Johnson, Drake & Piper, Inc., Appellant, v. United States of America, Appellee, 483 F.2d 682 (8th Cir. 1973)

Annotate this Case
US Court of Appeals for the Eighth Circuit - 483 F.2d 682 (8th Cir. 1973) Submitted May 17, 1973. Decided Aug. 23, 1973

Phillip H. Martin, Minneapolis, Minn., for appellant.

John A. Townsend, Atty., Tax Div., U. S. Dept. of Justice, Washington, D. C., for appellee.

Before VAN OOSTERHOUT, Senior Circuit Judge, HEANEY, Circuit Judge, and TALBOT SMITH,*  Senior District Judge.

TALBOT SMITH, Senior District Judge.

The case before us is a tax refund suit. It involves an effort by Johnson, Drake and Piper (hereafter the taxpayer) to increase its net operating loss for the year 1960, and, on the basis of such increase in loss, and the net operating loss carryback provisions of the Internal Revenue Code, to obtain a refund of income taxes paid for 1957.1  The losses themselves are not in dispute. The controversy concerns the taxpayer's belated attempt to amend its 1960 return.2 

During 1960 the contractor was engaged in the business of general contracting. It constructed buildings, highways, marine bridges, docks, and industrial foundations. Starting in about 1960 taxpayer commenced suffering serious reverses and it ceased doing construction work in about 1966.

The issue presented to us is whether the taxpayer erroneously reported its taxable income from certain long-term construction contracts in its federal income tax return for the said calendar year 1960. The alleged mistakes made by the taxpayer in so erroneously reporting such income fall into two different categories, as follows:

1) On Jobs 431, 429, and 426, preliminary book figures showing total net losses of about $553,000 were adjusted ("zeroed out") when the taxpayer originally computed its 1960 income.

2) On Job 413 the taxpayer reported a profit of some $89,000. It now seeks to report a loss of $35,000 in 1960 principally on the ground that Mr. Bacher,3  its chief accountant, had projected a "possible loss" of such sum.

It is the taxpayer's claim that under its method of accounting its full book losses were reported under such circumstances as were presented by the jobs in question. This, it asserts, was its consistent practice.4  It argues that the accounting entries made (except for the jobs in question) during the years 1953 to and including 1959 and 1960 establish that the book profit or loss reported by the taxpayer on all jobs during these years conforms with its accounting method, which includes the application of certain "rules"5  as to zeroing out, or the reporting, of book profits and losses, specifically, among others, that it was taxpayer's consistent practice to report the book loss on a job when that job was about 10% complete. It follows, argues the taxpayer, that since the reporting of the jobs in question did not conform to the said rules consistently theretofore followed, but rather diverged therefrom, through in large part, their auditors' use of the methods set forth in Accounting Research Bulletin No. 45,6  rather than taxpayer's former methods, such reporting involved errors which should now be corrected.

In support thereof the taxpayer, going back through the years, introduced evidence as to the accounting on job after job, which, assertedly, supported its position. The government, on the other hand, analyzing the same jobs, reached contrary conclusions, namely, that there was no inconsistency in the accounting, that book losses were zeroed out "when circumstances showed that no reasonable estimate of an overall profit or loss on the job could have been made" and that book losses stood unadjusted when, "because of serious uncertainties, leaving such book losses unadjusted resulted in the most reasonable reflection of operations to date on the percentage of completion method."

The 1960 jobs in question were highway construction contracts and were what is known as "unit-price contracts." Such contracts are to be distinguished from lump-sum contracts. A unit-bid contract is one wherein the contractor submits a price per unit (the cement work, for example, may be a unit) for each of the various categories involved. This type of contract is used where the final quantities of work cannot be determined with accuracy until final completion. Payments are made to the general contractor on a periodic basis for the work performed during the preceding period, normally a month. Since the contractor is paid only on completed construction units, termed "pay items," it is necessary that such units share not only the direct material inventory and labor involved but also the overhead costs and profit (margin) on the job. These latter elements may be apportioned ratably among all units, or, because of the inherent uncertainty as to precisely how many units in each category may ultimately be utilized, certain units may be "loaded." That is, units apt to be most certainly predictable, for instance, as to quantity, may be "loaded" with a disproportionate share of overhead and margin, thus resulting in what is termed an "unbalanced" bid. There are many reasons why a contractor would prepare and submit an unbalanced bid, e. g., "shrinkage."7  As Mr. Larson, senior accountant for Ernst & Ernst, testified, "The money is put on items that we know will not-should not shrink in quantity." (App. 479.) Although plaintiff's witness, Mr. Jesson, testified that taxpayer "tended to frown on that practice because it was too risky,"8  there was, we find, unbalancing to which we will later advert, in certain of the taxpayer's bids causing the periodic progress payments in such cases to bear a distorted relationship to the amount of work actually performed during the period.

