What, in your opinion, is the shape of the Long Run Phillips Curve?
Supplementary: if you use the words "full-employment" in your answer, can you do your best to explain what you mean by this. (This is not a "gotcha!" question. I know that's a tall order, because it's not always easy to come up with a simple theoretically and empirically watertight definition of theoretical concepts; many orthodox economists have equal difficulty explaining what they mean by the "natural rate of unemployment".)
I think that Modern Monetary Theorists want "orthodox" macroeconomists to engage them more. I'm not sure that I'm really orthodox, but anyway; I sympathise with their desire, and this is an attempt to meet it.
Here is my "rational reconstruction" of the MMT answer. By that I mean here is what I can best assume they believe if I want to make sense of what they say:
"The Long Run Phillips Curve in {inflation, unemployment} space is L-shaped. Alternatively, the Long Run Aggregate Supply curve in {Price level, real income} space is reverse-L-shaped. The price level, or rate of inflation, is exogenous with respect to Aggregate Demand until you hit "full employment" (or "full employment output"). Then it turns vertical. So increases in Aggregate Demand (for example, due to money-financed increases in government spending, transfers, or tax cuts) will cause output to increase and unemployment to fall, with no effect on the price level or inflation, until you hit "full-employment". Thereafter they only affect inflation."
That's what I used to believe in 1972, when I was studying A-level economics and politics. (I only got a D; but perhaps, I hope, it was because I did very badly on the politics part?)
In 1972 I couldn't define what I meant by "full employment". But now i think I can rationally reconstruct what I would or should have meant by "full employment". Here it is:
"An economy is at "full employment" when an increase in Aggregate Demand won't increase employment even in the short run."
To put it another way, I made no distinction between the Long and Short Run Phillips and Aggregate Supply curves.
OK. Now for the comments.
Worthwhile Canadian Initiative 
Nick, see Marc Lavoie’s ppt, especially slide 16, here for a depiction of the post Keynesian Phillips curve: http://aix1.uottawa.ca/~robinson/Lavoie/Presentations/en/DR06.ppt
The PKE position is that there are multiple rates of path determined growth, inflation and capacity.
vimothy: Thanks. Marc does address the question of what the Phillips Curve looks like. Personally, I am sympathetic to his view that there may be a range of natural rates (so the LRPC is vertical but “thick”). And you can build a formal model where you get that result. But that result does tend to be very “fragile”, in the sense that very small changes in the assumptions can make big changes in that result. I am less sympathetic to his drawing the Phillips Curve as having a stable non-vertical slope on either side of that range (but maybe that’s just the Short Run Phillips Curve?)
But I don’t think Marc classifies himself as a MMTer.
Begruntled,
Well said. I’ve said the same thing in a different way
Nick
In begruntleds model high inflation CANT lead to high unemployment because employment is always guaranteed.
A note about something I said earlier Nick.
My comment about messy was certainly vague so Ill try to be a little clearer. Textbook economics with its formulas and theorems and laws ends up forcing micro principals that often times dont survive the composition fallacy. They dont aggregate well for various reasons and it really doesnt seem to bother them because they view aggregation perspectives as socialism (in a sense). All that matters is that the formulas are neat and explain something that has been seen in economic behavior at one time (even if its in a widget factory). They dont want to get into the messy questions of whether this will really apply or be explanatory for an ENTIRE populace of widely varying desires and choices. Thats a PROBLEM!! None of us exist only as individuals and we are all very dependent whether we want to admit it or not on the collective. So if your ideas dont apply well to the collective and dont even consider the collective something to think about how can you really be relevant? The collective is messy.
Here’s my suggestions for Laws of economics
As worker
1) I will do extra work for you if you pay me enough extra (or if I really like you) but I reserve the right to change my mind about the arrangement in a couple months
As Boss
2) I will hire the guy who can do the same quality of work for the least amount (unless your my son)
3) I will lay you off as soon as my expenses get to high and your not my son
4) Getting educated is fun and will lead to better conversations at break time but there are no guarantees you will get a god paying job unless you work for your dad
5) Working hard will keep you at work longer and will piss your coworkers off
This article by James Galbraith discusses the fact that labor markets do not function like markets and the thinking on labor needs to be revised.
He uses the term “job structure”. Very interesting.
Click to access ppb36.pdf
Nick,
“The ratio MR1/MR2 (or P1/P2 under perfect competition) will depend both on consumer preferences (the shape of the indifference curves) and on the relative supplies of the two factors of production.”
Yes, but I only need it to not be exactly equal to the technical production characteristics. An epsilon of income effects, preferences, whatever — is all I need for the system (1) and (2) to be non-degenerate.
Whereas you are saying something much stronger — that P2/P1 is identically the same as MPK1/MPK2 on the 3 dimensional subset where MPK1/MPK2 = MPL1/MPL2.
And actually, you need something even stronger, because condition (1) — that MPK1/MPK2 = MPL1/MPL2, will cut a codimension 1 subset that in general moves as the capital and labor stocks vary, so you really need P2/P1 to be identically equal to MPK1/MPK2 on an open subset of the entire 4 dimensional space. Arguing that P1/P2 is a downward sloping function, or that there are similar influences is not enough to argue that MPK1/MPK2 is exactly the same function.
