Tuesday, 4 September 2012

That shocking UK debt chart

I've been meaning to take this chart apart for ages.

It first appeared towards the end of 2011 and caused a considerable stir. Writers around the world castigated the UK for its profligacy. And from some Americans there was almost palpable glee that the smug UK was in a worse mess than the US. The French were facing a ratings downgrade at the time, and at least one writer used this chart to justify Sarkozy's argument that it was the UK that should be downgraded, not France. Review after review highlighted this "excellent" chart, and went on to blame the UK's dominant world position in financial services for its debt crisis. And it's still doing the rounds. It popped up on Twitter again last night. Which reminded me that it really, really needs debunking.

Not one of the many articles featuring this chart that I have read questioned its accuracy or pointed out that Morgan Stanley had provided neither an explanation of how the chart was constructed nor information on its data sources.  Nor did anyone except Steve Keen notice that there were several other charts around at the same time which showed total debt from all sources, and Morgan Stanley's figure for UK financial sector debt was more than twice anyone else's. This chart from the ONS has private sector debt at 450% of GDP, of which about half is financial sector debt:

There is something else odd about the Morgan Stanley chart too. What on earth is "Europe"?  In this chart, the UK and Sweden are shown separately from "Europe", even though both are members of the European Union, and Norway is also shown separately even though it is part of the continent of Europe. This is to say the least confusing for people who don't understand European geopolitics. Presumably by "Europe" Morgan Stanley mean the Euro area, though that is inaccurate and misleading: Europe is much more than just the Euro area - in fact it is more than the European Union, too.

Now the Euro area has a sovereign debt crisis, does it not? Er, not according to this chart. The public debt of "Europe" is considerably less than Japan's and only slightly more than the US*. So as Joe Wiesenthal points out, "Europe" actually doesn't have a debt problem. Now that would be true if "Europe" were a full monetary and fiscal union like the US. But it isn't. It's a collection of sovereign states which adopted a common currency because they thought it would benefit their own economies, not because they wanted to share with others. It's a collection of sovereign states which have agreed to some common laws and a certain amount of power-sharing, but who jealously guard their right to fiscal self-determination. It's a collection of sovereign states who DO NOT ISSUE COMMON DEBT. There is no equivalent of federal debt in Europe. This chart is effectively comparing the federal debt of the US, UK, Canada etc. with the MUNICIPAL debt of the Eurozone (not even the whole EU). If we included the municipal debts of the US in its total debt pile, the US wouldn't look nearly so good, would it - since we know that several of its states are virtually bankrupt?

Anyway, to return to the UK. According to this chart it is by far the most indebted nation on earth and is crippled by the debt of its bloated financial services sector. Is this justified?

"Partly", I think would have to be the answer. Other sources show that household debt/GDP is indeed very high, at around 100% of GDP. And non-financial corporate debt is also very high at around 110%. Public debt actually isn't as bad, at just under 60% in 2011 according to HM Treasury (Morgan Stanley marks the debt to market, giving a higher figure of about 80% due to rising gilt prices).  But overall, excluding the financial sector, the UK does indeed have something of a debt problem.

We should remember that a high proportion of household debt is mortgages, because the UK has a very high level of home ownership and property is relatively expensive. Mortgages on primary residences are generally regarded as reasonably safe forms of debt, because people will cut discretionary spending to the bone to protect their mortgages. But a lot of people are financially stretched in the UK because of high levels of unsecured debt, rising unemployment and, perhaps more importantly, UNDER-employment (people forced to take part-time work when they want full-time), and high cost of essentials such as fuel and heating. Mortgage interest rates for many are at all-time lows because of the very low interest rate policy and unconventional measures pursued by the Bank of England, and it is probably fair to say that mortgage defaults would be far higher than they are if interest rates were at historically normal levels. Also, the UK has not suffered the catastrophic property market collapses of the US, Ireland and Spain: this is mainly due to housing shortages, but it is also due to capital flight from the Eurozone (London property is a favourite safe haven for the Greek and Italian wealthy) and active support for the housing market from a Government worried about funding for elderly care (the over-50s own much of the higher-value housing stock).

