Dr Keating om Solvens, Likviditet och Hållbarhet

Solvency, Liquidity and Sustainability

 Corporate bond markets, by and large, operate on the basis of equitable insolvency. An issuer or borrower must default on an obligation to make a payment and fail to cure this within an allowed period before the creditor may intervene; this intervention accelerates the terms of the debt which then becomes payable as to principal and accrued interest. Most solvency regulation operates on a different basis, balance sheet insolvency, which occurs when the current estimated value of assets for the institution is below the current estimated value of (non-equity) liabilities. In most cases, regulations specify in great detail the manner in which the values of assets and liabilities must be calculated.

The international financial regulatory system relies heavily upon the concept of risk-weighted assets and specifies required capital levels relative to the portfolio aggregate of these for individual institutions. This may be criticised at many levels. The institutional versus system approach fails to consider the inter-institution dependencies, which are very substantial for banks and have even acquired the description ‘contagion’. The use of risk as a management contrivance, when it is not observable, which means, inter alia, that complete contracts cannot be written upon it. It is also evident that the risk-weights of assets are not coherent across classes of financial institution or internationally. One of the more egregious examples is the treatment of long-term liabilities under newly proposed legislation. For banks the absence of long-term liabilities will be penalised while for insurers the presence of long-term liabilities will be penalised. A simple security, such as a corporate bond, may carry weights under different regimes which are multiples of one another depending upon the institutional ownership and the domicile of its operations. Perhaps though, the most important aspect of these weights lies in their incentives and signalling value. Prior to the crisis it was possible for a bank to report capital adequacy in excess of 8% of risk weighted assets, while simple leverage ratios of assets to qualifying capital were fifty-fold or more. It is clear that many, such as the zero weight for sovereign credits, the preferential treatment of mortgage instruments and the outsourcing of certain critical functions such as credit ratings, contributed meaningfully to the distortion of bank balance sheets and their instability in distress. Indeed, they spawned a complete discipline of regulatory arbitrage involving securitisation and shadow banking. The scale and extent of this arbitrage activity was very substantial. At the end of 2007, when measured as the ratio of risk-weighted to total assets, some banks such as Deutsche and UBS reported values below 20%, the UK banks were clustered in the 40%-50% range while some US banks, such as Wells Fargo and US Bancorp, were over 80%. Suffice it to say here that this arbitrage served to reduce the effectiveness of the regulatory balance sheet approach to insolvency and that the first signs of the system’s true difficulties arose as problems of liquidity. Risk weights may reduce regulatory capital requirements, but liquidity, as funding and funding risk, operates on the unadjusted value.

Some regulatory systems use a different basis – e.g. the application of ‘prudent’ assumptions in valuations, rather than risk weights, for pension funds. This is a fine example of the nature of financial regulation being of inspector’s quality. Here ‘prudent’ is described as conservatively biased and the problems and costs, which these distortions bring relative to the best estimate or expectation arising under the better definition of prudent as rational and well-informed judgement, pass without comment or analysis.

The difference between equitable insolvency and balance sheet insolvency is a question of liquidity. At a trivial level, it is obvious that in order to default or fail to make a payment, the borrower must have exhausted their current liquidity and be unable to obtain any further liquidity from selling or pledging as security any of its residual assets. By contrast, with balance sheet insolvency, the institution may have adequate liquidity to meet scheduled payments well into the future. In the case of pension funds this may be many decades. In fact one of the perverse consequences of calculating liability values as net present values is precisely that lower discount rates imply longer terms of payment sustainability for any level of funding. This incidentally renders meaningless the accounting profession’s idea that using a common discount rate makes pension accounts comparable. In fact, the term for which a given amount (or proportion) of assets is sufficient will also vary with the profile of the institution’s cash-outflow liabilities. This is illustrated for two pension schemes as Figure 1.

In this illustration both schemes are fully funded at today’s date with these funds being held as cash, generating no income. By the accounting standard both schemes are exactly solvent. The amounts held as cash are simply the net present values of these liabilities calculated using respectively 2.5% and 5% for each scheme respectively. We see that, with a five percent discount rate, scheme one has sufficient funds to meet commitments for seventeen years while scheme two has sufficient funds to satisfy twenty years.  When we move to using a 2.5% discount rate, the term of coverage extends to twenty five years for scheme one and thirty years for scheme two. Not even the differences are comparable. It is not at all obvious why the duration of coverage should be varied in this arbitrary way, nor, in the absence of further information about the future cash-flows emanating from assets, that the balance sheet has any substantial meaning.

Regulation which specifies a level of funding for a pension fund, or a level of liquid assets relative to short term liabilities for a bank, suffers a real problem which Charles Goodhart memorably described by the analogy of the solitary station taxi. This taxi though unoccupied is unavailable to any would-be passenger as local regulations require that there be at least one taxi waiting on the taxi-rank at all times. These regulatory funding or liquidity level rules deny their use as buffers and expose the regulated institution to the variation they were intended to mitigate. In the pension context above, ceteris paribus, payment of the current year’s pensions will result in contribution, or liquidity, demands to restore solvency; the buffer or capital value of the fund is effectively unavailable. Management becomes a question of short liquidity, remarkably similar to pay-as-you-go arrangements.

This question of future cash-flows lies at the heart of all financial analysis. A security, no matter whether bond or equity, draws its value from the future cash-flows which it will produce for its beneficial owner. Its market price is a far more complex affair rooted in the “animal spirits” of financial markets. One of the facets of liquidity is precisely that this is pledgeable future income or cash-flows. Pledgeable income and collateral are synonymous with liquidity for the purposes of this article.

It is perhaps possible to argue that the balance sheet view is accurate in the sense that equitable insolvency will ultimately result at a later date if all future incomes are pledged and capitalised in today’s asset values. But that is a very strong condition; it requires perfect foresight with respect to those future cash-flows and requires that they not be encumbered in any way which might limit their pledgeable value. This is far stronger than the standard assumptions of modern financial theory with respect to market values for traded assets. It is an assumption of perfect and complete liquidity.

It seems entirely implausible. The plant employed in production may generate one value in its first best use and something entirely different and much lower in second best use. In fact, some may actually represent liabilities rather than assets in other uses or scrappage. This is one of the reasons that liquidation of any enterprise is usually very costly.

It is also perhaps sensible to use a balance sheet view when liabilities may crystallise and become payable immediately. This is the classic situation of the bank run, where the balance sheet insolvency may rapidly induce equitable insolvency as demand deposits are called and term deposits not renewed.

A bank’s advances and loans are illiquid and may not be fully pledgeable because of a problem which arises from the role of a bank as a monitor. In this monitor role, the bank develops private information about a borrower. This is not available to any other banker and has proved problematic for those investment funds which have been developed to specialise in direct lending to the corporate sector in the wake of the banking crisis. The insolvency and liquidation of a bank may be seen to be economically costly in this regard.

It is, of course, possible for an institution to be in balance sheet surplus but unable to meet its due liabilities if these assets cannot be realised, by market sale, pledge or otherwise. This is traditionally the way in which the Bagehot doctrine of central banks lending only to solvent but illiquid institutions is posed. The difficulty with this lies in the inter-relation between liquidity and solvency; establishing whether an illiquid bank is solvent is a very old problem.

However, before moving to consider such issues in greater detail, it is worth considering another feature of publicly issued and traded debt securities, covenants. These are restrictions of various forms on the management of the business and its financial affairs. They come in many different flavours and can have markedly different effects with respect to the balance sheet and equitable insolvency time line. A simple negative pledge which denies the provision of explicit security to other creditors without a similar, or pari passu, treatment for existing bond creditors is perhaps the most common covenant. At first sight, this would appear to be no more than a concern with recoveries post equitable insolvency but in practice it shortens the time between balance sheet and equitable insolvency, since these security assets cannot be pledged for new finance. Some, such as an asset or interest coverage ratio may operate before balance sheet insolvency, precipitating that problem and perhaps even equitable insolvency. Such coverage ratios may make apparent sense in that they seem to enhance the security of the debt and enhance the recovery experienced post equitable insolvency, but if they do so at the cost of an increased likelihood of insolvency of either form they may be unwarranted and indeed counter-productive. It is also clear that covenants which bind prior to balance sheet insolvency may be strategically abused.

In general, the taking of security will hollow out an enterprise and reduce the external liquidity available to it. This has been extensively discussed in the context of the increased use of repo finance rather than unsecured deposits in the interbank market. This is a mechanism by which equitable insolvency becomes all the more likely. It is as well to make this point explicit: taking security will, in general, increase the likelihood of sponsor failure.

