# Chapter 1. The financial system outlook

Three major risks are set to shape the financial outlook:

1. Monetary policy normalisation: with the unwinding of ultra-low interest rates and the large-scale holdings of sovereign and private sector securities on central bank balance sheets, the transition may be volatile.

2. Financial sector vulnerabilities and the extent to which the recent finalisation of G20 reforms, including Basel III, has achieved the goal of a safe and sound financial system in the face of future stresses.

3. High indebtedness and leverage, especially related to China’s bank, shadow bank and wealth management businesses, and how well the Chinese authorities will be able to manage related risk.

All these risks have the potential to disrupt sustainable growth in the global economy. This chapter examines these three topics, concluding that financial system risk will be elevated in the period ahead.

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

## 1.1. Introduction

Since the global financial crisis of 2008, monetary policy has been focused on supporting both the financial system (in the early stages) and real economic activity in line with price stability objectives (in the later stages) with ultra-low interest rates and massive buying of debt securities by central banks, mainly in jurisdictions where global systemically important banks (G-SIBs) are located. Fawley and Neely (2013) states: “Initially, the Fed, BOE, BOJ, and ECB policies focused on restoring function to dysfunctional financial markets, but concern soon shifted to stimulating real growth and preventing undesirable disinflation”.1

Over two years ago, the US Federal Reserve (Fed) decided to begin the reversal of its accommodative policy stance via interest rates, given that the economy is operating closer to its capacity, which is when inflation pressures are more likely to arise. More recently, it has announced the reversal of quantitative easing. Inflation pressure has not yet emerged in the euro area and Japan, though the OECD Economic Outlook foresees a moderate pick up in the euro area over the next two years. While there have been signs of more inflation pressure in the United Kingdom, it has not yet begun the reversal of quantitative easing, since this is likely due to past depreciation of sterling.

The timing and scaling of reversals can also be linked to the recovery of the safety and soundness of the institutions that were at the centre of the crisis, and to any changes in the structure of markets that may warrant higher longer-term holdings on central bank balance sheets as a share of GDP compared to that which prevailed prior to the crisis. Banks in the United States have already reached that point which, coinciding with the sound shape of the economy, has supported the case for the Fed to announce a schedule for the gradual unwinding of large holdings of central bank assets. This shift to normalisation has already led to extreme movements in asset prices in the early part of 2018. This may be a foretaste of things to come, underlining the delicate balancing act required of central banks.

The beginning of monetary policy tightening in the United States in the fourth quarter of 2015 preceded the finalisation of the Basel III rules at the beginning of December 2017. The period of monetary support for the banking system would have been a good opportunity to fundamentally change the business models and governance of banks (to not mix commercial banking and investment banking, recommended by the OECD since the crisis). This opportunity was used only partially: capital requirements were raised but G-SIBs’ business models remained more or less the same as they were before the crisis. These banks have strongly defended their business models (limiting separation policies which lie outside the scope of the Basel process) but G20 reforms, including to over-the-counter (OTC) derivatives markets and liquidity requirements, offer mitigants to high degrees of interconnectedness.

The notional value of OTC derivatives, an indicator of the interdependence of G-SIBs,2 stood at USD 532 trillion in the second half of 2017, compared to USD 586 trillion in the second half of 2007, just prior to the crisis. This has fallen as a share of the recovered global economy over that period (by roughly one third). Credit default swaps– the most significant derivative type for interconnectedness – has fallen as a share of the total OTC derivatives market from 10.5% at its 2007 peak to 1.8% at end of 2017. In terms of OTC derivative counterparties, reporting dealers make up 15%, other financial institutions 80%, and non-financial customers 5%.

Clearing and margin rule improvements under Dodd-Frank and European Market Infrastructure Regulation (EMIR) have helped to reduce systemic risk for broker dealers, either by netting through Central Counterparty (CCP) clearers, or through margin requirements for un-cleared derivatives. CCPs cleared 60% of OTC derivatives by the end of 2017.

CCPs represent larger ‘nodes’ of interconnectedness, but initial design and subsequent implementation of work to strengthen CCP recovery and resolvability have helped address concerns on whether CCPs are sufficiently resilient. Nonetheless, default resolution and recovery will still depend on how members cooperate and the extent of panic in a future crisis.3

Thus the unwinding of central bank support for the interbank system is set to occur in an environment where interdependence risk has softened but remains. Other factors affecting conditions include:

• European banks have been reducing their non-performing loan (NPL) ratios but they are still too high.

• Markets have become used to easy policy, causing bonds and to some extent equities to become overvalued by end-2017.4

• Low yields have pushed some institutional investors into less-liquid, higher yielding securities.

• The US government has cut taxes which will increase its budget deficit, so that the private sector will need to absorb even more bonds in addition to the unwinding of central bank holdings accumulated during a period of quantitative easing.

Debt levels in the global economy have increased in a number of jurisdictions, and there is an interesting discussion of these issues in the latest OECD Economic Outlook (OECD, 2018). In last year’s Business and Finance Outlook (OECD, 2017a), there was some discussion about how bank regulation is shifting the structure of finance towards the non-bank financial system, both in OECD countries and in China. This year’s Outlook looks at this again in more detail with respect to China and at what the authorities there are doing about this issue. Off-balance sheet activities in China have expanded, causing debt levels and risk exposures to become dangerously high. In 2009, China expanded credit sharply to counter perceived potential effects of the crisis. Even though bank credit growth has since eased, responding to tighter policy, attempts to reduce a broader measure of credit growth have proved more difficult. This is because monetary policy tools such as interest rate ceilings, reserve requirement changes and quantitative credit controls can be avoided through off-balance sheet activity and market-based financing channels. Monetary policy tools may need to become more market oriented, given the ambitious policies authorities are pursuing to maintain solid growth and to reduce inequality in poorer regions, while also raising the ‘quality’ of GDP. This has resulted in greater use of macro-prudential tools and a number of other important reforms in Chinese monetary and regulatory policies.

The following sections cover:

• Section 1.2 – the prospects for interest rate normalisation and the unwinding of central bank securities in advanced economies, including the debate concerning how large central bank balance sheets should remain in the longer run.

• Section 1.3 – the finalisation of Basel III.

• Section 1.4 – developments in the Chinese financial system, including the high levels of debt, off-balance sheet activity and what Chinese policy is doing about the situation.

• Section 1.5 – general conclusions on the financial system outlook.

## 1.2. Prospects for the unwinding of ultra-low interest rates and central bank assets

The level of short (3-month) interest rates for selected countries where G-SIBs are headquartered is shown in Figure 1.1. The cut in rates to ultra-low levels from 2008 is unprecedented – particularly where negative rates apply. Only in the United States have interest rates begun to normalise. How fast the Fed will increase policy rates will depend on the outlook for growth and inflation, the strength of US bank balance sheets and the development of the US dollar exchange rate (since other central banks have not moved to raise interest rates in line). Only in the United States – where many of the securitised mortgages with Fannie Mae and Freddie Mac are held in conservatorship – is the economy strong enough, the banking system sufficiently profitable and the rebuilding of capital advanced enough to begin the normalisation of policy.

The size of central bank balance sheets for the United States, the euro area, the United Kingdom and Japan is shown in Figure 1.2. From around USD 3.2 trillion in total in January 2007, the central banks of the countries hosting G-SIBs increased this to some USD 15.0 trillion by the start of 2018. In fact, in order to deal with the crisis and exert targeted control over financing conditions for the broader economy (via the yield curve), central banks took USD 11 trillion of assets onto their balance sheets. Combined with low interest rates, investors have benefited from the liquidity-driven recovery in asset prices. Central banks, too, have benefited from an increase in the value of their holdings as a result of falling rates and tightening credit spreads for the large quantity of assets that they hold.

The raising of short-term interest rates toward more historically normal levels will have implications for bond markets, given the expectations theory of the yield curve.5 A return to historically normal interest rates will imply sustained lower bond returns from this point (see the bond simulation below in Figure 1.5). It would be reasonable to expect this adjustment to be accompanied by periods of volatility, as occurred in 1994 but for different reasons this time.6 With respect to central bank balance sheets, private portfolios will have to absorb the increased supply of assets into the financial system. The cash is obviously there to cover this mechanically and the economy is strengthening, but potential liquidity effects could prove to be significant.7 It remains to be seen:

a) whether banks and other financial institutions are ready to get along comfortably without the large liquidity cushions that have been the counterpart of central bank quantitative easing; and

b) how the post-crisis regulatory liquidity rules will affect behaviour in the face of central banks’ actions.

The bond market adjustment can have spill-over effects on equity markets though these on the whole do not appear to be excessively overvalued by historical standards. Figure 1.3 shows price-to-earnings (PE) ratios for the United States, the euro area and the United Kingdom in mid-2018.8 At end-2017, the US stock market was ripe for a correction, which duly followed in February 2018.9

The average price to expected earnings (PE) ratio is around 15 for most major equity markets, and is considered to be a ‘normal’ metric in the United States as it has mean-reverted to this level since 1881. Possibly reflecting the problems about banks’ assets in these markets and the implications for economic growth, PE ratios have become stuck at, or below, the level of 15 in the United Kingdom and Europe.

While equity markets have been a favoured asset class as interest rates declined over time, equity risk premiums are (surprisingly) not especially low. They have been able to maintain levels even somewhat higher than historical averages — seemingly making some allowance for excessively low bond rates when discounting future cash flows. The main risk to equities will be from an interruption to earnings growth from any significant slowing of the economy.

In the United States, where bank return-on-equity (ROE) measures are showing a healthy recovery across all large banks, economic growth is strong and tax cuts have been added to the mix. PE ratios expanded more than in other regions during the course of 2017. The Standard & Poor’s 500-stock index (S&P 500) average PE ratio for December 2017 was 18.5, not as high as the monthly peak of 24 at the height of the tech bubble at the end of the 1990s. A further correction of around 8% from the March level (based on forward earnings) would be required to return it to the normal PE ratio level of 15. These adjustments do not come about in a straight line. Bouts of volatility in both directions are usually present when markets are trying to find direction in a changing return environment. Derivative-based notes, algorithmic trading and index fund redemptions operate to increase the amplitude of volatility, particularly if views of investors and policy makers on the speed of policy normalisation and other economic issues are not in line. Some mechanisms, whereby index funds might play a role in increased volatility, are set out in Box 1.1.

The reversal of central bank policy cannot happen overnight. The impact on the bond market in expectation of this regime change has already been significant. This is why central banks will proceed in a gradual and well telegraphed way. Nevertheless, as discussed below, a number of disruptive mechanisms may come into play in the future.

Box 1.1. The potential of exchange trade funds (ETFs) to drive the US market

By the end of 2016, ETFs constituted 10% of US market capitalisation, 30% of daily trading volume and 20% of aggregate short interest (Ben-David et al., 2017). In the 6 May 2010 ‘flash crash’ 42% of US equity trading was in ETFs and this transmitted to underlying stocks and liquidity dried up (Borkevic et al., 2010). On 20 June 2013 ETF prices fell sharply due to the absence of arbitrageurs and authorised participants (Ben-David et al., 2017 and references therein). On 24 August 2015, ETFs experienced a run on prices causing them to move to steep discounts to their net asset value and to account for 83% of the trading halts (SEC, 2015). In February 2018, there was another significant test of the market. On this occasion, even with heavy trading, there were minimum net outflows from ETFs, with tight bid-ask spreads and apparently adequate liquidity.