Non-financial corporate debt, although very high, is not quite what it seems. Corporations also have unprecedentedly high cash balances. Exactly why they are carrying so much cash while also maintaining loan finance is a bit of a mystery (although it is sensible to have long-term loan finance coupled with a positive cash balance) but I would say has to do with worries about cash flow and access to working capital finance in the future, since banks are reducing unsecured lending. And that brings me to one of the major issues with this chart.

The financial sector's debt is, to a considerable extent, the savings of the household and corporate sectors, both in the UK and overseas. The size of these figures suggest that they include all forms of debt including domestic deposits (for comparison, the McKinsey chart below excludes domestic deposits from financial sector debt). It therefore shows the assets of the domestic non-financial private sector as the financial sector's debt, and in so doing gives a misleading picture of the UK's domestic finances. On this chart, every personal and corporate current account and deposit account balance is (presumably) included in "financial sector debt". This is also true for the other countries, of course - which should really make one rather worried about New Zealand. It has high household debt but apparently almost no financial sector debt. Do New Zealand companies and households have no liquid assets? And we should also take Japan's figures with a pinch of salt. Japanese households save like crazy, and this is reflected in relatively high financial and public sector debt levels (most of the Japanese government debt pile is owned by its population). But the real issue in the Japanese financial sector is asset quality, not debt: their banks are still hampered by high levels of non-performing corporate loans. The lack of granularity in this chart, and the absence of supporting data (or even attribution of data sources) makes it seriously misleading.

So is the UK's financial sector really as much of a problem as this chart suggests? Is financial sector debt really 600% of GDP? As Morgan Stanley do not give their data sources and the chart is no longer on Haver Analytics' archive list (and their website does not have a search facility) it is impossible to check how they arrived at this figure, but it is way out of line with other sources.  McKinsey, in their paper on the progress of global deleveraging, (link at bottom of post) gave UK financial sector debt as 219% of GDP:


This is consistent with the ONS figures from the chart further up the post, which was quoted by HM Treasury in its 2011 Budget report. And interestingly, McKinsey and Morgan Stanley both say they obtain their data from Haver Analytics. So why is the Morgan Stanley figure so much higher? Including domestic deposits can't possibly be the sole cause.

Grossing-up of interbank balances might be one possible explanation. Financial institutions lend to each other through the interbank unsecured and repo markets, and they also place deposits with and borrow from the central bank. The borrowing and lending of funds on a daily basis is the lifeblood of the financial system, and ensures that at the end of each day the financial system is fully funded overall, even though individual institutions are temporarily in debt to each other. This chart only considers debt. Could it be that interbank borrowings are included in Morgan Stanley's UK financial sector debt figures? If so, it is a gross error. Interbank borrowings should be netted, and only the difference (if negative) shown in debt figures.

But grossing-up interbank balances does not explain why the UK's financial sector is apparently so inflated relative to the rest. My guess is that Morgan Stanley have included in the liabilities of the UK financial sector debts incurred by UK subsidiaries of overseas financial institutions. In particular, they may have included derivatives issued and traded in London.

McKinsey say they do not include asset-backed securities in figures for financial sector debt, because that would cause double counting of the underlying, which is mostly domestic loans of various kinds. But they claim that Morgan Stanley do include them. By far the largest issuers of ABS are US financial institutions. Global Finance say that if ABS were included in figures for US financial sector debt, it would be 350-360% of GDP. Yet on the Morgan Stanley chart, US financial sector debt is shown as about 100% of GDP (as against 40% on McKinsey's chart).

Including in UK financial sector liabilities the debts incurred by UK subsidiaries of overseas financial institutions is perhaps an understandable approach, since these subsidiaries are usually UK-incorporated legal entities, but it creates considerable distortions. Just one example should be sufficient to show the distortionary effect: using this approach, the losses incurred by JP Morgan's Chief Investment Office in London earlier this year would have been counted as losses of the UK financial sector, not the US. And it raises the question who would be responsible if a systemically-important US financial institution suffered sufficient losses on its London trading activities to cause its failure.