As the job progressed it received constant scrutiny. The taxpayer used the calendar year, and the accrual method of accounting in reporting its income for tax purposes. Income for each job was reported according to the percentage of completion method.9  At year-end a "book profit or loss" was computed.10  This figure was then submitted to taxpayer's senior officials involved in accounting matters, particularly Mr. Gallagher (in the years after 1954) for analysis to determine whether such book profit or loss should be adjusted in the light of a re-estimation of the overall job profit,11  and of the impressions gained and conclusions reached from consultations and interviews with responsible officials close to the actual performance of the operation, as well as consideration, at times, of an independent "monitor" of the job, that is, an independent home office reestimation of the anticipated profit or loss, on the progress so far made on the job, whether significant or insignificant. The book profit or loss so determined was then subject to review by the taxpayer's independent auditors, Ernst and Ernst.12  The function of such auditors, as described by Mr. Kelly, the Ernst supervisor, was a review "to determine whether or not we could express an opinion on the fairness of the financial statement"13  We are cited to no exceptions taken as to the fairness of any statement or report before us. The reports and statements were those of the taxpayer. Mr. Larson testified that "all of the adjustments that we proposed are actually client's adjustments" and that "Somebody in the client's office would be aware of them and would either make them and we would approve them or would make them perhaps upon our suggestion, but it is a matter of the client's adjusting entries, not the auditors' adjusting entries,"14  although it was agreed, as well, that in event of disagreement income would be reported in accordance with the auditors' views.15 

Obviously, at this juncture, estimates16  have necessarily been involved, together with a high degree of professional judgment exercised with respect thereto. There must be a judgment made as to the "percentage of completion" profit based upon the most accurate estimate possible, at the time, as to the overall profit or loss to be expected from the job. Taxpayer concedes that "where a significant amount of work has not been completed" book profit or loss is not meaningful17  and it is zeroed out.18  Yet the line between significant and nonsignificant work is itself a matter of judgment. Similar exercise of judgment is involved in the projection of costs and in the estimation of ultimate profit or loss. We agree that the trial court's finding that "the zeroing out of losses was bottomed upon the good-faith judgment of those employees charged with that responsibility and its public accountant"19  and find it amply supported in the record.

Application of the judgmental principle prior to 196020  rather than any hard and fast "rule" of percentage of completion is illustrated by the Michigan Bell Telephone job21  in Grand Rapids in 1954, Job 335, where a book loss of over sixteen thousand dollars at approximately 24% completion, was zeroed out (employing an entry of "Adjustment to Deferred Costs-to Zero out"). This was done, testified Mr. Larson, of Ernst and Ernst, because the underbalanced or low-margin units had been produced first,22  a profit on the job was expected upon completion of the remaining pay units, and moreover, Bergstrom felt confidence in Mr. Victorson, "the man who was doing the job"23  and who was predicting an ultimate profit, not loss, on the job.24  It is clear from this record that judgmental factors as to the many imponderables in an on-going contract were exercised in and controlled the zeroing out process, rather than the application of "rules" urged upon us by the taxpayer.25 

We turn now to an analysis of the particular jobs in controversy in the first category, supra, Jobs 431, 429 and 426. Job 431, the Rochester, New York, Inner-City Loop, when about 38% compete,26  at the end of 1960, showed a book loss of $138,380.48. This loss was not reported but, rather, zeroed out. The Ernst and Ernst work papers27  indicated a possible ultimate profit due to unbalance in the bidding, whereby "most of the profit is in the concrete and paving, most of which is yet to be done."28  Mr. Gallagher in reviewing the situation, and after consultation with those on the job, had come to the conclusion, expressed in his memorandum Ex. C-9; that:

"Our year-end audit indicates a loss of approximately $138,000 on this job. We propose to zero this out as of 12/31/60 in expectation that the job will show a profit on completion in 1961."

