Coming Soon:

Now Available: Volumes I, II, III, and IV of the Collected Published and Unpublished Papers.



Total Pageviews

Saturday, February 19, 2011


One of the inevitable consequences of this move from calculation to planning is a corresponding loss of simple clarity of goal. When an entrepreneur is locked in cutthroat competition with other small producers, he has very little choice of feasible entrepreneurial goals. Short-term profitability is the condition of survival. Growth – for economies of scale, for technological innovation, and as a means of liberation from market forces – virtually is thrust upon him. At this stage, there is no room for what eventually comes to be celebrated as “industrial statesmanship”. A firm may flourish one season, and be driven to the wall the next. Once the firm has reached a certain level of size and control over its input and output prices, however, the managers of the firm [for at this stage, it is likely that the individual entrepreneur has been replaced by salaried managers] must choose among a variety of corporate goals: short-term profit maximization, longer-term profit-maximization, enhancement of the firm’s stock market performance [which is, of course, not necessarily identical with profit maximization], managerial stability, attractiveness to potential take–over financiers, resistance to a financial take-over, and so forth.

There is no calculus that dictates how managers are to decide among these goals. They are genuine alternatives, corresponding to different and incompatible managerial ambitions. In a fundamental sense, which I shall dwell on shortly, these choices are political rather than economic in character. What distinguishes them from what are ordinarily considered political alternatives is not their logical structure, but simply the identity of the constituencies to whom the decision-makers are accountable in the making of the choices.

The second means available to entrepreneurs for overcoming the tyranny of the market – incorporation into their enterprise of stages in the production process that were previously carried out by independent supplier firms – constitutes an even more important theoretical departure from the logic of capitalism and a corresponding movement toward a precursor of socialist planning. This time, let us use a simple hypothetical example.

Imagine a firm that has been producing cardboard cartons in a factory with 80,000 square feet of usable floor space. We may suppose that the firm buys the cardboard in large sheets from a cardboard producer, and then stamps, folds, and staples the cardboard into cartons. Cardboard cartons are the sole product of the firm, and abstracting from a number of issues to which we shall return, the accounting procedures used to keep track of costs and determine profit margins are quite elementary and straightforward. The firm’s accountants have made a few standard assumptions about the depreciation rates of the machinery, and we may simplify things by assuming that there are no tax considerations [accelerated depreciation, etc.] to complicate or warp the bookkeeping. Profit is then the difference between costs and revenues, and the profit rate is calculated on the value of invested capital, the profit margin on the difference between cost and sale price of each unit of cartons.

The market for cardboard sheets, we may suppose, is fluctuating and uncertain, and the carton manufacturing company’s manager believes that he can stabilize his costs and increase his profits by producing his own cardboard, rather than buying it on the market. Since there are 25,000 square feet of unused space on the factory floor, he allocates that space to cardboard production and puts a young, ambitions supervisor in charge of the new branch of the firm. The question arises: how shall the profitability of this cardboard manufacturing activity be computed?

There are two reasons why this question needs answering from the firm manager’s point of view. First, of course, he must decide whether it is better to produce the cardboard in-house or continue to buy it in the market. That question, we may suppose, is relatively easy to answer. Second, as the manager of a firm with two divisions, each headed by an ambitious divisional chief, he must evaluate the relative profitability of the two divisions in order to decide which division chief will be tapped for promotion when and if a new, higher-level job in the corporation comes open. In large, modern capitalist firms, this bureaucratic question of internal advancement is of much greater importance to everyone in management than the older question of profitability of the firm as a whole. Profitability becomes a key to managerial advancement, not merely to owner-satisfaction.

The new division manager, eager to make as good a record for herself as possible, argues as follows to the accountant, whose job it is to come up with profitability estimates at the monthly meeting of officers of the firm: “The depreciation on the factory building is exactly the same whether we run the cardboard-producing activity or not, as are the general overhead costs for utilities, telephone and so forth. Save for some small increase in operating overhead, the cardboard-production experiment imposes no new fixed costs on the firm at all. What is more, the bookkeeping, ordering, and shipping departments can absorb the relatively small flow of work generated by the new division without any increase in personnel or overtime costs. Hence, in figuring the input costs as against revenues for cardboard production, the only items that ought to be charged off against the new division are such direct costs as the new machinery required, the wages of the extra workers taken on to make the cardboard, and the cost of the raw materials.”