Authorised participants (APs) and high frequency traders arbitrage ETFs vis their underlying stocks. For example, APs short sell ETFs and buy the basket when the former trades at a premium to the underlying. At the end of the day, the short sales are covered by delivering the basket to the ETF in exchange for ETF shares. The reverse process happens when ETFs trade at a discount to the underlying. The case of ETF selling can be seen by following the arrows in Figure 1.4.

There is a large amount of academic research on ETFs. Ben-David et al. (2014) find that stocks with higher ETF ownership display significantly higher volatility. The Pan and Zeng (2017) study of corporate bond ETFs over the periods 2004-2016 suggests that liquidity mismatch can lead to mispricing, consistent with the idea that authorised participant arbitrageurs may stay on the sidelines during periods of market stress, due to concerns that mispricing will widen. This then becomes self-fulfilling due to the absence of arbitrageurs. Agarwal et al. analyse the consequences of having an extra layer of trading in ETFs on top of the trading that occurs anyway in the underlying stocks. They use the 2015 episode which includes periods of trading halts to ETFs affected by extreme price swings, allowing them to look at situations with and without the extra layer of trading. Controlling for index and mutual fund activity, they find a commonality factor in liquidity between ETFs and the constituent underlying stocks driven by the above market arbitrage activity. The commonality was not present during trading halts. They conclude that the ETF market reduces the ability of investors to diversify liquidity shocks due to the increase in the commonality in liquidity of stocks included in ETF portfolios. This and other research suggest there are significant policy issues that justify further analysis to improve understanding about risks that could be associated with ETFs.

Market mitigants are already in place should a more substantial stress conditions occur, through the structure and design of ETFs and funds more broadly. The ability for ETFs to pay in kind, or in the worst case, to use other tools such as gates, fees, side pockets or freeze fund redemptions, could help calm volatility. But these scenarios have not been tested in an environment of a general rush to exit securities markets.

### Potential impacts on bond valuations

The Fed has given a good indication of how it intends to proceed in reducing its holdings of securities as interest rises. It intends to do this by gradually decreasing its reinvestment in maturing securities – it will reinvest only amounts that exceed pre-set caps. These caps then are an estimate of the extent of the decline of its holdings (Federal Reserve, 2017). Assuming the Fed does not offset any of these maturity-driven declines with other active policy actions, the declines are scheduled to be:

• Treasury securities: USD 6 billion initially, and increasing every three months in steps of USD 6 billion over 12 months until the decline reaches USD 30 billion per month (in the last quarter of 2018).

• For payments of principal from its agency securities and mortgage-backed securities, reinvestment caps start at USD 4 billion and increase every three months in steps of USD 4 billion until they reach USD 20 billion per month.10

• The USD 30 billion and USD 20 billion caps will then be held in place until the balance sheet arrives at the level necessary to implement monetary policy efficiently and effectively.

An OECD bond yield model is used to simulate some possible impacts of this scenario in order to quantify the overall extent of price adjustment that may become necessary in the bond market. The main inputs and outcomes are shown in Figure 1.5. The dotted lines are estimates of the equilibrium bond rate tendency based on the factors included in the model – these are not forecasts as such, as there is uncertainty about future Fed policy and no account can be taken of market liquidity and trading influences.11 The scenarios are intended to give a broad idea of the sort of adjustment that may lie in front of the bond market in the period ahead.

The uncertainty about Fed policy arises in terms of the ultimate size of the balance sheet best suited to running monetary policy ‘efficiently and effectively’. The argument at one extreme is to keep the balance sheet at current levels on the basis of three arguments:12

• That there is strong demand for safe short-term securities in the financial system and this should not be left up to the private sector to provide (implying there is no need to restore the interbank market to its pre-crisis mode of operation).

• The Fed’s balance sheet is too large to control the Fed Funds Rate by varying the quantity of reserves, but interest rates can anyway be set by fixing the rate paid on excess reserves.

• Banks will be more willing to borrow in a liquidity crisis without ‘stigma’ if the Fed balance sheet remains large.

The announcement of the Fed caps profile suggests this view is not the most likely scenario. There were significant effects on spreads between mortgage-backed security yields and Treasuries in 2008 and the interbank market stopped functioning due to the serious liquidity crisis at the time. The quantitative easing policy was designed at first to address illiquidity in markets (dysfunction) and later to stimulate demand with interest rates at the lower bound. In the United States, this situation has now been satisfactorily reversed—it therefore makes sense to reduce Fed holdings that were associated with these earlier periods. Moreover, the bond model estimated at the OECD has important portfolio channels. If more securities are held on the Fed balance sheet then this gives rise to a reduced risk premium on holding such securities in the private sector. In looking at this process the Fed states: “The primary channel through which LSAPs appear to work is the risk premium on the asset being purchased. By purchasing a particular asset, the Federal Reserve reduces the amount of the security that the private sector holds, displacing some investors and reducing the holdings of others, while simultaneously increasing the amount of short-term, risk-free bank reserves held by the private sector. In order for investors to be willing to make those adjustments, the expected return on the purchased security has to fall. Put differently, the purchases bid up the price of the asset and hence lower its yield. This pattern was described by Tobin (1958) and is commonly known as the ‘portfolio balance effect’.”13 The OECD model used is in this spirit.

One of the contributing factors of the 2008 crisis was a general underpricing of risk, a situation that policy makers might wish to avoid in the future. Keeping risk premia artificially low will encourage markets to take on more risk than they should and perhaps also to believe that a rescue package similar to that following the crisis might always be forthcoming should it be needed in the future. For these reasons reducing the holdings of securities on the Fed’s balance sheet in the manner suggested by the caps approach would see risk premiums rise gradually to more appropriate levels which, for this reason, seems to be a sensible way to proceed.

The bottom panel of Figure 1.5 shows the scenario for Fed securities holdings consistent with its caps to reinvestment as a share of GDP (GDP being assumed to grow at trend). Two scenarios are shown for 3-month interest rates and the holdings by the Federal Reserve of Treasury securities:

1. Moderate scenario: shown with dotted lines in the period ahead is a moderate scenario where the short rates follow a path implied by Federal Open Market Committee views as set out in the minutes for the March 2018 meeting. With respect to Fed holdings, these are assumed to follow the pattern implied by the above caps until they reach 10% of GDP in early 2021. The Fed, therefore, is assumed not to return to the lower levels that prevailed prior to the crisis.14

2. Bearish scenario: shown with the solid line for the period ahead is a bearish scenario for bonds (that would require a faster period of growth and inflation risk ahead than currently implicit in official views). The Fed funds rate moves to 4%, and short rates sit slightly higher. On this view, and for illustrative purposes only, it is assumed that the caps profile is followed until holdings fall to around the 5% level that prevailed prior to the crisis. Short rates are assumed to rise gradually to a maximum of 4%.

In both of the above cases foreign holdings are assumed to stay constant in nominal terms and decline gradually as a share of GDP.

• Under the moderate scenario, which is shown with the grey dotted line in the top panel of Figure 1.5, the equilibrium 10-year bond rate rises to 3.8%.

• Under the bearish scenario the bond rate rises to 4.3%.

Other scenarios may also be considered, but these quite different views give some idea of the ‘digestion’ issues that may lie ahead for the bond market. These rate scenarios imply a significant amount of price adjustment in the more liquid Treasury bond markets, but it is worth stressing that the implied price volatility could be much greater for other illiquid bond markets in the adjustment process.15

### Possible disruptive effects

A number of effects that could make the path of adjustment more turbulent:

• Mortgage-backed securities prices move more than Treasury Securities in a rising interest rate environment: This is because borrowers of fixed-rate mortgages are less likely to prepay loans in order to refinance at a lower rate, which extends the average maturity of the bonds, making them more volatile. This is sometimes referred to as convexity and it produces more volatility in Treasury and swap markets (see Box 1.2).

• Volatility and derivatives: When volatility increases, the potential for losses by investors, speculators and bank broker dealers in the derivatives market also rises. Hedging is imprecise in these vast markets and problems always arise. It might be recalled, for example, that even three years after the financial crisis had broken out, the Morgan Stanley joint venture with Mitsubishi UFJ Securities – despite Morgan Stanley’s expertise – incurred a USD 1.76 billion loss.16 This resulted from imperfectly hedged derivative portfolios exposed to a small movement of the term structure of interest rates.

• Large block trades in illiquid securities: Large pension schemes, insurance companies, mutual funds and hedge funds may need to deal in volume and respond to possibilities of withdrawals in the event of a sell-off affecting fixed interest portfolios. Low interest rates incentivise buying of higher-yielding illiquid securities and structured products. While funds have flexibility to manage redemptions, including liquidity management plans, large-scale attempts to unwind less-than-investment-grade illiquid holdings could cause extreme volatility. Some of the issues are summarised in Box 1.1.

• Higher regulatory costs for market making: The regulatory changes under Basel III, changes in derivative margin and collateral rules, the Dodd-Frank capital requirements and the Volcker rule process (see Box 1.4) increase the cost of broker-dealer market-making activities. This has caused some pull-back of participation in collateralised agreement markets, reducing liquidity. The increased cost of renting space on a broker-dealer balance sheet increases the cost of all the ancillary trading linked to broker-dealers.17

• Inflation: there are still weak signs of a pickup in inflation at the start of 2018. While forward-looking indicators, such as current consensus expectations for core US inflation, suggest the rate moving close to the Fed’s 2% target in the second quarter of 2018 and recent historical data for average hourly earnings in the United States rose slightly in the twelve months to February 2018 (to 2.6%), and modest productivity growth running at around 1.1% should keep unit labour costs at an under 2% pace. The US core Personal Consumption Expenditure Deflator as a fundamental guide remains weak thus far. Of course, this could change. Any sign of inflation in the United States would cause a more serious rout in the bond market, because it would be seen as steepening the tightening curve. UK inflation in 2017 was related to exchange rate effects and this contribution should drop out in the absence of further depreciation.

• Brexit and other trade issues: the uncertainty about the outcome of Brexit negotiations and the impact it will have on the future of the City continues to hang over financial markets at a time when monetary normalisation is taking place. The UK and Europe are strong trading partners and London is the largest agglomeration for trade in financial services. Other potential trade disruptions also need to be avoided to help smooth the transition.

• US Treasury holdings by China: The three biggest holders of US Treasury securities, other than the Federal Reserve, are China, Europe and Japan, as shown in Figure 1.6. The major presence of China in the US Treasury market is potentially problematic for the above bond rate scenario. China itself has a number of financial problems to deal with related to high levels of debt, much of which is linked to property developers and off-balance sheet shadow banking (see Section 1.4). Local authorities are heavily involved and state-owned banks are the ultimate providers of funds. While it is difficult to develop scenarios, financial stress at home could bring about changes in the way China manages its foreign reserves portfolio. China could withdraw funds by selling Treasuries to deal with business defaults in China, loss of foreign reserves due to proposed capital account opening or, as some market strategists suggest, for geopolitical reasons.18 This would most likely require some adjustment in the above reinvestment caps proposed by the Fed or in the path of future interest rates. Otherwise pressures on interest rates would become greater than envisaged in the above scenario.