It was thus his decision that "in spite of what the book figure says there's no loss to report at that time."29  Such decision reflected the judgmental factors hereinabove described and applied.

Job 429, the Highbridge Interchange job, was zeroed out at approximately 21% completion at the end of 1960 although the job reflected a book loss of $129,766.51. Bacher's "monitor" (which apparently Mr. Larson did not see) indicated a possible $16,500 loss on the job (Ex. D-10). Ex. D-9, however, indicated that the high profit items were yet to be produced. "Under these [described] conditions," it states, "it was not possible to 'loadup' good items on early items, which contributes to the large spread between earnings and costs to date." The decision reached is found on a sheet apparently attached to Ex. D-12 ("Inter Office" on taxpayer's stationery, from New York Office to Gallagher) stating "Balance out to -0-." It was the opinion of Mr. Larson that the job (in spite of a field re-estimate of $100,046 profit) was more likely to be a break-even job at best, and that, based upon his understanding of the percentage of completion method, it would have been improper to leave the loss under all of the circumstances, since "if a job was projected to break even and end up with zero profit or loss" no profit or loss for that job for that particular year should be reported.30 

Finally, of this job category, Job 426 was a road constructed in Honduras, contracted for by the United States Bureau of Public Roads. A book loss of $285,206.36 was effectively zeroed out to show a small profit, $5,285.90 at about 22% completion. A great deal of difficulty had been experienced on this job in ascertaining the true profit picture. Mr. Jesson had written to Mr. Fredrichs, the divisional chief on the job, that "On this job we have a very large excess of expense over income to date. We are carrying the difference as work-in-progress so that we will not show any profit or loss."31  Mr. Fredrichs urged that "the job would show a profit of at least $100,000 on completion," pointing out that "they show a loss to date because they have already incurred a large part of their mobilization costs" (emphasis in original).32  For the auditors, Mr. Larson felt, because of "the uncertainty of the information relative to the job and the lack of information as to what the ultimate profit would be"33  that an effective zeroing out was in order. Mr. Gallagher also was of the opinion that uncertainties clouded accurate estimations of the profit or loss on the job, and the book loss was accordingly zeroed out.34 

The situation with respect to Job 413, the single job in the second category, supra, does not involve a zeroing out. Rather the taxpayer now seeks to "correct" the book profit actually reported for the year 1960 to reflect a loss of $35,000. It seeks to do this principally on the ground that a monitor prepared by Bacher bore the entry "Possible loss, $35,000."35  The government, on the other hand, stresses Gallagher's notations on the monitor, stating "Increase in quantity income per 2-1-61 Recap, $241,728. Revised profit $206,728, compares favorably to 2-1-61 Recap."36  It is argued by the taxpayer that the revised income figure does not take into account the cost of generating such income, casting doubt on the validity of Gallagher's conclusions. But Gallagher, not his subordinate, Bacher, was the authoritative representative charged with making definitive determinations of projected losses. Mr. Gallagher's responsibilities in this area were made clear by Mr. Jesson's testimony: "Well, in this kind of a situation where we had a monitor showing a loss, and a difference of opinions shown by the field, it would have been Mr. Gallagher's responsibility to make an evaluation of both documents" and come to a conclusion.37  Gallagher, then, after reviewing the various records, made his determination, based upon the totality of the best information available, and in so doing rejected his subordinate's "possibility." This he did in good faith (no fraud is charged) and we share the opinion of the trial court that the judgmental factors at the root of decisions made in 1960 are difficult of accurate and persuasive resurrection some seven years later. Such was not accomplished to the extent either of convincing the trial court that a mistake was made or this court that there is substantial error in the trial court's determination.

