The opinion
STEWART, Circuit Judge
(dissenting).
By use of the net worth method the Tax Court determined income tax deficiencies against the petitioner of about $800 for each of the years 1949 and 1950, and of about $1,000 for 1951. This result was largely accomplished by refusing to credit the petitioner with any cash on hand at the beginning of the taxable period.
*783 The petitioner, his wife and his mother all testified that he did have cash savings on December 31, 1948. This testimony was corroborated by two disinterested witnesses. No one testified to the contrary. There was nothing inherently incredible about this evidence. The petitioner was continuously employed at relatively high wages for fifteen years before the period in question; during much of that time he was unmarried and living with his parents, so that his expenses were low. He testified that his reason for saving was to go into business eventually for himself, and that is what in fact he did in December of 1951. In my opinion the Tax Court’s finding that petitioner had no cash savings at the beginning of 1949 was clearly erroneous.
My disagreement with the majority, however, amounts to more than simply a different view of the particular facts of this case. It stems rather from a belief that use of the net worth method in a case like this is not justified.
The petitioner was a full time salaried employee during the taxable period, except for the last month of the final year. He had some additional income during the period, consisting of rental from the upstairs apartment in the duplex where he lived, a $900 administrator’s fee and $328 of interest. He fully and accurately reported his salary and the rental and administrator’s fee. He fully and accurately reported partnership income for the single month of 1951 that he was in business. The Commissioner did not claim and the Tax Court did not find otherwise. 1
In the proceedings below the Commissioner made an effort to show possible additional sources of income. The petitioner adduced positive evidence to refute each source suggested. The petitioner’s evidence apparently satisfied the Tax Court, since it made no findings as to any likely source of additional income and concluded in its opinion that the Commissioner had failed to show such a source. Para. 55,121 P-H Memo T.C.
The Supreme Court has said, “Increases in net worth, standing alone, cannot be assumed to be attributable to currently taxable income. But proof of a likely source, from which the jury could reasonably find that the net worth increases sprang, is sufficient.” Holland v. United States, 1954, 348 U.S. 121, 137, 138 , 75 S.Ct. 127,136 . It is the view in the First Circuit that this language makes proof of a likely source, “an indispensable element of the net worth method.” Thomas v. Commissioner, 1 Cir., 1956, 232 F.2d 520, 526 . See also United States v. Donovan, D.C.E.D.Va., 1956, 142 F.Supp. 703 . We have not thought that these words in the Holland opinion stated such an absolute injunction. See Thomas v. Commissioner, 6 Cir., 1955, 223 F.2d 83, 86 ; Kashat v. Commissioner, 6 Cir., 1956, 229 F.2d 282, 285 .
But the Supreme Court’s language does clearly require some corroboration to support a finding that the alleged net worth increases were attributable to currently taxable income, and none was found by the Tax Court here. The net worth method of reconstructing income has typically been resorted to in cases involving entrepreneurial income of a substantial amount. In such cases the necessary corroboration can often be supplied by showing that the taxpayer’s disclosed business was capable of producing additional income sufficient to account for the alleged net worth increases. But in the case of a low bracket taxpayer working full time on a salary, corroboration must be found elsewhere. United States v. Ford, 2 Cir., 1956, 237 F.2d 57 ; United States v. Adonis, 3 Cir., 1955, 221 F.2d 717 .
I am convinced that the Tax Court in the present case has fallen into the very pitfalls so carefully pointed out by the Supreme Court in the Holland and companion opinions. Friedberg v. United States, 1954, 348 U.S. 142 , 75 S.Ct. 138 , 99 L.Ed. 188 ; Smith v. United States, 1954, 348 U.S. 147 , 75 S.Ct. 194 , 99 L. Ed. 192 ; United States v. Calderon, *784 1954, 348 U.S. 160 , 75 S.Ct. 186 , 99 L.Ed. 202 . To be sure those were criminal cases, and much of what the Supreme Court had to say in deciding them necessarily relates alone to criminal prosecutions. But the same limitations inhere in the net worth method wherever it is sought to be applied. It remains “a method that is itself only an approximation.” 348 U.S. at page 129, 75 S.Ct. at page 132. And whatever the forum, “its use in the ordinary income-bracket cases greatly increases the chances for error.” 348 U.S. at page 127, 75 S.Ct. at page 131.
The net worth method is a particularly formidable weapon against a taxpayer disarmed by the presumption of correctness attaching to the Commissioner’s determination. I have serious doubts about the use at all of a method so inexact to arrive at amounts so relatively small as are involved here. And I am convinced that in permitting the naked use of net worth figures without the restraining necessity of a corroborative finding of the kind suggested, there lies the possibility of serious injustice to honest taxpayers.
. From the outset the petitioner has conceded that he failed to report the interest income of $328.