By Henry Kaspar
One joy of blogging is that it provides an opportunity to present thoughts to a diverse, expert audience, and to see to whether the thoughts survive scrutiny. This post is different: I will ask readers for help with a puzzle that continues to bug me. The puzzle relates to the effectiveness of monetary policy in the liquidity trap.
There are two views on this issue, let’s call them the “traditional” and the “modern” view.
- The traditional view claims that in the liquidity trap monetary policy is powerless. If individuals want to hoard money rather than spend it, injecting more money won’t stimulate domestic demand. The most prominent proponent of the traditional view is of course Keynes. Among contemporary economists, this piece by Willem Buiter reads fairly traditional.
- The modern view has been developed by economists like Paul Krugman and Michael Woodford. It claims monetary policy can overcome the liquidity trap by creating inflation expectations that drive down real interest rates. If folks fear inflation would erode the real value of their cash holdings, this creates an incentive to stop hoarding. NGDP targeting builds on this, in essence it is a policy rule obliging the central bank to target higher inflation during depressions.
My trouble with the modern view is: how can the central bank create inflation expectations when the economy is stuck in the liquidity trap?
True, the central bank can flood banks with base money. It has no means, however, to ensure that banks pass the impulse on to households and corporations through credit creation – which won’t happen as there is no demand. And even when the central bank circumvents the banking system and targets directly the real economy through asset purchases (QE), chances are that households and corporations will either hold the receipt of these purchases as cash, or deposit the receipts with banks, who in turn deposit them with the central bank. I.e., whatever liquidity is injected comes straight back to the originator. With no means of reflating the real economy, however, the central bank cannot stimulate demand. And with no means to stimulate demand, it cannot create a shortage of goods and therefore inflation.
[P.S.: for the latter the central bank would need to buy goods rather than assets, i.e., conduct fiscal policy].
If the central bank cannot create inflation, and markets understand this, how could the central bank create inflation expectations? To use a game theoretic term, announcing a higher inflation target to combat the liquidity trap looks like a non-credible threat.
In short, I suspect the modern argument has a whiff of circularity: it assumes that (folks believe) monetary policy can overcome the liquidity trap, and from there derives that monetary policy can overcome the liquidity trap.
I know of three counter-arguments, but neither leaves me fully convinced:
1. At some point the liquidity trap will end by itself, and the central bank will regain traction. If the central bank promises higher inflation for then, this would reduce real rates already now, and thus help the economy out of depression. This link looks plausible but weak. Consider an economy like the U.S. that goes through a protracted balance sheet recession: credit constrained borrowers pay down excessive debts, transferring funds to creditors who have a higher propensity to save. Until debt is paid down to sustainable levels, private sector demand will be depressed – a process that can take years. The central bank could therefore promise higher inflation only for years down the road, changing savers’ calculus at best for very long-term investments – which, in turn, would arguably trigger a minimal increase in demand only, insufficient to eliminate the excess in desired savings. Cash hoarding would remain prevalent.
[P.S.: this concern is independent from the question of whether announcing a higher long-term inflation target would be time consistent and therefore credible]
2. The central bank could tie its currency to a foreign currency at a depreciated – and therefore inflationary – exchange rate. This is the Svensson mechanism. Essentially the central bank, unable do anchor inflation expectations itself, would temporarily adopt a foreign anchor. I see how this could work for a small, open economy – but not for large, relatively closed economies like the U.S. or the euro area. The required devaluation would be huge, and the countries whose currencies were supposed to appreciate against the dollar or the euro would not allow this to happen.
3. QE works not only through inflation expectations but also other channels, such as portfolio reallocation. I have no problem with the substance of this argument; one may label this Bernanke-QE as opposed to Woodford-QE. Yes, Bernanke-QE can work. But it is a different matter altogether: portfolio reallocation aims at reducing longer-term interest rates, and an economy with positive nominal longer-term rates is not truly in a liquidity trap.
Our commentator Alex F. once noted that macroeconomists can be divided into mechanics and metaphysicists (similarly Nick Rowe). Metaphysicists trust the power of expectations, mechanics need a physical transmission channel. He is right. I am a mechanic. I won’t believe into the success of a policy as long as I fail to understand how it works. And if there are more folks like me out there, fighting the liquidity trap through inflation targeting is doomed.
But I sense I am missing something obvious. Who can help?
@ H Kaspar,
I guess you do not expect someone to come up with a solution..
Anyway, one thing that is certain is that whatever solution the metaphysics com up with (and lets face it, economics is largely axiomatic so all it takes is to persuade rather than convince the others, create a new consensus and maybe that would result in expectations for a while). Again, whatever solutions the metaphysics propose MUST convince the mechanics, because they hold the accountable positions in the policymaking world. Maybe that narrows the field somewhat: the thing that mechanics and expectationists have in common, is probably the thing that could create expectations.
I know of one thing that achieves the opposite of what many people feel would need to be done right now: the policy mix adopted by the Dutch gvt: try everything to push housing prices down (despite the effects likely to result as per your recent post), follow disputable pension accounting rules to undermine consumer confidence further by mandating suggesting (private) pension cuts as a remedy for purely accounting deficits (which private sector employers and employees have agreed to ignore from 2014) and ignoring the strength of the public sector balance sheet when interacting with IMF and EU bureaucrats. But the gvt deficit remains in traditional “stimulus” mode (I mean the talk is gloomier than the action) If that works by pushing one of the few growing economies in EUR (and with strong fundamentals) into a recession (mainly by affecting consumer confidence), maybe the exact opposite works to achieve the opposite…Maybe you should invite the current chir of the Dutch Central bank (previously the top civil servant in the MoF) to explain this form of medication and ask him how to use his contraption in reverse gear. I am sure that would be exactly what we need, in the US and in the EUR on average.
PS as to remedy #2: it appears that the current “policy mix” (if one can call this mixture of improvisation, uncertainty and megaphone diplomacy so) has the (probably unintended) effect that it is hard for speculators and large official reserve holders to accumulate large amounts of EU denominated risk free assets (China likes to have more EUR but does not like to buy Italian bonds and it has no access to the ECB). Hence the EUR/USD may be weaker than it should be (FEER, REER). Maybe that will then also produce an inflationary effect. It certainly frustrates non-EUR producers of tradables and probably leads to greater output than there would have been if the EUR was rising like the JPY/EUR.
Wie wäre es damit: Die Zentralbank gibt eine zweite Währung heraus, die parallel zur Ursprungswährung läuft, von der Regierung zum gesetzlichen Zahlungsmittel erklärt wird und eine Umlaufsicherung besitzt(z.B. ein aufgedrucktes “Verfallsdatum”). Dieses Zweitgeld würde per Scheck oder Paket oder Helikopterabwurf direkt zu den Haushalten gebracht. In einem System mit zwei Währungen sagt Gresham´s Law nun, dass die weniger wertvolle Währung die andere verdrängt, weil alle Leute versuchen, diese schwache Währung loszuwerden. In diesem Fall wäre es die Währung mit Umlaufsicherung, da sie ab einem bestimmten Zeitpunkt nur zu bunt bedrucktem Papier wird. Man würde also quasi ein Splitting der Geldfunktionen erreichen: Das alte Geld übernimmt die Wertaufbewahrungsfunktion, während das Umlaufgeld die Transaktionsfunktion bekommt(Die Zweitwährung würde so zum reinen Transaktionsmittel, wie sich das die Vorkeynesianer gedacht haben). Das Problem der zu hohen Geldnachfrage – der Knackpunkt der Liquiditätsfalle – sollte so umgangen werden, da die Leute ja alle sofort losrennen, um ihr “schlechtes” Geld loszuwerden. Es wäre nebenbei auch ganz interessant, das mal in Griechenland zu versuchen, falls die aus dem Euro austreten. Da bräuchte man wegen dem Abwertungsdruck der neuen Drachme wahrscheinlich nicht mal eine Umlaufsicherung und könnte den Griechen auch erlauben, ihre Euros zu behalten(also keinen Zwangsumtausch zu machen), was dann en passant auch eine Kapitalflucht incl. bankruns verhindern würde.
Das ganze Vorgehen ist natürlich nicht problemfrei. Ich wüsste z.B. nicht, wie man das bilanztechnisch (in BWL-Hinsicht) und im Hinblick auf die Schulden behandeln müsste. Auch über den Effekt auf den Staatshaushalt müsste man nachdenken – was passiert mit den Steuereinnahmen, wenn alle das nur noch mit Umlaufgeld bezahlen? (Für Griechenland wahrscheinlich irrelevant, da die dann sowieso durch den Default müssten).
Was haltet ihr davon?
Gresham’s law greift nur bei festen Wechselkursen zwischen den beiden Währungen und auch nur wenn dieser falsch gesetzt ist, also eine Währung relativ zu hoch bewertet ist.
@ Jay,
If your Nebenwaehrung had a limited purpose as well as a limited life, it could be quite useful: suppose all people could do with it would be to pay taxes (and transfer it to their employers to pay payroll tax and to shopkeepers to pay VAT). That would achieve the same as a temporary tax cut but it would not be fiscal. Economically irrelevant (for some) but politically powerful.
@John Gresham: Huch, das hatte ich irgendwie anders im Kopf :) Aber würde der Effekt nicht der gleiche blieben? Wenn eine Geldsorte ein “Verfallsdatum” hat, würde ja alle Leute versuchen, dieses Geld auszugeben und das andere zu als Wertaufbewahrungsmittel zu horten.
@Rien Huizer: But it can become fiscal when the nebenwaehrung expires – when the government holds the nebenwaehrung at the expiring date. Or am I missing something?
@ Jay
Das klingt interessant; wie allerdings kriegt man den Switch zurueck zur alten Waehrung hin (welche die Leute ja weitehrin horten wollen)?
Sorry für die lange Wartezeit für die Antwort.
Also im Idealfall wäre die Liquiditätsfalle bereits gegessen, wenn man wieder zur alten Währung zurückkehrt, weil die WIrtschaft wegen der gestiegenen Ausgaben wieder wächst und die Geldnachfrage / Nachfrage nach sicheren Assets wieder abnimmt. Das ganze würde in den Tagen vor dem Wertverfall ähnlich einer Hyperinflation funktionieren: Da niemand die Nebenwährung halten will, nimmt die Kreislaufgeschwindigkeit immer weiter zu(bzw. die Geldnachfrage immer weiter ab). Dann könnte man durch einen dann veröffentlichten, festen Umtauschkurs von Nebenwährung zur alten Währung veröffentlichen und die Leute das Geld umtauschen lassen, ähnlich einem Währungsschnitt(und hoffentlich wären dann “die Schaufenster wieder voll” ;) ). Aber ich bin mir nicht sicher ob das Ganze so funktioniert wie ich mir das vorstelle, ich müsste es wohl mal modellieren…
@HK
“But I sense I am missing something obvious.”
I don’t think you do. I think you hit the nail on the head: “To use a game theoretic term, announcing a higher inflation target to combat the liquidity trap looks like a non-credible threat.” The only credible threat is the one you already mentioned: the central banks would have to buy stuff directly in the economy OR it can set a negative rate of interest. The latter could be done, I think (the former also, but I don’t think that this is politically feasible).
” … how could the central bank create inflation expectations?” Nein, das kann sie auch nicht . Sie kann auch nicht andere Erwartungen erzeugen.
Sie kann überhaupt viel weniger, als manche Leute ihr zutrauen, wenn es um Politik geht, also um Lenkung. aber sie kann sehr viel dummes Zeug anrichten, das sie selbst nicht übersieht: Sie wirft Bälle in das Spiel, ohne überhaupt den Flug und die Endlage zu kennen.
