Careful what you wish for
In a recent column Bernard Hickey suggested the following:
Taxpayers still face the risk of seeing bank losses socialised in future while today’s profits are privatised.
A more honest solution would be for the Government and the Reserve Bank to openly state that a bailout would not occur. Term depositers would demand a higher return to compensate for the higher risk, and it would remove the moral hazard that currently subsidises the profits of Australian banks.
Now lets be a bit careful here. Yes there is an implicit government guarantee of banks – in fact the way to look at it is through the lens of a deposit guarantee, when it comes to the big banks the New Zealand government will not allow depositors to lose out.
The logic behind this is the fear of a bank run. The reason we hadn’t had a financial crisis on a global scale between the Great Depression and 2007 was largely due to the implicit deposit guarantees that soveriegn nations had put in place during the 1930s. Even if these were not always “explicit” they helped prevent runs on the banking system, which engendered confidence and prevented financial crises. One key reason for the GFC during 2007-2009 was the sudden change in behaviour by governments – where they were showing themselves suddenly unwilling to provide this insurance on the basis of “moral hazard”. It is true that issues of moral hazard had helped to drive risky lending, but it was confusion around where the burden of debt fell and a lack of clarity around who was “implicitly insured by government” that led to a run on wholesale financial markets and the financial crisis.
The “solution” to any perceived issue in our banking system is not to get rid of deposit guarantees, and it is not to remove the implied subsidy that this may provide to the New Zealand banking system. It is to ensure that banks in turn face this cost during the rest of the cycle.
The RBNZ’s desire to set up the OBR is based on a desire to prevent bank runs, while also making who bears the burden of a bank failure “fairer”. They are trying to ensure that we don’t have a financial crisis, while minimising the cost associated with “moral hazard”. This is preferable to forgeting the lessons of the Great Depression and GFC, which is what we would be doing if we were to completely pull away from the implicit back stop of the banking system. Trust me, I don’t like implicit insurance for industries myself – but financial markets are one area where such things need to take place, and as Hickey says they have to take place in a transparent manner.
My understanding is that when these guarantees were first put in place there were dozens of banks in most countries, which limited the likely price to a government. When the banks consolidated – in the UK and Australia this now means four big banks – the likely cost of a guarantee became much higher. Hence the talk of banks being “too big to fail”.
The key thing to remember is that the size of the bank is endogenous with respect to the regulation – so if the banks are “too big” that is a suggestion that the implicit regulatory framework promotes banks being too big.
The idea here is that, given the implicit guarantee holds for large banks there is a return for being big, and so this is taken into account. Now this isn’t necessarily a bad thing – it could be a feature of the system rather than a bug.
Remember, in the GD people were complaining that banks were “too small” and as a result people were relying on very aggregate information in order to tell if their deposit institution was going to fail – so a pick up in failure rates led to withdrawals from banks that were otherwise fine, leading to bank runs and financial criises.
The issue isn’t the size of the banks – as the implicit guarantee is jusified in the case of BOTH banks that are “too small” and “too large”. The issue is making sure that any externality, or moral hazard, that comes from regulation is appropriately taken into consideration by financial institutions – this has been the driving force of changes to bank regulation since the GD (when deposit insurance was introduced).
It is a world of second-bests because of the idea of asymmetric, and imperfect, information in credit markets. Regulation has to, and does, take this into account. The reason the crisis has continued for as long as it has is because the EU and ECB seem both unwilling and unable to accept many of these lessons – which is tough given how much of the global financial system is tied to the European financial system.
Rod Carr’s RBNZ note from 2001 is on point.
http://www.rbnz.govt.nz/research/bulletin/1997_2001/2001jun64_2Carr27jun.pdf
Very nice – and also a good indication of how these issues were being clearly discussed prior to the financial crisis.
I’m with Hickey on this one. The whole idea of the business system is that there are profits and loses and banks have to face the same incentives from that system as any other business. Implicit insurance just creates huge moral hazard problems, the answer to which is banks going under in the same way as any other firm. Implicit insurance gives the incentive for greater risk taking by banks and thus the incentive to crate the very problems it is supposed to fix. Also the knowledge that there is no government guarantee for banks gives customers an added incentive to monitor banks more carefully. If there is the possibility of a bank run the banks themselves have the best information about what is going on and should have the incentive to so something about it. Insurance means they don’t. If bank A is in trouble because of a bank run, banks B, C and D should have the incentive to help feed the run so to stop it before it really gets started. The biggest danger to the banking system is the belief that banks are “too big to fail”.
I would say that the pre-GD experience in of itself justified deposit insurance – and being honest about that and trying to work out what would be “second best” in the financial system is best. As Sargent discusses here https://files.nyu.edu/ts43/public/research/phillips_ver_9.pdf
Yes, individual banks take more account of risk if they know there is no government backing, and yes the banking system as a whole has an incentive to co-operate when faced with a bank run – we observed both of these things in the pre-Fed US.
