Dealing with debt, financial regulation, and the lender of last resort

Previously we’ve talked a lot here about the lender of last resort function of a central bank.  In discussions a trade-off is often discussed, whereby having a lender of last resort can help to prevent financial crises when financial intermediaries are suffering from issues of “illiquidity”, but are not “insolvent” – however, the existence of a LOLR can in turn lead to moral hazard … where financial intermediaries and lenders are willing to take on “too much risk” and charger borrowers “too little”.

Although the discussion of these issues has a long history in economics, Thomas Sargent’s article (REPEC) on the issue – and the solution mentioned – are worth reading.

Now we live in a history dependent world.  Yes, we have had a global banking system willing to take on too much risk – due in a large part to issues of asymmetric information and an implicit solution subsidy of risk.  Yes, in this environment debt accumulated, and the existence of debt and the following credit constraints on people with “useful projects” that they could invest in is having a big negative impact – likely much bigger than any “social boost” that may have existed from the additional marginal projects that took place with easy credit.

We have an environment where people think there is a real risk of the failure of financial intermediaries.  Now if we understood “why” we could ask if there is a solution.  What are some reasons:

  1. There is too much debt.  If we saw this as an issue, we could convert bondholders into equity holders in banks.  After all, isn’t a government bailout really just a transfer from non-depositors to depositors in the bank?
  2. There is no trust.  There was a “capital stock” of trust that was built up between financial institutions, a stock that was destroyed and will have to be rebuilt.  This can be expected to keep hurting the efficiency of financial markets for a long time.
  3. Central banks/governments have lost credibility as lenders of last resort.  I think this is compelling – everyone talks about “too big to fail”, but exactly what that means, exactly what the “insurance” is, and exactly how the “insurance” is paid for are questions that are still up in the air.
For me the key issue is the last one – as the negative impact of the first two is “endogenously determined” by the credibility of the financial system stemming from regulation.  The fact we have a government monopoly for fiat money, and direct management, combined with a LOLR function, ensures this.

Note:  Anyone reading here since the crisis has been in full swing will think I lean strongly on the side of constant bailouts.  However, this is far from the case – I only started writing in favour of them once we already had a crisis where this needed to be the case.  In 2007, once the crisis had begun but a few months before Bear Sterns, I was writing like some sort of “purging our evils” Austrian economist.  After Lehman Brother’s collapse, even a half pint analyst like myself saw the issue of a lack of trust even when it was unclear whether there was a bailout or not 😛 (it was interesting reading these old posts, the lack of clarity around what was going on was even worse than I remember!).   By the time what happened was clear, we had switched our tune in favour of bailouts for that period of time – and even I found the ECB’s call to attack moral hazard in the middle of the crisis strange, even though I saw that as a big issue.

This was seen as a major issue coming into the crisis, it was viewed as a key issue during the crisis (I would argue that the efficiency of NGDP targeting would still depend on a banking system without bank runs), and now during these later stages of the crisis it is an issue receiving a lot of research and taken into account for regulation.  Mainstream economics has the tools to understand what has happened, and hopefully policies can be developed that ensure that the next economic crisis is something completely different.

  • HJC

    Matt, the mainstream mostly consists of models that don’t include money, let alone distinguish between inside and outside money. How can you be so confident in its toolkit?

    •  Hi HJC,

      The mainstream consists of models that try to illustrate what happens as the result of some set of primatives – namely how the actions of a set of individuals and institutions lead to a set of outcomes.  In that way they don’t tell us directly what will happen in reality – the “mainstream” as an entity gives us a set of narratives to understand what is going on.

      Given these narratives we can then appeal to data, and try to build a full understanding of what is happening – and only then can we attempt to describe “what” is going on and “what” could be done.

      Now money is a means of exchange, it either isn’t a distinct primative in any narrative that we need to paint – or it is, and it represents the payoff associated with whatever is going on in the model.  We can’t build a full specification of reality, but what we can do is to try to understand parts, through a series of different models, and then to try and use what is observed through data to build a fuller understanding.

