Keynes and econometrics

After the 1920s, the theoretical and methodological approach to economics deeply changed. Based on a criticism of Marshall’s work and legacy, a new generation of American and European economists developed Walras’ and Pareto’s mathematical economics. As a result of this trend, the Econometric Society was founded in 1930.

The constitutional assembly was held in  Cleveland, Ohio, during the annual joint meeting of the American Economic Association and the American Statistical Association. The Norwegian economist Ragnar Frisch played an important role in the Econometric Society that was founded to enhance studies based on the theoretical-quantitative and the empirical-quantitative approach to economic problems. In this way, the  founding fathers believed that  economic thinking could be as rigorous as the one that dominates the natural sciences.

At the 5th European Meeting of the Econometric Society, in 1935, Jan Tinbergen presented a paper on ‘A mathematical theory of business cycle policy’ that followed the Econometric Society’s guidelines. His causal explanation of the business cycle began with a priori economic-theoretical considerations about explanatory variables and then he proceeded to test a model.

In the late 1930s, John Maynard Keynes and other economists objected to this recent “mathematizing” approach. Keynes, as editor of the Economic Journal, wrote  a negative review of Tinbergen’s 1939 book A Method and its Application to Investment Activity. This book  presented an statistical testing of business cycle theories based on the application of the method of  multiple regression and  mathematical framing in the form of a specified model. At the core of Keynes’ concern lied the question of methodology. Recalling his own words:

Am I right in thinking that the method of multiple correlation analysis essentially depends on the economist having furnished, not merely a list of the significant causes, which is correct so far as it goes, but a complete list? For example, suppose three factors are taken into account, it is not enough that these should be in fact veræ causæ; there must be no other significant factor. If there is a further factor, not taken account of, then the method is not able to discover the relative quantitative importance of the first three. If so, this means that the method is only applicable where the economist is able to provide beforehand a correct and indubitably complete analysis of the significant factors. The method is one neither of discovery nor of criticism. It is a means of giving quantitative precision to what, in qualitative terms, we know already as the result of a complete theoretical analysis. (Keynes 1939: 560)

In this paragraph, it is clear that Keynes doubted the use of inductive methods of generalization and statistiicial inference to build economic theories because of the peculiarity of the economic systems characterized by:

  • a low degree of homogeneity,
  • a high degree of complexity
  • the lack of stability through time.

Indeed, on behalf of the peculiarities of the economic systems, Keynes highlighted that econometrics turns out to be a method not of testing or of discovery, but of measurement of selected variables.



Keynes, J. M.,  Professor Tinbergen’s Method, The Economic Journal, Vol. 49, No. 195 (Sep., 1939), pp. 558-577. Published by: Blackwell Publishing for the Royal Economic Society. Stable URL:

Tinbergen, J. A Method and its Application to Investment Activity. Geneva: League of Nations, 1939.

  1. Rob Reno said:

    Thank you. Wonderful resource.

    • Maria Alejandra Madi said:

      Thanks Rob for your comment!

    • Maria Alejandra Madi said:

      Thanks for reblogging the post!

  2. Excellent Maria — this make the point the econometrics is a tool of confirmatory data analysis – applicable when the model is already known. The second point, that we must already have picked up ALL relevant regressors, if the regression is to make sense, is made at length in my earlier post on “Choosing the Right Regressors“. This points out something which was obvious to Keynes, but is ignored by vast majority of practicing econometricians: Missing one significant regressor can easily lead to strongly misleading regression results.

    • Maria Alejandra Madi said:

      Thanks Asad for your stimulating comment. Your second point is extremely relevant. The problem of the selection of regressors – as you already wrote in earlier post – was also highlighted by Keynes in the review of Tinbergen’s book.

  3. Coming across Keynes while assessing the reliability of reliability theory, I was at the same time learning about sampling in quality control. When, five years later, I was introduced to C E Shannon’s Mathematical Theory of Communication (by then a quarter of a century old), I discovered that the sampling theorem and use of duplicate parts to improve reliability in spacecraft were based on Shannon’s mathematics and insight about the connstructive use of redundancy. Here I think the wording of Maria’s conclusion is missing the key word and Asad’s interpretation of it is correct: econometrics is measuring not the variables but the reliability of them. Induction, despite the assumption in Keynes’s and Popper’s time, is not about proof or disproof but quality control: theories being sufficiently reliable to be used in practice.

  4. Maria Alejandra Madi said:

    Thanks for your comments.

    In addition to the limitations of econometrics as a method of inquiry, the problem of the selection of regressors – as Asad already wrote in earlier post – was also highlighted by Keynes in the review of Tinbergen’s book.

    Keynes carefully analysed the legitimacy of Tinbergen´s implicit assumptions such as the completeness of the list of the significant causes and the independence and measurability of all the significant factors, The British economist believed that these conditions are far from being satisfied in the field of economics, Keynes´s concern was about those significant factors that are not measurable. The operation of econometrics does not cope with those economic problems where political, social and psychological factors that may be significant.

    Indeed, Keynes put in question if the statistical test can prove a theory to be correct or incorrect. It seems to his answer is NO.

    • The reliability connection is perhaps one teachers need to latch on to. If one draws a graph, not of reliability but of UNreliability over time, when a new piece of equipment (say a car) is introduced, the graph has the shape of a bath-tub: a steep drop as design faults are ironed out, then a relatively long period of reliable service, but at the other end of its lifetime a gradual rise in unreliability as the bits wear out. The same “bath-tub curve” applies to prices, which reflect the unavailability of goods when first introduced or becoming obsolete. It applies also to the rate of shipwrecks: much higher when starting from or leaving port.

      The behaviour of navigators and entrepreneurs however has significant differences, If the navigator has to divert to avoid on-coming traffic, he returns to his original course once the danger is past. The entrepreneur actively seeks high prices, building early failure into his products and staying on his new course only as long as it is profitable – with never a thought for passengers who signed up for his original product. (I’m saying this feeling very sore about an expensive Lenovo lap-top with Windows 10 software becoming inoperable soon after the warranty expires, an internet inquiry showing this to be far from an isolated case. However, it applies equally to the Brexit issue, brought about by money-grubbers turning the EEC into an American-style EU).

      I’m going to argue that economic pedagogy needs to start not with what to teach university freshmen but with school-children learning economic relationships from meaningful diagrams like the bath-tub curve, the Wheatstone’s Bridge equilibrium circuit and the PID negative feedback information circuits needed to follow a course in navigation and democratic control. The argument of Keynes’ General Theory is obvious when explained in terms of this last. If a Copernican revolution is to happen in economics, by the time anyone gets to university everyone needs to be taking for granted the monetary equivalent of the earth going round the sun, despite appearances.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: