implication of constitutional evasion, Congress has demonstrated its use, and it cannot be put aside as improbable. That it now has at least some measure of sanction from the Supreme Court of the United States is apparent from the following language of Mr. Justice Cardozo in his opinion in the Machine Company case holding the Social Security Act itself constitutional:

"The proceeds, when collected, go into the Treasury of the United States like internal revenue collections generally. Section 905 (a). They are not earmarked in any way.

*

* *

"The appropriations when made were not specifically out of the proceeds of the employment tax, but out of any moneys in the Treasury. * * *

*

"The proceeds of the excise when collected are paid into the Treasury at Washington and therefore are subject to appropriation like public moneys generally. * * No presumption can be indulged that they will be misapplied or wasted. Even if they were collected in the hope of expectation that some other and collateral good would be furthered as an incident, that without more would not make the act invalid. * * * This indeed is hardly questioned. The case for the petitioner is built on the contention that here an ulterior aim is wrought into the very structure of the act, and what is even more important that the aim is not only ulterior, but essentially unlawful. In particular, the 90 percent credit is relied upon as supporting that conclusion. But before the statute succumbs to an assault upon these lines, two propositions must be made out by the assailant. * * * There must be a showing in the first place that separated from the credit the revenue provisions are incapable of standing by themselves." (Chas. C. Steward Machine Co. v. Davis, 301 U. S. 548.)

It is perfectly obvious that a State may supplement unemployment funds with money taken from its general treasury; and it may supply its general treasury by taxes of a surprising number of kinds applicable to railroads. The various forms of taxes discussed by Mr. Justice Stone in the passage which we have quoted above from the Western Live Stock case are sufficiently illustrative, though the catalog so presented is by no means complete.

Against taxes of one or all of the sorts there mentioned, and against taxes of many additional sorts which the States have power to lay, and upon occasion to lay, this bill offers no pretense of protection, unless, indeed, it be claimed that protection is to be found in the provision of section 13 (g) which proposes to amend section 303 of the Social Security Act by providing that the Social Security Board shall not certify funds to any State "if it finds, after reasonable notice and opportunity for hearing to the State agency * * * (2) That such State is failing to afford reasonable cooperation with every agency of the United States charged with the administration of any employment insurance law."

This is a curious provision. It constitutes one Federal bureau a monitor over the States for the protection of another Federal bureau. The monitor bureau may summon a sovereign State to appear before it for trail of the issue of the reasonableness of acts of the State legislature for failure of the State legislature to administer the State power of taxation in such a way as the bureau sitting in judgment shall decide is "reasonable cooperation" with the other Federal bureau. The decision apparently is to be final, there being no provision for review. The bill does not pretend to define "reasonable cooperation" being content that the bureau be limited only by its own notions.

To find protection here against taxes not earmarked for unemployment relief one must indulge the assumption that the bureau may inquire whether the State is concealing the real purpose of the tax, an inquiry ordinarily beyond the power of the Federal courts, and must assume that the opinion of the bureau is to prevail over the announcement of the legislative assembly of the State and perhaps over the decisions of its highest courts.

If this is to afford protection to the railroads the lawfulness of the tax must not be permitted to hamper the bureau in the exercise of its supervisory function. It must sit in judgment not upon the question whether the tax laid by a sovereign State is lawful, but upon the question whether, lawful or not, it coincides with the view of the bureau as to what is "reasonable cooperation"; and thus its powers go far beyond those entrusted to any court under our governmental plan.

Of course such absurdity will not be within the intent of Congress if it adopts this bill; but that is merely another way of saying that this provision affords the railroads no protection against a double burden.

CONCLUSION

General unemployment relief is a lawful function of State government. It costs money now, and if the States succumb to demands for increased payments similar to the demand of railroad employees embodied in this bill, it will cost. 73643-38--16

vastly more.

The States have the power, and they will be compelled to exercise it, to get this money through taxation. Regardless of whether it is possible now to foresee the exact detail of future events or to demonstrate them by mathematical computation, at least the course of future events is obvious insofar as they relate to railroad participation in all public burdens, State or Federal.

It is clear from what has been shown above that there is only one way by which the railroads could be safeguarded against being compelled to contribute to the support of the separate State unemployment systems after their removal therefrom as proposed by this bill. That way would be for Congress, by valid enactment, in clear and unmistakable terms, to withdraw them entirely from the whole field of State taxation. Even if it be conceded that Congress, as an incident of its power to regulate interstate commerce, could prohibit all State taxation of interstate railroads (all of which are also intrastate railroads, exercising important rights and privileges granted by the various States), is thare any member of this committee who sincerely believes that such power will ever be thus fully exercised? Promises breed demands for more generous promises, as illustrated by this very bill. Some day these demands may be hushed by the voice of reason, but that time is not yet. Common experience indicates increased cost for State unemployment systems, and despite the treacherous protection pretended by the equivocal language of this bill, despite any language which might conceivably be put in the bill, the railroads must share these burdens when the States assume them, just as the railroads have always been called upon and expect always to be called upon to bear any other burden incident to the operation of State govern

ments.

The Nation is in the midst of a financial depression unprecedented in sudden approach and almost unprecedented in gravity. It has an unemployment roster for which the wisest have been unable to find a solution. The Senate Subcommittee on Relief and Unemployment on April 20 filed a preliminary report estimating that 12 million workers are totally unemployed and more than one-seventh of the population of the Nation are receiving public assistance. In the midst of these conditions the meager resources of industry are being depleted under the Social Security Act, not to combat the present depression or to temper present unemployment, but to provide against some other condition of unemployment which may arise at some indefinite time in the future.

Of all industry, perhaps the railroads most conspicuously arouse public concern. Certainly they are not in position to be selected for further experiment. This is no time for further experimentation. Let new experiment wait until the experiment now under way at least approaches completion; and let legislation which would add impetus to demands for further devotion of present resources to future needs wait until money is available to meet the needs of the present.

Mr. MARTIN. I might observe that the trend of tax legislation, for both railroads and their employees, under the Social Security Act, is to have them wholly support their retirement plan without any outside aid or contribution from the general taxpayers or anybody else. At the same time, they are paying their share of the general taxes to support the rest of the Social Security Act. I do not know that there is any way of curing that, but that condition does exist.

I have been reminded here this morning of a magazine article that I read 25 of 30 years ago, and I would give a thousand dollars to have it now. That was an article in which the writer predicted that eventually there would be substantially only two forms of government in this country, the Federal Government and the municipal governments, and that the State lines would practically amount to no more than county lines have heretofore amounted to. If we have not reached that time, it is coming. If this legislation is to go ahead, with the economic system we have in this country, that will be the situation.

Our corner grocery store and corner drug store are national industries, owned by gigantic national agencies. We are just driven more and more to place the country under Federal regulation.

We thank you for your statement, Mr. Souby.

This concludes the hearings for the opponents, and we are now ready to proceed with the rebuttal testimony. Mr. Latimer, the Administrator of the Railroad Retirement Board is present, and I presume there will be no objection to Mr. Latimer proceeding first. STATEMENT OF MURRAY W. LATIMER, CHAIRMAN, RAILROAD RETIREMENT BOARD

Mr. MARTIN. Before you proceed, I might say that I have a few questions to ask, some of which were given to me, but if you desire you may proceed without interruption.

Mr. LATIMER. Mr. Chairman and gentlemen of the committee, I would like to deal this morning with the question of the cost of unemployment insurance for railroad employees as provided in H: R. 10127. I will deal with the question in two parts, the first part relating to the cost of the benefits themselves, and the second part to cost of administration.

Any estimates as to the cost of unemployment-insurance benefits applicable to conditions at the present time must necessarily be based on data taken from different sources and pieced together, in part at least, by a process of estimation. All estimates as to the rate of unemployment are like any other estimates dealing with the future, in that they are subject to a margin of error. Since unemployment is a variable risk, falling off in years of prosperity and increasing in years of depression, a reasonable estimate ought to be based on data covering a period of some years.

The most comprehensive study of unemployment in the railroad industry was that made by the Federal Coordinator of Transportation covering the years 1924 to 1933, inclusive. That study as I believe, has already been pointed out in the hearings, was made largely from personnel records; the individuals studied were those on the pay rolls of the railroads included in the sample for the first pay roll periods of July 1924 and July 1929, and the last pay roll in December 1933. Because of the fact that the time elapsing between the dates of direct pay roll checking was so large, a fairly substantial number of short service employees were omitted from the Coordinator's study. For the purpose of determining the costs of an unemployment insurance system providing benefits for all those becoming unemployed, irrespective of the length of their railroad service, the Coordinator's study would have been seriously inaccurate. Since no such system has every been proposed, and since, on the contrary, all the proposals thus far formulated require fairly substantial employment as a qualification, or relate the length of benefit payments to the length of employment, or both, some of the omissions from the Coordinator's study may have no material effect on the cost of unemployment insurance except insofar as they may have added to pay rolls without adding to benefits. Of the employees included in the Coordinator's study, the proportions affected by unemployment during the years 1924-33, ranged from a low of 13.01 in 1928 to a high of 30.47 in 1932. That represents the number of persons becoming unemployed 1 day or more before becoming reemployed.