In the financial crisis, the US government bailed out AIG even though most of its losses were incurred in London. But Iceland, facing a similar problem, repudiated the debts incurred in London. There was considerable ill-feeling between the UK and Iceland over Iceland's behaviour towards its banks' overseas creditors, and at one point the UK prime minister used anti-terrorism legislation to prevent Icelandic assets being moved out of the UK: relations between the UK and Iceland are still strained, particularly as Iceland now appears to be experiencing economic recovery (unlike the UK). However, the UK did nationalise and dispose of the UK subsidiaries of both Kaupthing and Landesbanki, and compensated insured UK retail depositors with Icesave. So it could be that because of the Icelandic experience, Morgan Stanley have assumed that the UK would in future have to accept financial responsibility for debts incurred by UK subsidiaries of overseas financial institutions, and have therefore included those debts in the figures for the UK financial sector.

I don't think this is remotely reasonable. Iceland's behaviour is not generally considered by the international community to be a good blueprint for the handling of major banking failures. Economically, it was unquestionably the right decision for a small country whose banking sector incurred losses that would, if taken on to the public balance sheet, have overwhelmed it - as Ireland's did. But suppose the US did the same? Suppose it decided, in the event of collapse of the global derivatives bubble, to repudiate all debts incurred by overseas subsidiaries of US financial institutions? That's the kind of behaviour that starts wars.

The Icelandic "solution" should be recognised for what it was - sovereign default - and rejected outright as an acceptable approach to resolution of major financial crises. Instead, there needs to be acceptance by the international community that no one country should have to bear the entire cost of major systemic failure. And an international legal framework should be developed for assigning responsibility and distributing losses fairly. In the 2007/8 financial crisis, most of the losses were borne by the US. But if the US had behaved like Iceland, most of those losses would have been borne by the UK. That's what I think the Morgan Stanley chart is showing us - and it is a terrifying prospect. We must learn from it.

* Eurostat figures show the Euro area combined sovereign debt at 88.5% of Euro area GDP in July 2012.

Related links:

Debt and deleveraging: Uneven progress on the path to growth - McKinsey Global Institute (pdf download available)


  1. 'Europe' is the Euro Region. I think countries and regions are grouped by central banks, not a false view to things.

    Considering the total debt split into private, state and bank debt does make sense in general. A minimum that has to be provided. Anything else is misleading, better said, does provide a very limited perspective.

    Thank you for the McKinsey analysis. It is hard to find good reference data.

    You must be careful. In the future Eurostat will include the debt of those organizations that hold loans instead of the state, counties and communities too.

    All those numbers don't reflect ...
    All this does not hurt a lot. What will hurt is, when people realize - currently in discussion in Germany - that their pensions in 2030 will become reduced to a minimum pension**). In the area of pensions. EURO region is very specific, because of Germany. The German communities are almost broke, the pension for those who worked for the state are twice as much as the regular pension. In the bitter end the 3 'richer regions' will have to subsidize the rest of Germany and parts of Eurozone.

    Example 2800 EUR gross will lead to about 700 EUR, assuming the current relation between active workforce and not active people (pension) still 1400EUR on an avg. not inflated. You need at least 1100 EUR on an avg...

    The only problem Eurozone faces is the distribution of wealth and income. Most critical factor at the moment. Combine every fact from unemployment, demographic decline*), ... and you see the risk on a mid term already.

    *) The young people in the South of the Euro Region and those in East Europe provide the workforce that could help to ease the old-age poverty.

    In the bitter end the foreign debts count.

    Iceland worked for Iceland. Iceland is a very special case - think of the Japanese house wifes ...

  2. Hi Frances,

    as I use this graph in talks I thought I'd run you thoughts by my friend James Meadway the economist at the New Economics Foundation as he sent me it originally, this is what he says in a critique of your text.

