The Doom Loop

Andrew Haldane writes about equity and the banking system

In 1989, the CEOs of the seven largest banks in the US earned an average of $2.8 million, almost a hundred times the annual income of the average US household. In the same year, the CEOs of the largest four UK banks earned £453,000, fifty times average UK household income. These are striking inequalities. Yet by 2007, at the height of the financial sector boom, CEO pay at the largest US banks had risen nearly tenfold to $26 million, more than five hundred times US household income, while among the UK’s largest banks it had risen by an almost identical factor to reach £4.3 million, 230 times UK household income in that year.

How do we make sense of these salary increases? Easily, in fact. During the go-go years, bank profits reached spectacular highs. Bank shareholders remunerated managers for delivering these riches; CEO pay grew almost exactly in line with shareholder returns. Reality then intervened. The heart-stopping global recession of the last few years was largely induced by financial sector excess. The long-term costs of the crisis are likely comfortably to exceed a year’s global income.

The continuing backlash against banking, as evidenced in popular protests on Wall Street and in the City of London, is a response not just to the fact that the world is poorer, as pre-crisis riches have turned to rags, but to the way these riches were privatised, while the rags are being socialised. This disparity is nothing new. Neither, in the main, is it anyone’s fault. For the most part the financial crisis was not the result of individual wickedness or folly. It is not a story of pantomime villains and village idiots. Instead the crisis reflected a failure of the entire system of financial sector governance.

In the first half of the 19th century, the business of banking was simple. The UK had around five hundred banks and seven hundred building societies. Most of the former operated as unlimited liability partnerships: the owners-cum-managers backed the banks’ losses with every last penny of their own personal wealth. The building societies operated as mutually owned co-operatives, with ownership, control and liability all pooled. Financial sector assets amounted to less than 50 per cent of annual UK GDP.

Banks’ balance sheets were heavily cushioned. Shareholder funds – so-called equity capital – protected depositors from loss and often accounted for as much as half of the balance sheet. Cash, and liquid securities such as government bonds, enabled banks to meet their payment obligations to depositors. They accounted for about a third of banks’ assets. Banking systems maintained broadly similar arrangements across the US and Europe. This relationship between governance and balance sheet was mutually compatible. Owing to unlimited liability, control was exercised by investors whose personal wealth was on the line – a potent incentive to be prudent with depositors’ money. Bank directors – the major shareholders responsible for day to day management – excluded investors who didn’t have sufficiently deep pockets to bear the risk. Shareholders were firmly on the hook, and had a strong incentive, in turn, to make sure that managers didn’t step out of line. Managers monitored shareholders and shareholders managers. In this way, the 19th-century banking model kept risk-taking in check.

The global environment was changing, however. During the first half of the 19th century, rich countries were hungry for capital to finance investment in infrastructure, including railways. As long as capital in banks was restricted to a small number of unlimited liability partners, credit was constricted. But in 1826, the six‑partner restriction on UK banks was lifted, allowing them to operate with an unlimited number of shareholders as joint-stock banks. Ownership and control became legally distinct – a crucial historical shift. Nevertheless, shareholder discipline was still proving an effective brake on risky lending and there was pressure for further liberalisation. One obvious solution was to limit shareholder liability so that their losses were no greater than their initial investment. Once investors’ personal wealth was no longer in jeopardy, bank credit would free up. In the UK, the change came with the Limited Liability Act and Joint Stock Companies Acts of 1855 and 1856.

At first, limited liability status was not taken up enthusiastically by banks: they were reluctant to give up unlimited liability, which they regarded as a badge of prudence. But the collapse of the City of Glasgow bank in 1878, caused by speculative lending and false accounting practices, ended that. Eighty per cent of the bank’s shareholders were made destitute. The opinions of bankers, Parliament and public alike shifted quickly. By 1889, only two unlimited liability British banks remained.

Even then, some of the benefits of unlimited liability were preserved. UK banks moved to a regime of extended liability, whereby shareholders were liable for an additional fixed level of reserve capital in the event of stress or bankruptcy. This made for a deep pool of contingent capital, often amounting to 50 per cent of a bank’s assets. Extended liability was enough to keep bank shareholders and managers on their toes: risk monitoring incentives remained sharp and the appetite for risk blunt. But the balance started to tilt at the beginning of the 20th century. As banks grew in size, the vetting of shareholders became impractical. Then, during the Depression, the need to draw from the contingent capital pool had the effect of heightening panic among UK and US banks, rather than diminishing it. By the end of the 1930s, only six British banks still maintained reserve liability. The governance and balance sheets of banks were, by this time, unrecognisable from those a century earlier. Banks were now controlled by arms-length managers, no longer major shareholders, while ownership was held by a widely dispersed set of shareholders, unvetted and anonymous, their upside pay-offs unlimited but their downside risks now capped by limited liability.

