POSNER, Circuit Judge.

Section 162(a) of the Internal Revenue Code of 1954, 26 U.S.C. § 162 (a), allows the deduction of "all the ordinary and necessary expenses paid or incurred during the taxable year in carrying on any trade or business . . .," but this is qualified (see 26 U.S.C. § 161) by section 263(a) of the Code, which forbids the immediate deduction of "capital expenditures" even if they are ordinary and necessary business expenses. We must decide in this case whether certain expenditures made by Encyclopaedia Britannica, Inc. to acquire a manuscript were capital expenditures.

Encyclopaedia Britannica decided to publish a book to be called The Dictionary of Natural Sciences. Ordinarily it would have prepared the book in-house, but being temporarily short-handed it hired David-Stewart Publishing Company "to do all necessary research work and to prepare, edit and arrange the manuscript and all illustrative and other material for" the book. Under the contract David-Stewart agreed "to work closely with" Encyclopaedia Britannica's editorial board "so that the content and arrangement of the Work (and any revisions thereof) will conform to the idea and desires of [Encyclopaedia Britannica] and be acceptable to it"; but it was contemplated that David-Stewart would turn over a complete manuscript that Encyclopaedia Britannica would copyright, publish, and sell, and in exchange would receive advances against the royalties that Encyclopaedia Britannica expected to earn from the book.

Encyclopaedia Britannica treated these advances as ordinary and necessary business expenses deductible in the years when they were paid, though it had not yet obtained any royalties. The Internal Revenue Service disallowed the deductions and assessed deficiencies. Encyclopaedia Britannica petitioned the Tax Court for a redetermination of its tax liability, and prevailed. The Tax Court held that the expenditures were for "services" rather than for the acquisition of an asset and concluded that therefore they were deductible immediately rather than being, as the Service had ruled, capital expenditures. "The agreement provided for substantial editorial supervision by [Encyclopaedia Britannica]. Indeed, David-Stewart's work product was to be the embodiment of [Encyclopaedia Britannica's] ideas and desires. David-Stewart was just the vehicle selected by [Encyclopaedia Britannica] to assist . . . with the editorial phase of the Work." Encyclopaedia Britannica was "the owner of the Work at all stages of completion" and "the dominating force associated with the Work." The Service petitions for review of the Tax Court's decision pursuant to 26 U.S.C. § 7482.

As an original matter we would have no doubt that the payments to David-Stewart were capital expenditures regardless of who was the "dominating force" in the creation of The Dictionary of Natural Sciences. The work was intended to yield Encyclopaedia Britannica income over a period of years. The object of sections 162 and 263 of the Code, read together, is to match up expenditures with the income they generate. Where the income is generated over a period of years the expenditures should be classified as capital, contrary to what the Tax Court did here. From the publisher's standpoint a book is just another rental property; and just as the expenditures in putting a building into shape to be rented must be capitalized, so, logically at least, must the expenditures used to create a book. It would make no difference under this view whether Encyclopaedia Britannica hired David-Stewart as a mere consultant to its editorial board, which is the Tax Court's conception of what happened, or bought outright from David-Stewart the right to a book that David-Stewart had already published. If you hire a carpenter to build a tree house that you plan to rent out, his wage is a capital expenditure to you. See Commissioner of Internal Revenue v. Idaho Power Co., 418 U.S. 1, 13, 94 S.Ct. 2757, 2764, 41 L.Ed.2d 535 (1974).

We are not impressed by Encyclopaedia Britannica's efforts to conjure up practical difficulties in matching expenditures on a book to the income from it. What, it asks, would have been the result if it had scrapped a portion of the manuscript it received from David-Stewart? Would that be treated as the partial destruction of a capital asset, entitling it to an immediate deduction? We think not. The proper analogy is to loss or breakage in the construction of our hypothetical tree house. The effect would be to increase the costs of construction, which are deductible over the useful life of the asset. If the scrapped portion of the manuscript was replaced, the analogy would be perfect. If it was not replaced, the tax consequence would be indirect: an increase or decrease in the publisher's taxable income from the published book.

What does give us pause, however, is a series of decisions in which authors of books have been allowed to treat their expenses as ordinary and necessary business expenses that are deductible immediately even though they were incurred in the creation of long-lived assets — the books the authors were writing. The leading case is Faura v. Commissioner, 73 T.C. 849 (1980); it was discussed with approval just recently by a panel of the Tenth Circuit in Snyder v. United States, 674 F.2d 1359, 1365 (10th Cir. 1982), and was relied on heavily by the Tax Court in the present case.

We can think of a practical reason for allowing authors to deduct their expenses immediately, one applicable as well to publishers though not in the circumstances of the present case. If you are in the business of producing a series of assets that will yield income over a period of years — which is the situation of most authors and all publishers — identifying particular expenditures with particular books, a necessary step for proper capitalization because the useful lives of the books will not be the same, may be very difficult, since the expenditures of an author or publisher (more clearly the latter) tend to be joint among several books. Moreover, allocating these expenditures among the different books is not always necessary to produce the temporal matching of income and expenditures that the Code desiderata, because the taxable income of the author or publisher who is in a steady state (that is, whose output is neither increasing nor decreasing) will be at least approximately the same whether his costs are expensed or capitalized. Not the same on any given book — on each book expenses and receipts will be systematically mismatched — but the same on average. Under these conditions the benefits of capitalization are unlikely to exceed the accounting and other administrative costs entailed in capitalization.

