Winner of the New Statesman SPERI Prize in Political Economy 2016

Thursday 29 November 2012

Is it a sin for the central bank to help reduce debt?

There is a great deal of discussion about the appropriate goals for monetary policy. Should it just be targeting inflation, or should central banks also include the output gap in their objective function? Perhaps they should target nominal GDP. However, hardly anyone would dream of suggesting that monetary policy should help reduce government debt, by lowering interest rates or raising inflation. It is just accepted that we should only use costly (in welfare terms) increases in taxes, or costly cuts in government spending, to achieve debt reduction.

What I have called the consensus assignment for economies with their own currency is that the central bank stabilises demand and inflation, and the government through fiscal policy manages government debt. The first part of that assignment does not work at the zero lower bound (ZLB), although some economists find it difficult to admit this. But equally there are circumstances where the second part of the assignment is far from optimal, and where it is useful to have the central bank help reduced debt. Circumstances like when debt is much too high.

If we just think about the macroeconomics, where a benevolent policy maker operates both monetary and fiscal policy, then the reason is straightforward and pretty obvious. When debt is high, reducing real interest rates is a pretty effective means of getting debt down. If government debt is about the same size as annual GDP, then a reduction in the real interest rate on government debt of 1% is equivalent to raising taxes by 1% of GDP. If debt is only half the size of GDP, the same reduction in real interest rates is only equivalent to increasing taxes by half as much. The larger the stock of debt, the more effective interest rates are compared to fiscal instruments in reducing debt.

Will reducing interest rates to cut debt not allow inflation to rise? Of course we are compromising our control of inflation. Fiscal instruments can be used to reduce this cost: if we cut government spending as well as cutting interest rates, the net effect on inflation may be small. But the key point is that inflation is not the only thing that matters. Raising taxes or cutting spending to reduce debt is costly too. So it may be better, in terms of social welfare, to let inflation rise for a time to get debt down than to have taxes rise to do the same.

At this point you are probably thinking isn’t the whole point about the consensus assignment that monetary policy is much better at controlling inflation than fiscal policy? Yes it is in our benchmark models, but only when debt is not a problem. If we had some fairly painless way to control debt (i.e. lump sum taxes),[1] then outwith the ZLB, monetary policy is indeed the best way to control inflation.[2] But suppose instead that debt is much too high, and we need to bring it down (and there are no lump sum taxes). Suppose, as suggested above, that monetary policy is also the best way of getting debt down. In that case we are in a different world where comparative advantage applies.

If you are still having difficulties with this - and because the consensus assignment has become such a consensus you might well be - imagine if debt was ten times GDP. In that case it is so much easier to get debt down by reducing real interest rates, it would be crazy not to. Any advantage monetary policy had in controlling inflation would pale into insignificance. If monetary policy is only a bit better at controlling inflation and much better at controlling debt, the latter should have priority.

So if I’ve convinced you this is possible, what about in more realistic situations. Here I can say two things. First, in quite a lot of the DSGE work I have done with my co-author Campbell Leith, we found situations in which monetary policy was useful in reducing debt when debt was high. Here high meant something like 50% of GDP, which is pretty standard today. Second, in the real world there have been occasions when monetary policy was linked quite closely with getting debt down, such as after the second world war. (See, for example, Charles Goodhart here.)

The reason why the idea of adding debt reduction to monetary policy’s tasks seems so taboo today is that we are concerned about the non-benevolent government. A government that spends too much and taxes too little can create ‘fiscal dominance’, and an inflation rate that is permanently too high. In extreme cases it can lead to hyperinflation. As is so often the case with fiscal policy (like the specification of fiscal rules), it is the non-benevolent policy maker which determines how we think about things.

To some it will seem obvious that this should be so. Politicians can be trusted to do the wrong thing, particularly if it is in their interests. There are plenty of examples where this has happened in the past. However, many of us have been arguing that governments today seem if anything too eager to use fiscal policy to get debt down. We seem to have a twisted sort of fiscal dominance, where governments are reducing debt too rapidly, the economy is suffering as a result, yet monetary policy cannot effectively counter because we are at the ZLB. It therefore seems a little odd to be arguing that monetary policy cannot help bring debt down, because that would let fiscal policy makers off the hook. In countries like the UK, USA or the Netherlands, there is no hook, but policymakers are taking unpopular measures to get debt down nevertheless. Would they really do a complete about face if monetary policy lent them a hand?

