New Keynesians believe that Say's Law is false. But they want the central bank to try to make Say's Law look true.
New Keynesians believe that real business cycle theory is false. But they want the central bank to try to make real buiness cycle theory look true.
New Keynesians believe that the loanable funds theory is false. But they want the central bank to manipulate liquidity preference to try to make the loanable funds theory look true.
New Keynesians believe that the economy has zero self-equilibrating tendencies. But they want the central bank to make it look like the economy has powerful self-equilibrating tendencies.
New Keynesian model + New Keynesian central bank = "classical" economy + random error (because no central bank is perfect).
New Keynesian model + non New Keynesian central bank (like one that sets interest rates on a whim) = godawful explosive or implosive mess.
New Keynesians are engaged in a terrible secret conspiracy to try to make the economy look very different from how they think it really is. Except it's not terrible, and not secret.
Simon Wren-Lewis and Paul Krugman maybe don't need my help; but they have got it anyway.
"Why [is mainstream New Keynesian economics] brackish? Because it has retained the nonsense of marginal productivity distribution theory while discarding the foundations of Keynesian economics. The essence of Keynes’ economics was the liquidity preference theory of interest rates and rejection of the claim that price and nominal wage flexibility would ensure full employment. New Keynesians abandon both. They replace liquidity preference theory with loanable funds interest rate theory and they use price and nominal wage rigidity to explain cyclical unemployment."
No they don't. But they want the central bank to try to make it look like that.
(Which is not to say I think New Keynesian macro doesn't have big problems. But these aren't them.)
Worthwhile Canadian Initiative 
Nick!!… your first two lines are two of my favorites (You, DeLong, Sumner and Sadowski on #1, and Christensen on #2). The other two are new to me. This is great… now I can just store a link to this one page. Thanks!
“New Keynesians believe that the economy has zero self-equilibrating tendencies. But they want the central bank to make it look like the economy has powerful self-equilibrating tendencies.”
It sounds a little like using feedback to stabilize an unstable system.
Here’s the theoretical classical economy (first 12 seconds) left on its own. Here’s the actual economy with no central bank. Here’s the economy with a theoretically perfectly run central bank. Now instead of the computer deciding how much voltage to apply to accelerate that little cart, imagine that the Board of Governors of the Federal Reserve System (or any other CB leadership) do that instead: that’s the real economy with a real central bank. 😀 … I’m joking!… kind of.
A good post, which incidentally would make a good starting point for a critique of certain central bankers, particularly J-C Trichet. What’s wrong with these guys is that they think Say’s Law really is true, so they don’t have a clue how to make it so.
I cannot believe how poor Galbraith and Black’s response to Simon are. It’s like marking an exam written by a clever student who has completely failed to address the essay question.
Luis: I am trying to mentally reverse engineer what James Galbraith and Thomas Palley think a New Keynesian model looks like. And it’s nothing like what an actual NK model looks like.
Tom: I beat you to that one! my old post on inverted pendulums
If you want to understand, in a nutshell, New-Keynesian policy advice, look at where all macroeconomic theories start:
(gamma) dc/dt = i – pi – rho,
i is the interest rate pi is inflation, rho is a discount rate and gamma tells how strongly people change consumption with interest rate changes. (I’m leaving out precautionary terms from uncertainty.) If we integrate this DE we find two terms, the integral term and a term that tells us how consumption asymptotes at infinite time. In the New-Keynesian world, people focus on the integral and assume consumption will return to trend asymptotically. Then, a real interest rate that is too high drives consumption down today, and since the Fed sees a zero bound, the real rate can get stuck at too high of a level. This observation goes along way to explaining New-Keynesian policy advice.
For example, higher inflation lowers the real interest rate so anything that boosts inflation is viewed favourably – it lowers the real interest rate in the integral term and shifts consumption from the future to the present. New-Keynesian policy advice favours fiscal stimulus because higher inflation is good when confronted by the zero lower bound with a real interest rate that is too high. Pointless public works projects, which lowers the productive capacity of the economy all help! It goes on to explain the whole “managing expectations” open mouth policies by the Fed as well.
But all this comes from effectively ignoring the asymptotic term in the solution of the DE. If instead we use permanent income logic on the second term, we see that the problem might be long run declining productivity, and this observation gives us a completely different interpretation of the facts. What if all the disincentives from social programs that keep expanding and crony capitalism (rent seeking) that comes with excessive regulation is lowering the long run productivity? Then, the New-Keynesians are focusing on the wrong term in the solution to the DE! Long run supply problems will also lower today’s consumption. Today’s consumptions is not just set through all expected future interest rates, but also where we think consumption will be in the far off future. It’s this second term that is largely ignored by today’s economists (economists can be captured by their speciality, too).