Overall, the outlook for bonds and equities in 2018 and beyond is for much greater volatility and security price corrections. This began in early 2018, particularly in the United States, where valuation distortions for bonds and equities had become more extreme at the end of 2017. In part, the strong equity market in 2017 has been due to a better approach to dealing with the crisis that resulted in a stronger US economy and cutting the corporate tax rate, which caused increased enthusiasm in markets. Two roughly 1 000 point falls in the Dow Jones index in a single week during early February 2018 is indicative of two-way risks now being factored in.19 However, concerns remain about the longer-run resilience of the financial system following the re-regulation process (see Section 1.3).

Box 1.2. Mortgage bond convexity and fixed income volatility

The unhedged mortgage-backed securities (MBS), which the Fed will gradually push back onto the market, are very well known for their convexity. In other words, when interest rates fall, MBS prices rise less than Treasuries and, when interest rates rise, MBS prices fall by more than Treasuries. This happens because prepayments of mortgages rise in a falling interest rate environment (because of a lack of breakage costs for fixed-rate mortgages by borrowers), whereas extensions occur with rising rates (no-one wants to lock into higher interest rates so 30-year mortgages will be paid off much later than usual). The price of MBS moves less because it tends to shorter duration (i.e. expected maturity) in falling rates and by more because it tends to longer duration in rising rates.

The risk to portfolio values and structures is therefore very high in a rising rate environment and this leads fund managers to hedge their portfolios to keep duration at their required level. Duration hedging in a rising rate environment is achieved either by selling Treasury bonds, which directly exacerbates the effect of rising rates on Treasuries (makes the sell-off worse), or by dealing in the interest rate swaps market. In the swap market, this hedging can be achieved by paying fixed against floating – thus, when rates rise the fixed payer is protected from rising rates and the value of the contract rises with rates. Swap contract pricing therefore reflects the market’s expectation about future interest rates and is a useful tool for (speculative) investors. The floating rate payer faces a lot of risk in a tightening cycle and so the swap rate (i.e. the interest rate demanded by the floating rate payer) will rise if the outlook for tightening by a central bank like the Federal Reserve becomes more bearish. The spread between the swap rate and Treasuries reflects: (a) the (varying) credit spreads between banks and the safe sovereign bond (which widens in a crisis); (b) the supply and demand for swaps; and (c) swap market liquidity. There is an arbitrage to Treasury securities if the demand to be a fixed payer rises in a tightening situation.

In short, the duration hedging of MBS as short rates rise and the need to absorb more MBS in the market will exacerbate the rise in longer-term Treasuries and cause more volatility in the swap market. The impact may be softened in part by the limits on banks’ investment activities through the Volker Rule, and stronger capital requirements, which will weigh on wider demand for MBS as they return to the market. Central banks are broadly aware of this set of issues, but interest rates have been falling for a very long time and the test of how volatile this will prove to be lies ahead (e.g. Maltz et al., 2014).

### The need for liquidity in securities markets

There is considerable difference between the size and foreign holding composition of global fixed income markets. For example:

• In Q3 2017 the US fixed income market, according to the BIS, was USD 38.9 trillion with foreigners owning 26% of it; while for the Treasury market the equivalent numbers were USD 17.2 trillion with foreigners owning 36%.

• For Japan total debt was USD 12.6 trillion with government bonds at USD 9.4 trillion. Foreigners owned 11% and 6.1%, respectively.

• For China total debt was USD 11.2 trillion and sovereign debt was at USD 4.1 trillion. Foreigners owned 1.1% and 1.2% of these respectively.

The next largest markets for total debt are the United Kingdom (USD 5.8 trillion), France (USD 4.5 trillion), Germany (USD 3.6 trillion), Italy USD 3.3 trillion, Netherlands (USD 2.2 trillion) and Spain and Australia (both on USD 2 trillion). All of these latter markets are fairly open.

Central banks and sovereign wealth and pension funds would benefit from a wider choice of liquid investment grade-bonds around the world. The liquidity of a security is the ease with which it can be transacted without affecting its price, which requires market depth; i.e. that larger-sized orders do not move the market by very much. Even large bond markets can be relatively less liquid if the bulk of securities are tightly held and the free trading part of the market is small; or if capital controls are pervasive and only small amounts of bonds markets are tradeable by foreigners.

In the volatile environment expected in 2018 and beyond, as policy is normalised, liquidity considerations of very large pension and sovereign wealth funds may favour increasingly bonds denominated in the US dollar, euro and sterling, as well as large cap equities. This trend is likely to emerge in a significant part of the investor universe because large funds find it difficult to deal and manage to benchmarks in assets other than those that are very liquid. Norges Bank, which runs the largest sovereign wealth fund in the world, was recently asked by the Norwegian Ministry of Finance to examine all of its investment benchmarks with many of the above risks in mind. One of their key conclusions was stated as follows:

“Thus we find that the risk reduction that a long-term investor achieves by diversifying investments across countries and currencies differs between equities and bonds. In the long term, the gains from broad international diversification are considerable for equities but moderate for bonds. For an investor with 70% of his investments in an internationally diversified equity portfolio, there is little reduction in risk to be obtained by also diversifying his bond investments across a large number of currencies.

The benchmark index for bonds currently consists of 23 currencies. Our recommendation is that the number of currencies in the bond index is reduced. This will have little impact on risk in the overall benchmark index. We propose that the Ministry goes back to a specific list of currencies for the bond index rather than leaving this decision to the index supplier. The currencies on the list must be liquid and investable for the fund. The most liquid market for bonds is currently that for US Treasuries, followed by those for bonds issued by countries in the euro area and the United Kingdom. The Japanese bond market is large but far less liquid than those for the other currencies that currently have a substantial weight in the index. An index consisting of bonds issued in dollars, euros and pounds alone will be sufficiently liquid and investable for the fund” (Norwegian Ministry of Finance, 2017).

### Deepening and opening securities markets in Asia

The openness and depth of Asian bond markets in both local currency and dollar-denominated securities needs to be increased in economies that have reached an appropriate level of development, including by reducing restrictions on capital inflows and outflows that prevent building more liquidity in these markets—a view echoed by the former Peoples Bank of China governor Zhou at the 19th Party Conference (see section 1.4).

### Interest rate rises and the growth of corporate cash flow

Global policies and good practice offer measures to assist developing liquidity in Asian markets by focussing on market structures and specific investor and issuer activities.

For market structures such measures could include:

• increasing pre- and post-trade transparency to improve price discovery;

• derivatives markets to allow hedging of risk, and

• developing new electronic trading platforms.

Investor and issuer-focussed measures could include:

• standardisation of documentation and reporting;

• maturity dates that match liquid benchmark indices like treasuries;

• issuance that is eligible for inclusion in indices against which fund managers are measured; and

• initiatives that expand the investor base.

These measures, which require further progress in liberalising interest rates and foreign financial firm participation in the underwriting and market-making processes, would support the development of secondary markets. The process also requires opening up primary markets to foreign investors, including sovereign wealth and pension funds.

A look at what is going on in the corporate sector helps to explain the weakness of investment in the post-crisis period, which has been a factor in holding back the normalisation of monetary policy. Growth that permits rising real wages requires improving productivity. This, in turn, requires improved technology embodied in new capital expenditure. Since most investment is funded from cash earnings (not borrowing), it has been of concern that the growth of operating cash flow has been declining from the crisis to 2016, as shown in Figure 1.7. The data are based on a sample of 11 000 of the world’s largest listed non-financial companies.

Company earnings have been supporting equities and have grown faster than cash flow for some time, but this can be misleading. Whenever earnings run ahead of operating cash flow, one has to be concerned about their sustainability. This divergence usually occurs because of unsustainable elements, such as: changes in inventory (build-ups are a concern); transferring stock market fair value gains to the income statement; and gaps emerging between accounts payable and receivable. What is sustainable for growth is operating cash flow – which is always the backbone of capital expenditure.

In the latest data to 2017 there is some turnaround in cash-flow growth in the United States and Europe. Net interest costs and stable wages have been helping company cash flow, so it seems likely that the recent weakness evident in this firm-level data has been consistent with slowing demand and/or some margin compression. Technology and digitalisation are supposed to be revolutionising productivity – but these have not yet been sufficient to offset margin pressure trends. Indeed, to the extent that many companies are working in the digitalisation space (hardware, components, software, robotics, cloud computing, etc.), unit prices may be falling as these items are commoditised while the diffusion of the use of technology often takes many years (and sometimes decades) to lead to substantial productivity gains.

The signs of a cash-flow pick-up in 2017, if sustained, could lead either to more investment or allow firms to engage in increased share buybacks, though some estimates from private analysts using a different sample of companies show a more muted picture in 2017-18.20 It is of interest to note that cash flow is not yet accelerating in China where excess capacity in certain sectors has been centred. Increased pricing power in the US and Europe would support the case for more investment and a gradual normalisation of monetary policy. But much will depend on the progress being made in eliminating excess capacity in the global economy more generally, given the interconnected nature of supply chains.

## 1.3. Dealing with risks in the banking system through reforms

Jurisdictions differ in terms of the importance of capital market versus banking finance–the euro area and Japan are more bank-dominated and in the United States and the United Kingdom financing is more market-based, giving capital markets a greater role in financing the economy. Related to this, not all regions adopted the accepted wisdom for dealing with financial crises to the same extent. This requires the following four steps:

1. provide liquidity to the system;

2. guarantee deposits;

3. deal with the bad assets (get the bulk of them off balance sheets); and

Box 1.3. Risk-weighted assets: From Basel II to Basel III

Basel III attempted to deal with a number of issues related to flaws in the Basel II framework, and these were supported by work in the Financial Stability Board (FSB).The changes included inter alia:1

• The better quality and quantity of capital.

• Rules about how off-balance sheet items should be measured.

• The treatment of market risk and counterparty credit risk charges.

• Increasing so-called ‘bail-in’ bonds for resolvability (FSB TLAC work).

• Increasing the ‘granularity’ in the standard treatment of credit risk (more categories) and the risk sensitivity required in various models.

• The imposition of output floors (versus the standard approaches) which can help reducing problems associated with the internal ratings-based (IRB) approach.

• The singling out of G-SIBs for special treatment in the various leverage rules.2

• The treatment of credit valuation changes for counterparties in derivatives positions.

• The treatment of operational risk.

In revising Basel II, the approach has not been to abandon the basic model underlying the Basel framework, but instead to introduce more granularity and greater model sensitivity. Assumptions of the original Basel framework such as portfolio invariance and a single global risk factor have remained in the revised framework.3 Some constraints have been imposed to deal with the subjectivity of inputs, though this is still an issue that requires watching in cases where the advanced IRB approach is still present and where internal assessments are made.

1. Full details on these developments are on the BIS website: www.bis.org/bcbs/publ/d424.htm.

2. The RWA must be calculated as the higher of: (a) the approach that banks are approved to use (external ratings, standardised or IRB); or, (b) 72.5% of the RWA calculated by the standardised approach. This will apply to most aspects of Basel III: credit risk, counterparty credit risk, the credit valuation adjustment charge, market risk, and operational risk.

