by Matt Nolan
REVIEW: Capital in the Twenty-First Century
by Thomas Piketty
Harvard University Press (2014)
Capital, by Thomas Piketty, is a book that had to be written. Both he, and his translator Arthur Goldhammer, have come together to produce a text that is insightful, interesting, challenging, and clear to the public.
I don’t agree with all of the book (especially when it strays from its central thesis), have reservations about the descriptive conclusions, have strong reservations about the normative conclusion, feel that some logical conclusions were ignored for the sake of the central claim, and felt there are compelling alternative hypotheses that could explain the same data.
As an exploratory analysis of the amazing data set that Piketty, and others, has put together it was a well put together book. As an explanatory analysis, it provides a good starting point – but in truth it raises many more questions than it answers.
I have completed a detailed, 20 page review of the book here.
In section one I will summarise what I took to be the core thesis of the book. In section two I will discuss some of my concerns with assumptions in the first two parts of the book. In section three I will discuss part three of the book. In section four I will discuss part four. In section five I will discuss some alternate hypotheses that could have been discussed, and how they change the nature of policy conclusions. In section six I will conclude.
I will generally avoid small issues in the book which I disagreed with, but don’t have a substantive impact on the main thesis – each of the issues I will discuss are, in my view, centrally related to either the description involved or the appropriateness of the prescriptions that are put forward.
Summary of book and my concerns
Piketty is focused specifically on the two most aggregate factors of production: capital and labour. Capital in this context includes land, and includes residential property. Fundamentally, capital is seen as “the sum of all non-human assets that can be owned and exchanged on sum market”. Within this broad framework, Piketty is interested in asking what will happen to capital and labour income shares over the 21st Century, and what normative significance this has.
The core argument about what we can expect during the Twenty-First Century can be, arguably, boiled down to the following:
- In the long-run, the rate of growth, g, is given, and the savings rate, s, is determined primarily by social-cultural factors (and so can be seen as effectively constant).
- g is made up of a per capita growth rate and a growth rate in population. The population growth rate is going to fall. Therefore, g will decline over the long-term.
- With both s and g effectively given (“macrosocial” variables), and g lower in the future, the capital to output ratio (beta) must rise – what Piketty terms the Second Fundamental Law of Capitalism (SFLC), beta = s/g$.
- A dual argument that either i) the net return on capital, r, is effectively macrosocial, and can be taken as given or ii) the elasticity of substitution between labour and capital is greater than 1. With elasticity of substitution greater than 1, capital incomes share of total income will rise. Given the First Fundamental Law of Capitalism (FFLC) the share of income going to capital alpha = r*beta, an elasticity of substitution greater than 1 implies that the decline in r (determined in the market for savings and investment) will not be enough to offset the lift in beta from the SFLC.
- Historically these four points held, however the combination of two world wars, government appropriation and regulation, and a long but transitory lift in population growth led to a decline in capital income shares.
- Capital income is distributed more unevenly than wage income, and will continue to be following these shifts.
- Capital income, in “excess”, is less “moral” than wage income – the case for the immorality could be the change in the character of the individual/social groups, the threat to democracy involved, a social preference for egalitarianism, or a lack of opportunity/mobility.
- A coordinated global capital tax, given the prior assumptions of fixed s, and the response of r, can reverse this – and is normatively “right” policy.
This argument is based on a large set of historical data, and Piketty’s specific interpretation of it.
In parts One and Two of the book, Piketty argues for why capital to output ratios will rise, and why this in turn will lead to higher factor income paid to capital owners in aggregate. A derivation of the first result can be found in
Piketty (2010), while a discussion of optimal capital tax is in Piketty and Saez (2012).
Within this context, my concerns fall into two categories – concerns of omission, and concerns of truth.
- The omission (or potentially downplaying) of the fact that taxing savings/investment will fundamentally lower output.
- The fact a higher capital stock, and higher flows of savings per person, imply higher wages.
- A lack of definition of what “excessive” really means, and in what context, with relation to capital earnings.
- A focus on the heterogeneity of returns, without asking about the heterogeneity of individuals.
- The appropriateness of an elasticity of substitution greater than one.
- Treating savings (and potentially the real rate of return) as an exogenous “macrosocial” variable, rather than as something that is influenced by the choices and incentives of people in the economy.
- The use of adult personal income as a comparator to wealth.
- The use of adult personal income as a indication of inequality with reference to an egalitarian distribution.
- The use of the description of historical events, namely Britain post-Napoleonic wars, post WWII, and 19th Century views on poverty, justice, and merit.
- The strength of the value judgments around “earned vs unearned” income used to justify policy recommendations.
Let’s see if we can flesh out these concerns in a bit more detail (Review of Capital).
Note: In the review there were two things I directly said, but I’m not sure if my language was clear enough – but as I can’t access the file prior to posting I’ll just stay here. 1) In the example, the higher MPL from higher K drives up w and drives down r, that is why I state w/r is higher – my language isn’t quite clear enough in it. 2) when discussing the average as the minimum wage worker, the point is that in reality individuals/households “move between” income deciles a lot, making this an awful benchmark – again my language may not be clear enough.
If you catch logical mistakes in the review, I’d love to hear from you – the reason I wrote this without reading anything else was to ensure that I gave the book an honest appraisal, on the basis of my actual understanding. I’m more than happy to learn where my understanding is wrong.
Matt Nolan is an economist at Infometrics