Geldpolitik folgt der Realpolitik.
@ Ernst
Seriously right!
@ HK
I think the mechanical way to think about this truly underestimates the power of expectations, and multiple equilibria. And I blame the focus on interest rates, as my last post showed, for this. Sure, you should know your mechanics well, and maybe I need some more of that. But I think it is impossible to really understand monetary economics in a purely mechanical way (although in theory, it should be perfectly possible to do that, but it is very hard to model it).
The post of Nick Rowe that you linked to in my view is the best way to phrase and explain it. Maybe you can help me out by showing what it is in his post that you disagree with exactly.
@ Kantoos
What’s missing in Nick Rowe’s post is that Chuck Norris is handcuffed – he can’t beat anybody up. Thus the threat is not credible. A non-credible threat shouldn’t induce people to change their behavior.
@HK
are you just making a theoretical argument, or do you think Europe is in a liquidity trap? And if yes, what is the reasoning for that, because I need to explain this to one of my economics professors, who thinks we are not. :)
@ Johannes
good question ;-)
I reckon risk-free nominal interest rates would suggest that the euro area as a whole is close to the zero bound, but not yet there.
Cheers,
HK
1. You ask the right question: how can the central bank create inflation expectations when the economy is stuck in the liquidity trap. And you know the answer: it can’t because it is time inconsitent — we all know that as soon as the central bank creates inflationary expectations, the economy expands, and inflation continues to accelerate the central bank wil begin a proces of disinflation. The central bank as we now it can not credibly commit to debauch the currency.
2. But here are a few ideas for credible policies to get people to spend:
- implement a tax on money held in banks at a rate of 5% per month, increase the rate if neccessary
- announce a confiscatory change of currency to take place in two weeks, repeat as often as neccessary
- double government employment and lower retirement age by 10 years, furthermore each new government employee should be given a contract that states that if they’re ever fired they will receive a cash severance payment of 120x average monthly wage
- forbid holding of gold, silver, and foreign currencies under the penalty of death (also may need to do something about real estate, either nationalize it, or tax it at a confiscatory rate)
…there you go. That should do it…shouldn’t it?
Either that or wait till balance sheet problems have been sold to a point where demand for cash holdings decreases.
p.s. this just occured to me — and this is only half in jest — nominate Hugo Chavez and Fidel Castro to run your central bank…
And you know the answer: it can’t because it is time inconsitent — we all know that as soon as the central bank creates inflationary expectations, the economy expands, and inflation continues to accelerate the central bank wil begin a proces of disinflation.
I fear the problems start even earlier – the central bank cannot generate inflation expectations if the slump is sufficiently deep (thus an end to the liquidity trap is many years down the road) and the economy is large and/or relatively closed (thus the Svensson mechanism is beyond reach). There is simply no credible transmission mechanism.
Time inconsistency comes on top of it.
As for the proposals listed under 2., they aim at physically destroying the zero bound. This gives rise to interesting thought experiments, but not necessarily to implementable policies.
No transmission mechanism? Here is another thought experiment:
- the central bank announces that it will for an undefined but finite period top up every deposit made into every bank acccount by say 10% — i.e. if someone deposits 1000 Euros into your account the central bank will add another 100 Euros — in short order a lot of people should be shifting money between accounts AND spending some of the topup money because they realise what will happen when everybody will start doing the same.
Sorry to persists with more of such thought experiments. But the point is the tools are there, but the will to use them is not. We have no implementable policies because the political decision on who should take the losses accumulated on bank, company and personal balance sheets has not been made — we are poorer than we used to belive, now we must sort out who and by how much.
Once the decision on who takes what share of the loss is made, we can use inflation, deflation or other means (see chart below for a spectrum of historical examples) to arrive at a situation where demand for money will return to a state conistent with recovery.
The spectrum of tools for dealing with too much debt:
A <– B <– C D –> E –> F
Pro Creditor Pro debtor
A = debt slavery
B = confiscation of all assets and future earnings until debt is repayed
C = deflation
D = inflation
E = default on all debts or official cancelation of all debts
F = burn creditors at stake or organize pogrom of creditors
@ Pawel
Happy about thought experiments, but I find them inconclusive: why should folks spend these deposits if the profitable thing do is keeping the money in the bank and earn a high real return on it? And why should they anticipate that somebody else would start withdrawing these funds if it is in nobody’s interest to do so?
This seems the same circularity as pointed out above: if folks expect that the liquidity trap can be overcome the liquidity trap can be overcome. But they have no reason to believe this.
Pawel,
You forgot point 3. After doing everything in point 2:
3. Revolution!
Congratulations. You just managed to convince 80+% of the population that the current government is totally mad. Revolution occurs. The new Revolutionary Council announces a new currency and that all old debt will be null and void. Problem solved. :)
Snark aside. How do you expect any government being able to implement even half of your ideas and stay in office?
@ Detlef: Thank you, that is exactly the point: the things that would work (my extreme though experiment examples aside) are not politically feasible. Without a revolutionary break with the status quo in terms of solutions that are applied (i.e. somebody here & now deciding how to apportion the balance sheet losses to banks, companies and households) we will just need to wait till balance sheet problems get sorted out the old fashioned way…
@ H Kaspar: The thought experiment with the central banker would result in a debauched currency — if you and I kept start with Euro 1000 and keep depositing it in each others accounts, with the central banker topping up each deposit with 10% (or whatever sum) we would soon end up with millions — an obviously unsustainable path to follow once everybody starts doing the same. Just a convoluted variant of the helicopter drop I guess.
@ Pawel
People familiar with the policy repertoires envisaged when VIL had come to power unpreparedly in Russia (one of the craziest gambles ever undertaken by a German gvt) would go a lot further. But let’s restrict ourselves to measures that are constitutionally feasible and/or compatible with the Rechsstaat. Dictatorship economics is simple but unlikely to solve the problems here.
@HK: Thanks very much. Good overview of the debate.
I guess propenents of the view that central banks could do more would dispute your claim in 1. “This link looks plausible but weak.” If the central bank can credibly commit to sustained monetary expansion even when the economy is recovering (i.e. allow for temporary higher inflation), then this would support aggregate demand immediately. I am not so sure that the effect on private demand of lower real interest rates would as small as you think. After all, all households that are not overleveraged and all companies should be induced to increase spending, shouldn’t they?
@ David
I also sense that this is where the proponents of “committing to behave irresponsibly” would come in. But with what argument? Suppose you are a household, there is significant deflation and you don’t expect things to change earlier than in, say, 5 years. Until then you can hold your money safely in cash and earn a hefty real return. Why should you make your purchases now, rather than 5 years down the road when the opportunity cost of spending will be much less?
I hate to answer a question with another question… But if the central bank committed to raising, say, house prices by 5% per year and stated that it would buy as many houses as necessary to achieve that goal, do you think it would alter private sector behaviour? I believe it would (especially in the UK, perhaps less so in Germany…)
Is a commitment to alter the supply of base money to influence the general price level fundamentally much different? I don’t know the answer to this, so if anyone wants to chip in…
One more point. I suspect coordination with the fiscal authority is an important part of the mechanism that would shape expectations. To borrow Paul Krugman’s phrase, if the central bank is to be ‘credibly irresponsible’, the government must be complicit. Given how politically divisive inflation can be, this must mean the issue isn’t quite as technocratic as the ‘metaphysicists’ sometimes make out.
@ Dom
that’s an interesting thought, as higher house prices would repair many overindebted households’ and financial institutions’ balance sheets. One could hope that this would reduce debtors’ borrowing constraints, and thus enable those to spend who have a high propensity to do so – while standard monetary easing would rely mostly on changing the behavior of unconstrained savers with a low propensity to spend.
The flipside would be that central bankers would have to get versed in the details of writing rental contracts and delivering janitor services…
@ Kantoos, HK
HK put it well with the credibility argument. “Simply” influencing expectatios is like “positive thinking”. Tell depressed people they just have to think positively about the world – and they will perhaps get even more depressed.
Changing expectations is hard work – for individuals and communities. World views have to change, theories have to change, conventions have to change – this is not easily done. And then there is the actual state of the real world: How can somebody be optimistic if he has just lost his job, house, little shop, whatever – ask Greeks, US households etc. It’s like the film “Up in the Air” with George Clooney who flys around and fires people – telling them right afterwards what a great chance their new life could be. In short: While I wholehartedly agree that optimism and pessimism, different expectations etc. lead to good and bad equilibria, I would not agree that every equilibrium is possible all of the time. This is what people call “path dependency”.
Hk, isn’t your post actually saying that the people who are in the room with chuck norris are handcuffed or chained to their chair? Chuck can beat them up, but they still won’t be able to leave the room ;)
@ HK
I think that your qualm with the “modern view” is due to the fact that you don’t distinguish between the Krugman model and Eggertsson/Woodford model. Your skepticism is justified in the case of Krugman, yet not with respect to the E/G model. The E/W model does NOT assume that inflation expectations are controlled independent of some instrument variable (nominal interest rate PATH) that could actually determine inflation (the “mechanical view” as a necessary condition to exploit the “metaphysical view”). The model assumes that any time period with negative equilibrium real rates is finite and that, once the equilibrium real rate is positive, a zero nominal rate suggests increasing inflation. In a rational expectations model, the central bank can exploit this “mechanism” during the time period with negative equilibrium real rates (with policy rates at the ZLB): It can commit itself to keep nominal rates at zero level AFTER THE “CRISIS”, that is, when the Wicksellian rate turns positive again, whenever this may happen (in one year, ten years, fifty years, …). This, of course, raises expected aggregate demand and inflation expectations for the period following the crisis and, thus, increases present aggregate demand.
@ amv: isn’t this exactly what I summarized under counter-argument 1 – and also why this link looks weak, especially if the Wicksellian rate turns positive only many years down the road?
Cheers,
HK
amv,
If I´m indebted now (mortgage, consumer loans, credit card debts whatever) and trying to reduce my debts why should I make more debts now on the hope for higher inflation in 1,5 or 10 years?
Especially since nobody can guarantee me that I will have a job then. Even less a job with a wage that will rise as fast as inflation?
After all, inflation is only good for a borrower if my income rises faster than inflation. Given outsourcing, offshoring and globalization can the E/W model guarantee that?
Looking at the balance sheets of economic agents could be helpful. I think in your last post you did that and I felt it was quite enlightning.
The question would then be: can monetary policy lead people to changes in their targeted balance sheet? The problem is that in a deleveraging crisis some economic agents (the debtors) are forced to change their balance sheet quite drastically.
For the whole process to work half way smoothly other economic agents (the creditors) have to adjust their balance sheets accordingly in a voluntary fashion. Here expectations might play a role. The question is wether it is possible to change creditors’ expectations (through monetary policy) in a way that they are willing to change their balance sheet as drastically as the debtors are forced to. This adjustment could for example happen by creditors consuming their savings or trying to find other assets to save into, preferably the stuff that the poor debtors produce or have on their balance sheets. If the change in balance sheets of the creditors falls out of pace with the change of balance sheets of the debtors, deflationary forces will gain the upper hand.
In my model world there is also a financial intermediary between the debtor and creditor agents which will to large extent control the behaviour of the debtor agent but to a much lesser extent that of the creditor agent (since the creditor agent doesn’t owe anything to the bank but rather the other way round). Central banks even with semiconventional policies usually have some clout with the banks (the transmission mechanism here being quite “mechanical”).
The key issue about effective monetary policy would then be to reduce the speed at which the debtors are deleveraging for example through working the banking sector (mechanical) while at the same time do the very best to trick creditors into changing their expectations and hence their balance sheet (metaphysical).
