The 2008 global financial crisis had repercussions that still persist. The problems in the United States housing market spread to the real estate market in Europe, and to the banks with exposures to this market. Governments bailed out banks, cut spending and borrowed heavily. As of 2016, interest rates around the world are still low, and sustained economic growth seems elusive.
20.2 The 2008 Global Financial Crisis 529 The financial crisis was characterised by a lack of liquidity – particularly fund- ing liquidity – and a corresponding fall in the creditworthiness of firms and gov- ernments. Whilst the popular view is that the crisis is the fault of ‘the bankers’, it is important to understand both the background to the crisis and the particular risk management failures that caused it.
20.2.1 Causes of the Crisis The Role of China
A key role in the build-up to the crisis was played by China. Over the last few decades, the Chinese economy has grown very quickly. Much of this growth has been driven by exports to the West. The capacity for growth in China has meant that the rise in production has occurred in parallel with relatively low rates of inflation.
More importantly, the large flows of wealth into China – to pay for Chinese exports – did not prompt a strengthening of the Chinese renminbi. Such flows of funds normally cause a currency to appreciate significantly over time, but the People’s Bank of China has deliberately maintained a reasonably stable exchange rate with the US dollar. This maintained demand for Chinese exports, but also meant that increasing demand in the West did not lead to price inflation.
The increasing demand was stoked by persistently low rates of interest in the United States and elsewhere, and China played a part here too. The Chinese Gov- ernment needed to invest the income that it was receiving from exports, and a sig- nificant proportion of that income was used to buy US treasury bonds. The demand for these bonds was high enough that the price was inflated, and as the price of a bond increases its yield falls.
The Role of Housing Markets
The low rates of interest on government borrowing meant that banks were also able to maintain low interest rates. This meant that mortgages became cheaper, causing house prices to rise.
The Role of Regulation
The Gramm–Leach–Bliley Act allowed the commercial and retail banks to carry out investment banking activities, something that had been forbidden by the Glass–
Steagall Act. In normal markets, this could be seen as desirable since it allows banks to benefit from the diversification of carrying out both types of business.
However, in stressed markets, it means that catastrophic losses in the investment banking arm can adversely affect retail and commercial account holders. This is what happened to a large extent in the recent financial crisis, although several banks were also brought down by issues in basic retail and commercial lending.
The Basel I system of banking regulation also played a part. In particular, it gave banks an incentive to convert credit risk in respect of mortgages to market risk by securitising loans. The resulting vehicles are known as mortgage-backed securities (MBSs). Whilst this was a form of regulatory arbitrage, it was not discouraged, except through removing the need for arbitrage by improving the allowance for credit risk under Basel II. Indeed, the ability to package risk and spread it around the market was seen as an important diversification tool. Furthermore, the tranching approach used in collateralised mortgage obligations (CMOs) – the MBS version of a CDO – was seen as a good way to allow investors with different risk appetites to gain exposure to a single pool of risk. However, since the main buyers of these securities were other banks, they were linked to each other through exposure to the housing market in a way they had not been before. Furthermore, the holder of an MBS does not have the same level of information on the borrowers underlying the security as the bank that sold the mortgages does.
The Role of Incentives
Securitisation also had an adverse effect on incentives. Until the advent of securiti- sation, a bank making a loan would hold the risk for that loan on its balance sheet.
However, securitisation meant that once a loan had been taken on, the profit could be capitalised by packaging the loan – and much of the risk – into an MBS. This resulted in a reduced incentive to ensure the creditworthiness of borrowers and an overall decline in credit quality.
The pricing of CMO tranches was determined by credit rating agencies. Al- though there may have been concerns over the models used, as discussed later in this section, there was an incentive for banks to exploit this mis-pricing by retain- ing particular (under-priced) CMO tranches (usually the equity tranche) and selling the (over-priced) remaining tranches.
There were also incentive issues in terms of those structuring and trading MBSs and CMOs:
• a significant proportion of earnings was paid as bonus;
• bonuses were based on results over a relatively short period (often one year); and
• bonuses were based to a significant extent on team rather than individual perfor- mance.
The fact that a significant proportion of earnings was paid as bonus gave em- ployees an incentive to take significant risks to earn these bonuses – if bets yielded a positive result, a high bonus was paid; if they yielded a negative result, then the result was no bonus, but the minimum bonus was zero rather than a negative num- ber.
Because bonuses were often based on short-term results, there was an incentive
20.2 The 2008 Global Financial Crisis 531 to make short-term profits without considering the long-term impact of a trade. This was particularly true if the profit used to calculate bonus required assumptions in respect of future outcomes. In particular, if an MBS or CMO were bought or sold, then it would be possible to calculate the profit earned on this trade at the point it was created; however, since the ultimate profit depended on the extent to which the borrowers underlying the security were able to make payments, the final result may have been very different.
Finally, if bonuses were based on team performance, then there would have been little incentive to do anything differently from the rest of the team – a good (but different) individual result would not have had a huge impact on an individual’s bonus, but a bad (and different) individual result could have resulted in an individ- ual being fired. If an individual copied the rest of the team, then he or she would have shared fully in the good times and would have had protection in the bad – it is rare that an entire team is sacked.
The Role of Models
One common accusation in relation to the crisis is that the models used to price CDOs and CMOs were not sufficiently accurate. In particular, the Gaussian copula has been singled out. This is perhaps true for some market participants – particu- larly credit rating agencies – but many individuals recognised the short-comings of this copula when used in time-until-default models. Many also improved upon this assumption – but the bonus structure described above might have meant that many did not, if using a Gaussian copula gave higher expected profits and thus bonuses.
However, it is in any case important to recognise that even improved models cannot exactly replicate the world. As such, they should be treated only as guides to what might happen. Excessive reliance on the output of models by senior man- agement, and a lack of understanding of the models’ limitations, was at least as important as any shortcomings in the models themselves and perverse incentives.
Organisational Issues
Many banks were exposed to housing market risks both directly through mort- gages and indirectly through MBSs. However, it seems that senior managers did not recognise this concentration of risks. Furthermore, whilst some departments of banks were selling MBSs and CMOs, others were buying them. This concentration was similarly missed.
20.2.2 Evolution of the Crisis
Many of the mortgages sold in the housing boom were to sub-prime borrowers.
This meant that they had a higher-than-average risk of defaulting on those mort-
gages. If such a risk had been correctly priced through a sufficiently high interest rate, then this would not necessarily have caused a problem. However, as discussed earlier in this section, securitisation led to a reduced incentive to check creditwor- thiness and indeed to correctly price risk. The situation was made more precarious by the sale of mortgages with low initial rates that rose sharply after a couple of years. Since again the incentive was to sell as many mortgages as possible, risks were not properly explained to borrowers. This led, inevitably, to a sharp rise in mortgage defaults.
The widespread exposure to mortgages meant that many banks made large losses.
However, the complexity and opacity of some of the products used meant that many banks could not easily quantify their exposure to future losses. This made banks reluctant to lend to each other. Because banks rely on short-term funding to remain solvent, such a fall in funding liquidity left many banks on the brink of insolvency.
Some were actually wound up and many others required government funding to re- main solvent. The government assistance came in two forms. The first was to pur- chase certain illiquid assets from banks in order that they might exchange illiquid assets for liquid ones. However, some governments also provided cash in exchange for equity stakes in banks, in some circumstances going as far as complete nation- alisation. Whilst many of these stakes have been put back into private ownership, governments have often had to accept losses on their investments.
The reduced solvency of the banks meant that their own ability to lend was compromised. As a result, the liquidity crisis spread from the financial sector to the wider economy, as firms and individuals found it harder to borrow. When loans and mortgages were made available, the interest rates charged were higher, to com- pensate for the higher perceived credit risk, and also to help banks to rebuild their depleted risk capital. In addition, Basel III has increased capital requirements for banks, as well as introducing liquidity requirements. These have further reduced the appetite of banks for lending, particularly for illiquid, long-term projects.
The resulting slowdown in economic growth combined with the cost to gov- ernments of stabilising financial institutions has resulted in large budget deficits.
It also exposed the structural differences between various Eurozone governments, these differences being reflected in large differences – hundreds of basis points – in the cost of borrowing between governments.
20.2.3 Lessons from the Crisis
There are a number of important lessons that might be learned from this crisis, although commentators and stakeholders will have differing views over the extent to which each is valid. However, key ones in respect of risk management relate to:
20.2 The 2008 Global Financial Crisis 533
• the organisational structure of banks;
• the capital structure of banks;
• the structure of bank bonuses; and
• the use of models.
Organisational Structure
The Gramm–Leach–Bliley Act meant that banks profits could be smoothed over time. However, it also meant that commercial and retail account holders – who generally have a low risk tolerance – were exposed to excessive risks. Should this continue?
Within banks, ERM should have a much higher status, and CROs should have much more power. The CRF should be able to see all similar risks across a bank and should have the authority to stop undue risk being taken.
Those who design and work with complex models should be given a greater say in the use of those models and should be encouraged to make their limitations known.
Capital Structure
Banks will continue to borrow on a short term basis to fund risks with longer terms.
However, they should hold more capital in case the economic outlook changes adversely. The type of capital is also important – it should be liquid. Banks should also ensure that they have contingency plans in case normal sources of liquidity dry up. Both capital quality and liquidity are dealt with in detail in Basel III.
Having said this, capital – even liquid capital – is not an alternative to good risk management and should not be regarded as such.
Bank Bonuses
Bank bonuses should reflect the term of the instruments being traded. In particular, full bonuses should not be awarded before the risk inherent in any deal has fully run its course.
This principle should also extend to securitisations. If a bank transfers risk to the financial markets, then adequate risk should be retained to give the bank sufficient incentive to ensure the credit quality of the loans the bank sells in the first place.
In particular, this risk should be reflected in the term and structure of the bonuses earned by bank employees.
Finally, good risk management – to the extent it avoids large losses – should be as well rewarded as the ability to generate large profits.
Models Models are essential in risk management. However:
• models should be used as tools – their output should be used to help make deci- sions, but no more;
• those making decisions using the output from models should understand the model’s capabilities and limitations;
• if models are used for a purpose other than that for which they were designed, they should be used with caution; and
• there are some risks for which simple models are better than complex ones – more complex models are not necessarily better.