INNOCENCE LOST: ALCHEMY AND BANKING - The End of Alchemy: Money, Banking, and the Future of the Global Economy - Mervyn King

The End of Alchemy: Money, Banking, and the Future of the Global Economy - Mervyn King (2017)


‘Was ist ein Einbruch in eine Bank gegen die Gründung einer Bank?’ (What is robbing a bank compared with founding a bank?)

Bertolt Brecht, Die Dreigroschenoper (The Threepenny Opera), Act 3 Scene 3 (1928)

‘“Splendid financiering” is not legitimate banking, and “splendid financiers” in banking are generally either humbugs or rascals.’

Hugh McCulloch, Comptroller of the Currency, December 1863

In his account of the origins of the First World War, the historian Max Hastings quotes an exchange that took place in 1910 between a student and the commandant of the British Army staff college. Surely, the student suggested, only ‘inconceivable stupidity on the part of statesmen’ could precipitate a general European war. Brigadier-General Henry Wilson replied, ‘Inconceivable stupidity is just what you’re going to get.’1

Many would argue that ‘inconceivable stupidity’ was also the cause of the recent financial crisis - bankers were wicked and central bankers incompetent. Of course, the extraordinary events surrounding the start and evolution of the crisis took many by surprise, as did those that initiated the First World War. But the crisis of 2007-9 is not a story about individuals, entertaining though that portrayal might be. To see Dick Fuld, CEO of Lehman Brothers, or Fred Goodwin, CEO of the Royal Bank of Scotland, as the villains who undermined our financial system is a caricature that may satisfy our desire to name and blame - if difficult in practice to shame - but fails to get to grips with why so many people, so many banks, and so many countries took decisions that, with the benefit of hindsight, appear mistaken. Rather, it is a story about deeper forces shaping the constraints that governed the actions of individuals.

No doubt there were bankers who were indeed wicked and central bankers who were incompetent, though the vast majority of both whom I met during the crisis were neither. It would be both arrogant and complacent to assume that all the problems generated in money and banking arose because our predecessors, let alone our contemporaries, fell prey to ‘inconceivable stupidity’. Rather, like everyone else, they naturally responded to the incentives they faced. As individuals, they tried to behave in what they saw as a rational manner, but the collective outcome was disastrous. Because they could not affect the behaviour of others, all the key actors in the drama were understandably acting in their own self-interest - given the actions of everyone else. Since, for one reason or another, they could not cooperate with the other players, they all ended up worse off - an example of the prisoner’s dilemma.

Banks, too, faced a prisoner’s dilemma. If, before the crisis, they had exited the riskier types of lending, stopped buying complex derivative instruments and reduced their leverage, they would, in the short term, have earned lower profits than their competitors. The chief executive would likely have lost his job, and other staff defected to banks willing to take risks and pay higher bonuses, well before the wisdom of the new strategy had become evident. Even understanding the risks, it was safer to follow the crowd. The most famous - now infamous - statement of this dilemma was that of Chuck Prince, the CEO of Citigroup (at the time the biggest bank in the world), who said before the crisis, ‘As long as the music is playing, you’ve got to get up and dance. We’re still dancing.’2 He kept on dancing until the music did eventually stop in November 2007, and at that point he lost his job. Had he stopped dancing to the music of risk four years earlier, when he became CEO, I doubt that he would have survived that long. So it is less ‘inconceivable stupidity’ than the inherent difficulty of finding a way to a cooperative solution that was the main challenge facing bankers and other economic actors before the crisis. After the event it may seem easy to see how the crisis could have been avoided by some set of actions, but no one at the time had any incentive to take them.

During the crisis, I found that the study of earlier periods was more illuminating than any amount of econometric modelling (that is, the application of empirical statistical methods to economic data).3 The crisis did not exactly mirror any one episode in the past, but there were uncanny parallels with many such episodes. It was natural for American commentators to look back to the Great Depression of the early 1930s - the previous major episode of bank failures in North America - and Ben Bernanke, a long-time scholar of that period, brought to discussions among central bank governors a valuable historical perspective on the problems we faced. The collapse of the banking system in the United States in the 1930s was so severe that President Franklin Roosevelt, only a week after his inauguration in March 1933, announced a bank holiday (which lasted a week), shutting the banks to provide a breathing space during which confidence could be restored.4 In his first fireside chat on the wireless, he talked about ‘a bad banking situation’, and continued:

Some of our bankers have shown themselves either incompetent or dishonest in their handling of the people’s funds. This was of course not true in the vast majority of our banks but it was true in enough of them to shock the people of the United States for a time into a sense of insecurity and to put them in a frame of mind where they did not differentiate but seemed to assume that an act of a comparative few had tainted them all.5

In 2008 we too had a ‘bad banking situation’, and since then evidence of the repeated scandals of ‘a comparative few’ have indeed ‘tainted them all’.

Banks are part of our daily life. Most of us use them regularly, either to obtain cash, pay bills or take out loans. But banks are also dangerous. They are at the heart of the alchemy of our financial system. As described in Chapter 2, banks are the main source of money creation. They create deposits as a by-product of making loans to risky borrowers. Those deposits are used as money. Banks are able to transform short-term liabilities into long-term assets; in essence, they borrow short and lend long. They are, therefore, vulnerable to any crisis of confidence, real or imagined. The failure in 2008 of several large international banks, threatening not only to bring down the international financial system but to lead us back to another Great Depression, is hardly an isolated incident. Banking crises have been frequent down the years, occurring almost once a decade in Britain and the United States in the nineteenth and twentieth centuries. In one of the earliest crises, in Britain in 1825, many of the patterns apparent in the recent crisis could be seen almost two hundred years earlier: rapid expansion in overseas lending, a stock market boom, a central bank trying to restore price stability, a collapse of the banking system, concern about the interconnectedness of banks, the absence of effective regulation, official purchases of government bonds and drastic intervention by the central bank - the Bank of England - to lend to banks in distress. A familiar story

