More on the structural flaws of the Eurozone
The following note is in response to the call for evidence on ‘”Genuine Economic and Monetary Union” and its implications for the UK’ from the House of Lords’ European Union Committee, made on 24th April 2013.
The main points of the note are
- banks’ risks of future loss, and hence their need for capital to protect depositors, depend heavily on future movements in asset prices, since these asset price movements determine the value of loan collateral,
- ii. a single set of capital rules (such as the Basle III rules now being imposed on advanced country banking systems) cannot be appropriate for all countries at all times, and undermines the flexibility of national regulators to deal with specifically national problems,
- banks’ loan losses in the Eurozone periphery nations are so large that depositors can be repaid in full only by capital injections from the state which affect these nations’ credit rating,
- the interaction between banks’ solvency problems and their host nations’ sovereign default risk has created a vicious downward spiral of fiscal and monetary retrenchment, and losses of output and employment, and
- this downward spiral is much worse than in a traditional monetary jurisdiction because governments in a multi-government monetary union such as the Eurozone cannot borrow from the central bank
In these important respects a multi-government monetary union is highly dysfunctional compared with the one-currency-per-nation pattern found almost universally nowadays, except in Europe.
Asset price inflation and deflation, and banks’ capital requirements: an aspect of the Eurozone banking crisis
Since 2008 the so-called ‘periphery’ of the Eurozone (i.e., Italy, Spain, Ireland, Portugal and Greece, also known as ‘the PIIGS’ and ‘the Club Med’) has faced an unremitting financial crisis, in which the various countries’ banking systems have been the focus of concern. Banks have incurred losses that have been substantial relative to their capital. Regulators, including international regulators from the European Commission and the Bank for International Settlements, have demanded that capital must be rebuilt. With stock markets reluctant to provide new equity capital, governments have had to step in to fill the gap and so have become injectors of capital ‘of last resort’. However, these injections of capital have in turn had to be financed by the state and so have increased governments’ budget deficits. Unfortunately, even aside from the turmoil in financial systems, Eurozone budget deficits during the Great Recession have been well above the levels envisaged in the 1997 Stability and Growth Pact, one of the Eurozone’s defining documents. The burden of bank recapitalization has therefore interacted with a larger sovereign debt crisis in the Eurozone; it has been associated with a large-scale macroeconomic malaise of financial retrenchment, weak demand and high unemployment. That malaise has now lasted over five years and shows no sign of early resolution.
How did significant nations, with large numbers of undoubtedly capable central bank and finance ministry officials, get into a mess of this kind? And why have these nations’ banking institutions, which until the crisis had impressive records of profitable growth, been so badly embarrassed? My argument here will be that the fortunes of a nation’s banking industry are inextricably linked with one aspect of that nation’s macroeconomic fortunes, namely the rate of change in asset prices. More particularly, the prudent level of banks’ capital/asset ratios depends heavily on the prospective rate of change in asset prices.
Banks’ managements have always to take a view on that prospective rate of change, and they may ‘get it wrong’, with disastrous consequences to themselves, their shareholders and the nations in which they operate. But that does not mean they are necessarily to blame, for at least two reasons. First, they do not control the levers of macroeconomic policy, including critically the rate of change in the quantity of money, which determine the rate of change in asset prices. Secondly, compliance with a capital adequacy regime imposed by international bodies may give false confidence in the banking system’s sustainability. I will suggest that bankers have their reasons for basing corporate strategies on asset price expectations arising from past experience, even if past experience is not always a good guide to the future.
Keynes’ remarks on bank solvency and changes in asset price levels
In the depths of the Great Depression, in August 1931 soon after the Credit-Anstalt failure, Keynes wrote an article on “The consequences to the banks of the collapse of money values”. In that article he noted that banks lend against the collateral of ‘a multitude of real assets…which constitute our…wealth, buildings, stocks of commodities, goods in course of manufacture and of transport, and so forth’. In taking deposits from members of the public, they promise to repay in full (at 100 pence in the £, 100 cents in the dollar and so on). But, if those who borrow from the banks cannot themselves repay their loans, the ability of banks to honour this promise is compromised. The promise – or ‘guarantee’ (to use Keynes’ own word) – can then be honoured only from the banks’ own capital resources. In other words, banks’ capital acts as a buffer against expensive defaults in their loan portfolios.