Yet we hesitate to endorse the Faura line of cases: not only because of the evident tension between them and Idaho Power, supra, where the Supreme Court said that expenses, whatever their character, must be capitalized if they are incurred in creating a capital asset, but also because Faura, and cases following it such as Snyder, fail in our view to articulate a persuasive rationale for their result. Faura relied on cases holding that the normal expenses of authors and other artists are deductible business expenses rather than nondeductible personal expenses, and on congressional evidence of dissatisfaction with the Internal Revenue Service's insistence that such expenses be capitalized. See 73 T.C. at 852-61. But most of the cases in question (including all those at the court of appeals level), such as Doggett v. Burnett, 65 F.2d 191 (D.C. Cir. 1933), are inapposite, because they consider only whether the author's expenditures are deductible at all — not whether, if they are deductible, they must first be capitalized.

The legislative materials relied on in Faura relate to section 2119 of the Tax Reform Act of 1976, Pub.L. 94-455, 90 Stat. 1912. This provision allows publishers, until the Internal Revenue Service promulgates regulations, having prospective effect only, governing the deductibility of publishers' prepublication expenditures, to continue treating their prepublication expenditures as they have done in the past, notwithstanding Revenue Ruling 73-395, where the Service had ruled that such expenditures must be capitalized before being deducted. The enactment of section 2119 shows that Congress was unhappy with forcing publishers to capitalize their prepublication expenditures if they had been consistently deducting those expenditures immediately. But Congress stopped far short of requiring the Service as a general matter to allow such expenditures to be so treated. The House Report makes this clear: "Your committee believes that in view of the uncertainty with respect to the treatment of prepublication expenditures, the Internal Revenue Service should review this treatment and issue regulations to establish a uniform treatment of such expenditures for the entire publishing industry. This would allow interested taxpayers to have an opportunity to advise the Service about the practices and problems within the industry with respect to this matter. Since your committee is concerned about the retroactive application of Rev.Rul. 73-395 which would affect practices consistently followed by many taxpayers for years, your committee believes that any new rules applicable under regulations promulgated after the enactment of this bill should only have prospective application." H.Rep. No. 658, 94th Cong., 2d Sess. 337-38 (1976), U.S. Code Cong. Admin.News 1976, pp. 2897, 3234.

It is true that the Senate Report describes as "inequitable" the Service's refusal to allow professional authors to deduct their prepublication expenses immediately. S.Rep. No. 938, 94th Cong., 2d Sess. 404 (1976), U.S. Code Cong. Admin.News 1976, p. 2897. But the Senate's bill covered authors as well as publishers, while the House bill was limited to publishers, and in conference the House version was adopted. See S.Rep. No. 938 — Part II, 94th Cong., 2d Sess. 502 (1976). Thus the Senate committee's concern with the inequity of forcing authors to capitalize their prepublication expenses, even if it could be viewed as support for the Faura line of cases, did not carry the day in Congress. Moreover, the Senate committee's concern was in part just with the disparate treatment of authors and publishers under the House version of the bill, see S.Rep. No. 938, 94th Cong., 2d Sess. 404-05 (1976), a matter irrelevant to the merits of Faura. And so far as section 2119 itself is concerned, it is inapplicable to this case because the Tax Court found, correctly in our view, that Encyclopaedia Britannica did not have a consistent practice of immediately deducting the class of expenditures that includes the advances to David-Stewart.

Yet despite all this we need not decide whether Faura is good law, and we are naturally reluctant to precipitate a conflict with the Tenth Circuit. The Tax Court interpreted Faura too broadly in this case. As we interpret Faura its principle comes into play only when the taxpayer is in the business of producing a series of assets that yield the taxpayer income over a period of years, so that a complex allocation would be necessary if the taxpayer had to capitalize all his expenses of producing them. This is not such a case. The expenditures at issue are unambiguously identified with The Dictionary of Natural Sciences. We need not consider the proper tax treatment of any other expenses that Encyclopaedia Britannica may have incurred on the project — editorial expenses, for example — as they are not involved in this case. Those expenses would be analogous to author Faura's office and travel expenses; they are the normal, recurrent expenses of operating a business that happens to produce capital assets. This case is like Idaho Power, supra. The expenditure there was on transportation equipment used in constructing capital facilities that Idaho Power employed in its business of producing and distributing electricity, and was thus unambiguously identified with specific capital assets, just as Encyclopaedia Britannica's payment to David-Stewart for the manuscript of The Dictionary of Natural Sciences was unambiguously identified with a specific capital assets.