Doesn’t the “mechanical” outlook treat the liquidity trap as a brute fact: as if it just “is”? If one treats it as something caused, then changing what caused it changes whether one stays in the liquidity trap.
I find the “mechanical” approach unhelpful, since it seems to me obvious that expectations frame behaviour. That people will behave differently with different expectations. (Indeed, that “mechanical” interactions are not necessarily stable for that reason–see Goodhart’s law.) So, one can issue all the base money one likes, but if it does not change expectations it will have no effect on behaviour.
As Scott Sumner put some time back, if we put the head of the Zimbabwean central bank in charge of the Fed, that would change expectations. But issuing lots of base money and making it clear that the Fed will continue to target very low inflation, that does not.
It seems obvious also to me that expectations frame behavior. What is not obvious to me is that folks would believe in something they have absolutely no reason to believe in, only because the CB tells them so (and perhaps accompanies this with some irrelevant charades, such as large-scale asset purchases). I reckon in an entrenched liquidity trap folks should and would expect deflation – with no credible means of the central bank of changing this. If this is true, expectations management QE would not work.
As for treating traps as an “is”, no, not necessarily. As sketched above, I reckon liquidity traps are typically the result of debt overhangs that force liquidity constraints on borrowers. If that debt overhang is severe enough, desired S (to pay down debt) will be higher than desired I, even at zero nominal rates.
You’re right, massive write-downs of debt, for example, could eliminate the debt overhang and get the economy out of the liquidity trap, but this is an entirely different policy than monetary easing. More generally, I believe in the liquidity trap there are more effective instruments than monetary policy.
You are right. You indeed mentioned the mechanism. Sorry, it’s late and my brain evidently stopped working. I read your post piecewise during the day (due to my job) and overlooked your first counter-argument. This argument, however, rests on an implicit assumption: that the balance-sheet recession is independent of nominal spending expectations.
Assume – for sake of the argument – that expected nominal spending is on target. Assume further that some significant fraction of households is overindebted and engaged in reducing its debt burden. This suggests a transfer to their creditors, as described in your post. Since nominal spending is expected to be on target, the creditors’ relatively high “propensity to save” does not suggest a “leakage” (an old Keynesian term). They save because their rate of pure time preference is low relative to the ROI, which suggests investment demand. Thus, even though a fraction of households is overindebted, there are still agents populating our model (i.e. firms) that are willing to borrow (given the assumptions above). If the creditors’ time preference is high relative to the ROI, they will consume instead. In short, if spending expectations are on track, partial overindebtness leads to a change in the composition of demand. (A positive rate of pure time preference and productivity growth suggest a positive Wicksellian rate; negative productivity shocks may temporarily lead to negative equilibrium rates; I guess that we agree that in this case the E/W model applies). Thus, overindebtness alone cannot account for the balance-sheet recession you described in your last post. In a closed economy with anchored nominal spending expectations, the additional (investment or consumption) demand of creditors substitutes for the decrease in the demand of overindebted households.
A lack of demand occurs only if the creditors have a high liquidity preference (instead of a low rate of time preference). Quantity theorists would speak of lower velocity. Liquidity preference increases in macroeconomic risk. But macroeconomic risk is significantly mitigated if nominal spending expectations are on track. Thus, the scope of the balance-sheet recession depends on the central bank’s management of inflation expectations. It is therefore problematic to take the balance-sheet recession as given, and to wonder how the E/W expectation mechanism could ever apply. The point is that a credible commitment to keep nominal rates at zero beyond a negative-real-rate horizon simultaneously(!!!!) increases current nominal spending and mitigates the scope of the balance-sheet recession.
@ amv
Nice thoughtful post (as always). I reckon we disagree on:
– “assuming (for the sake of the argument) that nominal spending expectations are anchored.” One can’t assume that. This is what makes the argument circular.
– “lack of demand occurs because of high liquidity preference.” Low time preference plus a shortage of safe assets triggers the same outcome – and I’d argue this is what actually happens. I don’t believe balance sheet recessions have anything to do with folks “wanting” to hold money (deliberate hoarding occurs maybe in periods of intense financial strains, but these are short-lived). Balance sheet recessions result from forced income transfers from creditors to savers, thus high aggregate savings preference, coupled with a shortage of safe savings vehicles. Desired savings then spill over into money, and velocity drops.
(you’ll probably know Brad Delong and Nick Rowe discussed this some time ago)
Cheers,
Hk
@ zebulon
I read your comment after posting mine. We make a similar point, aren’t we?
@ Detlef
If you are an overindebted individual who pays off debt you don’t spend more because of the E/W mechanism. But those who receive those payments do.
@amv
That is right but if that creditor is a bank in need of balance sheet repair (the big exogenous elephant in our room is that banks (due to “incidental” regulatory change) are more or less forced (shareholder interests take priority) to shrink their balance sheets. Hence your shrinking real sector budget does not mirror relaxing someone else’s real sector budget constraint. And perceptions of wealth (rather than expectations of inflation-cum-rising nominal wages that are simultaneously suspect as Detlev mentioned) are a powerful yet hard to assess influence. Solving the banking problem will go a long way towards creating the conditions where expectations-based monetary policy could work (and then only if the consumption/residential real estate related forms of AD are desirable from a longer term economic structure perspective)
“Metaphysicists trust the power of expectations, mechanics need a physical transmission channel.”
Even if you think that expectations are crucially important (which they are :)), you would need some “physical” mechanism in order for expectations to form in the first place.
@ HK
1) The argument is not circular. In assuming on-target nominal spending, I just wanted to show that partial overindebtness does not necessarily imply an aggregate balance-sheet recession. Stated differently, a lack of aggregate demand is a necessary condition for an aggregate balance-sheet recession. In the first part of my post I did not specify how current nominal spending is determined. Think of fiscal policy, that is, the treasury engaged in credit-financed purchases of sufficiently many commodities (in this case the treasury is willing to borrow as a substitute for the overindebted households). If you accept that fiscal policy is able to keep current nominal spending on track, and that this significantly mitigates the balance-sheet recession, all that I have have to show is that monetary policy at ZLB is also able to determine current nominal spending. This is done by the E/W model. Here, current nominal spending is determined by the expectations of future nominal spending, and expected nominal spending is determined by the central bank’s commitment to keep the real rates below the Wicksellian rate (creating inflation expectations), once the Wicksellian rate turns positive.
2) Again, you assume that the collateral crunch is independent of nominal spending. Further, you write: “Low time preference plus a shortage of safe assets triggers the same outcome.” A safe asset is most often a liquid asset, and a liquid asset is most often a safe asset. A shortage of safe assets is most often a shortage of liquid assets. Note that I did not argue that the increase in liquidity preference is due to a change in the preference structure of creditors. I argued that – given the preference structure of creditors – an increase in macroeconomic uncertainty (and thus nominal risks in general) increases the “appetite for liquidity” (by analogy: given the risk aversion of agents, more uncertainty increases the appetite for risk). Again: I did not suggest that creditors woke up some day and felt like they wanted to hoard more, everything else given. I argued that they realized that macroeconomic uncertainty increased, and that – given(!) their preference structure – more liquidity will insure them against this increased nominal risk. A low rate of time preference suggests a low steady-state Wicksellian rate. Given the preference for liquidity, a shortage of liquid/safe assets suggest a high liquidity premium, that is, an increase in what is added to the Wicksellian rate to determined the nominal rate (in addition to the risk premium and inflation exp.). Supply and demand for liquidity determines spreads, not the Wicksellian rate that we discussed above.
Don’t get me wrong. I do not share the Market Monetarist believe that proper monetary policy can solve all of our problems. The collateral crunch could be mitigated, this for sure. But the relatively scarcity of safe/liquid assets is also due to fundamentally justified reassessments of risk (e.g. MBS in the US; sovereigns in Europe). To the extend that this is true, higher yields will persist as an optimal response of the market system, and they will persist even after your balance-sheet recession.
@ amv
I’m not sure I fully grasp your argument, so let’s go through it steps.
1) Nature of the shock
I have in mind an economy composed of borrowers and savers, where – as a result of a debt crisis – borrowers find themselves credit constrained. This transforms temporarily an economy with a relatively low aggregate propensity to save into one with a high aggregate propensity to save. If an economy’s propensity to save is high enough it will be in a liquidity type situation. For the sake of the argument assume folks in an economy want to save 50 percent of their income no matter what, but there are only investment projects with a real return of 20 percent. This has little to do with liquidity preference.
2) Slipping into a balance sheet recession
Now you’ll probably (and rightly) take issue with the “save 50 percent no matter what” assumption. High enough inflation (or NGDP) expectations would make savers save less. But this comes back to the question whether the central bank can generate these inflation expectations – or: keep expectations anchored in spite of the shock to borrowers. If savers observe that borrowers are credit constrained, understand that this will ceteris paribus trigger a sharp, deflationary drop in demand, and also understand that the central bank has no tool to at its disposal to create a physical shortage of goods (and thus inflation), savers have no reason to expect inflation and save less – if anything they should expect deflation and save more.
Is there an element in your argument that is not adequately reflected in the above?
3) Fiscal policy
Your comment about fiscal policy is interesting. If savers expect that the fiscal authority would be the residual buyer of goods (possibly financed by discounting the bills at the CB), and thus if needed create a shortage of goods that translates into inflation, yes, this should change savers’ calculus and make them spend more. Similarly, if savers expect borrowers’ debt to be written off, enabling borrowers to spend again, this should affect their calculus and spending pattern also.
In my view this suggests that affecting savers’ calculus – and therefore anchoring NGDP/inflation expectations – requires a credible policy instrument at hand to generate inflation, independent of what savers do themselves. Further, the institution that has the instrument as its disposal needs to be willing and able to use it. However, the central bank is unlikely to be that institution itself, at least not as long as its tool kit is restricted to QE.
Cheers,
HK
Typo: instead of “given the risk aversion of agents, more uncertainty increases the appetite for risk” read “given risk aversion of agents, more macro-uncertainty decreases the appetite for risk”.
This is indeed a conundrum. A coulle of additional thoughts (feel free to shoot them down)
- CB buying goods: if the goods are perishable, the buying will increase the price and if the goods afterwards perish in the handsof the CB, there is proper inflation; if the goods are not perishable, the price effect will be lower because of the potential future supply that acts as overhang. Only a transitory impact thus. So both options would seem suboptimal.
- CB giving free money to population: this will create liquidity and only non-domestic holders experience a value decrease of their money; the domestic population remains neutral as the inflation is compensated by more money. More chance that the population will spend it than the banks. Liquidity trap solved.
Hallo Herr Kaspar, nur ein kurzer Gedanke: Ich bin und bleibe sehr skeptisch, was die Möglichkeiten der Geldpolitik angeht, wenn sich die Wirtschaft in einer Bilanz-Rezession befindet. Es ist ja nicht allein die fehlende Kreditnachfrage. Die Situation ist meiner Ansicht nach vor allem dadurch gekennzeichnet, dass die “beleihbaren Pfänder” überreizt sind. Einfach formuliert: Die Geschäftsbanken generieren kein Geld für Leute, die nur noch über eine zweifelhafte Kreditwürdigkeit und zweifelhafte Pfänder verfügen.
Ist Japan dafür nicht ein “gutes” Beispiel?
