Divisia and M-3 estimates of the amount of money: two sides of the same coins (2 graphs, meta-statistic)

Estimates of the total supply of money are important for economic analysis. Economists have however developed different estimates of the total amount of money. Is one particular metric superior to the others, as William Barnett argues? Or should we use them in tandem as they show different aspects of reality? This blogpost argues that the last option is superior.

Monetary aggregates are important. According to monetary expert Charles Goodhart the analysis of monetary aggregates has however been neglected by main stream macro, which led him to write an article titled: ‘Whatever became of the monetary aggregates?’. He states that contrary to received mainstream macro wisdom monetary aggregates do play an important role in the economy while he also states, in true Post-Keynesian/Austrian spirit, that money creating ‘lending’ might be a more interesting monetary aggregate than the total amount of  money as such. Which is right. At present the M-3 ECB money aggregate shows a limited increase of the amount of money – but the quadruple accounting ‘counterparts’ of this growth show that this is despite a decrease of ‘domestic’ lending and because of lending to foreign companies (or governments? The statistics still need a little improvement). Which shows that the Eurozone is more depressed than the ECB data on total M-3 money lead us to believe – even though the bank seems reluctant to act upon this information.

The great advantage of the M-3 aggregate is that it’s based upon a quadruple accounting concept of money. My debt is your asset. Creating money ‘out of thin air’ also creates debt ‘out of thin air’. This process creates new economic entities consisting of a combination of a debtor and a creditor. Your economic reputation is as good as your creditors reputation – as banks discovered back in 2008.

M-3 and its counterparts enable analyses based upon this idea.

But like all metrics, M-3 has drawbacks. M-3 and its lending counterpart consist of different types of lending and money. In the case of money we can think of items like cash, overnight deposits and time deposits. Not all of these have the same velocity, an increase in cash will lead to a different change in the velocity of the money aggregate than an equal increase in time deposits.  A weighted average of different kinds of money will capture the change of the velocity of the aggregate better than just adding them together. Cash will get a relatively higher weight than time deposits. Such an average exists and is called ‘Divisia’ money (the weights will of course change over time). It’s a useful addition to our toolbox and shows, at the moment, low money growth in the USA, contrary to M-2 metric of the Fed!

Source

However. This metric is ‘sold’ to us by ‘Divisia good, M-3 no good’ statements like

“Conventional money-supply measures are not adjusted to account for differences in the degree to which various assets actually serve as money. Divisia measures, named after the early 20th-century French economist, Francois Divisia, make proper adjustments and thereby offer a more accurate picture of what is really happening to the money supply“.

No. Divisia does not offer a more accurate picture. It offers additional insight, which is something else. The conventional measures are based upon quadruple entry accounting, an indispensable analytical tool for any economists. As an example of the use of this tool we can cite Francis Coppola who recently wrote a very clever, intelligent, short and well written blogpost which not only debunked a lot of crap about the present monetary situation of the US of A but which also showed that Mr. Bernanke did use QE to increase the amount money in the non-financial private sector, which enabled this sector to deleverage without overt deflationary consequences.

Source: the graph was used by Francis Coppola but was originally made by Sober Look

Which leaves us with the question why the autor intellectualis of Divisia money, William Barnett, on the otherwise  excellent ‘Center for financial stability‘ site, which among other things hosts a whole load of clear and well researched country reports, tries to debunk conventional estimates again and again.

Barnett of course pushes his own name. But there is more to this. His statements are rooted in economic models which disregard endogenous money creation, hail ‘methodological individualism’ and which never used the power and precision of quadruple entry accounting in the first place. Any in-depth analysis of monetary data will of course show the limits of this approach. Economic actors create money and debt by mutual agreement, based upon economic decisions (which is of course not the same as ‘rational decisions’), decisions which lead to an endogenous increase of money and economic ‘molecules’ instead of individual ‘atoms’ (I really do not understand why so many Austrian economists do not love this idea – money and economic entities as the autonomous consequence of market action). There is nothing heterodox about this, it is clearly spelled out in the manuals on monetary statistics. To state this otherwise: Divisia money is an addition to the toolkit – but it has to be used together with balance sheet data on money and money creation: my borrowing is your lending and creates our money.