3. Portfolio invariance means that idiosyncratic risks can be identified, separated out, modelled and risk weighted (with different models depending on the levels of sophistication for obligors) to calculate capital charges for each that can be added up to derive total capital requirements. Because the Basel method applies to all global banks, this mathematical model also assumes a single global risk factor common to all banks (presumably the global macro cycle) – and has nothing to do with local jurisdiction risk issues.

Most jurisdictions achieved the first two steps, but the failure to take the third step in some jurisdictions has meant that central bank involvement in buying assets as liquidity policy has had to be larger and for longer to ensure that the interbank market continues to function (with large quantities of assets of uncertain quality remaining within the system).21 The euro area was partly constrained in dealing with bad assets by its state-aid rules.22 But the real issue has been its dependence on banking to finance the economy in the face of a significant quantity of bad assets. Thus, while banks’ non-performing loans (NPLs) have declined in most regions, they remain too high in some European economies and also in a number of emerging economies.

NPLs for large and small banks, distinguished by using a USD 50 billion asset base as a threshold, are shown in Figure 1.8. The size of the problem in parts of the euro area is larger than for other advanced economies, for both large and smaller banks. Since euro area growth has underperformed, and because the NPL problem there is significant, banks and policy makers have argued for a lower leverage ratio (LR) in order for their economy not to be disadvantaged compared with jurisdictions that have a larger role for capital markets.23 The United States, which followed the sequence of reform steps more rigorously, has had a stronger economy and has pushed through greater regulations onto its banks – particularly with respect to the LR via the Dodd-Frank Act.

While risk weighting has a role (see Box 1.3), it is at best a rough estimate of the sorts of losses that might apply in a crisis. The list of risks is large and has become even larger in recent years. These could include all manner of credit risks from traditional bank lending functions, but they also could include a vast array of risks from investment banking functions. Most of these risks arise from investment banking activities: they are interconnected in ways that differ in crisis periods compared to normal times and cannot be easily separated out and parameterised with portfolio invariance assumptions. These activities are not a significant part of traditional deposit banking in national and smaller regional banks.

The Basel III risk-weighting framework is not aimed at new risks that have been growing in recent years: misconduct risks and fraud, legal penalties for money laundering; operational risk, disruptive new technologies in the areas of blockchain and digital currencies and cyber risk.24 According to conduct costs data released by the CCP Research Foundation (2017), G-SIBs have paid USD 308 billion in fines between 2012 and 2016, and such issues have continued through 2017 and 2018. Global efforts to strengthen corporate governance and conduct are multiplying, including through the:

• OECD guidance on responsible business conduct for institutional investors (OECD, 2017b);

• FSB toolkit to strengthen governance frameworks to mitigate misconduct risk (FSB, 2018); and

• Growing adoption and expansion of individual responsibility and accountability and regimes, like those developed in the United Kingdom and Hong Kong, China.

Such efforts should help to improve conduct and compliance, and reduce losses and fines in the future, though it is still early days.

These risks add on to and interact with credit and interconnectedness risks and, like these latter risks, they cannot be predicted, risk-weighted and allowed for with any confidence. The Basel III approach to operational risk is illustrative. Operational risk is idiosyncratic and depends on trust, corporate governance, bank culture and due diligence. The size of bank income categories, multiplied by a history-of-past-loss parameters (the essence of the Basel III treatment), has little to do with these causes of operational and non-market losses.

### The leverage ratio

It is for these reasons that the OECD Secretariat has, on the basis of empirical research, argued that the simple leverage ratio should be given a more prominent role (see Blundell-Wignall and Atkinson, 2010). Banks prefer a RWA capital rule because it leaves them a degree of control over how much capital they need to hold.25 For the RWA ratios, the tougher Common Equity Tier 1 (CET-1) definition of capital is used, but the exposure measure in the denominator can be influenced by banks, thereby weakening its role as a binding constraint.

The LR is binding and can be used to ensure sufficient capital is held no matter where the origin of unpredictable losses. The LR limit in Basel III is, according to the analysis referred to above, too small; it uses Tier 1 capital in the numerator (not CET-1) and is less binding than it should be due to the nature of the deductions in the exposure measure. For the finalised Basel III, the LR is left at 3% for non-G-SIB banks, and introduces a LR buffer for G-SIBs set at 50% of their risk-weighted Higher Loss Absorbing Requirement (HLAR) versus RWA (see BIS, 2017).26 For G-SIBs, the required capital conservation buffer and the HLAR (along with the other requirements) is not onerous compared to the risks they run. The most risky G-SIB would have to hold only 10.5% of RWA as CET-1 capital and 4.75% of Tier-1 capital versus the LR exposure measure. The safest Basel-ranked G-SIB would face a LR of only 3.5%. Despite the greatest levels of support for banks in history, the 2008 crisis saw G-SIB losses well in excess of this range.

### Separation of investment banking from deposit taking businesses

OECD empirical evidence suggests by far the biggest issue for G-SIBs is the business model that mixes investment banking activities and extensive use of derivatives for creating complex investment products, and for regulatory and tax arbitrage purposes (Blundell-Wignall, Atkinson and Roulet, 2013). This creates leverage and contributes to financial risk. These uses of derivatives are not helping to fund the real economy (such as through simple hedging on behalf of clients). These activities should not be mixed with traditional deposit banking that cross-subsidises such risk-taking activities and causes a general under-pricing of risks.

Hoenig (2016) points out that US bank losses and Troubled Asset Relief Program (TARP) support amounted to 6% of their balance sheets in the crisis – notwithstanding the biggest monetary injections in history; the official large-scale buying of private sector assets; direct capital injections into banks; placing Fannie Mae and Freddie Mac assets into conservatorship and guaranteeing all of their assets; and the paying out of all AIG’s CDS liabilities to banks. Without these measures, the global financial system losses could have been larger.

OECD Secretariat research shows that the LR has a greater mitigating effect on G-SIB risk than the Basel RWA concepts, though this may change as Basel III comes into full effect.27

The United States and the United Kingdom have done a lot with respect to bank business models.28 The US Treasury has recommended further sensible alterations to the Volcker Rule following a review in 2017.

The EU countries, having dropped the Liikanen proposals (Liikanen, 2012), and Switzerland, having rejected separation from the outset, believe that a resolution regime is sufficient to deal with too big to fail (TBTF). In the United States, the Orderly Liquidation Framework for emergency circumstances and the requirement for living wills under the Dodd-Frank Act also goes in this direction. However, Duffie (2016) is sceptical of this approach at its current level of advancement because to have any chance of working, there are three basic requirements:

a) that the firm has enough bail-in securities;29

b) that the process of resolution should not lead to an early termination of contracts that the firm and its counterparties rely on for their financial stability across multiple jurisdictions; and

c) that policy makers act in a predictable and decisive manner.30

Some of these issues are being dealt with, but at the time of writing none have been settled in a satisfactory way. This is mainly because the amounts involved are large in a financial crisis:

• Deposits, covered bonds and other securities ineligible for bail-in are a large component of liabilities.

• The interconnectedness between G-SIBs is extensive across jurisdictions with different regulatory structures – and rules pertaining to the previous points differ.

More generally, the resolution approach to TBTF does not deal with reducing the subsidy to risk-taking for derivative activity in an ex ante sense. Indeed, the resolution of a bank through the bail in of unsecured bond holders makes it ex ante clear that counterparty positions of banks will be settled (even with any stay of early termination) at the expense of creditors first and the taxpayer second in cases where the seizing of collateral would be insufficient.

Box 1.4. The Volcker Rule and The Treasury Review

Section 619 of the Dodd-Frank Act, commonly referred to as the “Volcker Rule”, prohibits depository institutions from engaging in proprietary trading and from investing in covered funds (such as hedge funds and private equity) beyond small limits (Dodd-Frank, 2010). There are three tests for the blanket trading restriction:

1. instruments covered by the market risk capital rules cannot be traded for proprietary gain;

2. if the transaction would normally require the entity to be registered with the U.S. Commodity Futures Trading Commission (CFTC) or the Securities and Exchange Commission (SEC) as a dealer then it cannot be done (status test); and

3. the trade cannot be made for the purpose of short-term resale, benefiting from short-term price movements, realising arbitrage profits, and hedging any of the foregoing.

Treasury Review (2017), in response to an executive order, recommends removing this latter element because it is subjective and leads to conservatism and costly documentation.

The Volcker Rule ‘tests’ raised issues for market-making and, in the end, it was agreed that banks could continue this latter role provided that they did not build inventory beyond reasonable forecasts for serving expected client demand (which would cause excess inventory to arise).1 The Treasury review argues that this asks traders to forecast the impossible – particularly for illiquid OTC derivatives – and documentation is costly to firms. The review argues that such forecasts should not be required if the firm stands ready to buy and sell the instruments and if that appropriate hedging and compensation arrangements are in place.2

With respect to the ban on investing and sponsoring covered funds, the review suggests that it is too extensive and may not limit itself to hedge funds and private equity which are not defined clearly enough under the Volcker Rule. They argue that this might exclude the seeding of venture capital and other useful tools to support economic growth. It suggests longer seeding periods (from the current 1 year to 3 years).

The Treasury review recommends that all small institutions (less than USD 10 billion in assets) and even larger firms if they have less than USD 1 billion in trading assets (or the latter is less than 10% of its balance sheet) should be exempt from Volcker. The review goes as far as to suggest that any bank with an unweighted LR of 10% or more should be ‘off-ramp’ for the Dodd-Frank Act.3

The potential recommended changes follow on from the changes to section 716, which prohibit the granting of any US federal assistance to entities that are registered with the CFTC or SEC as swap dealers or major swap participants. In its original form, this rule would have limited a US insured depository institution’s exposure to the risks from engaging in such derivatives activities. Such activities were to be ‘pushed out’ into separate execution facilities. The provision was amended by being tacked on to a spending bill to ensure it passed both houses.4 Banks can now hold many important swaps on their balance sheet.5

The Treasury review rightly criticises market making restrictions. This is because the activity is about immediacy for clients (see Duffie, 2011). It is unreasonable to restrict quantities of inventories to levels forecast by some past rule of thumb about client needs. It is clear, too, that compliance costs are high, demonstrating ‘intent of a trade’ is difficult, and small uninvolved financial institutions (such as community banks) should be exempt.

1. Duffie (2016) points out that market making and proprietary trading are essentially the same thing.

2. This is consistent with Duffie’s (2012) views.

3. Consistent with the proposed Financial Choice Act 2017.

5. An exception is structured finance swaps, such as ABS swaps (see Warren, 2015).

It is worth noting that the idea of using Central Bank Digital Currency (CBDC) to manage systemic risk has gained traction in some jurisdictions, for example with Sweden’s eKrona project. CBDC could be made available to banks on a wholesale basis, or could take the form of electronic money made available directly to the general population, effectively providing retail deposit-taking accounts with a central bank. The constitutional referendum in Switzerland on the Vollgeld or ‘sovereign money’ proposal, which was held and failed to pass on 10 June 2018, might have achieved a similar outcome.

