The definition of capital and the difference between the productivity and the profitability of capital

Introduction
Reading the discussions about Piketty’s book Capital in the 21st century there seems to be considerable confusion about the concept ‘capital’, for three reasons.  First, the book is about 200+ years of ‘capital’. In this period, the (legal, cultural) concept of ‘capital’ changed which means that, to understand the book, a ‘historical’, non-homogenous understanding of capital is needed. Second, There also seem to be a habit among readers to mix up the concept of ‘capital’ of the asset side of the balance sheet (i.e. physical capital) and the concept of ‘capital’ of the liability side of the balance sheet (financial capital, i.e. all kinds of, often tradeable, ownership claims regarding physical capital, ‘physical’ capital nowadays increasingly consisting of intangibles).  This confusion is part of the state of economics, see for instance this Wikepedia entry about the marginal productivity of capital. Third, there is confusion about the nature of ownership rights and power. In this post, I’ll use the concept of the balance sheet as well as some data on long-term nineteenth century land rents to try to explain the concept in a more orderly fashion – using, among other things, Ricardian concepts and some data on ownership rights and land rents in the nineteenth century and the large difference differing ownership rights made to the return on capital (but not to the productivity of land). I presuppose that the reader has a little knowledge about ‘balance sheets’. Even experienced users of balance sheets might however be reminded that concepts of ‘property’, which are central to the entries of the balance sheet, do change in the course of time. In the Piketty’s book, the disappearance of ‘slave capital’ is the obvious, even glaring example. This disappearance of an entire entry on the national balance sheet of the USA is not just about accounting. All of us still know, about 150 years later, that it did not take the stroke of a pen but a bloody war to abolish slavery in the USA. This knowledge alone is already proof that the concept of property, and therewith capital, is not just a legal construct but also deeply embedded in a culture and large changes reverberate for centuries. The slave example also shows that property rights are also about power – power over physical capital (how should it be used, who should or are allowed to use it) and therewith also over people. To use a ‘Piketty’: ‘Gone with the wind’, a movie about the USA civil war, is a splendid story about (the demise of) rent income, the culture and the slave society of ‘the south’ of the USA. But ownership of ‘capital’ also gives one power over labour and labourers in non-slave societies which is a continuing reason for social conflicts, as the discussions about ‘flexible’ labour markets and ‘disciplinary’ 27% unemployment rates in the Eurozone show. The complicated and changing nature of ‘capital’ makes it difficult to define one of the favorite concepts of economists, ‘the marginal productivity of capital’. Below it will be argued that ‘the marginal profitability of capital’, a more nominal concept and which unlike ‘marginal productivity’ is more explicitly tied to changes in relative prices and the historical concept of capital in a society, is a better phrase.

 

2. How to define capital?

What is ‘capital’? Land? Roads, buildings and machinery?  The total value of the asset side of a balance sheet? Or the total value of the liability side of a balance sheet (arithmetically the same thing, but conceptually not)? Is it the discounted stream of income from capital, the discount rate being the return on some kind of capital (hmm… sounds tautological)? Or is it essentially a lever of power over the production process, including the way labour is used and profits or surplus (nowadays also named, when it comes to companies: EBITDA, Earnings Before Interest, Taxes, Dividends and Amortization) are distributed? Maybe a bit of all of this… Let me elaborate:

Around 1800 land was, as Piketty shows, in France and the UK and the ‘United States North’ easily the most important kind of ‘capital’ on the asset side of the national balance sheet – it accounted for about 50% of the total value (Piketty, p. 161). There was an exception to this rule of thumb, however. In the ‘United States South’, land was relatively less important because the value of slaves accounted for 50% of the total value of capital while the total amount of capital, compared with GDP, was twice as high as in the ‘United States North’. In the north, land, houses and other capital were as important (compared with GDP) as in the south, but virtually no ‘slave-capital’ existed over there. The abolishment of slavery in the USA took, as we all know, a bloody war. As a consequence, an entire and very important part of the total stock of capital disappeared – which destroyed a society. And which also shows that ‘capital’ and property are relations of power, backed by the power and the law and even the legitimacy of the state and deeply embedded in a culture and a society. It often takes a war or a revolution to change these relations. But sometimes they do change – and the social, political, economic, legal and cultural changes connected with these revolutions often reverberate for centuries. Think about the Reformation (which, in protestant areas, led to the loss of the extensive land holdings of the catholic cloisters and the legalization of ‘pure’ interest). Or the French Revolution (which was, to quite an extent, about reshuffling and guaranteeing ownership rights, including inheritance rights). Or the Russian one. Or Japan after World War II. Everybody still knows at least a little about these events. And it’s important to note that the concept and the value of entries on the asset side of the balance sheet are, in the end, dependent on the power and the laws of the state – while these same laws give power to the owners of the entries on the liability side, power over property and even to an extent power over people who, as a slave or a labourer or a business owner, use the assets to produce (yes, also power over business owners. Did you never hear a business owner complain about ‘the bank’?). Concepts and ownership of ‘Capital’ really define our societies. Balance sheets show, on the asset side, the different kinds of ‘physical’ capital and on the liability side the owners of this capital (remember that when you default on a loan, an entry on the liability side of your balance sheet, the bank can often occupy ‘your’ property). Which shows that ownership and capital (liability side of the balance sheet, let’s call this financial capital) has to be defined in a broad, long-term sense, including rights to the return on capital in the shape of interest or dividends, the possession rights of lenders and profit and inheritance rights. Financial capital is, to the very core, about power.
The liability side of the balance sheet mainly is a snapshot from the past: which financial ‘investments’ (including retained profits, loans and revaluations) financed the ‘physical’ items on the asset side and what does this mean for ownership rights (remember that every kind and act of financing comes with a reshuffling of state backed ownership rights). The asset side however shows the present: how much are these items worth (according to accountants), at this moment. The posts on the liability side define (in tandem with the legal system and the culture of a society) who gets the spoils and who calls the shots. This is the side which is important when it comes to the division of the national pie (well, of among other items the ‘EBITDA’ part of this, which some might call ‘surplus’). The posts on the asset side, however, comprise the ‘K’ used in productivity analysis (let’s call this physical capital): machinery, buildings, roads, bridges, land, copyright etcetera. This is the side which is important in ‘growth accounting’, where a weighted average of these posts is used to calculate the increases or decreases in the stock of physical capital. In the discussion about Piketty, economists use the asset and the liability side interchangeably. But these sides are not interchangeable. An increase in mortgage debt associated with an increase in prices of existing houses sold in the past does increase the flow of interest to the banks. But it does not increase the amount of physical capital (the amount, location and quality of houses), though the estimated value of physical capital might increase when we use a ‘mark to market’ approach to value existing houses, instead of, for instance, the historical purchase price. Summarizing: capital is not one thing – it is different things. Any definition of capital should take heed of this. When we want to analyse economic growth, we might want to use an index of the amount of physical capital. When we want to study inequality, it might be wiser to look at the liability side of the balance sheet, which can change even when the amount and composition of the posts on the asset side does not change. And we should of course look into the legal and cultural system. An unsolved problem mentioned by Piketty is for instance whether extremely high CEO salaries should be considered to be part of labour income or of capital income (bonuses are of course part of capital income, as they are part of EBITDA and the capital owners in the end decide about them). The total value of the posts on the liability side are of course equal to the total value of the posts on the asset side. But that does not mean that both sides can be used interchangeably.

Also, due to all kind of events:

* the value of the posts on the asset side can fluctuate wildly
* the owners of the posts on the liability side as well as the ownership rights embedded in these posts (like ‘rent control’) can change beyond recognition
* and the return on different kinds of capital (liability side) can change enormously.

