INNOCENCE REGAINED: REFORMING MONEY 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)

Chapter 7. INNOCENCE REGAINED: REFORMING MONEY AND BANKING

‘How is it possible to expect that Mankind will take advice, when they will not so much as take warning?’

Jonathan Swift, ‘Thoughts on Various Subjects’, 1703

‘We call upon every man who professes to be animated with the principles of the democracy, to assist in accomplishing the great work of redeeming this country from the curse of our bad bank system.’

William Leggett, Evening Post, 6 August 1834

For centuries, alchemy has been the basis of our system of money and banking. Governments pretended that paper money could be turned into gold even when there was more of the former than the latter. Banks pretended that short-term riskless deposits could be used to finance long-term risky investments. In both cases, the alchemy is the apparent transformation of risk into safety. For much of the time the alchemy seemed to work. From time to time, however, people realised that the Emperor had far fewer clothes than the Masters of the Universe wanted us to believe. Once confidence in the value of money or the soundness of banks was lost, there was a monetary or banking crisis. As Bagehot wrote in Lombard Street, ‘The peculiar essence of our financial system is an unprecedented trust between man and man; and when that trust is much weakened by hidden causes, a small accident may greatly hurt it, and a great accident for a moment may almost destroy it.’1 For a society to base its financial system on alchemy is a poor advertisement for its rationality. The key to ending the alchemy is to ensure that the risks involved in money and banking are correctly identified and borne by those who enjoy the benefits from our financial system.

How can we regain the innocence and trust in banking that, as described in Chapter 3, was lost over a long period in which crises became accepted as an inevitable feature of the financial landscape? Many English children used to be brought up on John Bunyan’s Pilgrim’s Progress, published in 1678 - a religious allegory telling the story of the trials and tribulations of Christian’s long and difficult journey from his home in the City of Destruction to find the Celestial City on Mount Zion, passing on the way through the Slough of Despond, Hill Difficulty, Vanity Fair, Hill Lucre and Doubting Castle. Much less well known and less well-read than Bunyan’s book is The Political Pilgrim’s Progress, published in 1839.2 It is the story of the journey of Radical from the City of Plunder to the City of Reform. In the City of Plunder:

Its people seemed active, industrious, and enterprising; but there appeared a singular custom amongst them, which greatly marred their social happiness and unanimity, and this was, that nearly one half of the inhabitants made a practice of putting their hands into the pockets of the other half, and taking their money from them. There was a law, indeed, for this singular custom … fortified by a thing called ‘government’ [which] always vehemently affirmed that the mode of making one half of the people work for and support the other half, was the very perfection of human wisdom.3

Like Christian, Radical faces many trials and tribulations while passing through many of the same places, particularly Vanity Fair. In that den of iniquity, with his friend Common-sense, Radical is shown the Paperkite-Buildings:

Here the people seemed to talk an altogether new language, different from anything that Pilgrim had hitherto heard in this metropolis. There was an everlasting chatter, like the incessant crowings of a rookery, about stocks, funds, omnium, scrip, debentures, rentes, metalliques, discounts, premiums, exchequer bills, shares, accounts, balances, advances, consols, India stock, bank stock, exchanges, settling days, bear and bull accounts, lame ducks, pressures, panics, long annuities, bar gold, bullion, coin, mint prices &c. The agitation and anxiety amongst the moving throng of the Buildings were exceedingly interesting. The people were all exchanging bits of paper, one with another; and this act was designated by the phrase of ‘the circulating medium’, on which many large volumes of books had been written, and which was considered as an occult science in that part of the country.4

But when he finally reaches his goal, Radical finds that in the City of Reform ‘there was no such thing as a stock exchange or a saving bank, or a bank note for any sum under FIFTY pounds’.5

Drawing inspiration from Radical, we might well ask how we can find our way to the City of Reform. The pretence that the illiquid real assets of an economy - the factories, capital equipment, houses and offices - can suddenly be converted into money or liquidity is the essence of the alchemy of the present system. Banks and other financial intermediaries will always try to finance illiquid assets by issuing liquid liabilities because they make profits by paying less on the latter than they earn on the former. That is why, although money is a public good, the bulk of its supply is provided by commercial banks. The problem is that the liquidity promised to investors or depositors can be supplied only if at each moment a small number of people wish to convert their claim on the bank into cash. Liquidity simply disappears if everyone wishes to convert their claim into money at the same time. What may be possible for a small number of people is self-evidently impossible for the community as a whole. And the problem is made worse by the fact that if a depositor believes that others are likely to try to take their money out, it is rational for him or her to do the same and get to the front of the queue as soon as possible - a bank run. Runs reflect the underlying alchemy and make the system unstable.

Liquidity is an illusion; here one day, gone the next. It reminds me of those attractive soap bubbles that one can blow into the air. From a distance, they look appealing. But if you ever try to hold them in your hand, they disappear in a trice. And whenever at the same time many people try to convert their assets into a liquid form, they often discover that liquidity has disappeared without trace. When there is a sudden jump in the demand for liquidity and investors rush to convert their claims on illiquid assets into money, the result is usually a crisis, exposing the alchemy for what it is. Liquidity is, however, only one aspect of the alchemy of our present system. Risk, and its impact on the solvency of banks, is the other. And in the recent crisis, concern about solvency was the main driver of the liquidity problems facing banks. When creditors started to worry that bank equity was insufficient to absorb potential losses, they decided that it was better to get out while the going was good. Concerns about solvency, especially in a world of radical uncertainty, generate bank runs. To reduce or eliminate alchemy, we need a joint set of measures to deal with both solvency and liquidity problems.

If a market economy is to function efficiently, businesses and households need a secure mechanism by which to pay their bills and receive wages and salaries. Ordinary current accounts are not vehicles for speculative investments and it is important that they have a stable value in terms of money, in which payments are denominated. But if a bank has assets that are highly risky, as many of its loans may be, then it is alchemy to pretend that deposits can be secure. So governments decided to guarantee deposits, first by creating deposit insurance and then in the recent crisis by extending blanket guarantees to all bank creditors. Because of their importance to the economy, and their political power, banks had become too important to fail. And the larger they became, the more likely it was that the government would bail them out in times of difficulty. Central banks lent vast sums to commercial banks. That stopped the rot, in the sense of removing the incentive to run on a bank, but at the cost of shifting the risk of the assets of banks on to taxpayers. In the case of Ireland, it almost bankrupted the country.

The toxic nexus between limited liability, deposit insurance and lender of last resort means that there is a massive implicit subsidy to risk-taking by banks. After the 1980s, when banking was liberalised, the degree of alchemy, and hence of subsidy, inherent in the risk and maturity transformation in the system increased. No individual bank could easily walk away from the temptation to exploit the subsidy. Each bank faced a prisoner’s dilemma. Only by running down its holdings of liquid assets, and financing itself as cheaply as possible by short-term debt, could it keep up with the rising profitability of its peers.

In short, compared with a century or even fifty years ago, banks have been financing themselves with too little equity and holding too few liquid assets. Before the crisis, equity was insufficient to absorb potential losses from the risks being taken, which meant that it was more likely that depositors or other short-term creditors would think about running on the bank in the wake of bad news. And in the event of a run, there were insufficient liquid assets to enable the bank to douse the flames by paying out. Even governments recognised that something had to change.

