« Do wage and price cuts increase (real) aggregate demand? | Main | The April Labour Force Survey numbers weren't all that good »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Nick,

OMO works by changing the level of excess reserves.

I’d be willing to bet that the long term correlation between the level of excess reserves and the level of the central bank policy interest rate is zero, in the case of both Canada and the United States, in normal non-QE environments. I would define this as pre September 2008 for the United States and up until recently for Canada.

On the other hand, I’d bet it would be possible to build a meaningful correlation function between changes in the central bank policy interest rate and the path of changes in the level of excess reserves during the time period around changes in the central bank policy rate, for both countries.

That implies to me that the true instrument of conventional monetary policy in both countries is the policy interest rate rather than the monetary base. Perhaps this is uncontroversial.

However, I think this would remain the case using NGDP futures targeting, in normal non-QE environments. In other words, I think the idea of excess reserves, or in a larger sense the monetary base, being an instrument is a myth or illusion in the case of NGDP futures targeting in normal non-QE environments. The effective instrument would remain the policy interest rate, where the market is now the FOMC determining that policy.

In a related way, I would guess there’s a chance NGCP futures targeting would fail operationally at the zero bound and/or with quantitative easing. Given the absence of the normal anchoring relationship between changes in the policy interest rate and the effect of OMO, the market would struggle enormously to determine the amount of quantitative easing required to stabilize the NGDP futures level. Because of this, it would become dysfunctionally volatile, I think.

Might an NGDP futures intrument trade somewhat like the Canadian dollar - eg. moved around by oil prices and subject to speculation? How would the Bank extract a signal from such market noise?

JKH:

One of the reasons I did this post was to break Scott's proposal down into its component parts.

Suppose for example that you disagreed with his OMO/base money instrument, and believed that the overnight interest rate was the only real instrument, and that monetary policy could only work (if it did) via that instrument. But you liked the rest of his plan. Then you would change the instrument to the overnight rate, and change the instrument rule to something like: "For every $100 bet that NGDP will fall below target, cut the overnight rate target by one basis point".

In fact, this is part of the topic of Scott's post on Greg Mankiw: http://blogsandwikis.bentley.edu/themoneyillusion/?p=1199

If it really is true that a short term interest rate is the only real monetary policy instrument, and that increasing the stock of base money can only work insofar as it lowers the overnight rate, then at 0% overnight rate, that instrument rule would hit the lower bound before it managed to raise the equilibrium NGDP futures price to target. If this fact were known, there would then be an infinite demand for bets on the NGDP futures market (as there would be if the Bank tried to target a futures price for $2+$2 at $5). But is an infinite stock of base money compatible with NGDP coming in below target, even if the nominal rate of interest were stuck at zero?

I'm not 100% sure if I understand you correctly on the correlation question. Are you saying:
1. There is no correlation between level of excess reserves and level of overnight rate.
2. There is a correlation between changes in excess reserves and changes in the overnight rate?

Or are you saying that changes in the overnight rate cause (are followed by subsequent) changes in reserves, rather than vice versa?

Anyway, this gets us back into the whole debate about the monetary transmission mechanism, as in my "mechanical metaphors", "gold price instrument", "medium of exchange" posts. My view differs both from yours and from Scott's, on these questions. You emphasise the interest rate, Scott the unit of account definition of money, I the medium of exchange definition of money.

brendon: If NGDP really were the target, then the single best indicator I can think of would be the NGDP futures price itself. If the NGDP futures price fluctuated relative to the true rational expectation of future NGDP, then there is a problem, of course. But it's not the same as using the exchange rate as the only indicator of future CPI. Because we know that, even if forex traders are rational, there are a lot of things that could cause the exchange rate to diverge from the rational expectation of future CPI. Real exchange rates change, when oil prices (terms of trade) change, for example, as you say. The Canadian analogy would be using the CPI futures market to target future inflation.

“But is an infinite stock of base money compatible with NGDP coming in below target, even if the nominal rate of interest were stuck at zero?”

No (i.e. I agree). Scott has made that point as well. My point is that I think it’s going to be very difficult for the market to find an equilibrium level of reasonable stability between zero and infinity.

“Are you saying:

1. There is no correlation between level of excess reserves and level of overnight rate.
2. There is a correlation between changes in excess reserves and changes in the overnight rate?”