For example,
Say
Y_1 = K_1^a L_1^(1-a)
Y_2 = K_2^b L_2^(1-b)
Then MPL1/MPL2 = MPK1/MPK2 on the codimension 1 set cut out by:
m = C/p
where m = L_1/K_1, and p = L_2/K_2, and C = (1-a)b/((1-b)a)
On that set,
(3) MPL1/MPL2 = Dp^(a+b) where D = [(1-a)/(1-b)]^(1-a) * (b/a)^a
“If the ratio P1/P2 were fixed, determined by preferences independently of Y1/Y2, then your equations would indeed do as you say (I think)”
No, P1/P2 just needs to be something other than
[(1-a)/(1-b)]^(1-a)(b/a)^a(L_2/K_2)^(a+b)
for some combinations of L_2 and K_2 within an open neighborhood of the equilibrium point. Certainly, it does not need to be constant. It can be anything other than the above messy function, and my claim would hold.
I think this is interesting — because if we were in the single consumer case, then you would be saying that the consumer’s 1/MRS(1,2) is exactly the same as MPK1/MPK2 away from the equilibrium points (they would of course be equal when (2) holds). In other words, in the presence of capital volatility, full employment is only guaranteed if there is a first order tangency condition relating the production substitution characteristics and utility substitution characteristics not just at the equilibrium points, but identically on an open neighborhood where (1) and (2) hold. Certainly if this were the case, it would be more advantageous for capital, since it would allow more points (K,L) to be potential equilibrium points — that choice is economically valuable, but will it always occur? If so, what are the mechanisms forcing the derivatives of micro-production functions to be exact copies of the derivatives of micro-utility functions at almost all points in the space of potential factor ratios?
So we can talk, in addition to wage flexibility, about production flexibility, in that over time, you might expect new techniques to be invented whose pairwise technical substitution rates get closer to the whole space of option, say by anticipating what consumer preferences would be in more and more likely equilibria.
But there is no reason to believe that with changing preferences and changing technologies, that you would always be in the maximal state where (1) and (2) are redundant. In that case, the production frontier would be less flexible in response to shifts in the size of the capital or labor stock.
And — correct me if I screw up the micro — you get the same situation with utility functions. By changing prices, you can get a single consumer with $1 and a fixed utility function to purchase any desired combination of two quantities (Q(A), Q(B)).
But as soon as you have two consumers with a total budget of $1, then it is no longer the case that by changing prices, any combination of quantities is possible, and you get a constraint between Q(A) and Q(B).
For example,
if the first consumer has utility = (x^2 + y)^(.4)
and the second consumer has utility (x + y^2)^(.4),
then MRS1(x_1,y_1) = 2x_1, with MRS2(x_2,y_2) = 1/(2y_2) = p_x/p_y
Both are satisfied on the (non-convex) hyperbola
() 4x_1y_2 = 1
And we can use this to bound total X in terms of total Y and vice versa. For example,
Multiply () by p_y and use p_y*y_1 <=1 to get
p_y = 4x_1 (y_2 p_y) <= 4x_1, so that
4x_1 > p_y, and then from the budget constraint:
p_x/p_y total X + (total Y) = 1/p_y <= 4x_1
Total Y <= 4x_1 – 2_x1(x_1 + x_2)
And similarly we can bound total Y in terms of X, etc. With more work, we could draw out the exact boundaries, but this is enough 🙂
Only in the degenerate case — e.g. both consumers have the same utility function, or each has a separable utility function, can we control Q(X) independently of Q(Y) by adjusting the relative prices of X and Y. The situation of a single consumer is not representative, qualitatively.
What you are arguing is that the “generic” micro-production function always has a similar separability property — such a claim needs to be justified, particularly if claims about wages and the labor market end up hingeing on obscure separability properties of micro production functions.
RSJ:
OK, there are two goods X and Y. In {X,Y} space, draw the Production Possibilities Frontier (PPF) that shows the maximum amount of Y that can be produced, for given X, given K and L, and given technologies. The PPF will be concave to the origin.
(Don’t trust me when I use the words “concave” and “convex”; I always muddle them!)
I need to assume just one consumer, or that all consumers have identical homothetic preferences (the MRS depends only on the ratio Y/X), if I want to work in 2-dimensions.
Each indifference curve is convex to the origin. (It must be convex to the origin otherwise it won’t satisfy the second-order conditions for utility-maximisation).
General equilibrium is that single point where the concave PPF is tangent to a convex indifference curve. (Remember there are an infinite number of indifference curve, each one drawn for a particular level of utility, so there’s always one that is tangent to the PPF.
In particular, the slope of the Indifference curve (the MRS) cannot be identically equal to the slope of the PPF (the MRT). They are equal only at one point, the competitive equilibrium.
There is a Budget Line that passes through that same point. It is tangent to the Indifference curve and to the PPF. The slope of the budget line is -Px/Py.
At that one point:
The slope of the budget line Px/Py equals the consumer’s MRS between Y and X, which satisfies the first order condition for Utility maximisation;
The slope of the budget line Px/Py equals MRT, which equals MPKy/MPKx which equals MPLy/MPLx, which is consistent with both firms maximising profits;
The consumer’s income, which equals WL+RK, which equals Px.MPLx.Lx + Py.MPLy.Ly + Px.MPKx.Kx + Py.MPKy.Ky will also equal (given CRS technologies) Px.X+Py.Y so the consumer can just afford to buy the goods his labour and capital produces. Ignoring any excess supply or demand for money (here’s where I slip in the assumption that Aggregate Demand is just right) his expenditure will equal his income.
I’ve probably got a couple of the ratios upside down (I normally do).
But what I have just done is sketched the existence of competitive equlibrium.
In graduate micro we spent a few weeks doing exactly the same thing with waaaaay more math, with an unlimited number of firms, goods, households, factors, etc. Somebody or other’s Fixed Point Theorem. I have forgotten it all, thank God! The above is all I can remember, because it was about all I really understood!
You would almost certainly understand those proofs better than me. All I have been trying to do is give you the economic intuition behind those math proofs. I hope it works, because that’s all I’ve got!
Nick Rowe,
It is not obscurantism to point out a series of academic work using more complex math than what you are presenting here in this debate. Simplistic assumptions and math will not deal with the problem. This is the problem with much economics that attempts with unrealistic assumptions and elementary math to deal with problems when there are measurement issues, heterogeneous non representative units or resources, indivisibilities and non smooth and non monotonic functions. The bottom line is that RSJ has a point, regarding the nonstable relation between resources and their renumeration independently of each other! Any mathematician will understand that and only economists are still not getting it because they still are attempting to analyze as if complexity does not exist.I would not say anything more about entropies from complexity that I use in my work because I will be called obscurantist!
Furthermore, in earlier comments I pointed out that the Long Run Phillips curve is unstable and shifting. I wanted to thank you for the opportunity to comment in your blog and the serious comments it receives!
Greg: “In begruntleds model high inflation CANT lead to high unemployment because employment is always guaranteed.”
But that just begs the question: “Could employment, be guaranteed? And if so, how?” For example, if the Long Run Phillips Curve had the “wrong” slope, a policy to create jobs financed by printing money could actually increase unemployment.
I wouldn’t call your list “Laws of Economics” in the same sense that (e.g. the “‘Laws’ of supply and demand” are “Laws of Economics”. Your “Laws” are more like “moral”(?) rules.
In my opinion, your Law 2 would reduce unemployment a lot, if it were generally accepted.
“As Boss
2) I will hire the guy who can do the same quality of work for the least amount (unless your my son)”
But I am in a distinct minority in liking your Law 2. It definitely breaks all current conventions. Your Law 2 would shock most people.
And the fact that it does break current conventions is one of the reasons why the labour market is indeed, as you say, different in many ways from other markets.
Panayotis: Dunno. Most of the time people criticise mainstream economists for using too much fancy math! Sometimes I think we can’t win.
Re: competitive equilibrium — yes I agree entirely, but I think we are talking past each other.
You are pointing out that given any (convex) feasible set, there will be a set of prices such that (in the PC CRS case), satisfy (2), namely MPK2/MPK1 = P2/P1 = 1/MRS(1,2) as well as (1) = MPL1/MPL2= MPK1/MPK2
I am pointing out that not all combinations (L1, L2, K1, K2) are in the feasible set.
And this is a no brainer. Write down a utility function, use the two production functions with a neq b, and solve for the condition that that MPK2/MPK1 = 1/MRS(1,2) as well as MPK1/MPK2 = MPL1/MPL2, and you will find that this restricts the possible values of (L1, L2, K1, K2). Then you should declare victory.
Only in very degenerate cases, such as MPK1/MPK2 = constant = MPL1/MPL2 = 1/MRS(1,2) will all tuples of L, K be in the feasible set.
Competitive Equilibrium results have nothing to say about what the feasible set is. You have to be told what it is, and use that as an input to the theorem. So my approach was to first treat as variables L1, L2, K1, K2, and then determine what the constraints on them were, assuming an equilibrium price existed. And you do get non-trivial constraints when your MRS, MPL, MPK are not constant. Write it out and see for yourself!
RSJ: You lost me (again!).
Given that 1 and 2 hold (i.e. both firms are maximising profits, and the consumer is maximising utility), then it is indeed true, as you say, “I am pointing out that not all combinations (L1, L2, K1, K2) are in the feasible set.” In fact, only one possible combination of (L1, L2, K1, K2) will satisfy both equations 1 and 2 (except in degenerate cases). [EDIT. and satisfy the conditions that L1+L2=L and K1+K2=K , I should have added]
If both firms are choosing the cost-minimising mix of K and L to produce a given level of output ( MPL1/MPL2= MPK1/MPK2 ) we must be on the PPF, and not inside the PPF. And this will (except in degenerate cases) require that both firms have the right mix of K and L. If you give one firm too much K, and the other firm too much L, they won’t be maximising profits, you will be inside the PPF, and production will not be as big as it could be, even with both K and L fully employed. You could get more Y and more X by rearranging K and L between the two firms.
“Only in very degenerate cases, such as MPK1/MPK2 = constant = MPL1/MPL2 = 1/MRS(1,2) will all tuples of L, K be in the feasible set.”
Agreed. [EDIT or nearly agreed] For example, if the two factors are perfect substitutes in production, like Y=Ky+Ly and X=2(Kx+Lx) for example, then the mix of K and L between the two firms won’t matter. And in this case the PPF is a straight line with slope -(1/2). [EDIT. And in that case, Py/Px must be 2, and the MRS=Py/Px condition determines which particular point on the PPF the consumer will choose]
Intuitively, except in that degenerate case, equation 1 determines the optimal mix of the two factors between the two firms, and ensures that we are somewhere along the PPF curve, then the MRS=MRT equation ensures we are on the utility maximising point on the PPF.
I know we must be talking past each other.
RSJ: I’m obviously doing a really bad job of teaching micro. My only excuse is that I never have taught it past the 1000 level, where we never do any math anyway! Sorry.
Try this: suppose you were the central planner in this economy? You were trying to choose {L1,L2,K1,K2} to maximise U(X,Y) subject to the two production functions, and the two resource constraints (that you can’t use more than the available labour or capital). Would you leave any labour or any capital unemployed? (Would either of the two resource constraints be slack?) No. Not if that resource had a positive marginal product in either good, and that good had a positive marginal utility.