The average annual percentage of employees affected by unemployment over a 10-year period was 18. That was shown by the Coordinator's study.

Since the Coordinator's study was completed, some additional information as to the rate of becoming unemployed in railroads has. become available as to the number of persons becoming unemployed, and it is naturally desirable to test some more recent information against the results obtained in the Coordinator's study. These data. are largely those collected by the Railroad Retirement Board in the course of the administration of the Railroad Retirement Act. For that purpose we have picked out two railroads, one a typical eastern railroad and one a typical western railroad. These two railroads employed during the year 1937 some 36,000 different employees, having a maximum at any one time of about 26,000 employees. We do not claim completely to be able to determine whether these railroads are exactly typical, because we have not been able to compare them with the railroad universe. These studies show that about 59 percent of the 36,000 employees who worked at some time during that year worked in each of the 12 separate months, so that, so far as the railroads are concerned, at least 41 percent must have become unemployed or been unemployed during the course of the year 1937.

Now, that 41 percent gross is subject to some correction in both directions: The first would be the employees whose service terminated through retirement under the Railroad Retirement Act; some died, and some became totally disabled, and, therefore, would not be eligible for compensation. Some of them did not serve the whole year, or came into the service after the beginning of the year, having been employed elsewhere in the early part of the year, and did not become unemployed, so far as the railroad industry was concerned, until the end of the calendar year. Our actuary submits as a reasonable estimate, that the total number of persons on the two railroads. who became unemployed for one or more days during the year is 40 percent. We have, therefore, in making estimates of the cost of paying benefits under H. R. 10127 pieced together the rates of unemployment from the 1937 data with the duration of unemployment data derived from the Federal Coordinator's study. A conclusion as to the reasonableness or unreasonableness of that figure, must be based to a considerable extent on the question of whether or not the rate of becoming unemployed in 1937 was reasonably representative of a period of time including good and bad years.

We have constructed a rate of becoming unemployed from I. C. C. data, which is necessarily imperfect for the purpose, but which is the only thing obtainable. The method followed was to summate declines in the index of employment declined using the months in which unemployment increased, year by year. These sums, for each year, were divided by the employment index for the year to arrive at a relative rate of becoming unemployed for the years 1924 through 1937. The results show that the rate of unemployment so calculated, was about 25 percent higher in the year 1937 than it was for the period from 1924 to 1933, and that it was, in fact, the same as for the 5-year period from 1929 to 1933, inclusive. The figures for comparison are, of course, based on comparable data, and though this does not represent the actual rate of unemployment, nevertheless, relatively speaking, I know of no reason to suppose that there is any more distortion of the figures than there would be in any others obtainable. Of course, in the latter part of 1937 the

decline of employment was probably swifter than in any corresponding short period of time, or for any period for which railroad employment data are available.

A second question relates to whether, or not, the duration figures on which the calculation has been made are reasonably typical. Another calculation was made based on the I. C. C. data, cumulating the net declines in the employment indexes for periods up to five months, and dividing the cumulations in each year by the employment index during the year to arrive at the relative amount of compensable unemployment. The results indicate that undoubtedly the duration of employment on the average was 33 percent longer in the period from 1924 to 1933 than in 1937; so if we had calculated the figure on the 1937 basis, undoubtedly the figures as to cost would have been too low.

The final question is as to the Coordinator's figures as to the duration of unemployment. This rate, as a matter of fact, indicates a longer duration of unemployment than do the figures prepared by the actuaries of the State agencies. They might be increased somewhat, however, if all of the short service personnel had been included, although that cannot be stated with any assurance. The duration of unemployment, of course, is affected not only by reemployment by railroads, but by employment elsewhere.

Some studies made by the Railroad Retirement Board with W. P. A. funds as to the duration of the reemployment of persons who were furloughed from railroad service, indicate that reemployment rate short of service railroad personnel is higher than for long service so far as jobs outside the industry are concerned. So I think it is not an unreasonable assumption that those two factors may cancel each other. At any event, the calculations show that we still have a fairly stiff rate of unemployment, so far as duration is concerned.

I would like to introduce at this point for the record four tables. The first table, dealing with railroad A, which is a typical eastern railroad, shows the number of employees, classified by the number of months in which they worked for the railroad during the year 1937, and by the amount of their annual earnings by $100 intervals.

(Table 1 is as follows:)