In praise of Wesley Clair Mitchell, economist
There are three main strands of Macro-Economic Measurement (MEM): National Accounting, the Flow of Funds and Business Cycle analysis. And one minor strand: the measurement of time poverty. This last strand is not yet macro but it does try to tack an estimated household production function to the national accounts data. Wesley Claire Mitchell, proud student of Thorstein Veblen and long-time head of the National Bureau of Economic Research is important for all of these and even stood at the cradle of two of them.
One pivotal publication of Mitchell is ‘The backward art of spending money’ (Mitchell (1912)) which stresses the importance of the household and the difficult art of spending money wisely as well as the pivotal importance of the mistress of the house for the welfare of the household (and society). I’m not going into this discussion but to avoid misunderstandings: stressing the pivotal importance of the household and the housewife is feminist. This article is entirely consistent with the work of the Levy institute on ‘time poverty’. It tries to estimate a kind of household production function and is based on the idea that “The predominant framework for measuring poverty rests on an implicit assumption that everyone has enough time available to devote to household production or enough resources to compensate for deficits in household production by purchasing market substitutes. … Ajit Zacharias argues that this implicit bias in our official poverty statistics threatens to undermine the Sustainable Development Goals”.
Another pivotal publication of Mitchell is ‘The Role of Money in Economic Theory’ (Mitchell (1916)). In this article he praises Veblen, lambasts the Austrian school and people like Jevon and praises Alfred Marshall as Marshall tried to introduce real prices and real production into economic theory while the a-historical work of the ‘marginalists’ amounted to little more than fruitless speculation about the supposed psychological behavior of mankind. Mitchell put his money where his mouth was and in 1921 went on to publish, with King and Macaulay, the milestone ‘Income in the United States: Its Amount and Distribution, 1909-1919’ which estimated time series of nominal and real income in the USA. Around 1920 there were quite some individual researchers working on this. In Bowley (1942) a very extensive literature overview can be found, subdivided by country. But this study (followed by much more detailed publications in 1922) sure was a step ahead and as such it was the foundation on which a student of Mitchell, Simon Kuznets, could work.
Next to this Mitchell worked on business cycles and published several works on this. Tjalling Koopmans criticized this as ‘measurement without theory’. However, reading Koopmans nowadays gives one the impression of a statistician who had mastered multiple regression and thus wanted all economics articles to be stated in a ‘multiple regression’ format, but who had never heard of principle component analysis. Modern economists using principle component analysis like Andrle, Brůha and Solmaz are, contrary to Koopmans, very favorable about this kind work of Mitchell.
When it comes to the Flow of Funds we can cite from the annual NBER reports of 1944 and 1948 which assigned Morris Copeland with the task of establishing the Flow of Funds and Milton Friedman with the task of writing a monograph on money which would result in his 1963 Friedman and Schwarz ‘A monetary history of the United States’ book (much better than Friedman’s work on permanent income and, as Rockoff argues, distinctly Mitchellian in approach):
In a 1945 NBER publication, The National Bureau’s First Quarter-Century, on pp.61-62 Copeland’s task (accomplished together with people from the Fed) was described as:
“About the circuit flow of payments and its relation to national income and output, our knowledge is exceedingly vague. We do know, however, that the flow of payments does not adjust itself automatically to the flow of goods men are able to produce and need to consume. Indeed, several theorists have argued that cyclical fluctuations in business activity are due primarily to recurring changes in the relative size of these two flows. The findings this investigation promise should put us in a far better position to diagnose our recurrent chills and fevers, and to seek remedies”.
This present reason given by the British Office of National Statistics to integrate the Flow of Funds with the national accounts still neatly mirrors this assignment:
“An understanding of the economic performance of the UK is especially important for effective policymaking and improving welfare. The non-financial accounts have long been extensively monitored as a health check for the economy, but they do not fully capture the build-up of financial risk. For instance, changes to the underlying resilience of the UK’s source of funding can impact the economy in a way that is not obvious from studying fluctuations in income or output … We have partnered with the Bank of England to address this, by enhancing the coverage, quality and granularity of financial accounts statistics for the UK” (ONS (2018)).
In the 1948 annual account of the NBER Friedman’s task was described as:
“Work on another monograph, dealing with the cyclical behavior of the money supply, its rate of turnover, and the condition of banks in different parts of the country as well as in the aggregate, will get actively under way this year. The study will go back to the Civil War, but will give special attention to developments in the sphere of money and banking since 1914 when the Federal Reserve System was instituted. Milton Friedman, Associate Professor of Economics at the University of Chicago, is in charge of the study”.
Friedman and Schwarz did not apply quadruple accounting methodology to their series and consequently, alas, did not manage to integrate their estimates of money with the Flow of Funds. This set the economics profession back by decades. Nowadays however, the Flow of Funds is the basis for the estimates of the amount of money as well as of flows of other money-like assets. They are also increasingly integrated with the National Accounts. While the work of Andrle, Brůha and Solmaz shows how methods for business cycle analysis can be applied to such integrated data. Data on the household production function as estimated by the authors of the studies on ‘time poverty’ can be added to this, to provide an overview of household prosperity consistent with the national accounts. It would not have surprised Mitchell. On the contrary. He would be amazed that this is not yet accomplished!
From: pp.2-3 of WEA Commentaries 8(3), June 2018