Banks are also highly visible. In New York, London and any major city, they dominate the skyline. Even on the other side of the world, far from the epicentre of the financial crisis, the names of the tallest buildings you can see from Auckland Harbour in New Zealand are dominated by global financial institutions: HSBC, Citi, Rabobank, ANZ, Westpac, AIG, Zurich, PWC, Deloitte, Ernst & Young. And the historic headquarters of English cricket at Lord’s in London is now sponsored by the American firm J.P. Morgan. Banks are visible not only in our physical but also in our financial architecture. They are huge players in the global economy. The biggest bank in the world is ICBC from China. In an age where inequality is prominent, it is remarkable how equal global banks are in terms of size. Among the twenty biggest banks in the world, the ratio of the assets of the largest to the smallest is little more than two to one.6

Taken together, however, these twenty banks accounted for assets of $42 trillion in 2014, compared with world GDP of around $80 trillion, and for almost 40 per cent of total worldwide bank assets.7 In the league table of countries that are home to the top banks in the world, China, France and the USA share the leading position with four each, followed by Japan and the UK with three each, and one each in Germany and Spain. Of the largest fifty banks in the world, ten are from China and the rest are in the major industrialised countries.8

Size isn’t everything, although it helps. As Macheath pointed out in Brecht’s Threepenny Opera, why rob a bank and risk imprisonment when you could start a bank and create money? Both robbers and founders are attracted to banks because that is usually where the money is. But in September 2008 the money wasn’t there. Banks were losing money hand over fist as people who had been willing to lend them large sums suddenly refused to make new loans. Unlike the frequent bank runs in the nineteenth century, when individual depositors occasionally panicked and withdrew their funds, the change of heart had come from other financial institutions, wholesale investors such as money market and hedge funds. Unable to replace these sources of funding, banks had to call in loans, sell assets, and, in some cases, seek funding from their central bank. Before the crisis, banks could borrow at the finest interest rates. During the crisis, the rates at which banks could borrow rose very sharply and in some cases funds were not available at any price - a case of innocence lost. Almost all the largest banks in the world received direct or indirect support (‘bailouts’) from their governments.

Much of the expansion of the banking system described in Chapter 1 resulted from the explosion of so-called derivative instruments (explained more fully in Chapter 4), based on the application of mathematics to finance. It is difficult not to be reminded of the Titanic - a marvel of modern engineering whose failure and sinking was unthinkable. But when the Titanic sank, at least only one ship was lost and not an entire fleet. In 2008 the whole banking system was at risk.

There has always been a fine dividing line between the respectable activities of traditional banking, managing the deposits of and lending prudently to established customers, and the ‘splendid financiering’ of those who cross the boundary into more dubious practices. While the chief executives of the world’s largest banks were reaching for the sky - literally, in their suites at the top of the new skyscrapers that housed their headquarters - some of their employees were at the same time plumbing the depths of banking by rigging the markets in which they traded. The modern exponents of ‘splendid financiering’ - investment banks - have been described as, on the one hand, inventing new financial instruments that are ‘socially useless’ and, on the other, of ‘doing God’s work’.9 With their global reach, their receipt of bailouts from taxpayers, and involvement in seemingly never-ending scandals, it is hardly surprising that banks are unpopular. Some bankers of my acquaintance are reluctant to admit their profession in polite company - certainly innocence lost. As the quotations at the head of the chapter illustrate, ’twas ever thus.

In 1873, the English political economist and editor of The Economist, Walter Bagehot, published a book that was to become a bible for those interested in how to handle financial crises. Lombard Street, a highly successful account of banking and the money market, explained how the Bank of England could and should have prevented earlier banking collapses by the provision of temporary financial support until the crisis had passed. As Bagehot knew only too well, banking crises are endemic to a market economy. Although his description of a central bank’s responsibility as a ‘lender of last resort’ has entered the textbooks, and was frequently cited as justification for their lending by central bankers during 2008-9, it is in need of updating. Banking has changed almost out of recognition since Bagehot’s time.

Changes in banking since Bagehot

For almost a century after Bagehot wrote his classic work, the size of the banking sector, relative to GDP, was broadly stable. But, over the past fifty years, bank balance sheets have grown so fast that the size and concentration of banks, and the risks they undertake, would be unrecognisable to him. J.P. Morgan today accounts for almost the same proportion of US banking as all of the top ten banks put together in 1960.10

The size of the US banking industry measured in terms of total assets has grown from around 20 per cent of annual GDP one hundred years ago to around 100 per cent today. In the UK, a medium-sized economy with a large international banking centre, the expansion is even more marked, from around 50 per cent of GDP to over 500 per cent. Most of this expansion has taken place over the past thirty years, and has been accompanied by increasing concentration: the largest institutions have expanded the most. Today, the assets of the top ten banks in the US amount to over 60 per cent of GDP, six times larger than the top ten fifty years ago. In the UK, the asset holdings of the top ten banks amount to over 450 per cent of GDP, with Barclays and HSBC both having assets in excess of UK GDP.

While banks’ balance sheets have exploded, so have the risks associated with them. Bagehot would have been used to banks with leverage ratios (total assets to equity capital) of around six to one.11 In the years immediately prior to the crisis, leverage in the banking system of the industrialised world increased to astronomical levels.12

How did banks find themselves in such a precarious position? Without a banking system our economy would grind to a halt with people unable to receive wages and salaries, pay bills, service loans and make other transactions. Banking is at the heart of the ‘payments system’. As with the supply of electricity, its importance to the economy is much greater than is reflected in the number of its employees or its contribution to GDP. Because of its critical role in the infrastructure of the economy, markets correctly believed that no government could let a bank fail, since that would cause immense disruption to everyone’s ability to make and receive payments. Creditors were willing to lend to banks at lower interest rates than would otherwise have been on offer because they were confident - correctly as it turned out - that even if things went wrong taxpayers would see them right.

When all the functions of the financial system are so closely interconnected, any problems that arise can end up playing havoc with services vital to the operation of the economy - the payments system, the role of money and the provision of working capital to industry. If such functions are materially threatened, governments will never be able to sit idly by. Institutions supplying those services are quite simply too important to fail.13 Everyone knows it. So highly risky banking institutions enjoy implicit public sector support. In turn, the implicit subsidy incentivises banks to take on yet more risk. In good times, banks took the benefits for their employees and shareholders, while in bad times the taxpayer bore the costs. For the banks, it was a case of heads I win, tails you - the taxpayer - lose. Greater risk begets greater size, greater importance to the functioning of the economy, higher implicit public subsidies, and yet larger incentives to take risk - described by Martin Wolf of the Financial Times as the ‘financial doomsday machine’.14 All banks, and large ones in particular, benefited from an implicit taxpayer guarantee, enabling them to borrow cheaply to finance their lending.

Although the implicit subsidy was not new, banks were able to exploit its existence to borrow more, and resorted to the use of ever more short-term finance from institutions (known as wholesale funding) in addition to deposits from individual or business customers. The average maturity - the length of life of a loan - of wholesale funding issued by banks declined by two-thirds in the UK and by around three-quarters in the US over the thirty years prior to the crisis - at the same time as reliance on such funding increased. As a result, banks ran a higher degree of mismatch than in the past between the long-dated maturities of their assets, such as loans to companies and households, and the short-term maturities of the funding to finance their lending, often from hedge funds.

Cheap funding fuelled lending. And as the banks got bigger, so did the implicit subsidy - the IMF estimated that in 2011 the cost of the implicit subsidy in the major economies was of the order of $200-300 billion. The bigger banks became, the more they were seen as too important to fail, and the surer markets became that the taxpayer would bail them out. But there are only so many good loans and investments to be made. Banks made increasingly risky investments. To make matters worse, they started making huge bets with each other on whether loans that had already been made would be repaid. As some of those loans went bad, the bets generated large losses. To cap it all, banks held only small quantities of liquid assets on their balance sheets, so they were utterly exposed if some of the short-term funding dried up. In less than fifty years, the share of highly liquid assets held by UK banks declined from around a third of their assets to less than 2 per cent.15 In the US the share had fallen to below 1 per cent just before the crisis.

The turning point came when the size of the balance sheet of the financial sector became divorced from the activities of households and companies. During the 1980s, a combination of deregulation and the invention of derivatives had two effects. First, it separated the scale of bank balance sheets from the scale of the real household and business activities in the economy. Lending to companies is limited by the amount they wish to borrow. But there is no corresponding limit on the size of transactions in derivative financial instruments, and most buying and selling of derivatives is carried out by large banks and hedge funds. The gross market value of derivatives has fallen since the crisis, but at just over $20 trillion at the end of 2014 it was still around one-half of the assets of the largest twenty banks in the world (and of the same order of magnitude as total lending to households and businesses).16

Second, the focus of much of banking changed from making loans, which requires a careful local assessment of potential borrowers, to trading securities, which involves a centralised operation to make and monitor transactions. The rise of US investment banks played a large part in that development. Commercial and investment banking, separated after the Glass-Steagall Act, were merged in large conglomerate banks after the repeal of Glass-Steagall in 1999 by the Gramm-Leach-Bliley Act. Standalone investment banks that were previously organised as partnerships turned themselves into limited liability companies - Morgan Stanley in 1975, Lehmans in 1982 and Goldman Sachs in 1999. Even traditional mutual building societies in Britain (the UK equivalent of US savings and loan associations) decided to convert into limited companies in order to be free to engage in a wider range of banking activities, and every single one that did so from 1990 onwards failed in the crisis.17

Those two developments altered the business model and the culture of our largest banks. Size became an objective because a bank that was clearly too important and too big to fail was able to borrow more cheaply, and even a small advantage in funding costs meant that it could offer cheaper loans to its customers. That enabled such a bank to expand more rapidly than its rivals in a virtuous circle of growth. So the size of the financial sector grew and grew. Along the way it included a massive expansion of trading in new and complex financial instruments, covering activities such as sub-prime mortgage lending. When I visited New York in December 2003, I found all the major banks worrying about whether they should either emulate Citigroup’s strategy of using its size to obtain a comparative advantage in funding costs or abandon the aim of global reach and try to become a niche player in particular markets. Inevitably, perhaps, when the crisis came it was Citi that required a large bailout. Although it fell from grace in dramatic fashion, the fact that it wasn’t allowed to fail vindicated the original strategy.

Before the crisis, banks paid large salaries and bonuses to people who created and analysed new products that could be sold to their clients. But not enough resources were devoted to assessing the riskiness of the balance sheet as a whole. With a growing proportion of bank activity deriving from the trading of complex instruments, it was difficult to work out how big the risks actually were. Certainly, no investor could have done so from the information made available by the banks in their accounts or any other published source. Nor did the banks themselves seem to understand the risks they were taking. And if that was the case, there was not much hope that the regulators, relying on information provided by the banks, could get to grips with the potential scale of the risk.

The expansion of trading rather than traditional lending also altered the culture of banking. New ways of making money relied on recruiting extremely clever individuals - mathematicians, physicists and engineers - whose job was to invent and price new financial instruments, and who were then lost to their former professions. Economists, who love being clever, applauded the application of mathematics to finance and the resulting explosion of transactions in derivatives. Sadly, the growth of trading led also to an erosion of ethical standards. There was a view that being very clever was a justification for making money out of people who were less clever.18 This attitude encouraged the arrogance of the traders who rigged and fixed prices in what were thought to be competitive markets. Almost all the major banks have been dragged into one or more misconduct scandals. Whether selling oversized mortgages to poor people in the US, selling inappropriate pension and other financial products to millions of people in the UK, rigging foreign exchange and other markets in which banks’ own traders were operating around the world, failing to stop overseas subsidiaries in places as far apart as Mexico and Switzerland from engaging in money laundering and tax evasion, there seemed no end to the revelations about what banks had been doing. By 2015, the total fines imposed on banks worldwide since the banking crisis ended in 2009 amounted to around $300 billion - a staggering figure.19

Bad behaviour by talented young men is often associated with the sporting world. But it applies equally to the world of finance, and is not a recent phenomenon. As long ago as 1905, the young John Maynard Keynes wrote to his friend Lytton Strachey: ‘I want to manage a railway or organise a Trust, or at least swindle the investing public; it is so easy and fascinating to master the principles of these things.’20 It is a small step from the expression of superiority in a private letter to the public expression of contempt seen on the trading floor. Keynes was, however, badly burned twice by misjudgements in the timing of his buying and selling activities. Towards the end of his investing career he mellowed and put less faith in speculative investments, writing: ‘As time goes on, I get more and more convinced that the right method in investment is to put fairly large sums into enterprises which one thinks one knows something about.’21 Perhaps the enormous losses banks incurred in the crisis, and the fines levied by regulators around the world, will bring a similar change of heart in banking.

Banks and other financial intermediaries create wealth by providing valuable services to their customers. But there is always the risk that they create the illusion of wealth - in the extreme case of the fraudster Bernie Madoff and his funds there was quite a long period when the perception of wealth was substantially higher than the reality.22 More generally, many of the substantial bonuses that were paid as a result of trading in derivatives reflected not profits earned in the past year but the capitalised value of a stream of profits projected years into the future. Such accounting proved more destructive than creative.

It is clear that the size, nature and culture of banking have changed considerably since the days of Walter Bagehot. To analyse the implications of those changes we need to step back and ask why banks are special.

What is a bank?

What is a bank? The answer may seem obvious. It was actor and comedian Bob Hope who said that a bank is ‘a place that will lend you money if you can prove that you don’t need it’.23 But most borrowers from banks do need the money. A growing economy requires new investment to add to the stock of capital (plant and machinery, inventories and buildings, including offices and housing) which, when harnessed to the labour force and supply of raw materials, produces the output that we call GDP. And that investment is financed from the supply of savings from both home and foreign savers.

Savings can be transferred to investors, whether businesses or households, in several ways. Businesses can retain profits to reinvest rather than distributing them as dividends; they can sell new issues of shares or bonds directly to savers; or they can borrow from banks. Equity, bond and bank finance are the essential building blocks of the methods companies use to finance themselves - although they can be combined in complex ways. Several factors influence the form in which savings are transformed into investment, including the tax treatment of different forms of saving and the willingness of savers to take risk. But perhaps the most important factor concerns the difficulty of assessing and monitoring the potential and actual profitability of investment projects. Equity finance (whether the sale of new shares or the retention of profits by companies) requires careful and continuous monitoring of a company’s activities. In contrast, one attraction of loan (whether bond or bank) as opposed to equity finance is that monitoring is required only when the borrower fails to make a scheduled payment.

Companies large enough to be quoted on a stock exchange produce annual accounts showing profits, as well as assets and liabilities, which are verified by independent auditors and scrutinised by an army of analysts.24 Savers who do not wish to rely on such public information can choose to invest in an array of financial intermediaries, such as mutual, pension and hedge funds or insurance companies. In that case, savers are relying on the judgement of the managers of the intermediary. Perhaps the most famous example of such a manager is Warren Buffett, whose company Berkshire Hathaway has an enviable track record of purchasing businesses that are well run and highly profitable. People who bought shares in Berkshire Hathaway, and held onto them, have seen the value of their investments steadily rise.25 Someone who invested $1000 in Berkshire Hathaway in 1985 would by the middle of 2015 have an investment worth $161,000, a compound annual rate of return of almost 17 per cent. By relying on the judgement of Buffett and his colleagues, such investors did not themselves need to monitor the businesses in which they were investing. Over the past century, financial intermediaries have grown enormously as the wealth of the middle classes has increased. Substantial amounts of wealth are now invested through pension funds, and even house purchases can be financed by sharing the equity in the home with an intermediary.

Banks are a particular type of financial intermediary which provide loan finance to businesses and households. They are particularly well placed to monitor their borrowers’ cash flows because they can observe the movements into and out of the bank accounts of the people to whom they lend. In most jurisdictions, the legal definition of a bank is an institution that takes deposits from households and businesses. A bank can raise finance by taking in or creating deposits, issuing other debt instruments and increasing equity invested in the bank. In the days of Walter Bagehot, banks’ assets mainly comprised loans and holdings of government debt and their liabilities consisted of deposits and shareholders’ equity. Banks today have more complex balance sheets, although the similarities remain. Consider a simplified version of the balance sheet of Bank of America at the end of 2014. Its assets comprised $139 billion of cash (including reserves held at the Federal Reserve), $867 billion of loans to businesses and households, and $1099 billion of financial investments, together with the value of buildings and other real assets. Total assets were $2105 billion.26 Liabilities comprised $1119 billion of deposits, $743 billion in other forms of borrowing and $243 billion of shareholders’ equity capital, giving total liabilities of $2105 billion. In fact, shareholders’ equity is simply what is left after all other liabilities have been deducted from the value of total assets.

