(dissenting).
The first section of the legislation which provides for “Refunds of Amounts Collected under the Agricultural Adjustment Act” repeals the pertinent parts of that Act. The second section, in effect, declares the policy of the new legislation. It provides that no refund shall be made of any amount paid as a tax “unless the claimant establishes to the satisfaction of * * * the Board of Review * * * That he bore the burden of such amount and has not been relieved thereof nor reimbursed therefor nor shifted such burden, directly or indirectly, (1) through inclusion of such amount by the claimant * * * in the price of any article * * *. (2) through reduction of the price paid for any such commodity; or (3) in any manner whatsoever * * * ” 1
A later section, 907(a), provides that if the claimant’s “average margin per unit of the commodity processed” was less during the tax period than during certain preceding and following periods, that fact is “prima facie evidence” that he bore the burden of the tax to the extent of the difference. Sections 907(b) (1) and 907(b) (2), in effect, define “margin” as the extent to which the cost of the commodity before processing, plus tax if any, is less than the gross sales value of the processed commodity. Section 907(e) provides that the “presumption” established by § 907(a) may be rebutted by “proof of the actual extent to which the claimant shifted to others the burden of the processing tax. Such proof may include, but shall not be limited to * * * (2) Proof that the claimant modified existing contracts of sale, or adopted a new form of contract of sale, to reflect the initiation, termination, or change in amount of the processing tax, or at any such time changed the sale price of the article * * * by substantially the amount *718of the tax or change therein, or * * * billed the tax as a separate item to any vendee, or indicated by any writing that the sale price included the amount of the tax * * * ”2
Claimant’s average margin per unit was less during the tax period than during the base period. This difference in average margins, multiplied by the number of units processed during the tax period, amounted to $230,511.31. The Board, “upon considering all of the rebuttal evidence,” ordered a refund of this amount. It was “unable to find any ‘proof of the actual extent to which claimant shifted to others the burden of the processing tax’ (Sec. 907(e) sufficient to warrant a change in the amount of refund found due according to the statutory presumption.” The Board’s decision was expressly “based upon the presumption which arises under section 907(a).”
About 85 per cent of claimant’s sugar was sold in the United States. On June 8, 1934, when the processing tax and the equivalent floor-stock tax3 of 53% cents a hundred pounds went into effect in the United States, claimant’s American selling agent immediately raised the price of claimant’s sugar by 55 cents. It raised it uniformly, at all points; San Francisco, Los Angeles, Portland, Seattle, and Tacoma. This increase was proved without dispute; but the Board’s extensive findings do not refer to it, or to the related fact the market price on the Pacific coast rose to the same extent on the same day. The fact that the tax did not go into effect in the Philippines4 until September 12 does not obviate the fact that upon the “initiation” of the tax in the United States claimant “changed the sale price of the article” in the United States “by substantially the amount of the tax.” With respect to the 85 per cent of claimant’s sugar which was sold in the United States, therefore, the “prima facie evidence” or “presumption” that, to the extent of the difference in claimant’s margins, it bore the burden of the tax, was rebutted in a way in which § 907(e) provided that it might be rebutted. With respect to this same 85 per cent of claimant’s sugar, claimant believed and repeatedly “indicated by * * * writing that the sale price included the amount of the tax.”5 The writings included contracts with some of claimant’s customers. The presumption was therefore rebutted in a second way, and also in a third way, in which § 907(e) provided that it might be rebutted. The remaining 15 per cent of claimant’s sugar was sold in the Philippines. On this sugar, claimant regularly “billed the tax as a separate item.” With respect to this sugar likewise, therefore, the presumption was rebutted in a way in which § 907(e) provided that it might be rebutted. And even independently of § 907(e), the presumption was “out of the *719case”6 as soon as proof was introduced which would support a finding that claimant shifted the entire burden of the tax. The evidence which I have just summarized would support, if not require, such a finding. It was confirmed by additional evidence which I summarize below. Both because of the provisions of § 907(e) and independently of § 907(e), therefore, in my opinion the evidence eliminated the presumption, as such, from the case. The difference in margins was not eliminated from the case, but it was only evidence, entitled to no artificial weight.7 It follows that the Board’s action in basing its decision upon the presumption was “not in accordance with law.” 8 The decision should have been based upon the evidence.9