So, is it the zero bound that is leaving the economy with interest rates that are too high, forcing up consumption growth and thereby depressing today’s level; or do consumers think that long run productivity will decline, and so consume little today because they expect to consume little tomorrow?
In economics, just like in physics, you need to explore the boundary conditions fully – especially the ones at infinity!!!
Avon: (Are you now, or have you ever been, John Cochrane? 😉 )
I am following you there. (But what does “DE” stand for?)
Fiscal policy works in NK models, not by raising the level of G (increasing G(t)), but by reducing the growth rate of G (cutting dG/dt), thereby raising the natural rate of interest (that interest rate compatible with C(t)+G(t)=Y(t) for all t, where Y(t) is “full-employment” output). I have made this point before. You are adding that increases in G(t) may reduce future Y(t), which could reduce the natural rate of interest today.
I think there is a bigger problem with NK models. There is nothing in NK models that ensures that C(t)+G(t) will asymptote to Y(t). Even if the central bank sets the right interest rate. It is just an assumption that is bolted on, that is not motivated by the model.
Yes, there are big problems with NK models, and policy advice. But the paleo-keynesians who are criticising them here don’t understand them at all.
Agreed about other problems and paleo-Keynesians (which is why I flip out every time I see ISLM logic to explain anything). DE is short for differential equation.
Yeah, I’ve read John Cochrane – his textbook is excellent (his treatment of the effect of a time varying risk premium and time series predictability is breathtakingly awesome). I read this stuff for fun, it’s not my day job. I think John Cochrane made this point about NK models sometime ago, but you see the logic of it beautifully in the first couple of chapters of his textbook on asset pricing. I really think we need much more work in economics that bridges the divide between financial economics and macroeconomic theory.
Avon: “But all this comes from effectively ignoring the asymptotic term in the solution of the DE.”
If the paleo-keynesians actually had a glimmer of understanding of NK models, they would focus right in on exactly that point. They would be (correctly) accusing the NKs of simply assuming that the economy converges to full employment in the long run, even when there is no mechanism in the model to get it there, even if the central bank is not stupid. The paleo keynesians could simply cream the NK’s, if they wanted to. Instead they bumble about in some imaginary 1950’s world.
Here’s me, pretending to be a paleo-keynesian, making that accusation about the asymptotic term.
Nick, you did indeed beat me to the pendulums… that was a nice video, highlighting the movement in the opposite direction first. But did you beat me to chimpanzee riding on a SegWay?
Also, have you seen the tripple inverted pendulum? The initial movement of the cart is prominent there. I’d like to see them balance a whole chain on end. 😀
“Avon Barksdale”… wasn’t that a character on “The Wire?”
But paleo-keynesians are nonsensical for so many other reasons. We need to start with fully recursive formulations respecting Ricardian equivalence to get anywhere. Without thinking in terms of dynamic programming, it is impossible to understand macroeconomics. To me, the permanent income hypothesis and the monetary foundations of inflation are as fundamental to economics as the discovery of the uncertainty principle was to quantum mechanics. Yes, this makes economics harder – we can no longer write down simple DEs with fixed points and declare victory. We need recursive motivations. No one said that the world would be easy to understand.
There is a wonderful parallel in physics. In the 1950s very few physicists knew anything about group theory. Group theory is pretty abstract stuff and it didn’t seem all that relevant to most of what was going on in physics. That all changed from the late 50s onward. You can’t do any physics any more without a rather in depth knowledge of group theory. The same needs to apply to economics programs – I see too many undergraduates (nearly all of them) who come to work for us who know little or nothing of dynamic programming. When I ask them how they could have studied macroeconomics without dynamic programming, they start by drawing ISLM plots. I tell them to forget nearly everything they’ve learned and then lend them my copy of Sargent and Ljungqvist.
Avon: then I’m screwed. Because I can’t do math. Ah well, I’m old, so it doesn’t matter. (And I do see people who can do math say some really daft things too!)
But look. Please read my paleo-Keynesian post that I link in my comment immediately above. I made the math extremely simple, so even I could understand it. Am I not getting at an important point there?
Avon,
“we see that the problem might be long run declining productivity, and this observation gives us a completely different interpretation of the facts.”
Yes, lowering the asymptotic natural rate is bad. And if you want to study that more carefully, there is nothing (other than mathematical complication) that prevents anyone from strapping a more complex supply side onto the NK model. But you don’t want to do that in a pure RBC or “New Monetarist” framework that effectively ignores the non-automatically equilibrating behaviours of the short rate which is critical for understanding the Fed’s control over the real rate and the short and medium term dynamics of convergence to optimality. Also, it highlights the additional liquidity trap risks of supply-side-damaging (natural rate lowering) government interventions. It also affects the set of available asymptotic solutions.