What impact did these changes have on banks’ incentive to take risks? The answer was provided in 1974, around a hundred years after the introduction of limited liability, by the Nobel Prize-winning economist Robert Merton, who showed that the equity of a limited liability company could be valued as if it were a financial option – that is, an instrument which offers rights over the future fruits of the company’s assets. This option has value – in the jargon, it is ‘in the money’ – provided a firm’s assets cover its debts. But the most extraordinary implication of Merton’s framework is that the value of those options can be enhanced by increases in the degree of uncertainty about the value of the bank’s assets. How so? Because while uncertainty increases both upside and downside risks, downside risks are capped by limited liability. For shareholders, the sky is the limit but the floor is always just beneath their feet. To maximise shareholder value, therefore, banks need simply to seek bigger and riskier bets.

The response to these incentives has been entirely predictable. Since 1880, the ratio of UK bank assets to GDP has risen roughly tenfold, and the increase has been particularly steep over the past thirty years, peaking at well over 500 per cent of GDP. The pattern in other developed countries has been similar, if less dramatic. The bets weren’t just bigger, but also riskier. During the 20th century, an alphabet soup of exotic and complex instruments, often known by three-letter acronyms, came to displace simple loans on banks’ balance sheets. These boosted banks’ returns. But if returns are high, risks are never far behind. Returns on bank assets were two and a half times more volatile at the end of the 20th century than at the beginning.

Finance has a further trick up its sleeve, a trick that at a stroke boosts both volatility and returns to the owners of a bank. Leverage, simply put, is borrowing against your capital stake. For example, if borrowing allows a bank to hold assets of 120 against capital of ten, then its leverage is 12. The beauty of leverage is that it effortlessly multiplies the amount shareholders receive as a return on their assets. Consider a bank that makes a 1 per cent return on its assets. By allowing leverage (assets relative to equity) of two, shareholders can double their money; with leverage of four, they can quadruple their money. And so on. Banks have been using this device for well over a century. As unlimited liability was phased out, leverage among banks rose from about three or four in the middle of the 19th century to about five or six at its close. Leverage continued its upward march when extended liability was removed, and by the end of the 20th century it was higher than twenty. In 2007, at its high-water mark, bank leverage hit thirty or more.

This strategy translated, by the arithmetical magic of leverage, into higher shareholder returns. Having begun the 20th century in modest single figures, equity returns to banks were, on average, close to 20 per cent by its end. At the height of the boom, bank equity returns touched 30 per cent. Higher leverage accounted for almost all of this. Bank managers no longer had to sweat their assets: they simply had to borrow against them.

The downside of this strategy is now only too clear. With leverage of two (UK banks in 1850), 50 per cent of your assets must go bad before your equity is wiped out and you go bust. But with leverage of twenty (UK banks in 2000), you will go bust if you lose only 5 per cent of your assets. Over the last hundred years, as returns to banking have increased so too has their volatility, rising by a factor of between six and sevenfold. In the recent financial crisis, UK banks’ shareholder returns fell from twenty-something to below zero in the space of a year.

In principle, market discipline ought to form a natural counterweight to these balance-sheet risks. Debt-holders in a bank, including depositors, ought to worry about shouldering increased risk, and should respond by raising the cost of funds or restricting their quantity, thereby restraining risk-hungry, excess-profit-seeking shareholders. During the 19th century, that theory fitted the facts. Depositor flight and bank runs followed when banks were perceived as fragile. But as the 20th century progressed, evidence of debtors exerting discipline over managers became increasingly patchy. Nowhere was the ineffectiveness of market discipline better illustrated than in the run-up to the recent financial crisis. At the same time as they were leveraging themselves up to the hilt, banks traded in debt markets as though they were riskless. Debtors should act as a brake on risk-taking, but in practice they served as an accelerator. The reason, once again, lies in incentives. When they face a crisis, it is dangerous for banks to have debtors take a hit. To do so may scare the horses, risking a stampede of deposits out of the door. Debtor discipline then has the effect of making a bad situation worse. Extended liability was abolished for just that reason. And the complex debt instruments issued by banks a hundred years later buckled under the same pressure.

In fact, making debtors shoulder the burden of risk in a crisis may have become harder over the past century. The structure of banking has been transformed during that time, in particular by the emergence of financial leviathans considered ‘too big to fail’. At the start of the 20th century, the assets of the UK’s three largest banks accounted for less than 10 per cent of GDP. By 2007, that figure had risen above 200 per cent of GDP. When these institutions hit problems, a bad situation can become catastrophic. In this crisis, as in past ones, catastrophe insurance was supplied not by private creditors but by taxpayers. Only they had pockets deep enough to refloat banks with such huge assets. This story has been repeated for the better part of a century and a half; in evolutionary terms, we have had survival not of the fittest but the fattest. I call this phenomenon the ‘doom loop’.