Official sector reforms

Since the crisis, the official sector has been hyperactive. Both at the national and international level, regulators have been tightening up on the freedoms given to banks in respect of how they finance themselves, their structure and their conduct. At the international level, a concerted effort has been made by the major countries in the G20, working through the Basel Committee of the Bank for International Settlements, to rectify some of the pre-crisis failures in regulation. The minimum amount of equity a bank must use to finance itself, known as its capital requirement, has been raised, and banks also have to hold a minimum level of liquid assets related to the deposits and other short-term financing that could run from the bank within thirty days, known as the liquidity coverage ratio. Regulators are also conscious of the need to look outside the boundaries of the traditional banking sector to see if elements of alchemy are appearing in the ‘shadow’ banking sector, and to conduct stress tests to see if banks are capable of withstanding the losses incurred due to particular adverse scenarios.

The Financial Stability Board, a group of officials from the G20, is leading work both on these investigations and on the challenge posed by the potential failure of banks that operate across borders. Nationally, regulators in countries such as Sweden, Switzerland and the United States have imposed additional capital requirements over and above the internationally agreed minimum. Countries such as the United Kingdom and United States have introduced legislation to separate, or ring-fence, basic banking operations from the more complex trading activities of investment banking.6 And most countries have either improved or introduced special bankruptcy arrangements - known as resolution mechanisms - to enable a bank in trouble to continue to provide essential services to its depositors while its finances are being sorted out and, if necessary, to facilitate a speedy transfer of depositors from a failing to a profitable bank. That represents a significant shift of opinion since before the crisis, when most countries were primarily concerned to ensure that their banking system was not weighed down by heavier regulation than in other countries. Ensuring a safer banking system is now seen as in a country’s self-interest. And national regulators, often working together, have pursued cases of misconduct by bank employees and levied substantial fines, as described in Chapter 3, as well as restricting compensation of bank executives.

Moreover, the market itself has imposed its own discipline on banks and other financial institutions. As a result, the banking system has changed a great deal since 2008. The largest banks have become smaller; the balance sheet of Goldman Sachs in 2015 was around one quarter smaller than in 2007. Investment banking is not as profitable now as it was when asset prices were rising in the wake of falling real interest rates. Many banks have cut back on the size of their investment banking operations and some, such as Citigroup and Bank of America, have sold their proprietary trading desks, which bought and sold investments on their own account, and turned themselves back into more traditional commercial banks.

Is all this enough? I fear not, and for one simple reason. Radical uncertainty means that sentiment towards financial firms can change so quickly that regulations which appear too burdensome one moment seem too lenient the next. The experience of 2007-8 illustrates what can happen. Let’s ask the following question: how much equity finance does a bank need to issue in order to persuade potential creditors that it is safe for them to lend to the bank? Before the crisis, the answer was hardly any at all. Markets were content to lend large sums to banks at low interest rates, even though banks were highly leveraged. After 2008, the answer was a very large amount. Not even the new higher levels of capital mandated by regulators were sufficient to ensure that markets were happy to restore previous levels and pricing of funding. The innocence that was lost during the crisis was proving very expensive to regain. For investors, the narrative about the wisdom of lending to banks had changed. So it is extremely difficult to know the appropriate level of equity finance a bank should be required to use in a world where alchemy is still a characteristic of the banking system. And the right answer can change from one day to the next. In 2012, the Spanish bank Bankia reported a risk-weighted capital ratio of over 10 per cent, well above the regulatory minimum; three months later it required a capital injection of €25 billion.

Two further aspects of current regulation are difficult to reconcile with radical uncertainty. The first is that official capital requirements are calculated with respect to estimates of the riskiness of different assets on a bank’s balance sheet. As described in Chapter 4, each type of asset is given a risk weight, agreed by international regulators, and this is used to calculate the overall amount of equity a bank must issue. Mortgage lending, for example, was thought on the basis of past experience to be relatively safe, and was given a low risk weight. Sovereign debt was believed to be so safe that it was given a zero risk weight, meaning that banks did not have to raise any equity finance in respect of such investments and so had no additional capacity to absorb losses on them. So complex did the system become that banks were allowed to propose their own internal models to calculate risk weights. It turned out that some banks had very different estimates of the riskiness of the same assets than others, undermining confidence in the fairness of the regulations.7

The people who designed those risk weights did so after careful thought and an evaluation of past experience. But they simply did not imagine how risky mortgage lending and the sovereign debt of countries such as Greece would become during the crisis, nor how large were the risks inherent in much more complicated financial instruments. Rather than lambast the regulators for not anticipating those events, it is more sensible to recognise that the pretence that it is possible to calibrate risk weights is an illusion. The need for banks to use equity to absorb losses is most important in precisely those circumstances where something wholly unexpected occurs and previous calculations of risk weights are irrelevant. That is why, during the crisis, a measure of equity relative to total assets was a much better indication of safety than equity relative to total risk-weighted assets. Risk weights in the design of capital regulation seem attractive at first sight, but they break in our hands when we try to use them. A simple leverage ratio is a more robust measure for regulatory purposes.

The second problematic aspect of current regulation concerns the requirement for banks to hold a minimum level of liquid assets - the liquidity coverage ratio - so that they can withstand an unusually high demand for the repayment of debt or deposits. I chaired the meetings that eventually agreed on the definition of the ratio. The discussions were overshadowed by one major conceptual problem. How could we define assets that were always liquid? Before the crisis, it seemed obvious that government debt was a safe and liquid asset. But experience showed that the sovereign debt of some countries was far from safe and liquid. Moreover, other countries, such as Australia, had managed their public finances so carefully that there was simply too little government debt outstanding to supply the demand for liquid assets. It seemed rather odd to penalise well-managed countries for not issuing large amounts of government debt! At heart, the problem was the failure to recognise that in a world of radical uncertainty only the central bank can create liquidity, and so liquidity regulation has to be seamlessly integrated with a central bank’s function as the lender of last resort.

None of this means that the extraordinary efforts of regulators in recent years to improve the system have been a mistake. But they are in danger of failing to see the wood for the trees. Regulation has become extraordinarily complex, and in ways that do not go to the heart of the problem of alchemy. The objective of detail in regulation is to bring clarity, not to leave regulators and regulated alike uncertain about the current state of the law.8 Much of the complexity reflects pressure from financial firms. By encouraging a culture in which compliance with detailed regulations is a defence against a charge of wrongdoing, bankers and regulators have colluded in a self-defeating spiral of complexity. No capitalist economy can prosper without sufficient certainty about the way that rights and obligations will be interpreted and enforced. Arbitrary regulatory judgements impose what is effectively a high tax on all investments and savings. The fact that it was in England that the Industrial Revolution began was in part the result of a stable and predictable framework for doing business with others. As the father of English commercial law, Lord Mansfield, put it in 1761 as the Industrial Revolution was gathering pace: ‘The daily negotiations and property of merchants ought not to depend upon subtleties and niceties; but upon rules easily learned and easily retained, because they are the dictates of common sense, drawn from the truth of the case.’9

Not many people can easily absorb and retain the totality of current financial regulation, and those who try are not left with the impression that it is common sense. The Dodd-Frank Act passed in the US in 2010 contained 2300 pages, with many thousands of pages more expected to cover the detailed rules that will follow, whereas the Glass-Steagall Act of 1933, which separated commercial and investment banking, covered a mere thirty-seven pages.10 In Britain, the Prudential Regulation Authority and the Financial Conduct Authority have combined rulebooks exceeding ten thousand pages.11 Such complexity feeds on itself and brings the system into disrepute. Efforts to comply with financial regulation are a barrier to new small firms trying to enter the financial sector, and, in advanced countries, result in the employment of several hundred thousand people. To employ such a large number of talented people to cope with complex regulation constitutes a large ‘deadweight’ cost to society.