Wie wurde die Bilanz-Rezession der Weltwirtschaftskrise ab 1929 überwunden? Meiner Meinung nach dadurch, dass die Regierungen Geld generiert haben, ohne auf Rückzahlung zu achten. Extreme Fiskalpolitik “dank” Aufrüstung und Krieg. Das bricht jede Liquiditäts-Falle. Bitte den nächsten Satz nicht falsch verstehen: Ich bin nicht sicher, wie lange die deflationäre und rezessive Periode der 30er-Jahre ohne Hitler gedauert hätte.
Also, was tun?
Dem ESM eine Banklizenz geben. Dort massenhaft Geld schöpfen und über Sozialtransfers (nicht!!! Kredite) an die Massen (nicht!!! die Elite) verteilen. Schon läuft die Inflation.
Gruß
SLG
I also find that discussion a little funny. How can you possibly fail in creating inflation if you can create money out of thin air. :) Just give it to someone who will spend it. Despite banks, companies and rich people, everyone will do.
@ HK
I think that our mutual ‘misunderstandings’ root in divergent priors. It would take a good conversation with at least one bottle of wine to arrive at a common language. Since my identity is available, tell me whenever you are in town. I bring the wine. Seriously, perhaps future posts will help us out.
@ amv
I sure take that! And vice versa, should you head in my direction (hkaspar@gmail.com for details)
Upon rereading your posts a few times though it seems to me that our main disagreement is on whether E/W can keep nominal spending/inflation expectations on target. You think yes, I consider this implausible, especially when the shock to AD is large. The quibble about liquidity preference vs. propensity to save is about how we characterize the shock to AD, I don’t think this is essential.
Cheers,
HK
Well, I am afraid it’s hard to find answers on HK’s question as long as you assume all market participants act rationale and care about any macroenomic theories.
I am confident that media and politician speeches do have much more impact.
People will save their money if they feell uncertain about their future chances to gain new income, meaning they are afraid to become unemployed. People will start to “exchange” money for other goods if they are afraid theat the money value may drastically decrease in near future.
Their views in regards whether the first or the later scenario is more realistic is not build upon any macroeconomic theories. 99% will not understand the factors which influence these theories. So they need to trust in other information. Those they get out of the media. And even if these information is as theoretical incorrect, they will believe in them and act respectively.
So you want to increase the spending while in a liquidity trap?
Just tell the people that most likely money value will decrease shortly.
Look at Greece, the fear that they will leave the Euro brings people to get their money out of the banks. However they don’t spend.Why? Because they assume if they get the drachma back, they will get a lot of drachma for their Euros at hand. So they assume that the money value will increase!
Make them believe that they will keep the Euro and that wages will not decrease even more but increase to push the economy. They will assume that the prices will increase as well and rather spend their money today.
I think this is a relevant observation. If two smart people like amv and hkaspar need at least one bottle of wine to determine whether a specific model does or does not circumvent the circularity argument, I am skeptical as to the real-world significance of those finer points :) Also, as far as I know, empirical studies on inflation expectations based on surveys find the general public much more backward-looking (or at best “present-looking”) than economic theorists would like.
That reminds me of a panel of MIT economists, in which Peter Douglas and Robert Solow suggested more fiscal stimulus or tax cuts (something like this).
Question Audience: …In our core graduate classes we teach rational expectations… tax cuts will not work since people spend less anticipating future tax increases.”
Douglas: “At MIT we don’t teach those models … I don’t think these strict models hold up at all.”
Solow: “That is much more polite than I would have said it…”
Here is the whole Panel (last question).
http://web.mit.edu/newsoffice/2011/federal-deficit-panel-1006.html?tr=y&auid=9754887
@ amv and @ hkaspar:
I highly endorse the modern view of influencing current macro outcomes by shaping long-term expectations and yes, in principle it should work. However I doubt the fact that agents in the E/W Model
- can smooth consumption over the necessary Time horizon once the Real equilibrium Rate is positive again
- agents have a safe asset for a period of lets say 5 or 10 years to transfer wealth
- do NOT price risk in the Valuation of assets. Risk does not play any role in the model, neither for future expected nominal spending nor for pricing assets. Agents Decide as if they are certain about the future.
- can shortsell and agents are not at least partially constraint by current income
By the Way:
- can monetary policy influence the fundamental Level of real risk?
- what is if the Risk-neutral nominal yields are expected to be Zero for a Prolonged period of time, but the relevant risky yields are not? In the E/W model, monetary policy only influence Long-Term Risk-neutral yields.
Best,
Wiknam
HK wrote: “And even when the central bank circumvents the banking system and targets directly the real economy through asset purchases (QE), chances are that households and corporations will either hold the receipt of these purchases as cash, or deposit the receipts with banks, who in turn deposit them with the central bank.”
Money demand need not be the same for all actors. Also, introspection makes it is plausible that money demand is finite for all individuals, given all other factors. (Why would it be infinite? Do you have an infinite demand for money, i.e. would you just hold on to a million, ten million, hundred million dollars in a liquidity trap?)
If the central buys assets (QE) and increases the money supply in the economy, _some_ actors will reach and surpass their desired level of money holdings and start spending their “excessively” held money on goods and services. This increases velocity and raises the price level for some goods, creating inflation. If the central bank announces a sufficiently large level of QE, which is assumed to be permanent, some actors will foresee this rise in prices and also spend some of their money before its purchasing power is lowered by inflation. This is the key channel, it seems to me.
Would people put their newly acquired money into the bank? Some surely would, if the bank deposits offered interest rates higher than expected inflation. This need not be the case. Indeed, if the real interest rate is negative (this being the cause of the liquidity trap), banks could and would offer nominal interest rates lower than the expected rate of inflation. Under these conditions, some actors would still find it preferable to put their money into banks. Other however would find it preferable to spend some money on consumtion now before inflation lowers its purchasing value.
Would banks deposit their money at the central bank? That depends on the interest offered by the central bank. Presumably at some rate of expected inflation, given a low level of IOR offered by the central bankes, they would rather invest it. Expected inflation would also raise future nominal gdp, thus raising return on investments and thus investment of the banks.
A crucial point is for people to expect that QE will not be reversed by the central bank as soon as inflation picks up. QE will only be able to influence inflation expectations if QE persists through the initial phase of rising inflation, until real resource underallocation is reduced. This would not be a plausible expectation if the central bank had a pure inflation target, since then it would have to combat inflation immediately. With a nominal gdp target however, such assumptions could be credible. Rising prices are the mechanism which induces people to raise their expenditures and so lead to a rising gdp. Hence the importance of a clearly articulated level-target of the central bank for inflation expectations and the ability of monetary policy to influence gdp.
@ wiknam
Good to hear from you, buddy! Sure, The E/W only talks about the Wicksellian rate. No risk premium there. This is something that applies to most pre-crisis models. I can only recommend this most excellent book which integrates macro and finance: http://books.google.de/books/about/The_Yield_Curve_and_Financial_Risk_Premi.html?id=65bTS117JmMC&redir_esc=y
But that is an altogether different question. If HK is right, then you don’t have to add complication to the E/W model … you can dispense it immediately.
@ Eclair
I always assume that the market in toto is much smarter than I am ;-)
@ Eclair
Actually, you raise an interesting point. It is a common presumption that RE means that agent must know the model, and that it implies that real-world individuals must be able to arrive at such model by individual learning. IMHO this presumption is mistaken. Like all other models, RE models predict general equilibria. But RE does not restrict our choice of convergence processes, the out-of-equilibrium behavior of the system as a whole. It would be heroic to assume that individuals with diverse believes and bounded information-processing capabilities could ever ‘learn’ a common model when confronted with their idiosyncratic perception of ‘reality’ (google ‘Mordecai Kurz’ and ‘rational beliefs’).
A better way is to think of ‘impersonal learning’ and ‘system intelligence’, an ‘invisible hand’ similar to the hand that coordinates an ant hill, or that is expressed by the collective intelligence of a fish swarm. As soon as HK and I bet on our models – and put money where our mouths are -, one of us or both would lose wealth and, thus, possibilities to influence prices. This wealth would increase the influence of those who – by chance and/or better knowledge – made relatively better model choices. If this impersonal process converges, and converges fast enough, we can think of real-world prices ‘as if’ they are a REE (EMH).
Conclusion: The fact that HK and I cannot agree on a model thus not suggest that RE modeling is somehow futile.
@ amv
Well, yes, the mechanisms you describe would undoubtedly work well in an ideal world. Just as the Austrian conviction that an economy with truly free markets (and no CB tempering with the interest rate structure) would pretty much always be maintained on its production possibilities frontier with all resources employed is perfectly in line with that ideal. In a world of that kind (which, among other things, would no longer require the institution of money) I would be a Hayekian :) I just happen to believe, with Keynes, that this is not the world in which we actually live, and that judging real-world economic and social institutions against the benchmark of that particular ideal can quite easily yield policy conclusions that are detrimental.
Incidentally, applying Kurz (1974) to modern financial markets leads you straight to the work of Chichilnisky since the early 1990s on endogenous uncertainty that reads quite prescient today. In 2006, e.g., she wrote (with Wu):
“New financial instruments [that] are introduced every day including indices, derivatives and innovative forms of government debt […] help manage risk and improve economic welfare. However, they can also increase macroeconomic volatility. The complexity of contractual obligation within a market can transmit individual risks and amplify them into correlated or collective risks. […] Since new instruments create new webs of obligations, financial innovation is the precipitating factor. The transmission of default from one trader to another and from one market to another transmits individual risk and magnifies it into collective risk. Although the individual risks are exogenous, depending on states of nature, the new collective risks are endogenous, depending also on economic behavior such as how many agents trade with those who defaulted.”
In other words, here you have risk, uncertainty and the potential for catastrophe created precisely as a result of the little ants’ collective “system intelligence” playing out – eventually, the ant hill collapsed.
Accordingly, with respect to rational learning theories, I personally agree with Colander and Rothschild’s (2011) assessment : “the whole exercise of looking so hard for tractable learning dynamics that can “rationalize” equilibria by stabilizing them strikes us as a being a tell-tale of misguided vision […]: it stems from a view that equilibrium should be inherently stable separate from the institutional structure that endogenously develops to maintain stability.”
The answer is by credibly committing to increase not only the current money supply, but also all future money supplies ( see Krugman, 1998; Bernanke, 1999; Friedman, 2000). In other words, a CB can increase inflation expectations by targeting a “higher” inflation rate.
If to be completely mechanical in solving economic problems you should admit that only measured physical-like variables can drive inflation (no expectations). In that case central banks are not able to get out of deflation by changing their rates or by QE. The case of Japan shows well that mechanicall speaking there is no “liquidity trap” but the overall price (say, the GDP deflator) has been decreasing since the late 1990s because of shrinking population and labor force: http://mechonomic.blogspot.com/2012/01/on-wise-monetary-policy-and-absence-of.html
Same is applicable to the US (the rate of participation in labor force has dropped by ~4% since 2001), Switzerland and some other developed countries. On the other hand, the level of labor force in the UK has been growing and thus inflation is larger than in the US.
Henry: “I am a mechanic”
No you are not. You are a fellow metaphysician, with a rival theory of his own. You are assuming that expectations must stay the same, since interest rates do not change.
Tell me: by what mechanism does a king get his orders carried out? (Assume I am a republican, and do not have the concept of a “king”.) Your reasoning is circular. You are assuming that people expect the King’s orders will be obeyed, and that each individual expects that if he does not obey the king’s orders, then other individuals will obey the King’s orders to punish him. I would laugh at your explanation, and accuse you of believing in fairies, or some sort of magical belief that one individual, who you call “King”, is a much stronger man than everyone else together. Which is obviously false, because we can all see that the King is just a man like everyone else.
What have interest rates got to do with monetary policy? Sure, changes in monetary policy may cause interest rates to change. But changes in monetary policy may also cause the price of peanuts to change. Why doesn’t the central bank just increase the price of peanuts, if it wants to loosen monetary policy? Just announce that the target price of peanuts will be higher. This is what Roosevelt did in 1932(?). Only he used gold, instead of peanuts. Because in 1932 people thought that gold was magic, like a king, and that peanuts were not magic. Nowadays a lot of economists think that interest rates are the magic king.
Could there be monetary policy in a world without interest rates? Or where usury laws kep interest rates fixed? Of course there could.
Interest rates aren’t the only possible king. We don’t have to choose a man who is mute to be king.
Sorry, Nick, but I don’t get your point. I mentioned (real) interest rates only once and in passing when describing the Krugman/Woodford mechanism, but I don’t see how this would be at the core (the “king”) of my argument. The core is the difficulty of a central bank to generate inflation expectations and thus discourage cash holdings, given that it has limited means to credibly generate a shortage of goods. Unless it buys goods (rather than assets) – such as peanuts – and therefore conducts fiscal policy.
To stay within your picture, in the trap the CB is a handcuffed, chained-to-the-wall Chuck Norris who tries to make folks move out of the room with the buffet into a dark, wet, cold cell by threatening to beat them up. I don’t see why they would move.
@ Nick Rowe
“Tell me: by what mechanism does a king get his orders carried out? (Assume I am a republican, and do not have the concept of a „king“.) Your reasoning is circular. You are assuming that people expect the King’s orders will be obeyed, and that each individual expects that if he does not obey the king’s orders, then other individuals will obey the King’s orders to punish him. I would laugh at your explanation, and accuse you of believing in fairies, or some sort of magical belief that one individual, who you call „King“, is a much stronger man than everyone else together. Which is obviously false, because we can all see that the King is just a man like everyone else.”
You’re going into deep philosophical / sociological territorry here: Why do we accept conventions although – ultimately – we know that they can be changed and are “just” conventions. Take the king: As long as enough people accept that he still is the king, your not following his orders means that you are beheaded – even more so if you say you are a republican (perhaps not in modern Britain, but take the middle ages). Your getting beheaded for not following the king’s order is what I would call a credible threat.
If you are a revolutionary you would have to convince enough people that now is the time to change conventions, behead the king and found a republic. This often means civil war and who will win is an open question ex ante. But in this particular case, changing the convention is not simple – it takes lots of conviction and blood. In general, although conventions are only constructions (which we as modern people know), they still are there and not easily changeable.
You can very well argue that debt is a convention and that you as a mortgage borrower simply don’t accept to pay service your mortgage but prefer to consume the money instead. This will of course boost growth. But lenders will not pleased. And as long as laws and especially police is backing the lenders, debt is a fact and deleveraging with its negativ impact on aggregate demand is also a fact.
On the other hand, going back to the king example: Historically over-indebtedness has been the number one reason for revolutions: Democratic Athens has been the result of fights over debt. And democratic Athens paid the formely endebted peasantry for their offices and their participation in democracy – which allowed them to have an income, spend it and expand the ancient Greek economy (however, the money for that was not created but extracted from Greeks colonies…).
Bottom line: Conventions can be changed but they are still facts of life you can not dispense with them easily, just by pretending that they are not there. So deleveraging will be going on and as long as the central bank is not willing to buy stuff directly and thus expand nominal output directly, its threat to expand nominal output is not credible.
I think the question to ask is, “what do the payoffs to actors have to look like to create self-fullfilling expectations.” Consider a firm making the decision to pull-forward spending on raw materials inventory to protect against rising prices. Say this actor is promised, by a “Chuck Norris” central bank, 4-5% inflation for two years, and 2% thereafter. His “downside” to foregoing inventory accumulation is essentially (roughly) 2-3% annualized for two years (the difference between higher promised inflation and the normal 2%). However, if accumulates inventory and the central bank is unable to deliver on its promise, the actor faces a “non-credible” central bank and the possibility of outright deflation. In this case, the downside is a sharp correction in prices: perhpas 20% or more in one year.
In other words, in a liquidity trap, the tail risk of severe deflation increases. A central bank promising a moderate and temporary increase in inflation creates an asymmetric payoff distribution: a small potential gain and a large potential loss from acting on the central bank’s promise. The way to change the payoff distribution is for the central bank to increase the tail risk of high inflation. It can do this only by significantly increasing or even suspending the long term inflation target. As Paul Krugman describes it, the central bank must appear to be “reckless” and risk the unanchoring of inflation expectations.
In short, I think expectations might be self-fullfilling. However, to stand a chance, they must create symmetric — or even inflation asymmetric — payoffs in order to accomplish this. Thus, there is a much higher risk that long term inflation expectations would be unanchored. There is no “free lunch”.
Henry,
Note that higher inflation expectations has successfully been used before to get an economy out of something like a liquidity trap. It happened with FDR in 1933. He talked up a higher price level extensively and then followed it up with devaluing the gold content of the dollar. Key to the story was FDR’s communicating his intent to increase the price level.
See Gautti Eggertson’s work on this period as well as the references found therein: http://www.ny.frb.org/research/economists/eggertsson/Great_Exp_AER.pdf
Also, I have some post that touch on some of your concerns.
(1) http://macromarketmusings.blogspot.com/2011/10/three-objections-to-ngdp-level.html
(2) http://macromarketmusings.blogspot.com/2011/08/how-would-monetary-stimulus-help.html
Thanks for the links, David.
I know the Eggertson paper, but I’m not sure I would put the 1933 episode into the same context as Gauti. As I understand it in 1933/34 the FDR abandoned gold, and the US was the last major industrialized economy to do so (Germany didn’t but introduced capital controls in 1932/33, which had much of the same effect). By un-tying the golden fetters FDR freed the dollar from overvaluation. Overvaluation was evident – as everybody else had already devalued – hence FDR’s act translated into immediate and strong depreciation expectations. These, in turn, translated into inflation expectations, which in turn allowed the economy to reflate.
In short, it seems to me FDR pulled a “Svensson” in the 1930s, not a “Woodford”. And it is not clear to me how one would pull a Svensson today (announce depreciation vis-a-vis the euro or the yen, economies that are close to a liquidity trap themselves? Depreciate against the renminbi, whose CB doesn’t allow it?)
[Somewhat relatedly, if I recall correctly the US ran large external surpluses in the 1920s - while in the 2000s it ran large external deficits. Thus, at least in the aggregate, the debt overhang cannot have been as grave in the 1930s as in the 2010s]
Btw, please don’t misread my post as an attack on NGDP-targeting. I am, as many, a bit on the fence with respect to the latter. I find the strong countercyclical component of level targeting very appealing, while I’m a tad concerned it would complicate expectation formation and thus weaken monetary transmission (everyone knows what wage growth to set with a 2 percent inflation target, it’s less clear with a level target). But my beef is not with level vs. inflation targeting, it is with the narrow technical question of how monetary policy could create inflation expectations in an entrenched liquidity trap.
Let me add an insight that I’ve gained from the discussion so far is – the most promising avenue may be for central banks to target debtors’ balance sheets and aim at shoring them up – either by trying to inflate debtors’ assets (houses), and/or by extending liquidity directly to them (a heli-drop targeting credit constrained folks).
Cheers,
HK
@ amv
About RE: I think it is not futile to think hard about expectations and also rational ones. However, in a world in which we do not agree about a common model but our actions and the expectations about our actions influence each other, there is no equilibrium. This is what Keynes nicely showed with his beauty contest. The Dybvig model is also nice in that respect: There are different equilibria according to the expectations of other people’s expectations – those equilibria can be good or bad. Here, rational expectations do not stabilise the outcome, but destabilise outcomes.
In such an environment the only thing that stabilises social interactions are conventions – conventions reduce the unpredictability of human action, thereby reducing the equiliria and the possible paths that the economy can take. In the Dybvig model, it the convention that stabilises outcomes is deposit insurance.
For Henry: http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/02/the-mute-king.html
(1) Raising the inflation target increases the expected profit on investment, even though there is a risk that the inflation may not materalize.
(2) Agree that devaluation isn’t an option for large countries.
(3) The problem with QE is that the central banks are buying safe assets. If they bought low quality bonds and stocks, that would help. Stocks have no “zero bound” limit.
Can a central bank always, always, without fail, generate inflation? No. Not unless it deliberately loses money – which is fiscal policy, not monetary policy.
Henry: Anybody can buy and sell peanuts. Anybody can buy and sell bonds. Anybody can borrow or lend money. Even you and I can do all that. That’s not where central banks have their power. If you start out by assuming that central banks only have power because they can buy or sell peanuts, buy or sell bonds, or borrow and lend money, then you don’t have a theory of central banks to begin with.
All other money is redeemable on demand into central bank money, and is the responsibility of the issuer to redeem that money for central bank money. Central banks control the time-path of the quantity of central bank money. What they choose to buy and sell with that money doesn’t matter (much), because anybody else can do that too. If the time-path of the quantity of central bank money doesn’t matter, then central banks are powerless under *any* circumstances.
@ Nick
Central banks control the time-path of the quantity of central bank money. … If the time-path of the quantity of central bank money doesn’t matter, then central banks are powerless under *any* circumstances.
I have no quibble with this, but fail to see how it would invalidate my argument – or better: address the questions raised in the article above.
It seems to me that in the liquidity trap the level of central bank money does indeed not matter much, and neither its time path, if the date by which folks expect to escape from the trap if far enough in the future.
Kings won’t help me resolve this, and not even queens. A transmission mechanism would.
Cheers,
HK
@ HK, Nick
I am sorry I couldn’t weigh in much. Time constraints.
But I think Nick is correct here: the central bank uses interest rates to communicate its preferred time path of central bank money (and thus, inflation). This communication channel is gone for now, so it needs other ways to do it. But I don’t quite see how an interest rate limitation should prevent it from doing so.
I view the expectations of people as a continuum: with unlimited asset purchases by the CB (corporate bonds, ABS, stocks etc.), you first scare off the gold buyers, next the home buyers, next the German investors, and so forth. But by scaring them into having inflation expectations, you set the wheel in motion. Some people start investing and buying, this raises asset prices and expectations of future demand, it lowers the value of the dollar and it raises the market clearing real interest rate, it relieves strained balance sheets and so forth. Then the wheel starts anew. How does an interest rate limitation prevent this from happening? Of course, you should support this with a level target of some sort, preferably an NGDP level target, but that was not what you were having trouble with, I know.
I don’t think your three arguments have done justice to this way of looking at the liquidity trap.
Finally, central bank money in a liquidity trap: of course, if you flush the system with temporary liquidity, it won’t matter. But make it permanent (and credibly so via a level target) and see how many people want to bet against you. I for one would consider it economically reckless (insane?) to bet on inflation against the only institution that can print money. Why would you bet against it?
@ Kantoos
with unlimited asset purchases by the CB (corporate bonds, ABS, stocks etc.), you first scare off the gold buyers, next the home buyers, next the German investors, and so forth. But by scaring them into having inflation expectations….
Why would this scare anybody into inflation expectations if there is no credible and strong transmission channel?
This is like a prison inmate telling the guards “I can break through prison walls, and once I’m out I’ll make sure you’re dead – so you better hand over the keys right now”. And to add weight to this claim he bangs his head against the prison wall until he’s a bloody pulp.
This would convince the guards to give him the keys?
I for one would consider it economically reckless (insane?) to bet on inflation against the only institution that can print money. Why would you bet against it?
I wouldn’t bet against anything – I would just ignore it, as long as I see no plausible mechanism by which the central actions create inflation. As I would ignore, at least for my investment and spending decisions, if the central bank governor shows up at the press conference with a mini skirt.