The manager of the old carton division, who knows a serious threat to his job when he sees one, flatly refuses to agree to this self-serving bit of accounting sleight of hand. What the cardboard-production manager has proposed will, if accepted, result in a dramatically higher rate of return to her division’s investment than can possibly be earned in his. How long will it be before the president of the firm gets it into his head to have them switch places with an appropriate readjustment of their salaries! He proposes that all of the fixed costs of the enterprise be allocated to the two divisions in proportion to the percentage of the floor space they occupy [a method of allocation that happens to favor him a bit more than the equally plausible proposal to allocate in proportion to the revenues generated by the two divisions].

The president of the firm, confronted with this disagreement between his two division managers, turns to the accountant, whom he asks for an objective, impartial, scientific ruling. And he now gets a distinct shock, for his accountant, who is as honest as she is competent, informs him that as a matter of fact [or, more precisely, as a matter of accounting theory] there is no objective, impartial scientific answer to the question: How shall the fixed costs of a joint-production enterprise be allocated among the several distinct productive activities of the enterprise?

This is a rather startling proposition, on which the entire argument of these remarks rests. I need therefore to spend some time explaining it clearly and indicating the grounds on which it rests. My guide here is a pair of monographs written by a Canadian Professor of Accounting, Arthur L. Thomas, and published by the American Accounting Association. The first monograph, “The Allocation Problem in Financial Accounting Theory,” appeared in 1969; the second, “The Allocation Problem, Part II,” was published five years later. Thomas seems to be something of an iconoclastic radical masquerading as a nerdy bean-counting number-cruncher. In the driest language imaginable, he argues ploddingly, painstakingly, but devastatingly that a central activity of financial accountants – the allocation of costs to the outputs of a firm – has no objective rational whatsoever, being based rather an arbitrary choice of one from among a number of alternative incompatible patterns of allocation. What this means, in plain language, is that an accountant, speaking professionally, cannot tell the management of a firm just what a unit of output costs and hence how much of the profit of the firm can be attributed to its sale.

Thomas himself draws no larger lesson from this extraordinary conclusion, but in what follows, I will try to show that it has the most profound implications for our understanding of the manner in which something very like socialism evolves in the womb of the capitalist firm. Incidentally, in his correspondence with me, Thomas seemed to evince a good deal of pleasure at the radical conclusions I drew from his work, but it goes without saying that when I venture beyond the confines of financial accounting theory I am entirely on my own.

The accountant is presented by the firm with the raw data concerning the costs of the firm’s activities – invoices for materials, bills for rent, electricity, and insurance, hourly wages and managerial salaries, inventories, and so forth. His job [if we imagine this to be Thomas himself as the accountant] is then to figure out how much of each of these factor costs to allocate or impute to each unit of salable output, and thereby to calculate the contribution to profits being made by each division of the firm. He is to do this as objectively and accurately as the data permit, not allowing his assessments to be influenced either by the hopes and ambitions of the company management or by the particular aspirations of one division of the firm rather than another. His conclusions can then be presented to the investing public as a neutral evaluation of the performance of the firm, and to the management as a sound basis on which to make corporate decisions.

There are three different features of the activities of large capitalist firms that defeat the accountant’s effort to find neutral, objective methods for allocating costs to individual outputs, or to the subdivisions of the firm responsible for their production. These are the phenomena of fixed capital, inventories, and joint production. Let us look at each of them briefly in turn.