    I think it makes one good point - where do these numbers come from? - but mixes it up with a whole load of less convincing economics.

    First, the blunt facts. Even on the alternative figures given the blogpost, UK financial sector debt is still seriously higher than that of other advanced economies. (The points about Europe and the eurozone are just pedantic, sorry.)

    UK national debt (what the government owes), on the other hand, is neither high by international nor historical standards. It's just inept to claim otherwise: it's below (for instance) the Rogoff 90% marker (Rogoff claims, on the basis of past experience, that any national debt over 90% starts to have detrimental effects on previous growth). It only tips over that point if you also throw in the interventions for the financial sector.

    Which is the second point. Debt matters beacuse it is a liability for the debtor - it is something on their balance sheet that is a claim on their future income, since the debt has to be repaid (with interest). On the other side, for the creditor, it is an asset - it is something expected to produce a stream of future income.

    But should the debtor be able to default on the debt, that liability can change. The creditor still wants repaying. But the debtor either cannot or will not do this. Should someone else step in they, instead, will hold the liability for the debt. This is more or less what happened during the 2007-8 crash: the bailouts were us stepping in to take on the liabilities of the banking system. That meant we, instead of the banks, were the ultimate holders of the liability.

    The problem with the analysis here is that I think it understates that element of the crisis. The banks may well have all this debt, and it may well be carefully balanced by (financial) assets on their balance sheets. But should there be another banking crisis - say arriving from the eurozone - we may find that we, not they, are the ones left holding the liabilities. The costs of the debt, then, will transfer to us.

    The blog touches on this (in order to dismiss the point) when it says that the high financial sector debt figure may include the UK-based subsidiaries of overseas financial operations, which it says is unreasonable, since the UK will not bail out what are in effect foreign banks. That's fine. Except that it seems equally reaosnable to assume (unless I'm missing something) that the Morgan Stanley figure includes the overseas subsidiaries of UK banks and financial institutions - which will be enormous, and whose liabilities will, other things being equal, fall onto the UK government in the event of a crisis.

    That's the real risk in the situation - not from government debt (neither particularly high, nor unmanageable) - but from financial system debt functioning (in effect) as additional government liabilities. We are most likely going to have to start writing some of this stuff off over the next few years, as it becomes clear capitalism is not returning to growth any time soon. This will be a messy, unpleasant process (just look at Greece!) in which the political battle will be to try and force the costs away from workers and wider society, and back onto the 1%. Fun times.

    Hope that's useful. Would be interested to see if you dig up anything more on the Morgan Stanley graph. I generally trust Zero Hedge (where I got it from) in these things, but it'd be useful to know how it was put together if it's being questioned.


    1. Noel,

      It is very evident from your comment that neither you nor James Meadway have properly read and understood my post.

      To take your points in order:

      1) I have never disputed that the UK's financial sector debt is higher than any other nation. In fact I produced two other charts both of which showed that it was (although I did note Global Finance's comment regarding the effect of including ABS on the US's financial sector debt). But there is a very considerable difference between 219% of GDP and 600% of GDP. Morgan Stanley's figures are way out of line with everyone else's and I was attempting to explain the discrepancy.

      2) It is not pedantry to point out that a chart is comparing apples and pears, as this one does when it compares US federal debt with Eurozone debt. Nor is it pedantic to point out that "Europe" has more than one definition.

      3) I did not say that UK public debt was high. I said that household and non-financial corporate debt were both high - which is true - and that public debt was actually not as bad. You have misread what I said and unfairly accused me of ineptness. You should retract that statement.

      4) Your lecture on the risks of debt is out of place on a post which is simply questioning the figures on one chart because they are out of line with other charts that purport to show the same thing. And it is patronising, frankly. I know every bit as much about the risks of debt as you and James Meadway do, as you would know if you had read my previous posts - but it is not the point of this post.

