Too big to fail

5 Results of the evaluation of the TBTF reforms

Have the policy changes that took place after the global financial crisis worked? Are banks now safer? Has the problem of TBTF been solved? In 2019, the Financial Stability Board started a large-scale ex post evaluation of its TBTF reforms to answer these questions.

The TBTF evaluation focused on the channels through which reforms are expected to operate: resolution reforms that provide public authorities with more options for achieving a resolution for banks, changes in the behaviour of banks, and changes in the pricing of bank risk in financial markets. Moreover, for the reforms to succeed, these mechanisms must be sufficiently strong to enhance financial stability.

It is important to note that successful reforms are not only measured against the frequency of financial crises or of actual bank failures. Rather, the purpose of the reforms is also to make sure that the impact of a crisis, if it happens, has been reduced. The greater the progress in implementing the reforms, the stronger these effects should be. The effects of resolution regimes work through implicit funding subsidies and the risk-taking decisions of banks. We can observe these effects even without observing an actual bank failing.

Let’s take a look at some of the findings of the FSB ex post evaluation.

Choices of banks

capitalization

The expansion of a bank’s capital base. Capital is the money that a bank has obtained from its shareholders and any profit that it has made and not paid out. Consequently, if a bank wants to expand its capital base, it can do so by issuing more shares or retaining profits, rather than paying them out as dividends to shareholders. Source: ECB.

The evaluation of the Financial Stability Board looked at several indicators that can provide information on bank behaviour. Figure 4 shows changes in the capitalization of systemically important banks, distinguishing three types of banks: globally systemically important banks (G-SIBs), domestically systemically important banks (D-SIBs), and a ‘control group’ of all other, smaller, non-systemic banks. Figure 4 shows banks’ capital relative to total risk-weighted assets and to total assets.

capital requirements

A rule requiring banks to meet or exceed a certain capital ratio. The ratio is calculated by dividing available capital by risk-weighted assets. This rule is necessary because banks take risks. For example, there is the risk of a borrower being unable to repay their loan (credit default risk) or financial market prices becoming unfavourable for the bank (market risk). In order to protect its creditors, the bank therefore needs sufficient equity capital to absorb any losses arising from this risk. Source: Bundesbank glossary. See also: leverage ratio (for banks or households).

Figure 4 shows that, for all banking groups, capital ratios have increased since the global financial crisis, which means that banks are now more resilient against shocks. This was demonstrated at the outset of the COVID-19 pandemic in 2020 when banks had buffers against losses.

The COVID-19 pandemic is different from the global financial crisis as the shock did not originate in the financial system. Governments and central banks intervened massively to support the economy. The FSB and the ESRB provide overviews of what happened.

The capital ratios of SIBs increased by more than those of other banks. So, in terms of the changes, their capitalization has increased faster. In terms of the levels, however, larger banks still have higher leverage, in other words, lower ratios of equity to assets. This may reflect differences in business models if, for instance, large banks are more diversified and therefore less risky.

And not only are systemically important banks better capitalized than before the reforms, they have also issued debt that can potentially be bailed in. In fact, most G-SIBs already (as of end June 2020) meet their 2022 final loss absorbing requirements described in Section 4.

profitability

A measure of the ratio of profit to capital invested. Profitability is sometimes used as a synonym for return on equity, which is calculated as the ratio of net profit to equity over a specified period, expressed as a percentage. Source: Bundesbank glossary.

The increase in bank capital has implications for bank profitability: profitability as measured through return on equity has tended to fall for systemically important banks relative to other banks. This may reflect higher capital, lower risk, and higher funding costs if implicit funding subsidies have declined. This might look like a negative side effect of the reforms: owners of banks or managers with compensation tied to the return on equity have lower incomes. But, from the point of view of society, the interpretation may be different: higher capital, lower risks, and lower implicit funding subsidies actually make the banks safer, lower the costs of potential bailouts for tax payers, and reduce distortions to competition. Hence, from a social perspective, lower bank profitability can very well be consistent with successful policies to address the TBTF problem.

bank lending

A loan where the lender (the bank) provides a borrower with a sum of money for a limited period of time. The borrower is obliged to pay interest to the lender. There are numerous types of loans, which are characterized by different maturities (date the final repayment is due), type and scope of collateral, or use (e.g. real estate credit, overdraft facility, instalment loan). Source: Bundesbank glossary.