While such a transformation of monetary systems is unlikely in the immediate future, it is a longer-term possibility given the level of institutional interest. Retail CBDCs could call into question the nature and purpose of traditional deposit-taking institutions, particularly those that fail to separate deposits from exposure to risker business activities.

### Conclusions on banking reform

The RWA approach to controlling leverage is improved under Basel III but still allows G-SIBs to alleviate standard capital rules via their internal risk and pricing models. This has the potential to corrupt the risk management process by linking internal bank processes to regulatory capital charges. Banks may be obliged to minimise important risk weights to improve the return on equity for shareholders. The imposition of finalised Basel III output floors are a big improvement, but still not as binding on leverage as they might be. Under the final Basel III, G-SIBs can still achieve a 27.5% cut in RWA by using their models compared to the standard approach for all categories.31 The standard approach for banks is similar to Basel II for external ratings for banks and companies. For mortgages, risk weightings based on loan-to-value (LTV) ratios are an important improvement because they introduce more risk sensitivity, though these need to reach nearly 90% before the old 35% risk weight is surpassed.

Financial crises cannot be predicted, and to ensure that central bank and taxpayer support in a crisis like that of 2008 will not recur in the future, an important part of the overall reform process was the introduction of Total Loss Absorbing Capacity (TLAC) of 16% of RWA by January 2019 and 18% by 2022 (and to be 6% and 6.75% of the leverage ratio denominator by those dates). Total regulatory capital may be applied to TLAC (but not the regulatory buffers) and other instruments issued by the entity are then added to this. These instruments must be: ‘paid in’; unsecured; non-callable; not redeemable; not subject to netting; at least one year to maturity; and exclude exposures to other entities in the G-SIB group. 32 TLAC facilitates single point of entry resolution. These instruments may be ‘bailed in’ in the event of resolution issues (requiring an implementation trigger, such as a credit default swap (CDS) spread and other supervisory metrics and judgements). TLAC makes resolution a more credible outcome, and therefore helps to reduce the TBTF problem.

OECD research has found that implicit guarantees ultimately financed by taxpayers matter and they distort financial incentives. Making bank failure resolution more efficient and foreseeing the bail-in of bank creditors is a promising initiative, even if it might take time to see market expectations adapt to the new frameworks (see, for example, Schich and Toader, 2017).

Financial regulation has been improved, but has not been crisis-tested, and it remains to be seen to which extent it has made the financial system resilient to absorb shocks like those of 2007/2008. In this sense it may be too early to conclude that the regulatory reform process is finished.

## 1.4. Risks building in the Chinese banking system

A more immediate concern may be risks in the Chinese financial system, especially in the banking sector.

The Chinese banking system consists of:

• The big four commercial banks: the Bank of China (BOC); the Industrial and Commercial Bank of China (ICBC); the China Construction Bank (CCB), and the Agricultural Bank of China (ABOC).

• Three large policy banks: the China Development Bank (CDB) which focuses on funding infrastructure, housing policy and industrial SOEs; the Agricultural Development Bank of China (ADBC); and the China Export-Import Bank (ExIm Bank) for international trade. The CDB and the ExIm Bank are also heavily involved in funding the vast Belt and Road Initiative (BRI) projects, which has picked up since 2013.

• A large number of smaller state-owned commercial banks.

• A small sector of joint stock banks.

Figure 1.9 shows the structure of banking in billions of US dollars in the top panel and as a share of GDP in the bottom.

In 2009, China expanded credit to state-owned firms to support investment spending in order to avoid the impact of the 2008 global crisis on the Chinese economy. This spending was supported in part by the big four banks, whose joint balance sheets expanded from 98% of GDP in 2007 to 109% by 2010. However, the bulk of the credit expansion came from the smaller state-owned banks that fund activity in regional planning zones. These banks expanded their balance sheets from around 82% in 2008 to 103% of GDP by 2010, an increase of some 21% of GDP. By 2017, the balance sheet assets of the Chinese banking system had reached USD 39.3 trillion, or around 310% of GDP. This does not include the off-balance sheet exposures of banks. Shadow banking (i.e. entrusted loans, trust company loans and un-discounted bank acceptances), plus wealth management products (WMPs), add another 63% of GDP in exposures. This is not all, since these numbers do not include the Dai Cha market – a parallel repo market for banks, brokerages and wealth managers.

Since interest rates in China are subject to guidance and are not fully market determined, there is a potential for a financial repression scenario. Financial repression refers to a policy framework that keeps the return on saving below the rate of inflation, reducing the real debt burden and allowing banks to lend cheaply to companies (McKinnon, 1973). In China, this is usually accompanied by a monetary policy based around window guidance, credit ceilings, and changes in reserve requirements as some of its main instruments.33 After the expansion of 2009, China tried to rein in credit expansion at a time when financial demands on the economy were becoming greater and more complex. This has led to the boom in off-balance exposures in shadow banking and WMPs as shown in Figure 1.10.

At the end of 2017, bank balance sheets added to 310% of GDP and have increased only modestly since 2010. But off-balance sheet activity has taken the total to 387% of GDP. This comes about because there is a fundamental inconsistency between Chinese development, which is reaching a more diversified and complex stage, and the quantitatively-constrained banking system. Banks are able to bypass restraints with a variety of off-balance sheet mechanisms (an issue raised by Governor Zhou, see below).

Companies are able to lend to each other through three mechanisms that involve banks as intermediaries only, and to which reserve requirements and ceilings on interest rates and bank credit do not apply.

1. With bank acceptances, companies can issue a bill that instructs the bank to pay XYZ amount to company ABC, provides the money for the payment and the bank acts as a guarantor (the acceptance liability replacing the money paid to the third party). This remains off the balance sheet of the bank unless the bill is discounted.

2. The company can also lend XYZ amount to company ABC directly, with the bank again acting as intermediary only.

3. Banks also administer trust funds on behalf of individuals and entities and may lend funds.

In all cases, the off-balance sheet activity of banks is motivated by these lenders achieving higher returns than are available via controlled deposit rates.

The Chinese interbank market includes both bank and non-bank participants (such as wealth managers and brokerages) and is a pledged market. This has the benefit of avoiding re-hypothecation which is a key feature of advanced-economies repos – where the latter require the transfer of ownership of the securities sold. With a pledged market, the ownership does not change hands and the bonds cannot be lent or used in other transactions. But this demand for a ‘re-hypothecation mechanism’ finds an outlet in the ‘Dai Chi’ market: a large informal repo market. Dai Chi is not based on legally-enforceable contracts – a bond is sold for cash, but not on an exchange, and there is an informal agreement to buy it back at a later date and at an agreed price. The risk is that if prices fall below the agreed price, the seller may walk away from the agreement because it is not legally enforceable. Nevertheless, some rough estimates suggest that Dai Chi could be as much as double the size of the formal interbank market (CNY 12 trillion in 2015; Kendall and Lees, 2017).

The interconnectedness of all of these markets does not sit well with a financial repression approach to policy, as attempts to control credit of banks via quantitative restrictions simply pushes the activity into shadow banking, WMP activities and interbank and Dai Chi repo markets. Banks, asset managers (often owned by banks), brokerages, and industrial companies all interact in these markets. There are overlapping ownership structures and, because of the involvement of SOE banks with WMP vehicles, it is likely that investors assume that these products are implicitly guaranteed by the state. Like the events in the lead up to the crisis in advanced economies, such implicit guarantees mean that risk will be underpriced, leverage will build up and the risk of a crisis will grow. The assumption that the state will bail out investors in these products has been validated time and again: the Sealand Securities bond default in December 2016 is a good example. This brokerage firm, owned by a local government, sold bonds in the Dai Chi market which had fallen well below the agreed buy-back price. The brokerage walked away from the deal and, to avoid panic, the supervisor stepped in and arranged anonymous repos, sharing the losses between various counterparties.

The size of the off-balance sheet activity in China is materially larger as a share of the economy than the securitisation that played such a large role in the 2008 financial crisis. China is using its macro-prudential assessment (MPA) framework to try to deal with these off-balance sheet processes – mainly through window guidance – but it is proving difficult to control.

### Comparison of advanced-economy G-SIBs and Chinese G-SIBs

Figure 1.11 compares G-SIB assets in advanced economies with those of the big four Chinese state-owned commercial banks, plus the three policy banks. The top panel shows the balance sheet assets of the United States, the United Kingdom, the euro area and Japan compared to those of the large Chinese banks, from just before the crisis to the present. UK G-SIB assets peaked in 2008 and have declined since then, both in dollar terms and as a share of GDP. The euro area G-SIB assets peaked around 2011 and have declined since the euro crisis. Over this same period, Japanese G-SIB assets have been rising moderately as a share of GDP. China, with a more quickly growing economy, has seen the absolute size of its large systemic bank assets grow very rapidly versus those of advanced economies. In 2007, the seven Chinese banks had balance sheet assets of around USD 4.1 trillion and 110% of GDP. By 2017, this had risen to USD 15.3 trillion and 120% of GDP, larger in nominal terms than any other jurisdiction. Large Chinese bank assets constituted 189% of GDP in 2017, some 103% of GDP higher than in 2007 – with most of this gain occurring with the expansionary policy push in 2009. As a share of the economy, only the United Kingdom (due to the role of ‘the City’ and a much smaller economy) is larger.

Chinese banks have become heavily involved in margin lending, as the number of retail investors borrowing funds to invest in stocks has multiplied. Wealth managers have also increased leverage in the stock market. The equity in these vehicles is small and, to achieve the high returns expected by investors, they have not only increased lending to real estate developers but also to buy stocks directly. There is a cyclical element to this. Thus, in the stock market boom between mid-2014 and 2015, WMPs were buying stocks and putting equity owners (banks) at risk.

Figure 1.12 shows the weighted average (by total assets) of the distance-to-default (DTD) calculated for the big four commercial banks versus medium and small banks. DTD is an indicator of solvency movements and its changes versus historical highs and lows are the main focus of interest rather than its absolute level. It is an outside calculation of the true value of bank assets (as opposed to book value), and zero does not necessarily mean default as banks and regulators may allow forbearance in dealing with solvency issues.34

The overall stock market peaked in May 2015, but bank DTD calculations began to fall from late 2014, anticipating the growing risk in the banking system.

As the economy slowed and the stock market crashed, authorities stepped in with a range of policy measures to limit trading and stop the collapse, possibly reflecting their concerns about possible systemic risk to banks:

• In mid-June, many company shares were suspended from trading.

• At end-June, interest rates in the interbank market were cut, and in August the yuan was devalued.

• In early July, big holders of stocks, including WMPs, were banned from selling shares, margin rules were relaxed, and IPOs were suspended.

By August 2015, the big four commercial banks were close to the DTD default point similar to their position in the 2008 financial crisis. The stock market intervention, devaluation and fiscal expansion measures to avoid adverse effects on the Chinese economy eventually turned the situation around. But, without these, the 2015 crisis might have required large recapitalisations from the state. In other words, the Chinese financial system, with its vast off-balance-sheet vehicles, is extremely risky. In 2017, the DTD has begun to decline again from high levels, and it is vulnerable to economic slowdowns and default accidents in the shadow banking and WMP sectors.