Which comes back to the article’s starting point: what exactly is the transmission channel?
@ HK
By “ignoring” it, you are betting against the CB, otherwise you would adjust your behavior… I think the best advice is: don’t bet against a committed CB that is willing to expand the amount of base money until it gets what it wants.
@kantoos: I also don´t understand where these inflation expectations should come from. Look at the Fed and imagine what the average person with interest in economics would think about it: They have increased the monetary base by ~ 300% without serious impacts on inflation, the exchange rate or inflation expectations. Why should another doubling do the trick? I mean, the financial guys say, “hey, the economy is still depressed, it didn´t work the last time, so why should it work now?” while some economists would add that without growth in broader money / credit aggregates(which we don´t see now), there will be no inflation. How do you change there minds?
@ Jay
one is tempted to think of the old “a hungry physicist, chemist, and economist with a sealed food can on a lonely island” joke.
The physicist tries to open the can with brute force. The chemist bathes it in saltwater. The economist states “I assume I have a can opener, and whoops…”
:-)
@HK: Thats the essence of it ;)
Actually, the central bank can create inflation: In asset prices. The bank is simply creating additional demand (QE, liquidity injection). This results in _lower_ interest rates. This may or may not kick the economy. The other side is combating Inflation, which generally works by ejection liquidity and raising requirements, which may or may not affect the inflation rate..
End of story.
The Central Bank is not and was never in control of the Inflation rate and therefore can not make any credible threats. The CB is, however, in control of liquidity…. by buying up the banks Assets or invest in them when nobody does (“lender of last resort”). Which makes it the best friend of the bankers. Thats all, nothing to do with inflation.
“The bank is simply creating additional demand (QE, liquidity injection). This results in _lower_ interest rates.”
Not if it works. If QE works stocks prices should go up, but risk-free bond prices should go down.
This is what I find crazy about central banks buying governments bonds. They are buying an asset that profits if the central bank fails. What message does that send?
To Nick’s point, it’s true that supplying liquidity is not a unique power of central banks. But it’s still a real power, and maybe the only power that matters in some cases.
I think I agree wih Kantoos here.
The point here is, to me, that if basically every single actor in the market does not believe inflation will rise, then and only then will it indeed not rise. And it must be the case even if, for instance, the inflation rate increases for some unknown, maybe unrelated, reason: everyone must firmly believe the CB will fail, and act accordingly. More realistically, the mechanism described by Kantoos (and by Yglesias some time ago, I believe) will take hold, and as some people start to believe, they start investing again, which increases the inflation rate a bit, and makes some other people believers, and there you go.
What seems interesting is that this mechanism, it seems to me, is simply a bank run.
Self-fulfilling expectations of people taking their money out of banks, created by some news on the financial papers, about the health of a bank or the coming of inflation.
>But make it permanent (and credibly so via a level target) and see how many people want to bet against you.>
“Permanent” supposes that enough people believe that at some point in time an event necessarily will occur. There are instances where people believe this and bet on it, even if their belief is wrong. Roulette is an example. Though there is no necessity for rouge ever to show up, they bet as if it were the case.
Similarly, one would have to show that people do spend as soon as they are convinced that the CB is in mode “permanent”.
If one cannot show this, it is just begging the question.
It would have been shown, if it had regularly happened.
In this case one does not need to know a transmission channel.
The magic of the CB is the transmission channel.
@Kantoos
Regarding “unlimited purchases”:
In a liquidity trap, a central bank engaging in QE will exchange a short term asset (Excess Reserves) for a risk asset. You argue that unlimited exhanges of this sort will create inflation. What if, instead of the central bank, the exchange is carried out by the fiscal authority? Say the U.S. Treasury offers to buy equities and corporate bonds in unlimited quantities. In exchange for these securities, Treasury will offer a s.t. asset — T-bills.
In each case above, the private sector ends up owning a s.t. asset; the public sector ends up with risk assets. In each case, the gains or losses from the larger public-sector balance sheet would accrue to taxpayers (all Fed profits/losses are remitted to Treasury and form part of the fiscal budget).
Given the above, would unlimited purchases of risk assets by Treasury lead to inflation? If so, would this be a monetary action, or a fiscal one? If its a fiscal one, then what is QE in a liquidity trap?
The monetary aspect of QE is a distraction. If central banks issued their own bonds instead of reserves, it wouldn’t change anything.
The key to being a “lender of last resort” (or “investor of last resort”) is having unlimited capital. (Technically central banks have very little capital, but it doesn’t matter since they don’t have to worry about being shut down by regulators).
@H.Kasper
“I won’t believe into the success of a policy as long as I fail to understand how it works. And if there are more folks like me out there, fighting the liquidity trap through inflation targeting is doomed.”
In a sense, I agree with you. Monetary policy works only if you believe in it. If you can manage not to believe in it, the central bank becomes powerless. And nothing will make this fundamental powerlessness go away. Without it, monetary policy would be mechanical. You seem to be asking for a mechanical explanation of a fundamental non-mechanical policy. No wonder, nobody can help you. (“mechanical” would be a policy that could work without people even knowing about it like Bernanke going out at night and buying goods and services till NGDP is on target. Even a heli drop is metaphysical. If people expect it to be reversed the next day, it won’t work.)
I like to think of the central bank as a coordination power. Let’s take your prisoner example:
“This is like a prison inmate telling the guards „I can break through prison walls, and once I’m out I’ll make sure you’re dead – so you better hand over the keys right now“.”
Imagine we are in a futuristic society, where each and every citizen can vote prisoners in and out, every day. If just one citizen (or say 10%) wants the prisoners out, he’s out. The voting is anonymous but our prisoner is a telepath. He knows what each and every citizen is voting, he knows their name, where they live, their family etc and he threatens every citizen, if they vote against him: “once I’m out I’ll make sure you’re dead.”
Now, how would you vote? If you vote him in but just one citizen votes him out, you are dead. Voting him in presupposes that you trust millions of citizen to do the same thing. But there is no coordination! The only guy in town trying to influence, to coordinate the citizens is our telepathic prisoner!
So, it’s not you betting against the central bank. It’s you betting that millions of co-money holders will do the same thing as you do, while the cb can actually influence them to do the opposite.
@ Alex F.
while the cb can actually influence them to do the opposite.
How……………….?
By having the central bank governor show up at the press conference with a mini skirt and twirl a hula-hoop?
[= in the trap about as effective as injecting base money]
P.S.: of course one anti-rational citizen would not suffice to get the “prisoner out” (=the economy out of the trap). This requires a sufficient number of folks with sufficient spending power to make desired I exceed desired S at the zero bound.
@ Alex F.
“Monetary policy works only if you believe in it. If you can manage not to believe in it, the central bank becomes powerless. And nothing will make this fundamental powerlessness go away.”
I disagree strongly. If you are a bank and you give out loans at a 3 % interest rate, increases of central bank intererest rate to 4 % mean that you make losses on new lending – and if maintained at this level for too long, you are bankrupt. Since a bank won’t give out new loans with this interest rate structure, real investment will fall and the economy will go into a slump. The central bank proves to be very, very powerful – and if you as a debtor or a bank don’t believe it, you are bankrupt.
However, it is not certain that decreases of short term interest rates will lead to more bank lending and more investment – especially if expectations about future profit rates of today’s investment decrease at the same rate or even faster than short term interest rates. The thing that investors have to believe in is that profits of their investment will increase relative to their financing costs – and the central bank has relatively little leeway of changing entrepreneurs expectations about future profits.
This is why the central bank is very powerful when it wants to slow down the economy – but not if it wants to restart it.
@ Alex F
“the central bank is very powerful when it wants to slow down the economy – but not if it wants to restart it.”
There is at least strong anecdotal evidence that CBs are good at braking the economy. That is according to some the main reason why competitive political systems with a universal franchise democracies need “politically” independent banks: the 19th century fear for universal franchise leading to waves of distributional effects (“taxing the rich”, playing santa claus with CB money etc) and especially for economically irrational socialists gaining power had a kernel of truth. In the seventies (before the information and transport revolution that unlocked the manufacturing potential of East Asia) there were severe inflation problems and only countries with independent central banks (or shadowing the currency of such a country) did well in controlling wage push inflation. Maybe that view has become unsophisticated but look at the journals from that period.
However, very little is known about how to design and implement a CB mandate (and operating rules/instruments) that can propel, rather than brake. In that respect I agree with you.
However again, the banking mechanism that you describe is not an adequate description of how modern banks interact with CB policy. They “lend” as much as they can based on models that aim at maximizing risk adjusted returns on capital. Given that many banks suffer from a share market price at which they cannot expand their capital enough to support existing loan demand when regulatory capital requirements increase (as is planned and likely to be implemented), the mark-up they require over their cost of borrowed funds (which varies ao according to that bank’s regulatory capital ratio and country of domicile and maybe also their short term rating (if different/lower)) is much higher than when they have surplus capital. The interesting question that we have right now is what is the interest rate that is relevant for neo-keynesian models: the rate at which the market provides marginal funds to your average weakly capitalized Eurozone bank plus a relatively high markup? Or the rate at which banks can borrow from the ECB plus a mark up that assumes banks being adequately capitalized and having a cost of capital at around their long term historical average?
And what can the monetary authorities control? Not the cost of regulatory capital and maybe not the rate at which banks can borrow from the market (which includes a risk markup)
Finally: as the weaker banks borrow more from the ECB (using marked-to-market collateral), they undermine their private creditors’ capacity to recover in the event of default and hence reduce their access to the private markets (where borrowing is unsecured except in the case of repos). The current financial system distress presents completely new challenges to whoever wants to use monetary policy via the interest channel.
@ Jay
Thanks for your late comment. If the Nebenwaerung could be used (only maybe) for tax payments, and assuming people would prefer to pay their tax with the Nebenwaerung, what would happen?
1. Gvt would have been paid. But gvt pays its expenditure with real money, so it depends on the bookkeeping….A creative gvt might be able to find solutions that are “better” optically (if the gvt has a subordinate position but benefits from asymmetric information, like most gvts in Euroland until the recent debt/deficit Hetze). Of course a Calvinistic gvt would simply increase the deficit and hence fiscalize. But so they would with the incidental tax cut, which might require parliamentary action.
2. Taxpayers have paid with free money (= equivalent to a one time tax cut but with different political and distributional effects/aspects) . Maybe this type of “covert fiscal stimulus would have different saving/consumption effects than an overt tax cut as per mainstream assumptions.
Of course I do not know, but given that I consider the real task of a politically rational (short term vote maximizing, not necessarily Machavellian and certainly not benevolent) gvt functioning under EU constraints to be oriented towards maximizing the apparent benefits for their own electorate, preferably at the expense of foreign gvts within the same union, maybe the Nebenwaerung would offer interesting perspectives. Anyone armed with a current ( for instance DSGE) model can design academically correct solutions that would satisfy mainstream economic experts. So that does not provide competitive advantage to politicians. The really creative ruler goes beyond that and operates with successful guile. That is probably why this would not really fly with supranationals, the ECB, etc.
But it is an interesting variant: instead of politically controversial tax cuts, issue a kind of food stamps (with an expiry date) that people can use to pay the gvt (and not save). I wonder if a covert incidental tax cut of this type would be less irrelevant than a standard one. Maybe not. Then there is lots of hope for populists.
Of cvourse your Nebenwaerung has other interesting aspects too. For instance the Bund could reprint Ostmarken, and allow the relatively poor residents of the fiscally poorer Laender to pay their taxes (maybe the incidental covert tax cut should be combined with an overt (but permanent) tax increase at the Land level) . The receiving Land would then redeem the Ostmarken at the Bundesbank of course. Maybe I am missing something here too
I agree, there are several problems regarding the estimation of the effects of such a policy in economic as well as in political terms. There could also be “misunderstanding effects”: The people don´t understand the process correctly, aren´t informed well or are driven by some kind of animal spirits(fear of a euro breakdown etc), which would lead to unforeseeable effects on consumption / investment spending etc. And I don´t think the implementation is politically feasible, because it could be seen as the beginning of the breakdown of the eurozone – exactly the impression Merkel and Sarkozy are trying to avoid.
PS: I have to thank for your elaborate comments :)
@ Alex F, Fali
My apologies, comments should have been directed at Fali, rather than Alex F. Maybe a small addition: traditional oligopolistic tendencies among banks are also much harder to control in the present environment than when banks have surplus capital. Another mar-up booster.
@ Rien Huizer
“In the seventies (before the information and transport revolution that unlocked the manufacturing potential of East Asia) there were severe inflation problems and only countries with independent central banks (or shadowing the currency of such a country) did well in controlling wage push inflation. Maybe that view has become unsophisticated but look at the journals from that period.”
You say it all: it’s about the wage push. Now, the wage push can NOT be controlled by central banks, but only by trade unions. Centralised trade unions in the Scandinavic countries, Austria, the Netherlands and Germany did effectively restrain from increasing wages at the same rate that decentralised unions in the UK and the US did. See here for more on that issue: http://www.people.fas.harvard.edu/~phall/PSGE_WP4_4.pdf
“However again, the banking mechanism that you describe is not an adequate description of how modern banks interact with CB policy.”
Yes, that’s right. Even the contractionary effect of central bank policy depends on bank capial AND on bank reserves. Banks with high capital and high reserves are much less responsive to changes in short term rates than banks that are just fulfilling their capital requirements. Still: High enough interest rates are likely to lead even well-capitalized banks into difficulties…
“The interesting question that we have right now is what is the interest rate that is relevant for neo-keynesian models: the rate at which the market provides marginal funds to your average weakly capitalized Eurozone bank plus a relatively high markup? Or the rate at which banks can borrow from the ECB plus a mark up that assumes banks being adequately capitalized and having a cost of capital at around their long term historical average?”
My answer: Dispense with the new-keynesian stuff altogether ;-). They assume that there is only one interest rate – and that this intererst rates influences consumption. In most models, there is no credit at all etc. I think those models led policy makers and the public astray.
“The current financial system distress presents completely new challenges to whoever wants to use monetary policy via the interest channel.”
I couldn’t agree more. That makes me even more skeptical about the ability of the central bank to increase economic activity. The only way – I think – is to give income to debtors – mainly households – that pay off their debts, thereby repairing bank balance sheets, which in turn makes rescue packages like QE less necessary, thereby giving private investors in banks more confidence.
@ FaLi/Rien
That makes me even more skeptical about the ability of the central bank to increase economic activity. The only way – I think – is to give income to debtors – mainly households – that pay off their debts, thereby repairing bank balance sheets, which in turn makes rescue packages like QE less necessary, thereby giving private investors in banks more confidence.
I agree also, but this seems to me a fiscal/regulatory rather than monetary task. With a household debt overhang, the most effective macro policy consists arguably of a scheme providing incentives for banks to write off household debts (possibly in part through debt-equity swaps that make banks benefit from future house price increases), coupled with a standing bank recap program and a temporarily expansionary fiscal policy.
By contrast, the effort to find – or dream up – a one-size-fits-all monetary policy solution, even though monetary policy’s effectiveness is severely impaired by the very debt overhang itself, looks to me like an unhelpful distraction in resolving the crisis.
Best,
HK
HK,
Yes. Maybe the problem is that no one knows how to deal with the twin private sector millstones right now: (1) the millstone of household debt (especially against a perspective of business cycle lethargy (there is a bit of an RBC sympathizer in me -not a concept useful for policy but certainly non-negligible). (2) the millstone of financial system distress.
We can find equitable and reasonably efficient solutions for the public debt problems (some seignorage, some public sector austerity -of course assuming a sympathetic monetary policy and maybe a non-docrinaire look at intra-EUR sovereign credit risks, potential for sharing and overall reduction). And that is a problem purely in the hands of policymakers, as long as they can either control or satisfy markets.
But there is no way to deal with the private sector problems that does not involve some form of involuntary transfers between households (which means violating private property rights, a cornerstone of our Rechtsstaat). That is difficult to design and even harder to execute. The literature on the economics of transition is filled with concepts that seem to be relevant for policymakers (like status quo bias). I think that one has to start with the banks, especially now the most urgent liquidity problems have been sedated. The banking problems are EUR-wide and some deal with bank shareholders (to make them accept that shrinking balance sheets is not in their long term interests, only in the private interests of of risk-averse managements) that rewards shareholder patience (but not an outright unconditional subsidy). Only then one can think of re-/overcapitalizing the banks and look at solutions for the private debts. Ideally banks should be prepared to lend today to borrowers who they thought acceptable five years ago (and whose situation has not changed significantky) . Gvts might reward banks that do that and punish ones that do not.
@ fali
Yr point re the role of cooperative/responsible trade unions well taken. Have seen that at first hand. But (at least in the case of The Netherlands), the simple fact that the CB was independent, respected and had a credible policy of maintaining the HLD/DM in a very narrow zone (backed up by very large FX reserves and an iron grip on the financial system (capital markets controls for instance, a new issue calendar, ban on non bank issue of tradable short term debt, no unregulated near banking etc) gave unions and employer organizations little scope for anything but cooperation (not that the workers were happy: there were instances of rogue unions emerging that did not want to cooperate, but those never gained large following).The NLG/DM link basically put the monetarist Bundesbank in control of Dutch monetary policy. Inflation is of course a monetary phenomenon, but syngergistic policymaking makes the job of both the CB and cooperative “social partners” or rather “elites”a lot easier. This despite the fact that in the early seventies Holland had a political economy problem later known as the “Dutch Disease” An enormous supply shock due to first the discovery of very large and easily exploitable natural gas reserves (with appr 75% of the revenue going to the gvt and second the rapid increase of energy prices of 1973. Which led to a very rapid increase in gvt and gas infrastructure spending in a very open economy. But the cost push inflation happening at that time was much less violent than in the US at the same time. Where the pre-Volcker CB was far from independent. The same in, eg, the UK and Australia.
So I guess that an insulated, tough and credible CB may well be a necessary (but not sufficient) condition to preempt/discourage destructive union policy and thus enabling social cooperation (with the occasional rhetorical noise) as per the “Rhineland” political economy model. The result is similar to paternalism but without the dictatorship.
The reason it’s hard to create inflation expectations when the central bank buys securities is the knowledge that it can extinguish the cash by selling the securities at a later date. So the solution isn’t difficult. Give away cash. I think giving everyone a few hundred to a few thousand would do the trick. The point is that somewhere between giving everyone $0 and $10 million (and much closer to $0, obviously), there is no possible way that inflation will take hold. No possible way whatsoever.
My bad. There is no possible way that inflation will NOT take hold if the central bank gives everyone enough cash. Just don’t give it to the banks – the banks know that the central bank is going to ask for it back (under the current system), so they hold it; waiting for the Fed to ask for it.
@fali, kasper
We are talkin past each other. If someone puts a gun to your head, “your money or your life”, this is not “mechanical”. It involves expectations. If you don’t believe he will shoot, you won’t obey. (Since cbs control m but no v (NGDP=m x v), there are always expectations involved. Last but no least they only control base money and as Huizer explains, banks are primarily capital constrained.)
Now, Kaspar’s problem is, he thinks the threat is not credible, because he doesn’t see a gun. Expectations alone can’t be enough, you need a supporting mechanism to be credible. I think the main supporting mechanism is the ability to create money. But imagine a central bank who can print an infinite amount of money but has to bury all of it deep in the ground and everybody knows that, could this cb raise NGDP only through expectations? Here Kaspar is right. If you don’t have a credible story how the money goes into NGDP, it won’t change expecations.
Kaspar’s challenge is this: tell me a credible story how a cb can raise inflation expectations only by buying assets.
Having read Nick Rowe again, I think both are right. Rowe’s story isn’t about inflation expectations, but about Tobin’s q and the wealth effect. You raise asset prices and firms will have an incentive to produce new assets raising NGDP. Higher asset prices repair balance sheets, people feel reacher and are going to spend more, again raising NGDP.
Tobin’s q and wealth effect, that’s the mechanism.
Only for the record, I’m the guy who favors heli drops, but not because I’m a “mechanic”. I want monetary policy to separated from banking. Helicopter drops/vacuum cleaner operations done by a new institution with no need for collateral. (Together with the abolishment of deposit insurance and government debt. That’s my holy trinity :-) Trying to raise asset prices by buying stocks is a terrible policy. If you create inflation expectations through heli drops, asset prices will go up too. That’s a much better way to do it.
@ Alex
sounds interesting, but I’m still not fully with you. Why would the wealth effect make people spend in a significant manner as long as they consider cash an excellent savings vehicle (as it buys them more bang for the buck some time the road)? If also believe to recall also a related discussion in the 1930s (the Pigou effect that argues the inverse though – deflation would increase the value of fixed income assets and make folks wealthier) with the result that any possible wealth effect would likely be too small to help in an entrenched liquidity trap (Eclair or Fali may help us here).
But this would seem an interesting and worthy undertaking: write down a model that convinces mechanics like me that there is a credible transmission channel that does not depend on ad-hoc and internally inconsistent assumptions about expectations.
And also for the record: I am already half convinced, thanks in large part to the discussion here. A “heli-drop” that targets credit constrained households (=those folks whose balance sheets are impaired) should work. Give cash to folks who used to spend, would like to spend, but who can’t spend because of a budget constraint imposed by concerns about their debt servicing ability. Loosening this budget constraint should be almost as effective as when the public sector spends itself. In fact this is akin to a government-paid debt bailout scheme that comes for (almost) free, as the government (through the central bank) makes use of idle resources. On the downside it’s a bailout and creates adverse incentives, and one needs to balance carefully the downsides of this against the upside of getting the economy faster out of depression.
All this suggests to me that one needs to think through very carefully what precise policies could work in a debt-overhang created liquidity trap and why. Just stating that “if the central bank creates unlimited amounts of base money all will be fine” means ignoring the liquidity trap rather than resolving it, and this is unsatisfactory.
Cheers,
HK
@ Axel F :), HK
Two quick shots from the hip :) Tobin’s q, yes, in theory I suppose that might work. In practise, what we observed in the past years, certainly in Germany, appeared to be a variety of “distorted Tobin’s q”, in that buying bubble-economy assets which saw their capital values rise continually (further bid up by those very foreign funds flowing in) was a lot more lucrative – at least in the short run – then investing in physical capital. Sinn e.g. has been arguing that because German financial investors got burned in those markets, they will now look for more reliable investment opportunities “at home”, and that therefore a golden decade is coming :)
On a related note, the wealth or (reverse) Pigou effect certainly did play a role in the run-up to the crisis, but again the question is if you really want a repeat story. Check out the Economist from 2002, “The Houses that Saved the World”:
“Last year America’s corporate sector suffered its deepest recession since the 1930s, with a massive plunge in profits and capital spending. Yet, despite rising unemployment and lower share prices, consumers continued to spend. How come? One explanation lies in the surge in house prices, where the Fed’s interest-rate cuts have certainly worked their magic.
Many economists were worried last year that the wealth loss from falling share prices would force consumers to cut their spending. Even after the recent recovery, American stocks are still worth 25% less than two years ago. Yet, as falling share prices made some households feel poorer, rising house prices have made many more feel richer. Over the past year average house prices in America have risen by 9%, their fastest-ever in real terms. Although American households as a whole have more of their wealth in equities than in housing, a relatively few rich people hold the bulk of the shares. For most people, housing is by far their largest form of wealth. Two-thirds of Americans own their homes, and gains in the value of those assets have encouraged them to keep spending.”
http://www.economist.com/node/1057144
And we all know how that worked out for everyone…
To be fair to the Economist, they ended their piece on:
“The lesson which consumers—and also many over-sanguine economists—have to learn is that spending cannot outpace income for ever. House prices have saved America and the world from a deep downturn, but they do not remove the need for consumers to take care over their balance sheets. Homes are only as sound as their foundations.” Amen.
@ Alex
“Since cbs control m but no v (NGDP=m x v), there are always expectations involved.”
Well, that’s what the textbooks tell you – unfortunately it’s wrong. Money is endogenous – banks give out credit and only look for reserves later. The central bank tends to accomodate reserve requests since not doing so would threaten the economy’s financial stability. In reality, loans create deposits which in turn lead to the creation of base money. But the central bank sets the interest rate for which it offers its reserves and here it has an effect (although muted by what Rien Huizer rightly pointed out).
So the central bank can NOT credibly commit to drown the economy in base money. It has not pistol! – except for what HK rightly pointed out: if the CB gave money to endebted people who could pay off their debt and spend the rest – then you get the economy going again. Under current (political) circumstances, this is not going to happen.
Further, the equation you cite is also not the whole story: it would have to look more like PT = v * M, where T are ALL transactions – not only those on newly produced goods (appearing in GDP), you have to add both transactions of already existing goods (e.g. things traded via ebay or already existing houses) and financial transactions. This is why P is not only the price for GDP goods but also prices of financial assets and used goods. So even if the central bank could (which it can not) give money directly into the economy (e.g. via a helicopter drop), people could also buy financial assets or only buy on ebay.
You argue that Tobin’s q and the wealth effect could lead to higher real spending, even if new base money is only used to buy assets (which is the case to some extent). Perhaps. But – as HK – I’m very skeptical that this is gonna happen under current circumstances. Tobin’s q should work because equity prices should reflect higher expectations about future firm profits. If however everybody knows that higher equity prices only reflect “artifically” higher prices and not possible higher future profits, firms should not expand their investment spending. By the way, the same applies to higher equity prices due to equity buyback – those buybacks are less likely to boost real investment but CEO pay – because their pay is linked to equity prices but not to real investment.
As far as households are concerned: The most important part of the wealth effect is housing since most households in the US own a house but not equity. However, the CB can hardly affect housing prices in a depressed economy in which people do not buy houses because they have to pay back mortgage debt.
So, neither does the CB has the power to influence the money relevant for the economy nor are rising asset prices likely to lead to more spending. I don’t want to say that there is absolutely no effect – things would certainly look worse if the central bank did nothing. But I am highly skeptical of the central bank’s power to get the US back on track alone.
@ Eclair
You did it again ;-).
For clarification concerning the question at issue we should clearly distinguish between the „classical” liquidity trap where people can spend but are not willing to do so and the “debt-overhang created liquidity trap” where they are willing but not able to spend.
The latter is a state of affairs where no expectations are involved.
A heli-drop of cash does change this state of affairs and suffices for an increase in spending.
The former is due to expectation and can be overcome by CHANGING expectation only.
This needs a mechanism, if the approach is to be considered scientifically sound.
In the case of the classical liquidity trap a heli-drop of cash will not induce changing expectations that lead to sustainable spending. Wealth effects are most likely, but are not sustainable.
I agree with what Éclair said.
Henry,
Let me introduce you to Scott Sumner:
“Indeed I know of no case where a central bank that wished to boost inflation and/or NGDP was unable to do so. ”
http://www.themoneyillusion.com/?p=7960
If all else fails a central bank could always start monetizing government debt or use a “helicopter drop” e.g. by financing a temporary tax cut with newly printed money.
Really, there are unlimited possibilities for a central bank to create inflation. And the easiest is probably to just use a self-fulfilling prophecy: If Draghi or Bernanke were to credibly anounce tomorrow that they would apply QE and monetize sovereign bonds or use other aggresive policies until they reach 5% NGDP growth or 5% inflation or X% unemployment, then inflation expectations would be set without requiring the FED or ECB to actually do anything.
@ JL
I wonder whether you bothered to read more of the article than the headline.
You wanted a physical transmission channel for inflation expectations.
I gave you one: the central back has the power to create inflation or deflation, hence it also has the power to control inflation expectations. It doesn’t get more physical than the (electronic) printing press.
But perhaps you aren’t a mechanist after all?
Feel free to ignore anonymous me. But Scott Sumner, Nick Rowe, kantoos and myself are all making the exact same point which you seem to ignore:
That trick that Zimbabwe pulled with the hyperinflation: every central bank can pull that same trick if they want to. Therefore no central bank is ever powerless when it comes to creating inflation or inflation expectations.
Monetizing government debt and spending: a foolproof inflation creator since ancient times.
@ JL
Monetizing government debt and spending – yes, this is foolproof, but it goes beyond pure monetary policy. This is fiscal policy financed with the printing press. If fiscal and monetary policy team up they can overcome one another’s (possible) constraint: lack of funding for fiscal policy and the liquidity trap for monetary policy, respectively.
This is Keynes’ point, of course, but in my perception not Sumner’s or Rowe’s. The latter seem to insist that monetary policy can somehow always and everywhere create inflation expectations on its own. To me this looks like ignoring the liquidity trap more than presenting a solution for it.
Cheers,
HK
True, Sumner and Rowe call it “asset purchases” instead of government spending.
But that’s mostly semantics, the following two scenarios are – ultimately – equivalent:
A) The central bank prints money to buy assets.
B) The central bank prints money to buy bonds from the government and the government uses the money to buy goods and services.
I described (B), Sumner and Rowe describe (A).
Assets are a particular class of goods. But Nick already said that “assets” could be expanded to include all goods the economy produces if necessary.
Using bonds as an intermediate step doesn’t change much. It provides a fictional separation between “fiscal” and “monetary” policy. But this separation is just an abstraction; it’s not real.
A bond held by the Fed is a bond that has been temporarily monetized: the interest flows straight back to the Treasury, like all Fed profits.
And if a bond owned by the Fed matures then the principal also flows back to the Treasury: debt monetization becomes permanent.
Zimbabwe doesn’t even make an effort to maintain this fictional separation:
Bank prints money -> government thugs spend printed money.
Do you consider Zimbabwe to be engaged in monetary or fiscal policy?
@ JL
in my view it makes a big difference whether the public sector buys assets or goods. Sufficient goods purchases create a goods shortage and therefore consumer price inflation. Observing potentially unlimimted goods purchases, rational folks woud have every reason to expect inflation.
Asset purchases can’t achieve this – unless the asset is a government bond, and the fiscal authority cooperates by spending the receipts.
As for Zimbabwe, it’s not clear to me how this is relevant here. Is or was Zimbabwe in a liquidity trap?
“Asset purchases can’t achieve this – unless the asset is a government bond, and the fiscal authority cooperates by spending the receipts.”
This is the essence of your concern, but it seems obvious that this is not true.
The central bank could buy business stocks and bonds causing the stock market to rise and companies to have more money. The central bank could buy mortgage securities leading to cheap, readily available mortgages and a rise in real estate prices. They could by commodity (e.g. agricultural) futures leading to a rise in prices for those commodities.
As for Zimbabwe, do you think that a liquidity trap could have reduced their ability to create inflation?
Or do you agree that the printing press is both a necessary and sufficient tool to manage inflation?
@ H. K.
The liquidity trap is the premise of the discussion.
I haven’t seen here anybody who has offered a convincing mechanism by which to get out of it.
A question:
Has anybody ever seriously questioned that the liquidity trap happens because of the expectation of falling prices and instead came up with a different explanation to think about?
I don’t want to say of course that people won’t hoard money, if they expect falling prices.
But I wonder, whether they must have this expectation, if they hoard.
@ JL
Just imagine you own a good X and your neigbour a good Y. Each of you paid EUR 1000 for their good.The CB decides to buy goods. Of course it cannot buy all kinds of goods and it uses a criterion “portability”. You and your neighbour bought the same garden bench but yours was anchored permanently to your house and your neighbour’s not. Your neighbour gets (because the CB wants to give a very strong signal) 20K EUR for his portable bench. You get nothing (but you keep your bench). I guess that approach would not be problematic?
This is of course abstracting from the real possibility that in EURland the CB (the ECB) uses a criterion taht the asstes from countries closest to disinflation will have priority. You live on the border between Austria and Italy and your neighbour in Italy hits the jackpot. He sells his bench for 50K and then pays you a visit (brings a bottle of very good Italian Merlot) and expreses his sympathy. He offers to have your bench cut loose en and pay you 5K for it. He claims he needs a replacement bench. You understand that he has found a way to sell TWO benches to the ECB (you have prejudices about Club Med people) and demand a lot more. He p[oints out that he can go to the local garden bench store and buy a new one for 1K. What he really wants to do is let you share in the bonanza from Frankfurt, because he likes his neighbour.
What will you do: call the EUR police? Help your neighbour digging?
I often wonder why the Central Bank, if it wants to cause some inflation in a liquidity trap, doesn’t just PRINT money. Literally. Not add to the monetary base by bond purchases, but by printing up money and then giving it away.
This would have certain advantages. First of all, you don’t accumulate a ton of bonds or other instruments that you will eventually have to dump. Secondly, you can target it geographically at areas where you want some inflation. Currently, for example, you want some inflation in Germany, to facilitate the adjustment of the price levels in the periphery. So what you do is you print up coupons and mail them to Germans. They are redeemable by any business offering consumer goods or services. (Mailing a coupon instead of cash or a cheque is intended to encourage spending rather than saving).
The ECB redeems the coupons by creating money.
The inflationary effect of this will be concentrated in Germany, the measure can be stopped at any time and it will certainly be more credible and effective in raising inflation expectations than ordinary open market actions or whatever the ECB does normally.
Plus, it’d be way more popular in Germany than any other ECB measures to fight the current crisis that I can currently think of.
Very great post on something I was bewildered myself for a while after reading Krugman etc. However, I must say that at least partly i was persuaded by point 1. So the better question (and something which could be modelled) is following: “What if iquidity trap never ends? Or what if its lenght is beyond time horizon of all agents?” I would argue that this is not a case of USA nowadays, but it might have been the case of Japan. Krugmans original arguments were that the have liquidity trap because of declining population. If so, then that was prospect for Japan for forseeable future, thus monetary policy is powerless for way too long for the agents to matter.
@ All
I thought that was an interesting post on Bernanke.
http://blogs.ft.com/gavyndavies/2012/02/26/the-metamorphosis-of-ben-bernanke/
@ Eclair
An interesting post indeed.
The insight for the discussion here:
>In the US examples of 2003/04 and after 2008, unconventional policy was intended TO PREVENT deflation from occurring, and so far it has succeeded in doing so.>
At times it may be difficult for a CB to determine whether the economy moves toward deflation and thus by misjudgment it may ease monetary restraints too much. Perhaps one could argue that this was the case in 2003/04.
However, the issue for monetary policy is avoiding the liquidity trap, rather than trying the obviously impossible of getting out of it.
Accordingly, emphasis in economics should be much more on this aspect.