Now check the first order conditions of that problem. You will find they are the same as the first order conditions for consumer utility maximisation and firm profit maximisation (they are the same as your equations 1 and 2). Therefore the competitive equilibrium duplicates the solution to the central planners problem (first theorem of welfare economics, more or less). And since the planner won’t leave resources unemployed, neither will the competitive equilibrium.
Does that help explain it?
Nick Rowe,
regarding your comment on “math”, my response is that I am competent but against its use for its own sake (out of knowledge)! Many economists base their analysis on elementary math presentation and given their limited competence on the tool they end up adopting simplified assumptions that fit their calculus! Their analysis is conditioned by their level of math application as their assumptions are correlated by their modeling. Even if their analysis cannot be refuted or confirmed by math proof, how can this be used to explain a complex reality? One should also realize that math presentation is a language imperfect in explaining itself and mathematical logic is not fully consistent as Wittgenstein showed in “Tractatus Logicophilosophicus”. Sorry, but the rest reads as Introductory Math for Economists that some of us have been bored to learn and teach in the past! Please save us!
Panayotis: I try to avoid math as much as possible. But sometimes you need it, unfortunately. Yes, math is imperfect. All languages are imperfect, English, graphs, as well as math. (Wittgenstein’s arguments in English were sometimes less than perfect models of clarity!)
“Sorry, but the rest reads as Introductory Math for Economists that some of us have been bored to learn and teach in the past! Please save us!”
Sure. What do you want me to do? Ignore RSJ’s argument? Refuse to speak math to him? (It’s the language he’s more comfortable in, so I’m doing my best to speak it, as well as speak English and graphs to him).
RSJ: I think I now see where you got hung up.
With Cobb-Douglas technology, the ratio of the two firms’ K/L ratios, i.e. k1/k2, is determined solely by the technology, independent of the total K/L ratio and consumer’s preferences.
With k1/k2 determined by technolgy, how does the weighted average of k1 and k2 adjust to match the total K/L ratio determined by the supplies of the two factors (and also meet the constraint of the consumer’s preferences)?
Answer: there are two degrees of freedom to meet those two extra constraints:
1. k1 and k2 can both increase (or decrease), provided they do so in the same proportions, to meet the total K/L ratio.
2. The relative size of the two firms can adjust, holding k1 and k2 constant, to match consumer’s preferences, and this will also change the weights in the weighted average of k1 and k2 that must equal K/L, so might require adjusting on point 1 above.
Ok, Nick, I think we are in agreement.
I was pointing out that with two firms, L1,L2 is in general a function of K1,K2.
But that does not mean that full employment is impossible, as you point out, there are degrees of freedom — in K — so that K1, K2 can adjust and full employment is reached.
But why would they? Each individual firm is in equilibrium, paying the market rate for labor and the market rate for capital, which are both equal to their marginal products. If a firm were to try to hire an additional worker — say a desperate worker who is willing to work for less than his marginal product — then the increase in L would cause MPL/MPK of the firm to change, bringing it out of equilibrium. You would need simultaneous coordination in which all the other firms adjust their L and K as well.
And with a large number of firms, the feasible sets become disconnected and horribly ugly. Even with 2 firms, in the simple example I gave with two CRS CDPF but with different exponents, and a separable utility function = log(xy), the feasible set was not compact. Add N firms, for N large, with all firms using the same 2 production factors, K and L, and your feasible production set becomes the intersection of N-1 convex surfaces — it could be anything. I’m sure that given changing utility functions and changing technologies, that there is always some (changing) magic combination of K_i’s that ensure full employment, but why would we exactly follow this ideal point, given the volatility of capital, without any lags or deviations?
And you can certainly imagine a situation in which K1/K2 is not always at the magic level — maybe foresight about future preferences or future interest rates is not perfect. Suppose that the actual level of K_i is equal to the ideal level plus some white noise term. Then the mean error of that white noise term is going to give you a mean “natural” unemployment rate, but due to capital market mis-coordination, rather than labor market rigidities, or a desire for leisure. Any deviation will force unemployment to occur, regardless of the wage rate, and this unemployment rate wont have anything to do with inflation — it will be the rate of unemployment caused by a less than perfect distribution of capital.
With the single firm model, this is not true — K1 can be anything, and you can adjust the wage to get full employment. But that’s just an artifact of the model, and cannot be applied to an economy with more than 1 firm.
Why don’t we hear more about the multiple firm intuition — unemployment due to a maldistribution of capital — rather than the single firm intuition of unemployment due to excess wages?
RSJ: “Why don’t we hear more about the multiple firm intuition — unemployment due to a maldistribution of capital — rather than the single firm intuition of unemployment due to excess wages?”
Good question. Maybe because only people like me can do intuition, and we can’t do the math for multiple firms; and the people who can do the math for multiple firms can’t do the intuition!
“…..You would need simultaneous coordination in which all the other firms adjust their L and K as well.”
Here’s my intuition: suppose we start at full employment equilibrium, then some new workers appear. They are unemployed. The market wage gets bid down, relative to rentals on capital. P.MPL is now greater than W, so both firms switch to a more labour intensive method of production (reduce k).
One puzzle is: what ensures that aggregate demand increases enough so that people want to buy the extra output that the extra workers can produce? Then you remember that this is implicitly a barter economy. There’s no money. Workers get paid in kind, using the firm’s output. So the unemployed worker goes to the shoe factory and gets paid in shoes.
” Add N firms, for N large, with all firms using the same 2 production factors, K and L, and your feasible production set becomes the intersection of N-1 convex surfaces — it could be anything.”
The mathEcon tell me it’s still convex, as long as you don’t have Increasing Returns to Scale. I was taught the proof as a grad student, didn’t really understand it at the time, and wouldn’t have a hope of explaining it to you now. But we (not me) still torture our grad students with it every year. Actually, your asking me this question is the first and only time in my life when my understanding this proof might have been useful. Maybe those MathEcon guys are doing something worthwhile after all.