Now with short interest rates at the ZLB, and in the US and UK with Quantitative Easing directly reducing interest rates on government debt, it would appear as if there is no conflict at the moment between keeping inflation on target and reducing real interest rates. (Although this does not stop some getting very worked up when the central bank helps to reduce current borrowing.) However there are two parts to real interest rates. To the extent that monetary policy could be more expansionary today, but at the cost of higher inflation, it could be helping to reduce debt. If it is not, and the government is benevolent, monetary policy is probably behaving sub-optimally.

[1] These are taxes which have no impact on incentives, and so do not reduce output when raised. Although Miles Kimball has written what seems like a lament to their absence, when Mrs Thatcher tried to introduce them in the UK it led to her downfall.
[2] This is also true if we are quite happy to let debt permanently rise and fall in response to shocks, as Eser et al show.


  1. There is nothing wrong with a CB helping the country it serves by using it moneyprinting capacity to generate income or stabilise the economy etc.
    What I am however missing is a proper pro con analysis (most of the time). And another thing is risk-assesment/-management. In combination with a lot of interpretation open effectively to personal preferences as well as political preferences iso economical reasons.

    Even in rather extreme circumstances like now one simply would have to see what are the pro and cons of different policy options, like writing off debt for instance.

    However it clearly looks that CBs have not a peoper oversight of things (aka they know only partly what the consequences are of the things they are doing). It simply looked they kept interstrates ingeneral substantially too low (which partly caused today's problems).

    Another point is that giving a country as a policy option CB effectively writing off state-debt makes it very vulnerable for political influence. What I mean is that certain politicians mainly from the left and/or populist part will simply take this one for political mainly short term reasons. And hardly look at the long term consequences.

    Another problem is what we see in Spain and Italy now. Countries wait with necessary structural reforms if the pressure is off. Letting say Spain off the hook by 'printing' it out means that structural measures most likley will happen later and the shock necessasy to make that happen likely gets bigger in that process.

    Again another is that effectively we would be playing with the blood of an economy. If things go wrong the troubles cannot really be overseen.
    With as added problem that the facts might not really change but the way markets look at them do. In other words the borderline can basically move overnight while the policies to adjust to that could take years.

    Might largely be solved by simply having rules publicised. How to act in what sort of a situation. More as guidelines of course. Markets can adjust to that in an early stage and if not acceptable you can reverse things in the planning phase and not in the operational one. Which likely much less fall out.

    But will be difficult. CBs basically missed this crisis. Some parts of that missing are excusable but other parts are simply not the RE bubble for instance and how that would play out in larger lines, simply looked pretty obvious but still as missed (imho more that they saw what was happening but didnot react in anyway properly).

    1. Spain and Italy can NOT print money - they are part of the Euro.

    2. You missed my point. My point is that structural change is postponed if given the opportunity. For which it hardly matters if a country is inside or outside the EZ.

  2. You seem only to be considering government debt. Is private debt not alsso relevant?

  3. This is a case where you need to make a distinction between policy actions and policy rules. If we knew in advance that the central bank was following a policy rule in which it would create higher inflation when debt was high, and lower inflation when debt was low, would that work? Because the knowledge that the central bank would act like this would raise nominal interest rates when the debt was increasing, and lower nominal interest rates when debt was decreasing, which would exacerbate the fluctuations in the debt.

    My guess is that it depends on the term structure. With long-term debt, using monetary policy to smooth debt might work. With short term debt, it probably wouldn't.

    NGDP targeting also has a debt/GDP smoothing effect, in a world where there are real shocks to RGDP.

  4. I think another possibility needs to be considered: what if a situation emerges in which the mandate of the central bank necessitates it to take treasury paper on its balance sheet forever, without ever unloading it; causing interest payments on this paper to be recycled to the treasury forver such that the paper is effectively already ripped up? What if Japan and the USA are currently in such a situation as a result of unproductive private debt buildup that confers a permanent disinflationary effect on the economy?

  5. This has been an endless topic of debate in Latin America and, in particular, Brazil where real rates have been extremely high by global standards. In Latin America, its pretty safe to assume that, in fiscal terms, governments can be trusted to be anything BUT benevolent. But, even if governments were enlightened and benevolent, tampering with monetary policy to achieve debt reduction would almost surely lead to greater price indexation (already widespread in the region), increase in risk premia (market would deman higher nominal yields to hold bonds, or migrate to inflation linked bonds, creating more price indexation) and loss of confidence in the currency

  6. Take the 3 countries mentioned.
    UK/BoE basically did that inflation 4% and low interest. Lowered real wages. Got some of the hot air out of the RE market via inflation plus kept mortgages affordable. Worked pretty well as far as I can see.

