Thomas Piketty has power that heterodox economists never had

One reason why Piketty makes conservative economists a bit flustered is that he’s coming from an unexpected direction. Traditionally, left-leaning economists have stressed the imperfections of markets and the adjustment problems which interested Keynes. Piketty thinks more like a Classical economist, emphasizing how well markets adjust over long periods of time. The whole kerfuffle about the elasticity of substitution comes down to that. In Piketty’s view, capital is incredibly versatile if you think in terms of generations rather than quarters. When I first studied economics we heard a lot about “elasticity pessimism” and it was usually the leftists who were the pessimists. But Piketty is a rather extreme elasticity optimist. I’m not sure that he’s right, but he’s the first writer to make me regret that I didn’t study Solow’s growth theory more carefully.

I’m not so sure. Piketty is, despite his protestations, firmly within the mainstream both in methodology and content; how strange is it for someone to get hired at MIT at 22, then move to a leading European social science department, publish many many papers in the very top journals, win the prize for Best Young European Economist, have his coauthor win the prize for Best Young American Economist, and then be seen as challenging the mainstream?

In the past 40 years, economics as a discipline has smoothly integrated game theoretical reasoning (totally absent in the economics of someone like Samuelson), behavioral (everything from high theory like Epstein-Zin to experiments), non-standard finance including lots of EMT deviations, a bunch of tools using networks developed by sociologists, history-laden political economy in the Acemoglu vein, the strong empirical turn toward quasi-experimental evidence, etc. This doesn’t look like a discipline that resists new ideas to me – indeed, the huge majority of what is published today in top journals would have been unpublishable in 1974.

True, most of this is work is done using mathematical theory, but this type of path dependence is reasonable: we have lots of existing models whose benefits and flaws are well known, and we have lots of stylized facts, so why would it make sense to avoid situating new work within that existing literature? Research doesn’t stand alone – it contributes – and the nature of the contribution depends on the existing base.

If you look at the response to Piketty within economics, his historical data-collection has always been very well-received. He was publishing in top journals with theoretical papers on redistribution right after doing his PhD (his 1995 QJE) and was publishing empirical/descriptive papers on the same topic a few years later (his influential papers on income distribution in France and the US). This gave him his reputation within the field. The problem with a book like Capital, though, is that the leap from this historical data to a projection as to what will happen in the rest of the century is much less well-founded. There simply isn’t enough theoretical or rigorous empirical evidence to support the claims Piketty makes about growing wealth inequality in the future, as tons of commentators within economics have pointed out.

The Lucas Critique – that a mechanism based on preferences, technology or constraints (institutional or other) must underlie any claim – is the most core of all core beliefs in econ. The parts of the book that do not appear to take that critique seriously – i.e., historical case study without reference to an underlying model extracting the causal factors – are not going to be influential within economics, period.

Sigh. No, we shouldn’t stop worrying about the concentration of wealth — sorry about that — but we should see it as one manifestation of a flaw in our makeup which has many possible manifestations. At some point, sooner rather than later I would say, needs to be taken far more seriously than a focus on economics alone allows.

As an amateur I’ve found the discussion surrounding K21 to be fascinating.

I wonder if part of it isn’t the timing along with the data and history. The Cold War has been over for a while. The good guys won. Now we could have Kuznets’s prosperity. Instead we had Reagan/Thatcher/Clinton/Blair, the triumph of neoliberalism and increasing inequality. Then we had the epic housing bubble and financial crisis at the capitalist center. It does appear that increasing wealth concentration causes instability and it didn’t clean up after itself in a very fair or thorough way. I believe Piketty’s suggestions about the future are open questions. Could see Keynes’s euthanasia of the rentier? Could wealth concentration cause a political reaction which leads to policies like a global wealth tax?

“… For reasons explained here, I eventually gave up the hope that economics (as I studied it in the 1990s) could give me that knowledge, and diverted to political theory/philosophy and later also ethics … “

Fascinating to read back through that thread from mid-2007, observing who is defending what segment of the economics profession, knowing as we know now that the Great Recession and associated financial crises are bearing down on the real world of human beings.

Gaffney’s review is pretty good, and I say that as someone who’s coming in from a completely different perspective. In particular, it’s not that I dislike the book it’s just that there’s all kinds of little inaccurcies, mistakes and misrepresentations all over the place. It’s neither precision and rigor, nor is it “a strong, historically-based, empirically-grounded, and theory-rich account “. History’s mostly wrong (see the big debate 20+ year old in Econ history on when wages began rising during industrial revolution). He gets Ricardo wrong (for reasons Gaffney mentions, and some others). He switches blithely between *returns* on a factor (say land rent) and the *shares* of a particular factor in income in a bit of sleight-of-hand-kind of way (paraphrase part on Ricardo: “growing income raised the demand for land so return to land rose. Something something something else, this growth in land’s share in income …” – wait a minute! If it’s income that’s raising demand for land and hence rent, then the share is not growing!) Misses all the other instances in economic theory where r vs g matters (Diamond and Samuelson?). And so on.

Admittedly I’m only about a half way through it, but so far it’s like I’m being promised that some great revelation is just around the corner but as I turn the page, it’s just more stuff about how important this revelation will be, when it’s revealed. Maybe my expectations were just too high going in.

mud man @ 8

I’m never sure when I haven’t chosen my words carefully enough. Then again, it’s hard to be sure until you see the reactions to them. I suppose that’s what conversations are all about.

It’s certainly true that the trend of human development has been toward more complex forms of social organization, and that economics broadly defined has always been part of that development. And, mea culpa, I suppose I was somewhat unconsciously echoing Marx in my characterization of capitalism. My point, though, was not to deny that we were mere foraging apes prior to 1740 or thereabouts, but only that what’s happened since, even if it hasn’t amounted to a genuine discontinuity, has seemingly reached an critical mass in the sense that, for the first time, human beings are in a position to determine the future of all life on the planet.

Pretty clearly, as a bundle of atavisms glued together by information and techniques laboriously compiled over centuries, we aren’t really masters of this technological genie we’ve let out of the bottle. While economists genuflect dutifully with the rest of us at that pious observation, few of them seem to really believe it, or to understand its relevance to their own discipline, even at this late date.

It’s certainly true that the trend of human development has been toward more complex forms of social organization

Random walks started near hard limits look like that, yes.

I think I speak on behalf of heterodox economists in general and say: We are going to keep on doing our work. You can read it, and perhaps be influenced by it. Or not. That’s up to you.

Anyway, since you just read the book, maybe you can help me with something I’ve been struggling with. Why does it matter if the profit rate is greater or less than the growth rate? What is it that will be true of an economy when r > g, that will not be true when g > r?

I’ve just started reading “Empire of Value” by Andre Orlean. Just translated this year. Much better than Capital in the 21st.

“In lucid, accessible language, André Orléan resurrects and explores the vital (but neglected) problem of value, grappling along the way with some fundamental defects of conventional theory. Through his mimetic hypothesis, the role of money emerges in the central role that both classical political economy and neoclassical economics denied to it. The Empire of Value is a bold argument, and a deep rejection of the justification for reliance on markets, except as a device for obtaining consent.” -James K. Galbraith

“The economics profession is presently experiencing a grave crisis of legitimacy.” First sentence in the introduction to Empire of Value. This proves to be an understatement, not a hyperbole.

@ 22

To me Piketty is saying that throughout Capitalism’s brief history, r has been about 5 percent consistently, which is kind of a weird political economical data point. But growth rates have varied (and per Sandwichman as have the quality of growth). As wealth concentrates, inequality increases and politics move towards the right, we find growth slowing and (per Sandwichman) the quality of growth and production slows, concentrating wealth and power further.

JW Mason @ 24

I think what makes it interesting is that the abstract blandness of projecting an asymptote into some distant shapeless future becomes kind of a rorschach test for economists. It is a subtle challenge to the ol’ law of diminishing returns, especially as applied to capital investment, and various other conceptualizations of capital that rattle around people’s cages. Gaffney (cited in a comment above) is sure that if only Piketty had been more careful in his definitions, land rent would pop out. (And, it would — Gaffney is not wrong.) It’s almost more interesting to read the reviews to see what people will do with it (and, often, how they will embarrass themselves).

Me, I go for apocalypse: something like North Carolina:
http://digbysblog.blogspot.com/2014/09/a-north-carolina-bridgegate-by.html

que_es @ 26

Yes, the observation has been around for a long time, but we haven’t made of it nearly as much as we should. We therefore accept as given what isn’t given at all, which strangles up our minds. As what’s not said gradually comes to supplant what is said, honest folks are forced into schizophrenia, and genial academics like Paul Krugman are compelled to offer their services as intellectual paramedics. All of which is unlikely to end well….

What is it that Piketty thinks will happen when r > g, but will not happen when g > r?

Let’s look at a sample of his comments (coming from various sections):

“If, moreover, the rate of return on capital remains significantly above the growth rate for an extended period of time (which is more likely when the growth rate is low, though not automatic), then the risk of divergence in the distribution of wealth is very high.”

“Are there deep reasons why the return on capital should be systematically higher than the rate of growth? To be clear, I take this to be a historical fact, not a logical necessity.”

“Let me now turn to the consequences of r > g for the dynamics of the wealth distribution. The fact that the return on capital is distinctly and persistently greater than the growth rate is a powerful force for a more unequal distribution of wealth. For example, if g = 1 percent and r = 5 percent, wealthy individuals have to reinvest only one-fifth of their annual capital income to ensure that their capital will grow faster than average income.”

@16 JQ I think you just elegantly and concisely stated why I’m struggling to get through your book.

I’m nervous about whether change can happen. Krugman was invited to “Rethinking Economics” and says something quite wrong and no one in the orthodoxy can bring themselves to see it. Then today or yesterday he hammers away at what I call the “tool theory of theory” which goes something like this:

a faulty syllogism does not correctly model the world, but it can be used as a tool to understand the world.

Say what? That’s what you get from the very best Swedish Memorial Bake-Off Winner?

Am I missing something, or is the Lucas critique just… stupid?

We treat things as black boxes. You watch the inputs and the outputs and try to predict.

If it works to open the black box and look at components of it that you can treat as black boxes, that might be useful. It might give you better results. If that’s hard to do, or if you try it and get bad results, then I think it’s not good to say that’s the only OK way to proceed.

If you understand the firing of neurons and can work your way up to creating and using memories and so on, that’s good. You might also get some good results treating that stuff as a black box.

If you understand all the details about how wood handles stress that will probably be useful for wood architecture stuff. But if you know enough to get by about that, the rules that work for carpentry and building, then you can get by.

I think Lucas also had the complaint that if you treat a whole economy as a black box, and you only measure inputs and certain outputs like unemployment and prime interest rate etc, we might simply not have enough data to predict very well. But even a good understanding of each of the components of the system does not tell you what you need to know about a whole economy. It’s kind of like, knowing everything about how the pieces on a chessboard can move — the rules for pawns and knights and such — will not tell you all you need to know to play chess. In that particular game it’s necessary, you can’t play chess unless you know those rules, but it isn’t enough. So probably after you think you know what to expect from each individual businessman etc, you still don’t have enough of a track record for whole economies to verify your macro predictions.