“Suppose that the actual level of K_i is equal to the ideal level plus some white noise term.”
OK. Good model. And suppose the white noise term fluctuates every week, and it takes more than a week to change K1 and K2.
We now have to distinguish the Short Run (less than a week) from the Long Run (more than a week).
The LR model is what we described, in expected value terms (or, roughly as we described, because it’s non-linear, so Jensen’s Inequality will apply). The SR model is now exactly like the SR model in ECON1000. Each firm has a fixed factor, capital. So the equation P.MPK=R does not hold in the short run. Only P.MPL=W holds, if labour can be changed daily. If W is flexible, we still get full employment, even in the SR.
You would have a great time doing a grad degree in Econ. (Or maybe not, I don’t know; but you would be good at it, because it’s all this sort of stuff, only taught by people who know it, and who know math.)
Jon said: “Employment is maximized at any stable inflation target.”
Maybe or maybe not. I’d rather concentrate on what happens if currency denominated debt is required for the (price) inflation target. For example, what if the fed wanted/allowed currency denominated debt to be created and wanted/allowed that debt to go into a housing bubble to get people to spend and employ people? I think I read somewhere that about 1/3 of the jobs created between about 2002 to 2008 were related to housing in the USA.
Jon said: “To deny this, really is to deny 30 years of progress.”
I really don’t see much progress in macroeconomics in the last 30 years, especially related to the proper understanding of the nature of currency denominated debt.
If there was, people could accurately describe what a liquidity trap actually is.
Nick said: “I’m crap at math, and my micro is very rusty. But remember, a load of guys who are very good at math and extremely good at basic micro theory have been ploughing this field for many decades. And they haven’t come up with a theory of unemployment from this stuff. You need to throw something else into the mix. Playing around with Cobb-Douglas production functions alone won’t do it.”
How about adding changes in retirement date(s) to macroeconomics?
Nick said: “One puzzle is: what ensures that aggregate demand increases enough so that people want to buy the extra output that the extra workers can produce? Then you remember that this is implicitly a barter economy. There’s no money. Workers get paid in kind, using the firm’s output. So the unemployed worker goes to the shoe factory and gets paid in shoes.”
How about what ensures that aggregate demand increases enough so that people want to buy the extra output that the same amount of workers can produce (positive productivity growth)?
Now add in money (as in currency and currency denominated debt).
It seems to me that the question should then be if positive productivity growth and cheap labor produce price deflation and there is a positive price inflation target and probably a positive real GDP target, what should happen?
RSJ said: “Why don’t we hear more about the multiple firm intuition — unemployment due to a maldistribution of capital — rather than the single firm intuition of unemployment due to excess wages?”
Let’s concentrate on bank capital. Let’s say it went up a little bit, but currency denominated debt went up a good bit more with the bankers not believing there would be many defaults.
The bankers were wrong, and there were many defaults. What happens?
I would have great fun being a grad student in economics. But I’m not sure how that would pay the bills 🙂 It does make me wistful, though.
But I object to this part:
“So the equation P.MPK=R does not hold in the short run. Only P.MPL=W holds”
I think the opposite occurs. Why?
Look at the prices of capital — they are much more volatile than wage rates. Investment is much more volatile than changes in employment. If anything, we live in an economy in which investors and firms try to ensure that
P.MPK=R
at all times by varying investment and interest rates quite rapidly — much more rapidly than wage rates and employment. And in fact labor hoarding occurs so that as capital is cut, employers try to hang onto the labor.
Why would this happen?
Because the capital is forward looking. If tomorrow, you stop believing that your plant — or project — will be profitable, then you shut it down that day. Even if it will continue to be profitable over the next 2 years. I’ve experienced this many times — profitable projects were cut now because management peered into the future and decided that year X from now, the project wont be profitable any more, and for the project as a whole to break-even, there was a requirement that it be profitable for 10 years, not just the current year.
And they were right to do this.
Anytime you have a factor of production that requires a greater time commitment, then utilization of that factor will respond more rapidly to changing outlooks that the factor that requires a lesser time commitment — provided that you have many firms. Just as a small change in outlook will cause house purchases to tumble, whereas candy bar purchases will remain unchanged.
And this is even more true for taking on new capital commitments. If you assume that depreciation/labor force growth is such that X new projects must be taken on each week in order to maintain the right capital ratios for full employment, then a change in outlook can quickly bring X down to zero. In the aggregate, deferral of capital investments is the same as liquidation — so for an economy with many firms and sizeable depreciation, I think that the aggregate K varies much more than employment in the short run, even though for a single firm “K” is assumed to be fixed in the short run.
I think I could make a case from the data that K moves first, and L follows.
Nick Rowe,
My point about math is that you should use it appropriately. You are employing elementary calculus and optimization theory for a problem that includes resources that you cannot measure like capital, and/or are heterogeneous like labor and mode of technology that is indivisible and discrete and functions that are smooth, continuous and monotonic. I pointed out the Samuelson reswitching problem. I agree with the conclusion of RSJ but not with the modeling you are applying to a problem recognized and settled partially long time ago against neoclassical theory!
RSJ: Yep. Implicitly, in that sort of model ‘K’ has to be interpreted as machines, that the firm rents by the hour, and are owned by households. Formally, there’s no difference at all between K and L in that model. And since no firm in that model is producing K, it’s not even really capital at all. It’s more like land.
Panyotis: it is just as easy/hard to measure capital as it is to measure labour. Just count the number of machines, like we count the number of workers!