      You and James have both overlooked the fact that the other two charts I quote both include overseas liabilities of UK banks, but they are still 400% of GDP lower than the Morgan Stanley chart. That is why I suggest that Morgan Stanley not only includes the overseas liabilities of UK banks, as the other charts do, but ALSO includes the UK liabilities of overseas banks trading in London - which would include rather a lot of the global derivatives bubble.

      I am very disappointed that you and James Meadway were so anxious to defend the chart and prove me wrong that you missed the main conclusion of the post, which is that the global financial system transcends nation states, and international agreement is needed to ensure that no one nation ever again has to bear the full cost of systemic failure. I would have thought you would be in sympathy with this conclusion.

      The provenance of this chart is really quite dubious. I spent a considerable amount of time trying to find the original sources for this chart, but to no avail. The furthest back I was able to go in the provenance chain was a paper from Mercatus which was presented to the US Congress: we know that Zero Hedge lifted the graph from that paper. The Mercatus paper has accompanying Bloomberg screen prints that show the derivation of the public debt figures - that's how I know that the public debt figure in the chart is marked to market. Have you and James never thought to query the fact that it shows UK public debt as 80% when the actual figure for end 2011 is 60%? But even in the Mercatus paper, there is no information on the data sources and no explanation of the 600%. I'm happy to provide a link to the Mercatus paper if you are interested.

      The chart does not seem to be available from Morgan Stanley themselves (I did try), and Haver Analytics provide datasets not final charts. Morgan Stanley created their chart from Haver Analytics data, just as McKinsey did: why the two charts are so different is the mystery I was trying to solve.

      The data sets for the other two charts I have quoted are both available, and both McKinsey and ONS explain how they have constructed their charts. I would strongly suggest that you stop using this chart and use McKinsey or ONS instead. You do your cause no favours by using unsubstantiated and inexplicable charts.

      If James Meadway has comments to make on my blog he should make them himself, not use you as his gofer.

  3. Thanks to everyone for give useful information.

  4. I tend to agree with Frances. Indeed, this is about the only sensible material I've found on the general net before drudging through academic journals (not much use so far). The two Rs have come in for much criticism since they divulged the dubious spreadsheet used in their pro-austerity paper.
    I teach university economics but remain a biologist at heart. I'm appalled at the lack of data available in economics and finance - one can take almost nothing at face value. R&R's famous (infamous?) paper reworked shows growth not negative but 2% positive after 90% debt/GDP.
    The actual figure of financial sector debt is likely less important than its quality, though if quality is lousy we'd prefer the lower figure. I have few ideas on how we can satisfactorily ascertain quality in today's accounting conditions. In small company analysis one can 'visit' the assets and determine the quality (£100K claimed for plant may turn into a liability of £10K to remove the scrap etc.)
    What we need to know is whether the 'debts' are profit making or a collection of Enron/RBS-style accounting frauds waiting to be revealed when the music stops and who will be holding the parcel at that point. I tend to gloom on this on account of the criminality rife in the financial sector, but lack anything like data on which to decide.
    The winding-up of a small sample of assets through real sales would tell us a great deal - but pigs might fly. When even peer reviewed material is deeply flawed and company accounts unreadable until we know claimed sales were repo 105s, the company is actually counter-party to its own claimed hedges and so on - the numbers just can't tell us the real position.

  5. I am curious about the assets held by financial institutions. Does anyone have a breakdown of what they comprise. For example, is a significant proportion of it UK housing mortgage debts?

  6. Sorry to be late for the party... I've always assumed that the Morgan Stanley chart did include interbank lending. While you're right that this would be "incorrect" and most charts would net interbank debt, the Lehmans fallout demonstrated that netting falls apart in the face of counterparty collapse. If you want a chart that shows debt exposure during happy times net the interbank lending, but if you want a chart that shows potential exposure should counterparties start falling...

    I'm just guessing obviously, as we never did get any solid information about the chart, but it would surprise me if risk managers at major institutions weren't wanting to track ALL debt these days, as events since 2007 show that we need complacency in theory like we need a bullet in the head.