Yes to both. Although for changes in the ON rate, changes in excess reserves are a round trip path, netting to zero.

“Or are you saying that changes in the overnight rate cause (are followed by subsequent) changes in reserves, rather than vice versa?”

No.

Nick: "But is an infinite stock of base money compatible with NGDP coming in below target?"

Answer: sure. We wouldn't be using money, we'd be using another kind of currency. Cigarettes perhaps.

As Cochrane keeps saying, nothing in economics rules out nominal explosions. Here's Sims making a basically similar point in a different context.

http://sims.princeton.edu/yftp/Targeting/TargetingFiscalPaper.pdf

I might surprise some people here, but I'm going to agree with JKH's point here:

"However, I think this would remain the case using NGDP futures targeting, in normal non-QE environments. In other words, I think the idea of excess reserves, or in a larger sense the monetary base, being an instrument is a myth or illusion in the case of NGDP futures targeting in normal non-QE environments. The effective instrument would remain the policy interest rate, where the market is now the FOMC determining that policy."

If most traders are Keynesians, (and it seems to me that in the real world that is how lots of people think about monetary policy), then I think they might focus on the fed funds rate, or other short terms rates, as their preferred policy indicator. This may seem odd, as the market would actually be determining the base, not the ff rate. But that's where the daily trading comes in--each day they get another shot--get to see how the base setting from yesterday influnced the ff market. Thus over a period of several days they can gradually nudge the ff rate in the desired direction. But I should also say that this is pure speculation on my part. They may well also look at stock and commodity markets, and lots of other indicators when deciding whether to go long or short.

I am still not sure I fully understand the meaning of terms like 'instrument' and 'short term target,' under different policy regimes. So I'll keep an open mind on this debate.

Where I slightly disagree with JKH is the zero bound. I think a zero bound would be massively unlikely under NGDP futures targeting, especially a 5% target. The reason rates fell to zero last fall is precisely because NGDP growth expectations plummeted in the fall. Banks don't usually lend money at zero percent when they expect 5% NGDP growth. So while I agree with JKH that the market would have a tough time determining the appropriate base in that environment, I think it is very unlikely to occur. BTW, if we did hit a zero bound, the market might also look at the other indicators that I often discuss, such as stocks, commodities, TIPS spreads, etc.

This also relates to my one disagreement with Nick. Nick presented the plan very fairly and accurately, but was mistaken in his assumption that the "wisdom of the crowds" issue was the key question at stake here. I did think that in 1989. But that assumes we are comparing NGDP futures targeting to a Svenssonian discretionary rule. Svensson argued the Fed should always adopt a policy stance expected to hit their target. But even the Fed admitted after October that NGDP was likely to come if far below their preferred level. Thus by far the most important advantage of NGDP futures targeting is that it would force a Svenssonian policy stance--force monetary policy to target the forecast. I would still support my proposal even if someone convinced me that the Fed could forecast NGDP 10% more accurately than the market--that's how big an issue this is for me. The loss of credibility last fall had far more damaging economic consequences than simply missing an NGDP target. It is when markets suddenly EXPECT NGDP to fall sharply below target for an extended period (1929, 1937, 2008) that we are in big trouble.

I haven't had time to look at rebeleconomist yet, but let me just say that short term rates would become more volatile if, and only if, the market felt that the extra volatility would help make NGDP less volatile.

AdamP, I favor NGDP targeting in dollar terms. So if the dollar became worthless, then NGDP would be infinite. Cigarettes might become the medium of exchange (although I don't seriously think this would happen), but not the medium of account for these contracts. I could say lots more about the "usefulness" of indeterminacy models, but I've already gone on way too long.

Thank you very much for doing this post Nick, and also thanks for all the useful comments from everybody. BTW, for those who think future targeting is "pie in the sky" I could also add that the greatest benefit of doing these thought experiments might be that it helps us to see other issues (like liquidity traps) in a new way.

Scott,

“I think a zero bound would be massively unlikely under NGDP futures targeting, especially a 5% target.”

This also may surprise some people, but I fully take your point here.

