A rational guide to cash management in politicised banking markets
There is not much in personal finance as emotionally charged as losing confidence in banks. This article brings together the various postings on the No Monkey Business blog at different stages of the banking crisis. Their purpose has been to be factual and rational, as a complement to instinct and habit and a counter to naked fear. As the crisis evolved, we have found ourselves changing our advice at different stages and adapting it to different people in different ways, but still rationally.
Why decisions need to be taken differently Advisers used only to have to worry about ‘idiosyncratic’ risk to their clients’ short-term risk free capital when deciding who they should use to hold cash. The odd bank might fail or at worse a cluster of banks who made the same bad bets or (less likely) suffered the same bad luck. Idiosyncratic risk could be managed by spreading deposits between separate banks inside the Financial Services Compensation Scheme level or by using a money market fund. Our advice has been obsessive about diversifying or insuring idiosyncratic risks as a matter of principle, and so dates back to the origins of the business not the origins of the banking crisis.
Why change this approach?
The optimality of each solution, assuming fixed customer preferences, changes when the relevant risk is ‘systemic’ rather than idiosyncratic, so affecting many or all banks rather than one or a cluster.
It changes yet again when governments take actions to prevent imminent systemic failure, if those actions make some banks safer than others and make those (or all banks) safer than diversified uninsured deposits in a money market fund.
Such government actions themselves potentially (or eventually) undermine government credit-worthiness, but probably not before short-dated government bonds have replaced all forms of deposit as the best risk free asset.
The measures of safety also alter if the risk is transformed for any reason from an improbable risk of large loss (typical of idiosyncratic failure) to a much higher probability of small loss (likely in systemic failure).
Most of these prompts for changes have already happened, as a result of the evolution of the crisis in banking itself and of the precise responses of government to mitigate the crisis.
Though the decision framework has changed, many of our clients have chosen not to alter their preferences. The common validation is still a rationalisation of essentially emotive impulses: the gravity of the consequences for everybody of a low-probability event makes it safe to discount that probability, or assume it does not exist. This version of the ‘safety of crowds’ cuts it for some people but not for everyone. It is these others who have changed their arrangements during the crisis. The most popular version of change has been to cut money market exposure, severely limit uninsured deposit exposure and move funds into short-dated gilts. Some of hedged their bets with a mix of all three.
Solutions for idiosyncratic risk A fundamental principle of managing risks to your capital is diversification. Modern portfolio theory teaches that it is irrational not to diversify idiosyncratic risks, such as when holding a large proportion of your wealth in one company (unless of course this really is an entrepreneurial rather than investment goal you are managing). That one company can do very badly, or even fail, quite independently of what is happening to other companies or to the economy as a whole, such as due to technological redundancy, fraud or bad management. This is what is meant by idiosyncratic risks.
Portfolio management makes diversification of idiosyncratic risks easy to implement because of the existence of packaged portfolios holding lots of securities, such as unit trusts or exchange traded funds. The risks you are then left with are those common to all investments of the same type, variously known as market risks or systemic risks. These cannot be diversified away.
When it comes to our cash, governments take away the need to diversify exposures to the risk of fraud or default by individual banks when they provide a state-sponsored deposit insurance scheme to guarantee or protect depositors. In the UK, the Government-sponsored guarantee scheme had a limit post Northern Rock of 100% of the first £35,000 per depositor per separately-regulated deposit-taking entity. This was later raised to £50,000 in response to further banking rescues. Beyond this level, idiosyncratic risk of default by a deposit taker can be avoided, not just diversified, by holding the maximum covered amount in different banks.
Uninsured idiosyncratic default risks can also be diversified by holding money in a money market fund, whose job is to diversify credit risk between many borrowers. These unitized funds may be positioned with equivalent systemic risk as you would yourself prefer, such as lending only for up to 30 days to strongly-capitalised banks or top-rated corporate borrowers in the ‘commercial paper’ market. For this, you accept that you will rarely earn more than the typical rate top banks need to pay for deposits and normally slightly less.
Money market funds may instead be positioned to chase higher returns than that, such as by depositing with riskier banks, by holding more in lower-rated non-bank commercial paper, or by lending for longer average terms and so taking on exposure to interest rate movements. The clue as to which type of fund it is may be in the name (such as ‘Enhanced Cash Fund’). In recent experience, many people did not start to find out which type of fund they owned until after the collapse of two French funds investing in Special Investment Vehicles set up to buy Collateralised Debt Obligations consisting of tranches of sub-prime mortgages. These non-bank vehicles typically funded their portfolios by issuing short-dated paper and rolling it over as it matured. Even some of the most conservative UK institutional money market funds held some exposure to short-dated commercial paper issued by these non-bank entities but no capital has been lost to date. The losses and closure of those French funds (made good by their bank owner), now more than a year ago, mark the start of the current banking crisis.
