Category: Interesting External Papers

The Bank of Canada has released the Bank of Canada Review: Autumn 2011 with major articles:

• The International Monetary System: An Assessment and Avenue for Reform
• Liquidity Provision and Collateral Haircuts in Payments Systems
• Extracting Information from the Business Outlook Survey: A Principal-Component Approach
• Modelling the Counterfeiting of Bank Notes: A Literature Review

The second article, on Liquidity provision, is by James Chapman, Jonathan Chiu and Miguel Molico, all of whom are bank employees. They explain:

The study presented in the following section argues that the central bank’s haircut policy can therefore directly affect liquidity in these markets and indirectly influence market participants’ choice of asset portfolios, as well as the pricing of credit and liquidity spreads. The central bank is concerned not only about its own exposure to credit risk, but also about the efficiency and stability of the financial system. Consequently, in setting its haircut policy, the central bank must consider the impact of the policy on the financial system and its participants.

A growing need for high quality collateral is forecast:

Policy-makers also face the challenge of a growing demand for high-quality collateral. Modern financial systems tend to utilize more collateral because of the increased private use of collateral, and because of the need to post additional collateral with payment and settlement systems. The G-20 countries committed to have all standardized over-the-counter derivatives contracts cleared by central counterparties (CCPs) by the end of 2012 to help strengthen the global financial system. Such an increase in CCP activity has the potential to increase the need for collateral. In addition, revisions to the core principles for financial market infrastructure, currently being considered by the Bank for International Settlements’ Committee on Payment and Settlement Systems and by the International Organization of Securities Commissions, will further increase the demand for collateral by financial market participants. The haircuts set by central banks are important parameters in determining the ability of financial systems to make the most efficient use of high-quality collateral.

The Bank of England has released its 2011Q3 Quarterly Bulletin with the following research articles:

• The United Kingdom’s quantitative easing policy: design, operation and impact
• Bank resolution and safeguarding the creditors left behind
• Developments in the global securities lending market
• Measuring financial sector output and its contribution to UK GDP
• The Money Market Liaison Group Sterling Money Market Survey
• Summaries of recent Bank of England working papers
  • An estimated DSGE model of energy, costs and inflation in the United Kingdom
  • The impact of permanent energy price shocks on the UK economy
  • Evolving UK and US macroeconomic dynamics through the lens of a model of deterministic structural change
  • Preferred-habitat investors and the US term structure of real rates

I was very interested in the second article, as it contains a defense of Special Resolution Regimes relative to bankruptcy; the politicians favouring of the former at the expense of the latter is a particular hobby horse of mine:

Commencement of insolvency leads to a freeze in the bank’s ability to make payments, which effectively results in the end of its business.(2) The sudden severing of these interconnections between a bank and the rest of the financial system and wider economy can have highly undesirable systemic effects. Individuals and small companies are entitled to compensation by the Financial Services Compensation Scheme (FSCS) for the first £85,000 of their deposits. But even a relatively short delay in the time needed by the FSCS to process and pay many deposit insurance claims can lead to hardship for households and businesses left temporarily without access to their savings. Disruption of this kind can undermine depositor confidence, potentially triggering contagion to other banks and endangering financial stability.

This fear on the behalf of depositors is a red herring. The authors would have us believe that delays by the deposit insurance corporation are inevitable, while delays by Special Resolution bureaucrats are non-existent.

The legal power to transfer some or all of the business of a failed bank to another company lies at the core of the United Kingdom’s SRR and of most other bank resolution regimes around the world.[Footnote]

Footnote reads: The transfer powers are called ‘stabilisation powers’ in the United Kingdom’s SRR and a ‘purchase (of assets) and an assumption (of liabilities)’ in the United States. The United States has had a bank resolution authority since 1933 and Canada since 1967. Transfer powers have also existed in Italy for some time and have been recently adopted in Germany.

I see no reason why these transfer powers can’t be incorporated into the regular bankruptcy process as a special case for defined institutions. Given bankruptcy, the regulator steps in as receiver and splits the bank with the objective of making the “Good Bank” as small as possible consistent with the purpose of maintaining financial stability. The “Bad Bank” holds all the common shares of the Good Bank and may sell them at leisure, although it may wish to do so on the weekend. I don’t see that this is different in practice from the aims of special resolution, or inconsistent with existing bankruptcy law. All that’s needed is an adjustment to bankruptcy law allowing this to happen when (i) the failed company is a bank and (ii) the receiver is the regulator.

Creditors, such as bondholders or other wholesale funders, that the resolution authority may have decided to leave behind in the residual bank do not enjoy these benefits [of being creditors of a solvent firm]. They must claim instead for repayment of their debts in the bank’s insolvency. But as is shown in the box on page 217, a decision to split the balance sheet in a way that fully protects depositors and certain other creditors could, on the face of it, put those creditors left behind in a potentially worse position than had the transfer powers never been used and the bank had been left to go through normal insolvency.

