The new ECB statistics on financial instability can and should be improved

Summary: the ECB is tinkering with new statistics on financial instability – among other things sectoral balance sheets. But unless these statistics are fitted into accounting or flow-of-funds models of the Eurozone economy and used for a pluralist analysis of the situation in individual countries next to the Eurozone averages  this will be of little avail. Fortunately, flow-of-funds models are already available at the ECB.

In October 2007 the European Central Bank (ECB) published a new series on Eurozone private debt which, when expressed as a % of GDP, showed an exponential increase which (of course, as it was exponential) could not be explained by changes in income and a lower interest rate. This was of course alarming. Unlike somebody like Steve Keen, who panicked and cried wolf after looking at exactly the same data and the same pattern of development for Australia, the ECB re-stated its believe in Robert Lucas style ‘rational expectation’ economics which equate financial stability with low and stable inflation and deregulated capital markets:

“assessing the historical pattern of household loan developments purely on the basis of the macroeconomic determinants of loan demand remains to some extent inconclusive, given that loan developments over the past two decades are also likely to reflect a number of structural influences, such as financial innovation and changes in mortgage market regulation, as well as the shift to a low-inflation and credible monetary policy environment in the euro area in the context of EMU”[1]

They clearly had not read their Georgist economics, which might have opened their eyes to the idea that a combination of easy credit ‘(financial innovation’) with a fast increase in (mortgage)lending is a marked characteristic of the expansion phase of a housing bubble. And Post-Keynesian economics might have opened their eyes to the possibility that a  ‘low-inflation … credible monetary policy environment’ might be the very prerequisite for such a bubble and also be one of the main reasons for an ever higher risk-appetite of banks as well as household. Instead, they sold out to the pied piper from Chicago. And remember – this was published as late as October 2007, when the housing bubbles in Spain, Ireland, the USA and the Baltic states already had popped. They of course should have done better.

Well, people do learn, it seems. Very recently, the ECB published a volume on monetary statistics which explicitly admits that low and stable inflation is  not the same thing as financial stability. While not a new idea (understatement) this is a step ahead for the ECB, surely as they want to gather more metrics on financial instability, for instance to avoid the danger of looking at averages only (hmmm – statistics 101), because of the necessity of including ‘shadow banks’ in the statistics (they have not done that already!?) and because of the importance of balance sheet analysis (hmmm – business economics 101). But is this enough? Are the new statistics good enough and will they show an increase of financial risks? NO, they aren’t. Sorely lacking from the publication is the incorporation of these individual statistics in one of the accounting, flow-of-fund models mentioned by for instance Dirk Bezemer:

“accounting (or flow-of-fund) macroeconomic models helped anticipate the credit crisis and economic recession. Equilibrium models ubiquitous in mainstream policy and research did not”[2]

These accounting models, which by the way are the grid designed to estimate the macro-economy, show, by necessity and in stark contrast to neo-classical ‘DSGE’ models, a complete oversight of financial flows or (in the case of the national accounts) the economy including sectoral balance sheets, while the metrics estimated in these models are ‘model-consistent’, again in stark contrast to neo-classical models which are based upon ‘utility’ but which do not use ‘utility’ to estimate these models. Or, to quote two neo-classical economists who for once seriously looked at the design of neo-classical models and the use of data (in this case monetary aggregates):

“The problem is that the Federal reserve and other Central Banks have not been producing data consistent with neo classical micro-economic theory”[3]

And the fact is: they are right. The metrics used in neo-classical models are not model-consistent – the models are about ‘utility’ and not about the real world, estimated by economic statisticians which give these models, when economists try to estimate them, always and by necessity an ad-hoc, incoherent character (why do you include this variable instead of that one?). The flow-of-fund models however are (when properly estimated) comnplete and consistent with reality and the economic metrics and the model are designed consistent and therewith do not have the incoherent, ad-hoc character of neo-classical ‘DSGE’ models. They do, in the case of the ECB monetary analysis (which uses such models!), already show (on a monthly basis!) the different kinds of loans leading to the creation of different kinds of money and the sectors which borrow or lend this money and they are consistent with monetary balance sheets of households, governments and companies. And unless the new statistics are incorporated in such a model the whole thing will be of little use.

Fortunately, the way ahead is relatively easy, as these models already exist. What has to be done? The ECB volume on  instability metrics spells out, again and again, the importance of balance sheet data. And indeed, chart 1 on p. 44 should be included in the textbooks (I could not copy it, this is a rough version):

Graph 1. Balance sheet assessment of various crises (crosses: source of vulnerability)

		Households	Firms	Financial	Central Banks	Governments
				Institutions
Argentina	1991	x	x	x
Sweden	1991	x	x	x
Mexico	1994		x	x
Argentina	1995			x
Japan	1995	x		x
Thailand	1997		x	x
Korea	1997		x	x
Indonesia	1997		x	x
Maleysia	1997	x	x	x
Russia	1998			x		x
Brazil	1999					x
Turkey	2000			x		x
Argentina	2001			x		x
Uruguay	2002			x		x
USA	2007	x		x
Iceland	2008			x
Ireland	2009	x		x
Greece	2011			x		x

Source: Bin Ibrahim, M. (2012), ‘Future challenges for monetary statistics in a changing environment – rethinking monetary analysis’in: ECB, Central bank statistics as a servant of two separate mandates – price stability and mitigation of systemic risk (Frankfurt).

Fitting these balance sheets into an accounting model is perfectly possible and will show the interconnectedness between for instance the household balance sheet and the balance sheet of the banks, for instance in the case when household lending increases. These data are already included in the modern national accounts and can be added to the present flow-of-funds monetary analysis of the ECB. Every month, the ECB already publishes flow-of-funds data on money creation, which show that ‘loans create deposits’ and which spell out which loans and which money are created (government lending, mortgages, consumer loans, business loans, M-1, M-2, M-3 and the like).[4] It does this however only for the entire Eurozone – an average! And the ECB volume clearly spells out the danger of only looking at such averages, i.e the ECB should supplement the ‘average’ analysis with at least an analysis at the national level (which would have shown that the high increase of mortgage lending in the summer of 2006 (+12% for the entire Eurozone!) was caused by a low increase in Germany – but increases of up to 30 and 40% in countries like Spain and Ireland). Add this to the 2006 bank lending survey which indicated that credit became more ‘easy’ as well as some Georgist and Post-Keynesian insights – and presto! The extreme dangers of the Irish and Spanish booms would have been identified.

The ECB can do better – and it should.

---

[1] ECB,  ‘Box 1. New Euro area series on MFI loans to households and non-financial corporations’ in: ECB monthly bulletin October 2007 pp. 17-19 (to see the exponential growth add the series on household and company debt); ECB, ‘Long-term developments of MFI loans to households in the Euro area: main patterns and determinants’ in: ECB monthly bulletin October 2007 pp. 67-84.

[2] Bezemer, D. (2009), ”No One Saw This Coming”: Understanding Financial Crisis Through Accounting Models”, MPRA Paper No. 15892, http://mpra.ub.uni,muenchen.de/15892/1/MPRA_paper_15892.pdf.

[3] Serletis, A. (2012), ‘Foreword. Macro-economics as a science’ in,  Barnett, W.A.’(2012), Getting it wrong. How faulty monetary statistics undermine the Fed, the financial system and the economy pp. , XXI. MIT

[4] See for instance (august 2006, more or less the height of the Spanish and Irish bubbles): http://www.ecb.int/pub/pdf/mobu/mb200608en.pdf

About these ads