A money market fund diversifies the idiosyncratic risks of both individual banks defaulting and even clusters of banks failing but they do not avoid these risks. Occasionally, their net asset values will fall below par (ie Â£1 or $1) as they book losses. But it is very occasionally. And if sponsored by a banking group there is a fair chance they will do what the French did and make good the loss. Even with clustering of losses (caused by shared lending errors by bankers on the same or even different banking markets, such as over-exposure to property or emerging markets), the risk of loss in a money market fund could rationally be viewed as a worthwhile contingent cost of outsourcing the job of diversifying exposures. This is really only likely to make sense where the size of total deposits means there is a genuine hassle involved in spreading them around in lots of £50,000.
For most retail savers, money market funds are unlikely to be a rational choice to diversify idiosyncratic risks as they will cost them about 0.5% pa for the privilege of spreading risks they could quite easily spread for themselves. An institutional fund (such as our clients use) ought not to cost more than 0.1% if depositing more than say Â£5m collectively for clients, which puts a much lower price tag on the hassle factor. When an investor’s cash is part of an investment programme, there may be a more genuine incentive to hold cash in a fund on the same platform as other investments and this could justify the cost. Be aware, however, that advisers and managers may prefer you to hold your cash in a form, such as a unitised cash fund, that enables them to attach a portfolio-based fee to it.
Systemic risks in banking Systemic risks mean slightly different things to different people but we should be able to agree that, whatever else they are, they apply to most participants in a system, such as in national or internationally-connected banking markets. These risks might therefore range from these markets ceasing to be able to perform their function for a time to mass failures in which holders of bank equity and loan capital (but not necessarily depositors) suffer partial loss or are wiped out.
Systemic risks crystallise for depositors only if losses on the bank’s asset book exceed their capital. If, by way of example, all sources of capital amounted to 7%, all losses were suffered in the loan book and the loan book represented half the portfolio, it would take a loss of more than 14% to cause depositors to receive less than full value in a liquidation.
The actual loss experience in a liquidation is likely to be greater because loans get sold quickly rather than worked out slowly, at levels that reflect the weakness of the seller. Where the failure risk is idiosyncratic, the size of the loss is likely to be greater as gross negligence or even fraud often plays a part. To the extent systemic risk is a function of the size of failure as well as the chance of failure, we might also assume that recovery rates on troubled assets and payouts to general creditors will be greater for UK banks than foreign banks operating in jurisdictions where foreign exchange reserves may not allow full payment to international creditors.
Systemic risks may also crystallize for depositors if the process of recovering their deposit, even if in full, is slowed down by the mechanics of administration or liquidation. If we assume the actual exposure to loss when the problem is systemic is less than when idiosyncratic, it is possible that the time value of money tied up will be as significant as the capital loss itself. There may also be specific knock-on effects of the enforced illiquidity, depending on the circumstances of the depositor.
The money markets as a source of systemic risk In the US and UK we are at the stage where the interbank market, in which banks deposit with or borrow from each other to balance their books daily, has nearly seized up because banks are fearful of other banks failing. So this type of systemic risk, affecting financial institutions rather than their customers, has arguably already crystallised.
In most affected countries, the response from government has been to give banks wider access to central banks’ funds. Central banks are tiny by comparison with commercial banks so they would not be able to do this on a large scale were it not for governments making funds available out of general revenue or debt issues or because the banking system is itself buying short-dated Treasury bills with its liquid assets because it is too scared to hold deposits in other banks. There are some signs that this situation is improving, at leasy for overnight money.
What is happenign here can be explained by means of a story told of a run on one of two banks in a small Cornish town in the 1930s. When and where may not be accurate and are not details that count but one that does is that the banks were next door to each other. As customers of the bad bank withdrew gold, which no bank would ever hold in anything like sufficient amounts to meet all deposits (just like notes and coin today), they went next door to deposit their little bag in the good bank. The good bank promptly passed the bag over the fence back to the bad bank, as an interbank deposit. In today’s version of the story, the good bank will not deposit with the bad bank and it requires the government standing at the fence to take the bag in its one name and lend it on to the bad bank. In both stories, the same bag goes round and round.
Whereas the scale of government intervention is shocking, it is essentially only the scale that has changed because governments necessarily have the institutional framework and structures for this intervention on a daily basis. Neither does the scale necessarily expose the taxpayer to great risk.
Recapitalisation Over the weekend of the 11th October, several European governments adopted co-ordinated but not identical measures to inject taxpayers’ money into the banks in the form of additional long-term capital.
Though the UK is attributed with leading, our impression was that Germany announced its intentions before the weekend. However, it was also before the weekend that the UK Treasury supposedly met with the angry leaders of four major UK banks who spelt out the urgent need for recapitalisation. Not so much co-ordinated as forced by these European actions, the US then decided to adapt the Treasury’s original plan and substitute capital subscriptions for the purchase of troubled assets (which was only possible, incidentally, because of amendments to the legislation demanded by recalcitrant Congressmen). Other governments restricted additional action to giving guarantees on senior bank debt.
Recapitalisation effectively overrides the Basel capital adequacy regime and sets new, higher levels of minimum capital, at the will of individual governments. In the UK version, banks must meet these new levels promptly, if not by private initiatives by the issue of new shares by rights, underwritten by the Treasury.
The new requirements represent about a 20% increase in capital. This is not trivial but it is also hard to argue it so large as to remove the risk of losses exceeding capital.
Accounting changes Following the same weekend, the International Accounting Standards Board also succumbed to pressure to give banks discretion to value assets at levels above market, instead of observing the ‘fair value’ accounting standard that forces banks to mark assets to market prices where there is some benchmark of public-market value. Fair value accounting has exacerbated the banks’ capital erosion because many securitized positions (the first losses to emerge) have readily obtainable benchmarks for achievable prices, but they are based on fire sales rather than fair value between willing buyers and sellers.
Bowing to reality in this way is not unprecedented. It follows the suspension of fair value reserving by UK insurers near the bottom of the 2003 bear market which threatened systemic solvency problems for the life insurance industry. Though the subsequent changes to the UK’s insurance company reserving were meant to avoid repetitions of a vicious cycle of selling prompted by accounting treatment, we do not think they will. But we are also sure that, if that happens, the new rules will also be quickly suspended.
Fair value accounting, or at very least the actions required as a consequence of changes in financial position resulting from fair value accounting, will be one of several regulatory features of this crisis that will be exposed to reform in good time and will be blamed for the unnecessary severity of the crisis.
Soft guarantees The UK is amongst several governments that have said they ‘will do whatever it takes’ to avoid systematic bank failures leading (by implication) to loss of uninsured deposits. The first test of this came quickly with the nationalisation in Iceland of Landsbanki, which had a deposit-taking license in the UK as Icesave.Â While there remains doubt about corporate and local authority deposits, retail deposits in excess of the FSCS limit were given a guarantee by the Chancellor within 24 hours of Landsbanki freezing accounts. The timing of any payments is, however, uncertain.
The Germans also announced guarantees of all deposits but this later turned out to be another form of soft guarantee, without any planned legislation.
The Irish Government triggered this with a hard guarantee but at some cost in ridicule as it could not make good on its promise if the losses were widespread. It has since retracted its offer in respect of sterling accounts but not before high volumes of cash had started to move from the UK to Irish banks. Obviously these were not the same owners who ridiculed the Irish guarantee.
The chance of systemic deposit losses System-wide losses assume the failure of banks on such a scale that governments are either unwilling or unable to fund all uninsured depositors, whether the guarantees are hard or soft.
Such a scale itself implies global integration via common loss experience, which in turn implies governments turn out to be powerless to prevent global recession, falls in security (particularly property) and inability of borrowers to meet their obligations. It is arguably a condition brought on by a prolonged recession, with its knock-on effects across all types of borrower, rather than by the largely forward-looking solvency concerns that have jammed up the banking plumbing so far. This is an important distinction, as it means the risk is not necessarily evident today.
Such a scale also requires there to be little distinction between the capital strength of banks. In the past, even large-scale clustering of failures led to a flight to sound banks, who did survive, rather than a flight from all banks. It is a moot point whether international banking capital standards and, more recently, government bailouts, have removed this distinction. There are virtually no banks of any size whose business model has depended on a conspicuously stronger and more liquid balance sheet. It requires attracting customers willing to pay a price, in lower interest, for that security. But that is pointless if you believe depositors will always be bailed out.