However, even with these solutions we experienced constant bank runs where the banking system was able to co-ordinate to avoid significant loss, but the impact on broader conditions was significant. The financial system is one area where the multiple equilibrium argument does hold.
Also I really don’t like criticising the size of banks in of itself – banks are big because of government policy, undeniably, but in the GD we justified bailouts because banks were “too small”. The size issue of individual institutions is a misnomer, it is the fact that financial firms fill an intermediary role across the entire economy, one that is strongly dependent on expectations, that is the key point.
I’m not sure that it’s consistent to argue both that “issues of moral hazard had helped to drive risky lending” and that there was a failure of the LOLR facilities. What is the evidence that implicit deposit insurance caused risky lending by non-lending or deposit taking investment banks like Lehman’s? How does LOLR, apart from lowering the cost of deposits and hence credit, encourage risky lending? Also, the reasoning behind stipulating capital requirements for banks and then asking for a insurance premium on deposits as well needs a bit more explanation. Why do we need both? (Comments to no-one in particular.)
Hola,
“I’m not sure that it’s consistent to argue both that “issues of moral
hazard had helped to drive risky lending” and that there was a failure
of the LOLR facilities.”
It is a matter of ex-ante expectations vs ex-post outcomes. Ex-ante we may experience choices that are due to “moral hazard” on the expectation of deposit insurance – ex-post deposit insurance may be optimal and we just decided not to do it. This gives a potential description of what may have happened.
“How does LOLR, apart from lowering the cost of deposits and hence
credit, encourage risky lending?”
That is exactly where the lending is “too risky” – as the “price” of credit is lower than it would have been (both in terms of the actual price, and in terms of the quality and monitoring of the lending). The existence of a backstop leads to a situation where there is a difference between the “social” price of a loan and the “private” price – that will be the argument used by those selling moral hazard.
“Also, the reasoning behind stipulating capital requirements for banks
and then asking for a insurance premium on deposits as well needs a bit
more explanation. Why do we need both?”
A good point IMO. One is setting a quantity, another is setting a price. I prefer the insurance premium because I prefer price mechanisms, as we have a better idea of the “cost” that a scheme may have socially, than we do about the right allocation or quantity that should occur. I use the same logic when I say I prefer a carbon tax to a cap and trade scheme.
If the crisis was mostly due to risky lending then no amount of LOLR would help – the banks would be insolvent.
The “price” of credit is the sum of the deposit rate and a credit margin. If deposit rates are too low, it does not necessarily follow that the credit margin itself is too low. Loan origination does not depend on deposit rates, only the assessment of the credit margin. The movement of deposit rates does not itself lead to risky lending as it is not a factor in the lending decision, only the borrower’s credit worthiness and their associated capital requirement. If there is overall too much lending then the CB should raise rates – the individual banks are not in a position to determine the social cost of the CB’s policy settings.
Given that banks exist to convert non-marketable and opaque private loans into marketable and liquid deposits, it would be very difficult to have a market-based insurance premium. Hence the capital requirements and ability of regulators to see deeply into the bank.
“If the crisis was mostly due to risky lending then no amount of LOLR would help – the banks would be insolvent.”
Even without moral hazard, we can still have the crisis being due to a failure of the LOLR function in terms of expectations – saying moral hazard exists merely provides the “trade-off” that we like to discuss.
So the moral hazard idea isn’t so much the trigger for the crisis -but a potential explanation of how we got to where we are. Of course if we were to use that explanation we have to ask why this wasn’t the case earlier! Whatever explanation we use needs to be consistent with an explanation of why the shadow banking system appeared 🙂
When it comes to the implicit subsidy I’m thinking along these lines:
http://www.voxeu.org/article/implicit-subsidy-banks
It is to do with the “counterfactual” cost. A clear piece of data we can look at to get an idea of that is credit default swaps, the implicit government guarantee had pushed these down to levels that were not representative of the true “risk of default”.
Even if banks are benefiting from cheaper funding due to an implicit subsidy, it’s still not clear that this subsidy is not passed directly onto the borrowers. The basic bank model is to borrow at LIBOR and lend at LIBOR + credit margin. Credit margin is a function of capital not funding. Changes in funding mostly change lending rates. So there may be a transfer of resources but not necessarily to banks.
Do you have evidence that CDS spreads where “too low”? Given your reluctance to label the NZD as “too high”, I be surprised if you were happy with this line of argument!
Surely none of the shadow banks expected to be bailed out.
Exactly!
That comment might sound a bit weird, but I agree we will largely see an implicit subsidy passed on to borrowers, then we can have a situation where borrowers are willing to take on “too much” risk relative to what they would at the “right” interest rate – and intermediaries may be willing to do “too little” screening.
