Longtime readers of this blog will know that I am a bit fascinated with the uses and misuses of Adam Smith’s work. For that reason, I am a big fan of Gavin Kennedy’s blog, Adam Smith’s Lost Legacy. Kennedy has been fighting against contemporary misinterpretations of Smith, with a special focus on the “myth of the invisible hand” (for an article-length summary of his position, see here). His work, along with Emma Rothschild, Warren Samuels, and others, has shown that the concept of the invisible hand emerged in 20th century economics, and was not central to Smith’s own writing. Smith used the phrase “invisible hand” just once in the Wealth of Nations, and it meant something more like “unintended consequences” than the harmonious workings of a perfect market. Emma Rothschild goes so far as to call it a “mildly ironic joke,” though I prefer to emphasize that the invisible hand was a common metaphor of the time that had little special significance for Smith. So, when did we start thinking of Smith as being the “theorist of the invisible hand” (and, around the same time, forget all his opposition to strict laissez-faire)? Gavin Kennedy’s historical work places the disjuncture in an oral tradition in England and the US in the late 19th to early 20th century. The myth remained somewhat confined until the publication of Samuelson’s extremely influential undergraduate text, Economics in 1948, which brought the myth to the masses (Kennedy 2010).
One way to see the influence of Samuelson’s publication, and to check the overall plausibility of the Kennedy-Rothschild-Samuelson (K-R-S) account of the myth of the invisible hand, is simply to check how often the phrase was referenced. Thankfully, JStor’s Data For Researchers tool makes this task trivial. Here’s a graph of counts of articles mentioning “invisible hand” over time:
Mentions of "invisible hand" in JStor's Corpus
There is a trickle of mentions leading up to the 20th century, with a takeoff following 1950 (and thus, a few years after Samuelson’s text was published).
Here’s a similar graph for mentions of “adam smith” AND “invisible hand”:
Mentions of "adam smith" and "invisible hand" in JStor's corpus.
Note that the trickles earlier on drop out – as “invisible hand” was a somewhat common metaphor in Smith’s time, it makes sense that other 18th and 19th century authors invoked it, but they did not appear to do so in reference to Smith. By the 1920s, and especially in the 1950s and after, we begin to see Adam Smith and the invisible hand coming together. A more detailed analysis is also possible in JStor, as we could look at the articles referencing both Smith and the invisible hand to see exactly how authors in the 1920s-1950s understood the phrase, and compare it to uses in articles from the 19th and early 20th century that did not reference Smith. Even in the absence of that more detailed analysis, I think we can safely say that the link between Smith and the invisible hand is a relatively recent creation. The evidence for the K-R-S position is strong.
Update: A commenter requested relative statistics. Here’s some data. 1090 articles published between 1770 and 1880 in JStor’s corpus mention Adam Smith. Of those, just *3* mention the invisible hand (and it’s not clear that they are referring to Smith’s invisible hand, I don’t have access to some of the early pamphlets). 1880-1900, it’s 8 out of 1319. 1900-1940, it’s 65 for both out of 4,243 for just Smith. By 1960-1980, it’s 514 mentions of both out of 7017 mentions of Smith. So, for clarity:
1770-1880 – .3% of articles mentioning Adam Smith also mention invisible hand
1880-1900 – .6%
1900-1940 – 1.5%
1940-1960 – 3.9%
1960-1980 – 7.3%
A (Budding) Sociologist's Commonplace Book 
Jay Livingston
/ March 7, 2012Google N-gram shows the same thing. A rise from 1800 to 1815 or so, a very gradual decline to 1960, then the boom. The graph for Adam Smith alone does not follow this same pattern. In fact, in the early 1960s, as the invisible hand becomes more and more visible, the trend in Adam Smith is down.
Paul Walker
/ March 7, 2012First your graphs are correlation rather than causation. I think you need to adjust for things like number of journals/articles in JStor over time. For example if 10% of all articles published 100 years ago mention the invisible hand and today 5% of articles published mention the invisible had but there are 100 times more articles published today than 100 years ago you will get your patten without there being any more interest in Smith or the invisible hand. Also you need a measure of general interest in Smith’s work. If more people today are interested in Smith’s work today no matter what their views on the invisible hand you could get your patten.
Also I posted this as a comment to one of Gavin’s posts a couple of days ago. But I feel it is relevant here as well:
“Let us assume for the moment that Smith never used the term “invisible hand” at all. Could we not still use to phase to describe the general thrust of Smith’s argument? As Eamonn Butler has noted,
“In fact, the critics read too narrowly. The invisible hand idea, as commonly understood, pervades Smith’s work, and would do so even if these two specific references had never existed. For the phase is a very convenient shorthand for Smith’s idea that human actions have unintended consequences; and that provided a few fundamental rules such as the principles of justice are followed, the self-serving actions of individuals can unintentionally produce a well-functioning and beneficial overall social order”
or as Craig Smith has written,
“It is the idea of the invisible hand, or more generally the idea of social evolution through unintended consequences, which represents Smith’s chief legacy to the modern world. The recognition that many of the most important human achievements are, as Smith’s friend Adam Ferguson observed, the results of human action, not the product of human design, is a profound lesson to us all. It is this observation which leads Smith to his deep scepticism towards ‘mean of system’ who would organise humanity to achieve noble ends.”
