates. Of the two the mutual association should have the greater appeal

because of its representative form of government, its consequent responsiveness to the will of its member-policyholders, and the divorcement from political interference which its scheme of organization renders possible.

Whatever may be the ultimate development of workmen's compensation insurance, it is reasonable to expect that society will insist upon the greatest security and the best service at the minimum cost to itself.

THE CALCULATION OF WORKMEN'S COMPENSATION

PREMIUM RATES

BY CLAUDE EDWARD SCATTERGOOD,

Assistant Secretary, The Fidelity & Casualty Company of New York.

As there are various agencies for the administering of workmen's compensation insurance, likewise there are varying conceptions of rates. The mutual company desires to return as large a dividend to its assured as possible; so the excess of the rate above adequacy, though a matter of some concern, has not the same bearing upon its calculation as in the case of the stock company; and a similar observation may be made with regard to the state fund, which also desires to return a dividend. Rates now in use by stock companies have been calculated with the idea of simply carrying the business without underwriting loss. It thus becomes necessary to set the limits of this discussion of workmen's compensation insurance rates, and they will be considered in the light of adequacy alone, with no reference to a dividend return.

Insurance is the creation of a fund by means of contributions of comparatively small sums by the many to pay comparatively large amounts of indemnity to the few. This is insurance in its broadest conception, and the contrast between the many who escape disaster and the few to whom it comes is narrowed in the case of workmen's compensation, because a large number of accidents are constantly occurring which individually occasion small payments, but which are heavy in the aggregate. Serious accidents causing death or permanent disability with heavy payments of indemnity are not so frequent as the slight accidents just mentioned, and occur to comparatively few of the employers contributing to the fund, but they too are heavy in the aggregate. It is against these serious accidents that an insurance system most aptly applies.

In order to establish a fund, contributions are secured in the form of premiums; and premiums are calculated by taking a unit of exposure to risk, multiplying that unit by the rate, and then multiplying the product so obtained by the number of units of exposure.

The theoretical exposure is the number of men working during a year reduced to the equivalent of the number of men working a standard number of working days during a year at a standard number of hours per working day. The practical exposure, and the one used, is the pay-roll of those men for a year. The pay-roll basis approximates the theoretical basis, and clears away the difficulty of men not working a full year or working at part time during portions thereof. If we can approximate the yearly wage of a man working a whole year, the number of men working a whole year can be obtained by dividing the total yearly pay-roll by the average yearly wage of one man. If the pay roll is $1,000,000 and the average yearly wage $600, the equivalent number of men exposed to risk for a year is 1,000,000÷600=1,667.

The practical exposure being the yearly pay-roll of an employer, a practical unit of exposure must be adopted, and this is $100 of payroll. The ratio of the benefits paid injured workmen or their dependents to the number of one-hundred-dollar units of pay-roll is the basis of workmen's compensation rates, and is called the "pure premium." It is that rate per $100 of yearly pay-roll which would be charged after all corrections had been applied, were there no expenses in the administration of the business. These corrections will be considered later.

As accidents vary in frequency and seriousness among employers in different lines of business it would be unfair to determine "pure premiums" for all pay-rolls of all classifications of employers combined. A generally recognized standard list of industrial classifications, about 1,500 in number, has been made, and rates for each of these classifications have been determined.

In order to determine as dependable pure premiums as possible it becomes necessary to secure the largest possible experience upon individual industrial classifications. It is unnecessary to enlarge upon or to illustrate the proposition that the larger the exposure to risk the more dependable the average pure premium per unit of exposure ($100 of pay-roll). The work of the National Workmen's Compensation Service Bureau, composed of about twenty stock companies administering workmen's compensation insurance, is probably the best example of the calculation of pure premiums by

1 See Standard Manual of National Workmen's Compensation Service Bureau, 13 Park Row, New York City.

reason of its being in possession of more workmen's compensation statistics (and all on a uniform basis) than any other organization. The companies comprising the bureau have for several years been reporting their statistics. These have been reported by states, within states by years of issue of policies, within such issues by manual industrial classifications, and within manual classifications, among other variations, by kind of benefit paid to injured workmen. To these statistics reported by the companies to the bureau has been added such additional experience collected by the insurance departments and industrial accident boards of various states as was on a basis uniform with the bureau's experience, and it is a matter of congratulation that the methods of the state bodies and the bureau in the matter of statistics have in a number of instances been practically uniform. The bureau being in possession of this experience has combined it by states, within states by year of issue, within years of issue by manual industrial classification, etc., thus producing as extensive and dependable experience as can be found. In such combinations corrections have been made to take care of varying cost of the workmen's compensation laws of the different states, so that all may be upon a known level.

Experience in the United States is in general tabulated upon a policy year of issue basis. This means that all policies issued, say in 1915, are considered as belonging to the 1915 unit. A policy issued in January, 1915, will expire in January, 1916, one issued in June, 1915, will expire in June, 1916, etc., so that the last policy to expire in 1916 on 1915 issues will be one issued on December 31, 1915. The indemnity and medical payments covered by these policies, when made, are ticketed with a 1915 issue mark, so that whenever, even during a long period of years, a payment is made on an accident covered by a 1915 policy it will be referred to the 1915 issue. The same procedure is adopted with payments incurred but not yet made on such accidents, commonly known as outstanding medical or indemnity. The pay-rolls on these policies are ticketed in a similar manner.

Thus a basis is established for computing rates that will take care of all medical and indemnity payments incurred during the year for which a policy is issued, so that when an employer pays his premium on a policy he knows it will cover all indemnity and medical aid for all accidents occurring during the period the policy covers. His premium for that year is a part of his business expenses and he