History abounds with examples. The run up in real house prices until 2008, financed by trillions and trillions of bank credit, the decline of interest rates on savings accounts engineered by central bank policies and the rise of the oligarchs in ‘emerging Europe’ are recent examples. The random nature of these events, not directly connected to any kind of changes in physical productivity of physical capital, calls into question the concept of the marginal ‘productivity’ of capital, which is, by necessity, a monetary concept (interest, rent, profits and dividends are utterly monetary variables). Should the money created by trillions of new mortgage loans be understood as a productive investment, even when the amount of houses does not increase? Does productivity of savings accounts decline when the central bank lowers interest rates? Does the marginal productivity of the stock of houses increase when (as is happening in the Netherlands) the government increases real rents with 8% in two years? Productivity is a physical concept. It’s the ratio of outputs to inputs (Wikipedia agrees). Often, outputs and inputs are weighed with prices (which are, of course, monetary defined variables), to calculate income. But changes in these prices do not necessarily imply that the physical relation between outputs and inputs changed, too. Also, changes in the amount of ‘claims’ on the liability side of the national balance sheet, be it because of changes in rents or an increase in mortgage loans, do not necessarily alter the level of production, though it might change the division of the surplus (QE is a recent example)

 

A historical example

A classical example of random changes in income related to property is of course the case of ‘Ricardian’ landrents. Increasing prices of agricultural products did lead to large increases in land rents and not to large increases of wages or mixed income of farmers renting land. Such developments tell us an important part of the dramatic story of progress, prosperity and setbacks and suffering in the nineteenth century. But they tell us very little about the story of changes in physical productivity of the land. In a blogpost, Brad deLong uses the idea of ‘reverse agricultural Ricardian rents in combination with urban Ricardian rents’ to state one of the possible critiques of Piketty’s idea that ‘capital’ matters as part of the possible story of progress and setbacks mentioned above:

“That the true historical drivers of the process are not the rate of profit r and the growth rate of the economy g that Piketty speaks of, but rather (1) the economic destruction of the relative wealth of the old European aristocracy as its landed rents collapsed under the impact of competition from the new regions of European agricultural settlement in the New World; and then (2) the rise of urban landed wealth in the form of location as population growth and economic density have outran transportation technologies, and congestion has become a first-order economic cost. This is the economic historians’ critique. It is one I still have to think about. It is clearly right: one of the things that destroyed European aristocratic oligarchies’ wealth was New World agricultural competition that reduced valuation ratios, just as the landed wealth of the owners of Mayfair, Paris, and Manhattan is a substantial part of the run-up in wealth. But is this a criticism of Piketty’s argument or just a mechanism through which it works? I am not sure…”

A related idea is voiced by Tyler Cowen when he quotes the article ‘Real inequality in Europe since 1500′ by Philip T. Hoffman, David Jacks, Patricia Levin and Peter Lindert: purchasing power of wages did increase, in the nineteenth century, due to lower prices of agricultural products caused by the same development which, according to DeLong, destroyed European aristocratic oligarchies (the first part of Churchill’s autobiography mentions this destruction, too). All these developments were connected to ‘return on capital’ – but had little to do with the productivity of land.

I can shed more light on this. Below, a graph of land rents in two of the three marine clay soil areas of the Netherlands, Zeeland and Friesland. The third area, Groningen, is absent as we do not have such series because the legal construct of ownership was somewhat different (‘beklemrecht’) – rents were not flexible, as in Zeeland and Friesland but totally fixed,which led to a totally different division of the surplus: farmers’ incomes instead of the incomes of land owners were high, until 1877. Without any dismal consequences for agriculture (and possibly even positive consequences, at the time it was often stated that Groningen agriculture was more ‘progressive’ because of this.

Sources: Priester, P. (1998), Geschiedenis van de Zeeuwse landbouw circa 1600-1910 (Wageningen); Knibbe, M. (2006), ‘Lokkich Fryslan. Landpacht, arbeidsloon en landbouwproductiviteit in het Friese kleigebied, 1505-1830‘ (Groningen)’ and (Friesland post 1830) sources cited in this book.
Up to 1877, the rent series show a ‘Ricardian’ increase of rents. Productivity of land might have increased a little in the nineteenth century, but this increase was limited at best and can’t explain the dramatic increase of rents (which, due to a leverage effect caused by the total stickiness of wages up to about 1870, was much larger than the increase of agricultural prices). The difference between Zeeland and Friesland can be explained by the Frisian specialization on dairy. And Brad deLong is right: the decline of rents after 1877 (a clear turning point) was caused by increasing ‘competition from the new regions of European agricultural settlement in the New World’, though the rebound of rents after 1899 gets to little attention in the literature and might explain part of Piketty’s high level of return of capital during the ‘Belle Époque’ (after WWI, rent income declined again, but this time not just because of declining prices of agricultural products but also because wages, which like all other prices had increased quite a bit during the war, declined much less than product prices). While Tyler Cowen is right too: lower prices of agricultural products did increase real wages, though this was only after (in the Netherlands) about 1880. Graph 2 shows per capita imports of food grains which led to a decline of the price level of agricultural products (until 1895).