We just have to do the best we can.

@33 “I think Lucas also had the complaint that if you treat a whole economy as a black box, and you only measure inputs and certain outputs like unemployment and prime interest rate etc, we might simply not have enough data to predict very well.”

This is pretty much the opposite of the Lucas critique! The assumption that agents can optimize their behavoir is at the core of the Lucas critique, If agents can’t predict economic outcomes how do they optimize their behavoir?

André Orléan on the legitimacy crisis in economics. From the introduction to The Empire of Value (2014):

The economics profession is presently experiencing a grave crisis of legitimacy. There was a time when it sought to provide sound guidance for democratic societies by improving the effectiveness of reasoned public policy. But now, through its own negligence, it has shown itself to be a source of confusion and error. It allowed a suicidal scheme of financial deregulation to be put into effect, without any prior attempt having been made to assess the scope of the risks involved or to devise appropriate precautions against them. Instead of awakening minds, economics has put them to sleep; instead of enlightening them, it has cast them into darkness. The disrepute in which the profession is held today stands in proportion to its own failure, which is extreme and without precedent.

The reaction of economists to the scathing criticism that has been directed at them is striking above all for its lack of intellectual courage. Even if a majority is prepared to admit that very harmful mistakes have been made, most economists also persist in warning against throwing out the baby with the bathwater. To be sure, they say, undue reliance on a type of modeling that recklessly overrates the virtues of competition, together with a dogmatic insistence on the hyperrationality of economic actors, is indefensible. But these shortcomings give a distorted picture of the discipline. Economics is perfectly capable of correcting its excesses, by drawing upon new fields of research such as multiple equilibrium theory and experimental economics, even neuro-economics. So say the economists. And yet instruction at the university level remains the same as it was before the crisis; research likewise proceeds on the same assumptions as before, using the same methods as before. However many newspapers and magazines announce the return of Marx, Schumpeter, and Keynes, the fact of the matter is that nothing has really changed.

None of this should come as a surprise. Science obeys its own rhythms. Economists are not like weathervanes, pointing this way or that with every shift in the winds; they cannot be expected today to teach the opposite of what they professed yesterday. Nor is economic theory a mere collection of recipes that can be sampled in response to changing tastes; it is a highly structured body of propositions built up from falsifiable hypotheses, rigorous methods of proof, and a vast archive of established results—what the historian and philosopher of science Thomas Kuhn famously called a paradigm. Kuhn showed that it is in the very nature of paradigmatic inquiry to resist challenges to its view of the world. For a paradigm to be overturned at a moment of crisis, not only must a persistent series of anomalies have been observed, in contradiction of the accepted wisdom, but, no less importantly, there must be a new paradigm ready to take the place of the old one. Now, the fact that an economic crisis brings previously unsuspected problems to light does not mean that fresh solutions are available on demand. It is true that economists today quote Keynes, Minsky, and Kindleberger more often than they used to. But this ought not fool anyone. No matter that economists now find it convenient to distance themselves from the neoclassical assumption of efficient financial markets, the theoretical framework that organizes their thinking and their teaching remains unchanged. It has been kept in place exactly as it was.

The present work proposes to make a new beginning. It proceeds from the conviction that the difficulties encountered by economic theory owe nothing to momentary circumstances, but are the consequence of a fundamentally mistaken conception of economic behavior.

#35

“I think Lucas also had the complaint that if you treat a whole economy as a black box, and you only measure inputs and certain outputs like unemployment and prime interest rate etc, we might simply not have enough data to predict very well.”

This is pretty much the opposite of the Lucas critique! The assumption that agents can optimize their behavoir is at the core of the Lucas critique, If agents can’t predict economic outcomes how do they optimize their behavoir?

I have probably misunderstood Lucas.

I thought the idea was that agents will attempt to optimize their behavior, and if we understand those agents well enough then we will successfully predict their attempts to do that, and that will let us predict the economy.

If agents can successfully optimise, then instead of developing theories of macroeconomics we should get some agents to optimise for the whole economy instead of just for themselves. Unless we are ready to assume that if they optimise for themselves that *will* optimise for the economy as a whole, in which case we don’t need no macroeconomics.

But if they will *attempt* to otimise, that will have effects that must be predicted. And there’s the problem that they learn from experience. If they can predict what the government will do, they will act on those predictions. So whatever your macroeconomic theory expects will happen, it probably won’t happen that way the next time around because the actors will do something different based on what they learned last time.

So what we need to do to get a working macroeconomics, is to predict how their behavior will change before it happens. And to do that we need a deep understanding of them.

But I may have misunderstood.

@OP To my mind, the econoblogosphere is, perhaps, the Road to Damascus waiting for its Saul. But that may be too ambitious.

In the history of physics analogy, Piketty is like Tycho Brahe who charted the location of every star in the sky, night after night after night. He had some bizarre ideas about theory, but the data collection laid the groundwork for his famous student Johanas Kepler. Kepler’s theories focused on the subset of physics related to orbits. Only then could Newton come along and generalize at a much bigger level.

In this analogy, economics will become an empirically useful endeavor when Piketty’s students’ students are being awarded their PhDs. Do we have to wait that long?

My current hope is Noah Smith. This post here (http://noahpinionblog.blogspot.com/2013/08/a-few-words-about-math.html) suggests he should see the problem and be a critic of the entire Neoclassical edifice. But he seems to have fallen for some bad apologia that suggest the problem is confined to macro, and, even there, prefers to defer to the elders.