### Chinese policy views and responses at the 19th Party Conference

It is very encouraging that then Governor Zhou Xiaochuan of the Peoples Bank of China (PBoC) spoke to many of the problems discussed above (see Zhou, 2017). He focused on the problems of structural imbalances in the real economy, excessive corporate debt, rising credit defaults in the bond market, inadequate corporate governance, lack of openness of the Chinese economy, and credit growth through innovative channels leading to the risk of financial bubbles and financial instability. In addition, he mentioned:

• Off-balance sheet financial activity was a contribution to the 2015 stock market volatility, and real estate price bubbles in some cities is related to excessive mortgage credit.

• Financial internet companies that are adopting strategies and products tantamount to Ponzi schemes (and without sufficient protection for consumers).

• Local government enthusiasm for growth is adding to financial leverage.

• Financial regulation is behind international standards, with a lack of clarity in responsibilities between different authorities.

• The financial system is not sufficiently market-oriented, and the lack of openness to the outside world leads to wasteful arbitrage profit trading and misallocation of resources.

This has been widely seen as heralding in important reforms many of which are already under way or implemented. While full implementation may take some time, some of the elements include:

• The China Banking Regulatory Commission (CBRC) and the China Insurance Regulatory Commission (CIRC), in order to better coordinate micro/institutional-level financial supervision functions, have been merged.

• Centralisation of macro-level financial regulatory functions in PBoC which will take a leading role in gauging systematic risks and financial regulatory reforms.

• To further strengthen China’s Macro Prudential Assessment (MPA) framework to include more financial products, in order to crack down on shadow banking activities and to provide safeguards against systemic risks. These have been already used to control leverage and off-balance sheet activities in 2017.

• Reflecting President Xi’’ desire to elevate risk prevention as a top priority, The State Council is to oversee financial stability and development, with a new Financial Stability and Development Committee playing a key role.35

In the longer run, Zhou suggests the need: to expand direct financing through better development of the capital markets (the more general need for which was suggested earlier in this chapter); to reduce intervention in the equity market; and to pursue capital account convertibility and the internationalisation of the yuan.

New Governor Yi Gang confirmed this orientation in a major speech in April 2018, announcing that China will encourage foreign investors to enter its trust, financial leasing, auto finance, money brokerage and consumer finance sectors. He also stressed that the opening-up should proceed apace with the reform in exchange rate formation mechanism, the advancement of capital account convertibility, and the reinforcement of financial regulation.

It would also be helpful to introduce greater transparency on China’s NPL issues. There has been some concern in the private sector that banks may shift weaker assets to WMPs with which they are connected so that official balance sheet NPLs may be understated. Clarifying and dealing with these issues, should they be found to be present would help to reduce financial stability risks.

## 1.5. Conclusions

There are parallels between China and advanced economies with respect to the interconnectedness of banks, wealth managers, the equity market and leverage, which spells a high level of risk for markets and institutions in the period ahead – though via mechanisms that reflect quite different levels of economic development. In advanced economies, interconnectedness is high because of the huge role of derivatives beyond that needed for hedging that helps fund the real economy, and because the transfer of securities in repo transactions allows re-hypothecation. Complex structured notes have been issued to clients seeking to beat the low-return environment. The risk in markets for large financial institutions and/or for investors is high. As experience of the crisis fades in the collective memory, bank separation proposals have been shelved, and the finalisation of Basel III has not given the leverage ratio the prominent role it successfully plays in the US regulation. Non-performing loan levels in Europe remain high, and their extent in China is obscured by the lack of transparency about which assets are sitting in off-balance sheet vehicles.

Against an improving economic backdrop, the US Federal Reserve is reversing low interest rate policies and is beginning to reduce their extensive holdings of securities. At some point other central banks will follow. To get back to 2007 levels, the central banks of the United States, Japan, the euro area and the United Kingdom would need to shed some USD 10 trillion in assets. It may be decided to hold more assets than prior to the crisis in the longer run, but even so the task of normalisation is likely to be on-going for years to come (especially since some central banks have not started yet). This process will, however, force shifts in asset allocations that will be associated with greater security price volatility in the period ahead. While China is not directly linked to these risks in advanced economies, due to the closed nature of its financial system, any problems there could see authorities shedding holdings of US securities which would increase liquidity pressures in advanced economies. Problems in China, should they arise, could be caused by credit events in the vast off-balance sheet exposures that have increased leverage risk.

The imbalances in global fixed-income market liquidity are problematic. Asian banks have a large presence in the United States and Europe, but in their own, often regulation-bound shallow markets, they are not playing an equivalent role in providing market liquidity to advanced economy investors. Liquidity events, volatility and complex market interactions via derivatives-based structured products (involving increased shadow banking firms) in advanced economies, together with the sizeable use of off-balance sheet vehicles in Asia, may combine to make for a more testing time in the period ahead.

## References

Agarwal, V., P. Hanouna, R, Moussawi C. Stahel (2017), “Do ETFs Increase the Commonality in Liquidity of Underlying Stocks?”, https://fisher.osu.edu/sites/default/files/etf_paper_osu_mar15.pdf.

Bank of England (2012), Review of the Bank of England’s Provision of Emergency Liquidity Assistance in 2008-09, as presented to the Court of the Bank of England, October.

Bank for International Settlements (2009), Comments Received on the Consultative Documents “Strengthening the Resilience of the Banking Sector” and “International Framework for Liquidity Risk Measurement, Standards and Monitoring”, www.bis.org/publ/bcbs165/cacomments.htm.

Bank for International Settlements (2017), “2017 list of global systemically important banks (G-SIBs)”. www.fsb.org/2017/11/2017-list-of-global-systemically-important-banks-g-sibs/

Ben-David, I., F.A. Franzoni and R. Moussawi (2018), “Do ETFs Increase Volatility”, Journal of Finance, forthcoming.

Ben-David, I., F.A. Franzoni and R. Moussawi (2017), “Exchange Traded Funds”, Annual Review of Financial Economics, https://0.annualreviews.org/doi/pdf/10.1146/annurev-financial-110716-032538.

Bernanke, B. (2016), “Should the Fed Keep its Balance Sheet Large”, Brookings Institution, Washington DC, September, www.brookings.edu/blog/ben-bernanke/2016/09/02/should-the-fed-keep-its-balance-sheet-large/.

Blundell-Wignall, A. and P. Atkinson (2010), "Thinking beyond Basel III: Necessary Solutions for Capital and Liquidity", OECD Journal: Financial Market Trends, vol. 2010/1, https://doi.org/10.1787/fmt-2010-5km7k9tpcjmn.

Blundell-Wignall, A. and Atkinson, P.E. (2008), “The Subprime Crisis: Causal Distortions and Regulatory Reform”, in Lessons from the Financial Turmoil of 2007 and 2008, Kent and Bloxham (eds.), Reserve Bank of Australia. www.rba.gov.au/PublicationsAndResearch/Conferences/2008/Blundell-Wignall_Atkinson.pdf.

Blundell-Wignall, A. and C. Roulet (2014), "Problems in the international financial system", OECD Journal: Financial Market Trends, vol. 2014/1, https://doi.org/10.1787/fmt-2014-5jxzmkg91s0t

Blundell-Wignall, A. and C. Roulet (2013), "Business models of banks, leverage and the distance-to-default", OECD Journal: Financial Market Trends, vol. 2012/2, https://doi.org/10.1787/fmt-2012-5k4bxlxbd646.

Blundell-Wignall, A., P. Atkinson and C. Roulet (2014), "Bank business models and the separation issue", OECD Journal: Financial Market Trends, vol. 2013/2, https://doi.org/10.1787/fmt-2013-5jzb2rhk9b6j.

Blundell-Wignall, A., Atkinson, P.E. and Roulet, C., (2012), “The Business Models of Large Interconnected Banks and the Lessons of the Financial Crisis”, National Institute Economic Review, No. 221.

Boissel, C. F. Derrien, E. Örs and D. Thesmar (2016), “Systemic risk in clearing houses: Evidence from the European repo market, European Systemic Risk Board, Working Paper No. 10, May.

Borkevec, M., I. Domowitz, V. Serbin and H. Yegerman (2010), “Liquidity and Price Discovery in Exchange-Traded Funds: One of Several Possible Lessons from the Flash Crash”, The Journal of Index Investing, Fall, vol.1 issue 2.

CBO (2018), “Bi-partisan Budget Act of 2018”, US Congressional Budget Office, www.cbo.gov/publication/53556.

CCP Research Foundation (2017), “Conduct Costs Project Report 2017”, London. http://foreigners.textovirtual.com/ccp-research-conduct-costs/274/221515/conduct-costs-project-report-2017-updated.pdf.

Dodd-Frank, (2010), Wall Street Reform and Consumer Protection Act, www.govtrack.us/congress/bills/111/hr4173/text

Duffie, D. (2016), “Financial Regulatory reform after the Crisis: an Assessment”, ECB Forum on Central Banking, 27-29 June 2016. www.darrellduffie.com/uploads/policy/DuffieSintraJune2016.pdf.

Dombrovskis, V.P. (2016), “Embracing Disruption”, speech to the European Banking Federation, 29 September.

Federal Reserve (2017), “FOMC issues addendum to the Policy Normalization Principles and Plans”, press release, 14 June, www.federalreserve.gov/newsevents/pressreleases/monetary20170614c.htm.

FSB (2018), Strengthening Governance Frameworks to Mitigate Misconduct Risk: A Toolkit for Firms and Supervisors, Financial Stability Board, www.fsb.org/2018/04/strengthening-governance-frameworks-to-mitigate-misconduct-risk-a-toolkit-for-firms-and-supervisors/.

Fawley, B.W. and C.J. Neely (2013), “Four Stories of Quantitative Easing”, Federal Reserve Bank of St Louis Review, vol 95, issue 1, January/February.

Gagnon, J., M. Raskin, J. Remache, and B. Sack (2011), “Large Scale Asset Purchases by the Federal Reserve : Did They Work ?”, FRBNY Review, May.

Haldane, A. (2012), “The Dog and the Frisbee”, speech given at the Federal Reserve Bank of Kansas City’s 36th economic policy symposium, “The Changing Policy Landscape”, Jackson Hole, Wyoming.

Kendall, R. and J. Lees (2017), “The Chinese Interbank Repo Market”, Reserve Bank of Australia Bulletin, June.

Liikanen, E. (2012), High-Level Expert Group on Reforming the Structure of the EU Banking Sector: Final Report; Brussels, October, http://ec.europa.eu/internal_market/bank/docs/high-level_expert_group/report_en.pdf.

Maltz, A.M., E. Schaumberg, R. Shiminov, and A. Strzodka (2014), “Convexity Event Risk in a Rising Interest Rate Environment”, Federal Reserve Bank of New York, March, http://libertystreeteconomics.newyorkfed.org/2014/03/convexity-event-risks-in-a-rising-interest-rate-environment.html.

McKinnon, R. I. (1973), Money and Capital in Economic Development, Brookings Institution, Washington, DC.

Medcraft, G. (2015) “IOSCO and FSB’s Work on Asset Management and Liquidity”, speech delivered at Investment Company Institute Global Asia Pacific Chapter Meeting, (Sydney, Australia), November. http://download.asic.gov.au/media/3450488/speech-to-ici-global-4-november-2015.pdf.

