The Financial Crisis of 2008
It’s awful. Why did nobody see it coming?
—Her Majesty Queen Elizabeth II1
The financial crisis of 2008 has had lasting effects on economic output and employment. Although economic growth has now returned to normal in the United States, with unemployment down to 5 percent and economic confidence restored, it took many years for output to recover to its precrisis level. Europe—admittedly struggling with other difficulties in addition to the financial crisis—is stuck in a complex economic situation with large-scale unemployment in the countries of southern Europe. The crisis has also placed a heavy burden on public finances, diminishing the ability of governments to intervene in any future crises.
August 9, 2007, was the date of the first intervention by the Federal Reserve and the European Central Bank. No one that day—economists included—remotely imagined that whole swaths of the banking system were going to be bailed out by governments. Or that five of the largest investment banks would disappear in their existing form (Lehman and Bear Stearns disappeared completely, Merrill Lynch was bought by the Bank of America, and Goldman Sachs and Morgan Stanley survived by asking to become regulated retail banks in order to receive liquidity assistance). Nor could anyone have predicted that successful commercial franchises like Citigroup, the Royal Bank of Scotland, and Union Bank of Switzerland (UBS) would need government support as a result of foolish risk taking; that an insurance company and two institutions guaranteeing real estate loans would receive around $350 billion from the US government; that a little more than a year later, the US government would have committed the equivalent of 50 percent of its GDP to the recovery efforts; that US and European governments would loan substantial sums directly to industry; or that central banks would use unconventional monetary policies, enter into an era of extremely low interest rates, and go far beyond their mandate by supporting governments.
In Europe, the UK, Belgium, Spain, Iceland, and Ireland all experienced massive banking problems.2 Some other countries, such as France, the Scandinavians, and Japan, fared relatively well in this context, although some of their banks benefited from the US taxpayer bailouts of institutions such as AIG, to which they were exposed. Was this relative stability due to the lessons learned from earlier errors, which led in the 1990s to the bankruptcy of Crédit Lyonnais in France and widespread banking crises in the others?
What caused the financial crisis? Have we learned its lessons? Are we safe from another? To tackle these questions, I will begin with an analysis of its causes before discussing the postcrisis situation. Finally, I will examine the responsibilities and role of economists in preventing crises. This chapter, which is a little more technical than the others, is also the only one that is not completely stand-alone: reading the preceding chapter, though not essential, would be advisable.
THE FINANCIAL CRISIS
There are many excellent accounts of the economic crisis,3 and here I can touch only briefly on the subject. One thing is for sure: the 2008 crisis is a textbook case for the theory of information and incentives courses taught in economics departments. At every link in the chain of transfers of risk, one of the parties had more information than the other (asymmetry of information), and this distorted the proper functioning of financial markets and their regulation.
Market failures due to asymmetries of information are a constant, although the introduction of new and often complex financial instruments, and participants’ and regulators’ lack of familiarity with them, certainly made the information failures worse. So we cannot explain the crisis by this cause alone. There were two other factors contributing to the market failures. First, inadequate regulation and a laxity in enforcement—especially in the US, where the crisis originated, but also in Europe—created incentives to take risks, ultimately at the expense of the public. Second, market failures and lax regulation had a bigger impact because the context had never been so favorable to taking risks.
AN EXCESS OF LIQUIDITY AND A REAL ESTATE BUBBLE
Crises often have their origin in a lack of discipline during good times. The Federal Reserve’s maintenance of abnormally low interest rates for several years early in the new century provided very cheap liquidity—at times the short-term rate was 1 percent. When combined with investors’ desire to find yields superior to the low market interest rates, this monetary policy fed the real estate boom.4
What’s more, in the decade before 2008, there was an influx of money looking for investment opportunities to the US. Financial markets in the United States are highly developed. They create many tradeable securities, which makes them attractive for people with money to invest. Some of the surplus savings from sovereign wealth funds in the Middle East and Asia, plus foreign exchange reserves accumulated by export surplus countries (such as China) were invested in the United States because they could not be recycled in their own financial markets. This international savings glut enabled financial intermediaries to invest in real estate. In turn, this strong demand for securities, along with favorable regulatory treatment not corrected until after the crisis, encouraged the securitization of debt. These macroeconomic conditions created a permissive environment that encouraged actors to plunge into the breaches created by market and regulatory failures.
Abundant liquidity and low interest rates led to a big increase in risky real estate loans granted to US households with only limited ability to repay them;5 these loans usually had a very low fixed interest rate for the first two years, followed by a variable interest rate with a high margin levied by the lenders. The lenders also often failed to verify the information (for example, about income) provided by prospective borrowers.6 The subsequent stagnation in real estate prices, which precluded the easy refinancing associated with rising property values, and a rise in interest rates in the mid-2000s led to defaults on these loans with upward-varying interest rates. Many households simply could no longer make their loan payments, whereas others, protected by American personal bankruptcy laws, decided to walk away from their loans when the market value of their houses fell below the outstanding balance on their mortgage. The danger posed by risky real estate loans was thus that a macroeconomic downturn might lead to repossessions, the eviction of property owners, and major losses for lenders when the property was put back up for sale on the market. The losses suffered by lenders were all the greater because other lenders were doing the same thing, leading to a decline in real estate prices.7
What was the US government’s reaction to this real estate lending? It took the same political decision that led to a banking crisis in some other countries (like Spain at the same time). The US administration encouraged more households to become home owners. During the 2000s, the government let the real estate bubble inflate, and, more damaging, allowed its banks to become exposed to it. It would have been better to reduce the tax subsidies available for home purchase (the tax deductibility of interest paid on real estate loans) and the implicit guarantees on home loans provided by semipublic agencies such as Freddie Mac and Fannie Mae. It would have been advisable to impose stricter borrowing criteria, such as capping the real estate loan-to-value ratio and the borrower’s annual debt repayment-to-income ratio. But the political imperative won out.