To protect themselves banks rarely lend more than the value of the collateral offered to them by the borrowers. In Keynes’ words,
banks allow beforehand for some measure of fluctuation in the value of…assets, by requiring from the borrower what is conveniently called a “margin”. That is to say, they will only lend him money up to a certain proportion of the value of the asset which is the “security”…The amount will, of course, vary in different cases within wide limits. But for marketable assets a “margin” of 20% to 30% is conventionally considered as adequate, and “margin” of as much as 50% as highly conservative
As long as the value of the collateral varies only a little, the banks are safe. If a borrower for some reason foolishly does not service the debt, the bank can seize the collateral, sell it and cover its deposit liabilities. But suppose that the value of the collateral falls by more than the margin. Again in Keynes’ words, ‘The horrible possibilities to the banks are immediately obvious.’ If borrowers in general cannot service their debts, the banks’ seizure of collateral leads to losses and depositors can be repaid in full only from banks’ capital. Indeed, if banks’ capital is insufficient to meet loan losses, the banks are bust and depositors will not receive back the full amount of their original claim (i.e., 100p in the £, etc.).
Evidently, in extreme conditions banks’ solvency can be undermined or even destroyed by large falls in the value of the assets pledged to them as collateral for loans (i.e., of the ‘buildings, stocks of commodities’ and so on mentioned by Keynes). The ‘horrible possibilities’ latent in such a situation were made actual in Cyprus earlier this year, when quite large banks – such as the Laiki Bank – had to declare insolvency and some depositors received less than 50 cents in the euro. So far banks in other Eurozone member states – including Greece and Portugal – have been able to repay deposits in full, but the debacle in Cyprus shows that this cannot be taken for granted.
Banking in an economy with rising asset values
So far I have been talking about banks in societies with falling asset values. Happily, the normal experience of modern economies is that asset values rise over time. The long-run upward tendency of asset prices reflects two key forces, the persistence of inflation, even if at low rates, and the growth of real output and incomes due in part to the increasing productivity of capital. For example, in the UK house prices in the second quarter of 2013 were 88½ times higher than in the second quarter of 1952, according to data prepared by the Nationwide building society. The compound annual rate of increase in house prices was 7.6%, somewhat above that of national income and average earnings.
It is clear that the solvency of a bank or building society involved in housing finance, and able to take residential property as collateral, benefits from an increase in house prices. If a mortgage bank insists on a maximum loan-to-value ratio of 80%, it does have some risk of loss even if house prices are rising steadily. A mortgagor may fail to service his or her loan in its first year, so that action has to be taken to repossess the property. The legal costs and the estate agency fees when it is resold may come to more than 20% of the house value, and allowance has also to be made for the bank’s internal management expenses and the financing cost of the loan. However, after a few years of 7.6% house price inflation, the loan-to-value ratio falls and the bank’s safety ‘margin’ becomes comfortable. Even with an ‘interest only mortgage’ (i.e., one in which no repayments of principal are made), the loan-to- value ratio drops to under 56% after five years and to under 39% after ten. At such levels of the loan- to-value ratio, it is almost inconceivable that repossession costs could exceed the margin of safety on the loan.
Banks and building societies compete in the housing finance market, with one aim being to attract first-time buyers who have only limited savings and cannot afford a big slice of equity in their first home. Competition for such borrowers is partly a matter of price, but also important is the loan-to- value ratio. The last paragraph explained how, in an environment of 7.6%-a-year house price inflation, an apparently quite risky mortgage with an 80% loan-to-value ratio in its first year improved in quality and after a decade was almost riskless as far as the lenders were concerned. Bank managements may be tempted to raise the loan-to-value ratio on first-time-buyer mortgages to 82%, and then to 85%, and then to 92%, and so on. The longer the period of rising house prices, the greater is the prospect that competition will encourage the industry to offer loan-to-value ratios – perhaps even of over 100% – that , at least superficially, are very dangerous. However, it is clear that – as long as house prices are rising strongly and are expected to continue doing so – even 100%-plus mortgages may in the end be serviced and repaid without difficulty. With a 7.6% compound rate of house price increase, a 110% loan-to-value interest-only mortgage becomes a 53% loan-to-value ratio after a decade.
Banks’ capital requirements in an economy with ever-rising asset values
How much capital do banks need in an economy where
- their assets are almost entirely loans to the private sector,
- the value of the houses, buildings and securities pledged to them as collateral is rising continuously, and
- loan-to-value ratios also leave a comfortable safety ‘margin’ of at least 20% ?