Fixed capital is capital that lasts longer than one cycle of production, which is to say, longer than it takes to produce one unit of output. The examples are endless – a robotic riveting machine on an automobile assembly line that can be used in the assembly of three thousand vehicles before it must be replaced, a power loom that wears out after being used to make ten thousand yards of cloth, an office copying machine that must be replaced on average every twenty thousand copies, an office building good for fifty years. Some of these items will have a resale value when it is time to replace them; others will simply have to be scrapped. Each of these factors of production costs the firm something, and that cost must in some way be apportioned to the units of output to whose production it contributes. To put the same point in other words, the accountant must select a schedule of depreciation for the productive factor. [I am entirely abstracting from the enormous complications introduced by the depreciation schedules stipulated by the Internal Revenue Service, which of course have nothing whatever to do with the actual wear and tear on the item being depreciated. These words are being written at my desk in the tiny pied-a-terre that my wife and I bought in Paris several years ago, which we rent out to readers of the New York Review of Books. Our apartment is in a seventeenth century building whose value has been appreciating steadily for the past four hundred years or so but when it comes time for me to fill out IRS Schedule E – Rents and Royalties – I am permitted to adopt the fiction that in twenty two and a half years, its value will have disappeared.]

Without going into the details of Thomas’ monographs, let me indicate some of the incompatible alternatives available to the accountant. He may assume that the fixed capital yields up equal parts of its purchase price to each unit if output produced with its aid. This was in fact the assumption Marx made in Capital. Alternatively, he may assume that the machine loses equal shares of its value in equal time periods – one tenth during each year of a ten year life, for example. This is not at all equivalent to the first alternative, because the machine may be used in the production of differing numbers of units in different years, either because of speed-ups and slow-downs in the production process, or because the machine has a breaking-in period before it reaches maximum efficiency, followed by a period of optimum functioning, after which as it wears out it is less and less capable of spewing forth product at the same rate. Each of these assumptions will yield a different depreciation schedule, and hence a different quantum of cost imputed to units of the output.

Alternatively, the accountant may adopt an entirely different approach, and by appealing to an existing market for second hand capital goods, impute to each year’s output the difference between what the machine would have sold for in that market at the beginning of the year, and what it would have sold for at the end of the year. Or – and this will yield a totally different set of numbers – he may compare what the company would have to pay for a second hand machine at the beginning of the year with what it would have to pay at the end of the year for a second hand machine one year older. But we have not even begun to ring the changes on this apparently elementary accounting problem, for the accountant may impute to a year’s output what it would cost the company to rent a machine for one year, even though the company has already purchased the machine, which is thus, from its point of view, a sunk cost. Indeed, Thomas actually considers the suggestion [which has been treated respectfully in accounting circles, to my surprise] that the company write off the entire cost of the machine in the year it is purchased, and treat it in all the subsequent years of its use as a free good, like the air that is required for the burning of fuel. The accountant may even, in any given year, choose whichever of these methods of allocation yields the lowest number, on the theory that he is thereby telling the management what its most efficient choice would be at any moment.

This litany of options available to the accountant does not begin to exhaust the logical possibilities. The problem faced by the firm, as Thomas makes clear, is that accounting theory offers no way of deciding which of them is objectively right – which accurately reveals to the management of the firm the true structure of its costs. Indeed, although the data with which the accountant works are, or can be made to be, factually accurate, there seems simply to be no answer to the question, “Just how much does each unit of output cost the firm to produce?” Thus, returning for a moment to our cardboard and box factory, the accountant cannot resolve the dispute between the division managers in a neutral and unbiased manner befitting a chartered accountant. Some of the many options are likely to favor the head of the box division, others will make the head of the cardboard division look better.

Similar problems arise when the accountant turns to the task of deciding how to carry on the books of the firm the cost of the stocks of material in its inventory. A firm acquires a stock of some input that it requires for its productive activities – coal for its furnaces, thread for its power loom, gravel for its cement mixers, diesel fuel for its trucks. Stocks characteristically are composed of units acquired at different times and at different prices. For the most part, stocks consist of homogeneous units – one gallon of diesel fuel is much like another, one bag of cement indistinguishable from another. [The homogenization of inputs, and the corresponding outputs, is , as Marx points out in Capital, one of the signal features of capitalism, and distinguishes it from the craft production that it replaces. It is accompanied by a like homogenization of the labor performed by the workers employed by a capitalist firm.]