It is obvious that the evidence thus reviewed presented sharp conflicts among the various witnesses. Differences were not confined to the lay witness. The taxpayer's experts, Professor Berryman and Mr. Ross, testified that it was the taxpayer's method, consistently followed, to report book losses in full when the job was more than 10% complete. The government sought to discredit the conclusion by arguing that the experts' review of the taxpayer's practices prior to 1960 was wholly inadequate to apprise them of how the taxpayer's accounting practices actually were implemented during the period critical to this case, and, moreover, that the "latest analysis estimated gain figures" employed by them were deficient as a basis for the conclusions reached. Similarly illustrative of conflicting testimony, the taxpayer criticizes Mr. Larson, Ernst's senior on the job who actually approved the taxpayer's accounting entries here in question, whereas the government supports his expertise and credibility. Mr. Jesson, on the other hand, who testified in support of the taxpayer's position, is criticized by the government as having "very little to do with implementing accounting practices" during the period under consideration.

It was the finding of the trial court, after consideration of a lengthy stipulation, exhibits, and testimony from both lay and expert witnesses, as well as voluminous records of past jobs and their accounting treatment,38  that taxpayer's method of accounting was a percentage of completion method, employing an adjustment of the accrual book profit or loss by the taxpayer's responsible accounting officials and auditors. Such adjustment was made when, in their good faith judgment, considering all of the pertinent circumstances, the said book profit or loss did not adequately or reasonably reflect the operations of the taxpayer. Accordingly, the adjusting entries to Jobs 431, 429, 426, as well as the non-adjustment of Job 413, were all consistent with the taxpayer's method of accounting, and, as well, reflected income for income tax purposes, and cannot now be changed.39  We agree.

It would serve no useful purpose and would only prolong this opinion for this court to consider in further detail the pro's and con's of the decisions reached in Jobs 335, 344, 345, 352, 347, 356, 358, 368, 401 and 406, among others. We have reviewed the arguments presented with respect to them but it is a legal truism that it is peculiarly the function of the trial court to resolve such conflicts as are here presented, and the court below, after weighing the competing versions of the taxpayer's accounting procedures, hearing lengthy testimony from many witnesses, reviewing a myriad of accounting worksheets, and engaging several witnesses in perceptive colloquies, made its findings of fact. They are presumptively correct. Wilson v. N. Y. Life Ins. Co., 250 F.2d 649 (8th Cir. 1958). We may, of course, set aside the judgment below if clearly erroneous, McAllister v. United States, 348 U.S. 19, 75 S. Ct. 6, 99 L. Ed. 20 (1954); Logan Charter Service, Inc. v. Cargill, 373 F.2d 54 (8th Cir. 1967). But upon the record before us, which we have reviewed in detail, we find no substantial error in the trial court's findings on the controverted issues, particularly that, "as a practical matter the profit or loss at the end of the year would be merely a question of judgment."40  Moreover, we are constrained to record, also, our agreement with the trial court's finding that "It seems clear from the evidence that the belated claim of losses during the year 1960 discovered by re-estimates was an afterthought of plaintiff's officials when subsequent to that year it became apparent that the four jobs in question, as well as other projects, were going to be heavy losers."41 

Under the view we have taken of the case it is unnecessary to discuss the additional issues argued by the parties.



Hon. Talbot Smith, United States Senior District Judge, Eastern District of Michigan, sitting by designation


The taxpayer contends that losses should have been reported on its tax returns as follows:

Percentage of Alleged Book Loss Not Job Number Completion Reported for 1960 431 About 38% $138,380.48 429 About 21% $129,766.51 426 About 22% $285,206.36 (net income reported $5,285.90) 413 About 59% $ 35,000.00 (net income reported $95,257.87- $6,036.23 reported in 1959, error, $130,257.87) ----------------- Total $588,353.35

"We must emphasize that we are not seeking to deny taxpayer the ultimate losses that were sustained on these jobs. We only say that taxpayer was not, under its method of accounting, required to take them in 1960 and may not belatedly do so now. Taxpayer could take losses on these jobs in years subsequent to 1960, but because of its serious business reversals they do not benefit taxpayer in those years." (Government's brief, pp. 56-57.)


The taxpayer's home office accounting department hierarchy comprised D. P. Jesson, Secretary-Treasurer and Financial Vice-President, E. R. Gallagher, Controller, E. F. Bacher, Chief Accountant, and Mrs. Carol Nordstrom, General Ledger Accountant. Of these individuals, testified Mr. Jesson, it was "Gene Gallagher who would have the best information from the job or from his own analysis as to whether these were other elements that should be taken into account" with respect to reporting profits or losses on jobs. App. 74


I.R.C Sec. 446(a) provides that taxable income shall be computed under the method of accounting on the basis of which the taxpayer regularly computes its income in keeping its books. It is further provided in the Treasury Regulations that a method of accounting which "reflects the consistent application of generally accepted accounting principles in a particular trade or business" will ordinarily be regarded as clearly reflecting income. Treas.Reg. Sec. 1.446-1(a) (2). It is specifically provided that all items of gross income and expense are to be treated consistently from year to year. As to the accrual method, employed in the case before us, it is further provided that the "method used by the taxpayer in determining when income is to be accounted for will be acceptable if it accords with generally recognized and accepted income tax accounting principles and is consistently used by the taxpayer from year to year." Treas.Reg. Sec. 1.446-1(c) (1) (ii)


"I. Book profit or book loss, as the case may be, was reported, without adjustment, except only where one of the specific exceptions listed below applied."

"II. Where a job was less than about 10% completed, neither a profit nor a loss was reported and the book profit or the book loss, as the case may be, was zeroed out."

"III. Where a job was more than about 10% complete but less than 25% complete a book profit was ordinarily zeroed out and no profit or loss was reported."

"IV. Where there was a book profit on a job more than 25% complete, taxpayer reported the book profit, not in excess of the approximate amount of the projected profit upon completion multiplied times the percentage of completion."

"V. Where the projected final profit or loss upon completion of a job was a loss which was greater than the book loss (or where there was a book profit), taxpayer reported the full projected loss in that year."


See Herwitz, Accounting for Long-Term Construction Contracts, 70 Harv. L. Rev. 449, 450 (1957): "Accounting Research Bulletins are promulgated by the Committee on Accounting Procedure of the American Institute of Accountants" and have been "long a source of thoughtful guidance to the accounting profession."


Shrinkage, as used in this context, is a concept to describe the following condition: The contractor, in preparing individual unit-bid contract prices, estimates the need to remove and replace with sand 2,000 cubic yards of soil. However, after completion actual removal and replacement totaled only 900 cubic yards. Thus shrinkage causes a reduction in profit realized and overhead distributed


App. 89


Herwitz, Accounting for Long-Term Construction Contracts, supra


From the gross income, comprised of the owner's payment estimates, plus unbilled estimates, less advance collections, was subtracted a cost figure, namely, the costs allocable to the year (including accrued costs) less inventory on hand at the end of the year, and less costs deferrable to a later accounting period. The result was (Stip. B-15) internally called the "book profit or loss" for the year. It was also referred to as a "preliminary working trial balance" since it might later be adjusted. At this point if we divide the total estimated contract price into the periodic payments made to that time the resulting figure will be the percentage of completion from the owner's standpoint, based upon the pay units completed


"During all years herein involved and for many years prior thereto including 1953, plaintiff, as part of its year-end accounting procedures, requested from each job superintendent or other job office personnel a current estimate of the projected profit or loss upon final completion of such job. Such projections would usually be prepared by the job superintendent, although in some cases they would be prepared by office engineers associated with the divisional office . . . [S]uch projections were usually prepared by job engineers or by office engineers using the estimate Summary form and were commonly called 'Job Recaps'." (Stip. B-16)


The Ernst employees who reviewed taxpayer's accounting entries and conferred with Mr. Gallagher, were George Kelly, supervisor from 1953 through 1960, Warren D. Bergstrom, senior on the job for the accounting years 1954 through 1956, and Willis E. Larson, the senior on the job for the accounting years 1958 through 1960