Changes in bank lending are another example of why a distinction between the perspective of individual banks and the perspective of the overall economy is important. Banks that face higher capital requirements have different options to achieve compliance. They can increase their capital ratios by raising more capital or by retaining profits. But they can also reduce lending and shift their balance sheets towards activities with lower regulatory risk weights. This route is particularly attractive for banks that have low capital ratios in the first place.

However, when you step back a little and look at the overall provision of credit to the economy, a quite different picture emerges: overall credit has not declined, it has increased!

The increase in credit is one key finding of the FSB’s TBTF evaluation. The Bank for International Settlements (BIS) publishes detailed information on the evolution of global credit.

When you look only at the market shares of systemically important banks in terms of total bank assets or lending, you see a declining trend (Figure 5). However, smaller banks have taken over some of the business of larger banks, as have non-bank financial intermediaries such as asset managers and hedge funds. So customers’ access to credit has not suffered.

The important question for regulators is whether risks have shifted to other areas of the financial system. There is, in fact, a lot of work going on at the FSB to look at exactly this question and monitor risks in non-bank financial intermediation—including the risk that non-bank financial institutions such as asset managers become TBTF and need to be regulated accordingly.

maturity transformation

The practice of borrowing money short-term and lending it long-term. For example, a bank accepts deposits, which it promises to repay at short notice or no notice, and makes long-term loans (which can be repaid over many years). Also known as: liquidity transformation.

Another term used to describe non-bank financial institutions is ‘shadow banks’. Banks engage in maturity transformation, and they are highly leveraged: they have far less equity in relation to their assets than other firms. (Read Section 10.8 of The Economy 1.0 to learn more about maturity transformation.) These two characteristics mean that they are inherently fragile. Most non-bank financial institutions—pension funds for example—do not share these characteristics. But some do, and these ‘shadow banks’ have the greatest potential to give rise to systemic risk. The FSB, therefore, monitors risks in these parts of the system. It described in March 2020 the role of non-bank financial intermediaries in the aftermath of the pandemic shock in financial markets.

But the bottom line is that one needs to look at changes in banking markets and the effects of regulations through the lens of social costs and benefits, not private costs and benefits. Also, one needs to look at the overall economy rather than specific market segments, and to consider how the system evolves over time.

Systemic risk

The indicators of the TBTF problem discussed so far focus on the balance sheets of banks, their lending behaviour, and their funding. They do not really tell us much about the risk that banks impose on the financial system. Recall that banks can create negative external effects for the financial system that individual banks do not internalize: if banks experience a negative shock to capital, they may fail to anticipate that other banks have capital shortages, too.

Measures typically used in the academic literature to measure systemic risk are related to the description of the TBTF problem introduced in Figure 1 (Section 2). That figure shows that banks can be systemically important if they are highly linked to the financial system—directly or indirectly—and if they are insufficiently capitalized. And their impact on the financial system depends on the fragility of other banks that may run into difficulties at the same time. Systemic risk can therefore be measured through the gap between potential losses in the banking sector and the capital that is available to cover such losses, and the contribution of an individual bank to this gap. The information needed, therefore, is: how big are potential losses in the financial system? How much capital is there to cover those losses? And what is the contribution of an individual bank?

Answering these questions requires the definition of a relevant market. That relevant market can be the domestic financial system, a regional (for example, the European) financial system, or the global financial system. If markets are not very integrated across borders, it may suffice to look at the domestic economy. But if the degree of financial integration is high—as in Europe—a broader definition of markets is needed.

Using such measures, the FSB’s TBTF evaluation concludes that, overall, measures of systemic risk have remained relatively stable over time, while the impact of G-SIBs has declined somewhat. Because of its mandate to monitor financial stability risks at the global level, the FSB evaluation report focused on changes in cross-border lending. Cross-border connectedness can be measured through the claims that banks have on borrowers in other countries or regions. Such measures of connectedness indeed reached peak values at the onset of the financial crisis and, after a sharp drop in 2008, have returned to or surpassed their pre-crisis levels.

The V-Lab provides data and further information on systemic risk around the world. The Bank for International Settlements provides statistics on cross-border connectedness.

Implicit funding subsidies

There have been changes in the behaviour of banks, and new resolution regimes have been implemented. Has this changed implicit funding subsidies? Answering this question is not easy because implicit subsidies are not observable. Hence, we need indirect measures to find out whether the subsidies have actually been withdrawn by the reforms, as intended. An intuitive way of assessing changes in funding costs is to compare the funding costs of larger and smaller banks. But this simple comparison does not do the trick; there may be many other reasons why lending to larger banks is less risky than lending to smaller banks. Larger banks may have better management or be more diversified than smaller banks, and, if so, this could also make them less risky. So to observe the effects of systemic importance on funding costs, it is necessary to isolate all other factors that could contribute to lower risk.

credit rating

A classification of debtors or securities in terms of their creditworthiness or credit quality. These classifications are generally carried out by credit rating agencies. The highest creditworthiness rankings are denoted as AAA or Aaa by the most well-known agencies, while less creditworthy ratings are denoted with different combinations of letters and numbers. Credit ratings agencies consider the credit quality of debtors or securities rated BBB- or better as ‘investment grade’. Those with a lower rating are classified as speculative; such securities are also referred to as high-yield bonds. Source: Bundesbank glossary.

credit ratings agency

A firm which collects information to calculate the credit-worthiness of individuals or companies, and sells the resulting rating for a fee to interested parties.

One approach to estimate implicit funding subsidies is to exploit different types of credit ratings. Some credit rating agencies issue a ‘support rating’, in which they assess the likelihood that the government will support a bank if it fails. They also issue a ‘standalone rating’, which assesses the strength of the banks assuming no government support.

funding cost advantages

The difference between a financial institution’s actual funding cost and its hypothetical funding costs if it were not benefiting from an implicit guarantee. Funding cost advantages are a measure of the implicit government subsidy accruing to systemically important banks and thus the expected bailout that bondholders expect to receive. See also: implicit guarantees.

Find out more Implicit guarantees and funding cost advantages of systemically important banks

Because implicit guarantees are quite important, let’s take a closer look at how they can be estimated. Let’s start from the simplified balance sheet of a typical bank. The right-hand side consists of equity and liabilities. Liabilities, in turn, comprise deposits and bonds (see Section 11.5 of the Economy). Funding cost advantages can be estimated by looking at the interest rates that these bonds yield. The funding cost advantage (FCA) of a financial institution i in period t can be measured as the difference between its actual funding cost and its hypothetical funding costs if it were not benefiting from an implicit guarantee:

\[\begin{align*} \text{FCA}_{i,t} &= (\text{Actual funding cost of TBTF bank})_{i,t}\\ &- (\text{Hypothetical funding cost if not TBTF})_{i,t} \end{align*}\]

Funding cost advantages are a proxy for the implicit government subsidy accruing to systemically important banks, and are thus the expected bailout that bondholders expect to receive. The funding cost advantage can be decomposed into the expected value of the implicit government guarantee, given by the probability of default, and the expected recapitalization by the government (that is, the money that governments provide to cover losses). The first component is the joint probability that the bank is both in distress and being bailed out. The second component measures the losses due to bank distress with and without a bailout (the ‘loss given default’ or LGD):

\[\begin{align*} \text{FCA}_{i,t} &= Prob(\text{distress}_{i,t}) \times Prob(\text{bailout}_{i,t} \mid \text{distress}_{i,t})\\ &\times (\text{LGD if not bailed out}_{i,t} - \text{LGD if bailed out}_{i,t}) \end{align*}\]

The probability of failure depends on bank-level choices related to bank risk. The probability of being bailed out, given that a distress event has occurred, \(Prob(\text{bailout}_{i,t} \mid \text{distress}_{i,t})\), is influenced by a number of factors. Regulations affect how easy it is to deal with a failing bank. The fiscal situation also has implications for the ability of governments to bail out financial institutions and provide support to the financial system in times of crisis. Section 22.13 of The Economy 1.0 (Section 12.9 of Economy, Society, and Public Policy) discusses how fiscal capacity constrains the types of policies that authorities can implement.

Figure 6 illustrates the evolution of the funding cost advantage of SIBs based on the factor pricing approach. This approach compares the return on a portfolio of SIB stocks with the return on a portfolio of stocks of other banks, taking into account other risk factors. This measurement provides an estimate of the equity market’s perception of TBTF risk. What is important in this figure are the patterns over time. Prior to the global financial crisis, funding cost advantages of large banks were positive, but relatively small. During the financial crisis, funding cost advantages increased, due to a rise in both the perceived probability of default and the perceived probability of bailout, given default. Following the crisis, funding cost advantages remained at a relatively high level. But after TBTF reforms were implemented, the funding cost advantages fell.

Policy evaluation: challenges and information sources

Trends in banks’ activities and funding costs can arise for very different reasons, not necessarily because of the reforms. Good causal evaluations need to take into account other factors: post-crisis reforms include not only TBTF reforms, but also other financial sector reforms; monetary policy used unconventional instruments such as asset-purchase programmes; and bank-specific and country-specific factors matter for observed outcomes. There is, therefore, a great deal of variation in the data. This raises the question of what drives the differences: is it the regulations? Is it other, unrelated factors? Or is it just randomness in the data?

As there is no ‘gold standard’ that addresses all empirical problems, it is useful to compare different studies to see whether a consistent picture emerges. Many empirical studies have, for example, looked at the effects of Basel III. The repository Financial Regulation Assessment: Meta Exercise (FRAME) contains standardized estimates of the impact of policies on economic variables such as the provision of credit. It was set up by the Bank for International Settlements (BIS). The initial version of FRAME, launched in 2019, focuses on the effects of capital and liquidity standards implemented under the Basel III reforms. Empirical papers estimating the implicit funding subsidies of banks have been added to FRAME as well.

In addition, there are many useful sources for empirical work on international banking. The International Banking Research Network uses datasets for different countries and a common empirical approach to assess the effects of regulations, shocks, and other policy measures on globally active banks. The International Banking Library provides an overview of studies on the activities of international banks as well as links to data sets on banks and banking regulations.

The effect on its funding costs of a bank being systemically important (the ‘SIB effect’) is illustrated in Figure 7. In this figure, the distribution of impact estimates is based on 106 estimates from 19 studies corresponding to sample periods between 2007 and 2018. Negative estimates represent funding cost advantages for systemically important banks over other banks (the control group). These estimates correspond to unexplained funding cost advantages, which cannot be attributed to any macroeconomic, financial or business-related factors, and are therefore interpreted as implicit guarantees.

The size of the estimated funding cost advantage for SIBs ranges from around zero up to 3.5 percentage points. In terms of economic significance, one can compare the magnitude of the funding cost advantages with, for example, the Aaa corporate bond yield, to determine its economic significance. An Aaa bond is very highly rated and is thus considered to be of low risk. In the same period, 2007–2018, Aaa corporate bond yields ranged between 3.4 and 6.3 percentage points. If investors did not expect large banks to be bailed out, their funding cost could increase up to 3.5 percentage points, which would imply an increase of more than 100% (i.e. 3.5/3.4 = 103%) for a 3.4% bond yield. Estimated funding cost advantages can therefore be highly economically significant.

The evidence collected in FRAME provides a number of insights:

• Most of the impact estimates are based on data for the United States, Canada, and European countries. Other markets, in particular emerging markets, are not very well covered by the existing literature.
• It confirms the evidence shown in Figure 6: on average, funding cost advantages estimates peaked during the global financial crisis (2007–2009), as shown by comparing pre-crisis and post-crisis estimates. This is not surprising, because this was the period when implicit subsidies turned into explicit subsidies. Moreover, the estimates of funding cost advantages during the post-reform period are smaller than those for pre-reform periods, which would be in line with the objectives of the reforms (but it is not proof of causation).
• During the global financial crisis (2007–2009) and the European sovereign debt crisis (2011–2013), systemically important banks had, on average, greater credit rating uplifts than they did in the years 2015–2017. This can be interpreted as a stronger belief in external support during periods of financial instability.

Exercise 2 Evolution of SIBs’ funding cost advantages

Use the FRAME repository to answer the questions below.

1. The table for impact estimates of the FRAME repository allows you to visualize the results of many studies. If you click on the TBTF proxy menu, you can query the results for ‘SIB effect’ and ‘SIB x Reform effect’, among others. Look at the graphs and their descriptions, and explain what the terms ‘SIB effect’ and ‘SIB x Reform effect’ refer to. You can find an even more detailed description in the repository report, which is available on the overview page.
2. Under the Breakdown options menu, you find different ways of disaggregating the estimates. Use the Regimes option and compare the ‘SIB effect’ in the pre-reform and post-reform periods. Do you notice a difference? Describe how the estimates differ between the periods.
3. Now break down the estimates by sample year. What was the average funding cost advantage for SIBs in 2007? Did the funding cost advantages for SIBs increase or decrease in 2008 and 2009? Recall what you have learned about implicit guarantee expectations. Discuss how this might have had an impact on the funding cost advantage for SIBs after government bailouts in the years 2008–2009. You might want to filter out some of the years by clicking on the legend of the interactive graph.