Two features of the balance sheet are striking. First, loans comprise only around 40 per cent of the bank’s assets. Second, although deposits clearly exceed loans, other types of borrowing provide a significant source of finance. There is clearly more to Bank of America than taking in deposits from households and businesses and making loans.

What is it that makes banks special? The distinguishing feature of a bank is that its assets are mostly long-term, illiquid and risky, whereas its liabilities are short-term, liquid and perceived as safe. Returns on risky long-term assets are normally much higher than the returns which the bank has to offer on its safe short-term liabilities. So banking is highly profitable. Unfortunately, the notion that a bank can offer safe returns on deposits that can be withdrawn at a moment’s notice by using them to finance long-term illiquid risky investments is, as common sense would suggest, generally false. The transformation of short-term liabilities into long-term assets - borrowing short to lend long - is known as maturity transformation. And the creation of deposits, which are regarded by the depositors as safe, into loans which, by their nature, are inherently risky constitutes risk transformation. Banks combine maturity and risk transformation. This is what makes them special. Moreover, when a bank makes a loan, it creates a deposit of equal value in the account of the borrower. That deposit can be withdrawn on demand and used to make payments. It is money. As explained in Chapter 2, most money today comprises bank deposits.

The transmutation of bank deposits - money - with a safe value into illiquid risky investments is the alchemy of money and banking. Despite innumerable banking crises, belief in the alchemy persists. Economists have shown great ingenuity in coming up with explanations of how the alchemy of money and banking works, and in suggesting some special synergy between bank assets and liabilities. The idea that risk transformation might be possible is based on the view that regional or national banks can diversify their lending across a large number of loans on which the risks are uncorrelated - that is, if a particular loan goes sour, it does not do so in circumstances when other loans also perform poorly. By making a large number of such loans, the bank supposedly avoids any real risk to the loan portfolio as a whole.

Unfortunately, experience suggests that the world does not behave in this way. In normal times, it may be true that the profits on loans to some businesses and households can be used to pay for defaults by others. When there are large swings in the economy, however, then the fates of different businesses move together. Such risk makes the value of the assets of a bank uncertain and prone to volatility. In the post-war period, most of the banking system’s large losses have occurred during times of recession and often on property loans that have gone sour.

People believed in alchemy because, so it was argued, depositors would never all choose to withdraw their money at the same time. If depositors’ requirements to make payments or obtain liquidity were, when averaged over a large number of depositors, a predictable flow, then deposits could provide a reliable source of long-term funding. But if a sizeable group of depositors were to withdraw funds at the same time, the bank would be forced either to demand immediate repayment of the loans it had made, resulting in a fire sale of the borrowers’ assets that might realise less than the amount owed to the bank - or to default on the claims of depositors.

Such a system is fragile for two reasons. First, most banks, especially in recent years, have financed themselves with rather small amounts of shareholders’ equity. The result is that even a small item of bad news about the value of a bank’s assets might put the value of deposits at risk. Concerns about the size of potential losses would leave depositors scurrying for the door. Second, even without bad news concerning the loan book, if any group of depositors started to withdraw their money it would be rational for other depositors to join the queue and get their money out before the bank had to start calling in loans, with the attendant risk of a loss to the depositors at the end of the queue. A run on the bank is self-fulfilling. And a run on its liquidity reserves can bring a bank down in short order.27

A bank run is likely either if there is a loss of confidence in the value of the bank’s assets or an unusually high demand for liquidity. In the first case, the run may be specific to a particular bank if, for example, there was a suggestion of fraudulent activity by its managers. In the second case, a change in perception of the value of bank assets more widely might lead to a suspicion of all banks and lead to a general increase in demand for liquidity. Only an increase in overall liquidity of the system will meet this demand. An individual bank will find it difficult to obtain more liquidity when all banks are scrambling for cash. In the extreme case where depositors flee the bank in large numbers, the resulting run on the bank can bring down not only that particular bank but create a sense of panic that spreads to other banks, leading to a collapse of the banking system. It was to avoid just such a panic that central banks lent hundreds of billions of dollars to commercial banks in 2008.

The basic problem with the alchemy of the banking system is that it is irrational for one person to place trust in a bank if others do not. And it is rational to be concerned about whether a bank can make good its promise to return a deposit on demand when radical uncertainty means that it is impossible to know how big might be the losses on loans. Four ways have been suggested to deal with this problem.

First, banks might suspend withdrawal of deposits in the face of a potential bank run. Banks could just shut their doors for a few days until the crisis has subsided. Of course, for a bank to close its doors, even if only because of a temporary shortage of liquidity, would send a signal that might lead to a loss of confidence on the part of depositors. It might only encourage a run to start sooner than would otherwise be the case, and if suspensions were regarded as potentially likely, then bank deposits would no longer function effectively as money.

Second, governments could guarantee bank deposits to remove the incentive to join a bank run. Deposit insurance is now a common feature of most advanced economies’ banking systems. Nevertheless, the insurance provided is not fully comprehensive, and on occasions that has created difficulties. In 2007, when the UK bank Northern Rock failed, individual depositors joined a bank run because their deposits were insured only up to a limit and not for their full value. The run stopped only when the UK government belatedly announced a taxpayer guarantee of the deposits. In 2008, financial institutions withdrew their funds from US banks or their new equivalents, known as ‘shadow’ banks, because those deposits were not insured. The Federal Reserve stepped in to provide a degree of guarantee to stop the run. The provision of deposit insurance is a subsidy and an incentive to risk-taking by banks. It is no panacea, although, as I will argue in Chapter 7, in a world of radical uncertainty some government insurance is probably inevitable.

