‘It’s not a question of one rotten apple, but a rotten barrel.’
Proverb
‘It should be clear that among the causes of the recent financial crisis was an unjustified faith in rational expectations, market efficiencies, and the techniques of modern finance. That faith was stoked in part by the huge financial rewards that enabled the extremes of borrowing, the economic imbalances, and the pretences and assurances of the credit-rating agencies to persist so long. A relaxed approach by regulators and legislators reflected the new financial zeitgeist.’
Paul Volcker, 20111
‘The long-term is what we are going to have for lunch.’
Trader’s view
‘Shadow banking, financial innovation, and intensive trading among financial institutions . . . gave us the credit cycle on steroids.’
Adair Turner, 20162
The brief answer to the question posed by this chapter is that from the 1960s onwards, governments gradually relinquished their control over banks and put their faith in market discipline; the banks used their new freedom to develop increasingly complex financial products to boost their profits, which they sold to non-banks and each other; these little-understood products, which were used mainly to finance real estate booms in the United States, Britain, Spain and Ireland, brought about the crash of the financial system starting in 2007, which spread to the real economy.
When, as an Oxford undergraduate, I opened an account at my local Midland (now HSBC) Bank in 1958, banking was safe and boring, like my estimable bank manager.
In those far-off days banking was divided into three sectors: retail banks, which took deposits and made loans to customers or businesses; building societies, which made loans for home buying; and merchant banks, which invested their clients’ money. The atmosphere of retail banking was summed up by the so-called ‘3-6-3’ rule: take in deposits at 3 per cent, make loans at 6 per cent, and be on the golf course by 3 p.m. Mortgage lending, done by building societies, was conservative, with borrowers having to put up substantial deposits. Merchant, or investment, banks were less ‘safe’, but in a repressed financial system, their activity was restricted. For example, exchange rates were fixed, so speculation in foreign currencies was small; the financial system was barnacled with capital controls. With tame banking, financial crises were rare, and system-threatening ones non-existent.
From the 1960s the trend was towards freeing finance from the straitjacket which the Great Depression and Second World War had placed it in. The development of the offshore euro-dollar market was an important breach in the wall of capital controls; in the 1970s banks were given the leading role in recycling OPEC surpluses. There was a relentless rundown in both capital and reserve ratios as the appetite for borrowing and lending grew, and the perceived risks fell.
As a result of deregulation, banks became ‘universal’. They were allowed to do everything: take deposits, lend to home-buyers and open investment departments. They became more concentrated, more remote from their customers. Capital was set free to roam the world. In the years of the Great Moderation, bank credit gained enormously at the expense of stock market issues as a source of business finance. Banks became part of an international financial network, holding each others’ assets. And these assets became more and more complex and opaque. Financial crises became more frequent and more severe.
In Ann Pettifor’s summary, orthodoxy handed two great powers to bankers:
first, the ability to create, price and manage credit without effective supervision or regulation; second, the ability to ‘manage’ global financial flows . . . out of sight of the regulatory authorities. By way of this shift . . . accountable public authorities handed effective control over the economy – over employment, welfare and incomes – to remote and unaccountable financial markets.3
Orthodox economists provided intellectual cover for the newly dominant vested interest.
The government’s role was not entirely permissive. Governments also encouraged banks to lend for political purposes. The root of the 2008–9 financial crisis lies in the American housing market and, specifically, in the government’s attempts to make home ownership accessible to low-income families. Building on the Community Reinvestment Act of 1977, President Clinton, in 1994, signed the National Homeownership Strategy, intended to help
moderate-income families who pay high rents but haven’t been able to save enough for a down payment; to help lower-income, working families who are ready to assume the responsibilities of homeownership but are held back by mortgage costs that are just out of reach; [and to] help families who have historically been excluded from homeownership.4
The government-backed institutions Fannie Mae and Freddie Mac played a large role in promoting the ‘American Dream’; over the years, they were ordered to increase the ratios of their loan portfolios in low-income areas. It was the combination of deregulation and government subsidy of bank credit to low-income households which proved toxic.
Behind the deregulation of banking were three propositions in theoretical economics.
Central bank models were based on a neo-classical fantasy world, with no financial frictions or default.5 Such a world has no need for commercial banks, and as a result central banks overlooked these earthly entities in their Platonic modelling. In the UK, for example, the Bank of England’s foremost macroeconomic model between 2004 and 2010 omitted the banking system from its grouping of key economic agents.