Of course, risky lending made it possible for less well-off people to become home owners. But many of these households lacked a proper understanding of the risks they ran if interest rates went up, or if real estate prices stalled, preventing them from taking out a new loan to cope with higher payments. Institutions making real estate loans have always played on households’ desire to own property to sell risky mortgage contracts. At a minimum, the US federal government should have helped ensure that there was symmetric information between lenders and borrowers about the risks, since few states regulated the conditions on mortgages or discouraged abusive practices.
The government can tackle the asymmetry of information between lenders and borrowers by giving additional information to borrowers, whose freedom of choice is then respected. Alternatively, borrowers can be protected in a more “paternalistic” way, although there are some dangers inherent in this approach. The justification for paternalism is that people yield to temptation and tend to overconsume compared to what they would choose in a cooler state of mind. This rationale underlies some kinds of policy interventions (see chapter 5). From this point of view, the state should set ceilings on the loan-to-value ratio on a property and on the borrower’s annual debt-repayment-to-income ratio and prohibit seemingly attractive “teaser rate” loans with very low promotional interest rates during their early years.
EXCESSIVE SECURITIZATION
Recall that good securitization requires two conditions: 1) the bank making the loans must keep skin in the game (i.e., enough of the risk of nonrepayment) on its own books to encourage it to monitor their quality, and 2) the rating agencies evaluating the investment quality of the loan portfolios must find it to their advantage to do due diligence. In the last chapter we saw that prior to the financial crisis banks retained too small a proportion of the risk to have enough of an incentive to grant only good loans.
In the United States, rating agencies are essential actors in the process of securitization.8 Recall that banks’ equity capital requirements depend on the riskiness of their assets. After 2004, when the Basel II accord was implemented in the United States, the banks were able to measure this risk using assessments made by the rating agencies. If a bank buys a securitized product, it will need to hold much less equity capital if the rating of the product is AAA than if it is BB. It is essential for government regulators to be able to trust the rating agencies, which are truly auxiliary regulators.
The main problem at the time was that the rating agencies were giving AAA ratings to securitized products that were much riskier than the AAA bonds issued by businesses or local authorities. Was this due to unfamiliarity with the securities—or due to a conflict of interest? It’s hard to say, but the agencies’ incentives were not fully aligned with the regulators’ objectives. The agencies received fees proportional to the value of securitized assets issued, thus creating an incentive to give higher ratings (just as if our salaries as professors increased with the grades we give our students’ papers). The desire to keep happy the investment banks that were major clients of the rating agencies was also part of the problem.
EXCESSIVE TRANSFORMATION
As the last chapter explained, a bank borrows short term to lend long term. This can expose it to a bank run, in which the bank’s depositors, fearing that it might turn into an empty shell, all try to withdraw their money at the same time. In the years before the crisis, many financial intermediaries—not only retail banks—took substantial risks by borrowing very short term on wholesale markets (the interbank and money markets). This strategy is profitable as long as short-term interest rates remain very low, but it exposes the bank to a rise in interest rates if it has not covered itself against this risk. If the interest rate is 1 percent and rises to 4 percent, the financing costs of an institution that is financed almost exclusively using short term borrowing (as was the case for the vehicles created to securitize real estate loans) roughly quadruples.
Banks that have no retail deposits are particularly exposed to this risk (since the introduction of deposit insurance, individuals’ deposits are very safe and therefore not subjects to runs). As we have seen, the five big US investment banks either went bankrupt or teamed up with retail banks to get support from the government. But retail banks, whose funding is, a priori, more stable than that of investment banks, had also increased their reliance on short-term wholesale funds.
This generalized risk taking through substantial maturity transformation puts the monetary authorities in a delicate situation. Either they do not act on interest rates to keep them low, in which case the financial system may collapse, or they keep interest rates artificially low and then indirectly bail out the vulnerable institutions. If they bail out these institutions, they validate the risky behavior, creating costs that I will explain below. Thus, the excessive transformation of short-term funds to long-term lending trapped the monetary authorities. This was particularly clear immediately after the crisis. (The problems today are different. This is true both for the central banks, which cannot make interest rates fall much below zero—because if they did, economic agents would prefer to hold cash, which would at least give them a zero interest rate, ignoring transaction costs—and for the banks, which suffer from low yields across all maturities.)
A WAY OF GETTING AROUND REGULATORY REQUIREMENTS FOR EQUITY CAPITAL
Regulated financial institutions (retail banks, insurance companies, pension funds, brokers) are subject to requirements regarding the minimum level of equity capital, as explained in the last chapter. For banks, the Basel accords set out the general principles at the international level. The idea is to maintain a buffer, the bank’s “capital,” which makes it likely that the bank can absorb most risks it faces. This protects the depositors’ insurer, the deposit insurance fund, and, if public funds are employed to bail out the bank, the taxpayer. For a given size of its balance sheet, the bank, on the contrary, has an interest in reducing its equity capital (or even, in extreme cases, minimizing it to match the regulatory requirement). In effect, holding less equity capital means a higher return for the shareholders who provide it.