App. 206


App. 320-321


Stip. B-19


"Problems [with respect to the measurability of percentage of completion] can occur in four areas: (1) Objective and certain determination of total revenue, (2) uncertainty regarding collection, (3) measurement of percentage of completion, and (4) cost estimation." Spiller, Financial Accounting. Rev.Ed.1971, p. 184


Appellant's Brief, p. 15; Jesson, App. 47, places this point of completion as "about" ten per cent


The term zeroing out simply means making a wholly arbitrary entry without any support and underlying documents in order to fall in line with established practices. (Jesson, App. 75). Larson stated as a purpose "to adjust the recorded profit or loss that should be reported, based on the estimated profit on completion." (App. 462.) It may be made in deferred costs, unbilled estimates, or through entries similar in purpose


Finding No. 18, App. 534. Stipulation H-1


The taxpayer urges that except for the zeroing-out on the three controverted jobs in 1960, no jobs were "ever zeroed out by the taxpayer" except where less than about 10% of the work had been completed


Job 335, Stip. H-4


Ex. L-4: "Sixty per cent of the billings worked to date have been on items with relatively low profit, steel, plumbing, electrical. We appear to have a start-up cost which can be justifiably deferred. Part of the contract calls for tearing down existing structures bid at a low margin."


App. 451


Ex. L-4


See, also, Stip. H-14, where a book loss at seven per cent completion was taken ("Torrential rains and floods have caused loss to date," Ex. L-18), and Stip. H-8 where a book loss at 12 per cent completion was zeroed out. In addition, Mr. Gallagher testified that up to between ten and thirty per cent "it was impossible to tell because of mobilization costs and other types of expense, possibly early return on unbalanced bid as to whether there was any income in the job." (App. 482.) Mr. Kelly was of the opinion that rather than any arbitrary percentage, zeroing out depended upon whether or not the job had had "a full season of operation by the end of the year" (App. 219) and Mr. Bergstrom testified that "I think it is primarily judgment. I never heard anybody fix a percentage." (App. 428.)


"At the end of calendar year 1960, on the basis of estimates allowed compared to the original contract price, Job 431 was 37.6% complete . . . . At that date, on the basis of costs incurred to date (excluding work in progress) compared to estimated total costs to completion (per 1/1/61 Recap) Job 431 was 46.6% complete" . . . (Stip. C-7)


Ex. G-28 (first and fourth pages)


App. 475


App. 498


Mr. Larson's testimony is explanatory of the procedures employed in the making of an adjustment:-

"As indicated in the memo, that adjustment was made, and I might say it was made not necessarily by us as auditors but by the client in connection with closing their year-end books-this entry and entries similar to this would be reviewed and discussed with the client before it was made. But the entry was made because without making the entry, there would have been a loss recorded on the job of approximately $129,000. Under the percentage of completion method of accounting, it would not be proper to leave the adjustment unmade. The memo indicates that the final job profit as shown in the 2/1/61 estimate of $100,000 was not necessarily reliable and although the job was 28 percent complete because of uncertainty on the estimate, the job was adjusted to show neither profit nor loss for that year." (App. 317.)

It was also his testimony (App. 321 and 328) that the entry made corresponded with the method of accounting used by the taxpayer, and that it was consistent with past practices.


Ex. F-5


Ex. G-23


App. 384


The situation with respect to Jobs 420 and 421, also 1960 jobs, wherein book losses were permitted to stand without adjustment, is not inconsistent with that reached in the jobs hereinabove discussed, since as "a matter of reasonableness" and "realistic and reasonable judgment" it was decided to let the losses stand


Ex. E-11


App. 151. See, also, Gallagher's letter, Jan. 20, 1961 (Ex. E-8), expressing the view that the job was in the black and the trend was good


App. 145. In preparing the monitor Bacher did not have the field recap (Stip. E-17) available to him, as it was to Gallagher in making his determination


Spiller, Financial Accounting (Rev.Ed. 1971) p. 163:

"The accounting for construction contracts actually can be quite complicated."


See App. 424-435, Cf., W. F. Trimble & Sons Co., 1 T.C. 482 (1943)


App. 531, Finding No. 14