One reason for this lies in the fact that, under UK insolvency law, depositors in the United Kingdom rank equally — or ‘pari passu’ — with other ordinary senior creditors and therefore should share any losses equally between them.[Footnote]

Footnote reads: This contrasts with some other jurisdictions, most notably the United States, where depositors rank ahead of the other creditors (so-called ‘depositor preference’).

Seems to me that if the problem is the seniority of depositors, this is most easily addressed through legislation changing the seniority of depositors. You don’t need a Special Resolution Regime to do that. Another means of achieving the same end is for the deposit insurer to put up all the funds required to cover to the insured depositors and give the insurer and the uninsured depositors a senior claim in the Bad Bank.

The authors conclude, in part:

Bank special resolution regimes are designed to address systemic risks caused by bank failure while freeing the public authorities from the dilemma of having to use public funds to bail out all of a bank’s creditors. By doing so, they offer benefits to a financial system not only at the point of use but more generally through their effect on the behaviour of banks and their creditors.

There’s nothing here that can’t be addressed by small adjustments to existing bankruptcy law. However, the most objectionable part of the article is contained in the box which describes their plans for adusting the resolution regime:

Augment the existing SRR by developing ways to restructure a firm’s balance sheet without splitting it into separate parts. There is currently much discussion around the possible use of a ‘bail-in’ tool to write down or convert into equity some classes of unsecured debt of a firm in resolution. This would enable the resolution authority to allow losses to fall on some creditors by reducing the value of their claims on the firm without having to deal with the operational and legal consequences of transferring some of the business to a purchaser. The practical benefits of such an approach may be significant articularly when dealing with large and complex banks with huge numbers of counterparties and contracts governed by different laws.

This business of giving the resolution authority the ability to change the seniority of claims in an insolvency by fiat, instead of in accord with existing bankruptcy law, is what really sticks in my craw. Discretionary “bail-in” provisions at the whim of the regulator are an affront to the rule of law.

The other article I found of interest was “Developments in the global securities lending market”, but this was a review of the topic and how it is changing, with little that was particularly new or controversial.

The Bank of Canada has released a working paper by Anna Pomeranets and Daniel G. Weaver titled Security Transaction Taxes and Market Quality:

We examine nine changes in the New York State Security Transaction Taxes (STT) between 1932 and 1981. We find that imposing or increasing an STT results in wider bid ask spreads, lower volume, and increased price impact of trades. In contrast to theories of STT imposition as a means to reduce volatility, we find no consistent relationship between the level of an STT and volatility. We examine the propensity of traders to switch trading locations to avoid the tax and find no consistent evidence that they will change locations. We do find evidence to suggest that taxes imposed on the par value of stock will result in corporations managing the par value in the direction of minimizing the impact of the tax on investors.

Section II of the report, “Regulatory History”, give a highly entertaining account of the history of STT with the New York state government attempting to collect as much revenue as it could, with the affected companies, investors and competitors attempting to minimize the figure. A lot of time, and highly skilled time at that, must have been burned up in these games – which is the most insidious effect of a targetted tax.

Similar fun and games have been observed elsewhere, particularly with respect to preferred shares, as discussed on the blog in the post Par Value.

The literature review in section III traces the debate from the beginning:

The earliest proponents of STTs, Keynes (1936) and Tobin (1978), argue that an STT will improve market quality. In particular, Keynes contends that chasing short-term returns, while potentially profitable to specific individuals, is a zero-sum game in terms of economic welfare. Since one investor’s gain is another’s loss and trading utilizes resources, the value-added through trading is negative. As a result, imposing an STT may increase welfare by reducing wasted resources. Second, since trading is speculative by nature, it potentially contributes to financial instability when trades are driven by short-term capital gains and not fundamental information. Keynes argues that an STT will curtail short-term speculation, and thereby reduce wasted resources, market volatility and asset mispricing. Consistent with Keynes, Tobin (1978) proposes a tax on foreign exchange transactions that would make short term currency trading unprofitable. He suggests that a transaction tax would “throw some sand in the wheels of speculation.”

I have a bit of a problem with this. In the first place, speculators help long term investors by providing liquidity when they wish to buy and sell; in the second place, market prices are a very important signal to issuers, who can choose to start new companies or issue additional stock in accordance with the prices. A dramatic (if rather unfortunate) example of this was the Tech Boom of the late 1990’s, in which all kinds of companies were able to get financing thanks to the influence of speculators on the price of Internet stocks. All good things can be taken too far!

After examining the data, the authors conclude (in part):

Our findings largely come down on the side of opponents of the tax who suggest that an STT will harm market quality. Since spreads have been shown to be directly related to a firm’s cost of capital, imposing an STT may hinder economic growth by reducing the present value of projected profits.