I haven’t read the whole thing, but maybe he meant that bigger families decrease each heirs’ share, and consequently increase their salaries/capital income ratio?

Yes, that’s what he meant, but that’s not “population growth” (which would affect g) that’s just “wealthier folks have more surviving children than non-wealthy folks” (which doesn’t really affect g)

Because if you start with a group with no wealth and you let the share of that group in total population grow then inequality increases. 0/N=0, but share of 0’s goes up. More generally, if number of surviving children decreases with wealth (or income) then higher population growth is going to mean more inequality, not less, contra Piketty.

Basically he’s trying really hard to provide a justification for what’s supposed to be the main contribution, which is that if r>g then we’re in trouble. g is pop growth + growth of average income. There might be a good case for the “growth of average income part”, I don’t know, haven’t seen anything convincing thus far. But there isn’t for the “pop growth” part. He’s just trying to force that square peg into the round hole with his hammer because everything has to be about r vs g. So you get nonsense arguments like the one he gives.

And the reason he resorts to this is because he wants to argue that we should expect g to be lower from here on now, and part of the reason for that is because pop growth is leveling off.

He’d have a stronger case (maybe) if he just kept pop growth out of it.

notsneaky @45
if number of surviving children decreases with wealth

At this point, we should really try to find the precise quotation I suppose. Still : I think there is the part were we don’t understand “population growth has an equalizing effect” in the same way.

For you it is about the effects of the fertility of two groups, rich and poor, and you find it doesn’t make sense. For me, it means big families will tend to dilute wealth – something I can imagine.

Ze Kraggash @ 48
In my picture, the X% (lets say “of everything”) grows at its financial rate, below the +100% in population. So surely it must go down, because doubling the population also double “of everything”, and X% just increased 6%?

… At this point I fear we are heading to YA Piketty Definition Party ; I _have_ to finish that book.

Back to my dejected heirs, the absolute should matter some, I think, because there is that threshold below which you need to work, and grow with the rest of us “g”.

@47-49

Ze Kraggash, yes that is exactly what I am saying.

You have:
(A) pop growth of wealthy – pop growth of non-wealthy
vs
(B) pop growth of wealthy + pop growth of non-wealthy (weighted by shares)

Piketty says that lower (B) increases inequality but what he really should say is that lower (A) increases inequality. But then that doesn’t fit this whole r vs g thing.

@50 Alex K., of course aggregate production functions do not exist, since they are abstractions. So there’s no “proofs” that they do not exist, just proofs that they can’t be derived from some underlying market structure and technology without strong assumptions. So what? Ultimately the question is whether or not they’re useful abstractions or not. Anyway, off topic really.

I got it. You were saying that the wealth of the top 1% (the capitalists) increased according to the rate of profit (r), while the wealth held by the bottom 99% (assuming it’s all derived from labor) must have doubled. It seems a bit counterintuitive to me, but I can’t think of a good rebuttal.

@notsneaky:
“[Aggregate Production Functions] can’t be derived from some underlying market structure and technology without strong assumptions. So what? Ultimately the question is whether or not they’re useful abstractions or not. Anyway, off topic really.”

I don’t think it’s off-topic: someone claimed that the theoretical apparatus in Pickety is essentially sound, being based in part on Solow growth theory.

I am saying that there is nothing theoretically sound about that theory: the attempt to explain growth (and whatever else) by using the two constructed variables of “capital” and “labor” runs into problems as soon as you try to settle on a consistent definition of “capital” and “labor” — a definition where the aggregated “capital” is somehow formed from dis-aggregated capital (and the same with labor). The problems are not merely that you can’t aggregate exactly — it’s that you can’t aggregate even to an approximation.

I am all for useful abstractions: if you show me that in a certain situation a few factors dominate, then I’m all for making a model using only those factors and then putting all the other factors under the umbrella of “error term(s).” But you should still do some work to show that the “error terms” are well behaved, and to do that you will likely need to argue by referring to the reality that you are trying to model.

I see no such efforts (never-mind successes) when it comes to justifying Solow growth theory.

———–
Also, here is some light reading related to aggregated production functions:

http://college.holycross.edu/eej/Volume31/V31N3P489_491.pdf

Alex K.,

Y[e] = f[aith](K, L[ittle]).

Aggregate Production Function Felipe, McCombie 2013

Felipe Website

Alex K: I wouldn’t mind defending the use of production functions in some manufacturing contexts.

Then, you’d be a fool. It was in trying, as a naïve twenty-something, to do exactly that, that I first realized just how brain-dead the concept was.

In manufacturing, everything is organized around technical control of the process and efficiency is a matter of controlling value of product quality and the expense of waste. You see anything about technical efficiency in the assertion that output is a function of input? You see any role for management. There’s a lot of management in manufacturing organizations.

The bottom-line is that output is not a function of input — it literally does not satisfy the mathematical definition of function: the same input can produce differing output. (The dodge, of course, is to assert that maximum output is a function of input, but that just begs the question of what is to be meant by, “maximum” in an operational context.)

I’ve seen economists informally describe the production function as a recipe, but the contrast with an actual recipe reveals the gaping problems. Think about baking bread: a baker and a baker’s tools: his spatula, his bowl, his board, his oven. Everything in the recipe — the eggs, flour, sugar, yeast — have been mysteriously omitted from the production function, but try to overlook that. Also, the gas to heat the oven. In what sense, can you be said to have maximized output from the combination of the baker’s time and the oven’s time? How do you account for the idleness of the oven at other times? If you drop an egg on the floor, how do you account for that instance of waste?

And, if you forget trying to analyze an actual plant, an actual baker’s kitchen and go with a theoretical economist’s retreat to extreme abstraction, in which K, L represent the whole vertical chain of production from the dust of the earth, what is the excuse for having “K” at all, since K is itself an assembly of products?. If you are going to follow the production process into an infinite regress, why not follow Sraffa into an infinite regress that eliminates K altogether?

The production function does not survive an hour’s critical scrutiny or application to a practical context. If you take the whole technical apparatus of production as the exogenous gift of engineers, and the social apparatus as the gift of managers, and confine your attention to doing linear programming to identify shadow prices, the best you can manage is a first cut at daily scheduling of operations.

It’s not like no one has noticed these problems. Nicholas Georgescu-Roegen famously tried to work entropy into the equation. But, much of the literature explores purely theoretical issues in a flurry of notation, and simply side-steps the criticism, like it has never been raised, or can be answered by an appropriate assumption.

Sorry, for the off-topic rant.

In manufacturing, everything is organized around technical control of the process and efficiency is a matter of controlling value of product quality and the expense of waste. You see anything about technical efficiency in the assertion that output is a function of input? You see any role for management. There’s a lot of management in manufacturing organizations.

The bottom-line is that output is not a function of input — it literally does not satisfy the mathematical definition of function: the same input can produce differing output. (The dodge, of course, is to assert that maximum output is a function of input, but that just begs the question of what is to be meant by, “maximum” in an operational context.)

Ah, that brings back memories. My junior year in college I tried to design an economy for a small southeast-asian nation using Leontieff matrices and I grafted a price structure onto it. Every turn of the game was supposed to be a month, so I figured I could put in supply-and-demand functions and run it 20 or 30 times to get the next turn’s results, and the economy would quickly get into balance doing that.

But it didn’t. I thought there must have been something wrong with my programming. It didn’t work even after I took out the drug economy that the players wanted, that got the whole country growing drugs and importing everything else they needed.

I guess the imports were part of the problem. Probably they needed a tariff structure, with 20 product groups they were getting stuck making two or three and importing everything else which put them at the mercy of the international market. But nothing I tried worked. My economy resolutely refused to equilibrate. I could set it so it equilibrated everything to zero, but that was it.

Later I realized that of course it was never going to reach an equilibrium no matter how long I ran it, it was making limit cycles and they all had different frequencies, and, well…. I’d taken some intro economics courses and I’d read a variety of other interesting things, and I believed that the invisible hand would not have palsy. But for me it did. To run the game I had to give up on the program make up economic numbers that kind of made sense. The players of course couldn’t tell the difference. I used an RNG to make random numbers, and then weeded out results that looked too implausible. The players were happy to accept anything that favored their own revolutionary groups, but argued about things that didn’t. I told them I tried to make the program follow laws of economics, but I didn’t always know the details of why things worked the way they did. Players were inventive at coming up with plausible explanations when they wanted to. But a lot of the time they paid more attention to the war.

Leontieff input-output models are neat. Oversimplified, of course, but if you don’t start doing anything until you have everything perfect, when will you start?

When I said above that this was “off topic” I was very much anticipating that this would turn into some discussions about production functions and all that. Not interested. Once you’ve danced that dance several times, it’s sort of boring and pointless. Suffice it to say that I actually think the Solow model is a very useful and practical model, which has a lot to say about the real world we live in. From “why do some poor countries grow faster than others”, to “why do some poor countries catch up with rich countries”, to “why does the growth rate of income slow down the richer you get”. to “why did the Soviet Union experience a burst of growth, followed by stagnation”, to “why was Western Europe able to recover so quickly from the devastation of WW2”, to a number of other relevant questions. Show me a model without production functions which can generate the same amount of insight and then we’ll talk.

What I wanted to do is to take Piketty on his own assumptions and analyze his argument while accepting his premises. So far, even under these conditions, it looks sketchy. If you’d like, if you’re the heterodox sort, feel free to just reject it all together because he says something about Solow. I don’t care, it’s your business.

And Sandwichman, I’m pretty sure I’ve told you this before, but the ability to look up synonyms in a thesaurus or a disposition for making snide comments is not actually a substitute for thinking. Or having something useful to say.

John Quiggin 09.23.14 at 5:56 am wrote:
“Piketty has pursued, much more successfully, the strategy I advocated here: roughly speaking, orthodox methodology + radical content.”

Good political strategy, poor scientific strategy. (Since the orthodox methodology is mostly bullshit.) Piketty succeeded by jettisoning most of the orthodox methodology and keeping just enough to get the orthodox to read the book.

An analogy is trying to describe evolutionary biology within the orthodox *Christian* theological methodology; it’s really pretty hard to do that, though the poor 19th century biologists tried. You really need to jettison as much of the bad methodology as you can.

“You see anything about technical efficiency in the assertion that output is a function of input? You see any role for management. There’s a lot of management in manufacturing organizations.”

There are important aspects of the management/ownership of a firm that are completely unaccounted for in a purely technical description of a firm. I think those aspects are mostly related to the embedding of the firm in a larger complex economic system: in disequilibrium, different owners and managers will have different opinions about the best course of action, and there is not (as in equilibrium) an unambiguous right answer, except maybe after the fact.

But as far as purely internal efficiency is concerned, I don’t see why in manufacturing situations you can’t account for it by adding another type of labor input called “management.” This is going to introduce various errors, but I don’t see them as unmanageable errors.

