Thursday, February 14, 2013

How can governments stop encouraging banks to be highly geared?

A reader of my book, Free to Flourish, is puzzled by a brief comment I made about fundamental weaknesses in the financial system. He asks whether the following passage implies the existence of a fundamental market failure with respect to the financial system:

'The underlying incentives that the system provides for participants to take risks with borrowed funds might even tempt saints to behave imprudently. Another outbreak of gambling with borrowed funds will become increasingly more likely as memories of the recent crisis recede, unless fundamental reforms are introduced. Required reforms include the removal of any implicit guarantees that any financial institutions are 'too big to fail' - by taking action to penalise rather than assist the owners of financial institutions which are at risk of default - and removal of distortions in tax systems which favour debt funding relative to equity funding' (Chapter 8).

I accept that there may be market failure in the financial system. There can be negative externalities associated with bank failures. If the failure one bank leads to loss of confidence in some other banks, there may be a market failure involved. Then again, there may not be. If the failure of one bank leads to loss of confidence in banks that have taken similar risks, leaving other banks unaffected, it would be reasonable to argue that the market is just taking appropriate account of new information. Nevertheless, at an aggregate level, I accept that central banks may be able to play a useful role in sustaining expectations of ongoing growth in aggregate demand when bank failures occur.

However, my concerns about the fragility of the financial system – as it exists at present – cannot be attributed to market failure.

The following hypothetical example might help to begin to explain the nature of the problem as I see it. Let us focus on two banks competing in a free market, without government interventions. Both banks are the same in nearly all respects, but while Bank A is profitable, Bank B is having difficulty competing for deposits. The reason for this is that the level of shareholder equity in Bank A is relatively high and potential depositors feel that the interest rate being offered on deposits in Bank B (the same as for Bank A) would not adequately remunerate them for the additional risks they would be taking. 

There are several options that Bank B might consider to become more competitive. For example, it could offer a higher interest rate to reflect the greater risks involved for depositors; it could reduce the risks in its asset portfolio (perhaps by having a higher proportion of its portfolio in relatively safe government securities); or it could issue more equity capital and become more like Bank A. The optimal level of equity depends on factors such as the riskiness of the bank's asset portfolio and the extent to which depositors require higher interest to compensate for risk.

Is this example plausible? Is it conceivable that it might be possible in a free market for a bank to be profitable with a relatively high level of shareholder equity? Many would argue that the example I have given is unrealistic because an equity risk premium must be paid for access to equity capital. On that basis, it is argued that banks with relatively high equity could be expected to have a relatively high cost of capital and thus to be less profitable than banks with relatively low equity.

Anat Admati and three of her colleagues provided a pertinent response to the suggestion that increased equity would increase funding costs for banks in their paper: 'Fallacies, Irrelevant Facts and Myths in the discussion of Capital Regulation: Why Bank Equity is Not Expensive'.  These authors draw attention to the Modigliani-Miller (MM) analysis which shows that increases in the amount of equity relative to debt financing simply re-distribute risk among investors. The total funding cost is determined by the total risk that is inherent in the bank's asset portfolio and is independent of gearing. In that context, any losses from using less borrowed funds must be offset by the correspondingly lower cost of equity capital.

The essential assumption of the MM analysis - apart from the assumption (discussed below) of no government intervention favouring either debt or equity funding - is that investors are able to take account of portfolio risk and gearing when pricing securities. Admati et al make the telling point that banks make this assumption in managing their risks.

So, what happens if we relax the assumptions of the MM model by introducing a tax system that encourages debt relative to equity, a government guarantee that banks will not be allowed to fail and protection for depositors? We should expect to get banking systems that are highly geared and fragile – like our current banking systems.

How can governments remove those distortions?  The obvious answer is just do it! However, removal of the tax distortions will require major tax reforms in countries that have classical company taxes. The problem in relation to government guarantees and protection of depositors is that announcements  that they will no longer apply are not likely to be credible (except when made by governments that are so heavily indebted already that further bank bailouts would be impossible in any case).

Does that mean that the best option is for governments to regulate bank behaviour to such an extent that bank failure becomes highly improbable? That approach would suggest that if Basel III is not restrictive enough to make bank failure sufficiently improbable, we should be prepared to move on to Basel IV, and then Basel V, and even to nationalisation of banking if necessary.