The Bank for International Settlements has released a working paper by Mathias Drehmann, Claudio Borio and Kostas Tsatsaronis titled Anchoring countercyclical capital buffers: the role of credit aggregates:

We investigate the performance of different variables as anchors for setting the level of the countercyclical regulatory capital buffer requirements for banks. The gap between the ratio of credit-to-GDP and its long-term backward-looking trend performs best as an indicator for the accumulation of capital as this variable captures the build-up of system-wide vulnerabilities that typically lead to banking crises. Other indicators, such as credit spreads, are better in indicating the release phase as they are contemporaneous signals of banking sector distress that can precede a credit crunch.

They explain:

We find that the variable that performs best as an indicator for the build-up phase is the gap between the ratio of credit-to-GDP and its long-term trend (the credit-to-GDP gap). Across countries and crisis episodes, the variable exhibits very good signalling properties, as rapid credit growth lifts the gap as early as three or four years prior to the crisis, allowing banks to build up capital with sufficient lead time. In addition, the gap typically generates very low “noise”, by not producing many false warning signals that crises are imminent.

The credit-to-GDP gap, however, is not a reliable coincident indicator of systemic stress in the banking sector. In general, a prompt and sizeable release of the buffer is desirable. Banks would then be free to use the capital to absorb writedowns. A gradual release would reduce the buffer’s effectiveness. Aggregate credit often grows even as strains materialise in the banking system. This reflects in part borrowers’ ability to draw on existing credit lines and banks’ reluctance to call loans as they tighten standards on new ones. A fall in GDP can also push the ratio higher. Aggregate credit spreads do a better job in signalling stress. However, their signal is very noisy: all too often they would have called for a release of capital at the wrong time. Moreover, as spread data do not exist for a number of countries their applicability would be highly constrained internationally.

We conclude that it would be difficult for a policy tool to rely on a single indicator as a guide across all cyclical phases. It could be possible to construct rules based on a range of conditioning variables rather than just one, something not analysed in this paper. However, it is hard to envisage how this could be done in a simple, robust and transparent way. More generally, our analysis shows that all indicators provide false signals. Thus, no fully rule-based mechanism is perfect. Some degree of judgement, both for the build-up and particularly for the release phase, would be inevitable when setting countercyclical capital buffers in practice. That said, the analysis of the political economy of how judgement can be incorporated in a way that preserves transparency and accountability of the policymakers in charge goes beyond the scope of this paper.

It’s a lucky thing that judgement is required for the process to work – otherwise there might be layoffs in the regulatory ranks! It’s also a lucky thing that details regarding the incorporation of judgement are beyond the scope of the paper, as otherwise one might have to examine the track record of the regulatory establishment in predicting crises.

I suggest that incorporating the judgement of the regulators into the process will have numerous effects:

• Increasing politicization of the regulatory bureaucracy
• Fewer crises (since the regulators will tend to err on the side of caution)
• More severe crises (since the range of judgments existing in the private sector will be replaced by a one-size-fits-all judgment imposed according to the fad of the day)
• increased uncertainty during a crisis, as both the sellers and the buyers of new bank capital will be unsure as to whether or not the buffers will be released.

I suggest that eventually we will regret the role of regulatory judgement to the extent that it is incorporated into the process.

I believe that the release of countercyclical buffers should be linked to the amount by which a bank’s write-offs have exceeded the norm for the past one (maybe two?) years. This would encourage (or at least mitigate the discouragement) of recognition of losses and provide a time limit for raising replacement capital (since the release will be effective for only one (maybe two) years.

The authors considered using bank losses as the anchor (building-up) indicator:

Aggregate gross losses: This indicator of performance focuses on the cost side (non-performing loans, provisions etc). The financial cycle is frequently signalled by the fall and rise of realised losses.

Although the use of this signal is not supported by the raw data, I suggest that:

• It provides the most logical link to the condition that one is trying to alleviate, and
• the knock-on effect of such a change (i.e., the way in which behaviour will be adjusted to account for the adjustement in rules) is in this case highly desirable – losses will be recognized faster.

As the authors’ graph shows, there’s no magic formula for predicting a bank crisis:

Click for Big

Although the credit-to-GDP gap is the best-performing indicator for the build-up phase, Graph 2 indicates that it declines only slowly once crises materialise. This is also borne out by the statistical tests shown in Tables 5 to 7. As before, bold values for “Predicted” highlight thresholds for which a release signal is issued correctly for at least 66% of the crises. The bold noise-to-signal ratio indicates the lowest noise-to-signal ratio for all threshold values that satisfy this condition.

None of the macro variables and of the indicators of banking sector conditions satisfy the required degree of predictive power to make them robust anchor variables for the release phase, ie none of these variables signals more than 66% of the crises. The best indicator is a drop of credit growth below 8%. This happens at the onset of more than 40% of crises and such a signal provides very few false alarms (the noise-to-signal ratio is around 10%).