“So, is it the zero bound that is leaving the economy with interest rates that are too high…”
Why can’t it be both? Even if raising G is unhelpful, a higher rate of inflation would be great. You wouldn’t know that from an RBC model. In fact you wouldn’t even know how to raise the inflation rate if you wanted to (see e.g. Williamson).
“In economics, just like in physics, you need to explore the boundary conditions fully – especially the ones at infinity!!!”
Maybe. In reality, nobody has a clue what lies out beyond 50 years, and most of our DSGEs are way to sensitive to boundary conditions 1000 or 1000,000 years out. Higher private discount rates tend to “tame” a lot of the irrelevant future dynamics, so rather than worrying about distant civilizations we should be worrying about credit constraints and declining concern about future generations. Cochrane’s obsessions, in particular, with the asymptotic trajectories of the NK model are basically angels on pinheads. The problem of choosing the asymptotic boundary conditions of a DSGE is not particular to the NK framework. Every RBC model has to make a similar set of assumptions. (And if you are working within a ratex framework, the fact that the economy has not already exploded is all you need in order to know that the asymptotic behaviour is bounded.)
Karsten: good comment. But: “(And if you are working within a ratex framework, the fact that the economy has not already exploded is all you need in order to know that the asymptotic behaviour is bounded.)”
I disagree. The fact that the economy has not already exploded is a fact that needs to be explained. Why don’t we observe “black holes”, where the economy implodes into hyperdeflation? If the theory cannot rule them out, yet we don’t observe them, that is a puzzle that contradicts the theory.
Nick, is your C(t) the same as Avon’s c (when he writes dc/dt)? Also, what are you beefs with the paleo-Keynesians? Also, dumb question here maybe, but Avon’s rho = the discount rate: for example, would the Fed’s overnight rate be an example of this, or is this a different concept?
Karsten:
I don’t think that the issues I talked about mean that there is something fundamentally incorrect with NK. I just think that focusing on one term and not the other skews policy advice. I also think that Cochrane is not simply obsessing about asymptotic trajectories – this is actually academically relevant and not just angels dancing on the head of a pin. Let’s use a bit more permanent income hypothesis logic here.
I freely admit that I am not an expert on DGSE models or NK models. I spend my time in math finance, I just read as much as I can of this stuff on the side to get a big picture idea of what’s going on. But this is what I do know. I find it distressing that modern macro focuses on issues, like “let’s add a touch of eye of newt here and toe of frog there” for monetary policy. We are entering a period in central banking of full discretion with cults of personality. This is nonsense and dangerous. Use our modelling to generate insight, but stick to a Taylor Rule or something similar and forget all this, “let’s just tweak inflation a bit” etc. Let’s take Kydland and Prescott a bit more seriously than we do. Activist monetary policy will not absolve the government of its fiscal sins or remove the sand it throws in the gears from excessive regulation. The idea that the central bank should act as some giant social engineering shop frightens me tremendously. Sooner or later we will pick the wrong DGSE model or some other such thing, trigger a depression like the Fed did in the early 30s and then we will shrug our shoulders a say, “Oh well, how could we have know better?”.
By the way, I apparently can’t type. DSGE, somehow my fingers like DGSE better when typing. Good thing I don’t actually put the trades through – my fat fingers would cause the next financial crisis!
Tom: “Nick, is your C(t) the same as Avon’s c (when he writes dc/dt)?”
Yes. Though Avon’s c is the log of my C(t), but that doesn’t matter much.
“Also, what are you beefs with the paleo-Keynesians?”
In this particular case, because they don’t know what NK economists are saying, and get it totally wrong.
“Also, dumb question here maybe, but Avon’s rho = the discount rate: for example, would the Fed’s overnight rate be an example of this, or is this a different concept?”
Totally different concept. Avon’s i is (roughly) the Fed’s overnight rate. Avon’s rho is our psychological impatience.
Nick, thanks! I was going to ask about the log thing: are the other rates logs too? (e.g. i)… so that you’re adding logs of numbers like 1.02 (say for a 2% rate), rather than multiplying them (I realize for small rates it doesn’t make much difference: adding 0.02 to 0.03, etc is approximately correct)? “discount” Whew!… that makes sense.
Can you give an example of an economist who’s a Paleo-Keynesian?
So here are the NORMATIVE problems with New Keynesians:
“New Keynesians believe that Say’s Law is false. But they want the central bank to try to make Say’s Law look true.”
OK, I suppose this might be desirable… but maybe it isn’t desirable. I think I’d like to see some normative argument as to whether we want it to be true or not. Arguably we want to have a general glut of most commodities (buffer stock! reserves! savings!), for example.