Thus can we not, in the above sense, apply the phase “invisible hand” to Smith’s work?”
Jay Livingston
/ March 7, 2012What I got from Dan’s post was not so much what “invisible hand” really means or what Adam Smith intended it to mean. What’s important is what it came to mean in the last 50 years. I read the post as being less about Smith and more about Paul Samuelson – specifically that the Samuelson’s version of “invisible hand” found great resonance in the late 20th century. And as the Google N-gram I referred to shows, it was “invisible hand” and not “Adam Smith” that drew the big increase in usage. (On converting the number of references into a ratio by counting all articles, it seems unlikely that the number of all articles took off in 1961 with the same acceleration as did “invisible hand,” which increased by a factor of 10 in 30 years. )
I’m sure there were other things about Samuelson’s text that made it such a hit, but its version of “invisible hand” apparently crystallized a set of ideas that was coming to dominate economic thinking.
Dan Hirschman
/ March 8, 2012Jay, I think you’ve captured the thrust of the post nicely. And yeah, I played with Google NGrams as well and found similar results to yours. NGrams’ metadata are a bit less reliable (especially when looking for a small number of mentions pre 1900), and here I think academic discourse is more interesting, so I stuck with JStor.
Paul, I don’t think the modern myth of the invisible hand is simply that human institutions evolved rather than were designed, or that unintended consequences are prevalent. The “invisible hand” is used a codeword for the perfection of the market, and impossibility of intervening in the market system to produce beneficial outcomes. While Smith may have believed that, in general, markets worked best when left relatively free from interference, he was too pragmatic and historical to turn that into dogma. Hence, Viner (1929) notes 60 places in the Wealth of Nations where Smith calls for more government action, in fields as wide ranging as finance, shipping and providing free education (to counteract the stupefying effects of division of labor in a society where factory work was beginning to dominate!).
As to correlation vs. causation, of course! But this here is correlation backed up by fine historical research by three of the best historians of economics. All I’m trying to do here is to show that the macro-level story is consistent with what they’ve been arguing all along. Also, in re: relative statistics, the story is much the same. Using JStor again, 1090 articles published between 1770 and 1880 in JStor’s corpus mention Adam Smith. Of those, just *3* mention the invisible hand (and it’s not clear that they are referring to Smith’s invisible hand, I don’t have access to some of the early pamphlets). 1880-1900, it’s 8 out of 1319. 1900-1940, it’s 65 for both out of 4,243 for just Smith. By 1960-1980, it’s 514 mentions of both out of 7017 mentions of Smith. So, for clarity:
1770-1880 – .3%
1880-1900 – .6%
1900-1940 – 1.5%
1940-1960 – 3.9%
1960-1980 – 7.3%
Run the numbers yourself. Adam Smith and the Invisible Hand start going together in the 20th century, esp. post-WWII, even as the number of articles skyrockets.
Paul Walker
/ March 9, 2012“The “invisible hand” is used a codeword for the perfection of the market, and impossibility of intervening in the market system to produce beneficial outcomes.”
As an economist I’m not sure what I’m sure what “perfection of the market” means. I would assume you mean the 1st and 2nd welfare theorems but they have nothing to do with Smith’s thought. As Mark Blaug has noted “[ … ] Smith’s faith in the benefits of ‘the invisible hand’ has absolutely nothing whatever to do with allocative efficiency in circumstances where competition is perfect a la Walras and Pareto; the effort in modern textbooks to enlist Adam Smith in support of what is now known as the ‘fundamental theorems of welfare economics’ is a historical travesty of major proportions. For one thing, Smith’s conception of competition was, as we have seen, a process conception, not an end-state conception. For another society, a decentralised competitive price system was held to be desirable because of its dynamic effects in widening the scope of the market and extending the advantages of the division of labour – in short, because it was a powerful engine for promoting the accumulation of capital and the growth of income.” (Mark Blaug 1996. Economic Theory in Retrospect. 5th edn. 60-1. Cambridge: Cambridge University Press.)
Something I thought of after posting the above is that I would guess that the use of the expression “invisible hand” is positively correlated with the interest in GE modelling. The growth in usage from roughly 1950-2000 would be consistent with this.
Also among the economists I have talked with about this, an obvious small sample problem here, the Butler/Smith type interpretation seems the most common one.
Dan Hirschman
/ March 9, 2012Paul,
I think we are in agreement, but I am now a bit lost. I agree with Blaug’s quote entirely – the “myth of the invisible hand” that Kennedy, Rothschild and Samuls are documenting is exactly this sort of tying of Smith’s work to the modern claims about the allocative efficiency of markets.
If you are engaged in a project to rescue the phrase “invisible hand” from its current association with arguments about the efficiency of markets, and recast it as a more institutional concept about the power of unintended consequences, good luck. But you have a very long slog ahead of you. For the moment, I think the more valuable project is disrupting the associations Adam Smith = Invisible Hand = Market Efficiency++.