A backtest of the Credit Gap as an anchor variable is encouraging:

Click for Big

Part of the authors’ conclusion is:

The analysis shows that the best variables to signal the pace and size of the build-up of the buffers differ from those that provide the best signals for their release. Credit, measured by the deviation of the credit-to-GDP ratio from its trend, emerges as the best variable for the build-up phase, as it has the strongest leading indicator properties for financial system distress. A side-benefit of using this variable as the anchor is that it could help to restrain the credit boom and hence risk taking to some extent.
…
A final word of caution is in order. Are our empirical results subject to the usual Lucas or Goodhart critiques? In other words, if the scheme proved successful, would the leading indicator properties of the credit-to-GDP variable disappear? The answer is “yes”, by definition, if the criterion of success was avoiding major distress among banks. As credit exceeds the critical threshold, banks would build-up buffers to withstand the bust. If, in addition, the scheme acted as a brake on risk-taking during the boom, the bust would be less likely in the first place. However, the answer is less clear if the criterion was the more ambitious one of avoiding disruptive financial busts: busts could occur even if banks remained reasonably resilient. In either situation, however, the loss of predictive content per se would be no reason to abandon the scheme.

One thing I will note is that it may be helpful to disaggregate the accumulation signal. As has previously been reported on PrefBlog, residential real-estate now makes up more than 40% of Canadian banks’ credit portfolios, vs. the more normal 30%. I suggest that it would be useful to examine the components of the credit gap, in the hopes that one or more of them will prove a better signal than the complete set.

It may also be the case that changes in the proportion of various components of aggregate credit are just as important as the aggregate amount itself – these changes could be indicative of dislocations in the economy that will have grievous effects if, as and when they return to normal.

The Bank of Canada has released the Summer 2011 Bank of Canada Review:

This special issue, “Real-Financial Linkages,” examines the Bank’s research using theoretical and empirical models to improve its understanding of the linkages between financial and macroeconomic developments in the wake of the recent global financial crisis.

Very important, I know – and it’s what the Bank’s job is all about! – but not my favourite topic.

Articles are:

• Introducing Multiple Interest rates in ToTEM
• The BoC-GEM-Fin: Banking in the Global Economy
• Bank Balance Sheets, Deleveraging and the Transmission Mechanism
• Mortgage Debt and Procyclicality in the Housing Market
• Developing a Medium-Term Debt-Management Strategy for the Government of Canada

I found the last article, by Marc Larson and Étienne Lessard, to be too general to be satisfying, but there were some interesting points:

Step 3 – Selecting optimal strategies In this step, a wide range of different financing strategies are reviewed, some of which may involve issuing debt in only some maturity sectors but not others. An optimization algorithm is then used to select those strategies with the best cost-risk tradeoffs, or the lowest cost for a specific level of risk. The output of this work is a curve that represents the most efficient financing strategies, similar to an efficient portfolio frontier, as well as the composition of the most efficient financing strategies.

Chart 2 and Chart 3 illustrate the results of the optimization exercise based on debt rollover as a risk measure. Note that the same exercise can also be performed using other risk measures. Chart 2 shows the efficient frontier of the optimal debt structures (lowest cost for a specific level of risk). Moving along this frontier from left to right shows how expected borrowing costs decrease—and rollover risk increases—as the government shifts the proportion of its borrowing program from long-term debt to shortterm debt. Chart 3 illustrates how the proportion of short-term debt in the optimal portfolio changes as one moves along the efficient frontier. Each colour represents a different debt instrument issued by the Government of Canada. As shown in this chart, low risk debt structures contain mainly long-term maturity instruments (10-year and 30-year nominal bonds and Real Return Bonds), while high-risk debt structures contain mostly short-term debt instruments (3-, 6- and 12-month treasury bills and 2-year bonds).

Click for big

I find the relationship between efficient amounts – as defined – of long nominals vs. RRBs to be fascinating. I suppose part of the reason for the relationship is that they are considered to have similar cost, but a big chunk of the RRB return is paid only on maturity; therefore they will have a lower quarterly rollover rate (less risk) at the same [relationship of] cost. Maybe! I’d have to look at the model in more detail!

However, rest assured that I will be quoting the bank’s conclusions on risk to support my “security of income” vs “security of principal” argument until you are all heartily sick of it, never fear! The risk of variance of investor’s income is exactly the flip side of the risk of variance of the Bank’s expenses.

The article on bank balance sheets by Césaire Meh uses a model to derive interesting relationships between counter-cyclical capital buffers and monetary policy. Note that “(demand-type) financial shocks … generate simultaneous downward pressures on inflation and credit contractions.” – i.e., a standard recession,while “(supply-type) financial shocks … cause credit contractions and upward inflation pressures,” i.e., a financial crisis / credit crunch:

Overall, these results suggest that the impact of countercyclical capital buffers on the transmission mechanism of monetary policy and, consequently, the nature of the coordination between these two tools, depend on the nature of the shocks experienced by the economy. Demand-type financial shocks pose no inherent trade-offs between stabilizing credit and achieving price stability. In this case, the use of countercyclical capital buffers eases the pressure on monetary policy, and less-aggressive movements in the interest rate would be required to achieve economic stability. Supply-type financial shocks, however, can generate a tension between stabilizing credit and price stability. In this case, activating countercyclical capital buffers could make it harder to stabilize inflation, and more-aggressive movements in the interest rate would be required. Under such circumstances, proper coordination between the two policy instruments will lead to a better policy outcome. [Footnote]

Footnote: Countercyclical capital buffers should be considered neither a substitute for monetary policy nor an all-purpose stabilization instrument. Rather, they should be viewed as a useful complement to monetary policy in a world in which financial shocks have become an important source of economic fluctuations.