      My confidence in the mainstream narrative that comes from this is because the mainstream narrative combined with policy failures that were not predicted provide a powerful way for understanding what has happened.  Given that we have a clear idea of where improvements can be made – clarifying the LOLR function, making financial institutions who are “too big to fail” pay for that insurance, etc etc.

      • HJC

        Hi Matt
        It sounds like you are putting the whole GFC down to the failure of central banks to perform their LOLR responsibilties! What about the role of the large increase of credit/inside money and asset prices? Bank failures and the need for LOLR came after these started to turn.

        •  Hi,

          I am indeed putting the majority of the GFC, and the following European soverign debt crisis, down to the failure of clarity and transparency regarding the LOLR function.

          Of course this is not the sole issue, and indeed there were other policy and market failures in the lead up to the crisis.  And although they should be looked into and policy improved, they were nothing like the failure of this role.

          In terms of the run up in asset prices and debt there are two clear things to keep in mind here:

          1)  There is a transfer cost, and a cost in terms of the distribution of debt between those with projects and others willingness to take on risk – given these issues such an event is a concern.  But the solution would be to ensure transparency in financial markets here, or worst comes to worst to let creditors know that they would effectively have to take on direct risk when things go wrong.

          2)  We can very much lay the blame for the accumulation of debt, rising asset prices, and all that jazz on a poor conception of the lender of last resort function of central banks!  Why?  Markets expected a lender of last resort, they expected insurance, but they didn’t have to pay for it – that along with issues of asymmetric information led to excessive risk taking, increasing demand for credit.

          Now again, this isn’t the whole story – but this narrative provides a compelling way of understanding the time leading up to, and the crisis.  IMO, we also need to take into account factors such as population demographics and financial market interventions in the developing world.

          However, although these things could well have caused a recession, weak growth, and some unemployment – the sheer failure of policy turned a potential slowdown into a full blown crisis.

          • HJC

            Do you view LOLR activities as any different from normal CB activities. To maintain their interest rate policy target they perform liquidity provision activities all the time, only on a smaller scale. Increased demand for credit explained by interbank settlement functions? Surely, that’s the tail wagging the dog. Remember, LOLR is supposed to be against good assets and not provided to insolvent banks.

            •  Heya,

              A LOLR function, even though it is only called upon rarely, is always active as part of a central banks activity – I see it as fundamentally seperate from monetary policy, but that is a different issue.

              I think the LOLR function is something that need to be transparent, and viewed as part of underlying financial regulation – the fact that the central bank is insuring against failure does imply there is a moral hazard problem, which needs to be dealt with in some sense.

              That would be where ideas such as risk weighted capital requirements, taxes on banks, etc all come in.

              We could exclude a LOLR function – but then we need to both SAY we won’t bail out ex-post (credibly) and ensure that we have sufficient regulation in place to deal with externalities (our systemic risk issues).  Banking regulation and an honest view of the LOLR function go hand in hand – and I think that has been widely accepted as the pre-crisis issue.

              In these ways the behaviour in asset markets and accumulation of debt were symptomatic of these failures – failures that economists have written heavily on in the past, making it a bit more surprising.

              • HJC

                If a central bank requires clearing banks to have balances (positive or negative) at the central bank (so it can conduct monetary policy) and also wants to maintain financial stability then LOLR must be available. Risk weights and bail outs are an entirely separate issue.

                • Ahhh fair call – I was using the term solely in terms of emergency lending, which is a bit inappropriate.  In so far as they are acting as a clearing house, and insofar as they are setting interest rates in a manner consistent with their inflation target I have no issue with this role.

                  However, the LOLR function can go further into emergency lending – and as you say it is a separate issue from the issue of monetary policy.  Furthermore, it was in the sense of emergency lending where they were a LOLR where there was policy failure – both in terms of it being taken into account in regulatory settings (given moral hazard), and given the ad hoc way it was dealt with during the crisis.

                • HJC

                  Sure, the LOLR is emergency lending but it is still only for interbank clearing and against good bank assets so it should not create a moral hazard and lead to risky lending.
                  Currently banks are well capitialised in the US and supported with liquidity in Europe. How is clarity of LOLR affecting the economic condition now?

                • Have replied above, as these comments are getting tiny

  • Hi HJC,

    I’m starting our comment thread up again here, as the comments keep getting smaller 😛

    “Sure, the LOLR is emergency lending but it is still only for interbank clearing and against good bank assets so it should not create a moral hazard and lead to risky lending.
    Currently banks are well capitialised in the US and supported with liquidity in Europe. How is clarity of LOLR affecting the economic condition now?”
    The full LOLR function, if taken to its most extreme (full deposit insurance) will lead to excessive risk taking by both intermediaries and creditors – as in the case of failure someone else pays, while in the case of success the full return is taken privately.  Just like in the Kareken and Wallace model mentioned by Sargent.

    Effectively, the central bank can’t commit to only backing “good” asset during a bank run – they recognise there are multiple eqm (the Diamond and Dybvig model mentioned by Sargent), and so recognising this depositors will act as if their savings are risk free, and the incentive for intermediaries to screen will be impaired as they don’t get the full “social” benefit of these actions.

    The European situation stems from the fact that:

    a)  A lot of financial transactions are currently denominated in euros, and through the Euro zone.b)  The disjoint between the existence of a supranational central bank and individual governments (some of which are highly indebted) indicates that there is a sizable chance of the failure of the euro, and of the potentially bankruptcy/default of national governments
    c)  A clear LOLR function for the ECB would ensure the existence of the euro and clarify the outcome regarding government default in the region – it tells us who the losers are.Once that uncertainty is gone, conditions will improve – after all this time it’ll still be ugly, but it will get better.It is the fact that borrowers who have potentially useful projects are credit constrained, and that lenders are unwilling to lend given uncertainty, that is directly retarding investment in Europe.  However, the uncertainty isn’t just around government debt – but around the value of that debt through the sustainability of the euro as a currency.  Given these points, we should be able to evaluate policy changes in Europe to figure out if things are improving, or getting worse – the ECB’s commitment to discretionary but unlimited purchases improved outcomes by making it more likely the euro would hold together for example.

    • HJC

      For the good assets idea I was working of Bagehot’s dictum. I will have a look at the paper you mention. If LOLR includes deposit insurance (I thought this would be a government, not central bank, function) then yes, obviously there’s moral hazard.
      The Eurozone is a special case because Target2 ensures that there will not be a balance of payments crisis as per a normal fixed exchange rate regime. The BOP problem will play out in the Euro central bank balance sheets.

      •  Indeed, Bagehot initial framing was good – but the type of intervention the market expects has an impact that he didn’t really cover from what I’ve read.

        And agreed regarding deposit insurance and the such being a central govt issue rather than a central bank one – however, it is one of those cases where the bank is better positioned to intervene, and where once again rule based policy could improve outcomes and transparency.  Central banks are really just a branch of government – just hopefully a relatively independent one.

        Agreed regarding the Eurozone to a large degree.  However, it isn’t just an issue of transfers on the balance sheet – there is also the view that there is an underlying run on otherwise good government bonds.  In so far as that is the case, having the ECB take it on is a good thing.

        • HJC

          I think that deposit insurance is more for the case of an insolvent bank than an illiquid one, so it could still be handled by Treasury (i.e. government but not central bank).
          If you check out Wynne Godley’s work, unconstrained (due to Target2) BOP problems translate into government deficits, they are not “good”. There is no stopping the the Euro break up by LOLR alone.

          •  If it was solely an issue of “ex-ante” insolvency then deposit insurance is a transfer – and probably an unfair one.  The only reason we support the idea of deposit insurance is because of the fear that firm facing illiquidity may be seen as insolvent.

            During the GFC, equity purchases were only really fair because the firms that were ex-post insolvent weren’t ex-ante – outside of that, there is a much stronger view about letting them fail. 

            I’m not saying the imbalance is good – I’m saying there are two elements to the crisis.  The imbalance between nations (the BOP issue), and the run on government debt (which isn’t based on long-term fundamentals).  The LOLR function solves the second – in terms of the first it is, again, just a potential unfair transfer that ignores the underlying imbalances.

            • HJC

              I suppose that my argument is that deposit insurance is probably not a necessary part of LOLR. As per Bagehot only banks that are solvent in “normal” time (I think this aligns with your ex ante) should get it. Otherwise bond holders should become equity holders etc.
              Sorry, I wasn’t clear about “good”. With uncompetitive economies in a fixed currency set up like the Euro, the deficits cannot be stopped. That was Godley’s insight.

              •  Isn’t the key thing here that banks that are “probably” solvent ex-ante could still experience runs – but since you do not know this you have to give said insurance more generally, implying that banks will take on more risk.

                It is an issue of the belief of what the central bank/government would do, combined with asymmetric information about bank quality – which is why transparency and a “time consistent” path for bailouts (or not) needs to be established.

                I was just trying to be more clear regarding what I meant by good – I don’t think we disagree 🙂  The main thing once again is whether we believe there is a run on government debt that is otherwise sustainable (even if still the result of an internal “imbalance”).

                • HJC

                  (Getting small again!)
                  You’re right, faced with uncertainty about bank solvency there could easilty be a run on an otherwise-solvent bank (ex ante or ex post). But perhaps with LOLR support against good assets this run can be allowed to run its course. The regulators can look at the bank internals and remove the information asymmetry, thus sending a solvency signal to deposit holders.
                  As for PIIGS debt, my argument is that it’s not sustainable due to the way the currency union is set up.

                •  I’ll move this to a brand new comment 🙂

  • Hey again HJC,

    I don’t think there is too much we really disagree on – its really just my interest in constantly talking that makes these threads go on as long as they do 😉

    With the bank run issue, the fact that we don’t know which assets are “good” or “bad” at the time of the run – combined with the fact that the expectation of the support of a LOLR implies a socialisation of some of the “risk” associated with assets that “go bad” – creates much of the issue here.  As soon as we are backstopping creditors who indirectly put their money into “bad assets” due to the inability to split them in real time we get a moral hazard issue.  My only real point would be that we need to make the conditions of support clearer, and accept that there will be occassions where this happens.  One the US did this, the GFC effectively ran its course – if it wasn’t for Europe the crisis would have been done by June 2009.

    Many policy makers in Europe agree that the debt is unsustainable, and also believe the PIIGS should have to deal with the cost of it.  If that is the case, let them default and/or leave the currency union – the current half/half solutions we have merely cause uncertainty.

    • HJC

      Yes, I agree that we mostly agree. I’m on board on the LOLR as cause of credit bubble etc, but it’s a really interesting idea. The main area of divergence is about the fundamental cause of the Euro crisis, but that’s a separate issue for another thread. Thanks

      • HJC

        Sorry, should read “not totally on board with the LOLR…”, whoops!

      • It also depends on what we mean by cause of the crisis – I agree that the structural issues were the underlying “cause”, and would have lead to hardship in Europe.  I just think a lot of the extra financial market issues which the rest of the world is stuck with is due to debates around who burdens the debt given lack of clarity around the LOLR.

        I’m sure we’ll discuss things another time – we’ve gone and filled up this comment thread nicely 😉 .  Its always useful for me to hear all these things at least 🙂

  • Pingback: Independence and credibility: The crisis and central banks | The Dismal Science()

  • Gerard saliot

    Nice information .
    Thanks for sharing.