How shall our accountant proceed in allocating the cost of inputs drawn from stock and thrown into production [as Marx liked to put it]? He may assume that the first unit withdrawn is the first unit that was added to the inventory, and he will then impute to the output whatever price was paid for it. This is the rule of allocation known in accounting circles as FIFO – First In, First Out. He will still have to adjust this price for the opportunity cost incurred by tying up that amount of capital in inventory, which in turn will require some choice from among the many alternative ways of computing the relevant rate of interest.

Alternatively, our accountant may opt for LIFO rather than FIFO – which is to say, Last In, First Out – or some average of the two ways of computing costs, and so forth. Since the price of stocks typically varies over the life of an inventory, it can make a very considerable difference to the profit or loss imputed to a unit of output which of these accounting conventions is selected. If I may echo the late unlamented Richard Nixon, it would be easy, but it would be wrong, to suppose that choosing one accounting rule and sticking to it will sanitize the accounting process and obviate these problems. It only takes a little mathematical imagination and some patience to construct examples in which, as a consequence of the pattern of variation of factor prices, the choice of FIFO over LIFO, or any of the other alternatives, can over time systematically advantage one division of a corporation over another.

For example, if a factor input has been falling steadily in price for some years, FIFO will make the division using that factor look unusually profitable, for the accounting convention will make its costs appear to fall. LIFO will make it appear that its profitability is in decline. If two divisions are in competition within the corporation, one using inputs whose price is falling and the other using inputs whose price is rising, the choice between FIFO and LIFO can have an inverse effect on their apparent relative profitability.

Adverting once again to our cardboard example, when the manager of the box division draws cardboard from the inventory being accumulated by the cardboard division, he must for purposes of keeping track of his profitability choose one convention for imputing the internal cost of the cardboard to his box output. Just as he lobbied for an allocation of the cost of space to his division that most advantaged him vis-à-vis his fellow division manager, so now he will try to persuade the accountant to select an inventory convention that has the same beneficial impact on his bottom line.

Which brings us to the third and most problematic of the sources of accounting ambiguity, joint production, for that is precisely what is at stake when it comes time to allocate the cost of the space in which the two divisions carry out their productive activities. Speaking generally, joint production is the use of a factor input to produce two distinct salable outputs. In this case, the input is the company’s building and the outputs are cardboard and boxes. An oil refinery usually generates an entire array of products from its processing of crude, as does a slaughterhouse from its transformation of beef on the hoof into an assortment of meat products, hides, and other outputs. It is not too much to say that in a modern corporation, joint production is the rule rather than the exception.

The accountant, in preparing an annual report of a firm, is called upon to allocate the cost of the inputs that are used jointly in the production process. As Thomas demonstrates, there is no neutral pattern of allocations that can determine how much of the cost of such a factor is to be allocated or imputed to each unit of the several outputs in whose production process it is employed. The problems are manifold, as should by now be obvious. Thomas cites as one example an attempt to distinguish the cost of a building in which are carried out production processes having multiple outputs from the cost of the land on which it is built. The problem is that tearing down the building would so significantly alter the land values in the neighborhood in which the factory is located that the sale price of the cleared land would in no way reflect the cost to be allocated among the several outputs.

Well, enough is enough. Even these remarks, which might be characterized as accounting lite, are more than any sensible layman would want to read. Why does all this matter to someone like myself who is trying to assess Marx’s analysis of the transition from capitalism to socialism?


wallyverr said...

While waiting for further installments, I wonder if the accounting allocation problem is analogous to the index number issue discussed in the blog some time ago. While the index number problem is in a sense insoluble, it doesn't prevent the use of index numbers to increase the rationality of economic policy making. Compare the wealth of information on the state of the economy available to the Fed and the Bank of England today compared to what was available in the 1930s. Similarly, perhaps, use of a flawed accounting framework would be better than no accounting framework at all. (Max Weber attached great importance to the development of accounting.)

Robert Paul Wolff said...

That is a very interesting observation that requires some thought and a lengthy answer. Let me defer it for just a bit. I will try to return to it.