I am stunned! You really believe that that all capital is the same?!?! All labor is homogeneous with no differences?!?!Not even the Cambridge theorists (US) will say that!
Nope! I did say “easy/hard” 😉 I mean it’s no easier or harder to assume that all workers are the same than that all capital goods are the same.
Exactly my point! That is why I question the math used! I sympathise with the fact that more advanced math can not be easily used and/or displayed in the blog. Mathematical logic can help under the circumstances.
“Formally, there’s no difference at all between K and L in that model”
Hmm, then why is it that P.MPK would not change in the short run but P.MPL would?
re: math
To P — I would read Joan Robinson’s essay “On the unimportance of re-switching”.
In any model, you need to capture simple effects via a stylized presentation. The issue is not whether all details are presented accurately, since that is not the role of models. Models are Gedanken experiments to gain some insight about the world. I claim that the PK’ers and a lot of the heterodox people have jumped the shark here. They looked at some of the neo-classical conclusions, disagreed, and then reject math as a result. I don’t think there is anything wrong with models per se, but rather with some assumptions behind the models and also their interpretation.
For example, Ricardian equivalence is nothing more than the government budget constraint — you could read it as a theory of crowding out, or you could read it as saying that government deficit spending always creates enough assets to buy the government debt. Mathematically, the two interpretations are equivalent, but for political reasons people choose to focus on one of them and ignore the other.
Models are not to blame for that. And a similar result holds for looking at labor market rigidities as the source of unemployment as opposed to capital market rigidities. That doesn’t come from models, but from politics.
“Hmm, then why is it that P.MPK would not change in the short run but P.MPL would?”
P.MPK would change, if L changed. Why did K not change in the short run, but L would? Only because we assumed it, just to see if it would make any difference. We could as easily have made the opposite assumption.
I agree with the rest of your comment. Except that Ricardian equivalence is the government budget constraint plus some other assumptions that tell you whether people will want to save or spend those assets that the government creates.
I was assuming that they would both change in the short run, and that this would lead to the white-noise caused unemployment. But if you assume only P.MPL would change, then there wouldn’t be white noise unemployment given enough wage flexibility.
For Ricardian equivalence, part of the assumption is that deficit spending must be financed by selling debt. So RE is telling you that households will in fact elect to save by buying that debt, allowing the deficit spending to occur in the first place.
It’s not saying that in addition to saving enough to buy the debt, that they will save even more — say that they will save 2X of what is deficit spent. Perhaps I’m missing something, but it seems like just the government budget constraint to me, and you could interpret it as a statement that governments will always be able to run deficits of whatever magnitude they want, as people will always elect to save enough to purchase enough bonds to finance whatever deficit spending the government chooses to do. I’m not saying that you should necessarily read it that way, but the statement that should the government run a deficit of X, people will elect to buy X worth of government bonds is pretty much an accounting identity.
Anyways I don’t mean to gum up the boards. Am sick now and blogging is better than watching television 🙂
RSJ: sorry to hear you are sick.
“…the statement that should the government run a deficit of X, people will elect to buy X worth of government bonds is pretty much an accounting identity.”
This is where you really have been badly lead astray by the MMTers. Some of the stuff they say might be right, or might be wrong; it’s an empirical and/or theoretical question. But when you misuse an accounting identity, and mistake it for an empirical or theoretical statement about how the world works, you are just logically wrong.
Let me just pretend that I think that Ricardian equivalence is wrong. It makes my job easier.
Yes it’s true (in a closed economy) that if the government sells bonds people buy those bonds. That’s trivially true by definition. If I sell you a car, you must buy a car from me. But that doesn’t mean you want to buy the car at the existing price. It doesn’t mean that an increase in supply has no effect on prices. I must persuade you to want to buy the car by lowering the price enough that you will want to buy it. Same with bonds. The government can only manage to sell the bonds if it can lower the price enough to persuade people to buy them. That means bond sales will cause the rate of interest to rise, which will reduce investment, persuade people to hold less money, etc. So selling bonds will have lots of real effects on people’s behaviour. Ricardian equivalence is false.
“I must persuade you to want to buy the car by lowering the price enough that you will want to buy it. Same with bonds.”
I guess I believe that for liquid instruments, the price is purely a function of expected return. Too much government spending can cause inflation, which would raise nominal rates. But inflation can arise even without too much government spending, in which case rates would also rise. So it’s better to just look at expected return — that is the “right” thing to look at. If that is influenced by quantity, then yes, quantity will affect the price, but only indirectly. I think this holds for open or closed economies.
This is only for liquid instruments, but the central bank ensures that government debt is liquid. And the ‘fair’ price is the liquid price. In the long run, I think all debt is liquid.
Nick said: “The government can only manage to sell the bonds if it can lower the price enough to persuade people to buy them.”
Put a 0% capital requirement on the gov’t bonds. Why can’t “banks” buy an unlimited amount of them at the current price?
IMO, the gov’t/the rich/the bankers will then get their real return by inflicting negative real earnings growth on the lower and middle class.
RSJ said: “the statement that should the government run a deficit of X, people will elect to buy X worth of government bonds is pretty much an accounting identity.”
Why can’t the federal gov’t (if it is the currency printing entity) run a deficit with currency (no currency denominated debt issuance)?
Hope you feel better.
Sure, the government can create money and spend it. Often without inflationary consequences. I claim it makes no difference, in that household wealth is unchanged whether the government offsets deficit spending with bond sales or whether it creates more money. Inflation, household wealth, and output will be the same in both cases, as will interest rates. I know this is an idiosyncratic view, but I’m sticking to it 🙂
I believe the MMT position is that it is better to issue currency than to issue bonds. I say that when assets are fairly priced, the effect is the same.