Norwegian Ministry of Finance (2017), “Bonds in the Government Pension Fund Global”, memorandum, 1 September 2011.

OECD (2018), OECD Economic Outlook, Volume 2018 Issue 1, OECD Publishing, Paris, https://doi.org/10.1787/eco_outlook-v2018-1-en.

OECD (2017), OECD Business and Finance Outlook 2017, OECD Publishing, Paris. https://doi.org/10.1787/9789264274891-en.

OECD (2017b), Responsible Business Conduct for Institutional Investors, Key considerations for due diligence under the OECD Guidelines for Multinational Enterprises, http://mneguidelines.oecd.org/rbc-financial-sector.htm.

Pan, K. and Y. Zeng (2017), “ETF arbitrage under liquidity mismatch”, European Systemic Risk Board, Working Paper No. 59, December.

Schich, S. and O. Toader (2017), “To be or not to be a G-SIB: Does It Matter? Journal of Financial Management, Markets and Institutions, Vol. 5, No. 2, pp. 169-192, www.jfmi.mulino.it/doi/10.12831/88826.

SEC (2015), “Equity Market Volatility on August 24, 2015”, Research Note, December, www.sec.gov/marketstructure/research/equity_market_volatility.pdf.

Treasury Review (2017), (Mnuchin, S.T. and C.S. Phillips), A Financial System that Creates Economic Opportunities: Banks and Credit Unions, Report to President Donald J. Trump, Executive Order 13772 on Core Principles for Regulating the United States Financial System, US Department of the Treasury.

UK Government (2011), Independent Commission on Banking: Final Report Recommendations, (Vickers Report), London, September, http://webarchive.nationalarchives.gov.uk/20131003105424/https:/hmt-sanctions.s3.amazonaws.com/icb%20final%20report/icb%2520final%2520report%5B1%5D.pdf.

Zhou, X. (2017), Report to the Nineteenth Party Conference, November, www.pbc.gov.cn/goutongjiaoliu/113456/113469/3410388/index.html.

Annex 1.A. Company data and sample description

Company data are based on the Bloomberg World Equity Index (BWEI). The sample includes all companies which have been listed in the BWEI over the period 2002-2017. 10 098 listed companies in 76 economies were selected (i.e. 6 506 in advanced economies and 4 592 in emerging economies according to IMF country group classifications) operating in 9 GICS industry sectors. Annual consolidated financial statements are collected on an annual basis, at the firm level and in current USD.36 The current primary source of this information is Bloomberg and some data are extracted from Thomson Reuters. All variables are winsorised at the 1st and 99th percentile levels to reduce the effect of outliers. Annex Table 1.A.1 presents the number of companies by country and sector.

To examine the financial characteristics of firms that succeed, the following financial variables are considered and are defined as follows:

• Capital expenditure: Amount the company spent on purchases of tangible fixed assets. It may include intangible assets when not disclosed separately.

• Operating cashflow: Amount of cash generated or used by a company in a given period solely related to core operations. It is calculated as the sum of net income, non-cash expenses (usually depreciation expense) and the changes in working capital.

• Return on equity (ROE): Ratio of net income to common equity. Net income is the profit after all expenses have been deducted. It includes the effects of all one-time, non-recurring, and extraordinary gains, losses, or charges. Common equity is the amount that all common shareholders have invested in a company.

• Cost of capital (COK): Weighted average (by the share of equity and debt in total assets, respectively) cost of equity and cost of debt.

Annex Table 1.A.1. Distribution of companies by economy and sector

Distribution by economy

Number of companies

Emerging economies

Number of companies

Australia

457

Argentina

17

Austria

25

Bahrain

2

Belgium

38

Bosnia-Herzegovina

14

808

Brazil

144

Cyprus (1)

22

Bulgaria

25

Czech Republic

6

Chile

43

Denmark

43

China

1407

Estonia

4

Colombia

14

Finland

48

Croatia

51

France

205

Egypt

35

Germany

208

Gabon

1

Greece

79

Hungary

9

Hong Kong, China

129

India

971

Ireland

30

Indonesia

114

Israel (2)

46

Jordan

9

Italy

91

Kenya

3

Japan

1099

Korea

515

Latvia

7

Kuwait

20

Lithuania

9

Macedonia

4

Luxembourg

6

Malaysia

226

Malta

5

Mexico

55

Netherlands

58

Montenegro

1

New Zealand

18

Morocco

8

Norway

37

Oman

5

Portugal

19

Pakistan

21

Singapore

57

Peru

17

Slovakia

8

Philippines

29

Slovenia

13

Poland

190

Spain

65

Qatar

9

Sweden

149

Romania

138

Switzerland

81

Russia

120

Chinese Taipei

192

Saudi Arabia

53

United Kingdom

365

Senegal

1

United States

2079

Serbia

39

South Africa

81

Sudan

1

Thailand

50

Turkey

98

Ukraine

21

United Arab Emirates

15

Venezuela

2

Vietnam

14

TOTAL

6 506

4 592

Distribution by sector

Sector

Emerging economies

Energy

607

220

Materials

838

870

Industrials

1 412

1 045

Consumer discretionary

1268

918

Consumer staples

402

447

Healthcare

653

308

Information technology

988

468

Telecommunication services

110

93

Utilities

228

223

1. Note by Turkey. The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Turkey recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey shall preserve its position concerning the “Cyprus issue”.

Note by all the European Union Member States of the OECD and the European Union. The Republic of Cyprus is recognised by all members of the United Nations with the exception of Turkey. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.

Source: OECD compilation.

Annex 1.B. List of economies by group

Two groups of economies are defined considering the IMF country group classification: advanced economies and emerging and developing economies.

Australia, Austria, Belgium, Canada, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hong Kong (China), Iceland, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, Luxembourg, Macau, China, Malta, Netherlands, New Zealand, Norway, Portugal, Puerto Rico, San Marino, Singapore, Slovak Republic, Slovenia, Spain, Sweden, Switzerland, Chinese Taipei.

## Emerging and developing economies

Afghanistan, Albania, Algeria, Angola, Antigua and Barbuda, Argentina, Armenia, Azerbaijan, Bahamas , Bahrain, Bangladesh, Barbados, Belarus, Belize, Benin, Bhutan, Plurinational State of Bolivia, Bosnia and Herzegovina, Botswana, Brazil, Brunei Darussalam, Bulgaria, Burkina Faso, Burundi, Cabo Verde, Cambodia, Cameroon, Central African Republic, Chad, Chile, People’s Republic of China, Colombia, Comoros, Democratic Republic of the Congo, Republic of the Congo , Costa Rica, Côte d’Ivoire, Croatia, Djibouti, Dominica, Dominican Republic, Ecuador, Egypt, El Salvador, Equatorial Guinea, Eritrea, Ethiopia, Fiji, Gabon, Gambia, Georgia, Ghana, Grenada, Guatemala, Guinea, Guinea-Bissau, Guyana, Haiti, Honduras, Hungary, India, Indonesia, Islamic Republic of Iran, Iraq, Jamaica, Jordan, Kazakhstan, Kenya, Kiribati, Kosovo, Kuwait, Kyrgyzstan, Lao People’s Democratic Republic, Lebanon, Lesotho, Liberia, Libya, Former Yugoslav Republic of Macedonia, Madagascar, Malawi, Malaysia, Maldives, Mali, Marshall Islands, Mauritania, Mauritius, Mexico, Federated States of Micronesia, Republic of Moldova, Mongolia, Montenegro, Morocco, Mozambique, Myanmar, Namibia, Nauru, Nepal, Nicaragua, Niger, Nigeria, Oman, Pakistan, Palau, Panama, Papua New Guinea, Paraguay, Peru, Philippines, Poland, Qatar, Romania, Russian Federation, Rwanda, Samoa, Sao Tome and Principe, Saudi Arabia, Senegal, Serbia, Seychelles, Sierra Leone, Solomon Islands, Somalia, South Africa, South Sudan, Sri Lanka, Saint Kitts and Nevis, Saint Lucia, Saint Vincent and the Grenadines, Sudan, Suriname, Swaziland, Syrian Arab Republic, Tajikistan, United Republic of Tanzania, Thailand, Timor-Leste, Togo, Tonga, Trinidad and Tobago, Tunisia, Turkey, Turkmenistan, Tuvalu, Uganda, Ukraine, United Arab Emirates, Uruguay, Uzbekistan, Vanuatu, Bolivarian Republic of Venezuela, Viet Nam, Yemen.

## Notes

← 1. See Fawley and Neely (2013). It should be recalled that in the early phases the aims were to reduce market stress After Lehman Brothers failed there was a run against money market mutual funds (MMMFs) as the buck was broken (values fell below a \$1 invested). The Asset-backed Commercial Paper Money Market Mutual Fund Liquidity Facility was launched, providing loans to depository institutions to buy high quality asset-backed commercial paper from MMMFs that needed cash to meet redemptions. Loans to SPVs followed and, on 25 November, the Term Asset-Backed Securities Loan Facility was launched: collateralised loans to eligible investors to buy asset-backed securities. In Europe, the ECB president made the early motivation clear: “the idea is to revive the market, which has been very heavily affected, and all that goes with this revival, including the spreads, the depth and the liquidity of the market. We are not at all embarking on quantitative easing” (www.ecb.europa.eu/press/pressconf/2009/html/is090507.en.html). Finally, the Bank of England states: “the Bank’s operations since 2009 under its Asset Purchase Facility, designed to improve the functioning of the commercial paper and corporate bond markets……. are sometimes characterised as the central bank acting as market-maker of last resort, but can be viewed essentially as just one of a gamut of actions a central bank may need to contemplate to address strains on the financial system” (Bank of England, 2012).

← 2. Note that other indicators are available. For example, the indicator-based measurement approach developed by the Basel Committee on Banking Supervision considers intra-financial system assets, intra-financial system liabilities, and securities outstanding at the firm level to assess the interconnectedness of banks.

← 3. CCPs concentrate interconnectedness of capital and liquidity. In the event of member defaults, matched books have to be re-established without suspension of activity which can be difficult in a panic event. The CCP operates as a waterfall: the defaulting member’s initial margin and default fund contributions are absorbed first, followed by CCP capital, and then non-defaulting member contributions. This is where the interconnectedness bites. Many actions are possible including tearing up contracts with defaulting members, auctioning contracts, forced loss allocation by the regulator, etc. There are extra-territoriality issues too, as jurisdictions have different views. In all cases co-operation between members and regulators will be essential.

← 4. This is perfectly consistent with the comment later in this chapter that equity risk premiums (ERPs) are not especially low. The cost of equity in a simplified model is e/p+g=r+ERP, where e is earnings (dividends), p is the level of the stock price, g is trend earnings growth and r is the risk free rate. When the risk free rate is very low the stock price may be very high, the cost of equity low, the ERP high and stock overvalued compared to normalised risk free interest rates.

← 5. Under the expectations theory of the term structure, for example, the long rate is the sum of expected future short rates plus a risk premium and equilibrium in this sense needs to be maintained. From a trading point of view, if short rates were expected to rise due to inflation expectation increasing, then longer duration selling would lead to a ‘bear steepening’.