The opinion of the court, like that of the Board, rests on the premise that since the “prima facie evidence” or “presumption” was favorable to claimant, claimant was not required to prove that it bore the burden of the tax but the Commissioner was required to prove the contrary; in other words, when claimant established its prima facie case the burden of proof shifted from the claimant to the Commissioner. This premise seems to me erroneous. I cannot reconcile it either with the settled meanings of the terms “prima facie evidence” and “presumption” or with the basic enactment in the second section of the statute that “No refund shall be made or allowed * * * unless the claimant establishes * * * that he bore the burden of such ámount and has not been relieved thereof nor reimbursed therefor nor shifted such burden * * The rebuttal section, 907(e), on which the court appears to rely, does not purport to limit that enactment. It purports, on the contrary, to limit the effect of the presumption. It is a legislative recognition of the fact that the particular types of evidence which it enumerates are, and others may be, “appropriate to overcome”10 the presumption. I think the spirit and purpose as well as the letter of the Act are violated by holding, as the court does, that a claimant who (1) immediately increases all his prices by the amount of the tax when it takes effect, (2) contracts with some of his vendees that he will refund to them the amount of the tax if the government refunds it to him, and (3) constantly recognizes, in other contracts and other writings, that he bears no part of the burden of the tax, should still be presumed to have borne a part of that burden.
The decision of the court rests upon the further premise that the Commissioner has not proved that claimant shifted the burden of the tax. This also seems to me erroneous. The other evidence in the case not only deprives the prima facie evidence of its artificial force as law, but also of its natural force as fact. It may be true that the shifting of a tax can seldom or never be proved to an absolute certainty and is not so proved in this case. But I think the fact that claimant shifted the entire burden of the tax is proved to so high a degree of probability that there is no rational basis in the evidence for any other conclusion.11 In my opinion, therefore, the Board’s decision should be reversed and the case need not be remanded for a decision upon the evidence.
The Act undertakes to make good to the taxpayer the burden of his tax payments. It does not undertake to compensate him for losses due to other causes. The fact that claimant’s margins declined and its business became unprofitable during the tax period is the only basis in fact for an inference that claimant failed to shift the burden of the tax. It is not a sufficient basis.12 There is no evidence whatever that the tax caused the margins to decline or the business to become unprofitable. There is strong, and I think conclusive, evidence to the contrary.
Claimant bought its raw sugar at Philippine market prices. Almost all of its American sales of refined sugar were made on the Pacific Coast and based, with certain discounts,13 on San Francisco market *720prices. If, follows that if the tax (1) increased the market price of refined sugar in an amount equal to the tax and (2) did not increase the market price of raw sugar, then claimant shifted the entire burden which the tax imposed upon it. Accordingly, to say that claimant failed to shift the entire burden of the tax is to say either (1) that the price at which refined sugar could be sold in the market did not increase, by reason of the tax, to the full extent of the tax, or else (2) that the price of raw sugar did increase, by reason of the tax, to such an extent that the spread between the price of raw sugar and the price of refined sugar did not increase to the full extent of the tax. Neither of these propositions is true.
The San Francisco market price of refined sugar was controlled by two large refineries.14 Accordingly on June 8, 1934, when the tax of 53% cents took effect in the United States, the market price of refined sugar in San Francisco went up from $4, where it had stood since May 24, to $4.55. The double coincidence, in time and amount, between the tax and the price increase precludes any other reasonable inference than that the tax caused the increase. Refined sugar continued to sell at $4.55 until November 25.15
On the other hand the market price of raw sugar for manufacture and export did not rise in the Philippines, where claimant bought its raw sugar, either on June 8 when the tax took effect in, the United States or on September 12 when it took effect in the Philippines. According to the record it stood at $2.2956 throughout May and June, rose a fraction of a cent in July, fell off a little in August, and again fell off a little, to $2.2253, in early September. According to the record it remained practically constant during the half-month which followed September ll.16
*721Thus the spread between the Philippine market price of raw sugar for manufacture and export and the San Francisco market price of refined sugar increased at once, by the amount of the tax, when the tax took effect in the United States, and did not decline for several months. The same is true of claimant’s margins. I think this precludes any other reasonable inference than that the market spread, and claimant’s margins, increased by reason of the tax to the full extent of the tax. In other words, claimant shifted the entire burden of its tax payments.