“New Keynesians believe that real business cycle theory is false. But they want the central bank to try to make real buiness cycle theory look true.”
This is definitely not desirable. A world where real business cycle theory is true is a nasty nasty world. Why would you want to try to make it appear true?
“New Keynesians believe that the loanable funds theory is false. But they want the central bank to manipulate liquidity preference to try to make the loanable funds theory look true.”
This is definitely not desirable. A world where loanable funds theory is true is one which underallocates money in a depression and overallocates it in a boom, systematically. Why would you want to make it appear true?
“New Keynesians believe that the economy has zero self-equilibrating tendencies. But they want the central bank to make it look like the economy has powerful self-equilibrating tendencies.”
OK, this is desirable. Self-equilibriation is great when you can do it, such as with a railroad train on railroad tracks (angled wheels). It doesn’t seem to be possible to make the economy self-equilbriating, of course — too many positive feedback loops.
But you see my point. Why try to make these things true when they are not desirable?
Nathanael: “A world where real business cycle theory is true is a nasty nasty world.”
Why do you say that? Ditto for loanable funds.
Start with a perfectly competitive Walrasian economy, with perfectly flexible prices. It is hit by shocks to technology, preferences, and resources, but responds efficiently to those shocks. The first theorem of welfare economics applies in that world. That is the world of (the simplest) RBC models, and the (simplest) loanable funds model.
The New Keynesians want to central bank to try to replicate the allocation in that model as close as they can. Imperfect competition (and imperfect information) means they can’t do it exactly, of course.
“Let’s use a bit more permanent income hypothesis logic here.”
The problem is that for infinitely lived agents with the PIH everything becomes about what happens in a billion years. It’s nonsense. This is why the boundary conditions matter so much and why anybody who worries about the exact details of the asymptotic behaviour is totally missing the boat. The point is to figure out how exactly to break the PIH in those models so as to remove the dependency on asymptotic expectations that we know aren’t relevant in the real world. This means risk premia, credit constraints and caring less about future generations than about ourselves.
The problem with any ratex DSGE (NK or RBC) solution is that the asymptotic behaviour is imposed as a boundary condition, so a trajectory cannot change stochastically from probably bounded to probably divergent (just like in most physics problems). It’s 100% one or the other. To deal with that we need models with solutions whose asymptotic behaviour is not uniquely determined by boundary conditions. (Think of a square root diffusion that can with finite probability disappear into an absorbing boundary (singularity) or not). My point is that a) high private discount rates seem like part of solution for making the asymptotic boundary conditions irrelevant and b) any analysis that hinges critically on boundary conditions at infinity is getting something critically wrong about what matters to current economic agents and what stabilizes the competitive equilibrium.
For example, an economy can have an asymptotically unsustainable non-Ricardian fiscal policy, but whether it is or not is determined by structural parameters of the economy that are assumed to be fully known to agents because ratex. If they are fully known then the economy cannot transition from sustainable to unsustainable, because the transition isn’t possible under ratex. For such a transition to be possible agents that are rational Bayesians without God-given prior knowledge of the asymptotic behavior might work, but it’s impossible under ratex. The usual neoclassical assumptions (which lead to things such as the PIH) are simply vastly to heroic, and aren’t even an acceptable first order approximation when trying to answer a lot of very important questions. Our first task is to take Kydland and Prescott a lot less seriously than we do.
Nick,
“The fact that the economy has not already exploded is a fact that needs to be explained.”
In practice, I agree with you. But in the NK model a divergent solution can’t be stabilized by a Taylor rule even temporarily, so any difference from a stable equilibrium would be quickly amplified and run away. So you can’t have a long period of stability followed by an explosion. With any reasonable error the solution would diverge very quickly. My comment only relates to the usual ratex DSGEs, and doesn’t relate to what could happen in the real world where unlike in model world there are no strict constraints on expectations switching from convergent to divergent.
Seems to me you are agreeing with when he says,
“Absent imperfect competition, price and nominal wage rigidity, and other market failures, the so-called new Keynesian model delivers Pareto optimal outcomes.”
You are claiming New Keynesians believe CBs can deliver Pareto optimal outcomes. This is the problem according to him.
*Seems to me you are agreeing with Palley when he says
Sorry.
How are interest rates set in NK models?
Nick,
What I meant to conclude is that the divergent solutions of the NK model aren’t candidates for our current economies because they would have exploded a long time ago. That means either that economies can’t have deflationary spirals or just that the NK model isn’t exactly right. But it doesn’t mean that a debt deflation black hole can’t happen in the real world. The solution, I’d guess, is a better supply side model. As Keynes pointed out, eventually capital destruction raises the natural rate so greatly that it exceeds the real rate so we don’t get a black hole (just a Great Depression).