It is quite common to over-collateralise exposures. Bonds may be secured by real estate whose appraised value must remain at, say, 125% of the debt, securities offered in repo may be subject to ‘haircuts’ which limit the advance to less than the current market value of those securities, traded futures and options may be subject to an initial margin. All of these serve to reduce the degree of leverage available from the instrument; they are explicit induced liquidity imperfections. It is notable that regulators usually specify considerable conservatism in the setting of margins and haircuts.

One of the most frequently reported empirical regularities of equity markets is that they deliver returns in excess of those to be expected from dividend discount pricing models. The standard explanation or interpretation of this is ‘myopia’, or excessively high implicit discounting by investors. Such phenomena could equally well be explained by the presence of unpledgeable future income within the enterprise. The presence of unpledgeable income for bond securities would have a different effect – it would postpone, or perhaps eliminate, equitable insolvency and should be expected to lower the cost of debt issuance or market yields.

There is a very substantial literature on another aspect of corporate finance which is relevant; the principal – agent problem. Management may divert income for its own purposes with such phenomena as empire building and other perquisites; this is unpledgeable income but is, in general, observable. There is also another principal agent problem which is unobservable; the failure of management to exert effort to maximise income potential. This is colourfully referred to as ‘shirking’. There are parallels to these issues at the level of an economy. If shirking and the misrepresentation of income are substantial it may make sense to apply sales taxes rather than income taxes or indeed to utilise the inflation tax in order to maintain and support debt service obligations.

It is though clear that at the level of the corporation, in times of distress management may forego some or all of their ‘perks’ in order to preserve their employment in the ongoing business and indeed they may be expected to exert additional effort to generate further income. Obviously broader economic conditions will be relevant to the success of these strategies. In times of economic distress and low economic activity the effort strategy is clearly more difficult to successfully implement than when this distress is induced by an idiosyncratic company specific circumstance. However, the central point here is that the enterprise can expect to generate more revenues than are priced by the market or by discounted present value techniques.

Under the balance sheet view, insolvency occurs prior to equitable insolvency and is more likely than it. This implies that the yields that we observe in bond markets, which are of the equitable form, are lower than would apply to balance sheet insolvency. This is a problem for solvency based regulation as it illustrates the fact that this regulation is itself not market-consistent. It suggests that regulation overstates the likelihood of the failures which these regulations were intended to avoid and this may be economically costly.

It is well-known that the banking system creates liquidity in the sense that every loan creates an equivalent deposit. There is no net wealth creation. One criticism of the banking system of the past few decades is that it has been better at producing new claims on wealth rather than wealth itself. The question arises as to whether the banking system can do this without limit and that reduces to the question of the extent to which these claims on debtors can be satisfied from the wealth of the community. Put simply, if debt exceeds wealth then it is just not possible to extinguish all debt by wealth transfer. The magic of banking lies in the fact that it relies upon not just present wealth but also that which will be created in the time prior to the maturity of the debt. And indeed the possibility that further refinancing may be feasible at the maturity or repayment of the debt. As recently as the 1970s it was common in international banking circles to focus upon self-liquidating project finance; investments which were expected to pay their debt service cost and generate enough funds to repay the initial advance. Sinking funds and other structural devices for spreading repayment through time were commonplace. It is also obvious that the debt bearing capacity of a society is a function of the term of the debt.

It is clear that, left to its own devices, the banking system can create more credit that can be supported by the wealth which supports it. Consideration of a banking system which lends only for consumption makes this point obvious. The debts in this system would clearly have lower value than the wealth from which it was generated and could not be fully repaid other than through further refinancing of the debtor. We will revert later to this latter situation, which is often referred to as a question of confidence.

Holmstrom and Tirole have demonstrated that under uncertainty banks may not produce adequate liquidity and economically valuable projects may go unexploited. This introduces a role for the central bank as the insurer of systemic liquidity, and in extremis its role as lender of last resort.

There is an important further dimension to this – the source of the wealth. When this is domestic, the system is closed. We owe it to ourselves in aggregate. Debt service and repayment are then purely a redistribution issue among the members of the society. When this wealth is international in original, debt service lowers the income which is available for domestic consumption and wealth generation.