At that point I begin to see red. If we are not dealing with a market failure, why are we attempting to displace the market? Is it really necessary to put the entire banking system into a regulatory strait jacket, with all the inefficiencies that involves, in order to live with the consequences of past regulatory failure? Would it not be possible for governments to make a credible commitment never to bail out another bank if they were prepared to spell out punitive action to be taken if regulatory agencies assess banks to be at risk of default? For example, why not announce plans for pre-emptive action to install administrators to restructure banks if they are assessed to be at risk of default?  


With the benefit of comments from kvd and Jim Belshaw (see below) it is clear that the line of argument presented above is not as clear as it could be and contains some unnecessary red herrings.

1.      The definition of banks. For the purposes of this discussion, the distinguishing characteristic a bank is that it is a company with relatively low shareholder equity and a relatively high proportion of debts repayable on demand. Later in the post, my focus is narrowed to financial institutions with deposits guaranteed by governments and/or viewed by governments as 'too big to fail'.

2.      The definition of externalities and market failure. The discussion in the paragraph immediately following the quote from Free to Flourish raises issues concerning the technical definition of externalities and market failure that are a largely a red herring from the perspective of the general line of argument I am developing here. All I needed to say was that while I acknowledge that there may be a case for government intervention based on the existence of market failure, that is not the basis for my concerns about the fragility of the banking system. (Nevertheless, the discussion is raising interesting points. There might be something wrong with our definition of market failure if new information about bank solvency that leads to the collapse of the banking system does not qualify as evidence of market failure. The question that kvd has raised about whether there is a case for government guarantees to cover use bank facilities for every day transactions using is alsoof interest to me. I will try to follow that up in a subsequent post.)

3.      My hypothetical example involving Bank A and Bank B. The example seems to have clouded the point I was trying to make, rather than illustrate it. The point the example was intended to illustrate is that in a free market banks would not have an incentive to seek ever-greater leverage. The rate of return on shareholder funds may rise as leverage increases, but depositors and shareholders would have an incentive to take account of the increasing risk of bank insolvency. As leverage increases the cost of borrowing additional funds could be expected to rise (i.e. the interest rate on deposits would need to rise). And at some point the increase in expected return on shareholder's funds will not be sufficient to compensate shareholders for the increased risk of failure of the firm.

4. Should the Australian government continue to guarantee banks deposits? That is the title of a later post in which I discuss issues raised by kvd.


Anonymous said...

Winton, you mention your own 'seeing red' and I guess this post raises the same sort of feeling in me.

First, and if I'm wrong I apologise, you do not define what you mean by 'bank'; then you state 'Bank A is profitable' without defining just what you mean by the term 'profitable'.

Second, your acceptance that 'the market' should play any part in the securitisation of depositors' funds (alongside equity participants) offends against my own beliefs.

Third, I would not seek in any way to regulate or limit the rich investing their money in any way they wish. But government failure to differentiate between the basic needs of their populace, and the desires of a relatively small, select group of players - that I find a complete abrogation of a basic government role - more specifically, a responsibility.

By all means let's limit government involvement and guarentee - but let's first more clearly delineate what it is that government should be obliged to protect.


Winton Bates said...

Hello kvd

Thanks for telling me about your feelings. I mean that - writers need feedback about the feelings of their readers.

I didn't define what I meant by a bank. The distinguishing characteristics are corporations that have relatively low shareholder equity and a relatively high proportion of debts repayable on demand. Later my focus is narrowed to financial institutions with deposits guaranteed by governments and/or viewed by governments as 'too big to fail'.

Bank A has normal profitability i.e. profitability similar to that of other corporations facing similar risks.

I don't understand your second point. What is it about the idea that the market should have a role that offends against your beliefs? Are you suggesting that depositors should not be expected to take account of differences in the risks involved in placing their funds in different institutions?

Your third point seems to be that the government has a responsibility to protect the public against the potentially harmful activities of a small group of players in financial markets. I agree in principle. The problem is how do we define harm and what should be done about it.

The public should be protected against fraud. If a gambler borrows money and promises to repay on demand, is that a fraudulent promise? Perhaps it is if his only source of funds to repay his debt is his potential earnings from gambling. The issue is what should we do about such situations? Should we attempt to regulate all financial transactions in great detail? Or should we focus on punishing fraud in order to deter such behaviour?

Anonymous said...