“Think of a square root diffusion that can with finite probability disappear into an absorbing boundary (singularity) or not.”
Exactly. This is a great example. The parameters of square diffusion will tell you whether you stay positive almost surely. The asymptotics of the solution split into classes and it depends crucially on the parameters of the theory. You get different theories with different parameters.
“The problem is that for infinitely lived agents with the PIH everything becomes about what happens in a billion years.”
Yes, but this is an allegory for “long” time behaviour. It is not crazy to think that people might suppose that the future will have low productivity. For those models to capture something about reality, all we need is for the current generation to think that 30 years from now won’t see the same productivity gains that we saw in the last 30 years. 30 years may be asymptotic enough. Remember, all these models are parables for something far more complicated.
“My point is that a) high private discount rates seem like part of solution for making the asymptotic boundary conditions irrelevant.”
Yes, but then your analysis of the current US growth path is that the US population had a sudden outbreak of thrift. Really, does that even track as a serious explanation for what has been going on for the last six years or so? I think Casey Mulligan’s work (sand in the gears, expanding state) gets to the issue much more closely than the requirement for models to have very high discount rates (people actually do buy houses, and make long term investments, so very high discount rates can’t be the whole picture).
“any analysis that hinges critically on boundary conditions at infinity is getting something critically wrong about what matters to current economic agents and what stabilizes the competitive equilibrium.”
No. This matters. The sensitivity to boundary conditions tell you something about how the model will break and require other ingredients if the assumption of things like high discount rates fail to explain reality. Is it really high discount rates or is it a change in permanent income? In physics, you see things like this all the time – if you extrapolate certain theories like quantum electrodynamics to arbitrarily high energy, the theory runs into a pole and the coupling constant explodes. It means that new physics has to enter before the pole, you just don’t have a full description of Nature.
Again, I am no expert by any means on DSGE or NK. But a framework based on focusing on intertemporal substitution effects with no income effects, while interesting for generating insight, will have limitations. Let’s be honest and critical about what NK models can’t do and what other possible explanations exist. And let’s not use these models to give politicians an excuse to do what they always want to do, which is spend other people’s money!
“New Keynesians believe that the loanable funds theory is false. But they want the central bank to manipulate liquidity preference to try to make the loanable funds theory look true.”
But isn’t this the same as loanable funds, just with the central bank being the price setter. In your NK simple model C(t)/C(t+1) = (1+n)/(1+r(t)), so if r does not equal n then the system is out of equilibrium, because savings are not at their equilibrium level. The interest rate has to adjust to clear the market, isn’t this loanable funds?
What would the NK model look like with free banking?
AH: I should probably have been more precise. There is loanable funds at whatever level of income we’re currently at; and there is loanable funds when income is at the full equilibrium level. For example, in an ISLM model, if the economy is always “on” the IS curve, then loanable funds in that first sense is always true (as is liquidity preference, if it’s always on the LM curve too). But loanable funds in that second sense means the interest rate where the IS curve crosses the LRAS curve.
“30 years may be asymptotic enough.”
That’s my point. So if you are worrying about how exactly some solutions diverge (Cochrane), you are really missing the point. You only need to worry about how well your model captures the economically relevant future and prices (less than 100 years). With some reasonable private discount rates the rest is irrelevant.
“Really, does that even track as a serious explanation for what has been going on for the last six years or so? I think Casey Mulligan’s work…”
Reasonable people will differ. I happen to think it’s the principal cause.
“people actually do buy houses, and make long term investments, so very high discount rates can’t be the whole picture”
In real terms house prices have returned close to zero over the past 100 years. But when you add net rents on top of that you have a pretty high yielding investment (being a landlord has been a good business). Other long term investments like equities also have very high returns relative to the risk free rate. They are only long term because much of the yield arrives in terms of growth, not income.
“The sensitivity to boundary conditions tell you something about how the model will break and require other ingredients if the assumption of things like high discount rates fail to explain reality.”
Yes, but do they fail to explain reality? I’d say it’s the standard NKs and Kydland/Prescott RBCs without OLG type discount rates or liquidity constraints that fail.
“a framework based on focusing on intertemporal substitution effects with no income effects, while interesting for generating insight, will have limitations”
Sure. But leaving out market failure in intertemporal trade is also a very serious mistake. And like they say… there are a lot of Harberger triangles in an Okun gap. In my experience it’s a lot tougher to get a fresh water economist to consider even the possibility of deficient demand than it is to get a salt water economist to consider tax efficiency. RBC types never add price stickiness, but NK types often add capital and even tax effects. I’ll take the policy advice of Mankiw or Woodford any day over Prescott, Lucas or Fama.
Avon and Karsten: I’m enjoying your discussion.