The answer is ‘very little’. Of course, they need some capital to anticipate the occasional delinquent borrower, with the attendant costs of taking the collateral and selling it. But it should be clear that – even if there are a large number of such delinquents – the 20% margin means that loan losses should be negligible. Indeed, several examples could be cited of banking systems operating on extremely low capital cushions after long periods of general asset price appreciation. In the late 1980s Japanese banks had equity-capital-to-asset ratios of about 2%, as the relentless boom in Japan from the 1950s had been accompanied by continuously rising asset prices and negligible loan losses. In 1960 the USA’s money centre banks (such as the New York-based businesses that belonged to the major clearing houses) had an equity-capital-to-assets ratio of 9.0%. Twenty years later, after the prosperity of the 1960s and much of the 1970s, that had dropped to 3.6%.
The conclusion I have just drawn may seem shocking. As I write, international officialdom, with the full backing of the media and interested academic experts, is pressing for banks to maintain ever- larger capital cushions. But I have just said that in a macroeconomic context of rising asset prices, which has in fact been normality since the 1930s, banks can protect their depositors with extremely low capital ratios.
Two further points need to be developed before moving to the next stage in the argument. First, some banks may have boards of directors with a handful of old fogeys who can remember the last big crisis. These stodgy, ultra-safe banks could of course stay aloof from trends in their industry. They may notice that equity capital cushions are becoming exceptionally thin and simply decide that they will keep their ratios close to traditional long-run norms. They might seek to uphold – say – a 6% equity- capital-to-assets ratio rather than the sub 3% equity-capital-to-assets ratio that is emerging among their competitors. The decision to ‘opt out’ of the race may prove correct in the end, if asset prices turn downwards and overleveraged competitors face heavy losses. But, until that happens, a decision by one bank to persevere with a high capital-to-assets ratio is itself risky. It implies either a poor return to equity and hence disgruntled shareholders or an expensive lending policy, which may alienate customers. (Banks’ return on capital is equal to the average return on assets multiplied by the inverse of the capital-to-assets ratio. A bank with a capital-to-assets ratio above the industry average can achieve the same return on capital as the industry as a whole only by having a higher average return on assets, i.e., by charging more for its loans.)
Second, when they determine their corporate strategies, bankers have to take a view on future asset price movements. It is easy for uninvolved commentators to regard the situation as static, with capital requirements geared to expectations of loss arising from the occasional delinquent. But that is not the real world. Banks’ loan portfolios are in constant flux and yet also have an indefinitely long life, since the banks must aim to replace maturing loans with new ones that may extend out 20 or more years into the future. Their solvency, and their ability to make a promise to repay depositors in full, depends on the existence of a safety margin over the value of the collateral lodged with them and an implicit assumption that the value of the collateral will not collapse while the loans are outstanding. One of the most important risks that banks take is to operate in the belief that macroeconomic policy will be conducted with sufficient skill to prevent asset price slumps.
Ireland as an example
I have already cited some examples of banking systems (Japan in the late 1980s, the USA in the early 1980s) that had become lightly capitalized after long periods of asset price gains. I now want to bring the subject closer to current Eurozone concerns by looking at Ireland and, specifically, its housing market. Ireland’s accession to the then European Economic Community in 1973 was accompanied by a drastic opening of trade to the rest of Europe and indeed the whole world. The next 30 to 35 years saw extremely rapid growth of output and employment, as its productivity levels became among the highest in the continent. This 35-year period also experienced massive appreciation in the values of houses, farmland and commercial property, the assets that are standard collateral for bank loans. Ireland’s banks, which had been sleepy and provincial organizations in the 1960s, provided fabulous returns to shareholders in these decades of economic dynamism. Because asset prices were rising almost without interruption, bank suffered little by way of loan losses. Bank managements came to assume that asset prices would keep on increasing in future.
Although Ireland did have one or two years in the 1980s when house prices fell fractionally, the relentless character of the asset price gains is clear if five-year periods are examined. The chart below shows the change in house prices over five-year periods beginning in 1980. (Note that the 1980 value relates to the five years from 1975.) There is not one single negative value before 2009.
I have explained that bank managements must have a view on the future change in asset prices when they form their lending strategies. What in 2006 was the sensible and realistic central house price assumption to be made by bank managements in Ireland? Given their past experience, what number would banks’ loan officers have used for the likely change in house prices in the five years to 2011? Surely the answer would have had to be positive. A one- or two-year reverse like that in the 1980s might be envisaged, but – looking ahead to 2011 – nothing in the collective memory of the Irish financial sector could justify a forecast of a large fall in house prices.
But that is what occurred in practice. The chart above shows the collapse in house prices both in Ireland as a whole and in Dublin by itself from the peak in the second quarter of 2007; it brings out the suddenness and abruptness of the change in housing market conditions from the middle of 2007. In little more than two years house prices nationally fell by about 30% and in Dublin by itself they collapsed by 45%. Keynes warned in his 1931 article about the ‘horrible consequences’ to the banks if asset price movements were much larger than expected. Needless to say, by mid-2009 the entire Irish banking system was insolvent, although the legal formalities and tidying-up continue four years later.
The question now becomes, ‘what level of banking system capital would have been required in spring of 2007 to maintain Irish banks’ solvency in view of the imminent housing market catastrophe?’. The answer is ‘more capital, relative to assets, than has been held by any banking system in an advanced economy since the 19th century or indeed than has ever been demanded by regulators of a previously profitable banking system’. (And the Irish banking system was unquestionably profitable, in the eyes of most observers, in early 2007.) At present international regulators – including regulators under EU auspices working with officials from the BIS and the International Monetary Fund – are visiting the Eurozone periphery countries. They are trying to persuade or cajole national authorities to recapitalize their banking systems to Basle III standards, with the Basle III standards supposed to make bank resilient to shocks. Most important, in the regulators’ judgement, the purpose of the Basle III standards is to ensure that ‘in the next crisis’ governments will not have to recapitalize insolvent institutions. The essence of this regulatory process is standardization on an international norm, in the belief that a particular ‘norm’ is both optimal and non-discriminatory between nations.
My point is that the whole exercise is misconceived. If the Irish banking system had complied with Basle III in mid-2007, if indeed it had over-complied with international norms in the manner of the Vickers Report gold-plating of Basle III now being imposed on the UK banking system, that would not have stopped Irish banks going bust in the two/three years from mid-2007 given that house prices and other asset values were about to crash by 30% – 45%. More generally, the appropriate level of banking system capital, relative to assets, varies hugely from nation to nation, and from period to period, and the imposition of a uniform set of capital rules for all nations is a mistake. Most of the time, when asset prices are rising gently, banks can operate with quite light capital buffers. The occasional recession may inflict quite serious loan losses and wipe out some capital, but a couple of years in which all operating profits are retained (i.e., the shareholders receive no dividends) plus a rights issue should be sufficient to rebuild capital to an adequate level. The only circumstances in which the state should be involved in bank recapitalization (with the ultimate purpose of protecting depositors) is when an asset price crash has had the ‘horrible consequences’ for the banks that Keynes wrote about in 1931.
But that asset price crash ought to be avoidable by the sensible conduct of monetary policy, while the sensible conduct of monetary policy is the task of the central bank, not of the commercial banks. At any rate, these matters are both complex and specific to particular nations. They are specific to particular nations not least because different nations have different trend growth rates of output, different demographics and so on. In my opinion the imposition and enforcement of a single set of international rules is not the way ahead. (Do I need to observe that those nations that do not belong to the BIS [and hence do not have to comply with the wretched Basle III rules] are at present much happier economically than those that do? Isn’t that obvious?)
As I have explained in this note, it is easy to describe one set of circumstances in which banks can – very reasonably – have very low capital-to-asset ratios. It is also to describe another set of circumstances in which banks need to have such immense capital resources that no sane private-sector investor would buy their equity (and when in fact either depositors must lose money or the state must step in). I agree that this second kind of situation should never be allowed to develop. But the job of avoiding this second kind of situation falls, above all, on a nation’s monetary authorities. These authorities should prevent asset price crashes, and of course the preceding asset price booms, by the appropriate conduct of monetary policy (i.e., in my view, by maintaining a more or less constant low rate of growth of the quantity of money, since over the medium term the rates of change of the quantity of money and national wealth [or ‘the total value of all assets’] are related). The appropriate conduct of monetary policy is not the responsibility of individual banks or bankers.
Can nations go bust because their banking systems are insolvent?
I do not have time here to explain why the operation of the Eurozone in its first eight years, from 1999 to mid-2007, was accompanied by such a dangerous and extreme asset price boom in its peripheral nations. Ireland was only one example of the boom and the troubles that came with the subsequent bust. I also do not have space to consider the interaction between the closure of the international inter- bank market in August 2007 and the subsequent financial turmoil. However, there is another feature of the Eurozone that needs comment. This feature is both unique to a multi-government, multi-nation monetary union like the Eurozone, and has made a solution more elusive and problematic.
In a standard monetary jurisdiction – that is, in a single sovereign nation state – a currency is issued by one central bank which is answerable to one legislature and responsive to one government. Further, a number of commercial banks maintain accounts at the central bank in order to complete settlement between themselves, and both the commercial banks and the central bank are concerned that the claims of depositors (who use the banks to carry out payments services) should be honoured in full. The regulatory machinery, including such arrangements as deposit insurance, is specific to the one nation and the one banking system. The central bank is banker to the government, which – in the extreme – can require the central bank to lend it whatever sums it (the government) commands. In other words, the government – let it be repeated, the one government of the one nation – has the power of the printing press. This set of institutions has a weakness. It carries the risk that the government will abuse its power, by printing too much money and causing rapid inflation. But it also has an important strength. Since the government can borrow without limit from the central bank, it can always honour debts expressed in its own currency. Within its own borders, for debts expressed in the local currency, the government can never go bust.
This characteristic of a standard monetary jurisdiction is of great value in an extreme financial crisis of the kind now being experienced in the Eurozone periphery. The discussion has demonstrated that, after a severe asset price crash, banks’ loan losses can exceed their capital a few times over. If the depositors are still to receive 100 pence in the £, 100 cents in the euro or whatever, the state must step in and recapitalize the banks. In a standard monetary jurisdiction that is easy enough, because the government can borrow from the central bank and use the loan proceeds to finance the capital injection. Sure enough, the ratio of public debt to gross domestic product is likely to rise sharply. But, as long as the public finances are otherwise in good order and the debt is easily serviced (i.e., with low interest rates), nothing very dramatic will ensue. The banks can be recapitalised, depositors are repaid in full, the government can service its debt, after a few years the banks can be privatized again, and life goes on as before.
Unfortunately, the Eurozone is not a standard monetary jurisdiction. Under the terms of the Maastricht Treaty governments cannot borrow without limit from the central bank. Indeed, they are supposed not to borrow from the central bank at all. As a result, governments can go bust within their own borders. In most countries the capital stock is worth about five times national income and the banking system’s claims on the national population in local currency rarely exceed 150% of national income. However, in exceptional asset price booms the capital stock may be valued at seven or eight times national income, while the debt of the domestic private sector to the banks may be more than twice GDP. The asset price bust may consequently result in loan losses to the banking system that exceed its capital by 50% – 100% of GDP. Something of this sort does seem to have happened in Ireland, Cyprus and Greece. Spain may be better placed, but it faces the same kind of difficulty. The governments of the Eurozone periphery have therefore to make capital injections into the banking system that are enormous relative to GDP.
But how can the government raise the necessary funds? Because they cannot borrow freely from the central bank, the task of bank recapitalization – the task, in other words, of making sure that depositors are paid back in full – is far more painful than in a standard monetary jurisdiction. In a monetary union like the Eurozone banking insolvency interacts viciously with sovereign default risk. This vicious interaction is not found in the one-currency-per-nation system which is normal in the modern world.
The Eurozone suffers from basic design flaws
If Europe’s political leaderships are intent on keeping the Eurozone in being with its current membership, this paper does not have any easy answers. The nations of the Eurozone periphery have crippled banking systems, with loans losses that are higher than – and indeed sometimes are multiples of – the equity capital they held before the crisis. Officialdom’s response has been to demand that banks rebuild capital, if necessary by capital injections from the state, but the amounts involved are substantial relative to already strained public finances. The Eurozone is so structured that governments can go bust within their own borders. The banking problems therefore aggravate the fears of sovereign debt default that have beset these nations in the last few years.
In my view the nations in the Eurozone periphery could solve their macroeconomic problems more easily, with less damage to output and employment, if
– they did not belong to the Eurozone and their governments could therefore borrow freely from the central bank in order to finance bank recapitalization,
– they had their own currencies and could therefore devalue them to a level necessary to generate an external payments surplus large enough to honour debts to international creditors, and
– they did not belong to the BIS, and had the freedom to set capital and liquidity rules for their banking systems which would again let the banks expand lending to domestic private sectors.
I am well aware that the ruling elites of such countries as Ireland, Portugal, Greece and so on regard membership of the Eurozone, the EU and the BIS as vital badges of national respectability. That is bad luck for the people who have to endure the result, which is prolonged macroeconomic trauma.