I would note though that this may be small, or may not exist as an incentive at all – it is meant to represent the “cost” of having a lender of last resort. I would like to see more actual counterfactual research as well on that issue.
“Do you have evidence that CDS spreads where “too low”? Given your
reluctance to label the NZD as “too high”, I be surprised if you were
happy with this line of argument!”
I was definitely loose there – but we could argue they were low relative to the risk of default. As with the idea of the exchange rate it is more useful to ask “why” a price is what it is rather than use the “too low” or “too high” term – so I see your point 🙂
“Surely none of the shadow banks expected to be bailed out.”
That could well have been the case, which really just makes them more vulnerable to bank failure.
I’m enjoying this book at present:
http://www.amazon.com/Misunderstanding-Financial-Crises-Dont-Coming/dp/019992290X/ref=pd_sim_b_1
It discusses the ideas in a clearer and more consistent way than I do 🙂 . I would note that he attacks economists in a way that is a bit unfair, but it is because he is an economist attacking the hybris of economists with respect to determining the “right counterfactual” – economists have a great discipline to analyse these issues, but we didn’t put all the “core elements” in our frame of analysis … we pretty much missed the existence of the shadow banking sector.
Tbf, I only started working as an economist atthe start of 2007, so I’m sure I have an excuse 😛
Sounds like we almost agree!
In terms of too much borrowing, I would argue that it’s the CB’s job to monitor credit growth and adjust rates accordingly. I don’t think the argument is strong that banks are culpable for the excess lending or assessment of systemic risk. It’s possible that there could be “too much” lending even though banks are assessing individual credit worthiness correctly. (I’m assuming that there are regulatory capital requirements.)
CDS spreads encompass a lot more that actuarial loss rates: counterparty risk, mark-to-market risk, basis risk, liquidity etc. It’s quite hard to unbundle the “risk of default” from them.
I’ve read a few of Gorton’s things, he tends to focus on the repo market, is that the case in this book? He’s especially interesting given his involvement in the AIG mess.
My point about the shadow banks was that surely they didn’t behave irresponsibly (supposedly) because of any moral hazard.
I also think that there were some vital elements missing from the models…
Aha, I see. Indeed, when economists bang on constantly about moral hazard and the such we aren’t really saying anything about anyone in the economy behaving “inappropriately” – we are really just saying that the price agents are facing differs from the full “social price” associated with the action.
Everyone involved is being responsible given the prices they face, however there is something unhinging the price from the full value associated with it – and government guarantees are one avenue for this.
Yar, the book is mainly sitting in the repo market – but unlike his other work he is focused on yelling at economists for failing to apply their knowledge and models appropriately. He accuses economists of failing to explain “why” we had the quiet period – which is a bit of an exageration, but at the same time a fair point.
“I also think that there were some vital elements missing from the models…”
The thing that isn’t communicated very well by economists is that there isn’t “a model”, there are a set of partial models that we have to integrate when discussing ideas. The way we do this integration is part of the intellectual hybris that Gorton rails against – it isn’t the economic methods that are broken, it is the practicioners. I think its a good point.
But in case where the problem is too much (systemic) credit growth, even though banks are pricing credit correctly, then the social price is misaligned due to inaction by the CB. Government guarantees are not needed.
The “why” is interesting but does he have an answer? I suspect that the repo market is more of a symptom than a cause. The 0% haircuts were dodgy but it’s the same complacent calm before the storm seen in other markets.
There is, however, what is referred to as the “canonical” model, which doesn’t have credit money or banks at all. Although some extended versions may have these things retro-fitted they are not a natural part. They are always a bit ad hoc. For example Krugman’s patient and impatient agents paper.
“But in case where the problem is too much (systemic) credit growth, even
though banks are pricing credit correctly, then the social price is
misaligned due to inaction by the CB. Government guarantees are not
needed.”
The central bank sets the marginal rate of borrowing for banks in a way consistent with its inflation target – we can still have a systemic issue when borrowers and/or lenders expect, ex-ante, that in the face of default they won’t bear the full loss associated with that default. In this view banks are an intermediary, and the only way they would directly be “blamed” is if we had a model where their incentive to screen loans was comprimised.
The “core” model sitting any any economists head usually treats banks and financial institutions as passive essentially, and I would say that any explanation I make, and any implicit model I use, would continue to use that assumption – banks are intermediaries.
If you mean by intermediaries the commonly-held view that banks take existing deposits and lend them to borrowers then that’s not a good assumption.
In reality banks give the borrower the right to swap an IOU from the borrower for one of their own, and when this happens the borrower spends the bank-issued IOU because it is commonly accepted (as money). The borrower chooses the amount to draw and the timing, the bank sets an upper limit. Most limits are only used about 50% so in theory a bank’s balance sheet could double in size overnight. There is no conceivable way that there are deposits sitting ready for this.
The central bank rate doesn’t just affect the marginal cost, but the cost of the back (existing) book of loans as well. All loans, while drawn, need to be funded (but not before).
An intermediary in the sense that it is a borrower doing the borrowing and a lender doing the lending, and a banks role is to reduce the information asymmetry between the parties.
Fair enough, it leaves out the interesting bits, but obviously I need to prove that the interesting bits are also relevant! I suspect that it’s important that banks, as manufacturers of credit, act as more than mere mediators between borrowers and lenders. Cheers!
The key thing for me is that we now that, if the central bank can help set the price, in a way that is consistent with their inflation target (which in turn represents underlying economic conditions), then banks will lend to all borrowers that will match that expected price. And as intermediaries they have an incentive to do things such as compete for customers and screen loans.
What other roles can they be seen as having, I’m interested 🙂
If borrowers can draw their lines of credit at their discretion (think credit cards for companies not mortgages) but at rates largely set by the CB, then the broad money supply is only very partially under the control of the CB. It is more a function of borrowers’ “animal spirits”, for want of a better term. As long as (CPI) inflation expectations are contained, credit growth is largely unconstrained (by the elements of the standard models, anyway). I think this is very interesting and surely important.
Yes. With inflation expectations anchored and the central bank managing monetary policy in a way consistent with keeping the economy at potential, broad money isn’t necessarily clearly defined. We also need expectations by borrowers to get out asset prices.
In that case, credit growth and asset prices rely heavily on expectations. This is fine, but it is only “welfare relevant” in so far as it does one of two things:
1) Influences movements from potential/inflation (which will then be taken care of by monetary policy). This is the view that asset price inflation can be used to predict future inflation.
2) It risks a systemic event (this is an issue of financial stability).
In this view, we don’t care about any transfers that occur due to bubbles, changes in credit conditions, or the such. We only care about the impact on the broader economy and the risk of a systemic event (which could move us to an “inferiror eqm”).
Now this isn’t to say that the transfers and misallocation of investment aren’t important – just that the central bank doesn’t have a mandate, or necessarily enough ex-ante information, to deal with this type of “second order” issue. Of course, this view may be changing with people discussing active risk weighting management by central banks through their financial stability arms (rather than just setting passive risk weights, they may adjust them based on views of misallocation).
In that case we are essentailly saying we trust the public sector with some issues of the allocation of captial rather than the private sector – we may as well make the assumption clear there!
Still – that rant is sort of how my interpretation of the issues goes 🙂
I think you’re on the right track. The traditional IS/LM “crowding out” stabilisers for preventing over-consumption or -investment aren’t very relevant when there are limited constraints on broad money growth. The risks of unnoticed mal-investment or credit-financed imports causing a systemic event must be very high.
Yar, but they are subject to the “expectations of private agents”, which are both unforecastable and unobservable ex-ante.
Economists tend to shy away from assuming preference shifts – but this isn’t because we don’t think they exist, it is because we like to explain as much as possible about the interrelationship between other things, and figure out what measureable variables we should measure, before hopping over to a “trivial preference shift”. It is the reductionists in us.
This is also the reason why:
1) We need to think about “insurance” for the financial system given its exposure to expectations and the systemic risk that entails. This is the reason why the financial system is more regulated than any other.
2) Ways we can create knowledge about preferences, or make them observable, seems pretty useful. Hence the push for research in behavioural and neuroeconomics – the key thing that is missing from them IMO is still the observability issue.
The modelling of expectations is a big area where we could say economists aren’t provide a full description of the world and risks – however, it is something we think about its just that its a very hard issue. The fact that this needs work doesn’t undercut everything else that is models – it just shows that economics is still being “built”.
I agree that the ultimate credit constraint is private agents’ expectations but what disturbs me is that the current set of models use constraints (fixed money supply or money multiplier) that simply aren’t justifiable. We don’t know the credit growth boundaries and therefore how far asset bubbles can grow before agents’ expectations “correct”.
I’m not sure that modern models actually have monetary aggregates in them a lot of the time – asset prices come in, but there is normally a “no-ponzi game” assumption that helps to anchor what we are guestimating expectations are.
The key thing is, the model we use depends on the question we are asking. Monetary policy models aren’t supposed to discuss issues with the allocation of assets and investment, as those aren’t policy relevant for monetary policy. As a result, the models used don’t touch on that issue.
In other areas of macroeconomics, they will use other models that are meant to answer those specific questions – and there is potentially a role for such things in terms of the financial stability area of central banks.
The history of economies with “growth” has been filled with bubbles, they should be treated as a element of what occurs but their relevance does depend on what you are trying to “do”. Furthermore, our ability to do something about “expectations” is far from clear – we could change the mix of mood enhancing drugs we put in the water perhaps!