Throughout the tax period, claimant frequently declared that it was shifting the entire tax burden.17 Accordingly it made contracts with some of its customers to pay over to them any tax refund which it might obtain in respect to the sugar which it sold them. Apparently claimant did not make many such contracts; but customers with 'whom it did make them paid the same prices as claimant’s other customers. One of the anomalies of this case is that if claimant obtains a refund on the theory that it did not pass on the entire tax burden it will, because of its contractual recognition of the fact that it did pass on the entire tax burden, pay over to some of its customers so much of the refund as relates to their purchases.
In the course of the tax period, the price of refined sugar fell as low as $4.20 and rose as high as $5. On the average it approximately retained the June 8 tax increment of 55 cents. The last pre-tax price, $4 on June 7, 1934, was 40 or 50 cents lower than the price had been a year previously. During the tax period, the price (less tax) did not entirely regain what the price had lost before the tax went into effect. Accordingly the average price, less tax, during the tax period as a whole was lower than the average price during the preceding year or two years. The decline in claimant’s margins was largely due to this fact. To attribute this fact to the tax would be to assert that if there had been no tax, the range of prices from June 8, 1934, to January, 1936 (the tax period) would have been the same that it was during the previous years. There is no basis for any such assumption. It cannot rationally be assumed, for example, that although the actual price ($4.55) less tax (53.5 cents) between June 8 and November 25, 1934, was $4,015, if there had been no tax the price would have ranged between $4.40 and $4.75 as it had done in the corresponding part of 1933; for the price on June 7, 1934, was only $4, whereas on June 7, 1933, it had been $4.40 or $4.50.
In the tax period as a whole, the price of raw sugar was somewhat higher, on the average, than it had been during the previous two years. But neither the claimant nor the Board nor this court suggests that the tax was or could have been responsible for any increase in the price of raw sugar. All the evidence is the other way. Throughout the parts of the tax period in which the increases took place, the amount of the tax remained stationary. In the aggregate, the increases exceeded the amount of the tax by more than 50 per cent.18 Thus there was no coincidence, either in the time or in amount, between the tax and the increases in the price of raw sugar. Sharp increases in that price were not confined to the tax period. At no time during the tax period did the price of raw sugar rise quite as high as it had risen a year before the tax period. In May, 1935, during the tax period, the price went up to $3.11; but in July, 1933, before the tax period, it had gone up to $3.17. The record does not show, and we need not inquire, what did cause any of the increases in the price of raw sugar. The causes may have been, for example, reduced acreage, poor crops, wage increases, or combinations among sugar growers. The historical evidence that the tax caused none of the increases is re-enforced by theoretical considerations. The tax was no part of the costs which producers of, or dealers in, raw sugar had to meet. Though a substantial tax on processing necessarily tends to cause, as it caused in this case, a substantial increase in the price of refined sugar, its normal tendency would seem to be to depress the price of raw sugar.19
With respect to the 15 per cent of claim*722ant’s product which it sold in the Philippines, the case is not materially different. (1) According to the record, the market price of, refined sugar in the Philippines between September 12, 1934, when the tax of 53% cents took effect there, and the end of September was $4.1593. This was exactly 53% cents higher than the price between September 1 and September 11. It seems apparent that the increase was due to the tax. Afterwards the price fell to $3.85, but rose again to $4.07, during the tax period. There is nothing to suggest any connection between the tax and these later fluctuations in the price of refined sugar. There was nothing unusual about them.20 While they were taking place, the tax remained stationary. (2) There is nothing to suggest that the tax caused any increase, at any time, in the price of raw sugar for manufacture for Philippine consumption. Between September 12 and the end of September that price, $2.38, was about 8 cents higher than it had been between September 1 and September ll.21 But larger price changes than that were frequent. From July to August, 1934, while the tax was not yet in effect in the Philippines and remained unchanged in the United States, the price of raw sugar had increased about 37 cents, from $1.91 to $2.28. In March, 1935, during the tax period, the price went as high as $2.54; but in February, 1934, before the tax period, it had gone as high as $2.92.
Though the court thinks the Commissioner’s proof insufficient, the court does not suggest and I do not know how there could ever be clearer proof that, despite a decline in a processor’s margins, he shifted the entire tax burden. In specifying means by which the presumption based on margins might be rebutted, “we do not think that Congress was attempting to require the impossible.” 22 Even if the evidence were insufficient to compel an affirmative finding that claimant shifted its entire tax burden,23 the result of the case should be the same. For the Act of Congress does not require the Commissioner, in order to defeat the claim to a refund, to prove that claimant shifted the burden of the tax. It requires the claimant, in order to establish its claim to a refund, to prove that it did not shift the burden of the tax. For that conclusion, the evidence furnishes no rational basis. To say this is to- contradict neither the Board nor this court. There has been no finding that the evidence alone is sufficient to support that conclusion. The principle of administrative finality,24 as I understand it, does not require or permit this court to sustain a decision which has no rational basis in the evidence and is based upon an error of law.
§ 902, 49 Stat. 1747, 7 U.S.C.A. § 644.
Italics supplied.
“Upon the sale or other disposition of any article processed * * * that on the date the tax first takes effect * * * is held for sale * * * there shall be levied, assessed, and collected a tax * * * equivalent to the amount of the processing tax which would be payable with respect to the commodity from which processed if the processing had occurred on such date.” 48 Stat. 31, 40, § 16(a), 7 U.S.O.A. § 616(a).
Claimant’s refining was done in the Philippines, and the taxes now in suit were paid after September 12, 1934.
For example: “The above price is based on Government Processing Tax at the rate of 53% cents per 100 pounds, and any change in this tax to be for the account of buyers.”
“Above price of $4.25 basis San Francisco includes present Compensating Tax.”
“ * * * if we recover any processing taxes levied on sugar sold to you, on which said tax is included in the price, any amount recovered will be returned to you * * * ”
“By agreeing to return any taxes we have paid and which may at some future time be returned to us, we are doing no more than could be expected * * *" “The processing tax is a manufacturers tax and the wholesale buyers do not pay it, but only pay our wholesale price which has been increased by the tax. * * * »
“ * * * all buyers should claim by the end of the year on P. T. Form 71 for taxes paid * * * provided always that they bore the burden of such taxes, and did not pass them on to their buyers.”
All the invoices which claimant sent to its American agent included the tax as a separate item, though the bills which the agent furnished to customers did not. Though the prices named in the invoices were not the prices at which the sugar was actually sold, they represented, as the Board found, what claimant “expected it might receive from the sale of sugar.” The invoices were not bills, but they were of course “writings.”
Del Vecchio v. Bowers, 296 U.S. 280, 286, 56 S.Ct. 190, 193, 80 L.Ed. 229.
Cf. American Law Institute, Model Code of Evidence, Rule 704.
49 Stat. 1750, § 906(g).
Com’r of Internal Revenue v. Bain Peanut Co. of Texas, 5 Cir., 134 F.2d 853, certiorari granted 320 U.S. 721, 64 S.Ct. 36, dismissed 64 S.Ct. 633.
Anniston Mfg. Co. v. Davis, 301 U.S. 337, 356, 57 S.Ct. 816, 825, 81 L.Ed. 1142.
Cf. Williams v. United States, 99 Ct. Cl. 203, 48 F.Supp. 647.
Cf. Vennell v. United States, D.C., 36 F.Supp. 646; Id., D.C., 38 F.Supp. 381, affirmed, 3 Cir., 122 F.2d 936; Insular Sugar Refining Corp. v. United States, 99 Ct.Cl. 345, 49 F.Supp. 319.