It is also relevant that the commissioning of the manuscript from David-Stewart was somewhat out of the ordinary for Encyclopaedia Britannica. Now the word "ordinary" in section 162 of the Internal Revenue Code has two different uses: to prevent the deduction of certain expenses that are not normally incurred in the type of business in which the taxpayer is engaged ("ordinary" in this sense blends imperceptibly into "necessary"), Deputy v. du Pont, 308 U.S. 488, 495, 60 S.Ct. 363, 367, 84 L.Ed. 416 (1940); Trustees of Graceland Cemetery Improvement Fund v. United States, 515 F.2d 763, 778 (Ct.Cl. 1975), and to clarify the distinction between expenses that are immediately deductible and expenses that must first be capitalized, Commissioner v. Tellier, 383 U.S. 687, 689-90, 86 S.Ct. 1118, 1119-20, 16 L.Ed.2d 185 (1966). (A merging of these two distinct senses of the word is a possible explanation for the result in Faura.) Most of the "ordinary," in the sense of recurring, expenses of a business are noncapital in nature and most of its capital expenditures are extraordinary in the sense of nonrecurring. Here, as arguably in Idaho Power as well — for Idaho Power's business was the production and distribution of electricity, rather than the construction of buildings — the taxpayer stepped out of its normal method of doing business. In this particular project Encyclopaedia Britannica was operating like a conventional publisher, which obtains a complete manuscript from an author or in this case a compiler. The conventional publisher may make a considerable contribution to the work both at the idea stage and at the editorial stage but the deal is for a manuscript, not for services in assisting the publisher to prepare the manuscript itself. Yet we need not consider whether a conventional publisher should be permitted to deduct royalty advances made to its authors as current operating expenses, merely because those advances are for it recurring business expenses because its business is producing capital assets. Idaho Power, though factually distinguishable, implies one answer to this question (no), Faura another (yes), and section 2119 still another (sometimes). But the principle of Faura, whatever its soundness, comes into play only when the expenditure sought to be immediately deducted is a normal and recurrent expense of the business, as it was not here. And section 2119 is inapplicable.

There is another point to be noted about the distinction between recurring and nonrecurring expenses and its bearing on the issue in this case. If one really takes seriously the concept of a capital expenditure as anything that yields income, actual or imputed, beyond the period (conventionally one year, United States v. Wehrli, 400 F.2d 686, 689 (10th Cir. 1968)) in which the expenditure is made, the result will be to force the capitalization of virtually every business expense. It is a result courts naturally shy away from. See, e.g., Briarcliff Candy Corp. v. Commissioner of Internal Revenue, 475 F.2d 775, 785 (2d Cir. 1973). It would require capitalizing every salesman's salary, since his selling activities create goodwill for the company and goodwill is an asset yielding income beyond the year in which the salary expense is incurred. The administrative costs of conceptual rigor are too great. The distinction between recurring and nonrecurring business expenses provides a very crude but perhaps serviceable demarcation between those capital expenditures that can feasibly be capitalized and those that cannot be. Whether the distinction breaks down where, as in the case of the conventional publisher, the firm's entire business is the production of capital assets, so that it is literally true that all of its business expenses are capital in nature, is happily not a question we have to decide here, for it is clear that Encyclopaedia Britannica's payments to David-Stewart were of a nonnormal, nonrecurrent nature.

In light of all that we have said, the contention that really what David-Stewart did here was to render consulting services to Encyclopaedia Britannica no different from the services of a consultant whom Encyclopaedia Britannica might have hired on one of its in-house projects, which if true would make the payments more "ordinary" in the Faura sense, is of doubtful relevance. But in any event, if that is what the Tax Court meant when it said that David-Stewart was not the "dominating force," its finding was, we think, clearly erroneous. We deprecate decision by metaphor. If the concept of a dominating force has any relevance to tax law, which we doubt, an attempt should have been made to operationalize it, as by computing the ratio of Encyclopaedia Britannica's in-house expenditures on The Dictionary of Natural Sciences to its payments to David-Stewart. If the ratio was greater than one, then Encyclopaedia Britannica could fairly be regarded as the dominant force in the enterprise. Although this computation was never made, we have no doubt that Encyclopaedia Britannica was dominant in the sense that, as the buyer, it was calling the tune; and it was buying a custom-made product, built to its specifications. But what it was buying was indeed a product, a completed manuscript. This was a turnkey project, remote from what is ordinarily understood by editorial consultation. While maybe some creators or buyers of capital goods — some authors and publishers — may deduct as current expenses what realistically are capital expenditures, they may not do so, at least outside the scope of section 2119, when the expense is tied to producing or acquiring a specific capital asset.

Encyclopaedia Britannica urges, as an alternative ground for sustaining the Tax Court's decision, that the payments to David-Stewart were immediately deductible as research and experimental expenditures under 26 U.S.C. § 174(a). This ground was not considered by the Tax Court, and it would be premature for us to consider it without the benefit of that court's views. The Tax Court can on remand consider it and any other unresolved issues.

REVERSED AND REMANDED.