It would be a bit of a surprise if expectations of long run growth suddenly fell 6 years ago. If, that is, we were talking about expectations of the supply-side. Beliefs about future technical change, for example, shouldn’t change so quickly. (And beliefs about current or future government decisions shouldn’t change much either, on a global scale.)
But suppose there really were no reliable process that would get the economy towards its supply-side potential? Suppose aggregate demand suddenly fell because people’s expectations of future aggregate demand suddenly fell, and there could be multiple rational expectations paths, and nothing to prevent a sudden change in beliefs? Just a different-coloured sunspot?
Karsten:
Yes, reasonable people can disagree. I don’t think that NK models are useless or anything – they provide lots of insight and opportunities for thinking about good questions. There are very smart people involve with all of this. Woodford is a very bright guy and so is Mankwi. And then again, so are Prescott, Lucas, Fama, and Cochrane.
What I don’t like is the social engineering that goes on with the policy advice. If really bright guys have quite divergent policy advice, it suggests to me that we don’t understand things well enough to micromanage the economy through some kind of social engineering experiment. I think the default option should be to do very little, use no discretion, and stop pulling on the levers of the state. The Fed created a disaster in the 1930s when they managed without really understanding what they were doing. We should apply a bit of Hayekian humility from time to time. Appealing to various academic camps based on what are actually highly stylized conditional models is not a basis for handing over the keys to economic production. We might think we’re smart, but we’re not that smart 😉
Again, I am not an expert on RBC, DSGE models etc. – as I said I read this for fun (I’m a physicist turned quant and my focus is on derivative pricing and stochastic processes), but I enjoy it along with the discussion here.
“Other long term investments like equities also have very high returns relative to the risk free rate.”
OK, now you are getting into my area and this is a bit of a funny business. This really is a statement of the equity premium puzzle and time varying risk premia in the pricing kernel. And being a landlord has not necessarily been good business either – this a single asset class that will have a lot of idiosyncratic risk so it’s not clear to me why this is really that great of an investment strategy. (Does it decompose as a separate factor inside the stochastic discount factor or project onto the value premium? Probably not, so why should it see on average a high return? What’s the systematic risk premium? But that is for another discussion.) I don’t mean to suggest that NK models aren’t very interesting. There is lots to learn from studying them, I get that.
Nick:
Casey Mulligan has a great exposition on how the expanding welfare state played a large role in triggering the recession in 2008 (the recession had to come before the financial crisis). So, as the US starts to add social programs that look more like Europe, expectations of future lower productivity will reflect those changes. So, given that the US has not snapped back to tend, it might suggest a shock to permanent income, and that realization has set in during the last recession and hence the new lower growth path. If the US ends up with Canadian productivity over then next few decades, permanent income will look like a good explanation.
New Keynesians don’t think says law is false. They accept that a dearth of money corresponds by say to a glut in production. Few modern Keynesians accept overproductionist or underconsumptionist descriptions of what’s going on. An ideal central bank doesn’t return us to says law. Rather it prevents money from being non neutral by being, a la Say, in correspondence with everything.
Avon Barksdale:
“Casey Mulligan has a great exposition on how the expanding welfare state played a large role in triggering the recession in 2008 (the recession had to come before the financial crisis). So, as the US starts to add social programs that look more like Europe, expectations of future lower productivity will reflect those changes. ”
The assertion that the recession came before the financial crisis is factually incorrect.
It’s conventionally agreed that the financial turmoil reached its ‘hot stage’ sometime in August 2007 (when BNP Paribas told its investors they wouldn’t be able to take out their money from two of its funds). Even before that institutions specializing in sub-prime products had become insolvent.
The Great Recession started in December 2007 (you can check the NBER list of contractions).
Besides, I think that concepts such as Ricardian equivalence or PIH vastly exaggerate certain mechanisms (inter-temporal optimization, I don’t think there’s a substantial number of people who’d make their saving/dissaving decisions based on the government’s budget deficit or debt projections) and totally neglect other mechanisms (current and short-run income, for the great majority of people this is actually the single most important criterion for spenging/savings decisions, in my view), which I think play a much more important role.
One could, for instance, argue that expanding the welfare state would make the economy more, and not less, productive (available data point to the government being much more efficient at providing insurance than the private sector, especially in the health sector) and that the prospect of this happening should boosts expectations.
ZDENPR:
“The assertion that the recession came before the financial crisis is factually incorrect.”
No, the recession came first. People didn’t decide to start defaulting on mortgages because they thought it would be fun. The recession started in 2007 and the way mortgage contracts had been written set the stage for a catastrophe.