“Think about baking bread: a baker and a baker’s tools: his spatula, his bowl, his board, his oven. Everything in the recipe — the eggs, flour, sugar, yeast — have been mysteriously omitted from the production function, but try to overlook that. Also, the gas to heat the oven. In what sense, can you be said to have maximized output from the combination of the baker’s time and the oven’s time? How do you account for the idleness of the oven at other times? If you drop an egg on the floor, how do you account for that instance of waste?”

All of the above seem pretty weak criticisms unless I’m missing some deeper point.
The eggs, flour and sugar are the inputs. How exactly are they unaccounted for? The point is not so much to maximize output as it is to maximize profits; that point in the production set that maximizes profits will have its own particular combination of idleness for the oven and the baker. The dropped egg can be seen as the result of the quality of labor plus some random error term.

I don’t see much problematic about this. If the criticism of aggregate production functions would be only that they don’t correspond exactly to reality then that would be a criticism not worth making. The problem is that there is no good grasp of the errors involved in such aggregation procedures — and from negative aggregation results we know that there is no good reason to expect such bounding of the errors involved.

(I agree with your other point, about expressing a micro-production function as f(K,L) — there are indeed deep aggregation problems in such a procedure too)

@65

Thesaurus? I suppose you’re referring to “mumpsimus.” That’s classic Joan Robinson on the Cambridge Capital Controversy. She said she found it in a dictionary.

“I was delighted to find in a dictionary the word mumpsimus, which means stubborn persistence in an error that has been exposed.”

@68 Like South Korea? I guess beyond the point where serious discussion is possible and we’ll just get the usual knee-jerkin’. Too bad, it was looking good for awhile.

@73 Exactly like South Korea. Behind the North when the war ended (I’ve heard) and afterward treated like Japan and Germany and everywhere else where we fought a war.

Proxy wars were fought not by Americans, but rather, our proxies. See, eg, all of Central America, Haiti, most of South America, many Arab nations, Angola especially but other parts of Africa, China.

Places where we fought actual wars: Canada, Mexico, Europe, Japan, Korea, Vietnam, Iraq, Philipeans.

@74:

Aside from the apparent fact that you got your numberings wrong…

I myself wouldn’t try to conflate the U.S.S.R. and Japan in some sort of singular totalizing explanatory scheme, based solely on economic factors, likely over-generalized and ahistorical, without regard for the political dimensions, including the geo- part. But YMMV. However, hermeneutically speaking, just as one always chooses one’s past precedents based on present concerns, so one can choose one’s “insights”.

It means that there are lots of different “usefuls”. And I find Sandwichman, among others, very “useful”, in his erudite and acerbic way.

notsneaky @ 65 The Solow model appears to work as a growth accounting framework for algebraic reasons that have little enough to do with its conceptual conceits.

Not true, although it’s a common misconception because of that “Bah Humbug” paper that keeps getting circulated in the heterodox econoblogosphere. That paper is junk and just makes a basic mistake.

Also growth accounting is one thing, Solow model another. Accounting makes no predictions, models do.

Asshole.

Suffice it to say that I actually think the Solow model is a very useful and practical model, which has a lot to say about the real world we live in. From “why do some poor countries grow faster than others”, to “why do some poor countries catch up with rich countries”, to “why does the growth rate of income slow down the richer you get”. to “why did the Soviet Union experience a burst of growth, followed by stagnation”, to “why was Western Europe able to recover so quickly from the devastation of WW2″, to a number of other relevant questions.

I think a lot depends on whether you consider “I dunno, probably something to do with technology” to be an impressive answer.

@yabonn, but this is not about their incomes, it’s about their wealth, their assets. Someone with a good income but no assets and living paycheck to paycheck has 0 wealth. So, it probably has something to with these new people accumulating their wealth, how fast it’s happening compared to the accumulation of wealth by the capitalist segment.

Ze Kraggash @ 87
Well, yes you need lots of assumptions and hand waving from here : saving rate for the poor, type of wealth of the rich (bonds and housing probably won’t react the same to population growth) etc. Accepting all that (and more I didn’t think about) one can maybe get what P means with his g and r. For example, your example at 48 of X% staying the same is just, in this case, a case of the r=g.

@88 A production model that didn’t have random variation would be a very bad one. As would one that assumed that principals are fully informed about the quality of agents. That’s why thousands of papers have been written on these topics (I’ve written 50 or so myself).

Economic activity is incredibly complex, and any attempt to understand it is fraught with difficulty. That’s true whether you call yourself heterodox, orthodox, mainstream or nothing in particular. Are you saying anything more than this?

#90

I don’t get the role of g. For example: as a laborer, I could rent a house all my life and accumulate 0 wealth over 30 years. Or I could buy a house, pay about the same amount monthly, and accumulate $200K of wealth over 30 years.

I know a guy who got a good job and put money into a house. But after some years there was a recession and he lost his job. Lots of people lost their jobs, and housing prices dropped. He would not be able to meet his payments and would lose the house. A man came to him and offered him $5000 for his $60,000 equity in the house. He got angry. “It’s worth $60,000 and you want it for $5000?” The man shrugged. “That’s my offer. If you’d rather you could get nothing.” He took it. When I met him, 10 years later, he was renting an apartment and he rode his bicycle to work because he did not want the expense of a car.

If only he had bought a house that cost only $60,000 he might have done much better. But if he moved into a neighborhood where he could get a $60,000 house, not unlikely it would have burned down.

“it is off-topic, and I must limit myself. ”

Well, OK then. It’s unlikely though, that you will ever be able to show something as strong as “production sets can’t be used to analyze any important aspects of the production process.”