The Bank of Canada has released a working paper by Jason Allen, Ali Hortaçsu and Jakub Kastl titled Analyzing Default Risk and Liquidity Demand during a Financial Crisis: The Case of Canada:

This paper explores the reliability of using prices of credit default swap contracts (CDS) as indicators of default probabilities during the 2007/2008 financial crisis. We use data from the Canadian financial system to show that these publicly available risk measures, while indicative of initial problems of the financial system as a whole, do not seem to
correspond to risks implied by the cross-sectional heterogeneity in bank behavior in short-term lending markets. Strategies in, and reliance on the payments system as well as special liquidity-supplying tools provided by the central bank seem to be more important additional indicators of distress of individual banks, or lack thereof than the CDSs. It therefore seems that central banks should utilize high-frequency data on liquidity demand to obtain a better picture of financial health of individual participants of the financial system.

Essentially, the paper provides further evidence that the bond market is excitable:

In contrast to public measures of bank risk (such as CDS prices), which varied widely both crosssectionally and in the time-series, we show that Canadian banks’ bidding behavior in the liquidity auctions as well as their behavior in the payment system and overnight interbank market was largely indicative of low risk. The fact that overnight market remained quite active throughout the crisis – total loans transacted stayed virtually unchanged while prices actually fell – indicates that participants did not believe there were significant liquidity or counterparty risks. Similarly, Afonso, Kovner and Schoar (2010) find that the overnight Feds Fund market remained active following the collapse of Lehman Brothers, although they do find evidence of increased counterparty risk. We argue that the impact of various liquidity-providing actions undertaken by the cental bank and federal government during 2008 might have led to a surplus of liquidity in the market, resulting in overnight unsecured loans transacting even below the target rate. In contrast, we find that during the asset-backed commercial paper (ABCP) crisis in the summer of 2007, when these extraordinary liquidity facilities did not exist, interbank rates did increase. Similar to Acharya and Merrouche (2010) this suggests liquidity was more scarce at this time. In neither episode, however, do we find evidence of an increase in counterparty risk.
…
We will provide evidence that in the case of Canadian financial institutions, for which CDS spreads varied substantially both in the cross-section and in the time series, the corresponding variation in the short-term spreads was lacking. We will argue based on further indirect evidence that it therefore seems that in case of fairly illiquid CDS markets, the short term spreads recovered from the bidding behavior in liquidity auctions may provide a more useful source of information about counterparty risk and potential financial trouble in the turbulent times of a crisis.
…
The situation in Canada was remarkably different. While the default probabilities of individual banks implied by the prices of their respective CDS contracts followed a similar pattern as their European counterparts, the Canadian banking system showed very limited signs of stress (or increase in liquidity demand) in 2007 and in the first half of 2008. In particular, the Canadian banks were much less willing to pay a premium above the reference overnight rate to obtain liquidity from the Bank of Canada. Figure 4 depicts the aggregate bidding functions in each auction that the Bank of Canada conducted before September 2008. It suggests that virtually all banks judged that even if they would not have their demands in these auctions satisfied, they could secure the liquidity elsewhere and hence were bidding at, or very close to, the reference overnight rate (OIS).

Using the standard
formula (see e.g., Hull (2007)):

Pr (Default 5y)_T = 100 * (1 – (1/(1 + (cds_T /10000)/(1 – recovery))^T ))

we can recover the risk-neutral default probabilities implied by the CDS prices. For the largest institutions, banks A,E and K, the implied default probabilities (assuming 40 per cent recovery rates) went from close to zero to over 15 per cent. Given the increase in default risk implied by the CDS contracts on the banks in our sample we might expect an increase in liquidity hoarding during the crisis.

Click for big

Christopher Condon and Robert Schmidt of Bloomberg report that:

Money-market mutual funds would be forced to create capital buffers equaling 1 percent to 3 percent of assets to protect against losses under a plan now favored by staff at the U.S. Securities and Exchange Commission, according to three people briefed on the regulator’s deliberations.

Top SEC officials, seeking to make money funds safer, prefer the plan over another capital buffer idea crafted by Fidelity Investments and calls to eliminate the funds’ stable share price, said the people, who asked not to be identified because they weren’t authorized to speak publicly. The concept is based on recommendations submitted to the agency in January by university economists known as the Squam Lake Group.

Readers will remember the Squam Lake proposal on contingent capital, which I didn’t like, but was a whole lot better than OSFI’s idiotic approach.