Third, the benefits of limited liability - the system we take for granted today in which shareholders’ losses are limited to their initial investment - could be revoked for the shareholders of banks. In modern times, limited liability originated in the nineteenth century as companies sought to find capital on a scale adequate to finance the expansion of railways and new industrial plants. In the case of banks, unlimited liability would mean that shareholders were responsible for all losses, providing assurance to depositors and making a bank run less likely. In the United States, between 1865 and 1934 (when deposit insurance was introduced), bank shareholders were subject to ‘double liability’ - in the event of failure they were liable to lose not merely their equity stake but also an additional amount corresponding to the initial value subscribed for the shares they held, which could be used to repay depositors and other creditors. During this period, claims on shareholders in failed banks were quite successful in producing resources to pay out depositors.28 Following numerous bank failures during the Great Depression, however, limited liability was extended to banks, and reliance placed on deposit insurance to deter runs.

In Britain, the death knell of unlimited liability came somewhat sooner with the failure of the City of Glasgow Bank, the third largest bank in the UK, in October 1878.29 The bank had been badly managed and its accounts falsified. Following a report that revealed serious problems with the accounts, other banks withdrew their support for City of Glasgow and it closed its doors. Later that month, the directors were arrested and charged with fraud. With a promptness that seems remarkable by today’s standards, the directors went on trial the following January, were convicted and sent to prison. It was left to the shareholders to meet the costs of supporting the depositors. Not only individual shareholders but also the trustees of private trusts which owned shares in the bank were personally liable for the losses, and 80 per cent of them went bankrupt. The plight of the shareholders aroused widespread public concern and the government introduced legislation to permit banks to convert to limited liability status.

The Economist magazine backed the legislation, citing the difficult position in which many innocent shareholders found themselves: ‘an examination of the share lists of most of our banks exhibits a very large - almost an incredible - number of spinsters and widows, a considerable sprinkling of Clergymen and Dissenting Ministers, professional men, and others, whose occupations do not appear likely to have enabled them to accumulate much wealth … Out of the whole number more than one third are women.’30 Admirable though such diversity might seem today, it was seen then as evidence of the vulnerability of small investors. Spinsters and widows, let alone dissenting ministers, could not be expected to monitor and control bank executives. By August 1879, with commendable speed, the Banking and Joint Stock Companies Act, covering little more than three pages, had been passed by Parliament, requiring the publication of audited accounts and permitting banks to take advantage of limited liability.

It seems impossible to imagine now that unlimited liability could be restored. Yet limited liability in a bank with only a small margin of equity capital means that the owners have incentives to take risks - to ‘gamble for resurrection’ - because they receive all of the profits when the gamble pays off, whereas their downside exposure is limited. That is not in the best interest of the company as a whole. During the crisis, the chairman of one of Britain’s largest banks told me of his frustration that he was supposed to act solely in the interests of shareholders and not also in the interests of bondholders and depositors. After the crisis I asked a senior member of Goldman Sachs why he believed that their attitude towards risk management was an example for others to follow. The answer was simple: for a century Goldman Sachs had been a partnership and so every partner had an incentive to help monitor and manage the risks of the firm, a culture that had survived its transition to limited liability company status in 1999. In investment banking, the partnership form of business has much to be said for it. Those who manage other people’s money are more careless than when managing their own.

Fourth, central banks could replace lost deposits by providing official loans to a bank experiencing a run. It would, in the phrase that has become only too well known, act as a ‘lender of last resort’. If banks experienced a run then the central bank could supply liquidity to enable each individual bank to meet its depositors’ demands until the panic subsided and confidence returned. And if the crisis were a purely temporary one and the bank were solvent - in the sense that its assets could be sold once the crisis was over for an amount greater than the value of its liabilities - this ‘lender of last resort’ policy would achieve the objective of stabilising the banking system.

Flooding the system with liquidity has been seen by many economists, officials and politicians as the answer to almost any financial crisis. But it is never easy to distinguish between a liquidity and a solvency problem. In only a matter of days, a shortage of liquidity can turn into a solvency question. Banks will always claim that their problems result solely from illiquidity rather than a fall in the value of their assets. And the distinction between liquidity and solvency is one that may be observable only after a detailed examination of a bank’s balance sheet, difficult for the authorities and impossible for investors. In September 2007, the consensus was that the crisis was solely one of liquidity. But it quickly became clear that it was in fact a crisis of solvency, and that a solution would require the combined efforts of central banks and taxpayers.

The failure before the crisis was a lapse into hubris - we came to believe that crises created by maturity and risk transformation on a massive scale were problems that no longer applied to modern banking, that they belonged to an era in which people wore top hats. There was an inability to see through the veil of modern finance to the fact that the balance sheets of too many banks were an accident waiting to happen, with levels of leverage on a scale that could not resist even the slightest tremor to confidence about the uncertain value of bank assets. For all the clever innovation in the world of finance, its vulnerability was, and remains, the extraordinary levels of leverage. Pretending that deposits are safe when they are invested in long-term risky assets is an illusion. Without a sufficiently large cushion of equity capital available to absorb losses, or the implicit support of the taxpayer, deposits are inherently risky. The attempt to transform risky assets into riskless liabilities is indeed a form of alchemy.

Risk in the system and interconnectedness

Given that banks are inherently fragile, it is not surprising that there have been frequent banking crises. The failure of any one bank can cause serious problems for its depositors, other creditors, and its shareholders. But what really matters for the economy is the risk inherent in the banking system as a whole. Risk in the system is not measured by an average of the risks of each individual bank. The simplest way to see this is to consider an artificial but striking example. Borrowing short to lend long creates risk. Suppose that there are one hundred banks in the economy. The first bank issues demand deposits (that is, deposits that are instantly convertible into cash) and invests in securities issued by other banks with a maturity of six months. The second bank issues liabilities comprising securities with a maturity of six months and invests in bank securities with a maturity of twelve months. And so on. The one hundredth bank issues liabilities with a maturity of forty-nine and a half years and invests in non-bank securities with a maturity of fifty years. Bank regulators are asked to check that each bank is not exposed to an excessive degree of maturity transformation. Since, in this example, that is only six months for each bank, the regulators report a satisfactory compliance with restrictions on the degree to which banks are allowed to borrow short and lend long. But the banking system as a whole is creating demand deposits and investing in long-term securities with a maturity of fifty years. There is massive maturity transformation in the banking system as a whole. This example is clearly artificial - although before the crisis banks did transact extensively with each other in similar ways - but it illustrates the need to examine the system rather than individual banks.31

The question of interconnectedness is not restricted to links within the banking system. Since the crisis a great deal of attention has been paid to the need for banks to issue liabilities that can absorb losses in bad times without the need for taxpayer support. This is the basis of calls for banks to issue either more equity or special ‘bail-inable’ bonds that can be converted to equity capital when substantial losses arise. Such proposals have much merit, but they beg one very important question. Who are the people or institutions who hold the equity and debt that are supposed to absorb losses? If they are pension funds or insurance companies, which play a crucial role in the management of household savings, then it is far from obvious that they are best placed to absorb the risks generated in the banking system. This takes us back to the fundamental question posed by the alchemy of the banking system - we all want deposits to be safe and yet we also want to finance risky investment projects. How do we square the circle? In other words, we need to look at the financial sector as a whole, and not just the formal banking system.

In the run-up to the crisis, new institutions grew up to form a so-called ‘shadow banking’ system. In the US it became larger in terms of gross assets than the traditional banking sector, especially between 2002 and 2007, largely because it was free of much of the regulation that applied to banks. There is no clear definition of what constitutes ‘shadow banking’, but it clearly includes money market funds - mutual funds that issued liabilities equivalent to demand deposits and invested in short-term debt securities such as US Treasury bills and commercial paper.

Money market funds were created in the United States as a way of getting around so-called Regulation Q, which until 2011 limited the interest rates that banks could offer on their accounts. They were an attractive alternative to bank accounts. Such funds - and hence the owners of their liabilities - were exposed to risk because the value of the securities in which they invested was liable to fluctuate. But the investors were led to believe that the value of their funds was safe. By the time the crisis hit, such funds had total liabilities repayable on demand of over $7 trillion. And they lent significant amounts to banks, both directly and indirectly through other intermediaries.

It was because they were a significant source of funding for the conventional banking system that the Federal Reserve took action to prevent the failure of money market funds in the autumn of 2008 when, after the failure of Lehman Brothers, concern about the ability of such funds to hold their value led to a run on them.32 In Europe and Japan, money market funds did not grow to the same extent because banks had more freedom to pay interest, and since the crisis, unlike in the US, such funds have had to choose between being regulated as banks or becoming genuine mutual funds with a risk to the capital value of investors’ money.

At one time or another, almost all non-bank financial institutions have been described as shadow banks. In some cases, such as special purpose vehicles set up by banks themselves, the description is merited. Those vehicles were legal entities, such as limited liability companies, set up for the sole purpose of issuing short-term commercial paper, not dissimilar to bank deposits, and purchasing longer-term securities, such as bundles of mortgages, created by the banks themselves.33 In essence, they were off-balance-sheet extensions of banks, and during the crisis many were taken back on to the balance sheet of their parent bank. Entities such as hedge funds and other bodies carrying out fund management are also sometimes described as examples of shadow banking. But since they do not issue demand deposits, the comparison with banks is much less convincing. The challenge posed by shadow banks is to ensure that institutions engaging in the alchemy of banking are regulated appropriately, and I shall return to this issue in Chapter 7.

Financial engineering allows banks and shadow banks to manufacture additional assets almost without limit. This has had two consequences. First, the new instruments created are traded largely among big financial institutions and so the financial system has become enormously more interconnected. The failure of one firm causes trouble for the others. This means that promoting the stability of the system as a whole by regulating individual institutions is much less likely to be successful than hitherto. As in the stylised example described above, maturity and risk mismatch can grow through chains of transactions among banks and shadow banks without any significant amount being located in any one institution. Shadow banks posed as big a risk to the stability of the financial sector as conventional banks. Second, although many of these positions even out when the financial system is seen as a whole, gross balance sheets are not restricted by the scale of the real economy, and so banks and shadow banks were able to expand at a remarkable pace. When the crisis began in 2007, no one knew which banks were most exposed to risk. Almost no institution was immune from suspicion, the result of the knock-on consequences so eloquently described by Bagehot when he wrote:

At first, incipient panic amounts to a kind of vague conversation: Is A.B. as good as he used to be? Has not C.D. lost money? and a thousand such questions. A hundred people are talked about, and a thousand think, ‘Am I talked about, or am I not?’ ‘Is my credit as good as it used to be, or is it less?’ And every day, as a panic grows, this floating suspicion becomes both more intense and more diffused; it attacks more persons; and attacks them all more virulently than at first. All men of experience, therefore, try to ‘strengthen themselves,’ as it is called, in the early stage of a panic; they borrow money while they can; they come to their banker and offer bills for discount, which commonly they would not have offered for days or weeks to come. And if the merchant be a regular customer, a banker does not like to refuse, because if he does he will be said, or may be said, to be in want of money, and so may attract the panic to himself.34

The risk premium - the rate of return required by investors over and above the return on safe government debt of the same maturity - on loans even to large banks rose very sharply during the crisis. For most banks the risk premiums on their unsecured debt had more than trebled by October 2008 relative to their levels at the start of 2007. All banks, irrespective of the precise nature of their business and balance sheet, were tarred with the same brush. Even today, risk premiums are higher than before the crisis.