Of course, I didn’t really argue against this; I just presented a conditional, as I think you acknowledged, without exploring the probability attached to such a conditional under your proposal. I don’t claim to fully understand the proposal in all its detail yet, so I haven’t really formed a final opinion on it. But I do think your point as represented in the above quote is extremely important to the value of your overall argument. It seems logical to me that if targeting nominal GDP is a good idea, and if such a futures market is an effective mechanism for fully engaging Fed policy together with market expectations around this idea, then the zero bound would be a less probable event - perhaps not impossible, but less probable. Qualified by those “ifs”, that conclusion just seems intuitive and powerful to me. And I can visualize it working through the proposed mechanism of the futures market in the kind of conditions that we had last fall, because there would have been a massive one way bet that would have mechanically forced the Fed to flood the base and allowed the fed funds rate to drop much more quickly, which would have at least reduced the probability that policy would end up getting stuck at the zero bound. If I’m not mistaken, I think that’s been your point, more or less, all along.

Speaking of cigarette economics:

Prison Economy Spirals As Price Of Pack Of Cigarettes Surpasses Two Hand Jobs [NSFW]

http://www.youtube.com/watch?v=K2IYIJc1f00

Scott: "AdamP, I favor NGDP targeting in dollar terms. So if the dollar became worthless, then NGDP would be infinite. "

Well, if the dollar was no longer in use then NGDP would be undefined not infintite (something like the difference between 0/0 and 1/0, although neither of these is a number and neither is infinity so Scott's statement is still nonsense). NGDP measured against the cigarette numeraire might well be below target (probably would be).

Just to clarify: in the paranthetical I'm saying that infinity is not a number. I'm not saying 1/0 is not infinite.

OK Adam, NGDP would approach infinity as the dollar's value approached zero. BTW, wouldn't money always have some scrap value? Pennies are zinc. Federal Reserves notes might have collectible value, or use as fuel. If you are assuming I don't take the risk of cigarettes as money seriously, you are right. My fear is that cigarettes will be banned.

JKH, Again, I think you have it exactly right. I do think the key benefit of NGDP targeting would have been its effects before the severe crisis became obvious in October. I think traders would have seen increasing risk of recession in the late summer, and aggressively sold NGDP futures, sharply expanding the base and reducing rates. If this policy was successful, expectations might (and I emphasize might) have stayed strong enough to prevent falling to the zero bound. So I fully agree with your points, and agree that my slight criticism in my previous response wasn't really directed at what you wrote, but more at inferences that readers might have drawn--as you did not claim a liquidity trap would have occurred, just that if it did it would presented some added complexities for the market.

But Scott, can't we agree that the fed just might, maybe, stop the program before it had bought up all outstanding treasury debt? Or, more importantly, isn't in likely that agents in the economy might think that?

Just to continue from my last comment, let's assume we can agree that the answer to the last question in my 2:49 above is yes.

Now, Scott has claimed that his futures targeting scheme makes the liquidity trap non-existent (here's a quote from Scott in a comment on Nick's mechanical metaphors post: "remember that NGDP futures targeting would immediately eliminate the liquidity trap").

The key here is that Scott is claiming the trap never happens in the first place. The basic condition of the liquidity trap is a negative full employment real interest rate, this condition is determined by the full emplyoment consumption path and has nothing to do with money. So how is it true that the futures scheme eliminates the liquidity trap? Well, presumably because the monetary base would automatically adjust to a level that breaks the trap. Then, anticipating sufficient monetary expansion, agents never feel the need to reduce expenditure in the first place and thus the negative real rate doesn't result in deficient AD.

Thus, eliminating the liquidity trap rests on two interconnected things. First, their must be some path for monetary base that breaks the trap with absolute certainty. Secondly, to prevent the trap from springing in the first place agents must believe, before the fact, that the fed will allow that path to be followed (and not step in and abandon the futures scheme).

On the first point here is Scott responding to comments on his own blog: "There is no serious economist in the world that thinks a yen dollar rate of 1000 would have been ineffective in moving Japan out of the liquidity trap a few years ago."

This statement has a problem. The problem is that, as stated, it's not true (unless you want to claim Lars Svennson is not a serious economist!). Here's Svennson (http://www.princeton.edu/svensson/papers/Tokyo509.pdf) explaining that it takes much, much more than just the FX devaluation to break the trap, you need the currency devaluation in conjuction with a price level target and you have to maintain the devaluation until the price level target is hit. Notice that the monetary expansion must continue past the point where the trap is first broken and agent's must believe the expansion will continue to even break the trap in the first place.