As we saw in Chapter 4, complexity is an inefficient way of coping with radical uncertainty. Can we do better?

More radical reforms

Frenetic activity among the official community cannot conceal the fact that, although much useful repair to the fabric of regulation has been made, nothing fundamental has changed. The alchemy of our banking system remains. Since the bank bailouts in most advanced economies were huge, it is surprising that more has not been done since the crisis to address the fundamental problem. The scale of central bank lending to a relatively small number of financial institutions was so large that one could have paraphrased Winston Churchill by saying that never in the field of financial endeavour had so much been owed by so few to so many - and with so little radical reform. Of course, governments, financial regulators and central banks are all well aware of the nature of the problem, but official efforts to tackle it are in stark contrast with more radical ideas proposed, albeit never implemented, by earlier generations of economists.

Even though the degree of alchemy of the banking system was much less fifty or more years ago than it is today, it is interesting that many of the most distinguished economists of the first half of the twentieth century believed in forcing banks to hold sufficient liquid assets as reserves to back 100 per cent of their deposits. They recommended ending the system of ‘fractional reserve banking’, under which banks create deposits to finance risky lending and so have insufficient safe cash reserves to back their deposits.12 The elimination of fractional reserve banking was a proposal put forward in 1933 as the ‘Chicago Plan’.13 The proponents of the plan included the brilliant American monetary theorist Irving Fisher and a distinguished group of economists at Chicago such as Frank Knight, Henry Simons and Paul Douglas; later support came from right across the spectrum of post-war economists, ranging from Milton Friedman to James Tobin and Hyman Minsky.14 Interestingly, John Maynard Keynes was not part of this group, largely because Britain did not experience a banking crisis in the 1930s and his focus was on restoring output and employment.15 More recently, a number of economists have proposed variations on the same theme: John Cochrane from Chicago, Jaromir Benes and Michael Kumhof from the IMF, the British economists Andrew Jackson, Ben Dyson and John Kay, Laurence Kotlikoff from Boston and the distinguished FT commentator Martin Wolf.16

There are two ways of looking at these radical approaches to banking reform, one by focusing on the banks’ assets and the other on their liabilities. The essence of the Chicago Plan was to force banks to hold 100 per cent liquid reserves against deposits. Reserves would include only safe assets, such as government securities or reserves held with the central bank. In this way there would be no reason for anyone to run on a bank, and even if some people did withdraw their deposits there would be no incentive for others to join them, because there would always be sufficient funds to support the remaining deposits.

So far, so good. But who would perform the many functions that banks carry out today, especially lending to businesses and households, so enabling them to build factories and purchase homes? In other words, who would finance the transfer of existing assets and bear the risk involved in financing new investment? That relates to the liabilities of banks under such radical reforms. If deposits must be backed with safe government securities, then it follows logically that all other assets, essentially risky loans to the private sector, must be financed by issuing equity or long-term debt, which would absorb any losses arising from those risky assets. As a result, this approach would, in effect, separate safe and liquid ‘narrow’ banks, carrying out payment services, from risky and illiquid ‘wide’ banks performing all other activities.17 It would be illegal for wide banks, including the ‘shadow’ banking sector, to issue demand, or even short-term, deposits.18

The great advantage of reforms such as the Chicago Plan is that bank runs and the instability they create would disappear as a source of fragility. The Chicago Plan breaks the link between the creation of money and the creation of credit. Lending to the real economy would be made by wide banks and financed by equity or long-term debt, not through the creation of money. Money would once again become a true public good with its supply determined by the government or central bank.19 Governments would not have to fight against the swings in money creation or destruction that automatically occur today when banks decide to expand or contract credit. It was the sharp fall in credit and money after 2008 that led to the massive expansion of money via quantitative easing. As Irving Fisher put it, ‘We could leave the banks free … to lend money as they please, provided we no longer allowed them to manufacture the money which they lend … In short: nationalize money but do not nationalize banking.’20 And the clarity and passion of Fisher in the 1930s are echoed in the arguments of John Cochrane and Martin Wolf today. Such reforms would indeed eliminate the alchemy in our banking system, which the official reform agenda fails to tackle.

So why hasn’t the idea been implemented? One explanation is that it would eliminate the implicit subsidy to banking that results from the ‘too important to fail’ nature of most banks. Banks will lobby hard against such a reform. To protect the system of making payments, as crucial to the daily functioning of the economy as electricity is to our daily lives, governments will always guarantee the value of bank accounts used to make payments, and it is therefore in the interest of banks to find ways of putting risky assets on to the same balance sheet as deposits. More importantly, however, eliminating alchemy in this particular way has some other disadvantages. First, the transition from where we are today to complete separation of narrow and wide banks could be disruptive, forcing a costly reorganisation of the structure and balance sheet of existing institutions. It would be easy for the banking community to portray such a move as unwarranted interference in the management of private banks, and even the much more limited ring-fencing adopted in the UK has come under attack for precisely this reason.

Second, the complete separation of banks into two extreme types - narrow and wide - denies the chance to exploit potential economic benefits from allowing financial intermediaries to explore and develop different ways of linking savers, with a preference for safety and liquidity, and borrowers, with a desire to borrow flexibly and over a long period. Constraining financial intermediation would mean that the cost of financing investment in plant and equipment, houses and other real assets would be higher. The potential efficiencies in using different ways of bringing savers and investors together would be lost by legally mandating a complete prohibition on the financing of risky assets by safe deposits - provided that we could find other ways, as I discuss below, of following a path that would lead to the end of alchemy.

Third, and most important of all, radical uncertainty means that it is impossible for the market to provide insurance against all possible contingencies, and one role of governments is to provide catastrophic insurance when something wholly unexpected happens. Ending alchemy does not in itself eliminate large fluctuations in spending and production. In a world of radical uncertainty, where it is possible that households and businesses will make significant ‘mistakes’ about the future profitability of investment, there is always a risk of unexpected sharp changes in total spending.

Ensuring that money creation is restored to government through the requirement for narrow banks to back all deposits with government securities does stop the possibility that runs on the banking and shadow banking sectors will transmit shocks at rapid speed right across the financial sector, as happened to such devastating effect in 2008. But the risk from unexpected events is then focused on the prices of assets held directly by households and businesses and on the solvency of wide banks. It would be possible for governments to stand back and allow the prices of real assets and claims on those assets, including the prices of the bonds and equity of wide banks, to take the hit. As discussed in Chapter 5 and again in Chapter 9, one of the most difficult issues in monetary policy today is the extent to which central banks should intervene in these asset markets - either to prevent an ‘excessive’ rise in asset prices in the first place or to support prices when they fall sharply. It is difficult because the case for intervention rests on the view that the central bank knows better than other people when market prices reflect ‘mistakes’. I am not sure that their track record justifies an optimistic judgement of the ability of central banks to see the rocks ahead and steer the economy around them. Providing emergency money to meet a sharp jump in the demand for liquidity and central bank reserves is one thing; impeding a move of the economy, and asset prices, to a new equilibrium quite another.