Size of the banking sector

The size, concentration and riskiness of banks have increased in extraordinary fashion over recent decades - as described above - and would be unrecognisable to Bagehot. A banking and financial sector is essential to economic growth. Many poor countries need a growing financial sector as they develop, and many poor people around the world will need greater access to financial facilities. But is the banking sector now too large?

Size has advantages, and not just for big banks themselves. A system comprising a few large nationwide banks may be more resilient than one consisting of a large number of small local banks. In the 1930s, the United States saw the failure of almost 10,000 banks - 4000 in 1933 alone. It was the regional fragmentation of American banking that proved so vulnerable to the downturn because there were limited opportunities to set losses in one area or industry against profits in others. The loss of so many banks was devastating to small businesses in general and agriculture in particular, and exacerbated the depression in the economy. In contrast, the UK did not experience a banking crisis during the Great Depression. One reason was the tradition of national branch banking with its greater resilience and, it is fair to say, a high degree of oligopoly, which restrained competitive pressures on banks to lend in more optimistic times.

A similar benefit from size can be seen in the experience of Canada, a country with a long history of remarkably few banking crises.35 Canada has a small number of banks (now five) and is not an international financial centre.36Those two factors have stood it in good stead and explain the comparatively good performance of Canadian banking during the recent crisis.

Those advantages of size relate to the composition of the banking sector and its concentration. There are three reasons to believe that, before the crisis, the banking sector as a whole became too large. First, because bank deposits are used as money, their failure can prevent people paying bills and receiving wages, so undermining the payments system. That is why governments are unwilling to allow banks, other than small ones, to fail. Banks, as we have seen, are too important to fail. The implicit guarantee of bank liabilities amounts to an effective subsidy, allowing banks to raise finance more cheaply than other entities in the private sector. Depositors and others who lend money to banks believe that they are in effect lending to the government.37 The belief that when in trouble banks will be bailed out by the state because they are too important to fail leads to an implicit subsidy, which means a larger banking system than is justified by the underlying economics.

Second, many of the examples of high personal remuneration, especially in the form of bonuses, in the financial sector reflect not high productivity but what economists call rent-seeking behaviour. In other words, the remuneration is far higher than is necessary to persuade people to work in the industry. Financial markets are places where delusion and greed find common cause. Many of the transactions in complex financial instruments are zero-sum - a clever trader makes money out of a less clever one. Such activity diverts talent from professions where the social returns are high, such as teaching, to those, such as finance, where the private return exceeds, often substantially, the social return.

Third, financial capital is attracted into the industry by the appearance that there are high profits to be made. Sometimes these are overstated, especially on long-term financial contracts. As mentioned earlier, a common way of exploiting normal accounting conventions for derivative and other complex transactions was to report as current income the present value of expected future cash flows, even though they had not yet been received. Some of those practices reminds me of certain football clubs, which would sell the right to future gate receipts and season ticket sales to outside investors and then report the proceeds as money available for current spending. Indeed, the practice of converting a stream of future profits into a capitalised present value was one of the ways in which Enron, the American energy company that went bankrupt in December 2001, misled the market.38 The exaggeration of the profitability of complex transactions represented by these accounting practices provides misleading incentives to the use of capital in the banking industry.

For those reasons, when the crisis hit, banks had grown to a size where it was risky to the system as a whole to allow them to be subject to normal market disciplines, where they had become almost impossible to manage (as the growing revelations of misconduct by most large banks have revealed), and where they were immune to the usual due process of law out of fear of the consequence of their failure for the financial system as a whole. Banks had become too big to fail, too big to sail, and too big to jail. And in some countries, the size of the banking sector had increased to the point where it was beyond the ability of the state to provide bailouts without damaging its own financial reputation - for example in Iceland and Ireland - and it proved a near thing in Switzerland and the UK. The fate of the Bank of Scotland (which in 2001 evolved into HBoS) and the Royal Bank of Scotland, founded in 1695 and 1727 respectively, encapsulates the problems of British banking. Between 2000 and 2008, the exposure of HBoS to commercial property rose by 600 per cent, resulting in a potentially large loss if asset prices were to fall. RBS’s exposure to commercial property also rose rapidly, at an annualised rate of 21 per cent, and the fragility of its balance sheet was exacerbated by the ill-timed acquisition of the Dutch bank ABN Amro in the autumn of 2007. Both had to be rescued by the taxpayer, in the form of the Bank of England and the government, in 2008. Millions of passengers in Europe arriving at airport gates emblazoned with the letters RBS are reminded daily not of the successes of UK banking but of its failures.

By 2008, too many banks were an accident waiting to happen. When the banking system failed in September of that year, not even massive injections of both liquidity and capital by the state could prevent a devastating decline of confidence and output around the world. So it is imperative that we find an answer to the question of how to make our banking system safer.

The attempt to link the creation of money, in the form of bank deposits, to the financing of long-term risky investments seems attractive in normal times because the alchemy it represents conceals the true cost of the arrangement. Only when the inherent fragility is revealed by a crisis do the costs become apparent. The demand for a safe reserve of purchasing power in the form of money can increase hugely and suddenly in times of crisis. But banks and even governments cannot create safe assets. To understand the demand for safe assets, and the extent to which our banking system as presently constituted can respond to that demand, requires a deeper analysis of the nature of uncertainty.