In defense of the MMT paradigm, they are well aware of the inflationary risks of deficit spending, but correctly argue that in many cases there is excess capacity that can absorb deficit spending without causing inflation. In any case, the choice of whether to deficit spend or not should be a political choice, based on public purpose, rather than a choice made by the central bank or the bond markets.
RSJ,
I agree with Joan Robinson which by the way I have met when she was very old. The reswitching issue is that the relations are more complex and can reverse and cannot be examined only in the locality that is convenient for us!It is not because is important for the measurement of capital issue. My point is that the math we use must be consistent with the assumptions relevant to the reality of the situation we examine and not what assumptions fit our simplified model.This is circular thinking leading to local solutions.
Bingo. We should democratically elect the people who govern the Bank of Canada (or the Fed, or whatever).
BTW, I was at the MMT meeting in DC that I suspect inspired this blog. Click my name-link for more.
RSJ: I believe the MMT position is that sovereign bond sales (for a currency issuer) merely shuffle extant private sector assets (public sector liabilities) but do not change the size of either balance sheet, and are certainly unneccessary for enabling Government expenditure.
All Government expenditure is done via marking up accounts, and all taxation is done via marking down accounts. There are no goblins shuttling wheelbarrows of gold around, it’s just numbers in excel.
Therefore, why bother to issue debt at all? You need to drain bank reserves to hit a positive federal funds rate. If the Fed is happy with rates at zero, it can stop issuing Treasuries and there will be no impact on anything. In today’s reserve rich environment I believe that is even more true.
So, in sum: all Government spending is currency issuance. Bond sales do not “off set” this issuance, they merely alter its term structure for purposes of mechanically setting the FFR.
RSJ,
Accounting balances are a static representation, usually in matrix form of a situation at a period of time and not time itself. Its a “black box” with no behavioral dynamics to it. You can be stock flow consistent but you need to add behavioral relations.
As about your statement of RE is circular and has no meaning. I sympathise with the MMT position but it says something more. Particularly, that there no future tax liabilities to cover and repay any debt because the debt issue is demanded for net financial savings purposes. This is behavioral. Similarly, there is no crowding out effect as in the presence of a banking system any reserves created by the central bank can be used to validate credit expansion in response to increasing effective demnad. If there are unutilized resources, rates will not have to change as inflationary pressures are contained. This is also behavioral.
Panayotis: “Accounting balances are a static representation, usually in matrix form of a situation at a period of time and not time itself. Its a “black box” with no behavioral dynamics to it. You can be stock flow consistent but you need to add behavioral relations.”
Yep. Exactly. Especially that last part about needing to add behavioural relations. (For once we totally agree on something.)
Welcome back Winterspeak! I enjoyed and agreed with a couple of your recent posts. Wish you allowed comments. I especially wanted (off-topic) to ask you about your rather cryptic statement that the Eurozone does not insure deposits. I had a guess at what you might mean, but I wasn’t sure. Is it because the individual countries can’t create money, so their insurance promises aren’t credible?
Mandos: I saw that all you guys were off partying in DC. But it was reading Billy’s blog that originally inspired this question.
Nick: Your wish is my command! Had to move blog to a new host, and I’m seeing if blogger comments work if I run on blogger itself.
You guessed at my meaning exactly re the Eurozone.
Winterspeak: That’s an important insight re Eurozone deposit insurance. Only the ECB can credibly insure deposits (if it breaks the rules). But will it? Let’s hope it’s not tested empirically. But I’m not optimistic.
P,
MMT’ers do not believe that there is no crowding out effect. That is a gross misrepresentation.
The difference is operational and political– most people have a fear of currency issuance and assume that in all cases, each dollar issued causes some inflation, raises interest rates, and is harmful.
MMT’ers argue that instead of assuming that deficit spending is always harmful, that the government should deficit spend up until the point where inflation exceeds some publicly agreed upon target. And this may require large amounts of deficit spending or large amounts of surpluses (depending on the circumstance), but that the reaction function that determines how much deficit spending the government engages in should be based on these inflation/output trade-offs as decided by the political system, rather than medieval fears that all printing of money is harmful, or that this should be left to cloistered experts in the CB.
The second point is that they do not want the CB to engage in this reaction function, but the fiscal arm — i.e. true “helicopter” drops of money, or rather not giving the money away but paying for goods and services. Whereas the CB is only allowed to purchase assets at market prices, and therefore monetary operations cannot stimulate the economy in the way that fiscal operations can.
I.e. with a monetary “stimulus” — e.g. lowering borrowing costs — you are requiring the private sector to stimulate itself, by encouraging more debt growth and deficit spending by members of the private sector.
With fiscal stimulus, the government engages in the deficit spending, which does not leave a debt overhang in the private sector.
I believe that because of this, a government that repeatedly resorts to monetary stimulus will end up with debt deflationary pressures, as the private sector eventually reaches a point that it is not willing to self-stimulate via borrowing anymore. This is another form of “crowding out” that should be talked about more, since we are experiencing it now in the U.S., whereas we are not in a situation of experiencing fiscal spending crowding out.
So the general principles are:
use fiscal policy rather than monetary policy
deficit spend based on politically agreed upon inflation/output trade-offs
do not offset bond sales with deficit spending, but issue more currency
None of the above suggests that there is no crowding out, but rather the MMTers are aware of the inflation crowding out that happens with deficit spending as well as the balance sheet crowding out that happens with monetary stimulus.