The Board found that claimant’s “actual proceeds derived from the sale of processing tax-paid sugar were substantially less than the theoretical gross sales *720value used in the computation of the margin for the tax period * * * ” But it does not appear that the tax had anything to do with this fact. The Board did not find, and the evidence does not show, that claimant came nearer to realizing the “theoretical gross sales value” during other periods. On the contrary the Board found, without limitation to the tax period, that “The basic price of refined sugar on the Pacific coast is established by California & Hawaiian Sugar Refining Corporation, Ltd., and Western Sugar Refinery, the two largest refineries on the west coast. Under instructions and authority received from International Suchar Corporation, Balfour [petitioner’s agent] sold petitioner’s sugar at varying discounts, ranging from 5 to 15 cents per cwt., under the prices established by these two companies. This difference in price was necessitated principally by the fact that petitioner had for sale only granulated sugar in bags of 100 pounds each or in 100-pound bags containing units of 5, 10, or 20 pounds each and did not have for sale assortments such as cubes, brown sugar, and powdered sugar. * * *
“The sales of sugar by petitioner to International Suchar Corporation and Refined Syrups, Inc., * * * were not arm’s-length transactions. Petitioner’s prices to those two purchasers were always substantially lower than the prices then obtaining for the petitioner’s sugar on the Pacific coast.” Board findings 18, 22.
The Board did not segregate the proceeds of claimant’s sales to its two affiliates, or inquire to what extent the decline in claimant’s margins may have been caused by larger arbitrary discounts to them during the tax period than during the base period.
Note 13 supra.
The opinion of the court points out that the price of refined sugar did not decline when the tax was withdrawn. But the effect of the imposition of the tax, and the effect or lack of effect of its withdrawal, are two distinct questions. Only the first concerns us. The record does not show what kept the price up when the tax was withdrawn. But it is common knowledge that prices may remain up for a time, after removal of the pressure which caused them to rise, simply because purchasers have become accustomed to the prices and the “traffic will bear” them. This tendency is of course most marked when control of the sellers’ side of the bargain is concentrated in a few hands. The Board found that “The basic price of refined sugar on the Pacific coast is established by California & Hawaiian Sugar Refining Corporation, Ltd., and Western Sugar Refinery, the two largest refineries on the west coast.”
Tho record does not show any Philippine price of raw sugar for manufacture and export between September 27, 1934, and January 26, 1935. Claimant did no processing for export, and the record perhaps implies that none was done, between September 11, 1934, and the end of January, 1935.
Note 5 supra. Cf. United States v. Jefferson Electric Mfg. Co., 291 U.S. 386, 405-406, 54 S.Ct. 443, 78 L.Ed. 859.
In May, 1935, the price of raw sugar was 82 cents higher than when the tax took effect in the United States and 89 cents higher than when it took effect in the Philippines. The tax was 53% cents throughout the tax period.
If the increase in the price of refined sugar had been due to an increased demand for refined sugar, that demand *722would of course have tended to increase the demand, and therefore the price, of raw sugar. But this fact has no bearing on the question of the effect, upon the price of raw sugar, of an increase in the price of refined sugar due to a tax (or other expense) which is imposed upon refining or refined sugar but not upon raw sugar. Although the demand for sugar may be relatively inelastic, a substantial increase in the price of refined sugar, such as this tax caused, must tend in some degree to reduce the effective demand for refined sugar. This in turn tends to reduce the demand for raw sugar, and therefore the price of raw sugar.
In 1933, when the tax was not in effect, the price fluctuated between $3.66 and $4.16.
The price was $1.77 in May and $1.-75 in June. The tax took effect in the United States on June 8.
Anniston Mfg. Co. v. Davis, 301 U. S. 337, 355, 57 S.Ct. 816, 81 L.Ed. 1143.
Cf. United States v. Jefferson Electric Mfg. Co., 291 U.S. 386, 405-406, 54 S.Ct. 443, 78 L.Ed. 859.
Dobson v. Com’r of Internal Revenue, 320 U.S. 489, 64 S.Ct. 239; Com’r of Internal Revenue v. Heininger, 320 U.S. 467, 64 S.Ct. 249.