But there is a somewhat more compelling argument for the provision of catastrophe insurance to financial intermediaries. It stems from the importance of debt finance in the financing of the real economy. Since the crisis it has become fashionable to obsess about the role of debt - and it is indeed a good idea not to owe too much in case an unpleasant surprise leads to financial embarrassment. But debt has a special role, especially when held against collateral; that is, when the lender takes a claim on assets owned by the borrower that act as a guarantee of the loan in the event of a failure to repay. The lender will typically lend only a proportion of the value of the collateral, applying a haircut to cover the risk that if the borrower defaults the collateral will still be sufficient to repay the debt. The more liquid and less volatile the collateral, the lower the value of the haircut.

Debt finance of this kind means that the lender does not need to monitor carefully the twists and turns of the venture to which the loan was extended - which may be well-nigh impossible in the case of small or complex businesses - but only the value of the collateral. That is why many small businesses find it difficult to obtain equity finance and their owners have to pledge the value of their home as collateral when borrowing. It is also why banks find lending to students unappealing: it is difficult to monitor their ability to repay and they can offer little by way of collateral. Collateral is valuable precisely because its value does not depend upon the borrower’s creditworthiness. The extension of debt backed by collateral helps to overcome the pervasive lack of information about borrowers’ creditworthiness.21 It oils the wheels of those parts of the economy that other sources of finance, such as equity, cannot reach. And one of the key roles of financial intermediaries is to lend against collateral.

Collateralised borrowing is, therefore, an important feature of the financial system, and it will survive the elimination of the incentive to run on deposits and other short-term unsecured debt. Although wide banks cannot create money in the form of deposits, they can still borrow short and lend long. In both cases they use collateral. They lend to households and businesses against real assets, and they borrow against financial securities created for the purpose, which give the impression to purchasers of bank debt of being liquid and safe but ultimately are backed by the long-term loans and other assets of the bank. A large quantity of paper claims on underlying assets has been constructed to satisfy the demand for collateral. In this way, even wide banks create a degree of alchemy. When unexpectedly bad news arrives, collateral falls in value and is perceived as more volatile and less liquid than before. Lenders will want more collateral to continue or roll over existing loans. Borrowers, whether businesses or banks themselves, may be forced to sell assets in order to replace withdrawn loan facilities, and the attempt by all borrowers to obtain sufficient collateral creates a multiplier effect, driving down asset prices further. All this is very much worse in a banking and financial system where runs can occur. But it would not be entirely absent in a world of wide banks. It would create a demand for governments or central banks to provide catastrophic insurance in the form of supporting the value and liquidity of collateral. Undoubtedly, one of the motives for the bailouts of the creditors of banks and other financial firms in 2008 was the conviction that failures of such firms would cause the financial system as a whole to freeze up and contract the availability of credit to the real economy. As former US Treasury Secretary Timothy Geithner wrote in his memoirs, ‘the only way for crisis responders to stop a financial panic is to remove the incentives for panic, which means preventing messy collapses of systemic firms, assuring creditors of financial institutions that their loans will be repaid, and … exposing taxpayers to more short-term risk’.22

Whatever the merits of the actions taken in 2008, there is no doubt that an observer could say wryly, ‘I wouldn’t start from here.’ Runs on conventional and shadow banking systems alike led to a collapse of both. The size and cost of creditor bailouts were increased significantly by the inadequate amounts of equity available to absorb losses in the banking system. And attempts to provide liquidity insurance through central bank facilities as the lender of last resort failed to penalise banks that took advantage of such support, not least because to collect insurance premiums when paying out on the policy is rather late in the day and might have made matters worse. The system in place before the crisis provided many incentives for banks to structure themselves in a way that made a crisis more likely - and that is exactly what concern about ‘moral hazard’ means. Standing ready to do whatever it takes to keep the financial system functioning is not enough. The system itself has to be designed carefully in order to reduce the frequency and severity of crises. There is a case for the provision of catastrophic insurance - but not unconditionally and not in the way that was forced on policy-makers in the circumstances of 2008.

Some commentators have taken issue with concerns about moral hazard, arguing, by analogy, that fire departments put out fires started by people who smoke in bed. But as a society we supplement fire services by strict regulations to make it less likely that fires will start. We need to do the same in the financial sector. Too much thinking about how to respond to crises, especially in the United States, has focused on throwing money at the problem once the fire has broken out. We need to anticipate the problem. There is wisdom in the Roman saying ‘Si vis pacem, para bellum’ - if you want peace, prepare for war.

Can we find a way of retaining the attractive feature of the Chicago Plan - that it ends bank runs - while at the same time reducing alchemy in the wider banking system? Or, to put it another way, can we reduce the cost of eliminating alchemy?

A new approach - the pawnbroker for all seasons

The way forward is to recognise that the prohibition on the creation of money by private banks is not likely to be sufficient to eliminate alchemy in our financial system. Radical uncertainty means that the provision of catastrophic insurance in some circumstances is desirable. Bagehot’s concept of a lender of last resort is, in some key respects, outdated. He understood that it was impossible in a crisis to tell whether a bank was or was not solvent, but that would not matter if the central bank could lend against ‘good collateral’. In his day, and until relatively recently, banks held large amounts of government securities and secure private commercial paper on their balance sheets. Good collateral was in plentiful supply. But when banks ran down their holdings of liquid assets, that all changed. The result was that in the crisis there was not enough good collateral and central banks had to take ‘bad’ collateral in the form of risky and illiquid assets on which haircuts, often large ones, had to be imposed to avoid risk to taxpayers. In consequence, central banks could lend only a proportion of the liquid funds that a bank might need. As described in Chapter 5, central bank lending encumbers the balance sheet, reducing the collateral available for other creditors, thereby encouraging them not to roll over their loans to the bank.

The essential problem with the traditional LOLR is that, in the presence of alchemy, the only way to provide sufficient liquidity in a crisis is to lend against bad collateral - at inadequate haircuts and low or zero penalty rates. Announcing in advance that it will follow Bagehot’s rule - lend freely against good collateral at a penalty rate - will not prevent a central bank from wanting to deviate from it once a crisis hits. Anticipating that, banks have every incentive to run down their holdings of liquid assets and to finance themselves with large amounts of debt, and that is what they did. It is not enough to respond to the crisis by throwing money at the system to douse the fire while reciting Bagehot; ensuring that banks face incentives to prepare in normal times for access to liquidity in bad times matters just as much.

It is time to replace the lender of last resort by the pawnbroker for all seasons (PFAS). A pawnbroker is someone who is prepared to lend to almost anyone who pledges collateral sufficient to cover the value of a loan - someone who is desperate for cash today might borrow $25 against a gold watch. Since 2008, central banks have become used to lending against a much wider range of collateral than hitherto, and it is difficult to imagine that they will be able to supply liquidity insurance without continuing to do so. In the spirit of not letting a good crisis go to waste, I think it is possible to build on two of the most important developments in central banking since the crisis - the expansion of lending against wider collateral and the creation of money by quantitative easing - to construct a new role for a central bank as such a pawnbroker. I stress this point because so many proposals for reform create alarm among bankers, and often therefore governments, since they are a step into the unknown. In contrast, the idea of PFAS is a natural extension of measures already introduced.

When there is a sudden jump in the demand for liquidity, the pawnbroker for all seasons will supply liquidity, or emergency money, against illiquid and risky assets. Only a central bank on behalf of the government can do this. But it will do so within a framework that eliminates the incentive for bank runs. The idea of the PFAS is a coping strategy in the face of radical uncertainty.

Inspiration for the principle of a PFAS can be drawn from the American journalist William Leggett, who wrote in an article in the New York Evening Post in December 1834:

Let the [current] law be repealed; let a law be substituted, requiring simply that any person entering into banking business shall be required to lodge with some officer designated in the law, real estate, or other approved security, to the full amount of the notes which he might desire to issue … Banking, established on this foundation, would be liable to none of the evils arising from panic; for each holder of a note would, in point of fact, hold a title-deed of property to the full value of its amount.23

The aim of the PFAS is threefold. First, to ensure that all deposits are backed by either actual cash or a guaranteed contingent claim on reserves at the central bank. Second, to ensure that the provision of liquidity insurance is mandatory and paid for upfront. Third, to design a system which in effect imposes a tax on the degree of alchemy in our financial system - private financial intermediaries should bear the social costs of alchemy.24

The basic principle is to ensure that banks will always have sufficient access to cash to meet the demands of depositors and others supplying short-term unsecured debt. The key is to look at both sides of a bank’s balance sheet. Start with its assets. Each bank would decide how much of its assets it would position in advance at the central bank - that is, how much of the relevant assets the central bank would be allowed to examine and which would then be available for use as collateral.25 For each type of asset the central bank would calculate the haircut it would apply when deciding how much cash it would lend against that asset. Adding up all assets that had been pre-positioned, it would then be clear how much central bank money the bank would be entitled to borrow at any instant. Because these arrangements would have been put in place well ahead of any crisis, there would be no difficulty in the central bank agreeing to lend at a moment’s notice. The assessment of collateral, and the calculation of haircuts, have become routine since the crisis and would become a normal function of a central bank as a PFAS. The amount which a bank was entitled to borrow against pre-positioned collateral, added to its existing central bank reserves, is a measure of the ‘effective liquid assets’ of a bank.

The second step is to look at the liabilities side of a bank’s balance sheet - its total demand deposits and short-term unsecured debt (up to, say, one year) - which could run at short notice. That total is a measure of the bank’s ‘effective liquid liabilities’.26 The regulatory requirement on banks and other financial intermediaries would be that their effective liquid assets should exceed their effective liquid liabilities. Almost all existing prudential capital and liquidity regulation, other than a limit on leverage, could be replaced by this one simple rule. The rule would act as a form of mandatory insurance so that in the event of a crisis a central bank would be free to lend on terms already agreed and without the necessity of a penalty rate on its loans. The penalty, or the price of the insurance, would be encapsulated by the haircuts required by the central bank on different forms of collateral. Just as motorists are compelled to take out third-party car insurance to protect other road-users, so banks should be made to take out a certain amount of liquidity insurance in normal times so that they can access central bank provision of their liquidity needs in times of crisis.

Consider a simple example of a bank with total assets and liabilities each equal to $100 million. Suppose that it has $10 million of assets in the form of reserves at the central bank, $40 million in holdings of relatively liquid securities and $50 million in the form of illiquid loans to businesses. If the central bank decided that the appropriate haircut on the liquid securities was 10 per cent and on the illiquid loans was 50 per cent, then it would be willing to lend $36 million against the former and $25 million against the latter, provided that the bank pre-positioned all its assets as available collateral. The bank’s effective liquid assets would be $(10 + 36 + 25) million, a total of $71 million. It would have to finance itself with no more than $71 million of deposits and short-term debt.

Banks would be free to decide on the composition of their assets and liabilities, allowing variety and experimentation in the types of business they transacted, all subject to the constraint that alchemy in the private sector is eliminated. The PFAS adds a desirable degree of flexibility to the Chicago Plan.

It would be possible, and sensible, to implement the scheme gradually over a period of, say, ten to twenty years. The existing degree of alchemy can be calculated as the excess of effective liquid liabilities over effective liquid assets as a proportion of the total size of the balance sheet. Suppose that in the above example the bank today had liabilities of $50 million of deposits, $35 million of short-term unsecured debt (with a maturity of less than one year), $10 million of long-term debt and $5 million of equity. Its effective liquid liabilities would be $85 million, leaving a shortfall from its $71 million of effective liquid assets of $14 million and a current degree of alchemy of 14 per cent. With a twenty-year transition, the bank would be required to reduce that degree by 0.7 per cent each year so that by the end of the transition period it would completely satisfy the rule that its effective liquid assets exceeded its effective liquid liabilities, and there would be zero alchemy. During the transition it would probably be sensible to retain existing prudential regulation, and the ring-fencing restrictions imposed in recent legislation, partly to see how they worked and partly as an incentive for the financial sector to see the transition through. Because the PFAS builds on some of the extraordinary developments in central bank balance sheets, now is the ideal moment to begin reform. It may take something like twenty years to eliminate the alchemy in our system completely, but there is no reason to delay the start of that journey.

As with any reform of this kind, the scheme would apply to all financial intermediaries, banks and shadow banks, which issued unsecured debt with a maturity of less than one year above a de minimis proportion of the balance sheet. That is an arbitrary figure, and open to debate. A key challenge is to ensure that alchemy does not simply migrate outside the regulated sector, and end up benefiting from an implicit public subsidy.27 No doubt there would be other practical issues to resolve, but the reason we employ high-quality public servants is to solve such problems and not allow lobbyists to use them as an excuse for resisting principled reform.

The idea of the pawnbroker for all seasons builds on both the tradition of the lender of last resort and the experience of the crisis of 2008. It has six main advantages.

First, the proposal recognises that in a real crisis the only source of liquidity is the central bank, supported by a solvent government, which can convert illiquid assets into liquid claims.

Second, the idea provides a natural transition to a state in which the alchemy of the present private sector system of money and banking is eliminated.

Third, it avoids the choice between either the status quo or the extreme radical solution of 100 per cent reserve banking with all lending financed by equity. It allows banks and other financial intermediaries to choose for themselves the structure of their balance sheet and how to relate particular types of assets to the structure of their liabilities. In so doing, it offers a way of promoting competition in the financial sector while restricting the degree of alchemy. Compared with the Chicago Plan, it lowers the cost of eliminating bank runs.