Now, this is important because it pertains to the second point, the point about agents a priori beliefs. The monetary expansion needed to break the trap may be truly huge and very long lasting, or at least the market might believe it would need to be truly huge and long lasting. Couldn't it be that agents would reasonably worry that a monetary expansion big enough to break the trap would be so large that the fed would abandon the futures scheme before the trap was broken? After all, a permanent yen dollar rate of 1000 would be quite unpleasant for the Japanese. I stress that what the fed actually would do is not the issue, it's what agents believe before the fact.

And there's the issue, if agents, a priori, doubt the fed will allow the monetary expansion to continue until the trap is broken then they will still reduce their expenditures. We end up in a situation where the futures scheme is engineering a large expansion of the base but the trap springs anyway. The level of base blows up with no real effect (at least not right away). In that case maybe the fed does step in, validating the original doubts.

Just to pre-empt Scott's counter argument I'll add my next point, related to the one just above.

In order for the futures scheme to completely "eliminate the liquidity trap" requires more than just the fed staying the course no matter what levels the monetary base reaches. In order for agents not to want to change their ACTUAL expenditures the agents themselves must have 100% confidence that the market will get the right base. Is that really likely?

After all, in the market there are just as many people doubting whether the stock market is sensible after the current rally as there are doubting if the fed will do the correct thing. How much doubt would it take for people in the real world to cut back on discretionary spending, even just for a while?

Really, the reason I'm so sure that the plan wouldn't work as Scott describes is because the logic of the plan is absolutely, 100% identical the new Keynsian Taylor rule type policy plan. Scott presumably would deny this but it's true.

The logic of the NK Taylor rule policy scheme goes something like this:

We have a Phillips curve that relates current inflation to expected future inflation and the output gap. In theory the Taylor rule also eliminates the liquidity trap by stabilizing expected inflation.

If expected inflation falls then this changes current inflation (via Phillips curve), thus rates fall more than one for one, so real rates fall. Thus current inflation is increased back to target and the original expectation is shown invalid. In the model, of the central bank is commited to the rule then aggregated demand is completely stabilized.

Now, compare the logic of Scott's plan to the NK logic. In NK we have the causal chain:

change expected inflation causes a change to intrest rates causes a change to current inflation which re-establishes intended expectations.

Scott's plan:

change expected NGDP growth causes change to monetrary base causes a change to current NGDP growth whic re-establishes intended expectations.

The logic is identical with inflation = NGDP growth and interest rates = monetary base. Now we can argue over transmission mechanisms (rates vs base) or the target (inflation vs NGDP growth) but that is not what I understand Scott to be arguing. He has said several times if you don't like the NGDP growth target like he does then fine, target inflation.

But how can he be so sure his plan will "eliminate the liquidity trap" when the Taylor rule doesn't? Agents might quite reasonably trade in the future but still have enough doubt in the ability of the market to hit the target (perhaps because they simply don't believe monetary policy has any effec in the liquidity trap) that they reduce their own expenditures.

Remember, it's not about whether monetary policy actually works in the liquidity trap, it's about whether agents believe it does before the fact.

I am sorry if this question has been answered somewhere already; if so, please refer me to the answer:

What asset would the OMO use and how would the price be set (eg by auction)?

It strikes me that this is key to whether there is a liquidity trap or not. I have long thought that the issue of monetary policy implementation is an important omission from academic study of monetary economics - Krugman does not seem to get much beyond "t-bills". Hopefully, present events will change that.

Adam:

I am trying to re-state your point above in my own words, and my own way of thinking. Tell me if I have got substantively the same point, or if I am off in a totally different direction:

Suppose Scott targets 5% NGDP growth. Suppose the LRAS curve is moving right at 3% per year. Suppose the natural real rate of interest is minus 3%. Then we have a problem, because to keep on the LRAS we would need 2% inflation, but this would require nominal interest rates of -1% to have the real interest rate at -3%. And we can't have negative nominal interest rates.

Am I making the same point, or a different one?

(Probably a different one. I need to re-read your above two posts more carefully, and try to wrap my head round it. My head is working slow these days.)

By the way, further to my comment at 07.25, I have long suspected that Fed market operations (in particular coupon passes) contributed to Greenspan's conundrum by unnecessarily adding demand for long-term treasuries. Now that the Fed has accepted the idea that they can use the asset side of their balance sheet to target easing, perhaps they should review the effect that their market operations had during the boom.

Rebel: From what I have read of Scott's view, I think he would answer that it doesn't really matter much what the central bank buys in an OMO, and how it does it. Scott sees the monetary base as central, because the monetary base constitutes the unit of account, and all other monetary aggregates are redeemable for the monetary base. So OMO is just a mechanism to increase or decrease the monetary base. The central bank could buy or sell bicycles, antique furniture, in principle, and it wouldn't make much difference.

Scott will no doubt correct me.

On your general point, I tend to agree with you. We keep saying "We are in a liquidity trap, where nominal rates are zero", but we have lots of different interest rates, and a lot of them are nowhere near zero. And the biggest problem with current macro models, the "Neo-Wicksellian" ones, is that they only have one interest rate, and do not model the stocks of money and bonds, and so have nothing to say about interest rate spreads, and how those spreads might change if the stocks of various financial assets changed through being bought or sold by the central bank. And I find it very hard to believe that those spreads would be exogenous with respect to the central bank's actions.

Changing the stocks of those assets in public hands should affect the spreads, if they are not perfect substitutes; expectations of future monetary purchases and sales likewise should affect spreads; and expected future AD should affect the spreads.

This was a point I was trying to make a few months back, here http://worthwhile.typepad.com/worthwhile_canadian_initi/2009/03/the-return-of-monetarism.html and here http://worthwhile.typepad.com/worthwhile_canadian_initi/2009/02/is-lm-and-two-wedges-understanding-the-second-wedge.html and elsewhere.

Or are you arguing something different?

Nick,

Yes, I am arguing something similar to you (I like your wedge explanation by the way), except that I would broaden it out beyond interest rates.

An interest rate is just another way of expressing the price of debt, although there are other substitutes for money - eg gold - for which the price would be conventionally expressed as a rate of exchange against money. If base money is supplied by purchases that are so highly diffused as to have an insignificant effect on the prices of the assets bought, then increases in money supply can be expected to have the greatest effect on the prices of money's nearest substitutes, with a progressively attenuated influence on the prices of more distant subsitutes. This may not be the most efficient way of stimulating activity. Alternatively, targeting purchases on particular assets can address particular problems - including boosting bike manufacturing if desired! Readers may find my wife's pictorial representation of this idea amusing.

The links in comments seem to have been disabled. The url to which the last sentence of my comment at 09.44 refers is: http://reservedplace.blogspot.com/2009/05/pictorial-comparison-of-qe-in-japan-and.html

test

Yes, posting comments has been a little strange for me too, recently. Sometimes it won't appear, and I have to re-post. But the original link worked for me, Rebel. I think bob and Adam have been having trouble too, judging by their "test" comments.

Nick,

That's adifferent point that applies generically. If the bank targets 2% inflation but the full employment real rate is -3% then we have the same problem. Althoug, that might be an argument in favor of a Taylor rule that looks at GDP and inflation seperately since then we'd see real GDP falling short and lower rates, allowing inflation to go above target.

My point is a theoretical one about endogenous expectation formation. The logic of how Scott's plan would avoid the liquidity trap in the first place is identical to how NK-Taylor rule models avoid the traps. NK-Taylor rule models also have the theoretical property that AD is completely stabilized by controlling expectations, it's just that none of models designers is bold enough to believe that it would work so flawlessl in practice. Thus, it seems reasonable to ask if Scott's plan might be subject to the same deficiencies. I'm trying to articulate how it works, that's not easy. Woodford also struggles with this in his book and my critique of Scott is identical in principle to Cochrane's critique of NK models.

I'm not, btw, arguing the idea is a bad one. I'm trying to point out that it's not really very different from the NK-Taylor rule type models and it fails to solve any of the problems with that class of policy design.

Rebel: Yes, I agree. We are talking about all asset prices, and all goods prices too. For example, we can think of stock prices as an interest rate (either the earnings yield, E/P, or the dividend yield) but when we hear the words "interest rates fall" we don't normally think "stock prices rise", but it's really exactly the same thing as bond prices rising. And for assets like gold, bicycles, houses, bulldozers, and real investment goods, we find it even harder to think of their prices as being the inverse of an interest rate (like the ratio of house rents to house prices) but that is exactly what they are.

I like your wife's cartoon, as expressing one particular monetary instrument. But I would draw all the glasses on a flat table, with little pipes connecting them all together. So it doesn't matter where you pour the liquid in, *eventually* all the glasses will reach the same level.