Having a strong explanatory power in some important situations is all that is required in order to justify keeping such production models in one’s toolbox.

@85 I think a lot depends on whether you consider “I dunno, probably something to do with technology” to be an impressive answer.

But it’s not just that is it? If you think that “X causes Y” and somebody comes along and shows you that “X cannot cause Y, because then we wouldn’t observe what we actually observe, it’s got to be something else” then I think that’s useful. Surely eliminating incorrect answer increases one’s understanding. As the man said, it’s better to know that you don’t know, than to not know and not know you don’t know.

(in fact, and some folks here might agree, Economics could probably use more of these Solow-type “I don’t know what the correct answer is, but what you got right now sure as hey ain’t it” kind of results. And usually the heterodox celebrate these kinds of results since they show up limitations of Economics. But here the “K” comes with a lot of historical baggage so accepting that “K” ultimately doesn’t matter all that much produces just too much dissonance)

It comes up a lot in defenses of Neoclassical Economics and basically says: “Ok, there are valid criticisms, but you can’t say it’s totally wrong. Nothing is totally wrong.”

But what if it’s totally wrong?

Then you have a counterexample to prove it’s wrong to say nothing is totally wrong.

Neoclassical economics is not “totally wrong.” But neither is it totally wrong to say that “tomorrow the weather will be much like today.” You don’t need a weatherman to tell you the latter.

Neoclassical economics is limited in its capabilities and much, much abused by people for ideological and professional prestige purposes. The worst part is the bullying of heretics in the name of neoclassical economics by a self-appointed orthodox priesthood.

In her post at the top of this thread, Ingrid linked to an earlier post, A liberating exit, in which she wrote, in part:

In 2001, I was one of the three Cambridge University PhD students who wrote a little piece called Opening up economics, which was originally signed by 27 economic, business economics and development economics PhD students to call for an ‘opening up’ of the economics discipline. In fact, originally we were four writing the proposal, but one dropped out since she became too worried that writing this piece might jeopardize her chances of getting her degree. In fact, I think it’s fair to say that many of us were scared, and some fellow PhD students didn’t want to sign because they worried about how their supervisors would react. So far for the notion of a free market of ideas.

The current discussions about the credibility and validity of the claims made by heterodox economics seem to some extent to miss the point of what I understood to be the main reason for anger and resentment among heterodox economists: the effective and very strong marginalisation of anybody who is not doing either fancy empirical work (advanced econometrics or experimental techniques), or theoretical work in the form of applied mathematics. Any other kind of work is not considered of decent quality, is not published in the journals which are perceived as publishing the highest-quality work, is not regarded as valid material for a job talk, and so on and so forth. In addition, the culture of mainstream (read: almost all) economics departments is such that many, possibly most, economists regard most of the other social sciences as journalism—and philosophy is just words, words, words (as one fellow PhD student once told me, “everyone can use words, it doesn’t mean anything”). The dominant views on interdisciplinary research follow from these attitudes: yes, good interdisciplinary research is valuable, but most work done in other disciplines is not real science, so we can’t collaborate with them.

That says it all.

the effective and very strong marginalisation of anybody who is not doing either fancy empirical work (advanced econometrics or experimental techniques), or theoretical work in the form of applied mathematics.

The first part may be true. But the second – “or theoretical work in the form of applied mathematics” – is just absolutely not true. Maybe, just maybe it was true… 20, no, 30 years ago. Maybe you even have to go back to 40.

These days there are very few people who are doing “theoretical work in the form of applied mathematics” (which IMO is actually a bit of a shame). There are probably more Economic Historians out there than people theorists of this sort. Actually, scratch that word “probably” and replace it with “obviously”.

If we were having this discussion 20 or 30 years ago I’d say “well, that’s not true but I can see how you think that since a lot of big names you hear about are theorists” but even that is not true anymore. Look at the list of John Bates Clark medal winners. In the past ten years I see only one who could possibly fit that description (Acemoglu). Ok, if we stretch it a bit, you got Matthew Rabin in 2001 (but hey, his “theoretical work in the form of applied mathematics” is all about showing that standard economic theory on uncertainty *doesn’t* work) To get to a theorist before that you have to go back to Kreps in 1989. So out of the last 16 winners you got 3 who do this sort of thing.

This is a common and completely erroneous misconception. Part of it may arise just from the different definition of what constitutes “theoretical work in the form of applied mathematics”. For some people that means “yuck, it’s got some equations in it”. But that’s not a serious definition.

Economics these days is 95% about “interesting data” and/or fancy econometrics. And even the “fancy econometrics” are not always about mathematical acumen per se but about being good at figuring out what kind of estimation strategy is appropriate to a given problem.

Look Sandwichmen. I’m not your buddy. In fact, some of your previous comments here and elsewhere directed at me have been downright creepy and stalker-ish. Your comment above was clearly a “I know where you live” kind of a threat or attempt at intimidation.

Cut it out. I would appreciate if you did not address me by my name (I have no problem with other commentators, who haven’t creeped me out like you, have doing so), I would appreciate it if you stopped making coded references to personal aspects which have nothing to do with the conversation, and I would very much appreciate it if you stopped making comments “On a personal note”.

I don’t know how others perceive it, and maybe they think it’s all cool because your world view is closely aligned with theirs so you get a pass, so a bit of nasty assholery is fine as long as its directed at one of them “evul economists”. But from where I’m sitting it looks like you’re obviously *trying* to act like a low life creep. Stop it.

