A Comparison Between Traditional Banking and Market Based Finance

I've been working on a paper using stylised balance sheets to look at the interaction between traditional banking and market-based finance.  The aim is to examine the reasons for shifts between these two intermediation models before and during the crisis and the implications for liquidity and solvency across the sector.  In this post, I just want to give a brief overview of the differences between the two techniques.

The diagram shows some simplified balance sheets.  These are not necessarily supposed to represent distinct entities, so much as different business models.

The assets of traditional banking consist of a loan portfolio and a holding of high quality liquid assets (HQLA).  The liabilities will be short term unsecured deposits with the balance forming capital (there will be other borrowings of a longer term nature, but we can ignore those here).

 The loan portfolio will be long dated assets which will generally be fairly illiquid, in that it is difficult to sell them at short notice for a price close to balance sheet valuation.  Given the short term nature of its deposit liabilities, the traditional bank is therefore running a maturity mismatch risk.

To deal with this mismatch, the bank maintains a holding of HQLA.  This might consist primarily of government securities, but central bank reserves would also be included in this.  In the event that it faced deposit withdrawals in excess of the rate of run-off on its loans, the bank would hope to cover this out of its HQLA holding.

The traditional bank is therefore running a kind of liquidity insurance.  It is offering liquidity to all of its depositors but covering this with a much smaller quantity of liquid assets.  Its risk is that more depositors wish to withdraw their deposits at once than it can cover out of liquid assets.

With market based finance the assets are all marketable instruments.  I have illustrated this here as securitised loans.  The underlying loans may be similar to those held as illiquid assets on the traditional bank's balance sheet.  The process of securitisation transforms these into a standardised tradeable instrument.  This means they can be bought and sold much more easily.  With sufficient market depth, there will be a continual market price for these securities and a holder will expect to be able to sell at close to that price.

Having liquid securities as its assets allows an entity using market-based finance to fund itself using secured financing, for example repo.  The repo funding provided by investors is effectively secured by underlying assets (the securitised loans).  The value of the security provided is slightly higher than the amount of funding and, critically, is adjusted every day for market movement.  This is only possible because the collateral is in the form of marketable securities with a transparent market price.  The level of additional security provided therefore only needs to cover possible day-to-day price movement.

In the market based finance model, all of the assets are used as collateral[1].  There is no pool of unencumbered liquid assets like the HQLA held by the traditional bank.  However, no such pool is needed, because the market based finance model is providing liquidity in a different way.  Rather than rely on a type of insurance as in traditional banking, the liquidity transformation relies on the marketability of the underlying assets.  Whereas a traditional bank would dip into its pool of unencumbered liquid assets to deal with a high level of withdrawals of its funding, an entity using market based finance would simply look to sell down its regular asset base.

Both business models face liquidity risks, but they manifest themselves differently.  The liquidity risk in traditional banking is that the HQLA holding is insufficient to cover withdrawals.  The liquidity risk in market based finance is that the liquidity of the assets fails.  For both models, the robustness of their liquidity strategies may change.

A depositor with a traditional bank is taking credit risk on the bank.  They do not know the intricate details of bank's balance sheet, so they are relying on the skills of the bank's management in making good business decisions.  They will however draw comfort from the level of capital which provides a cushion against loan losses.

A provider of finance in the form of secured funding takes much less risk on the borrower itself, relying instead on the collateral provided.  Some degree of over-collateralisation is needed to meet rating requirements or to cover haircuts, but generally the market based finance model requires lower levels of capital.

A number of developments within the financial intermediation sector in the run-up to the crisis and during it can be traced to the differences between these two business models.   These include changes in liquidity and solvency levels, changes in lending appetite including credit criteria, and changes in credit spreads.

---

[1] There will repos with a variety of different counterparties.  Distinct collateral is used for each counterparty; they do not share in the same collateral pool.

What's OK and What's Not on Loanable Funds and the Natural Rate of Interest

I've read various things recently on the theory of loanable funds and the natural rate of interest, so I thought I'd say something on it.

I want to start by looking at how we might illustrate a market of loanable funds in a typical demand and supply graph, with quantity on the horizontal axis and some benchmark rate of interest on the vertical axis.  The supply curve then shows the amount that people wish to save at each interest rate, other things being equal.  The demand curve shows how much people wish to borrow at each interest rate, other things being equal.

I think this graph makes a kind of sense.  However, we have to be careful.  Normally, we might use such a graph to show how a change in either demand or supply would lead to a change in price (the interest rate) so as to restore equilibrium.  Let's suppose that this graph showed demand and supply for apples instead.  Then, if the demand curve shifts to the right say, we might expect a rise in the price of apples, changing quantities supplied and demanded until they were equal again.

Bear in mind here the assumption that all other things are equal.  In fact this is an assumption that is almost certainly incorrect.  An increase in demand for apples requires changes in demand and supply in some other market.  You can't just demand more apples; you have to demand more apples instead of something else - bananas say (or leisure time).  So the demand curve for bananas will also move, with implications for the price of bananas.  That in turn will impact on the demand for apples.  So clearing in the apple market is not brought about exclusively by a change in the price of apples, but by changes in all markets.

This is particularly important when thinking about the loanable funds market.  Critically, one of the things assumed constant is the level of income.  So the supply curve shows the amount people wish to save at any given level of income.  Now we need to consider what happens if people wish to save more.  On the face of it, this would lead to a shift in the supply curve to the right.  However, like with the apples, you can't just save more - there has to be some counterpart, which in this case must mean spending less.  And spending less results in lower incomes.

This means that if we want to consider an increased desire to save, we cannot simply represent this as a rightward shift in the supply curve leading to a fall in the rate of interest.  In fact, the income implications of an increased desire to save may result in both the demand and supply curve shifting leftward.

So we need to be careful about applying the standard story to an interest rate clearing the market for loanable funds. 

However, if that were all, there would still be some value in drawing a demand and supply graph for loanable funds.  We might show curves that represented demand and supply on the assumption that all other markets were clearing, including in particular the employment market.  This would then tell us what level of interest would prevail in that situation.  It would not tell us how that level of interest came about (and we know to be suspicious of the idea that it arises from supply and demand pressures in that market), but we would know that if every market was clearing, then that must be the rate that would hold.

This rate would then be the natural rate of interest - the rate of interest at which planned savings equals planned borrowing, under conditions of full employment.  You might want to invoke rates of time preference or marginal efficiency of capital, but this is not really necessary for a natural rate to exist.

Nevertheless, it is not clear that such a rate does exist.  If we make certain assumptions about household preferences or production functions, we can certainly show that there is a rate which meets this condition.  But these assumptions are completely ad hoc, and there is no reason to believe they reflect the real world better than some different assumptions.  It is entirely possible that when we draw out the demand and supply curves for loanable funds, that we find that there is no rate for which demand equals supply under full employment.  For example, the graph may look like this.

If that were the case, then full employment would be impossible.

Much of the current consensus approach in economics relies on the idea that full employment (and price stability) can be achieved by setting the market interest at the natural rate.  It may not matter how such a rate is determined, but is clearly essential for this that such a rate does in fact exist.  It is far from clear to me that it does.

On the whole, I find objections to loanable funds and the natural rate of interest overdone.  As theoretical concepts, I think they have their place, if used correctly.  However, I am very sceptical of the idea that the monetary policy is all about matching the market rate to the natural rate.  This is mainly because I have doubts about the stability of the relationships involved.  But at a more basic level, it's also because I don't think we can take it for granted that there is in fact any interest rate that can achieve clearing in all markets.  

The Option to Monetise

Nick Rowe has a very confusing post about sacrificing goats.  To be honest, I didn't really understand the post, but reading some of the comments and his replies, as well as following his link, made it a little clearer.  So hopefully, I've got I right what he is trying to say.

The point relates to the possibility of funding government deficits with newly created money, rather than bonds.  Nick's point (I believe) is that, given certain assumptions about how the central bank conducts monetary policy, it doesn't make any difference how its funded.  If we assume that the central bank is targeting a certain level of NGDP and acting in the market to achieve that (including buying and selling government debt), then any change in deficit funding will be exactly offset by a change in central bank actions.

If my interpretation of his point is correct, then I'd agree with it.  However, I wanted to look at a slightly different point (but related, I think).

It is sometimes suggested that funding deficits with newly created money is a way of addressing rising levels of public debt.  The concern is that debt must eventually be refinanced, and if there is an ever increasing amount, this refinancing may eventually become impossible at an acceptable price.  At this point, the government will be forced into raising taxes.  This then leads to Ricardian Equivalence concerns.  If debt is inevitable going to lead to higher taxes in the future, then people may feel they need to start saving for it now.

Regardless of the question of how realistic this is, monetisation of debt seems to avoid it.  Newly created money never has to be repaid, so there will never come a day when taxes have to be levied to pay for it.  So maybe monetary financing of deficits is the way to avoid the problem of rising public debt.

The point I would like to make here, is that if this is correct (and I'd say it is), then it is not actually necessary for the government to carry out the monetary financing.  All that is required is the knowledge that they are prepared to do so at some point in the future, if necessary.  In other words, once you have admitted the option to monetise if appropriate, you have already removed any constraint that a rising debt level imposes.

If we assume that public debt carries a higher rate of interest than public money, then the debt / money mix will impact on net current transfers.  But otherwise, what really matter is whether debt monetisation is an option at all, whether it is actually used is less important.