    Holland. Euro so cannot have an independent policy, so the issue is purely theoretical. Policywise they are very German. High inflation is very unpopular, not as much as in Germany but close.
    2 other points.
    Pensions low interest also caused pensionreserves to drop considerably. Leading to lowering pensions and a lot of discussion thereon. Combined with inflation worries (now close to 3%, much higher than the usual 2% and in the middle of a crisis allthough partly caused by a VAT hike) this would make even lower interest politically a no go. And the pensioncapital destroyed was probably higher than the extra growth would have been anyway.
    Politics have heavily determined the outcome of policy decisions. They increased the price drop of RRE by partly abolishing the tax deduction for mortgages (and destroyed 50 Bn RE-value (in 1 year) and brought many people under water for a few Bn extra tax revenue, will be something like 2 Bn is my estimate. From an economic pov completely idiotic however politically apparently necessary. Similar with some other measures. The only one I could understand is that some air had to be let out of the entitlements and they used the crisis for that. Pretty clear politics in Holland are simply too stupid or too focussed on short term political gain to work with a scheme like you propose.

    US is a different case again because of the reserve status. Also there pensions look to be a huge problem.

  7. Simon,

    Are you asking if it's OK for the CB to move the real rate away from the (socially optimal!) natural rate? Or are you saying that the socially optimal rate (i.e. the natural rate) declines with increasing government debt? Or is this a disequilibrium proposal in which the government gets to pull off a one-off price level hike in order to bring debt levels back into line?

    1. It is about adjustment to steady state, and policy that creates either surprise inflation or a real interest rate 'gap' (difference from natural rate) as part of the adjustment.

    2. Simon,

      OK. But under the usual assumptions the natural rate path *is* the path of the real rate that intertemporally optimizes social utility. So you are assuming something different. Whether the CB ought to deviate from the natural rate is not a moral, but rather a technical question. Why is the RBC equilibrium real rate not socially optimal?

      It sounds to me like you are saying that we can take advantage of the market implied probability measure for the rate path being "wrong." Eg. the CB has the market believing it's Taylor rule following and the CB can arbitrage that by doing something inconsistent with that implied measure. (Now we're outside of the usual ratex assumption.) As Nick points out though, it seems suspicious that there are gains to be had in the long run from tricking the market. Increased risk premia on term debt seem inevitable. Is it possible that this is a game where the optimal strategy involves the occasional "default" (in the purely game theoretic sense)?

    3. There is no tricking the market here. The natural rate is the rate which would occur under flexible prices. Debt can still be too high under flexible prices. So the natural rate is not socially optimal, if departures from it can increase social welfare by getting debt down.

  8. This comment has been removed by a blog administrator.

  9. The practical counter-cyclical insights of this line of thinking are immense ---- and they have previously been addressed by a handful of economists ranging from Australia's Richard Wood the the USA's Dean Baker. Baker periodically suggests this AND adds a detail which you omitted. Namely, that the monetary authority could minimize the inflation risk, if and when there is one, by raising the reserve requirement. What a difference it would make for economic stability if governments with their own currencies would automatically engage in monetized deficit spending every time a serious recession emerged.

  10. When I studied introductory macro during the sixties the idea that the system could be left in equilibrium at below full employment for years in a row if not decades, was considered too risky by the politicians and socially unacceptable by the vast majority of economists living at the time. Too risky because at that time there was central planing as an alternative and unacceptable because we knew or so we though, what needed to be done to restore full employment under the context of sticky wages.

    On this basis I must say that I find it depressing to hear arguably the cream of the profession asking on the face of unacceptable rates of employment "Why is the RBC equilibrium real rate not socially optimal?" Or to argue against a modest increase in the rate of inflation as timidly proposed by the author of this article on the grounds that it would be " playing with the blood of an economy" or would imply to " move the real rate away from the (socially optimal!) natural rate".

    In my opinion, it worked in the past and I cannot see any reason why it should not work again. The problem is purely ideological and political.

  11. "In that case we are in a different world where comparative advantage applies."

    In other words, it is the world where we should apply what Mundell called the principle of effective market classification: Policies should be paired with the objectives on which they have the most influence.

    The problem is, policy makers' benevolence is not so effective tool for fiscal discipline as is the independency of the central bank. Therefore, from the above principle, the central bank should also take care of fiscal discipline, along with debt reduction. But these are contradictory tasks - I suppose no central bank would be able to manage both.


Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time. In addition, I cannot publish comments with links to websites because it takes too much time to check whether these sites are legitimate.