I reported in May that the SEC was grinding ahead on MMF regulation; readers will remember that I have a long-standing interest in the topic.

The Squam Lake Proposal for MMF regulation points out:

In the past, without the prospect of government guarantees, whenever money market funds threatened to break the buck, it had been common for their managers to bail them out in order to preserve the franchise values of their fund management businesses. Between August 2007 and December 31, 2009, at least 36 U.S. and 26 European money market funds received support from their sponsor or parent [footnote] because of losses incurred on their holdings of distressed or defaulted assets, as well as the costs of meeting the redemption demands of investors through sales of assets. Going forward, if sponsors believe that their funds will receive government support, their incentive to bail out their own funds may be substantially reduced, particularly given the squeeze on profitability associated with exceptionally low money-market interest rates.

Footnote: See “Sponsor Report to Key Money Market Funds,” by Henry Shilling, Moody’s Investor Service, !ugust 9, 2010; The forms of support included capital contributions, purchases of distressed securities at par, letters of credit, capital support agreements, and letters of indemnity or performance guarantees.

They propose a capital buffer:

Thus, as an alternative to floating NAV, a second broad approach, which we focus on below, preserves the stable NAV structure but enhances its safety by requiring sponsors to establish contractually secure buffers that could absorb at least moderate investment losses to their money market fund investors. This is akin to a capital requirement for stable-N!V funds; The President’s Working Group Report (2010) describes various alternatives, including some forms of liquidity facility or insurance that are consistent in spirit with this approach, but it does not make a specific recommendation. [footnote]

Footnote: See “Report of the President’s Working Group on Financial Markets: Money Market Fund Reform Options,” October 2010, [published by the Treasury] which suggests that money market funds continue to pose systemic risk. Among the alternative policies described in the President’s Working Group Report are: conversion of all funds to floating NAV; a private or public insurance scheme for stable-NAV funds; a rule by which large redemptions would be paid in kind (that is, with a portfolio of assets held by the fund); a two-tier system of both floating-NAV and stable-NAV funds under which stable-NAV funds would be required to have some support mechanism; a two-tier system under which stable-NAV funds are only available to retail investors; a rule forcing stable-NAV funds to convert to special purpose banks, holding capital and having access to lender of last resort facilities, and for which depositors would have some insurance coverage.

The Squam Lake group proposes:

The manager of a stable-NAV money market fund must provide dedicated liquid financial resources that, in combination with those represented by the assets of the fund class investors, are sufficient to achieve a net buffer of “X” per dollar of net asset value. These additional resources are to be drawn upon as needed to support fund redemptions at one dollar per share until the fund converts to a floating-NAV or until the buffer resources are exhausted. That is, at the end of each business day, the combined resources available to fund investors represented by the sum of dedicated additional sources and the previous day’s marked-to-market per-share value of the fund’s assets must exceed 1+X per share held as of the end of the current day. The fund must convert to a floating-NAV fund within a regulatory transition period, such as 60 days, in the event that the fund manager falls out of compliance with this buffer requirement.

They do not formally recommend a buffer size (that is, the value of X in the proposal) but indicate that 3% is a good place to start discussion:

When setting the size “X” of a required buffer, regulators may wish to consider the amounts by which money market funds have broken the buck in the past, or the amounts per share that fund sponsors have contributed in order to prevent them from breaking the buck. In the two-day period following Lehman’s bankruptcy, the Reserve Primary Fund reported a minimum share price of 97 cents.9 Had redemptions not been halted by the Reserve Fund’s sponsors, a fire sale of additional assets could have caused significant additional losses. A buffer of at least $0.03 per share would therefore have been necessary to prevent the Reserve Fund from breaking the buck.

Another consideration in determining the size of a buffer requirement is the concentration of fund assets among the debt instruments of a small number of borrowers. As of June 2010, for example, the top 5 exposures of U.S. prime money market fund assets, were all to European banks, with each of the 5 banks representing an exposure of at least 2.5% of aggregate fund assets.

Frankly, I think this is unnecessarily complex and specialized. Money market funds are, essentially, banks. They should be regulated as banks.

Update, 2011-8-3: The Fidelity plan is a little different:

Given that tepid response, the SEC is discussing other ideas such as those suggested by Fidelity Investments, which opposed the notion of a liquidity bank in its comment letters to the President’s Working Group.

Under the Fidelity proposal, money market funds would create a capital reserve or an “NAV buffer” by charging investors more over a period of time, said Norman Lind, head of trading for the taxable- and municipal-money-market desks at Fidelity Management and Research Co., the investment adviser for Fidelity’s family of mutual funds.

The SEC would work with fund boards to determine a range that a fund should keep for capital reserve, he said during a panel discussion.

“Let’s say you retain five basis points per year and you accrue that over time,” Mr. Lind said. “The idea is that once you have a buffer in place … you stop charging that fee.”

Unlike the ICI’s proposal, Fidelity thinks that its idea is simple to implement and doesn’t require regulatory changes, Mr. Lind said.