Fourth, it solves the moral hazard problem associated with the conventional lender of last resort. Banks will be required to take out insurance in the form of pre-positioned collateral with the central bank, so that, when required, liquidity can be provided quickly and cheaply on demand. There will be no need to apply a penalty rate on lending during a crisis because the disincentive to rely on the provision of central bank liquidity is provided by the haircuts on collateral. And the central bank can assess the collateral in normal times and not, as happened during the crisis, be forced to make snap judgements about collateral when the storm arrives. No doubt in normal times there will be pressure on the central bank to set haircuts to favour politically popular types of bank lending and intense lobbying by banks to lower the haircuts. But central banks are in a stronger position to resist such calls in normal times than after the crisis has hit. Moreover, a distortion of haircuts would not change the fact that with guaranteed access to central bank liquidity under PFAS, the incentive to run on a bank, or not to roll over short-term debt, would disappear. So the stigma of the use of LOLR assistance (see Chapter 5) would be much less.

Fifth, it exploits today’s abnormal circumstances by incorporating two of them into a permanent feature of the PFAS. First, by creating money through quantitative easing, central banks have greatly expanded their balance sheets. The creation of this emergency money has raised the proportion of liquid assets on bank balance sheets. For example, the reserves of US banks held with the Federal Reserve rose from only 1 per cent of their total assets before the crisis to just over 20 per cent by September 2015.28 This new higher level should be maintained, if necessary by central banks continuing to purchase government debt when existing holdings mature. Second, the infrastructure built up within central banks to assess and manage collateral during the crisis should be maintained as a permanent feature. This feature is already part of the regular operations of the Bank of England and the Reserve Bank of Australia.29

Sixth, regulation would be drastically simplified, comprising just two provisions: the PFAS rule that effective liquid assets must exceed effective liquid liabilities and a maximum value for the permitted leverage ratio. Most other capital and liquidity regulation could be abolished at the end of the transition. This would bring an enormous benefit in terms of simplicity and allow large amounts of extremely complex regulation to be discarded.

Each country could choose its own path to the end of alchemy, and there would be no need for international agreement on the details of banking regulation. The principle of the Basel regulations was to impose minimum standards on all countries, but the case against alchemy is a natural one, and countries should be allowed to adopt it at their own rate. The ability of a central bank to operate as either a LOLR or PFAS depends on the solvency of the state. A government that is insolvent, or cannot print its own currency, cannot easily support a banking system in a crisis. Regulation is gravitating naturally back to nation states because the provision of liquidity insurance inevitably involves fiscal risks. That is one reason why resolving the problems of the banks in the euro area is proving so fraught in the absence of fiscal and political union. It is likely that governments will increasingly feel that to be able to regulate banking activity appropriately, they must require all banks operating within their borders to be separate subsidiaries and not branches of foreign banks regulated overseas.

The essence of a successful pawnbroker is the willingness to lend to almost anyone against extremely valuable collateral. In 2008, banks had very few ‘gold watches’ and plenty of broken ones, and central banks were forced to lend against inadequate collateral in order to save the system. Before the next crisis it would be sensible to make sure that the banking system has sufficient pre-positioned collateral, including central bank reserves, to be able quickly to raise the funds to meet the demands of fleeing depositors or creditors who had decided not to roll over funding. Unlike a traditional high-street pawnbroker, central banks will want to take collateral that is more difficult and time-consuming to evaluate than gold watches. The PFAS rule is a strong incentive for banks to bring collateral to the central bank before a crisis.

The biggest problem in asking the central bank to act as a PFAS is the challenge of determining appropriate haircuts. If it is difficult to calculate risk weights for bank assets, why should it be easier to calculate haircuts? The answer is that the purpose of the two calculations is very different. For one, the aim is to compute the overall risk of a portfolio of different bank assets which requires knowledge of all the possible outcomes and the correlations of returns on different assets. For the other, the more limited aim is a rough and ready calculation of the discount at which an asset pledged as collateral could be sold when a bank could not repay the central bank, allowing for the fact that in a crisis the central bank would hold the collateral until more normal times had returned to financial markets. Once set, haircuts should remain unchanged for, say, three years and not be altered frequently. As with any pawnbroker, the central bank should be conservative when setting haircuts and, if in doubt, err on the high side. The size of the haircuts on different types of collateral are the equivalent of an insurance premium that banks are required to pay for access to liquidity on demand so that they are not exposed to runs and can cope with wholly unexpected shocks. They are the tax on alchemy. The size of that tax should reflect the cost to taxpayers of providing the implicit insurance in giving banks a call on liquidity on demand in return for pre-positioning collateral.

In a world of radical uncertainty there is no mathematical way of pricing such insurance. On most bank assets, other than reserves with the central bank, haircuts should be large. Under the Chicago Plan - a limiting case - they would be 100 per cent. With a PFAS, haircuts reflect the ability of the central bank to hold collateral while a crisis persists and dispose of it in more normal times. They will reflect the volatility and illiquidity of the assets, and once set, they cannot be altered during a crisis or they would not be a committed source of liquidity. So when setting haircuts on different assets, central banks need to be sure that they can absorb potential losses on collateral - although they would not need to liquidate quickly their holdings of assets brought to them in a general panic. Unlike the financial institutions requiring liquidity, a central bank could afford to wait until conditions returned to something closer to normal.

But that might still involve permanent changes in the relative prices of different assets, and haircuts need to take that risk into account. They are, therefore, likely to be much higher than in the normal commercial provision of short-term lending. And on some assets they may well be 100 per cent. In recent years, central banks have lent at haircuts ranging from only a few per cent to 60 per cent or more, depending on the type of asset in question. Setting large haircuts in normal times is the PFAS equivalent of taking the punchbowl away as the party is getting going. And it is important that central banks do not see their role as underpinning liquid markets in particular assets. It is not the role of central banks to subsidise the existence of markets that would not otherwise exist. For assets that are complex or obscure, a useful heuristic for setting haircuts would be to learn from Dennis Weatherstone of J.P. Morgan (see Chapter 4). If the central bank executive responsible for the management of collateral could not understand the nature of the asset in three fifteen-minute meetings, then the haircut would be 100 per cent.

From time to time banks will fail - indeed failure is part and parcel of a prosperous market economy. With a PFAS, banks at risk of failure would have a year in which to be reorganised. There would be no panic rescues over a weekend, no dramatic tales to retell in subsequent memoirs. Bank resolution - a special bankruptcy regime for banks - would be simpler than it is today because deposits (liabilities) and the collateral lodged with the central bank (assets) could be lifted out of the failed bank, and the deposits transferred to another bank together with the liquid reserves for which the collateral was pledged. That would enable the resolution authority to sort out the rest of the bank without serious disruption to its depositors.

In essence, when a bank fails it needs to be separated into its ‘narrow’ and ‘wide’ components. Sorting out the bankruptcy of a large, complex bank is a costly and messy business, as the failure of Lehman Brothers in 2008 showed only too clearly. People made careers out of overseeing the process, and it is still going on. It is important that regulators make banks recognise the true state of their balance sheet sooner rather than later. One lesson from the Finnish and Swedish banking crises of the early 1990s, and also from the decade-long problems in the Japanese banking system, is the importance of recognising losses early and promoting transparency about the true state of the balance sheet. A lack of transparency means that banks have the ability to drag their feet in recognising losses. That means that loans are tied up in businesses with few profitable investment opportunities, thus denying financing to companies with the ability to expand. It is a recipe for stagnation.

The debate about whether banks are ‘too important to fail’ boils down to a simple question. Are banks an extension of the state, as they are in centrally planned economies, or are they part of a market economy? If the latter, then to correct for the social costs they impose on society in a crisis, banks should be made to take out compulsory insurance through the PFAS and have sufficient ‘loss-absorbing capacity’, on the liability side of their balance sheets, to reduce the implicit taxpayer guarantee to bank creditors. Only equity finance guarantees that capacity.30 The introduction of a PFAS would aid the gradual evolution of expectations towards the view that banks will not be bailed out. It is a question of creating a banking and financial system in which governments feel little incentive to step in and bail out failing firms.

Before the crisis, banks used too little equity and owned too few liquid assets. The right response is to require banks to use more equity finance and meet the PFAS rule. A minimum ratio of equity to total assets of 10 per cent would be a good start, compared with the 3-5 per cent common today. A century ago the ratio for many banks was 25 per cent!31 If the amount of equity finance is low, then any item of adverse news makes it more likely that short-term creditors will desert the bank, and shareholders have an incentive to take risks in order to ‘gamble for resurrection’, because large losses will fall on creditors or taxpayers.

The PFAS rule is not a pipe dream. Some central banks have already moved in that direction. For example, the Bank of England has for some while encouraged banks to pre-position collateral as way of obtaining liquidity insurance. In the spring of 2015, the value of collateral pre-positioned with the Bank was £469 billion and the average haircut was 33 per cent. Together with reserves at the Bank of £317 billion, the effective liquid liabilities of the banking system were £632 billion, compared with £1820 billion of total deposits.32 There was still a substantial degree of alchemy, but around one-third of deposits were backed by ‘effective’ liquid assets and the idea of gradually eliminating alchemy through a PFAS is realistic. Alchemy can be squeezed out of the system by pressing from the two ends - by raising the required amount of equity and keeping central bank balance sheets, and hence bank reserves, at broadly their present level. That would allow the PFAS rule to complete the job. Far from being a radical and unrealistic objective, the elimination of alchemy could be achieved by building on actions that were taken during the crisis and the adoption of the PFAS rule. The idea is new; the means of implementation isn’t. There is a natural path from today’s ‘extraordinary’ measures to a permanent solution to alchemy.

The future of money

In a world of radical uncertainty, the demand for liquidity can change in an unpredictable and unexpected manner. Of the three roles for money described in Chapter 2, that of supplying liquidity is satisfied by the pawnbroker for all seasons. But do we really need money for the two other roles - to enable us to buy ‘stuff’ and to ensure a stable unit of account, or measuring rod, to value production? Could innovations in information technology make money redundant in respect of those two roles?

We no longer need cash to buy ‘stuff’ and even the use of cheques to make payments has been rapidly declining. We use electronic transfers instead. So should we stop issuing paper money? There would be some advantages. A large proportion of banknotes, especially those of large denominations, are used for illegal transactions, both to evade tax and for other criminal activities. In the United States, over $4000 in notes and coin circulate for every man, woman and child.33 In Japan, the figure is almost double that. More than 75 per cent of those holdings are in the form of notes of the largest denomination, the $100 bill and the ¥10,000 note.

There is clearly a strong demand for anonymity when making payments. Much of that has been eroded in respect of electronic payments with counter-terrorism surveillance and the introduction of regulation to prevent money laundering that requires the disclosure of large amounts of information to governments. So the demand for paper money is unlikely to disappear quickly, and anonymity for illegal transactions is the opposite side of the coin of individual privacy.

Nevertheless, electronic payments are the way of the future. Even old-fashioned bank robberies are diminishing - they almost halved in the US between 2004 and 2014 - to be replaced by an explosion of cybercrime.34 At present, electronic transfers simply move money from one bank account to another - convenient but not revolutionary - and banks then clear payments with each other through their own accounts at the central bank. In principle, two parties engaged in a transaction could instead settle directly by a transfer of money from one electronic account to another in ‘real time’.

A step in that direction was the creation of bitcoin - a ‘virtual’ currency launched in 2009, allegedly by one or more individuals under the pseudonym of Satoshi Nakamoto. Ownership of bitcoins is transferred through bilateral transactions without the need for verification by a third party (necessary in all other current electronic payment systems). Transactions are verified by the use of a software accounting system accessible to all users.35 The supply of bitcoins is governed by an algorithm embodied in the software that runs the system (with a maximum number of twenty-one million). If you can persuade someone to accept payment in bitcoins, then you can use them to buy ‘stuff’. The price of bitcoins in terms of goods and services, or currencies such as the dollar, is determined in the market. Without any public body setting the standard for bitcoins as a unit of account, their price is highly volatile - less than $1 when launched, a peak of over $1100 in December 2013, and back to below $400 in late 2015.36 With no one standing ready to redeem them in terms of any other commodity or currency, bitcoins are a highly speculative investment. They have no fundamental value: their price simply reflects the value that bitcoins are expected to have in the future.

The integrity of the algorithm determining the supply of bitcoins is vital. An indication of what can go wrong when confidence in that process is lost is the fate of a related venture, the auroracoin, a digital currency in Iceland. As an alternative to government-issued paper money, auroracoins were circulated in Iceland by a private entrepreneur in March 2014 through a ‘helicopter’ drop to every citizen listed on the national ID register. Within a few months they had lost over 96 per cent of their launch value.37 Moreover, as described in Chapter 2, with any private fiat money new entry can undermine the value of existing currencies. What is to stop some new group of programmers from launching a digital currency under the name ‘digidollars’? The aggregate supply of digital currency cannot be controlled by any one issuer, which is why governments have nationalised the production of paper currencies.

Digital currencies attract those who would like to make payments anonymously.38 As an innovation in payments systems, and hence as a means of payment, bitcoins have generated genuine interest. But as money, they are more akin to a form of digital gold - appealing to those who distrust governments to control the supply of money but highly volatile in value.39

But why use money to make transactions when computers offer the possibility to exchange goods and services for wealth? With high-speed electronic transfers it is becoming feasible to transfer stocks and shares and other forms of marketable wealth from the buyer to the seller instead of money, so enabling buyers and sellers to avoid holding some of their wealth in a form that earns little or no interest. Pre-agreed instructions embedded in computer algorithms would determine the sequence in which financial assets belonging to the purchaser were sold and used to augment the financial assets of the vendor, also in a pre-agreed sequence. Assets used in this way could be any for which there were market-clearing prices in ‘real time’. Someone buying a meal in a restaurant might use a card, as now, but the result would not be a transfer from their bank account to that of the restaurant; instead there would be a sale of shares from the diner’s portfolio and the acquisition of different shares, or other assets, to the same value by the restaurant. The key to any such development is the ability of computers to communicate in ‘real time’ to permit instantaneous verification of the creditworthiness of the buyer and the seller, thereby enabling private sector settlement to occur with finality. There would be no unique role for something called money in order to buy ‘stuff’.

Electronic transfers of wealth in ‘real time’ sound attractive because they economise on the use of money. But electronic messages that carry the instructions to make payments travel not instantaneously but at a rate bounded by the upper limit of the speed of light. Admittedly this is fast - around 186,000 miles (300,000 kilometres) a second - but not fast enough to avoid problems. Some professionals in financial markets, such as high-frequency traders, have invested heavily in microwave technology that is even faster than fibre-optic transmission to enable them to deploy the tactic known as front-running (see Chapter 4). Regulators are already concerned about this type of behaviour; imagine how much more serious front-running would be if all payments were made by selling and buying financial assets. Since transactions can never be literally instantaneous, I think it unlikely that we will move in the foreseeable future to a system of making payments that is entirely divorced from some form of money. Bank accounts, either to make anticipated payments or to hold a liquid reserve of generalised purchasing power, will be with us indefinitely, and so therefore will be the need for a pawnbroker for all seasons.

One way to reduce the demand for deposits, and to mitigate the scale of operations of the pawnbroker for all seasons, would be for central banks to allow anyone at all to open an account with them for purposes of making payments to others and to hold liquidity. At present, central banks could not cope with a large influx of customers, although new technology would make the task of handling so many accounts easier to imagine. It is fair to say that anonymity would be unlikely to be a characteristic of such a system. Such a development would simplify and reduce risks in the system of money transmission. International transfers could then be cleared through central banks. But it is unclear why, when we can use the pawnbroker for all seasons to provide liquidity to a competitive banking system, such a benefit would outweigh the costs of foregoing the advantages of competition in the provision of customer services. Whatever arrangements come to dominate how we make and receive payments in future, it is already clear that it will be necessary to guard the computer systems used for settlement purposes as carefully as the gold at Fort Knox is guarded today.

If we ever moved to a world in which we bought and sold ‘stuff’ by transfers of wealth, then neither money nor central banks, in their present form, would be needed.40 The need to control the supply of money would be replaced by a concern to ensure the integrity of the computer systems used for settlement purposes. Radical uncertainty means, however, that there will always be a demand for liquidity as a reserve of future purchasing power, and the ultimate source of liquidity is the central bank. Moreover, there is one additional role for a public body that it makes sense to give to the central bank - the regulation of the unit of account. Even if money disappears as a means of payment, we will always need a stable unit of account to price goods and services.41

There is an enormous advantage in all of us agreeing to use the same unit of account determined centrally rather than allowing a profusion of different monies. We rely on weights and measures inspectors to ensure that retailers who use yards (or metres) and pounds (or kilograms) as the units of length and weight define them in exactly the same way. It would be a great inconvenience if what was meant by a ‘yard’ was different in New York than in San Francisco or London. We can just about cope with the translation from yards to metres and from Fahrenheit to Celsius, but think how difficult it would be to order goods from different parts of the world if each locality had its own definition of the unit of length or weight. Converging on a common unit has enormous value, in the same way that a telephone is useful only if others can use the same network (if, in the jargon of economists, there are ‘network externalities’). It is striking that Article 1, section 8 of the United States Constitution conflated the standards for physical weights and measures with regulation of the currency: ‘The Congress shall have power … To coin money, regulate the value thereof … and fix the standard of weights and measures.’ As George Washington said in the very first annual presidential message to Congress, ‘Uniformity in the currency, weights, and measures of the United States is an object of great importance.’ And as declared in Magna Carta eight hundred years ago, ‘There shall be but one Measure throughout the Realm.’42

It may be that we could allow anyone to issue their own money, as advocated by the economist and Nobel Laureate Friedrich Hayek, trusting in competition to ensure that we would all decide to use the money of the ‘best’ issuer (the supplier most trusted not to abuse their ability to print money).43 But competitive monies have arisen rarely, and usually only in situations where government money is either absent (as, for example, in the use of cigarettes as currency in prisoner-of-war camps) or badly managed (as in periods of hyperinflation). Despite the abolition of foreign exchange controls, competition among national currencies has not reduced the dominance, within each economy, of a single public currency. There are exceptions, but they are few and far between. The abandonment of their own currency by Panama and Ecuador, and their adoption of the dollar (see Chapter 7), is one example. In Germany, Notgeld (see Chapter 5) was an example of a thousand flowers blooming - but they faded when a viable national currency was re-established. Network externalities make it difficult for competing currencies to emerge.44

A single unit of account requires collective decisions as to the definition and management of that unit. Unlike measures of length and weight, where a physical definition is determined and monitored by inspectors of weights and measures, the value of money requires a degree of discretionary management to avoid the costs of excessive volatility in output and employment. It is pre-eminently a task for a central bank given a mandate to meet an inflation target over the long term, as explored in Chapter 5.

Whatever happens to money as a way of buying ‘stuff’, it will always have a future as the only true form of liquidity. There will always be a demand for the liabilities of the central bank and for a stable unit of account. And so central banks will retain their ability to set interest rates and the size of their balance sheet. It may be tempting to imagine that technological innovation will mean that the successors to Bill Gates (the founder of Microsoft) and Steve Jobs (the founder of Apple), although not issuing their own currencies but instead providing a way of exchanging stocks and shares for goods and services, will put the successors to Ben Bernanke and Janet Yellen out of business. But the management of our system of money and banking requires collective decisions.

I hope that by now the reader will have been persuaded that only a fundamental rethink of how we, as a society, organise our system of money and banking will prevent a repetition of the crisis that we experienced in 2008 and from which previous generations suffered in earlier episodes. A major failing before 2007 was that the monetary policy framework designed to deal with good times and the lender of last resort framework for bad times were not properly integrated. In the crisis, massive support was extended not to save the banks but in order to save the economy from the banks. The role of pawnbroker for all seasons demands a great deal of central banks but, as with monetary policy, they would be exercising discretion within a clear framework designed to cope with radical uncertainty.

Before the crisis, hubris - arrogance that inflicts suffering on the innocent - ran riot, and changed the culture in financial services to one of taking advantage of the opportunity to manage other people’s money rather than acting as a steward on behalf of clients.45 But ‘the true steward never forgets that he is a steward only, acting for a principal’.46 The maxim ‘my word is my bond’, which underpinned the traditions of the City of London for many years, means little if those words are incomprehensible. The sale of complex financial products by people who only half understand the risks involved to those who understand even less is not an attractive advertisement for the financial services industry. What kind of person takes pride in parting a fool from his money?

I have explained the principles on which a successful reform of the system should rest. It is a programme that will take many years, if not decades. There is time to put in place such a programme to protect future generations. But the reform of money and banking will not be easy. Most existing financial institutions, and the political interests they support, will resist strongly. At the end of 2014, Paul Volcker criticised the ‘eternal lobbying’ of Wall Street.47 As Radical found when he left the City of Plunder in his quest for the City of Reform:

the party who had long enjoyed the privilege of putting their hands into their neighbour’s pockets, were composed of two sects, one called the ‘outs’ and the other the ‘ins’-the ‘outs’ maintained that there was no such city as that of ‘Reform’; while the ‘ins’ thought there might be such a city as ‘Reform’, but it was at such a long distance, and the road was so intricate and beset with brambles and thorns, that it was dangerous for anyone to set out on such a journey just at this peculiar time.48

Keynes’s optimism that ‘the ideas of economists and political philosophers, both when they are right and when they are wrong, are more powerful than is commonly understood’ has flattered academic scribblers ever since.49 It might be more accurate to reverse Keynes’s famous dictum about the influence held by economists over practical men to ‘economists, who believe themselves to be quite exempt from any practical influences, are usually the slaves of some defunct banker’.50 But there is nothing special about finance that requires us to abandon rational argument and leave our future in the hands of the gods of finance.