Sources: Dutch trade statistics; Pilat, D. (1989), Dutch agricultural export performance (1846-1926) (Groningen); CBS (2001), Tweehonderd jaar statistiek in tijdreeksen 1800-1999 (Heerlen/Groningen); Knibbe, M. (2007), ‘De hoofdelijke beschikbaarheid van voedsel en de levensstandaard in Nederland, 1807-1913‘ in: Tijdschrift voor sociale en economische geschiedenis 4-4 pp. 71-107.
As Dutch nominal wages also started to increase after about 1870, it’s clear that the latter half of the seventies and the eighties were a ‘labour bonanza’,which is also shown by the labour share in GDP, which increased from about 38% around 1870 to around 58% in the nineties, to decline again to 48% around 1910 (E. Horlings, J.P. Smits (1997), ‘A comparison of the pattern of growth and structural change in the Netherlands and Belgium, 1800-1913′ in: Jahrbuch für Wirtschaftsgeschichte p. 83-106. First part of this article here, second part here (yes, annoying). The article also contains information about different kinds of capital in Belgium and the Netherlands in this period. Loads of data can also be found in this book.)
The point: the increase of rent income was not caused by changes in the ‘marginal productivity of land’ but by price changes caused by mega-historical developments, in true Ricardian fashion. The (increasingly) absentee landowners in Friesland had not much to do with this – they just reaped their windfall ‘profitability of financial capital’. Even when we take into account that the value of land will have increased too, which means that the return on capital (liability side of the balance sheet) increased less than rent per hectare, available evidence still shows a whopping rate of return until 1877. According to neoclassical thinking, the marginal productivity of capital is equal to additional output ’caused’ by using an additional unit of ‘capital’ (physical capital) as well as to the return on ‘capital’ (financial capital). But this mixes up the asset side and the liability side of the balance sheet. The marginal productivity of capital should be defined as the (value of) additional output caused by an additional use of a physical unit of capital as defined on the asset side of the balance sheet. The rate of return on capital (let’s call this the profitability of financial capital) is however dependent on the surplus (i.e. value added minus wages and mixed income), which in a multi product world shows wild swings independent of additional use of physical capital but is highly dependent on (often erratic) relative prices (for instance: oil prices) and which is also dependent on the law and culture of a society. These changes in relative prices are at the heart of the famous Cambridge controversy, which showed that when capital is not just one homogenous something and relative prices change you can’t aggregate the different entries of the balance sheet into one amorphous kind of jelly. Relative price changes are especially large when it comes to non-reproducible assets, like land, oil, clean air and the like. There was a good reason why classical economists differentiated ‘land’ (i.e. oil) and ‘capital’ (the oil drilling platform). And there was a bad reason why neoclassical economists started to treat these different items into one badly specified and inconsistent aggregate. We should not mix up the asset and the liability side (or different kinds of assets and different kinds of liabilities) and use them in tandem, while even when we aggregate only the asset side care should be taken to gauge the influence of different weighing schemes. There are overriding reasons why balance sheets have two sides and multiple entries. Economists should look at all of these sides and entries as well as at the legal, political and cultural aspects of a society when studying capital, growth, inequality and the division of the spoils. That’s what economics is about.

Addendum: for Germany, ‘Die Ehe der Maria Braun’ is a nice example of a movie about a situation somewhat comparable to the situation at the end of ‘Gone with the wind’ – while checking this I noticed that Fassbinder, director of dozens of movies, was only 37 when he died. Unbelievable. My favorite: Querelle.