I don’t get it, frankly. Who owns the buffer?

The Financial Stability Oversight Council (comprised of an alphabet soup of US financial regulators) has released its 2011 Annual Report.

There are many items of interest in this excellent report; one issue that I find interesting is the regulation of money market funds:

The stable, rounded $1 NAV fosters an expectation that MMF share prices will not fluctuate. However, when shareholders perceive that a fund may suffer losses, each shareholder has an incentive to redeem shares before other shareholders, causing a run on the fund. Such redemptions can accelerate the likelihood of a break-the-buck event to the extent that the fund’s asset sales to meet redemptions significantly depress the market value of the fund’s remaining assets. In such a scenario, the ability of early redeemers to receive the full $1 NAV is essentially subsidized by the losses absorbed by remaining shareholders.

Click for Big

MMFs invest in assets that may lose value, but funds have no formal capital buffers or insurance to absorb loss and maintain their stable NAV. When losses do occur, MMFs have historically relied on discretionary sponsor support to maintain a stable NAV and preserve the franchise value of fund management businesses (Chart D.2). That support may come in the form of capital contributions or the purchase of assets that have lost value, for example.

Sponsors do not commit to support an MMF in advance, however, because an explicit commitment may require the sponsor to consolidate the fund on its balance sheet. Thus, although investors ostensibly bear the risk of an MMF breaking the buck, sponsors have in the past borne that risk themselves, fostering the perceived safety of MMF investments. Moreover, the uncertainty about the availability and sufficiency of such support during crises, and the fact that many MMFs lack deep-pocketed sponsors, contribute to their susceptibility to runs.

Expectation of Government Support

Given the unprecedented government support of MMFs during the crisis in 2008 and 2009, even sophisticated institutional investors and fund managers may have the impression that the government would be ready to support the industry again with the same tools.
…
Although these new rules are a positive first step, the SEC recognizes that they address only some of the features that make MMFs susceptible to runs, and that more should be done to address systemic risks posed by MMFs and their structural vulnerabilities.

The report takes a much more reasonable view of the Flash Crash than did the highly politicized SEC report:

During periods of violent price movements, market liquidity can evaporate as hedging strategies to protect against market risk become strained or directly amplify the price movements. For example, in the October 1987 equity market crash, portfolio insurance programs were designed to sell when prices declined; in fact, they were set to sell at an increasing rate, thereby accelerating the market decline. Similarly, in the flash crash of May 6, 2010, liquidity evaporated and market functioning deteriorated rapidly. Regulators have added circuit breakers in equity markets to mitigate such dynamics (see Section 5.3.4), but this event illustrated the potential fragility of market liquidity, particularly in areas characterized by rapid innovation and change in market behaviors.

The role of exchange traded funds (ETFs) during the flash crash has focused attention on these products. The rapid rise of ETFs has been driven by the attraction of gaining liquid exposure to less liquid asset classes—such as commodities and certain emerging markets—without having to execute trades directly in less liquid markets (Chart E.1). However, the liquidity of ETFs depends heavily on the support of market makers and on market functioning in the underlying asset. The relationship between ETF turnover and market volatility bears further analysis, and regulators must continue to monitor the development of more complex products in both U.S. and foreign-domiciled funds that might heighten liquidity concerns.

One item of great interest to me was developments in the idea that insurance companies should be regulated at the consolidated level – there is not a single mention of this in the report, so for the moment I will assume that this highly desirable reform has been dropped. Instead, the report discusses the new Federal Insurance Office (FIO), which is just another micromanaging job-creation scheme.

The Committee on the Global Financial System, a unit of the Bank for International Settlements, has released a Working Group Report titled Fixed income strategies of insurance companies and pension funds:

Insurance companies will be affected to a greater extent by the introduction of Solvency II, a comprehensive risk-based regulatory framework to be phased in from 2013. Solvency II and other risk-based regulatory regimes in Europe require that assets should be marked to market and that liabilities be discounted at risk-free rates (possibly augmented by an illiquidity premium). Solvency II also requires insurers to hold loss-absorbing capital against the full range of risks on both their asset and liability side to weather unexpected losses with a probability of 99.5% over a one-year horizon. While the latest quantitative impact study by European insurance regulators suggests that the majority of insurance companies will not face an imminent need to raise new equity, they may rebalance their asset portfolios in line with the new risk charges. The proposed changes tend to make it more expensive to hold equity-like instruments, structured products, and long-term or low-rated corporate bonds, whereas government bonds and covered bonds will receive relatively favourable capital treatment.

A related concern is whether life insurers and pension funds can maintain a long-term investor perspective. Factors contributing to this concern among market participants include the steep regulatory risk charges and short horizons to be used for assessing solvency and for addressing funding shortfalls. Prospective volatility in financial statements under international accounting rules may also limit the scope for taking long-term or illiquid positions without any concern for short-term fluctuations in their value. As is the case for institutional investors more generally, these factors tend to encourage a shift away from long term investing in risky assets, in addition to the ongoing trend toward more conservative asset allocations in the aftermath of the financial crisis. This could alter the traditional role of life insurance companies and pension funds as global providers of long-term risk capital. A partial retreat of institutional investors from the long-term and/or illiquid segment of the credit market could reduce the private and social benefits the sector generates through long-term investing, and the extent to which it mitigates the procyclicality of the financial system.