The supervisors in charge of financial regulation have a complicated task. For one thing, bank balance sheets and financial techniques are continually evolving; for another, supervisors have limited means to carry out their supervisory tasks or to attract the most talented staff (who have the choice of working instead for the regulated institutions, insurance companies, investment banks, or rating agencies). The supervisors’ task is also complicated when they compete with each other. Prior to the crisis, banks in the United States could sometimes choose their regulator by defining their principal activity in order to obtain the most lenient supervision (for example, selecting “real estate” meant they would face a less intrusive regulator). Incidentally, the very fear of a regulatory race to the bottom in terms of equity requirements by national regulators motivated the Basel accords, which set an international minimum standard.
Before the crisis, many financial institutions exploited the flaws in the regulators’ analysis of their risks to understate their need for capital and thus increase the return on their own equity capital. For example, they made loans requiring little capital to be held against them to vehicles containing assets they had themselves securitized,9 even though the risk for the financial institutions was equivalent to what it would have been if they had left the securitized loans on their balance sheets; put more simply, moving mortgages from the balance sheet to another entity, which it insured against refinancing problems through a line of credit, reduced the bank’s capital requirement substantially. In the end, the regulators were unable to, or didn’t know how to, restrain such dangerous behavior.
EXCESSIVELY VAGUE BOUNDARIES OF THE REGULATED SPHERE AND A SOMETIMES-UNHEALTHY MIXTURE OF PUBLIC AND PRIVATE
Simply put, the prudential regulation of banks involves give and take. Retail banking is supervised, involving equity capital requirements and other constraints. In exchange, it gets access to central bank liquidity and to deposit insurance, two factors limiting its exposure to risk. Deposit insurance dissuades small depositors from fleeing if its financial difficulties become public. Thanks to the central bank’s provision of liquidity, the retail bank can calmly choose between selling assets at a reasonable price and reconstituting its own funds by issuing new equity. Unregulated banks (known as “shadow banks”—investment banks, hedge funds, money market funds, and private equity firms) do not have this privilege. At least, not in theory.
The 2008 crisis showed that failure to regulate the mutual exposure of the regulated and unregulated sectors can compel the authorities to rescue unregulated entities by pouring in capital, buying up assets, or simply keeping interest rates low. So the unregulated sector had access to the taxpayers’ money and central bank liquidity without having to submit to the discipline of prudential supervision. This interdependence between the regulated and unregulated sectors is illustrated by the debate about the US authorities’ refusal to rescue Lehman Brothers in 2008. Taxpayers’ money had already been used to save another investment bank, Bear Stearns.10 A few days after Lehman Brothers went bankrupt, the US government also bailed out another large unregulated entity, AIG, a large insurance company that had become, de facto, an investment bank. Later, much more public aid was given to retail and investment banks. It is difficult to estimate the cost of this aid at the time that it was granted. In the United States, it proved, ex post, to be modest: the banks ended up paying back most of the funds they received. Obviously, things could have turned out worse, as they did in some European countries.
To return to the case of AIG: a priori, there is nothing abnormal about rescuing a large insurance company. However, AIG’s insurance activity was viable and capitalized separately, precisely to protect it from the collapse of the parent holding company engaging in risky activities. The holding company could have gone bankrupt without serious consequences for its insurance business. Although it seems odd that AIG’s holding company was able both to escape supervision and to have access to the taxpayers’ money because of its bad management, the interlinking of this institution with regulated institutions—through over-the-counter markets in derivatives, for example—created a systemic risk that “justified” its rescue.11
The boundary between the public sphere and the private sphere was as blurred as the border between regulated and unregulated sectors. In September 2008, two semipublic real estate credit agencies, Fannie Mae and Freddie Mac, which insured or guaranteed 40 to 50 percent (by 2007, 80 percent) of the outstanding real estate debt in the United States,12 were rescued. Real estate again was the problem, but these two companies were anomalies. Since they were private, their profits did not benefit the taxpayer.13 On the other hand, they had a US government guarantee (in the form of lines of credit with the US Treasury) and were counting on the general belief that the government would bail them out if they got into difficulties. So it proved. Once again, profits were privatized, losses nationalized. Ultimately, they had not been rigorously regulated.14 Strikingly, these agencies still play a major role in guaranteeing real estate loans in the United States.15 In Europe, by comparison, the European Commission limited this phenomenon by successfully using the State Aid Law to prevent EU member governments16 from subsidizing private businesses through implicit state guarantees.
THE NEW POSTCRISIS ENVIRONMENT
The crisis left at least two legacies: low interest rates and the search for new forms of regulation.
HISTORICALLY LOW INTEREST RATES
This particular legacy was supposed to be temporary. Very soon after the crisis began, the US, European, and British central banks provided much liquidity and thereby reduced interest rates to close to zero—in other words, to negative levels allowing for inflation (that is, in real rather than nominal terms). Japan has had an interest rate below 1 percent since the mid-1990s; in 2017, it is zero. Interest rates in Japan and Europe are expected to remain close to zero for a while, while the United States is beginning to raise rates very cautiously.
Low interest rates in downturns have a clear rationale. In particular, low short-term rates allow financial institutions to refinance themselves at low cost and ease the financial sector’s problems.17 In the end, only the state can provide liquidity for the economy. It can do two things that markets never can: first, mortgage the future revenue of households and businesses (even those that don’t yet exist), or more precisely, the taxes that the public authority will levy on this revenue. This sovereign power of taxation underlies the state’s role in macroeconomic regulation. The ability to tap into the future income of economic agents allows the state to issue national debt and to provide liquidity to the banking system.18 The state can jump-start banks and businesses today in exchange for an increase in taxes tomorrow. Second, the central bank can create inflation, and in that way change the real value of contracts that are denominated at face value, both loans and nonindexed wage agreements. (Currently, central banks struggle to create expectations of even moderate inflation, so this second approach has no effect.)
The primary goal of providing liquidity is of course not to save banks that imprudently get themselves into difficulty and need refinancing; it is to keep alive the financial intermediaries so essential for the economy to function. Small and medium-sized businesses do not have access to financial markets (they cannot issue bonds or commercial paper to finance or refinance themselves) because they do not have an established reputation in the financial markets, have few assets to pledge, and are not diversified. They depend on banks to monitor them and ensure that the collateral these businesses offer is of good quality. When the banks are in difficulty, small and medium-sized enterprises are the first to suffer, as we have seen in all credit-crunch events.
Yet low interest rates, no matter how necessary they may be in a crisis, are not without costs:
• They lead to a massive financial transfer from savers to borrowers. In fact, this is exactly what a monetary bailout of banks is intended to do. But low rates please other investors, not just those in the regulated banks, because a decrease in interest rates increases the price of assets such as property or shares (the future yields on these assets are attractive in relation to the low yields offered on the bond market). This redistributes wealth because the owners of these assets, whether regulated or not, receive more when they sell them.19 Low rates thus have gigantic redistributive effects, some desired and some not.
• Financial bubbles tend to emerge when interest rates are low, as we saw in the preceding chapter.
• Low interest rates encourage financial institutions that have guaranteed higher yields to their customers to take additional risks. This is a problem in Germany, for example, where insurance companies have promised investors in life assurance funds yields as high as 4 percent. The yield on ten-year German government bonds, which varies between 0 and 1 percent, makes it very difficult to make good on this guarantee unless the funds are invested in high-yield and therefore riskier (even “junk”) bonds.20
• Low short-term interest rates could lay the foundation for the next crisis by encouraging banks to borrow even more short-term funds. This argument is currently less persuasive for two reasons: 1) “quantitative easing” has had just as big an influence on long-term rates, which are now often as low as short-term rates, so banks’ prospects are not as attractive as their low financing cost would suggest, and 2) regulators are currently setting up liquidity requirements so as to limit banking institutions’ short-term indebtedness.
• There is a fifth cost: when we get to zero nominal rates, they can fall no further, because people will prefer to hold cash, which keeps its nominal value (that is, it offers a nominal rate equal to zero).21 This is what economists call the Zero Lower Bound (ZLB). The central bank can no longer boost the economy by lowering interest rates to negative levels; this can quickly lead to a recession and unemployment. In this situation, central banks have to turn to a toolkit of complex, imperfectly mastered instruments22 that I will not discuss here.
LOW INTEREST RATES OVER THE LONG TERM?
Until the crisis, the consensus among macroeconomists was that we were in a period called “the great moderation.” Monetary policy, sometimes accompanied by fiscal policy (in a “policy mix”), seemed to have done remarkable work during the twenty years preceding the crisis. It aimed at price stability by targeting a stable target inflation rate (2 percent, for instance), and adjusting monetary policies to reflect the economic situation and unemployment level. Today, this consensus regarding the primacy of monetary policy no longer exists, in part because this policy mix is not viable at the Zero Lower Bound.
What if low interest rates were not merely a temporary phenomenon associated with the crisis? What if we were fated to live for a long time in an economy with low interest rates, in which monetary policy is unable to reenergize the markets and prevent recessions and unemployment—a phenomenon that is a piece of what is sometimes called “secular stagnation”?23 Economists disagree as to whether this is the situation today. What is certain is that there has been a decrease in interest rates on safe assets (say, government bonds) since the 1980s. In real terms (that is, once inflation has been deducted), these interest rates were around 5 percent in the 1980s, 2 percent in the 1990s, 1 percent until the Lehman Brothers bankruptcy in 2008, and about minus 1 percent since then. What are the reasons?
The first structural reason concerns the supply and demand for safe assets. If there is little supply and much demand, the price of these assets will necessarily be high. For a financial asset, a high price corresponds to a low yield (intuitively, the owner of the asset pays a high price if he only acquires the right to get a low yield in the future). There are several other symptoms, in addition to the low interest rates, of this excess demand. Before the 2008 crisis, there was a frenzy of securitization whose goal was to create safe financial assets (although ultimately this securitization created risky assets, as explained above, which was not how it was described at the time). Another sign is the emergence of bubbles.
The demand for safe assets has increased. First because of an overall higher level of savings. Since emerging economies (such as China) and countries with raw material revenues (such as the petroleum-producing countries when the price of oil was still high) did not have developed financial markets, they tried to invest their money in developed markets. This led to the “savings glut” mentioned earlier. Another factor contributing to higher savings is the increase in inequality, because well-off households save more than poor ones. Greater savings tend to reduce the yield paid to savers.
Savings have also exhibited a “flight to quality”: a shift in savings composition in favor of safe assets. Since the crisis, tougher prudential regulation has penalized risk taking. As a result, banks, insurance companies, and pension funds now have more appetite for safe assets, that require them to hold relatively little capital. Individuals are also taking refuge in uncertain times in safe assets. The French invest almost 85 percent of their long-term life assurance savings in “euro funds” (that is mainly in bonds issued by governments and by highly rated companies); most of these euro funds are guaranteed in nominal terms (i.e., there is no risk of a loss of principal). They do not tend to invest in riskier assets, such as shares.
The supply of safe assets seems to have decreased: diversified real estate portfolios and the sovereign debt of OECD countries, which used to be considered completely safe, are now risky. This has led to a significant worldwide fall in liquidity. According to Ricardo Caballero and Emmanuel Farhi, the supply of secure assets had fallen from 37 percent of worldwide GDP in 2007 to 18 percent in 2011.24
Finally, lower population growth is often invoked to explain low rates.25 Demography has complex effects, but many researchers agree that it is a factor. For example, it decreases labor supply relative to capital, and thereby the yield on capital, causing interest rates to fall. In a pay-as-you-go pension system (that is, with pension benefits financed by contributions levied on active workers rather than funded by pension funds’ assets), the demographic slowdown also translates into a reduction in the relative number of active workers, an increase in private savings to offset the reduction in pensions, and, ultimately, lower interest rates.
So it is possible that low interest rates may be here to stay, in which case we will have to rethink macroeconomic policies.
THE NEW REGULATORY ENVIRONMENT
Nothing is without risk. Although we need to respond vigorously to the failures of regulation, and to reduce the frequency and scale of crises, we cannot completely eliminate the danger that they will happen. Just as a person who has never missed the beginning of a movie, never been late for a meeting, and never missed a train is probably overcautious, an economy in which people acted in such a way as to make a crisis inconceivable would probably be one functioning far below its potential. To avoid all crises we would have to constrain risk taking and innovation. We would also need to invest in the short term rather than over the long term, because the long term is more uncertain, and so riskier. The goal is therefore not to completely eliminate crises, but rather to get rid of incentives that encourage economic agents to adopt behaviors harmful to the rest of the economy. This requires limiting the externalities the financial system imposes on savers and taxpayers.
Prudential regulation and supervision are more art than science, because it is in fact difficult to assemble the data needed to measure precisely the effects predicted by theory. Yet there are still some general principles we can use, although it is hard to quantify the relevance of each of them. In 2008, a number of economists, including myself, recommended26 protecting regulated institutions against the risk of contagion from the unregulated sector; increasing their levels of equity capital and putting greater emphasis on liquidity; making regulation more countercyclical; monitoring the pay structures of senior bank officers; allowing securitization, but supervising how it is used; monitoring the rating agencies; rethinking the “regulatory infrastructures”; and, in Europe, creating a supervisor on the European level within the ECB (this has since happened—see chapter 10). What is the situation today?
HAVING YOUR CAKE AND EATING IT TOO
Regulators, central banks, and governments have been forced to intervene to rescue financial institutions they did not regulate through bailouts, buying up toxic products, and loosening monetary policies. As we have seen, one illustration of this phenomenon is the recent history of US investment banks, particularly Bear Stearns and AIG. The only denial of aid to an investment bank (Lehman Brothers) created a serious panic in financial markets, halting consideration of further private sector bail ins.
The fear of systemic risk played too big a role in the formulation of public policy. This was partly because there was a lack of transparency about mutual exposures between financial institutions. The regulators had little information regarding the exact nature of the mutual exposures and the counterparty risk involved in the trades in over-the-counter markets. More generally, it is almost impossible for a regulator to calculate mutual exposures, direct and indirect, in the global financial system, especially since some of the financial institutions concerned are either not regulated or are regulated in other countries.
Thus, it is a question of keeping as many harmful financial products as possible out of the public sphere. In this case, the public sphere corresponds to the regulated sphere, which in theory is the only one that can be bailed out. Reforms have been implemented to make this containment more likely in practice.
The first of these reforms is the standardization of the products and their trade on organized (rather than over-the-counter) markets. Although it is important for the financial system to be able to tailor products to various specific needs, this makes it much more difficult for supervisors to assess the corresponding commitments and valuations. There is clearly no question of banning financial innovation or tools suited to specific needs, but the migration of regulated intermediaries toward standardized trade on exchanges should be encouraged (through a judicious choice of equity capital requirements). Unregulated intermediaries should, of course, remain free to put the accent more on over-the-counter trading. As explained in the preceding chapter, businesses and banks primarily need insurance contracts against straightforward risks—changes in macroeconomic variables such as exchange rates or interest rates, and the default of counterparties to which they are most exposed.27 These standardized products can be traded on derivatives platforms that limit mutual exposures: regulators need to have a clear picture of regulated institutions’ exposure to default by another institution. The use of well-capitalized clearing houses that demand guarantee deposits from their participants, along with the centralization of supply and demand, could potentially achieve this.28 The new, postcrisis Basel III accord has moved in this direction. It penalizes over-the-counter contracts by demanding that more capital be held against them. It would be desirable to go further. At the same time, clearing houses must be subject to strict prudential rules too, otherwise the regulator would merely be diminishing the risk of (direct) bank failures by increasing the risk that the clearing houses default.
A more drastic version of the idea of insulating the retail banks is one that structurally separates retail banks from investment banks, a proposal made in different forms by Paul Volker, the former head of the Federal Reserve in the US, and by Commissioner Liikanen for the European Union. It was stated in its most drastic form by John Vickers, a prominent British economist, in the recommendations of the report from the UK’s Independent Commission on Banking.29
THE COUNTERCYCLICAL CHARACTER OF EQUITY CAPITAL REQUIREMENTS
There are good theoretical justifications for a countercyclical solvency ratio, that is, capital requirements that are higher during booms but fall during a banking crisis. For one thing, periods when banking equity capital is in short supply go hand in hand with a credit crunch, making life difficult for businesses depending on the banking system; in particular, small and medium-sized businesses either must pay high interest rates or are refused loans. For another, policymakers should aid the financial system at times of scarce liquidity, especially for those liquidity shocks that are rare (as this makes it too costly for the private sector to hoard liquidity just in case). Loosening solvency constraints in such periods is one way to provide this aid, alongside monetary policy.30 The Basel III agreement has provided for banks’ capital buffer to be countercyclical, reflecting macroeconomic conditions.
THE REGULATION OF LIQUIDITY AND OF SOLVENCY
Before the crisis, there was no unified regulation of liquidity, whether in the Basel accords or at the European level, and the liquidity requirements placed on banks were low. In theory, regulators should enforce both a liquidity ratio and a solvency (or capital) ratio, but the practice is more complex. It is notoriously difficult to construct a good measure of the liquidity of a financial intermediary. The liquidity of a bank depends, on the asset side of the balance sheet, on the possibility of reselling, when necessary, securities (treasury bonds and bills, certificates of deposit, securitized products, stock shares, bonds) without too much discount (this is market liquidity). On the liabilities side, it depends on whether the institution can raise funds (such as sight deposits or certificates of deposit) rapidly and on reasonable terms (this is funding liquidity). A bank’s liquidity also depends on its reputation, which affects the value of the assets it tries to sell and its ability to raise new funds.
The Basel Committee is putting the final touches on two new ratios: The Liquidity Coverage Ratio requires banks to hold liquid assets, such as treasury bonds, in an amount equal to or larger than its net cash loss over thirty days in case refinancing could not make up for a massive withdrawal of (essentially uninsured) deposits. The Net Stable Funding Ratio has a similar flavor, but looks at a horizon of one year.
Calculating equity capital requirements will always be a work in progress. The correct level depends on the risk the regulator is prepared to tolerate, on the volatility of the economic environment, on the quality of supervision (are the rules being properly applied?), on the composition of the bank’s assets and liabilities, and on the danger of activities migrating into the unregulated, shadow banking sector. With the necessary data hard to get, it is difficult for an outsider to estimate with precision the right level of equity capital. There will always be some trial and error, but we know this much: banks were not holding enough equity capital before the crisis.
Basel III has increased the requirements: The required Tier 1 capital31 rose from 4 percent to 7 percent, to which can be added the countercyclical buffer (when there is high credit growth in the economy) of between 0 and 2.5 percent. An extra capital requirement of up to 2.5 percent is required from banks that are deemed systemically important. The total requirements for Tier 1 and Tier 2 capital rose from 8 percent to up to 13 percent. There is also a new minimum leverage ratio, the philosophy of which is based on a vision that supervisors have an extremely limited ability to gauge the risk; according to current proposals, banks will need to have at least 3 percent of their risk-unweighted exposures (on and off balance sheet, as well as those related to derivatives and securities financing) in Tier 1 capital, with a possible surcharge for large banks deemed to be systemically important.
Will that be enough?32 It’s difficult to say, but the increase in capital requirements is an important step forward.
The current reforms tend to be “macroprudential”: they are rooted in the idea that the solidity of a bank depends not only on its own equity and liquidity, but also on the solidity of other banks. There are many reasons for this.
Banks can be interdependent through their mutual exposure, which raises the fear of contagion if one of them goes bankrupt; they are also dependent on each other in a more indirect way, because if they encounter difficulties at the same time, they will try to sell off their assets concurrently. The wave of orders will cause asset prices to fall (in so-called “fire sales”) and reduce each bank’s market liquidity.
Bank failures have different consequences in times of crisis than in periods of stability. There is more likely to be a systemic impact if the other banks are also affected by a macroeconomic shock. Furthermore, a possible bailout by the taxpayer is more expensive if the government has already been forced to rescue other banks. Finally, we have already observed that, in the case of excessive transformation (borrowing short and lending long) by financial intermediaries, the central bank will have no choice but to reduce interest rates. All this implies that a bank should hold more capital when its strategy means its risk of failure is strongly correlated with macroeconomic shocks.
REMUNERATION
Remuneration in banking circles is the subject of two debates. One concerns the amount: the amount of remuneration in finance, in particular in the US and the UK, is high. High levels of remuneration in themselves do not justify ad hoc treatment of the financial world: whatever the state’s preferences about redistribution, it should redistribute income through taxes, not decide whether a banker deserves less than a television anchor, a successful entrepreneur, or a soccer player. The other debate concerns whether high remuneration packages reward good performance or, on the contrary, create bad incentives. Large bonuses received by managers33 who later fail, stock options received before share prices collapse, or golden parachutes34 that reward underachievement are shocking not only from an ethical point of view, but from the point of view of efficiency. These outcomes do not create good incentives.
The bonus culture is pertinent both to the excesses of finance and to the question of inequality. Systems for remunerating managers are in fact too often focused on overachievement and the short term, and thereby encourage excessive risk taking. This is particularly true when there is a small risk of extreme loss (“tail risk”). A risky strategy that is profitable with a probability of 95 or 99 percent, but may produce a catastrophe otherwise, then secures (most of the time) a generous remuneration for the manager, and leaves the high—but improbable—losses to shareholders, creditors, and taxpayers.
Why are shareholders likely to go along with such compensation policies? The first answer is that they also profit so long as the downside risk is not realized, even if they would lose their shirts if it did. A second answer is that the banks tend to give priority to short-term remuneration to attract talented employees. This was particularly apparent during the years preceding the 2008 crisis. Unrestrained competition for talented employees is conducive to bonuses and short termism; bonuses do not result only in higher-than-average pay, but also a wide dispersion of profit-based compensation among managers or traders (because the increase in remuneration due to competition to attract and keep talented employees occurs through variable remuneration, not fixed salaries).35
Why do banks’ creditors agree to lend under such conditions? They do not always know about the risk taking, but, above all, banks are able to take these greater risks thanks to the state’s explicit or implicit safety net, which enables them to continue raising new funding even in the face of bad omens. That may be why finance is different. The television anchor, the entrepreneur, or the soccer player do not call on public money when they get into financial difficulties.
Thus, it seems legitimate for the state to regulate compensation in the private sector, at least the part of it that is liable to be bailed out with public money. The state can insist on compensation schemes that induce bank managers to take a long-term perspective. A case in point is deferred compensation, in which the managers’ compensation is vested over time and granted only when it becomes clear that the managers’ performance was not a flash in the pan.36 In addition, the second pillar of Basel II allows regulators to require an increase in equity capital if the system of remuneration encourages short termism and excessive risk taking. Of course, deferring remuneration by just a few years may not be enough: certain risks materialize only much later, because they are taken over a long period (for instance, the risk of longevity in life insurance). Over very long periods (ten years, say) it is difficult to distinguish the contribution made by one manager from that of successors.37 A compromise must be found.
Finally, it is entirely possible that the remuneration committees of banks indulge senior management. It’s not clear that this is peculiar to finance. Problems of governance exist in all sectors of business. Regulation specific to finance cannot be grounded in this argument alone.
Opponents of regulating bankers’ remuneration have two arguments:
The first relates to the importance for a bank, as for any business, to be able to attract the best talents to lead it. Let us suppose, to pursue this argument, that a bank could increase its value by 0.1 percent by attracting a CEO who is a little more talented than others. If this bank has a value of a hundred billion dollars on the stock market, this increase represents a hundred million dollars. The bank would rightly be prepared to pay a lot to acquire the services of this more talented manager.38 A variant of this argument maintains that the bank has no choice because of the competition from unregulated intermediaries, such as hedge funds and private equity companies, offering generous remuneration to those they consider to be the best managers. Failing to match these employers would deprive retail banks of the best talent.
The second argument is that the excesses of finance will not be corrected by regulating pay and bonuses. Hubris may be as important as a cause of dysfunctional behavior as is profit.39 Recall the immoderate ambition of CEOs such as Richard Fuld, who wanted his bank, Lehman Brothers, to beat Goldman Sachs; and Jean-Yves Haberer, who wanted to transform Crédit Lyonnais into a world leader; or the personal ambition of rogue traders such as Jérôme Kerviel (Société Générale) and Nick Leeson (Barings). If hubris is the main driver of risk taking, regulating remuneration would have little effect, and only classical prudential supervision could limit it.
In conclusion, questions about excess pay seem to go beyond the regulatory framework for finance. They raise the general question of the level of redistribution the government wants to see, whether in banking or in any other sector. The question of the structure of remuneration and incentives seems to be more specific to banking, insofar as the failure of a bank can lead to the demand for public funds. Controlling remuneration that encourages risk taking and is oriented toward the short term must therefore be part of the supervisory framework.
Basel III created some relevant guidelines (the exact regulations depend on how these guidelines are implemented in different countries). They reduce the proportion of variable pay (for example, variable remuneration such as bonuses is not to exceed fixed remuneration); in addition, the guidelines introduce a deferral period (usually three to five years) to penalize behavior that is profitable in the short term but costly in the long term. As in the case of the increased equity capital requirements, these reforms are difficult to calibrate. But they seem to be heading in the right direction.
RATING AGENCIES
The crisis also raised questions about the rating agencies. These agencies play a central role in modern finance by informing individual and institutional investors, as well as regulators, about the risks of financial instruments. They failed to do this in the case of subprime mortgages. The major argument in favor of regulating the rating agencies is that over time their judgment has become an integral part of the regulatory assessment of risk, and that they earn major revenues from this. The capital requirements of regulated institutions (banks, insurance companies, brokers, pension funds) go down significantly when they hold highly rated debts. The privilege that rating agencies enjoy has to be counterbalanced by supervising their methodology and conflicts of interest. On the other hand, there is no basis for regulating the activities of rating agencies that are not involved with prudential regulation (unless these activities lead to a conflict of interest).
Basel III and the new prudential regulations for insurance companies (“Solvency II”) retain the principle of using ratings to estimate risk, while regulators in the United States are now far more circumspect about using ratings.
REGULATORY INFRASTRUCTURES
The crisis put in question not only the regulations, but also the supervisory institutions that apply the regulations. Could supervisors take prompt corrective action before getting to the point of either closing or bailing out a bank? Could coordination be achieved between different national supervisory authorities or between the authorities in several countries? When it comes to international cooperation, the main problem is transnational financial institutions. The systems for guaranteeing deposits and for transferring assets, and the laws governing bankruptcies, differ from one country to the next. Supervision (the monitoring and implementation of capital requirements) and the management of crises (bailing out institutions or accepting their bankruptcy, buying up toxic assets, and so on) offer textbook cases of free riding and countries’ ability to game the system. Unfortunately, I do not have room to explore this important issue further here.
Is THE FINANCIAL SYSTEM NOW SECURE?
As I have said, the current state of our knowledge and, in particular, the limited availability of the data that would allow supervisors (or economists) to calibrate capital and liquidity requirements precisely should encourage us to be humble. However, as long as the reforms are implemented and not derailed, the financial system will prove to be less risky than it was before: the Basel III reforms seem to be headed in the right direction. An increased requirement of equity capital, the introduction of a minimum liquidity ratio, the inception of macroprudential measures in the form of countercyclical equity capital buffers, a greater use of centralized exchanges instead of over-the-counter markets, institutional reforms (for example, the creation of the European Single Supervisory Mechanism)—all are genuine improvements.
There are still, however, major areas of risk. Some of these are connected to the macroeconomic environment; they are based on slower global growth, more volatile financial markets, and the challenge of how to exit low–interest rate policies without compromising growth. Other concerns stem from the combination of geopolitical risk and local economic conditions—for example, in Europe political shocks such as the UK’s Brexit vote, the political uncertainty over the European Union, the structural weakness of certain economies, the significant proportion of unproductive loans still on European (especially Italian) banks’ balance sheets, and the intimate connections between banks and sovereign states.40 There is uncertainty about how China will transition from a catch-up economy to one on the frontiers of technology and institutional design (including managing its credit bubble and reforming financial markets). In the emerging economies, overindebtedness in foreign currencies (usually in US dollars) may put businesses and banks in difficulty if the local reliance on commodities (natural resources, agricultural products) is associated with inadequate risk management.41 Finally, economists still do not know enough about how prudential regulation ought to operate, including the extent to which investors should be held responsible for their investments in regulated institutions (i.e., bailed in, in case of default)42 and, of course, about the proper calibration of capital and liquidity requirements.
I will end by discussing a particular issue: shadow banking. As regulation becomes more rigorous, banking activities tend to migrate toward “parallel” banks that are either lightly regulated or not regulated at all. There is no objection to this as long as the migration does not take place at the expense of vulnerable actors (small depositors and small and medium-sized enterprises) or taxpayers. Now, as we saw in 2008, the shadow banking sector can benefit in practice from public liquidity and bailouts. At the time, this was because the regulated banks were exposed to the shadow banks if they got into distress, either directly through liabilities owed them by shadow banks, or indirectly because the latter might trigger fire-sale prices for some types of assets, thereby making it harder for regulated banks to raise cash by selling their own assets; but we can imagine other factors that could lead to resorting to public finances in case of bank distress, for example if individuals put their money in shadow banks, or small businesses started to depend on borrowing from them (both true today in China).
WHO IS TO BLAME? ECONOMISTS AND THE PREVENTION OF CRISES
In the end, the financial crisis of 2008 was also a crisis of the state, which had been disinclined to do its work as a regulator. Like the euro crisis discussed in chapter 10, the 2008 crisis had its origin in the failure of regulatory institutions: failure in prudential supervision in the case of the financial crisis, and failures of state supervision in the case of the euro crisis. In both cases, lax supervision prevailed as long as everything was going well. Risk taking on the part of financial institutions and countries was tolerated until the danger became obvious. Contrary to what many people think, these crises were not technically market crises—the economic agents were reacting to the incentives they faced, and the least scrupulous among them exploited gaps in the regulation to swindle investors and take advantage of the public safety net. Rather, the crises were symptoms of a failure of national and supranational state institutions.
Economists have been roundly reproached for not having predicted the crisis43 and even for being responsible for it. In reality, most of the causes of the financial crisis were connected with hazards that had been studied before it occurred: asset bubbles, the impact of excessive securitization on the issuers’ incentives, the growth of short-term indebtedness and the possible lack of liquidity in financial institutions, poor measurement of banking risk, the moral hazard of rating agencies, the opacity of over-the-counter markets, the drying up of markets and the disappearance of market prices, herding behavior in financial markets, and the procyclical impact of regulation.
Nonetheless, although academic research provided some keys to understanding several of the factors that led to the crisis, it had little success in preventing it. We must acknowledge that economists had little influence during the period leading up to the crisis. Four factors contributed to this situation:
First—and this is an essential point—it has to be understood that economists will always be more comfortable identifying the factors likely to lead to a crisis than predicting whether it will occur, or on what date, just as a physician will be more comfortable identifying factors that might cause an illness or a heart attack than in saying exactly when they will occur.44 Just like epidemics and earthquakes, financial crises are difficult to predict, but we can identify likely causes. Since financial data are very imperfect and the world is continually changing, there will always be great uncertainty about the magnitude of the effects concerned, not to mention the self-fulfilling factors (like bank runs45) that are, by definition, unpredictable because they are, in Keynes’s words, based on “the feeling … in the mind of the investor.”46
Second, the diffusion of academic knowledge was very piecemeal. The blame for this falls both on the researchers, who often did not make the effort to share their knowledge and make it more operational, and on policymakers, who pay little attention to gloomy warnings from economists when things are going well. Researchers cannot expect policymakers to read technical articles (even if knowledge is often transmitted by economists working for regulatory authorities); they have to extract the essence, make research comprehensible, and show exactly how to make use of it. These are things that top economists are often loath to do, because they prefer to devote their time to creating rather than disseminating knowledge—not to mention the fact that their academic reputation depends on the approval of their peers, not that of policymakers. To facilitate the dissemination of scientific knowledge, it can only be beneficial to train excellent applied economists who will work for regulators rather than embarking on academic careers, and will share research insights in conferences organized with regulators, central bankers, and bankers.
Third, almost all researchers were unaware of the extent of the risks that were being taken in the financial sector; for example, they did not know the amount of off-balance-sheet commitments or the size and correlations of over-the-counter contracts. To be sure, supervisors had only partial knowledge too; but outside their small circle, very few knew what was going on. Should academic economists have been better informed? I have no good answer to that question. On one hand, it would have been useful if policymakers had listened to economists. On the other hand, economists specialize: research and teaching are distinct from applied economics, even if they nourish each other.
Fourth, a few economists, either by inner conviction or because of conflicts of interest, underestimated the importance of financial regulation or oversold the virtues of over-the-counter markets or of financial innovation. Their arguments were quickly exploited by interested parties. Charles Ferguson’s 2010, well-researched film Inside Job, while certainly polemical, shows the dangers of complicity between researchers and the subjects of their research. The issue of conflicts of interest is not very different from the problems that arise in other sciences when private or public interests intrude into the world of research. The difficulty is immediately obvious: those with the information that could have the greatest relevance for public policymaking are often connected with those who have a stake in regulation. There is no miracle solution. To help mitigate this problem, most research groups, universities, and public organizations now have an ethical charter requiring researchers to declare potential conflicts of interest. This is useful, but researchers must ultimately be bound by personal ethics.