Intermediate Goods and Duplication

Historically, as Perlman and Marietta explain, various traditions of national income accounting arose in several countries in pursuit of different purposes. The German and UK accounting efforts were focused primarily on harnessing production in the service of imperial objectives, while the design of the National Bureau of Economic Research project in the US was more related to social welfare purposes. During World War II, however, the US Commerce Department's income accounting model evolved toward the goal of maximizing war production.

The clash of purposes – whether those were spelled out or not – came to a head in an exchange of views between Kuznets and economists from the Commerce Department, which took place in the August 1948 issue of The Review of Economics and Statistics. Kuznets had been asked by the journal to review the integrated set of national income and product accounts, which appeared for the first time in 1947 as a supplement to the Commerce Department's Survey of Current Business. "His review was unexpectedly devastating," according to Perlman and Marietta.

Kuznets was convinced that meaningful measures could only be defined "in relation to some end-goal of economic activity." He acknowledged the practical difficulties and the necessarily subjective element to such evaluation. The implicit goal of the Commerce Department's National Income and Product Statistics, according to Kuznets, was measuring the total output of the economy, not the nation's economic welfare. A subtext of Kuznets's criticism of the Commerce report was that once a particular approach became institutionalized, it would constrain future discourse about methods and policies.

In response to Kuznets's criticism, the Commerce economists maintained that their approach avoided moralistic judgments and that its emphasis on material production addressed the important national concerns arising out of the Depression and wartime mobilization. One moralistic judgment it couldn't escape, though, was the tacit assumption that current institutional arrangements were value neutral.
The defining technical issue that arose in Kuznets's critique and the Commerce economists' defense was the issue of "duplication" or the double counting of intermediate products. Kuznets emphasized, "nothing in the system of accounts compels one to define the totals as they are defined in the report." What is needed in order have "any basis for analysis" are criteria for distinguishing "final from intermediate products" and the report, in Kuznets's view did not provide those criteria.
"We assume that the final goal of economic activity is provision of goods to consumers, that final products are those turned out during the year to flow either to consumers or to capital stock (for the ultimate benefit of future consumers), and that everything else, by the nature of the case, is intermediate product whose inclusion in the output total would constitute duplication."
Kuznets objected most pointedly to the inclusion of all goods purchased by government as final products, arguing that a large part of government spending was devoted either to services to business or to law enforcement and military operations. He pointed out that those activities were not direct services to consumers but antecedent costs of maintaining society at large. To view government itself as the ultimate consumer, Kuznets warned, would be "fetishism, with dangerous implications that should be obvious."
"The huge amount of resources employed under government auspices and the tremendous visible flow of tangible commodities from such employment, particularly during wartime, may impress one strongly with the "reasonableness" of including governmental activity fully and the "unreasonableness" of omitting the overwhelming proportion of it by using the criterion we suggest. How can one exclude all this tangible performance from a total of a country's output? But it must be clearly recognized that the total we are seeking is that of product, of the end-result of activity – not of the volume of activity itself. Despite a great deal of activity, the product from the viewpoint of a clearly defined end, in this case services to ultimate consumers, may be quite limited."
The Commerce economists replied that it is not possible to draw a precise line between final and intermediate products and it is not feasible to examine each purchase by consumers, government or business to ascertain whether it was for use as a final or intermediate good. That's why it was essential to establish simplifying conventions that would enable compilation of national statistics, because "any measure of total production must be somewhat conventional."
"It must overlook the fact that the expenditures of individuals in their business capacity are influenced by their standards as consumers, and that the expenditures of consumers are influenced by their activities as producers. It must overlook also the fact that the conditions under which work is performed have an important bearing on the welfare of individuals…
"But though conventions are needed it is evident, as we have said, that different institutional arrangements can cause similar things to be classified as either final or intermediate and that some corrections of the data for these differences are possible."
The last paragraph could bear translation out of the passive voice. The Commerce economists were, in effect, conceding that Kuznets had a point but appealing for the reader to "trust the experts" to make the appropriate adjustments when needed. After maintaining that it would be too difficult and too subjective to establish explicit criteria along the lines Kuznets was advocating, the Commerce team tendered (albeit passively) their expertise oblige as a consolation. 

The meaning of "duplication" in social accounting may best be understood by analogy to the Red Queen's race in Lewis Carroll's Through the Looking Glass.
"Well, in our country," said Alice, still panting a little, "you'd generally get to somewhere else – if you ran very fast for a long time, as we've been doing."
"A slow sort of country!" said the Queen. "Now, here, you see, it takes all the running you can do, to keep in the same place. If you want to get somewhere else, you must run at least twice as fast as that!"
"Altogether," George Jaszi, one of the Commerce economists wrote in 1958, "this seems one of the most difficult subjects in national income theory." These difficulties, Jaszi continued,
"…are chronic ailments for which no general remedy seems to me possible in the framework of national output measurement, although I do not want to exclude dogmatically the possibility that at least some of the problems could be solved if we adopted a radically different approach."
Furthermore, the difficulties pertain not only to government purchases but also to consumer purchases in the light of technological change, institutional change and changes in consumer needs or tastes. One can almost hear the economists waving their arms and protesting, "there's nothing new here. Sure, the GDP isn't perfect, but it's the best we have and it does, after all, give us an idea of productive capacity – of the economy's ability to meet social needs if it was redirected to that purpose."

But that's simply not true. The capacity of the economy to produce the particular output that it produces is no indicator of its capacity to meet genuine consumer needs because current output doesn't take into account the costs of transition and barriers to conversion. Factored into the current aggregate of national income, for example, are the salaries of lobbyists paid to ensure that things don't change in a more progressive direction. Special interest pleading, after all, is a large part of what has been growing! Furthermore, current critiques of the growth imperative rely on a premise quite compatible with the Commerce economists' contention that "some corrections of the data for these differences are possible." The "embarrassing left" premise is that those necessary corrections have not been made and that the discrepancies have become untenably large – large enough to discredit the desirability of even more of the same.

Jaszi was correct in calling attention to the infelicity of the term "duplication" or double counting for something whose effects are only similar to double counting. The problem is not that the value is actually counted twice but that it is a value that ought not to be counted even once because it doesn't add anything to final consumption. Often the goods or services are compensatory for a prior or contemporary loss – the substitution of a costly good for a free good that is no longer available because of congestion, pollution or social fragmentation. Hueting refers to these kinds of entries in national income accounts as "asymmetrical" rather than duplication. Bartolini discusses how the rise of negative social and environmental externalities itself functions as an engine of measured GDP growth.

It should be non-controversial that "one cannot draw simple inferences from changes in output to changes in consumer satisfaction if there are concurrent shifts in needs, technological conditions, or institutions." That would be "to ask more of such measures than we can reasonably expect." Those are not criticisms. They are statements by Jaszi in defense of the Commerce Department methodology. Yet drawing simple inferences about consumer satisfaction is precisely what economists do when they assert the necessity of growth. When they do so they are not referring to the necessity of "some other kind" of growth but the necessity of more of what the current measurements measure.

These reflections on what it is that grows don't take into account the deliberate embezzlement of the system for the expressed purpose of "siphoning off a part of the annual increment in the gross national product" to pay for armament expenditures whose purpose was partly to enlarge that national product. That is addressed in Chapter 11 of my Jobs, Liberty and the Bottom Line.

Finally, there is the opinion – identified with Harrod and Domar – that economic growth is essential for maintaining full employment. That assertion, it can be readily shown, is based on the tacit assumption that other policies for maintaining full employment are not, or cannot be pursued and thus is no more than a circular argument.