“current and short-run income, for the great majority of people this is actually the single most important criterion for spending/savings decisions, in my view”
That might be your intuition, but it’s wrong. People don’t behave that way. They may not solve some dynamic program to exact optimality, but they don’t behave like Keynesian drones either. If what you said was true, we could solve unemployment by increasing inflation – this strategy failed miserably in the 1970s. So, it must be the case that at least some forward looking thinking is going on by the public.
“One could, for instance, argue that expanding the welfare state would make the economy more, and not less, productive (available data point to the government being much more efficient at providing insurance than the private sector, especially in the health sector) and that the prospect of this happening should boosts expectations.”
This is almost never the case. The government does not use prices to coordinate economic activity so it is very difficult for the government to maintain efficiency (and I mean Pareto efficiency here). Available data does not point to the government being more efficient at providing insurance – giving the insurance company an army is not the road to efficiency. People forget that the US does not have a private health insurance market – it is a highly regulated and government contorted mess. It creates very high marginal tax rates on the poor and regulation with tax incentives for employers destroys any chance of a guaranteed renewable individual health insurance market. I am not in favour of Obama’s health care plan – I think it will make the problem worse – but that doesn’t change the fact that the current US health insurance situation is an awful mess. (A topic for another conversation).
Avon,
“What I don’t like is the social engineering that goes on with the policy advice. If really bright guys have quite divergent policy advice, it suggests to me that we don’t understand things well enough to micromanage the economy through some kind of social engineering experiment”
Like setting interest rates? The freshwater guys think rates policy is basically irrelevant. It would a total disaster to let a guy like Prescott anywhere near the NY Fed. I can’t think of any similarly dangerous case of extreme wrongheadedness on the Keynesian side. The two sides just aren’t equivalent. There is something pathological about the ability of the RBC guys to ignore the evidence.
“they don’t behave like Keynesian drones either”
Just to be clear, the New Keynesian agents are intertemporal optimizers (labour/consumption smoothers).
Nick,
“Just a different-coloured sunspot?”
I agree 100%. “Supply” and “demand” are both inextricably linked to hopes, dreams, fears and every other contagious idea and human emotion and the idea of some objectively knowable optimal future supply side is ridiculous. Some day economics will have to grapple with the consequences of real world complexity and uncertainty. The Austrians have lots to contribute on this front…
Karsten:
OK, this is silliness. Prescott is not some random idiot who ignores evidence. Seriously, grow up – a different interpretation of the facts does not make you pathologically stupid.
And yes, the social engineering experiment is awful. Today, the Fed acts more and more in discretionary mode and buying securities that are not just treasuries. I don’t think people take the Great Depression seriously enough – it was caused by the Fed through and through. My grandmother was reduced to wearing a burlap sack as a skirt because economists in the 30s got it so wrong.
In macroeconomics, the models are just that, models. I don’t care how fancy or how many bells and whistles get attached, they are stylized facts about what the economy is doing. This is not physics, fermions and bosons don’t change their mind. Economic freedom goes hand in hand with peace and prosperity. I don’t want supposed smart men or women running the show in any real sense. Fine, if you want an inflation target, do it through a simple rule, known to all, and move on. The Fed does not need thousands of economists to feed input into some beast of a decision making process. I don’t care how smart anyone thinks they might be, using this kind of discretionary power risks disaster. It will be captured.
Look, if this was just some silly academic debate, fine, I get, the acrimony gets high because the stakes matter so little. But when we lay our hands on the ship of the state – which is really about removing economic freedom from people – I have little time or patience for social experiments. These are people’s lives, not simply talking points for a tenure track competition.
Nick,
Avon’s comment about physics and group theory got me quite interested:
“Quantum mechanics showed that the elementary systems that matter is made of, such as electrons and protons, are truly identical, not just very similar, so that symmetry in their arrangement is exact, not approximate as in the macroscopic world. Systems were also seen to be described by functions of position that are subject to the usual symmetry operations of rotation and reflection, as well as to others not so easily described in concrete terms, such as the exchange of identical particles. Elementary particles were observed to reflect symmetry properties in more esoteric spaces. In all these cases, symmetry can be expressed by certain operations on the systems concerned, which have properties revealed by Group Theory, a rather obscure branch of mathematics that had previously been mainly a curiosity without practical application.
Physics uses that part of Group Theory known as the theory of representations, in which matrices acting on the members of a vector space is the central theme. It allows certain members of the space to be created that are symmetrical, and which can be classified by their symmetry. It is found that all the observed spectroscopic states of atoms and molecules correspond to such symmetrical functions, and can be classified accordingly. Among other things, it gives selection rules that specify which transitions are observed, and which are not. These matters are so commonplace in spectroscopy that the fact that they are extraordinary and wonderful is hardly realized.”
(from http://mysite.du.edu/~jcalvert/phys/groups.htm)
Hmm.