That "eventually" is the Quantity Theory insight. But why doesn't it all happen at once? Why are there short-run "Cantillon effects", which depend on where the money enters?

"God invented time to stop everything happening at once" (was that Einstein?)

We have sticky prices and wages to stop everything happening at once. But we draw a very sharp distinction between "goods prices" and "asset prices". "The rate of interest adjusts instantly, but the price level adjusts slowly". That's wrong. Houses are assets, but their prices seem to be short-run sticky to me. Why not other assets too?

And it's not just sticky prices that stop everything happening at once. There is a delay between my getting new money and my spending it. (If there weren't this delay in spending money, we wouldn't hold stocks of money at all, when money is rate of return dominated; but we do.) So the flows of money won't happen instantly. So all the glasses don't reach the same level instantly. Which means it does matter where the new money enters the system, in the short run. Which means it matters what the central bank buys.

Adam: thanks. I understand you better now. But yes, it's not an easy point. I need to think it over.

Nick,

Nice idea to have the glasses in a line! In that case perhaps the viscosity of maple syrup would be needed to represent the slow progress of levelling. Mmmn!

I agree with you about asset and goods prices, and would make the same point about the measurement of inflation. But even central bankers like to be liked, and goods are what people buy while assets are what they own!

AdamP, I have to admit that I am not worried about the Fed having to buy up all the Treasury debt. But if they felt that, for political reasons, they could not, then they would obviously have to stop futures trading, or else incur huge capital losses. But I see that case as being so unlikely that it is not even worth worrying about. My own view is that there would be nothing wrong with buying up the entire stock of T-bonds if the public's demand for cash was that great.

BTW, your hypothetical could apply to any regime. What if under a gold standard the Fed had to buy up the entire world stock of gold. Would we be willing to spend that much? Is that the sort of hypothetical that would be relevant when deciding whether to adopt a gold standard? I doubt it.

A liquidity trap is a negative full employment real interest rate? So if the nominal rate is 20%, and the expected inflation rate is 21%, and we are at full employment, does that mean we are in a liquidity trap? I've never seen that sort of definition before. In any case, I assumed the term 'trap' referred to monetary policy ineffectiveness. So that's my definition.

I have a new post today bashing the Taylor Rule. I don't see my plan as having any relationship at all to the Taylor Rule. The Taylor Rule uses an interest rate instrument, which becomes ineffective at zero rates. My policy does not. More importantly, I have a forward-looking policy, whereas the Taylor Rule is a backward-looking policy. That is a crucial difference.

I wouldn't be so sure that Cochrane would agree with you; he told me (by email) that he liked my futures targeting idea. That's not to say he endorsed it, but he didn't see any reason why it wouldn't work.

Svensson never denied that a 1000 yen exchange rate would spring a liquidity trap. Maybe he would deny it, but I doubt it. I would never argue that a modest depreciation in the yen would definitely spring the trap. And obviously if the BOJ returned to deflationary policies some time after the 1000 rate was established, then Japan might fall back into deflation. Anything is possible. Thus if the new path was 1000, 900, 810, 720, etc, Japan would clearly slip back into deflation, I agree.

Rebeleconomist, Nick's right, what you buy is a second order issue.

Nick, First let me say that the scenario you present is extremely unlikely in my view. Negative 3% real rates is pretty weird for an economy with a positive 3% real growth rate. But if it did occur, then I presume that owners of every single asset with more than a negative 2% expected real rate of return would sell those assets to the Fed. A better option might be to set the trend rate higher than 5%, if the Fed actually thought there was any likelihood of such an equilibrium. BTW, if such a possibility is a big problem for my proposal, it would be a massive problem under the current regime, far more difficult to deal with than under futures targeting.

Scott,

Nick said that you would say what the central bank buys is a second order issue, which is not quite the same thing as Nick saying it himself!

Anyway, aren't Cantillon / limited participation effects the classic reason that monetary policy "excites" activity?

Scott, on this blog, as on your own, you need to read the comments in order to address them intelligently. I said a negative full employment real rate is the basic condition of the liquidity trap. The whole rest of my comment is a discussion of whether or not the trap actually happens.

Rebel: yep, but I think Scott meant to say that Nick was right about what Scott would say. (Did I say that right?) ;)

I'm not sure I would say it myself, at least not in the short run. Glasses and maple syrup.