In Solvency II, both assets and liabilities are marked to market, ie fair valued. The present value of liabilities, or technical provision, is defined as the amount an insurer would have to pay to transfer its insurance obligations immediately to another willing buyer. It consists of the best estimate, the present value of the expected future cash flows (net payments to policyholders), calculated on a specified discount rate curve (term structure), and the risk margin, which is an additional premium above the best estimate. How the discount rate is constructed is of considerable importance given that the risk margin, and the present value of liabilities, will increase when this rate decreases.

The current expectation of how this rate will be constructed, is the swap curve (excluding credit risk), augmented by an illiquidity premium for those obligations coming due more than a year ahead. (This was also the discount rate used in Quantitative Impact Study 5.) In addition, an extrapolation (technique) towards a fixed rate (ultimate forward rate) will be used to get the discount rates for longer maturities than those available from market rates in different countries.

The motivation for adding an illiquidity premium, according to the proponents, is that there is general acceptance that the valuation of corporate bonds should take into account risk spreads in the discounting of future cash flows.35 Bond spreads during the crisis far exceeded the cost of credit risk mitigation (CDS spreads), and therefore included a substantial component pricing in illiquidity. The important role the illiquidity premium played in the valuation of assets, while liability cash flows continued to be discounted at the risk free rate, was largely responsible for a substantial shortfall in insurers’ balance sheets. The application of the illiquidity premium on the liability side would aim to reduce this valuation mismatch to avoid situations where insurers are forced to dispose of illiquid assets. The financial condition of insurers would be improved by allowing them to discount liabilities at a higher rate when markets are illiquid. Such a countercyclical mechanism might even mean that insurers would be willing to take on additional illiquid assets in a period of market distress, depending on whether the change in their net asset-liability position would improve their capital position.

Government bonds. Since European government bonds in domestic currency are classified as risk-free under Solvency II, there is a clear regulatory incentive to increase exposure to this asset class, including to euro area periphery debt. However, major insurance companies also rely on internal risk models that account for spread and default risk on sovereign debt. As in the case of banks, this would tend to moderate the incentive to shift toward high-yield sovereign debt even if the overall demand for sovereign debt is likely to rise. On balance, however, one may expect greater demand for long-dated sovereign debt which, all else equal, will further contribute to low long-term interest rates. In addition, insurers’ efforts to reduce their duration gaps tend to reinforce the demand for long-dated government debt from an ALM perspective.

Any sizeable shifts in the government bond space may lead to noticeable financial market implications, given the volume of government bonds on the balance sheets of insurers (as well as on those of pension funds, see Section 2.4). Depending on initial conditions and current bond holdings, further shifts into government bonds may well produce downward pressure on yields, although differentiation across issuer countries is likely to occur.

Corporate and covered bonds. The impact of Solvency II on the corporate bond market is also potentially significant. Historically, insurance companies constitute a key investor base, holding more than 30% of the corporate bond supply.48 Solvency II will impose capital charges on corporate and covered bonds that did not exist under Solvency I, although internal models at large insurers had taken into account credit risk before Solvency II was developed.

Under the standard formula, Solvency II capital charges have relatively steep duration and credit slopes which can be expected to lead to some portfolio adjustments. The capital requirements for corporate and covered bonds are calculated by multiplying a rating-induced shock factor with the duration of the bonds. A BBB-rated bond with a duration of 10 therefore requires 25% (=2.5%*10) in equity capital before diversification benefits. This formula appears to penalise long-term bonds since credit spreads at the long end are less volatile than those at the short end. The credit slope is similarly steep.49 Corporate bonds with a low rating effectively attract a capital charge similar to that of equities.

Under the instrument-specific capital requirements in Solvency II, the following
investment allocations generate the same capital requirement under the standard formula:

• 100% in covered bonds (AAA-rated) with a duration of one year,
• 20% in covered bonds (AAA-rated) with a duration of five years, and the rest in EEA government bonds,
• 13.3% in corporate bonds, AAA-rated, with a duration of five years, and the rest in EEA government bonds,
• 8.6% in corporate bonds, A-rated, with a duration of five years, and the rest in EEA government bonds,
• 1.6% in corporate bonds, B-rated, with a duration of five years, and the rest in EEA government bonds,
• 1.5% in “global equities” and the rest in EEA government bonds,
• 1.2% in “other equities” and the rest in EEA government bonds.

What it looks like to me is that the net effect will be to shift funding risk from the financial economy to the real economy. Once this is in place look for the next financial crisis to be propogated much more thoroughly to the real economy, with lots of firms finding they have debt coming due that cannot be rolled.