Some further thoughts on expansionary sovereign creditworthiness shocks

Just a quick follow up to my post questioning the result in Paul Krugman’s model that a sovereign creditworthiness shock is necessarily expansionary in an economy with an independent monetary policy and flexible exchange rates.   I was a little surprised by the ease at which some commentators suggested that questioning this result is absurd.   My suggested extension to Paul’s model is really just the addition of a “balance sheet channel,” but where the balance sheets in question are those of the banks.  The argument for such a channel has been best made by Ben Bernanke.   See, for example, this speech.   An extract:

[T]he cost and availability of nondeposit funds for any given bank will depend on the perceived creditworthiness of the institution.  Thus, the concerns of holders of uninsured bank liabilities about bank credit quality generate an external finance premium for banks that is similar to that faced by other borrowers.  The external finance premium paid by banks is presumably reflected in turn in the cost and availability of funds to bank-dependent borrowers.  Importantly, this way of casting the bank-lending channel unifies the financial accelerator and credit channel concepts, as the central mechanism of both is seen to be the external finance premium and its relationship to borrowers’ balance sheets.  The only difference is that the financial accelerator focuses on the ultimate borrowers–firms and households– whereas financial intermediaries are the relevant borrowers in the theory of the credit channel.

The question then is how a sovereign creditworthiness shock could affect the perceived creditworthiness of domestic banks.   I focused in the previous post on the potential role of the sovereign as a capital backstop for the banks.  But other channels are also potentially present.  The credibility of liability guarantees will depend on the creditworthiness of the sovereign.   Also, to the extent that the banks are holding government bonds, capital losses will erode the banks’ capital. 

There is the question as to whether conventional or unconventional policy could offset contractionary effects of a sovereign creditworthiness shock.   Vasco Cúrdia Michael Woodford, for example, have examined how monetary policy rules should be altered where “financial frictions” are present, though their focus is on borrower rather than bank balance sheets (see here).   But the zero lower bound will presumably limit what even an extended monetary policy rule can achieve.  Maybe QE could help offset the contractionary effects of the induced shocks to bank balance sheets.  (This is separate from the potential role of QE in avoiding the shock to sovereign creditworthiness in the first place.)   But Paul himself has questioned the potency of QE as a stimulus tool.

Update:  Krugman responds to Rogoff’s response here.   (I am spending too much time on this, but it is fascinating.)

I just want to pick up on one point.   Paul asks:

On the economic question at hand, Rogoff seems to be playing bait and switch. Wren-Lewis and I both asked why, exactly, a country like Britain should fear a Greek-style loss of confidence. Rogoff, however, spends most of his piece arguing that Britain would, in fact, be hurt by a euro collapse. Of course it would. So? This doesn’t even seem relevant to Rogoff’s own argument for austerity policies: British austerity might be “insurance” against loss of confidence, but it makes no difference either way to the probability of disaster in Europe. What’s going on here?

Both Paul and Ken agree that a euro collapse would be directly bad for the UK economy.   They clearly disagree on the likely implications for UK creditworthiness.   But there is a further disagreement related to the post above.   If it did occur, Paul believes that the creditworthiness shock would itself be expansionary, and so help offset the direct contractionary impulse of the euro zone shock.   I think Ken believes it would be compounding.   The differences here can affect the perceived value of “insurance.”



Paul Krugman on Expansionary Creditworthiness Shocks

An interesting debate has been taking place following Ken Rogoff’s recent FT piece on the risks of UK sovereign default.   [Update: Rogoff responds to Krugman here.] Simon Wren-Lewis has argued that the risk is overblown given that the UK has can conduct an independent monetary policy (see here).   In extremis, the Bank of England could simply print money to cover debts as they come due.   Also, QE could be used to lower the cost of funding to the government.   (See also the argument of Giancarlo Corsetti and Luca Dedola how co-ordination between the central bank and the treasury could ensure the creditworthiness of the government even without the need for the central bank to act as an unlimited lender of last resort.) 

A more provocative response to Rogoff comes from Paul Krugman.   Using a stripped down New Keynesian open economy model (pdf) he argues that the government’s loss of creditworthiness could actually be expansionary in a liquidity trap.   I hasten to add that this analysis is not meant to apply to a country like Ireland, but only a country with its own independent monetary policy and a flexible exchange rate.  

On the face of it, this looks too good to be true.   But in the topsy-turvey world of the liquidity trap strange things do seem to happen.    In this case, though, I do think it is too good to be true. 

The key to the expansionary result in Paul’s model is what happens to the exchange rate.   Using a standard interest rate parity condition with a risk premium, an increase in the premium leads to a downward jump in the exchange rate.   (Essentially, the exchange rate must fall sufficiently to create expectations of future appreciation that are sufficient to maintain the interest rate parity condition.)   With nominal rigidities, the nominal exchange rate depreciation is also a real exchange rate depreciation, and hence the expansionary force. 

So what then is left out?   I think one problem comes with the assumption that that the central bank can still set the domestic interest rate following its usual monetary policy rule (constrained by the zero lower bound).   What the stripped down model leaves out is that central bank is only directly affecting the overnight lending rate between banks, and also possibly future expectations of this rate.  

This is where the experience of Ireland and other peripheral euro zone economies is relevant.   Stressed banks cannot access funding at the central bank’s policy rate.   This is what has led to the fragmentation of bank funding markets across the euro zone.   The weak creditworthiness of the sovereign is one of the factors affecting the creditworthiness of the banks, as the sovereign remains the primary capital backstop to the banking system.   And the high funding costs of the banks is keeping up lending rates to the real economy.  

I don’t see a reason why a similar dynamic could not play out in the UK.   Even if UK banks are in better shape than peripheral-country banks, a creditworthiness shock to the government would still be likely to affect their funding costs and thus push up borrowing rates for households and firms.  How the balance between the expansionary force from the exchange rate and the contractionary force from credit risk would play out I don’t know. 

Just to reiterate, this is not an argument that the UK faces a serious risk to its creditworthiness.   I find Simon Wren-Lewis largely persuasive on that – even if I can’t help niggling doubts.   But if a creditworthiness shock did occur, it seems unlikely that all would be for the best in the real economy.  

The Government’s Balance Sheet after the Crisis: A Comprehensive Perspective

My colleagues at the Fiscal Council, Sebastian Barnes and Diarmaid Smyth, have produced a new report detailing a comprehensive approach to the State’s balance sheet.  The report is available here. 

I just want to underline here the value of the type of comprehensive approach developed by Sebastian and Diarmaid for policy analysis.   Policy analysis can often be distorted by taking an overly narrow view of the government accounts, and in particular focusing only on the General Government accounts.  Some examples:

In banking-related discussions, focus is often on the need for new State-provided capital and the consequent implications for Gross Government Debt.  But even if the new round of stress tests do not reveal the need for new capital, any newly revealed losses are real losses in terms of the net worth of the State.  On the other hand, additional capital required to meet more ambitious capital adequacy ratios need not lead to a reduction in net worth. 

In the discussion of “stimulus programmes”, attention is often on forms of financing that run down State financial assets or use “off-balance sheet” forms of financing.   The latter are typically off-balance sheet only to the extent that the liabilities – actual and contingent – are outside of the General Government accounts.   They are very much on a more comprehensively measured balance sheet for the State.

Finally, in the discussion of the benefits from the promissory notes transactions, most attention is given to the impact on the General Government deficit, including the roughly €1 billion reduction in the deficit for 2014 and 2015.  This particular reduction is largely an accounting fiction that follows from the limitations of General Government accounting.   It is the result of the high interest rate that was paid on PNs.  Sebastian and Diarmaid’s analysis shows the irrelevance of this interest rate from a comprehensive balance-sheet perspective (see also this recent post by Seamus Coffey).   However, they show that the transactions are likely to generate a significant net present value gain when account is taken of the effects on all the relevant entities on the comprehensive State balance sheet, although the size of that gain is uncertain and depends on such factors as the choice of discount rate and the future evolution of the risk premium on Irish debt. 

More on self-defeating adjustment

Ashoka Mody offers another thoughtful perspective on Ireland’s crisis resolution effort on today’s This Week programme on RTE radio (audio here; scroll down to last item).    Although he covers a good deal of ground, much of which I agree with, his main recommendation is that Ireland should consider scaling back its fiscal adjustment effort.   As I note in a piece for the Business section of today’s Sunday Independent, reasonable people can disagree on the optimal speed of the adjustment given the tradeoffs involved.   But Ashoka raises the old issue of the adjustment effort being self defeating.   At the risk of repetition of past posts, I cannot see how the evidence in the Irish case bears this out.   There is, of course, wide agreement that fiscal adjustment lowers growth.  But it is a significant step from here to a claim that the adjustment is self defeating on its own terms of improving the fiscal situation.   I think it is worthwhile to continue to scrutinise this claim.   

The self-defeating hypothesis comes in a number of forms depending on the focal outcome variable.   For convenience I label the main forms as: the strong form (underlying primary deficit); the semi-strong form (debt to GDP ratio); and the weak form (creditworthiness).   Ashoka’s main focus is on the semi-strong version, but it is worthwhile to briefly review the case for each.  

Ireland’s underlying General Government primary deficit has fallen from a peak of 9.3 percent of GDP in 2009 to a projected 2.5 percent this year.   This occurred despite the massive non-austerity related headwinds the economy faced.   All else equal, if there had been no adjustment from 2009, simulations show the underlying primary deficit would have reached 14.5 percent of GDP this year.   The underlying actual deficit and debt to GDP ratio would have been 20.5 and 160 percent of GDP respectively.    Of course, something would have given before these ratios were reached, but it is useful to consider the crude counterfactual in assessing what has been achieved.  (These simulations assume a reduced-form deficit multiplier of 0.5 and an automatic stabiliser coefficient of 0.4.  The underlying model being used is sketched here. ) 

Turning to the semi-strong form (Ashoka’s main focus in the interview), it is well known that for multipliers around unity fiscal adjustment can lead to a short-run rise in the debt to GDP ratio.   The reason is that the negative effect on the denominator can swamp any positive effect on the numerator.   But this static analysis can give a very misleading picture of the impact of the adjustment on the path of the debt to GDP ratio.  

This is most easily seen using the classic equation for the evolution of the debt to GDP ratio: ∆d = (i – g)d-1 + pd, where d is the debt to GDP ratio, i is the interest rate, g is the growth rate, and pd the primary deficit as a share of GDP.   The debt to GDP ratio may rise initially due to the adverse effect on g.   But the effect of this year’s adjustment on g should be temporary while the impact on the primary deficit should be permanent.   The permanent effect should quickly swamp the temporary effect, leading to an improvement in the debt to GDP profile as a result of this year’s adjustment.   (One thing that could overturn this result is that today’s adjustment leaves a long negative shadow on the level of GDP (what is known as hysteresis), but that would again require that today’s adjustment causes the primary deficit to rise rather than fall.) 

Simulations for Ireland show that an additional €1 billion of adjustment in 2014 would only lead to a higher debt to GDP ratio in 2015 for values of the reduced-form deficit multiplier greater than 2.3 – above any reasonable estimate for the highly open Irish economy.   Reduced austerity might well improve growth performance, but it seems highly unlikely that it would improve Ireland’s debt dynamics as Ashoka claims. 

Turning finally to the weak form, it is possible the direct adverse growth effects from fiscal adjustment on investor confidence could swamp any positive effect through the deficit and debt.   This might happen, for example, if investors worry that the deeper recession would undermine political support for the adjustment effort.   It is thus possible that despite the positive fiscal impacts, creditworthiness (as measured by secondary market bond spreads) might deteriorate.  This is ultimately an empirical question.   But the fall in Ireland’s 10-year bond yield from 14 percent in mid 2011 to below 4 percent today hardly suggests that Ireland’s fiscal adjustment effort has been self-defeating on even this weak definition.  

Response to Shay Begorrah

With advance apologies for a too self-referential post, I think it might be useful to put my response to commenter Shay Begorrah from Tuesday’s thread on the front page.   Shay clearly sees me as an unabashed fiscal hawk, where more deficit reduction (and more expenditure-focused deficit reduction) is always better.   As this is not at all my view, and given my official role through the fiscal council, it might be worthwhile to explain my actual view a bit. 

During 2009 and the first half of 2010, the main thrust of my fiscal policy posts on IrishEconomy was that the fiscal adjustment was too frontloaded (see, e.g., here).    At the time I assumed that Ireland’s credtiworthiness was reasonably robust.   I underestimated how fragile Ireland’s creditworthiness  was in the context of monetary union with an underdeveloped lender of last resort to governments.   My basic approach to fiscal policy is Keynesian.   But since the middle of 2010 it has been clear to me that fiscal policy must steer a difficult course between supporting demand and sustaining the borrowing capacity of the State.   Not least, the latter is essential to ensure that the fiscal adjustment can be phased in any sort of reasonable way, and thus to sustain the essential protections and services of the welfare state. 

The idea that there is a trade off between demand and creditworthiness is challenged from both right and left.   From the right, the idea of an expansionary fiscal contraction is invoked: discretionary deficit reduction would then increase demand, further enhancing the deficit reduction.   I do not read the available evidence as supporting this contention (see, e.g., the important work from the IMF).   From the left, the idea of self- defeating austerity is invoked.   This comes in different forms: discretionary deficit reduction does not actually reduce the deficit; discretionary deficit reduction does not reduce debt to GDP ratio; and discretionary deficit reduction does not improve creditworthiness.    I do not believe that such self-defeating effects are borne out by the Irish experience (see Box C, p. 46 here).  

Another thrust (obsession?) of my more recent posts has been the critical importance to more robust creditworthiness of strengthening the crisis resolution mechanisms – and more narrowly the fiscal lender of last resort function – for the euro zone.   The political economy of such developments must been seen in the context of weak solidarity between what is an association of nation states, where there is a reluctance to risk transfers and a fear of moral hazard.  While I believe the aggregate fiscal stance of the euro zone has been significantly too tight from an optimal fiscal policy perspective, the strengthening of fiscal rules must also be viewed in terms of the political economy of strengthening the LOLR function.   The fiscal compact, for example, must be seen in this light. 

Finally, on the relative measures of expenditure- and tax-based adjustments, my prior belief around 2009 based on the literature was that expenditure-based adjustments were less contractionary than the tax-based alternative.  (Of course, in deciding on the mix of adjustments, other factors besides the macro effects – such as fairness – are important.)  From the more recent evidence, I have become much less convinced of the macroeconomic superiority of expenditure-based adjustments.   This issue was discussed in the first fiscal council report in October 2011 (see Box 4.1, p. 36, here; a more general review of the multiplier literature is given in Appendix E, p. 93, here).   The inconclusiveness of the evidence is the reason that the council has focused on an overall deficit multiplier in our analysis.