Lesson 8 of 8
The hard part
Published statistics measure something adjacent to the production structure, not the structure. Hayek disowned the simplification that made it countable.
By Javier Sanz
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
In one minute
- Hayek disowned the average period of production in 1941, calling it a simplification that evaded essential problems and whose tools were almost useless in complicated cases.
- Lachmann argues capital has no natural measure at all, because unexpected change alters the value of every combination.
- The nearest published statistic, gross output, counts sales between industries. That is adjacent to the structure and is not a measure of distance from consumption.
Everything so far has been a picture. This lesson asks what happens when somebody tries to find the picture in data, and the answer is that the honest version of this theory is harder to test than its friends usually admit.
What the nearest published series actually counts
What gross output includes
- Sales to final users, which is what value added captures.
- Sales to other industries, the intermediate inputs.
- So it counts the transactions between firms that value added nets out.
- Published quarterly by industry, alongside the GDP by industry release.
What it does not tell you
- How long any process takes from first input to final sale.
- Which industries sit near consumption and which sit far from it.
- Whether a rise came from more stages or from more sales per stage.
- Anything about the specificity of the capital involved.
Both are sales measures by industry. Neither is a time measure.
The author's own retreat
The clearest statement against measuring the structure comes from Hayek. In the preface to The Pure Theory of Capital, published in 1941, he describes the most important of the inappropriate simplifications he had used earlier as the attempt to introduce time into capital theory in the form of one relevant interval, the average period of production.
He goes on to say that this supposed simplification evaded so many essential problems that replacing it raised a host of new questions, and that some of the tools taken from the earlier writers were almost useless in analysing more complicated situations.
That is the man who drew the triangle, ten years later, saying the countable version of it does not work.
Lachmann's stronger claim
Lachmann goes further in Capital and Its Structure. His argument runs through heterogeneity: because unexpected change alters which combinations are viable, and because value is affected by every such change, it is impossible to measure capital, and capital has no natural measure.
He still wants a generic concept and a way of classifying capital goods, but he explicitly gives up on a common denominator. If he is right, a number for the length of the structure is not a hard measurement problem. It is a category error.
What gross output can and cannot do
The Bureau of Economic Analysis publishes gross output and defines it as the value of goods and services produced by the nation's economy, principally measured from industry sales or receipts, including both sales to final users and sales to other industries.
That is genuinely useful and it is genuinely not the thing. It counts the inter-firm transactions that value added removes, which is why people reach for it here. But the count of transactions is not the count of stages, an industry's position in the accounts is not its distance from a shopper, and nothing in the series carries time.
A rise in gross output relative to value added is consistent with a longer structure. It is equally consistent with the same structure and more outsourcing.
What is left that a reader can do
Not much at the level of an economy, and that is the honest answer. At the level of a company there is more: a filing says what a firm sells, who buys it, how long its projects take, and what it has committed to build. None of that is a measurement of the structure and all of it is evidence about one firm's place in it.
The applied courses in this academy go that way on purpose, because the company is where the filings are.
Check yourself
What did Hayek say in 1941 about the average period of production?
- That it was the most useful tool in capital theory
- That it was an inappropriate simplification which evaded essential problems, leaving the older tools almost useless in complicated cases
- That it should be measured quarterly
Show the answer
Answer: B.
He says it in the preface to The Pure Theory of Capital, about his own earlier work, and says replacing it raised a host of new questions.What is Lachmann's argument against measuring capital?
- That the data are not collected
- That unexpected change alters which combinations are viable and affects value, so capital has no natural measure
- That only physical units count
Show the answer
Answer: B.
He still wants a generic concept and a classification, but explicitly abandons a common denominator, which makes a single length of the structure a category error rather than a hard estimate.Why is gross output not a measure of the production structure?
- Because it is published too late
- Because it is a sales measure by industry that carries no time dimension, so a rise is equally consistent with more outsourcing as with more stages
- Because it excludes intermediate inputs
Show the answer
Answer: B.
Its own definition is industry sales including both final and inter-industry sales. Position in the accounts is not distance from a shopper.Sources
- Hayek, The Pure Theory of Capital, 1941, preface, on the attempt to introduce the time factor as a single average period of production being the most important inappropriate simplification, on it evading essential problems, and on some of the earlier writers' conceptual tools being almost useless in more complicated situations.
- Lachmann, Capital and Its Structure, 1956, chapter 1, on unexpected change altering capital combinations and value, on the impossibility of measuring capital and its lack of a natural measure, and on the need for a classification in place of a common denominator.
- Bureau of Economic Analysis, glossary entry for gross output, for the definition as the value of goods and services produced by the nation's economy, principally measured from industry sales or receipts including sales to final users and sales to other industries.
- Bureau of Economic Analysis, GDP by Industry, for the release in which gross output is published alongside value added.
What this leaves out
- This is a heterodox tradition. The mainstream account of the same phenomenon differs, and nothing here is a forecast or a view on any security.
- No figures from the gross output series are quoted here, because a number would imply the series measures the thing this course is about, and it does not.
- Attempts have been made to test claims about the production structure using published statistics, and they are contested on both sides. This lesson names the measurement problem rather than adjudicating those attempts, and the course on the case against the theory takes the empirical record up directly.