May have to make use of this when returning to the topic of asymmetric redeemability.
🙂
Avon: “No, the [US] recession came first.”
Scott Sumner did a post a few months back, with the data, saying basically the same thing.
“The recession started in 2007 and the way mortgage contracts had been written set the stage for a catastrophe.”
Can you expand on that last bit? (On the way mortgage contracts had been written).
“If what you said was true, we could solve unemployment by increasing inflation…”
I do not follow you there. It seems to me there are two unrelated questions: the AD side (whether C is a function of current or permanent income); the AS side (whether an increase in AD will reduce unemployment).
Karsten: “Like setting interest rates? The freshwater guys think rates policy is basically irrelevant.”
I strongly suspect that at least some freshwater guys are deeply confused about central banks setting interest rates. If you start with a model with perfect price flexibility, then central banks can not really “set” interest rates. They can set the growth rate of the money supply, for example, which determines nominal interest rates as a consequence. Higher expected growth rate in M means higher expected inflation and higher nominal interest rates, but with next to no real consequences. But they cannot mentally translate from that world into a world of sticky prices where central banks can, temporarily, set nominal interest rates, but where setting a temporarily higher nominal interest rate means lower money growth and falling inflation.
“I agree 100%.”
I am very pleased to hear that. I really wish that NK economists would explore this possibility, which is already there in their models (or would be if they didn’t ad hocly assume it away). A good NK macroeconomist, who took Cochrane’s sort of analysis seriously, could really make something of this.
I see Prescott has come up here a bit. I’ve noticed that Mark A. Sadowski is not a fan of him, so I save links some of what he has to say about his work, for example:
http://www.themoneyillusion.com/?p=14137#comment-154228
Also Sumner quotes Prescott here:
“It is an established scientific fact that monetary policy has had virtually no effect on output and employment in the U.S. since the formation of the Fed,”
The comments section includes some people dumping on Prescott after that, so Scott defends him (a little):
“Everyone, Prescott is a fine economist. But like many fine economists he’s not a good monetary economist.”
And finally, Sadowski summarizes his views on Edward Prescott (after I dig up a few of his prior juicy quotes):
“And this especially applies to Edward Prescott, whose whole life’s work (in my opinion) is just a long series of glorified horse manure piles (pick any paper at random).
And you can quote me on that.”
JKH,
Nature is even crazier. All particles belong to the representation of group and it the action of the group that gives us the fundamental forces. When the tau lepton was discovered in 1977, we immediately knew that we would find two other quarks, the bottom and top, from group theoretic reasons. This same reasoning told us we would find the Higgs boson, too. It is an amazing achievement to understand the symmetries of Nature – you know how Nature has to come out even before you discover the complete picture!
BTW, these techniques are also used in math finance. The link between stochastic differential equations and partial differential equations can be understood through differential geometric constructions (there are some wonderful option pricing results that we can get from understanding the group structure of the problem). OK, way off topic, but it is so beautiful and fun!
“And finally, Sadowski summarizes his views on Edward Prescott (after I dig up a few of his prior juicy quotes):”
And this is exactly why economics is called the dismal science. Notice how physicists don’t do this – the work of Nobel Laureates in physics never gets described as manure. Seems like a bunch of prima donnas to me.
Avon, Sadowski rarely says anything w/o a reason. I invite you to ask him why he feels that way. I’m sure he’ll tell you. You can generally find him at Sumner’s. If you address your comment to him, he’ll see it and reply 95% of the time. Sumner has great respect for Sadowski as an empiricist. Or at least read some of his specific comments about his work (my first link). Here’s what Nick Rowe had to say to Mark once: “you are the best econoblog commenter out there.”
What did chemists and physicists have to say about Linus Pauling’s “work” on mega doses of vitamin C?
“What did chemists and physicists have to say about Linus Pauling’s “work” on mega doses of vitamin C?”
That had nothing to do with his Nobel Prize – and yes he went crazy later in life. It happens.
“That had nothing to do with his Nobel Prize”
Well is it possible that Sumner is correct? I.e. whatever his other talents, Prescott is not a good monetary economist?
Nick,
Mortgages were originated with all kinds of weird clauses, balloon payments, etc. If a recession hit, default would be a big problem, and everyone knew that. The only way the MBS market pricing worked is if everyone believed that the government would make things whole if things went bad, which they did (Bear Stearns) – and then didn’t (Lehman Brothers) – and then did again (AIG). It was nonsense and we got a near classic run on debt. (Duffie’s book on how big banks fail is a wonderful read.)
From what I understand, traditional Keynesian economics tells us there is a permanent trade-off on the Phillips curve between employment and inflation. If people only think about current income, then inflation can “trick” them into working more. The 1970s invalidated that approach.