@106 There are two related moves that lead to the bitter disputes exemplified above. Krugman’s favorite is the “nothing is ever wrong” move that I call the “tool theory of theory.” All theories are wrong but some ar useful. But what Krugman never seems to see is that a tool that only clarifies thought can be used to get to any conclusion the user wants. That, of course, makes it a great way for bad people to kill poor people, which is what we see in Greece.

The other move is one that Nick Rowe makes a lot: we have a theory for that. Say somebody is pointing to asymmetrical information as causing problems for a particular theory. Nick will respond, but Neoclassical econ knows about that! Lots of theories include it. But this misses the point. The question isn’t whether each bit of reality is reflected in at least one mainstream theory. The question is: do you have a methodology that makes sure your models include asymmetric information when it plays an important role? And, no, there is no such method. It’s totally ad hoc. So, once again, theory gives answers that kill people.

Straight forward errors that kill people. In the face of that, snark is healthy.

I think Bruce Wilder in 105 is totally correct; I was always very worried at how badly economics described the world that I used to work in, before I got into the operational research literature and Stafford Beer, but his books make it clear how much you really need to think about if you’re interested in really modelling a production process (and in his “Principle of Sufficient Variability”, he seems to have the sketch of an argument that state-contingent sets and correspondences aren’t going to solve the problem either)

Immunity from criticism = if reality doesn’t conform to theory, it’s because one of the “other things” wasn’t “the same.”

Thornton Hall @ 102 There are two related moves that lead to the bitter disputes exemplified above. Krugman’s favorite is the “nothing is ever wrong” move that I call the “tool theory of theory.” All theories are wrong but some are useful. . . .
The other move is one that Nick Rowe makes a lot: we have a theory for that.

Do you have a category label for JQ’s move, as quoted below?

JQ: Although production functions are used as a kind of shorthand in various contexts (and are often used wrongly in such contexts) the idea that production technology can be described by functions rather than state-contingent sets and correspondences was obsolete by the early 1950s (see Arrow, Debreu, Shephard etc).

No hint, of course, of what problems “state-contingent sets and correspondences” solved for Arrow, Debreu, Shephard etc or how this would relate to the difficulty of understanding the economics of the technical and managerial organization of an actual manufacturing firm.

@118 I’ll have to think about it. But in @119 the words “vast literature” are a dead give away of the “we have a theory for that” move.

John Quiggin @ 11:29 pm: the problem it solved for Arrow, Debreu etc is precisely the one you raised, that the same inputs can produce different outputs, and vice versa.

Oh, for the love of mike!

The problem I raised was that production functions analyze only one type, and therefore, one portion of production efficiency, that relating to the allocation of factors, aka allocative efficiency.

That leaves other types of efficiency in the organization of production unidentified and unanalyzed.

The point about output not being a function of inputs was to highlight the insufficiency of the production function as a theory of production. If you don’t even have enough terms in the equation to establish a functional relationship, that’s a conclusive indication you’ve left something out.

None of this is hidden in basic neoclassical theory. The production function becomes a function by introducing an assumption that technical and managerial efficiencies are achieved (aka output is “maximized” for a given set of inputs). But, one really finds out how limiting this concession to analytic convenience is, when one goes about trying to understand the organization of a manufacturing operation, which is what I was saying to Alex K.

The suggestion that I was denying the existence of the vast literature in operations research, management studies, etc. seems bizarre to me. I’ll let Thornton Hall administer the appropriate snark. I do think dsquared’s reference to Stafford Beers, a founder of operations research, was exactly on point, regarding what I was trying to get at.

Beers applied concepts of cybernetic control from feedback to problems of business and production organization, and his ideas may indeed form a useful contrast to the approach of economists to efficiency. Beers got a little weird for my taste, frankly, but I got a lot out of his sketchy introductions. Ross Ashby’s little book, An Introduction to Cybernetics, ought to be a foundational classic for economics. The model of control that Ashby laid out gets a lot of uncredited application in economics, including principal-agent analysis; it should be applied to understanding production processes, imho.

Lol

@125 Sturgeon’s Law applies, in more ways than one.

BW:
“What the production function is not is a satisfactory theory of production, which can be usefully applied in an operational analysis of an actual industry or firm. It doesn’t form a useful descriptive analysis.”

Production functions (or more generally, productions sets) do not even pretend to be a complete description of a firm (say). Production set theory has a clear and hard to miss label of “black box theory of production.”

You can mark it as a youthful mistake that you tried to use such a black box theory to explain the details of the managerial sausage making (i.e. how firms go about achieving various optimizations, optimizations which are just assumed by production function theory).

But just because you made such a mistake it does not mean you have somehow proved that a black box theory of production is not useful in various contexts. For some analytic purposes (say, the analysis of the viability of some investment), what is of interest is mostly what can be achieved with various technologies and inputs — that in order to achieve those results you also need certain managerial talent and certain operational procedures is something understood, but not part of the analysis.

Maybe you wanted to say that the use of black box theory of production tends to marginalize more detailed analyses of productive organizations. Or something. That would have made some sense. But to say that production set theory has no analytic use is clearly absurd. (On the bright side, you did avoid being “not even wrong”)

——————-

To bring this back on topic, the criticism of aggregate production functions (AGFs) does not fall into the category of gratuitous criticisms which view any abstraction from the complexity of the real world as invalid. Rather, the problem is that in the use of AGFs there is no evident process of abstraction whatsoever, let alone any rough estimation of the errors introduced by the abstraction. As soon as you try to clarify what AGFs abstract from, you get into unsolvable theoretical problems.

Bringing this back to Pickety, Kevin Donoghue is probably right that there are open questions in his empirics and I’m also right that Pickety uses no solid theory. Other than that, he seems to have written a pretty popular book.