Consider the effects of the too-big-to-fail problem on risk-taking incentives. If banks know they will be bailed out, those holding their debt will be less likely to price the risk of failure for themselves. Debtor discipline will therefore be weakest among those institutions where society would wish it to be strongest. This encourages them to grow larger still: the leverage cycle isn’t merely repeated, but amplified. The doom loop grows larger. The biggest banks effectively benefit from a disguised, and growing, state subsidy. By my estimate, for UK banks this subsidy amounts to tens of billions of pounds per year and has often stretched to hundreds of billions. Few UK government spending departments have budgets this big. For the global banks, the subsidy can reach a trillion dollars – about eight times the annual global development budget.

We have arrived at a situation in which the ownership and control of banks is typically vested in agents representing small slivers of the balance sheet, but operating with socially sub-optimal risk-taking incentives. It is clear who the losers have been in the present crisis. But who are the beneficiaries? Short-term investors for one. More than anyone else, they benefit from a bumpy ride. If their timing is right, short-term investors can win on both the upswings (by buying) and the downswings (by short-selling) in financial prices. Bank shareholding has become increasingly short‑term over recent years. Average holding periods for US and UK banks’ shares fell from around three years in 1998 to around three months by 2008.

Bank managers have benefited too. In joint-stock banking, ownership and control are distinct. That means managers may not always do what their owners wish. They may seek to feather their own nests by making decisions that boost short-term profits and thereby justify an increase in their own pay. Such decisions may also increase banks’ vulnerability to shocks. In an attempt to avoid this problem, shareholders have sought to align managerial incentives with their own. One way of doing that, increasingly popular over the past decade, has been to remunerate managers not in cash but in equity or using equity‑based metrics. This can generate hugely powerful pecuniary incentives for managers to act in the interests of shareholders. At the peak of the boom, the wealth of the average US bank CEO increased by $24 for every $1000 created for shareholders. They earned $1 million for every 1 per cent rise in the value of their bank. But such equity-based contracts also set up some peculiar risk incentives. In the 19th century, managers monitored shareholders who monitored managers; in the 21st, managers egged on shareholders who egged on managers. The results have been entirely predictable. Before the crisis, the top five equity stakes were held by the CEOs of the following US banks: Lehman Brothers, Bear Stearns, Merrill Lynch, Morgan Stanley and Countrywide. We know how these disaster movies ended.

The evolution of banking as I have described it has satisfied the immediate demands of shareholders and managers, but has short-changed everyone else. There is a compelling case for policy intervention. The best proposals for reform are those which aim to reshape risk-taking incentives on a durable basis. Perhaps the most obvious way to tackle shareholder-led incentive problems is to increase banks’ equity capital base. This directly reduces their leverage and therefore the scale of the risks they can take. And it increases banks’ capacity to absorb losses, reducing the need for taxpayer intervention. Over the past few years, this case has been pushed by regulatory reformers. Under the so‑called Basel III agreements struck in 2010, banks’ minimum equity capital ratios will rise fivefold over the next decade, from 2 per cent to close to 10 per cent of assets for the largest global banks. That is a significant shift. Will it be enough?

Recent academic studies suggest not. A 10 per cent capital ratio translates into bank leverage of roughly 25. So even once Basel III is in place, an unexpected loss in a bank’s assets of just 4 per cent will be enough to render it insolvent. History is full of such unexpected bad news, from wars to shocks in oil prices. Basel III is a good starting point, but may not be the finishing line.

An alternative, more radical approach would be to tackle the problem of bank governance and control. Since the crisis, a number of proposals have been made – increasing board expertise and the power of risk committees, for example. These are steps in the right direction. But one look at the star‑studded cast of non-executive directors on the boards of failed financial institutions is to realise that to wish for better is to wish upon a star.

What else might be done? As at the start of the 20th century, the prevailing ownership and control models in banking were the public limited company and the mutually owned co-operative. Under the first, ownership and control are vested in a small minority of stakeholders, with rights assigned according to weight of portfolio: it is an equity dictatorship model. Under the second, ownership and control are vested in a much wider set of stakeholders, with rights unrelated to weight of portfolio – a stakeholder democracy model. Both have their problems. For the public company, it is risk and profit-seeking by the minority. For the mutual, it is a lack of financial incentive to curb risk, as those with the majority voting rights typically have the smallest financial stakes. But it isn’t difficult to conceive of governance models that combine the best bits of both.

To give one example, voting rights could be extended across a wide group of stakeholders, but weighted by stake. Governance and control would then be distributed across the whole balance sheet, curbing the profit-seeking incentives of the equity minority, while weighting voting rights by size of portfolio to avoid the inertia of mutuality. Bank governance would then be a wealth-weighted democracy, a hybrid of the mutual and joint-stock models. This would seem like a radical departure. But in many respects it would be a return to the incentive-aligned structure of 19th-century banking. Over the past century, the imbalance between those who have garnered the profit from banking and those who have taken on the risk has increased. It isn’t only banks that have gone bankrupt during this crisis. Households, companies and even countries have borne the brunt. Putting equity, social and financial, back into banking is essential if the financial system is to be durably repaired.