← 6. The 1994 bond market sell off was caused by complacency on the part of investors about the likelihood of a tightening scenario. While this is less likely in cash markets this time, due to a more careful focus on Fed communication, the composition of investment is very different. There has been a heavy substitution into lower grade and illiquid fixed income markets. This presents new forms of bond risk for the normalisation of rates.

← 7. Because interest rates are very Iow in safe assets (cash and government bonds), fund managers and pension funds have moved into illiquid longer-term assets, including emerging market debt—as documented in previous Business and Finance Outlooks. These markets are easy to get into, but would require large price discounts in a bear market scenario. This has been a major concern at the Financial Stability Board (contingency plans for how to deal with a redemption scenario).

← 8. Price-to-earnings (PE) ratios are proxied by using the IBES S&P 500 Composite 12-month forward for the United States, the IBES MSCI Europe 12-month forward for the euro area and the IBES FTSE UK 12-month forward for the United Kingdom.

← 9. PE ratios can become higher than sustainable when interest rates are very low and earnings are in the process of returning to normal growth. See note 4. This point coincides with the value of market capitalisation rising versus GDP, the longer-run ‘Buffet’ valuation metric.

← 10. The caps are not binding until maximum caps are reached.

← 11. Specifically, it is the co-integrating equation estimated in Blundell-Wignall and Roulet (2014) using up-dated data until end-2017. This equation is found to be co-integrated with error correction tests. A risk premium shift crisis dummy variable is included from 2008.

← 12. See Bernanke (2017), and the paper and views reference therein.

← 13. See Gagnon et al. (2011), p. 3. The LSAP acronym is Large Scale Asset Purchases.

← 14. See Gagnon et.al (2011) for a discussion of why and how QE was implemented, and Bernanke (2017) for a discussion of the case for a larger than historical balance sheet.

← 15. A 10-year on-the-run Treasury with a coupon of 2.7% could expect to lose 9% of its value on the more moderate scenario and 13-14% of its value on the more bearish scenario. Mortgage securities and junk bonds could expect to lose more than this.

← 16. See Wall Street Journal: www.wsj.com/articles/SB10001424052748704071704576276350338480020. Many other cases may be cited, including the case of Societe Generale for even larger amounts.

← 17. Duffie (2016) makes this argument with respect to the US supplementary leverage ratio.

← 18. Zhou (2017) refers to significant corporate leverage (above international benchmarks for ‘problems’ and defaults). He also presents the unusually strong view to open the capital account. His views are set out in section 1.4.

← 19. The correction in the US stock market triggered an increase in volatility -- as measured, for example, by the Chicago Board Options Exchange (CBOE) Volatility Index, VIX. Algorithmic traders were able to switch positions to the short side very quickly, exacerbating the sell-off and the spike in the VIX. This hurt investors in exchange traded products that had been betting on low volatility of that index continuing. Investment banks, having spotted business opportunities, had created products to give investors better returns in the low-interest-rate environment. They were betting that the VIX would stay low because it had stayed low for such a long time. These structured notes allow the investor to buy the future at a discount to a future date (normal backwardation), and take advantage of the roll at maturity every month or quarter. But when volatility spiked (in what was probably the largest 1-day spike in history of the VIX from 17 to 37, a 116% rise, on 5 February 2018), there was a rush to get out causing covering buy-trades and ‘whipsaw’ effects in futures.

← 20. Societe Generale’s top-rated analyst is reported (in early 2018) to be focusing on declining growth in operating cash flow in the United States as a negative signal for equity markets and earnings, see for example: www.marketwatch.com/story/why-a-slowdown-in-cash-flow-growth-is-bad-news-for-stocks-2018-01-22.

← 21. Bernanke (2016) sets out quite neatly the obvious reason for all of the bond buying following the crisis in terms of the lender-of-last-resort function. In a liquidity crisis banks would have to dump assets at any price (fire sales), exacerbating their losses. There was a need to lend cash against assets such as treasuries and guaranteed mortgage bonds. Very clearly this has nothing to do with inflation and the usual channels for monetary policy. A debate has emerged about the extent to which this can or should be removed.

← 22. OECD (2016) provides an excellent discussion of options to deal with NPLs and the circumstances when state aid rules might be adjusted due to ‘serious economic disturbance’ considerations.

← 23. There are numerous examples of this in European bank submissions in the Basel III consultation process. Policy makers usually support their national champions without making it too explicit. But for Europe, Dombrovskis (2016) clearly stated the competitiveness objectives versus other regions, like the United States, as the main reason for objecting to output floors: “We want a solution that works for Europe and does not put our banks at a disadvantage compared to our global competitors.”

← 24. Blockchain and digital currencies are difficult to factor into a financial outlook for the year ahead. There may be surprises, but the effects are expected to occur of a longer run horizon. Banks are experimenting with the new technologies, including central banks. But the share of transactions via these channels is still very small. In next year’s Business and Finance Outlook these issues for the longer run will be taken up in more detail.

← 25. This is evident in bank submissions during the consultation processes undertaken in the Basel III process. See Bank for International Settlements (2009). The most common theme is support for risk weighting and criticism of simple leverage ratio definitions. In previous publications the OECD has drawn attention to letters to clients of major banks pointing to the ability to reduce risk weighted assets as targets linked to return on equity.

← 26. For the finalised Basel III, the LR is Tier 1 Capital divided by an exposure measure. The final version of Basel III leaves the permitted ratio at 3% for non-G-SIB banks, and introduces a LR buffer for G-SIBs set at 50% of their risk-weighted Higher Loss Absorbing Requirement (HLAR). For the HLAR, G-SIBs are divided into 5 buckets (5 is worst, through to 1 the best). By 2019, banks must hold a base 4.5% CET-1 ratio versus RWA plus a capital conservation buffer (CCR) of 2.5%, for a total on average over the cycle of 7%. The HLAR for G-SIBs consists of: 3.5% additional CET-1 capital for the level-5-assessed G-SIB bucket, then 2.5%, 2.0%, 1.5% and 1% (for the level-1 bucket). The most risky G-SIB would have to hold only 10.5% of RWA as CET-1 capital. For the leverage ratio base of 3%, (the less demanding Tier 1 versus the exposure measure), this translates to 4.75% only for the riskiest G-SIB, and then 4.25%, 4%, 3.75% and 3.5% (for the level 1 ‘safest’ G-SIB). Dividend distribution constraints apply to banks below any of these requirements.

← 27. See Blundell-Wignall, Atkinson and Roulet (2012) and Haldane (2012). The 2017 update of this work shows that these results have remained stable for G-SIBs. This evidence is also consistent with the idea that G-SIB risk can be reduced by separating certain investment banking functions for those G-SIBS with high interconnectedness risk, as in Blundell-Wignall and Roulet (2013).

← 28. Only a few European countries have implemented aspects of the initial Liikanen separation proposal, and the United States has reduced the impact of the ‘swaps push-out’ aspect of the Volcker Rule—only a limited number of swaps are subject to the rule (https://dealbook.nytimes.com/2013/05/23/banks-lobbyists-help-in-drafting-financial-bills/). The United Kingdom has introduced a ring-fencing policy (the so-called Vickers reform), see UK Government (2011) and is closest to early OECD proposals, which were referenced in the interim consultation process. It aims to ring-fence the domestic retail business from international financial shocks and to limit taxpayer costs for losses given default and applies to all banks with more that GBP 25 billion of core retail deposits. The rules come into full effect in 2019.

← 29. In the case of the EU/UK a statutory bail-in power applies to a broad range of liabilities subject to certain exclusions.

← 30. The Bank of Portugal case, with respect to the resolution of Banco Espirito Santo into a bad bank and a new ‘good bank’ Novo Banco, is illustrative of how EU law works in practice. Specifically chosen international investor’s bonds (not European) were bailed in from the new bank to the bad bank, causing them to collapse in value (because they were not backed by good assets). Litigation on a range of issues, including EU competition law, is currently under way. The summarised history of this saga can be found in: https://ftalphaville.ft.com/2018/01/19/2197893/the-novo-banco-debacle-and-the-rule-of-law-in-europe/.

← 31. Which is only a little less than compared with Basel II regulatory relief from the IRB approach. This compares to the 2004 quantitative impact study for Basel II which showed a regulatory benefit for risk weights achieved by using the IRB approach (compared to the standard approach) of -21.9% for banks and financial firms; -61% for residential mortgages; -6.5% to -74.3% for retail; and -21.9% to -41.4% for corporates and commercial real estate (Blundell-Wignall and Atkinson, 2008).

← 32. Deposits, derivative liabilities, debt with derivative links (like structured notes), tax liabilities, encumbered securities, and securities where bail-in can be legally challenged are all excluded.

← 33. China, though it is experimenting with an interest rate corridor approach in the interbank market.

← 34. The distance to default (DTD) is an important measure for monitoring banks because while the book value of liabilities can be observed the true value of bank assets cannot. In principle if the value of assets falls below liabilities the banks is in a default position, but only the reported book value of assets is reported not their saleable value. The formula to calculate the distance-to-default is derived from the option pricing model of Black and Scholes (1973) and is set out as follows:

${\mathrm{D}\mathrm{T}\mathrm{D}}_{\mathrm{t}}=\frac{\mathrm{log}\left(\frac{{\mathrm{V}}_{\mathrm{t}}}{{\mathrm{D}}_{\mathrm{t}}}\right)+\left({\mathrm{r}}_{\mathrm{f}}-\frac{{\mathrm{\sigma }}_{\mathrm{t}}^{2}}{2}\right).\mathrm{T}}{{\mathrm{\sigma }}_{\mathrm{t}}\sqrt{\mathrm{T}}}$

Where: is the Market value of bank’s assets at time t; ${\mathrm{r}}_{\mathrm{f}}$ , is the risk-free interest rate; is the book value of the debt at time t; , is the volatility of the bank’s assets at time t; and $\mathrm{T}$, is the maturity of the debt. However, the market-value of assets (Vt) and its volatility ( ${\mathrm{\sigma }}_{\mathrm{t}}$) have to be estimated. Equity-holders have the residual claim on a firm’s assets and have limited liability. Equity can be modelled as a call option on the underlying assets of the bank, with a strike price equal to the total book value of the bank’s debt. Thus, option-pricing theory can be used to derive the market value and volatility of bank’s underlying assets from equity’s market value (VE) and volatility ( ${\mathrm{\sigma }}_{\mathrm{E}}$), by solving some further equations (see Blundell-Wignall and Roulet, 2013). Obviously it follows that an actual default and the DTD at zero are not the same thing because banks and regulators may allow forbearance in dealing with solvency issues. Hence the DTD is an indicator of solvency issues and its changes versus highs and lows historically is of some interest for analysts monitoring banks without inside information.

← 36. The items on the balance sheet represent stock variables, and elements from the income statement as well as the cash flow statements represent a flow. Bloomberg provides the option to collect the information in current USD values. Bloomberg, for example, reports items on the balance sheet using the exchange rate set on the date of publishing; income statements and statements of cash flow items are reported using the average exchange rate for the period. Thomson Reuters on the other hand uses the WMR Spot Rate set on the date of publishing for items on the balance sheet, income statement and statement of cash flows.