Hi Winton

What's not to like about a system where the government establishes an absolute monopoly (currency) then subdivides that monopoly into four parts (big banks) yet seeks to limit its responsibility towards depositors who are basically using said monopoly as a means of 'transaction', rather than 'investment'?

As far as I can work out our banks total funding sources might be analysed as long term external 25%; short external 20%; shareholders 7% - leaving depositor funds 48%. Of that 48% I believe roughly half is in the form of fixed term deposits, with terms of up to two years.

It is the other half of that 48% (if you accept my rough figures) that I believe should be subject to government guarentee; market rates/risks on all other sources are not my concern. With the foregoing as background, I would answer your specific questions as follows:

1)See answer to question 2

2)That is exactly what I'm suggesting.

3)It is not a fraud if the gambler is up front about his intended 'investment'. Relate this to banks, or companies generally, and that is why we have government bodies to (i) establish and (ii) enforce, monitor, rules of disclosure and governance. It is the second bit in which governments appear to be either deficient or incompetent.

4)What should we do? Your two alternatives are not 'either/or'; the government has a responsibility for both. But I'd add a third. The government should recognise a responsibility toward those who use banks in the ordinary course of transacting their affairs by legislating for a complete divorce between the needs of those people, and the rest of what banks get up to these days - what used to be simplistically termed 'merchant banking', but which is now more akin to 'magic mirror money manipulation'.


Winton Bates said...


If a firm promises to repay a debt on demand but doesn't do it, that seems to me to be very close to fraud. I'm not suggesting that the chief executives be sent to debtors prison, but if we want to live in a society where people keep their promises those who make promises they can't keep should not escape without penalty and be free to repeat the process.

I still think I am on the right track in arguing that if governments can't help themselves from encouraging risk takers it is better to try to deal with the consequences by targeting the 'dangerous drivers' rather than imposing a very low speed limit on everyone.

I have some sympathy with the idea of protecting mum and dad(and grandma)from the evil bankers but I'm not sure they actually want that protection. They seem to prefer to live dangerously and deal with firms like Pyramid and Banksia rather than the four pillars.

Anonymous said...

I have some sympathy with the idea of protecting mum and dad(and grandma)from the evil bankers but I'm not sure they actually want that protection. They seem to prefer to live dangerously and deal with firms like Pyramid and Banksia rather than the four pillars.

Winton, nothing I said was directed to these sorts of 'mum and dad' investors - other than the need to oversee good governance, and prevent outright fraud.

Rather, the people who deserve protection by guarantee are those who cannot avoid using bank facilities simply to engage in normal life. They have no other means of shifting funds from A to B other than depositing it in a bank. They are not investors; they are transactors. They represent nil risk to the banks, and provide their funds interest free.

But referring to investors, if you wish to put it in terms of 'lowering the speed limit', then I'd suggest what is actually needed is increased competency within the various governments already in place. We are not overburdened with regulation at present; we are disadvantaged by lack of competent oversight of what exists.


Winton Bates said...

The question of transactions is interesting. There might be potential for government regulation to reduce transactions costs, but I doubt whether that actually happens in the real world.
I suggest we continue this discussion later, after I have read and reviewed 'The Banker's New Clothes'by Anat Admati and Martin Hellwig. This book, just published, might provide some useful insights into problems of bank regulation.

Jim Belshaw said...

Winton, a clarificatory question. please. What is the profitability of Bank B compared to A?

Winton Bates said...

Hi Jim
I'm not sure. Whatever the profitability of firm B, it isn't sustainable. In retrospect it might have been better to follow comparative static conventions and assume an initial equilibrium with all firms making normal profits and then for the owners of firm B to have the bright idea that it might be able to improve profitability by becoming more like firm A, but with a lower ratio of equity to debt.

The point I am trying to make is that it is plausible that in a free market there would not necessarily be an incentive for banks to have a low level of shareholder equity.

Jim Belshaw said...

I understood the point, Winton. I was just struggling a bit with the analysis, trying to untangle the issues in my mind. On the surface, it applies to all institutions in the business of borrowing money so that they can lend it; banks are one category. All sell debt.

Focusing just on vanilla finance, With a lending institution profitability depends on the difference between cost of funds and interest received on loans adjusted for bad debts. So long as that margin is positive, profits will increase with gearing, but so will risk to the institution.

Additional equity capital allows the institution to lend more in terms of the capital plus gearing, so profit goes up. To the degree that the additional capital improves risk perceptions among lenders, the institution can borrow more at any given interest rate, the same at a lower interest rate.

It's not just absolute profits that are important, but return on shareholders' funds. To what degree will the additional equity increase or reduce the return on shareholders' funds?

In all this, it's reasonably easy to specify a mathematical case in which an increase in bank equity may increase both absolute profits and return on SHFs.

I have been speaking of vanilla lending. Financial institutions also make money from sale of financial products and fees. These transactions may also involve some combination of lending or use of the balance sheet. In the case of the Sydney Harbour Tunnel, for example, Westpac Project and Advisory Services put together a package for which it received a fee. This flowed straight to income. Westpac also lent money to the project. All this created an incentive to do deals now for immediate profit ignoring longer term risk.

Now how does all this relate to market failure? I'm not sure that it does. A competitive market implies a degree of failure among participants. Just because a bank or a whole series of banks goes down even to the point of collapsing the economy does not mean that there has been market failure.

The real issue here to my mind is one of externalities, not market failure. Financial institutions in general and banks in particular are regulated because of the broader costs of failure.

Winton Bates said...

I think your analysis does help untangle the issues. I can restate the basic question behind my example as follows:
If return on shareholder funds increases with leverage, what is it that stops companies from seeking ever-greater leverage? As you imply, the risk to the survival of the firm increases with leverage. This could be expected to have two implications. First, as leverage increases the cost of borrowing additional funds could be expected to rise. Second, at some point the increase in expected return on shareholder's funds will not be sufficient to compensate for the increased risk of failure of the firm.

As discussed above, the analysis has to be modified to take account of tax and government guarantees (in the case of banks).

Your mention of private funding of infrastructure raises questions in my mind about the recent failure of firms involved in some major projects. It seems obvious to me (with the benefit of 20-20 hindsight) that their funding model relied too heavily on debt - given the inherent risks involved in these projects. Perhaps the risks were less obvious at the time, but I still wonder whether there are still factors in our tax system (despite dividend imputation) that might encourage such firms to rely excessively on debt funding.

With regard to your final point, when I think of market failure I tend to think in somewhat narrow economic terms of it being attributable to existence of externalities or public goods (which are just an extreme case of externalities).

I have to admit that it is difficult to fit some apparent market failures (e.g. speculative bubbles) into that framework. For example, if an epidemic of excessive confidence leads to higher leverage in many firms and results eventually in widespread business failure, is that evidence of market failure? I find it hard to put my finger on a particular externality that might lead to that result. An epidemic of irrationality doesn't seem to fit easily within a definition of externality.
In any case, the fundamental issue is whether there is anything a government agency could do that would result in better outcomes. I think it is appropriate for people at the RBA to be thinking about this question (as they have been).

kvd: I am now feeling somewhat guilty about suggesting that we postpone our discussion. If you are still reading, please feel free to re-join this discussion.
In any case, I will try to pick up the threads in my review of the Admati book.

Jim Belshaw said...

It's interesting, Winton. Looking at you comment on market failure, I made a distinction between market failure and externalities. You treat externalities as an example of market failure. Your definition is consistent with that put forward in wikipedia. I am now wondering about the value of such a broadly defined concept.

To spell this out a little, there are many different interacting markets. A financial institution operates in multiple markets, each of which can be analysed separately. It borrows money, it lends money to people or firms engaged in multiple activities in multiple markets, it provides packages of services, it competes for people, it is concerned about the market for its shares.

Take kvd's point. Banks provide transaction services. We put money with them and then use that money to fund day to day activities. Banks make a profit on that service through charges and interest etc received on loans made from such deposits. This is a distinct market place with its own dynamics and indeed its own regulatory structure. It is also an imperfect market place because of the importance of networks.

In simple terms, a market is a place where buyers and sellers meet. An efficient market is one where prices move such that goods are cleared. In thinking of markets, we equate efficient markets with competitive markets. This gave rise to regulation to try to prevent abuse of market power, to enforce competition, to overcome market failure.

This concept of market failure is quite independent of the concept of externalities. Take the heroin market place. This is clearly an imperfect marketplace. Those imperfections come about because Governments believe that heroin poses health risks, adverse outcomes that should be avoided or reduced. Whether those interventions are effective is beside the point. The point is that we have a clear distinction in this case between market failure and externalities.

Now factor in multiple interacting markets. Assume that each market is efficient and competitive. There is no market failure at the individual market level. However, problems can still arise because of the interactions between markets.

Consider the bank transactions case. A bank fails because of its activities in another market. Of itself, this says nothing about market failure in that market. The concept of efficient markets does not imply that all buyers and sellers will benefit or avoid mistakes. The effect of the failure, however, is to damage the transactions market place in a particular place. That market place has failed. Now we have an example of market failure induced by interactions between markets. Again, we can distinguish between market failure and externalities. The bank collapse was not a market failure. However, its impact on the other market was an externality that led to market failure.

The question becomes more complicated by the way we so often define markets in geographic or jurisdictional terms. However, I am going to leave my comment here until a later time, for it brings us into another disucssion area between us.

Winton Bates said...

I agree with you that the concept of market failure is of limited use. I think Harold Demsetz made a good point that the relevant choice is not between an existing imperfect market and an ideal norm of a perfect market, but between likely outcomes under current institutional arrangements and a proposed alternative set of institutional arrangements.

At the same time, some arguments for government intervention are stronger than others. For example, arguments for industry assistance based on externalities (e.g. those associated with research or training) can be more powerful than those based on the fact that an industry employs people, or exports, or displaces imports. Again, the power of arguments to intervene to ensure competition depend on contestability rather than just observation that there is a divergence from assumptions of perfect competition.

In some instances there may appear to be strong grounds for government intervention for whatever reason, but the proposed method of intervention may produce worse outcomes. For example, in the case of heroin, even if we accept the argument that potential consumers should be protected from themselves, we still need to consider the effects of existing laws in encouraging organised crime, corrupting law enforcement agencies and providing an inducement for users to turn to crime (including drug dealing) to finance their habit. We should also consider the question of whether sending users to jail enhances their lives.

In the case of the transactions services provided by banks, it is obviously very important for a high level of trust to exist among all the participants involved. When some institutions begin to doubt whether other institutions can be trusted, transactions costs go up astronomically. That might suggest a case for governments to intervene to provide insurance, but before jumping to that conclusion we should consider how government intervention might affect the incentives of the players. If government involvement displaces private insurance and increases incentives to take risks - and a great deal of regulation is required to offset that - it might be more trouble than it is worth. I don't have enough information at present to make strong assertions one way or another on this question.

Anonymous said...

Winton, Jim, this very interesting discussion is now in territory 'beyond my paygrade' - as someone recently excused himself.

My interest was initially piqued by what I referred to as the 'securitisation' of a significant part of the funds sources available to banking institutions - namely those funds deposited in the ordinary course of getting on with one's life. If you accept my figures, this amounts to somewhere north of 20% of the funds available for them to pursue their objectives.

While I would be the last to suggest any of the 'big four' are in danger of collapse, I do think that in your higher level analysis of 'marketplaces' and 'risk assessment' it begs the question as to just what is represented by the 20+% of unsecured creditors (because that's effectively how depositors' funds are treated; and that's why there were recent queues outside various high street banks and building societies in the UK) which I termed 'transactors'.

My simple point remains that these funds should be regarded more as the old fashioned 'Trust Fund' one sees in any solicitors' practice. Yet that is not where they presently sit in calculation of leveraging. Within that they are subsumed in those funds available to satisfy any higher-secured obligation. Except for shareholders, they are in fact last in the queue, along with any other trade creditor.

When one thinks of such funds, Winton mentions the 'mum and dad investor'; the implication being that the sums are small, difficult to manage, an annoyance really in terms of transaction costs.

But when I think of those funds I'm referring to my working cheque account (which varies from $2000 to $25,000 quite regularly - and for which I dearly pay fees) and also a present property transaction I'm involved in (gross about $500,000). These funds are sloshing around in the banking system, available (God forbid) at any time for our banks to satisfy secured creditors. Come a crunch, my funds are essentially an unsecured interest free loan to my bank, available for them to pursue (did you term it?) enhanced shareholder returns.

Too much regulation involved to protect such funds? I'd suggest a reclassification of such funds as first charge government backed liability. Would that would necessitate a recalculation of the risk attaching to other funding sources? Yes, and so be it; the market will decide that.

I'll leave the wider issues to you both. My interest was restricted to this specific area.


Winton Bates said...

kvd: Thanks for clarifying the point you are making. I need to think about it.