In order to keep the main model ‘as simple as possible’, the Bank preferred to look at financial issues through ‘separate, more specialised models’.6 But this approach placed theoretical elegance above understanding, and left the Bank unable to identify a burgeoning financial crisis in time, with an ensuing series of inaccurate forecasts and missed targets. The UK was not unique; central banks the world over paid scant attention to banks as a source of credit creation in their models, a fact Mervyn King has labelled ‘a source of embarrassment, both intellectual and practical’, for the profession.8
Figure 59. The Bank of England’s main economic model7
This oversight was particularly pernicious as it supported the misconception that central banks controlled the money supply. This led to a revival of the crude Quantity Theory of Money. The result was that few politicians were aware that every loan creates a deposit (money). Even bankers themselves did not fully understand their crucial role in the economy.9
Bad theory led to bad practice. The authors of Where Does Money Come From?, written precisely to clear up these muddles in money, hit the nail on the head:
A review of the arguments at the time makes clear that the theoretical support for such deregulation was based on the unrealistic assumptions of neoclassical economics, in which banks . . . perform no unique function and are classified as mere financial intermediaries just like stockbrokers. This does not recognise their pivotal role in the economy as the creators of the money supply . . . [S]ince the 1980s, bank credit creation has decoupled from the real economy, expanding at a considerably faster rate than GDP . . . evidence that an increasing amount of bank credit creation has been channelled into financial transactions. This is unsustainable and costly to society, as it amounts to resource misallocation and sows the seeds of the next banking crisis.10
One hugely important macroeconomic implication arises from treating banks as mere intermediaries between savers and investors. This is that markets determine a natural rate of interest which delivers full employment – precisely the claim which Keynes set out to refute.
The efficient market hypothesis (EMH), made popular by Eugene Fama (1970, 1976) is the application of rational expectations to financial markets. The rational expectations hypothesis (REH) says that agents optimally utilize all available information about the economy and policy instantly to adjust their expectations. The implication of this is that shares are always correctly priced on average, because investors adjust their buy/sell actions instantaneously and accurately to any newly released information.
Thus, in the words of Fama, ‘I take the market efficiency hypothesis to be the simple statement that security prices fully reflect all available information.’11
An ‘efficient’ market is defined as a market where there are large numbers of rational, profit-maximizers actively competing, with each trying to predict future market values of individual securities, and where important current information is almost freely available to all participants. In an efficient market, competition among the many intelligent participants leads to a situation where, at any point in time, actual prices of individual securities already reflect the effects of information based both on events that have already occurred and on events which, as of now, the market expects to take place in the future. In other words, in an efficient market at any point in time the actual price of a security will be a good estimate of its intrinsic value.12 [my italics]
There are different versions of the efficient market hypothesis. In its ‘weak’ form, investors make predictions about current prices using only historical information about past prices (like in adaptive expectations). In its ‘semi-strong’ form, investors take into account all publicly available information, including past prices. In its ‘strong’ or ideal form, investors take into account all information that can possibly be known, including insider information.
Several problems can be identified here. First, while it is recognized that share prices may fluctuate randomly round their ‘correct’ values, they will not do so permanently. If there happened to be over-valued or under-valued assets, the very action of investors trying to sell/buy them to make a profit would work as a self-correcting mechanism. This important feature of the efficient market hypothesis postulates that markets are self-correcting and thus self-regulating, with government attempts to improve on this bound to be distorting.13 In this way, the efficient market hypothesis is essentially the modern manifestation of Adam Smith’s ‘invisible hand’.
There is a paradox here. On the one hand, the theory says that there is no point in trying to profit from speculation, because shares are always correctly priced and their movements cannot be predicted. But on the other hand, if investors did not try to profit, the market would not be efficient because there would be no self-correcting mechanism. There is a joke about two economists who spot a $10 bill on the ground. One stoops to pick it up, whereupon the other interjects, ‘Don’t. If it were really $10, it wouldn’t be there anymore.’ Therefore, efficient markets actually depend on participants believing that the market is inefficient and trying to outperform it.14
Secondly, if shares are always correctly priced, bubbles and crises cannot be generated by the market. Systematic market failure – e.g. Alan Greenspan’s ex post explanation of the crash as ‘under-pricing of risk worldwide’ – cannot happen. As Eugene Fama himself put it: ‘I don’t know what a credit bubble means. I don’t even know what a bubble means. These words have become popular. I don’t think they have any meaning.’15
This attitude leached into policy: ‘government officials, starting with Alan Greenspan, were unwilling to burst the bubble precisely because they were unwilling to even judge that it was a bubble’.16 The EMH made the identification of bubbles impossible because it ruled them out a priori.
Thirdly, under market pressures, financial innovation would only lead to increased efficiency. There was naught to be gained from reining in the machinations of banks and financiers as they built up an increasingly complex system.
A dose of realism, or even a cursory knowledge of history, would have told these savants that markets do not work in this way.17 Moreover, there were sound theoretical reasons to distrust an unfettered financial system. Hyman Minsky, an economist whose work was completely ignored until after the crash, argued that financial stability leads inevitably to financial fragility, as optimism turns to ‘speculative euphoria’ and markets become ‘dominated by speculation about sentiments and movements in the market rather than about fundamental asset values’.18
But these arguments had no place in the neo-classical hegemony and so, despite its glaring theoretical gaps, the EMH became the intellectual underpinning of financial market deregulation.
Mark-to-market accounting aims to estimate the ‘fair value’ of an asset by reference to its current market price, rather than what it cost the investor to buy. If an investor owns ten shares of a stock bought for $4 a share and that stock now trades at $6, its mark-to-market value is 50 per cent more than its book value.
But mark-to-market accountancy offers an accurate measure of value only if markets never get it wrong. In fact they often do. During a boom, confidence pushes up the market prices of stocks and other assets. The ensuing increase in reported wealth further bolsters confidence, and encourages investors to take more risks and increase their lending and speculative activities. A feedback loop develops between investors’ perception of the economy and their decisionmaking, increasing the likelihood of a speculative bubble. Once this bursts, asset prices fall back down and investors find themselves ‘over-leveraged’. Similarly, mark-to-market accounting in a bust can put undue pressure on banks, as the panic and mistrust engendered by a downturn can mean that markets undervalue perfectly healthy assets. In this way, mark-to-market accounting increases volatility by artificially enlarging and contracting balance sheets. But its doompotential was blithely ignored.
Value at risk modelling was used by banks to assess the amount of risk they faced on their portfolios. A VaR measure takes a given portfolio, a time horizon and a probability level p, and spits out a threshold value of loss for that portfolio, representing a ‘realistic’ worst-case scenario. For example, if your portfolio has a one-day 1 per cent VaR of $1 million, this means that 99 per cent of the time your portfolio will not fall in value by more than $1 million over a one-day period. VaR measures were popular, as they condensed lots of risk modelling into a single, easily comprehensible figure.
VaR modelling is deeply flawed. It overlooks the worst risks by ignoring scenarios that are less likely to happen than some arbitrary threshold, lulling bankers (and regulators) into a false sense of security. As a prominent hedge-fund founder puts it: ‘This is like having an airbag that works all the time, except when you have a car accident.’20 VaR measures tell you nothing about the risks beyond this threshold, which can be the difference between losing $1 million and going bankrupt. Its ‘confidence intervals’ are totally unwarranted.
Figure 60. VaR modelling19
VaR measures assume that we have the correct probabilities for all future outcomes. This is a risible assumption. In August 2007, the CFO of Goldman Sachs claimed that they were ‘seeing things that were 25-standard deviation moves, several days in a row’,21 meaning that Goldman’s models estimated that a single day’s losses of this magnitude happened once every 1.309 x 10135 years. By way of comparison, this number is fifty-two orders of magnitude larger than the upper estimate for the number of particles in the universe.22 Either Goldman were dreadfully unlucky, or their models were crazy.
What explains this failure? First, VaR (and other financial models) were calibrated using short-term data collected exclusively from the relatively stable period immediately before the crash, which proved useless once the boom subsided. But, more fundamentally, humans do not face a knowable distribution of probable outcomes; we face genuine, radical uncertainty, or ‘unknown unknowns’. Undue faith in mathematical models, which cannot account for this sort of uncertainty, left the financial sector ignorant of the dangers it was running. The doomsday scenarios weren’t doomy enough.*
Banks borrow short to lend long. This makes them vulnerable to any collapse in the value of their assets. Since banks are the lifeblood of a modern economy, their lending and borrowing have generally been subject to state regulation. Banks have been restricted in whom they can lend to and how much they can borrow. They have been legally separated by the type of assets they are allowed to own, and the type of liabilities they are allowed to incur. These regulations are designed to ensure their solvency and liquidity.
A bank, like any business, is solvent when it has enough assets to cover its liabilities. If a bank’s debtors start to default on their loans, the value of the bank’s assets falls and threatens to make the bank insolvent. To guard against this was the purpose of capital adequacy regulations.
Liquidity, on the other hand, is the ability to meet one’s short-term obligations; the bank has to have access to enough cash to pay its depositors and other creditors on demand. There has been a relentless run-down in liquidity ratios. For example, in the mid-nineteenth century banks had to hold 60 per cent cash against liabilities. This was limited to 12.5 per cent in 1981. Then it was abolished. So banks increasingly relied on borrowing to meet claims on them.
A bank can be illiquid but solvent if it owns more in assets than it owes, but has a cash-flow problem where it cannot borrow cash or sell its assets in time to meet its payment obligations. A liquidity crisis is much less serious than a solvency crisis; temporary funds from the central bank can alleviate a liquidity crisis, but are of no use if the bank is insolvent.
Still, the two are somewhat related. In the period 2007–8, confusion over who was solvent and who was not meant that banks stopped lending to each other, drying up their principal source of liquidity; this led to the bank run on the UK’s Northern Rock in 2007. Similarly, illiquidity can force a bank into insolvency if its financing costs exceed the interest it receives on its assets, or if it has to ‘fire-sell’ its assets in order to pay its debts on time.
A bank’s leverage is the ratio of its debt to its equity. It can be expressed either as a ratio of assets to capital (say, 25:1) or as the percentage of assets that are backed by capital (4 per cent). Leverage ratios are thus almost identical to capital adequacy ratios, but without the risk-weighting element, which permits banks to hold less capital for supposedly safer assets. In the run-up to the crisis, banks relied on borrowing, and not their own equity, to finance their acquisition of assets, pushing up their leverage.
Banks have an incentive to maintain as high a leverage ratio as possible. Leverage magnifies the possibility of a gain; banks can expose themselves to more risk – and thus more reward – with a smaller amount of their own capital. In the good times, banks increase their leverage as both lenders and borrowers are optimistic about their undertakings.
However, leverage also magnifies the danger of loss. If banks are leveraged at 25:1, then a fall of more than 4 per cent in the value of their assets wipes out their capital. Thus, in a downturn, if the price of an asset that is widely held by hedge funds and banks falls, the institutions’ balance sheets worsen. Banks respond by selling off their assets in order to ‘deleverage’ themselves. This causes the price of assets to fall further, which kicks off another cycle of selling, and so on.
Immediately before the crisis, leverage (debt to equity) ratios of major banks reached over 30:1. This was by no means a historical anomaly. The difference this time was the extent of embedded leverage. While balance sheet leverage was comparatively low in the run-up to the crisis, even when adjusted for risk, embedded leverage was much higher. Embedded leverage measures a bank’s total exposure to risk compared to its equity holdings, whether or not this risk appears on its balance sheet. It was by holding assets off their balance sheet through the use of a range of innovations that banks increased their leverage beyond the regulated limits. As we shall see in Section V, it was the massive increase in embedded leverage that brought the financial system to grief.
By deregulation we mean a weakening of controls over the lending and borrowing activities of the banking system. Three important signposts on this ‘deregulatory’ road were:
1. The gradual dismantling of the capital controls that had limited the flow of money between countries, i.e. to whom it was possible in the world to lend money. The idea was to free up world savings so that they could be channelled to ‘their most productive uses across the globe’.23
Capital account liberalization was part of a broader project of financial globalization, managed by a number of international organizations with increasing power. Foremost among these was the International Monetary Fund. Dominated by the interests and free-market ideologies of rich countries, the IMF imposed a ‘neo-liberal’ agenda on developing countries and forced them to ‘open up’ their capital markets to the outside world.
As a result, foreign money flooded into poorer countries. Though some of it financed foreign direct investment in infrastructure, much of the money took the form of speculative capital flows. These flows, otherwise known as ‘hot money’, allowed financiers to make short-term profits by moving capital from country to country, but brought an unprecedented level of financial volatility to developing economies. Heeding the IMF’s commandments, many countries experienced financial crises accompanied by severe downturns, most notably in East Asia and Latin America.24 Recently, the IMF’s research arm has issued a sort of a mea culpa, recognizing that while the ‘growth benefits [of capital account liberalization] are uncertain, costs in terms of increased economic volatility and crisis frequency seem more evident’25 and that ‘full capital flow liberalization is not always an appropriate end-goal’.26 But this came too late, and long after economists working outside the orthodox position had come to the same conclusion.
2. Controls on types of bank lending within countries were stripped away. After the Great Depression, the US Congress passed the Glass–Steagall Act in 1933,27 which separated commercial and investment banking in order to protect key banking services (taking deposits, making everyday loans) from the risks of ‘casino banking’. The Clinton Administration overturned this in 1999 with the Gramm–Leach–Bliley Act, allowing both sets of activities to be done by the same bank. In the UK, Thatcher’s ‘Big Bang’ of the 1980s, which allowed banks into the mortgage market and commercial banks to merge with securities’ houses, had similar consequences, and the model of the do-it-all, universal bank emerged. Regulatory loosening led to the loosening of moral restraints. Whereas traditional banking focused on developing long-term relationships with customers and providing them with a good service, investment banking is marked by an emphasis on short-term opportunism and risk-seeking behaviour.28
The effect of the deregulation of lending was to undermine the public utility aspect of banks: they became free to do whatever financial business they wanted. They grew so large that their failure would have a devastating impact on the rest of the economy. And so, when the financial sector collapsed under the weight of its own greed, governments were forced to spend billions of dollars on rescuing them. The public ended up footing the bill for a crisis that wasn’t its fault. Most gallingly, banks receiving government support continued to pay out for executive bonuses, while the rest of the country put up with austerity in the name of fiscal prudence.
‘Too big to fail’ banks provide a textbook example of moral hazard, which arises when people take on excessive risk once they know they will not pay the price for their risk-taking if things go awry. If bankers know that the state will absorb their losses, they are free to gamble to their heart’s content and are rewarded for their efforts: ‘no industry has a comparable talent for privatising gains and socialising losses’.29
3. The weakening of capital and eligible collateral requirements for lending and borrowing. Basel I (1988) and Basel II (2003) set minimum capital requirements for making loans of different riskiness.30 The logic seemed compelling. Suppose a bank held 10 per cent capital against its loan portfolio. A 10 per cent default on its loans would wipe out its capital. Under Basel I, internationally important banks were required to hold 8 per cent capital against all their unsecured loans (e.g. to small businesses), and much lower percentages for secured loans (e.g. for residential mortgages). If the risk of default on certain loans was judged to be low (as was thought to be the case for mortgage-backed securities with a AAA credit rating), they were given a 20 per cent risk-weighting; the capital banks needed to hold against these loans was only 1.6 per cent of their value (20 per cent of 8 per cent). For AAA-rated mortgage-backed securities issued by Fannie Mae and Freddie Mac (the US government-subsidized mortgage-brokers) the required capital was 2 per cent.
The perverse effects of these attempts to mitigate the riskiness of lending is easy to see. Banks came to prefer mortgage-backed securities to any others, because the risk of cluttering up their balance sheets with non-performing loans seemed so small. Creditrating agencies gave them AAA ratings for the same reason. The underlying assumption that the risks of the various types of loans were known was rarely questioned.
In another key move by the US’s Security and Exchange Commission at the same time, AAA-rated mortgage-backed securities were made eligible as collateral for bank borrowing. So they could be both part of the liabilities against which banks had to hold capital, and part of the capital to hold against the liabilities.
The result of these various deregulatory moves was an explosion of bank debt to finance loans to eager home-buyers tempted by nugatory interest rates and deposit requirements.
There is never a one-to-one relationship between theory and policy. Policy is an art, and an impure one at that, heavily contaminated by ideology, political beliefs and special interest lobbies. Nor was deregulation simply one-way traffic. Often it was a matter of replacing direct by indirect controls, or replacing one set of regulations by another.
Nevertheless, there is clear evidence that the financial theories described in Section II above influenced the policy of bank deregulation. The UK’s Financial Services Authority was commendably honest about how the efficient market hypothesis furnished it with the following intellectual assumptions in the run-up to the crash:
• ‘Market prices are good indicators of rationally evaluated economic value.’
• ‘The development of securitised credit, since based on the creation of new and more liquid markets, has improved both allocative efficiency and financial stability.’
• ‘The risk characteristics of financial markets can be inferred from mathematical analysis, delivering robust quantitative measures of trading risk.’
• ‘Market discipline can be used as an effective tool in constraining harmful risk taking.’
• ‘Financial innovation can be assumed to be beneficial since market competition would winnow out any innovations which did not deliver value added.’31
From these, it followed that:
• ‘Markets are in general self-correcting, with market discipline a more effective tool than regulation or supervisory oversight . . .’
• ‘The primary responsibility for managing risks lies with the senior management and boards of the individual firms, who are better placed to assess business model risk than bank regulators, and who can be relied on to make appropriate decisions about the balance risk and return . . .’
• ‘Customer protection is best ensured not by product regulation or direct intervention in markets, but by ensuring that wholesale markets are as unfettered and as transparent as possible.’32
In short, the evolution of the regulatory system created an unprecedentedly permissive environment for the financial sector. Governments put their faith in the wealth-generating power of financial innovation, and ignored the danger that their products might become too opaque to understand. The structure of these instruments became so complex that their potential for damage became unmeasurable and untraceable. If market participants cannot understand the products they are trading, this undermines the whole edifice of efficient markets. When the crisis hit, uncertainty about the value of these products and the level of entanglement between different financial institutions brought the entire sector to its knees.
The rise in embedded leverage was made possible by securitization.
Securitization is the generic name for the use of financial engineering to transform illiquid assets into liquid securities. It is the process of bundling together illiquid assets such as car loans, student loans, credit card debt, mortgages and so on to form ‘asset-backed securities’ (ABSs) which are then sold to various investors for ‘cash’. All of these assets have in common the fact that they are associated with a cash flow – borrowers have to repay the loan which backs the security to the security’s buyer. The explosion of securitization was made possible by the huge increase in computing power,33 and was a classic case of banks creating credit money out of almost nothing.
The different kinds of ABS are:
This is a type of asset-backed security that is secured by a collection of mortgages. It made possible the banks’ ‘originate and distribute’ model of mortgage-lending, which played a fateful role in igniting the crisis. Instead of holding home loans on their books, banks created packages of different mortgage loans of different levels of riskiness, which were sold on to investors, often long term. Many of these securities were backed by American sub-prime mortgages – loans to high-risk borrowers.
The motivation behind pooling mortgages together was to decrease the default risk of the entire portfolio; i.e. to ‘diversify’ risk away and reduce the effect of ‘statistical outliers’. Portfolios, as a whole, were considered to be of little risk because their returns came from a wide variety of mortgage-owners. The default by any single borrower would not have an enormous impact on the portfolio as a whole, so the dividends to the investors were assumed to be backed by stable cash flows. Thus, mortgage-backed securities were given a low risk weighting and allowed to be used as collateral, as detailed above.
However, this relied on the assumption that defaults on mortgages are not highly correlated with each other, whereas at the time there was insufficient historical data on the rate of mortgage defaults, especially on sub-prime mortgages. It turned out that mortgage defaults were highly correlated, including geographically.* So in the run-up to the crisis, the risk of MBSS was significantly under-priced.
Collateralized debt obligations form a distinct but overlapping category from MBSs. CDOs can be backed by any form of debt – mortgages, corporate bonds, even other ABSs – and are split into ‘tranches’ of varying risk and maturity, so as to offer investors more choice. The top tranche, called the ‘senior’ tranche, was entitled to the first payments, although it yielded the lowest returns for the investor, due to being low risk. The bottom tranche yielded the highest return, but it would only be paid what was left over from paying the other tranches.
The development of credit default swaps (CDSs) massively increased the scope and destructive power of securitization. CDSs are similar to insurance policies in that one party (the buyer) pays a regular fee to another (the seller) who will pay out in the case of the loan defaulting. As such, CDS contracts were an additional way of removing risky assets from banks’ balance sheets and freeing up capital to be used elsewhere.
The key difference with a traditional insurance policy is that the buyer of a CDS does not necessarily have to hold a corresponding loan on its books, i.e. anyone can purchase a CDS, even buyers who do not hold the loan instrument. This is called a ‘naked’ credit default purchase. So an institution could purchase a CDS against another institution, even if it has not made a loan to it, essentially betting that it would default.
Clearly, this exacerbated the problem of moral hazard: banks were insured, so they did not have an incentive to prevent loans from defaulting. Rather, they had an incentive to lend to increasingly less creditworthy clients because these would pay high interest rates. Greek Prime Minister Papandreou compared purchases of naked CDSs to buying fire insurance on a stranger’s house and hoping that it will go up flames; their purchase incentivized financial arson.
The reason is that because naked credit default swaps are ‘synthetic’ – meaning that there is no underlying asset on the institution’s books – there is no limit to how many can be sold. In addition, these too can be securitized; e.g. the ‘Abacus’ CDO consisted of a portfolio of credit default swaps. As a result, in 2007, the gross value of CDSs far exceeded the ‘real’ value of the bonds that backed them: ‘At the peak of the CDS market in mid-2007, there was at least $60 trillion of CDS outstanding . . . The underlying bonds (against whose default the CDS provided insurance or on whose default the CDS permitted bets to be taken) were a small fraction of that $60 trillion.’34
In other words, the majority of CDS derivatives were bought to place bets, not for insurance purposes, and this made the financial market very unstable. It is also an example of the increase in intrafinancial intensity during the run-up to the crisis – the growth in trading activities between financial institutions greatly exceeded that of interaction with the real economy.
CDSs removed a large part of the risk from short-selling – betting on a stock going down in value – because they provided insurance against the bet turning out wrong. They thus encouraged speculation on the short-side. But while buying a CDS contract carried limited risk but almost unlimited profit potential, ‘selling CDS offers limited profits but practically unlimited risks’.35 In the run-up to the crisis, the banks made the crucial mistake of ruling out the possibility of the insurer having no money to pay the claims. This is what happened to the insurer AIG: it had to be bailed out for $182 billion by the Federal Reserve in 2008 due to not having enough capital to survive the wave of claims being made against it.
The fact that the insurers could themselves default meant that the banks had not effectively offloaded their risk via the use of credit default swaps.
Banks set up legal entities called Special Purpose Vehicles (SPVs) to hold risky assets off their balance sheets. Legally, the SPV, not the bank, was the issuer of the securities. The idea was that, once a bank transferred its risky assets to its SPV, it could effectively count them as off its balance sheet and take more risk – e.g. issue more loans – without breaching leverage rules. SPVs were what enabled banks to conceal their embedded leverage.
Asset-backed securities, including MBSs and CDOs, had to be moved off banks’ balance sheets using SPVs before they could be sold to investors. SPVs then repackaged asset-backed securities into different tranches to create even more complex securities. What investors failed to notice, however, is that even the safest ‘senior’ tranche was not risk-free. This is because SPVs engaged in ‘re-securitization’. Take the example of CDOs. An SPV could take tranches from two CDOs and bundle them together to form a new CDO of CDOs. The outcome was a new financial derivative called ‘CDO-squared’, which held the rights to the repayments of the assets of both original CDOs. And it was possible for this CDO-squared to be combined with another CDO, which itself might have been combined with another CDO, and so on. So one could have CDOs-cubed and greater. The result was a compound embedded leverage – it became impossible to measure the risk of a given CDO and know where the risk lay.
The funding of SPVs was itself problematic. SPVs usually funded their purchases of long-term assets (loans, mortgages and so on) by issuing short- to medium-term debt in the form of ‘asset-backed commercial paper’ (ABCP). The ABCP was seen as low risk by investors, because it was backed by the assets held by the SPV, and so the interest rate on it was low. SPVs thus used short-term loans with low borrowing costs to purchase long-term assets with higher rates of return to make a profit. And whenever the ABCP matured, the SPV would usually issue more ABCP to ‘roll over’ its debts.
But this maturity mismatch meant that SPVs were very vulnerable to liquidity crises. When demand in the commercial paper market dried up in 2008 because investors lost confidence, SPVs were forced to call upon other lines of credit: namely the ‘revolving credit line’ set up by the parent bank. This meant that the parent bank was strained twice during the crisis: it incurred losses on its own balance sheet, and it had to pay out the losses incurred by its SPVs.
As an example of the complexity involved in analysing some CDOs, the ‘Aquarius’ CDO structure had a total of 180 issues behind it. Each of these issues had, on average, 6,500 loans at origination, so the Aquarius CDO had exposure to about 1.2 million loans.
Hence Warren Buffett’s cry that derivatives are ‘financial weapons of mass destruction’.36 Their structure became so complex that risk became unmeasurable and untraceable. The whole financial system came to depend on the proliferation of these ‘unknown unknowns’.
Credit rating agencies (CRAs) became notorious for awarding AAA ratings to bank loans soon to be defaulted. Financial derivatives were given top ratings although they were far more risky than, say, German government bonds given the same endorsement. Many reasons have been adduced for these false prospectuses (see Appendix 11.1). But the principal one is that the agencies were paid by the issuers of the securities (the banks), not by the investors.
Figure 61. Securitization issuance trends in the UK37
These misleading ratings contributed to the crisis. The Asian Development Bank Institute explains: ‘The growth of the international financial markets over the last twenty years would have been unthinkable without CRAs. Only because of the availability of clear, internationally accepted indicators of the risk of default were investors willing to invest in international securities.’38 CDOs and other derivatives were so complex that investors essentially became overly reliant on credit rating agencies. The rating agencies also contributed to the volatility of the market by giving over-generous ratings in the upswings and overly pessimistic ones in the downswings.
Figure 61 shows the huge volume at which securities were sold in the UK, especially ABSs, MBSs (both residential and commercial), CDOs and collateralized loan obligations (CLOs).
The complex web of financial instruments created during the years of the Great Moderation was supposed to immunize the self-regulating financial system from the danger of collapse. How wrong could one be! Boom and bust in real estate were central to the US sub-prime mortgage crisis of 2007, the Irish crisis of 2010 and the Spanish crisis of 2012. Major banks in Belgium, France, Germany, Ireland, the Netherlands, Switzerland and the UK all faced insolvency as the price of their mortgage-backed assets fell below the price of their liabilities.
The opaqueness of the financial tools multiplied opportunities for fraud: the doyen of financial probity, London, became the world centre of money-laundering. Crooked auditors and lawyers were employed at large salaries to hide tainted money in special accounts. Since there was no anticipation of collapse, no precautions were taken against it, or thought given to the consequences of financial collapse and bank bail-outs for the finances of governments. The huge expansion of budget deficits caused by the collapse was key to the disablement of fiscal policy we have already discussed in Chapter 8.
The main theoretical mistake behind securitization was the assumption that securities are always liquid: they can always be sold quickly and without (much) loss. This being so, securitization spreads risk round the system; the spread of risk reduces the risk faced by any individual institution; this enables them to lend more; this enables the public to get more and cheaper credit. And cheap credit, as we have seen, was the substitute for rising wages and social security.
The relationship between theory, practice and policy is a perennial issue in political economy. In the case of banking, this relationship turned toxic, as neo-classical economics, deregulation and financial ‘innovation’ worked together to precipitate financial crisis. The efficient market hypothesis discounted the possibility of financial crises happening. Regulators turned a blind eye to the build-up of stress in the banking system because they believed in the efficient market hypothesis. Banks used their new freedom from control to create evermore opaque financial instruments (‘securitization’). All three pledged themselves to the service of mankind.
Spiralling out of control, the financial system collapsed. Since banks are the source of the credit on which the economy depends, the economy collapsed in their wake.
Credit ratings agencies were meant to provide serviceable estimates of debtor risk, but failed spectacularly in their task.
The principal reason is that they were paid by the issuers of securities (the banks) and not by investors. Banks shopped around for the best credit rating, which put CRAs under commercial pressure to rate banks’ products favourably. This gave issuers undue power over their own assessments. While minimizing the risk of banking default before the crisis, the rating agencies exaggerated the risk of government default after it. They were servants of the vested interests that paid them.
Further, the market is dominated by three firms: Moody’s holds 40 per cent market share, Standard & Poor’s 40 per cent and Fitch 14 per cent. Barriers to entry are very large, mostly because a CRA needs to have a good reputation to succeed. So the agencies are under little pressure to rethink their methodology; more importantly, concentration creates a single mindset.
CRAs have been accused of outright ‘blackmail’ to solicit new business. A case in point is the scandal involving Moody’s and Hannover Re, a German insurance company. Moody’s provided the latter with an ‘unsolicited’ rating and issued it a letter saying that it ‘looked forward to the day Hannover would be willing to pay’. When Hannover refused, Moody’s continued providing ratings and downgraded them gradually over the years. In 2004, Moody’s rated Hannover’s bonds as ‘junk’, and this caused the firm to lose $175 million in market capitalization, despite its buoyant assessments from other agencies.39
Apart from the fact that they were paid by the sellers of the securities they were rating, the following technical factors contributed to the agencies’ poor performance.
1. CRAs failed to acknowledge the crucial distinction between highly complex structured financial products and simpler corporate and government bonds.
2. They lacked sufficient historical data, especially relating to mortgages.
3. They under-estimated correlation in defaults. As a result, CRAs thought that securitized products (backed up by portfolios of loans) carried little risk, and indeed that they were less risky than these underlying loans considered individually. Many collateralized debt obligations were rated AAA when they were backed by a pool of only, say, BBB-rated loans; the strength of the CDO was supposed to come from its structure. Furthermore, the CRAs assumed a maximum 5 per cent decline in national housing prices.
4. Combined with the above, the use of normal distributions in VaR models privileged ‘thin tails’ (for which an extreme event is unlikely) over ‘fat’ ones (for which an extreme event is much more likely).
5. CRAs failed to allow for the impact of moral hazard on lending standards.
6. Giving the financial sector access to their rating methodology made it easier for issuers to ‘structure to the rating’. The UK Financial Services Authority’s Turner Review of 2009 says: ‘the practice of making the [CRAs’] models . . . transparent to the issuing investment banks also created the danger that issuers . . . were designing specific features of the structure so that it would just meet a certain rating hurdle’.40
* The systemic under-estimation of risk continues. The latest bank doomsday shock scenarios seriously under-estimated the exchange rate swings caused by the real-life events of Brexit and the election of Donald Trump.
* Defaults in one area were presumed to be uncorrelated with defaults in another, which ceases to be the case in the event of a national housing market crash.