I hope the EU is not tested either. But there you have it, exogenous currency in action! I’m sure there are austrians somewhere talking about how they got it almost right ; )
RSJ: Small quibble, but I think MMTers would say that is that CB cannot engage in the stimulus effort you describe as CBs cannot add to the private sector’s net financial assets, they can merely alter their term structure. This is why (as I said above), MMT do not believe that bond sales “offset” (or “sterilize” to use an old gold standard term) currency issuance.
Winterspeak,
I don’t want to re-argue all these points here, but I was engaged in lengthy debates over at Billyblog in which they argued that the payment of interest on government bonds enriched the private sector above and beyond the deficit spending, and that therefore bond sales (might be) more inflationary than net currency issuance. Because of this, they argued that selling bonds was the unfair granting of “annuities” to private sector households.
My position was that whenever a financial asset is sold at market prices, you cannot assume that the interest received is more than the opportunity cost of the funds necessary to buy the asset, and that therefore it is completely irrelevant whether bonds are sold to the private sector or not. The private sector is enriched by the deficit spending, not any subsequent financial asset sales or purchases, provided that all assets are sold at market prices.
Aggregate household financial net-worth will be exactly the same regardless of whether the government has a policy to offset some or all of that deficit spending with bond sales.
Many pixels were spilt on this, because on the one hand, people “counted” the money — always a dangerous undertaking, as you need to “value” the money rather than count it — and said “look, before households have $X and next year they have X+ rX — they must be rX richer! Government is sending them rX, that, in aggregate, they would not have. And I was trying to point out the whole concept of a growing economy, in which all financial assets are expected to grow by rX, that in fact all assets will grow by rX, — hence r is the risk-free rate — regardless of whether the quantity of “cash” grows by rX, or who sends what to whom.
…and I’ll continue with this thought because I think it has a lot of relevance for some of my other discussions with Nick.
The belief that it doesn’t matter whether the government is sending rX interest payments to households, or whether the households are receiving rX from the business sector seems to suggest a theory of crowding out — that households have a choice of investing in the private sector or in government bonds, and 1 dollar of government debt purchased “crowds out” 1 dollar of private sector debt. This is false — in the exact same way that the loanable funds theories are false.
It is false because of the possibility of balance sheet expansion — which is something I’ve never seen allowed for in any RA model (although I’ve only looked at a handful). Balance sheet expansion screws up equilibrium boundary conditions, and so you must assume that the debt –> 0, which is not really observed. Balance sheet expansion is a bit like increasing returns in that way.
In general, I don’t think economists have taken balance sheet expansion into account, and so because of this, they believe that an increased desire to save must result in lower interest rates. But this is not true if households take the money that they saved, and use it as collateral to short a security instead of going long. You can equally go short or long — you can buy a put option or a call option. There is no reason to believe — a priori — that increased purchases of financial assets will always result in more call options being purchased than put options.
This is why unlike the goods markets, the financial asset markets do not have a downward sloping interest rate curve with respect to the amount of money saved. And in fact periods of high volume transactions tend to be market crashes — something that is not observed in the goods markets.
This is not just an artifact of financial engineering. Even if it were impossible to buy put options or to short securities, in a world of endogenous capital, you can “short” capital by purchasing the net capital producing firms, or go long capital by purchasing the capital consuming firms. For example, if you believe that housing is over-priced, you can buy stock in the homebuilders, lowering their cost of funds, and bringing more houses onto the market, therefore you are net seller of houses, and your actions raise the yield of the house as an asset class.
In the financial markets, increased purchases can result in downward pressure on yields or upward pressure on yields — it depends on what the expected return is. But if you view the financial market as just another good market, then increased purchases will only result in falling yields, and your loanable funds theory will hold, and your (financial) crowding out theory will hold as well.
There is only a “real” crowding out, in the sense that there is a finite capacity to produce real goods. The capacity to produce financial assets is infinite. Assuming liquid markets, there is never any shortage of money to buy a bond, as you can always borrow to the buy the bond, and you will as soon as the price of the bond is less than the expected return. Then you will make money off the spread. And in the same way, you can always short the bond, as soon as the price of the bond is greater than the expected return.
Back to the situation of government bonds — or really all bonds — in aggregate they are always purchased via balance sheet expansion. The household sector is a net holder only of equity. All bonds are purchased solely for financial engineering purchases, with some households being (directly or indirectly) net short and others being (directly or indirectly) net long, and any bond interest payments, whether those payments are made by government or not — exactly cancel to give no benefit to the household sector.
In a closed economy, aggregate household sector net worth exactly tracks the net present value of the equity of the private sector, and is not affected by the quantity or yield of bonds at all. But this can be a non-intuitive result, and without this view, you will start believing that it is possible to “run out of cash” to buy bonds, and so if the government tries to sell too many, then private sector bonds must increase in yield, regardless of what happens to expected return. They may increase in yield, but only because of the “real” crowding out — not because of the size of the auction.
RSJ: I don’t want you to re-argue anything, just pointing out the standard MMT perspective (which I think is correct, although I’m always open to learning more).
If you have a link to the right thread in billyblog I’d appreciate it.
Your point on whether interest paid on Treasuries is really fiscal expansion or not is a good one. A simple reading of MMT would argue that it is, although it could be wrong (for the reasons you lay out). There is an interesting extension that you may have already had on whether Treasury rates are really at a “market” price. In normal conditions they are (although they don’t need to be) and under QE who knows. Still, the Treasury currently sets quantity to manipulate price (instead of setting price and letting quantity float) and that sounds like “market price” to me.
I don’t know what you mean by “RA” model, but I agree, I have yet to see an economist demonstrate any understanding of what a balance sheet is — with the exception of MMT. It’s certainly possible that even they have not pushed this far enough.