‘Unwilling’ suggests governments will be tempted to make a rational trade-off that favours spreading the social cost of losses amongst those who chose to hold uninsured bank deposits (individuals or companies) rather than just tax payers. This is not unthinkable if the losses arise as a consequence of a prolonged recession.
‘Unable’ suggests governments (or the tax base) may be too small to fund all depositors. In theory, a government can do ‘whatever it takes’ when it has control of its own currency, as Zimbabwe does, and the UK does, but Ireland (and any other euro zone country) does not.
Government securities as the last resort Since you cannot physically remove large amounts of notes and coin from a bank, your only means of moving capital out of the banking system as a whole is to use it to buy goods that might act as a store of value alternative to currency or to buy government-issue securities and instruments. Many of these currency alternatives are extremely volatile and may themselves be subject to bubble conditions that will not withstand a slump (gold, art).
‘Real assets’ are, in the long run, better than government securities with anything other than a very short maturity if the way losses are spread across society includes printing money and, eventually, higher inflation. Capital flight should logically be expected to target short-dated conventional gilts and index linked gilts but not long-dated bonds.
In the UK, the yield on 12-18 month gilts has fallen back during this most recent phase of the crisis from around 4% to just over 3%. This implies a yield give-up, or ‘cost of safety’, relative to one-year bank deposits of between 2 and 3%. The cost of safety is slightly less relative to overnight deposit rates but these could easily average much less over the next year than 5 or 6%, cutting the ex post cost of safety. Neither the size of the yield drop nor the level at which it has recently settled suggests a widespread and price-insensitive public panic. But the yield give-up is sufficient to suggest that some people do not fully trust the government to prevent or make good all depositor losses, whether banks fail or not.
As long as some risk of outright loss or frozen liquidity is thought to remain, it does not require any strong emotional impulse to move into short gilts at 3%.
The implications for money market funds If you accept either i) that the existence of soft guarantees makes the risk of deposit loss in failed banks very small or ii) that the amount of deposit loss in the event of systemic failures is likely to be quite small, then it follows that the probability-weighted loss of capital in a money market fund is no less than, or might even exceed, that in a bank.
The risk in a fund is a function of the increased chance of loss when exposures are widely spread. Though this deals effectively with idiosyncratic risks, there is a more significant probability-weighted loss when several sterling deposit takers, including banks in foreign jurisdictions, and maybe several non-bank issuers of commercial paper, fail together.
Theoretically, the exposure to idiosyncratic risk remains and still calls to be diversified, which is what the money market fund does. But in the present situation even losses from an idiosyncratic bank collapse may be anathema to the Government because to the public that would be indistinguishable from fresh evidence of systemic failure.
The probability-weighted risk of loss in a money fund can also be compared with the safety-first yield on short gilts. If a money fund lost, say, 3%, it would wipe out the interest give-up on short gilts. In practice, the trade-off could be even more finely balanced because money market rates that drive money fund yields might fall sharply in a recession. If rates averaged, say, 4%, it would require only 1% of losses in a money market fund. This might be equivalent to perhaps a dozen discrete losses, assuming none is recoverable.
Even a very small loss could pose practical challenges for money funds, if most holders have not appreciated this risk exists. They might then trigger heavy redemptions as funds flee to the safety of the bond market. Similar moves in the US led to government guarantees being given on existing money fund holdings. This was always a bigger problem in the US because money funds are far more widely used by retail savers and investors, often in conjunction with clients’ relationships with a broker or bank.
There is also a possibility that the risk-averse institutional funds run by Aviva and Barclays, which have less retail money in them, get tainted by the impact of larger losses than we have described here in funds aimed at retail savers (or their advisers) who were uninformed yield chasers.
Conclusions We think the balance of advantage has changed and that cash management choices also therefore need to be challenged, if not actually changed.
We think this particularly militates against money market funds, even though these are our preferred normal means of diversifying idiosyncratic risk. The choice is then between returning to the banking system (partly a vote of confidence in those soft guarantees) or paying the price of safety in the gilt market.
All choices are likely to be made partly if not mainly on the basis of what helps people sleep at night, which inevitably reflects their own beliefs and emotional biases. Rational information may modify strong emotions but not overwhelm them.
Even in the absence of strong emotions, wealthy individuals who make decisions on the basis of after-tax, risk-adjusted returns may still choose perfectly rationally to move to short gilts even if such action may seem indistinguishable from panic.
We see nothing intellectually lazy or irrational in opting to spread these different types of exposure, as many of our clients have done.