By the way, here's a *Canadian* economist writing today in my local paper, who above all people ought to be using your maple syrup analogy, but all he says is the water is viscous! http://www.ottawacitizen.com/Business/Inflation+flood+build+your/1586707/story.html

Rebel: follow-up. Actually I would say that sticky prices/wages are the main reason money has real effects in the short run. Cantillon/limited participation/lags between getting and spending new money effects (are these all fundamentally the same thing??) will also cause real effects in the very short run.

Scott says: "And obviously if the BOJ returned to deflationary policies some time after the 1000 rate was established, then Japan might fall back into deflation. Anything is possible. Thus if the new path was 1000, 900, 810, 720, etc, Japan would clearly slip back into deflation, I agree."

Now suppose that the 1000 yen dollar rate was established and from day one everyone EXPECTED the BOJ to return to deflationary policies as soon as the trap was broken. Suppose the EXPECTED path on day one was 1000, 900, 810, 720, etc. What would happen then?

Actually, on subject of whether or not Scott's plan actually fights recessions here's a question. I stress that this has nothing to do with what I was saying above...

Suppose that we are operating under Scott's plan. Furthermore, suppose we do find ourselves with a negative full employment real interest rate. The monetary base undergoes a huge expansion and so far NGDP growth isn't falling.

Now, is it that the real rate is negative?

Suppose we reach a point where exactly half the population is 100% sure that the monetary expansion thus far is not yet sufficient and trades in the futures market, betting that NGDP growth will be way below target.

However, on the other side, the other half of the population is 100% sure that the monetary expansion thus far will induce a massive inflation and bets equally in the futures market that NGDP growth will be way above target.

Suppose that both sides trade in equal amounts and so actual monetary base doesn't change. Now, those that fear a great depression will reduce discretionary spending as much as they possibly can. And that's half the population remember.

On the other hand, those that fear hyper-inflation probably don't increase expenditures. They just hold euros or gold.

So, with half the population reducing their demand might we end up in a recession with deficient AD? How would Scott's plan prevent it?

sorry third paragraph should be:

Now, is it obvious that the real rate is negative?

Rebeleconomist, I agree with Nick's later point, the way currency is injected isn't that important, it has real effects because of sticky wages and prices.

AdamP, You may think your questions are clear, I certainly don't. I have no idea what you mean by 'basic condition'. Are you saying negative full employment real rates are a necessary conditon, a sufficient condition, or both? Whatever your answer, I disagree. I don't think it is necessary or a sufficient condition.

Regarding your point about the Fed buying up the entire stock of debt: You asked me to admit that such an outcome was far-fetched, that in the real world the public wouldn't expect the Fed to do something so extreme, and hence the policy would have no credibility. Then you proceed to give me one hypothetical after another that is far more far fetched:

1. Suppose base demand was so high the Fed had to buy up the entire national debt.
2. Suppose the Japanese yen went to 1000, but then was expected to suddenly appreciate after hyperinflationary depreciation.
3. Suppose the public's inflation forecast wasn't the normal bell shape, but a double hump (camel hump).

I don't doubt that you can come up with all sorts of bizarre hypotheticals where my plan might produce strange results. I'll make your life easier by giving you another. Why rely on the camel hump idea? Suppose most people think NGDP will grow by 5%, but it will actually decline by 5%. Sure, if people are that weird, then NGDP futures targeting won't work very well. Do you have any actual, plausible, real world scenario where the idea will not work?

I still think you focus too much on interest rates. Depreciation of the yen will raise Japanese prices Period. End of story. (as Krugman likes to say.) It will do that regardless of interest rates, regardless of expectations. As long as the Japanese government keeps depreciating the yen they will continue to get inflation. If you ask me what happens if they stop, what happens if they don't do the policy I recommend, then yes, the inflation will stop. What is so weird about me proposing a policy that will work if they stick to it, and won't work if they abandon it. Has there ever, in all of economic history, been a policy where that was not true? What if the discretionary regime we have today goes back to the real bills doctrine next year? Won't that be bad? Is that a reason not to have a discretionary regime today?

You may have a point, and it may be obvious to you what your point is, but it isn't obvious to me. Before you call my responses "unintelligent," you might work on improving your questions. When a single question is spread over many different separate posts, it isn't always easy to follow.

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad