On December 21, 1937, you asked that I consider ways of raising (1) $500,000,000 and (2) $1,000,000, in addition to present tax receipts. On December 29, I submitted a general memorandum on the subject, and promised more details in a few days. In a subsequent discussion, however, we agreed that I had better use my time on some other problems until this particular became pressing.

I am taking up this problem again in the present memorandum, since the imminent passage of the farm bill, the pressure for additional unemployment relief funds and (over the longer term) the proposed increase in armaments might easily increase expenditures by $1,000,000,000 a year.

THE POSSIBLE BUDGET DEFICIT. -- The budget for 1939 estimates expenditures at $6.2 billions, exclusive of debt retirement and also exclusive of $0.7 billions to be credited to old-age, etc., Thus a Federal "borrowing deficit" of $0.3 billion results, which, however, does not lead to a sale of Federal bonds to the banks and the public, if the States' unemployment trust fund is in a position to absorb as much as the budget message estimates (i.e., $0.6 billion).

If we take a pessimistic view, we can conceive of the recovery and relief appropriation for 1939 being raised by about $1.0 billion above the $1.1 billion total estimated for that year in the budget. That 1939 total includes $1.0 for W.P.A. and "Emergency relief Program," compared with $1.3 for the same item (all in W.P.A) for 1938. Perhaps a better comparison is between the 1939 $1.1 total and the (1938) $2.0 billion total for all "recovery and Relief." The point is, the 1938 total will apparently be increased by about $0.3 billion, making the 1938 total $2.3 billion -- compared with the $1.1 billion estimated in the budget for 1939. An increase of $1.0 billion or even $1.2 billion in the 1939 total therefore appears to be a possibility.

The 1939 budget allocates $0.6 billion to the agricultural adjustment program. No one seems to know how much more the new farm bill will cost, but I have seen an estimate of $0.3 billion reported in the press. If business conditions are so bad that the extra billion is needed for relief, revenues may well fall below the $5.9 billion estimated for 1939. Suppose there was a decline of $0.5 billion below this estimate -- then the effect of the added relief load (using the extreme figure of $1.2 billion), the added farm expenditure ($0.3 billion) and the decline in revenue ($0.5 billion) would be an increase of the Federal "borrowing" deficit to $2.3 billion. The sale of Federal bonds to the banks and public would, as noted above, be less than this, by the amount that the states' unemployment trust fund could absorb. It might not be able to absorb nearly the $0.6 billion estimated in the budget, however, if business continued poor. Possibly about $2.0 billion would have to be borrowed from the banks and the public.

These figures omit consideration of the added armaments expenditures, which apparently will not greatly exceed (if at all) $0.1 billion in 1939).

POSSIBLE SOURCES OF ADDED REVENUE -- Under these circumstances what would be the best way to raise (a) $500,000,000 (b) $1,000,000,000 -- assuming, for the sake of argument, that either of these sums should be raised?

For reasons given in my memorandum of December 29, the choice should be restricted to the personal income taxes, the death and gift taxes, the beer tax, the taxes on tobacco products other than cigarettes, and the miscellaneous excise and stamp taxes. A manufacturers' sales tax and import duties on coffee, tea, cocoa, etc., are possibilities, and so is an increase in the corporation normal tax, but none of these appear desirable. Increase in the payroll taxes, customs duties generally, the gallonage tax on spirits, and the tax on small cigarettes are probably still less desirable (but the spirits tax might well be increased if the administrators of the tax think that this could be done without losing revenue through bootlegging, restriction of demand of spirits generally, etc.).

ESTATE AND GIFT TAXES. -- First choice seems to me to be a series of measures to get more money from the death and gift taxes. The chief problem here is the lag between enactment of the change and the inflow of receipts. If the additional money is to be raised, not to relieve immediate pressure on the Government bond market, but rather to give assurance of progress toward a balanced budget, increases in these taxes are specially suitable. They strengthen confidence in the fiscal policies of the Government without at once exercising a deflationary effect. But if an immediate deflationary effect is needed (as when the market for a government's bonds is already showing signs of grave weakness), these increases do not of course act fast enough.

I think we can assume that there will be no need for an immediate deflationary effect in the near-term future, and I therefore think the lag in increased receipts would not be a marked disadvantage, and might even be an advantage. However, if the matter is left to drift for a year or so, the answer might have to be different.

A. The steps that in my opinion could be taken at once to increase the revenues from the death taxes and gift taxes are as follows (in addition to changes already recommended by the sub- committee):

1. Increase the rates of the estate tax and the gift tax, and lower the specific exemption from $40,000 to $20,000. The increase in death tax rates could be somewhat like the recommended in the Haas report of last August, Vol. V. p. 24(a). This schedule was approved in the Blough-Shoup report of last September (pp.M-16 to M-19).

Under the present schedule a net estate, BEFORE deducting the specific exemption of $40,000, pays no tax. Under the proposed rates and exemption it would pay $600, or 1-1/2 per cent. /1/ An estate of $100,000 now pays $4,200; it would pay $9,000. An estate of $500,000 now pays $80,400, or 16.1 per cent -- it would pay $164,900, or 33.0 per cent. /2/ An estate of $1,000,000 now pays $211,000; it would pay $443,500. An estate of $5,000,000 now pays $1,900,000, or 38%; it would pay $2,922,700, or 58%.

This increase in rates and lowering of exemption might produce an added revenue of somewhere around $400,000,000 in a normal year.

If serious attention is to be given to the probability of increasing the estate tax rates and/or lowering the specific exemption, I suggest that the statistical section be started at once on the making of estimates of yield (and not for a "normal year", but specifically for the fiscal years 1939, 1940, and 1941, assuming the legislation to become effective as of a given date -- say June 30, 1938 -- and assuming only moderate business improvement in those years).

2. Repeal the $40,000 insurance exemption. This might add $10,000,000 or $20,000,000 to the revenue, assuming the new high rates to be in force.

3. Eliminate the option to value property as at death or one year later. (I do not have any means at hand to estimate even roughly what this would mean in revenue increase, but for the sake of saving something, suppose it adds $20,000,000 a year).

4. Increase the rates of the gift tax to correspond with the increase in the estate tax, and lower the specific exemption to $20,000. This might yield an added $10,000,000 or so.

5. Include in the decedent's taxable estate all property over which he has power of appointment either general or limited (at present we tax only if the power is general and is exercised). This would probably yield more than $10,000,000. It appears that this provision would be held constitutional except possibly where the power was extremely limited.

6. Include in the decedent's taxable estate under Sec. 302(c) the proceeds of insurance on the decedent's life where he has retained no incidents of ownership in the insurance (he cannot change the beneficiary, cannot borrow on it, etc.). This might yield $10,000,000 a year easily. (At present the gift tax applies to the premiums when paid). It appears that this provision would be held constitutional, at least for policies taken out after 1918.

7. Remove the exemption now granted to non-resident aliens on transfers (estate tax and gift tax) of Federal Government bonds. Possibly $5,000,000 a year is optimistic.

Altogether these changes in the death and gift tax might add almost half a billion dollars to the Federal Revenue in the fiscal year 1941. Possibly the increase in 1940 would be about $0.2 or $0.3 billion.

B. In addition to the steps listed above, there are others that seem clearly desirable and that would likewise increase substantially the death tax and gift tax revenues. They are put in this second group, however, because they might be more difficult to enact speedily. They need to be worked on further to smooth out technical problems. Thus they might imperil the rest of the program too much to be considered for action this year.

1. Coordinate the gift tax and the estate tax by starting the progression of rates on the estate where they left off with the last gift. This would increase revenues substantially, but I have no way at present of computing even an approximate amount.

2. Restrict the exemption to charitable bequests, perhaps as suggested in the Haas Report (Vol. V. p. 20, second paragraph).

3. Reduce the relief now given to property previously taxed within five years, by graduating the relief as suggested in the Haas Report (Vol. V, p. 16, first paragraph).

4. Impose some kind of accessions tax on remainderman interests when they mature (this point will be developed by Harriss in his forthcoming memorandum on death and gift taxes).

5. Remove deduction (now allowed where state law allows it) for support of dependents. Possibly $2,000,000 or so involved.

INCOME TAXES. -- If revenue is needed more quickly than the estate and gift taxes will bring it, the personal income tax seems to me to be the best source. An increase here would probably have a smaller deflationary effect on purchases of consumers' goods than any other increase that would bring the money is an quickly, and I am still assuming that for the time being such deflationary effects are not needed. From this point of view, increases in taxes on the well- to-do and the wealthy -- say those with incomes above $10,000 or $15,000 - seems called for. But it is also desirable to avid the deflationary effects that come from decreasing willingness to take changes and invest. In a depression almost every new investment (whether in an old or new industry) involves a seemingly large element of risk, either for (1) the new investment or (2) the already existing investment that has to be pushed out further in the risk area when new senior capital is put in - for example, outstanding common stock, when bank loans are made or new bond issues are floated. A clue to proper action is given by this supposition: An invest for thinks of possible new income as being superimposed on the margin of his existing income, and hence subject to the marginal tax rate (the bracket rate), of the effective rate. Thus if a man who is already receiving $50,000 taxable income is asked to invest with the possibility of adding $6,000 to his income, he tends to think of the $6,000 as being subject to a 25 per cent Federal income tax rate (the bracket rate on $50,000 to $56,000). If his tax is raised by increasing only the rate in, say, the $20,000-$40,000 rage, his willingness to make this new investment is probably not decreased. Over the long run the increase might have a dampening effect, since over the long run he may plan his investment program more as a whole, but this memorandum is concerned primarily with short-run tendencies.

These speculations are too vague to show just where the rates might best be raised, but I think they indicate that somewhere in the middle or lower-middle brackets rather than in the upper-middle or higher brackets is the place. It might even be advisable to lower, at the same time, they very highest surtax rates. The net result could be an increase in total tax, even for the very wealthiest taxpayers. These taxpayers would, however, have their tax increased by a smaller percentage than taxpayers in the lower ranges.

In my memorandum to you on November 26, 1937, on p. 8, there is given a hypothetical "Schedule X" of surtax rates than more or less fulfills the conditions set forth above. It might raise from $100,000,000 to $200,000,000 added revenue in the fiscal year 1940 and somewhat less in fiscal 1939, if passed by June 30, 1938. If more revenue than that is needed, I suggest that Schedule X be strengthened in the $10,000-$20,000 range rather than in the $20,000- $50,000 range.

As indicated in recent conversations, it is desirable to allow a carry-forward of business net losses for six years or so, and also a long carry-forward of capital losses against capital gains. If this were done, an increase in the corporation normal tax rate might be desirable at the same time the surtax rates were raised as suggested above.

FOOTNOTES

/1/ In these computations it is assumed that the exemption is constant, not vanishing as recommended by the Haas report, and that the bracket schedule, not a totality schedule as recommended by the Hass Report, is used. These assumptions imply, not necessarily disagreement with these features of the Haas report, but a belief that if a program is to be put through, without much delay, to raise substantial amounts of revenue, it will not prove feasible to put through, at the same time, changes in such points of technique. /2/ From here on the tax due is before credit for state taxes. Roughly it may therefore be regarded as the total of Federal and state death taxes.