Why Thomas Piketty is Wrong About Inflation and Interest Rates

I have pointed out on here recently that Thomas Piketty’s views on public sector debt are wholly un-Keynesian. Well, we should also point out that his view of inflation and interest rates are also fairly un-Keynesian. Piketty basically thinks that the reason that governments have been able to run persistent government deficits is due to consistent inflation which erodes the real interest rates governments must pay on their debt. This may be true, but the conclusions he draws from it are altogether incorrect and, again I must stress, not the conclusions a Keynesian economist would draw. Piketty writes,

The inflation mechanism cannot work indefinitely. Once inflation becomes permanent, lenders will demand a higher nominal interest rate, and the higher price will not have the desired effects. Furthermore, high inflation tends to accelerate constantly, and once the process is under way, its consequences can be difficult to master: some social groups saw their incomes rise considerably, while others did not. It was in the late 1970s—a decade marked by a mix of inflation, rising unemployment, and relative economic stagnation (“stagflation”)—that a new consensus formed around the idea of low inflation. (p134)

In his book Money, the Post-Keynesian economist Roy Harrod brings to the reader’s attention a part of Keynes’ monetary theory that is not widely appreciated today. Namely, that Keynes thought — contrary to what the vast majority of working economists today assume — that the expected rate of inflation and, indeed, the actual rate of inflation have no effect on baseline interest rates. Harrod writes,

In the Keynes scheme the prospect of inflation has no tendency to raise the rate of interest. This springs from his contention that to understand the nature of the market rate, one must fix one’s spotlight firmly on the relation between cash and bonds. The point here is that cash itself is as liable to erosion by inflation as the promises to pay cash. If the choice is between holding cash and holding promises to pay cash, there is no difference whatever as between these two assets in regard to the prospect of inflation. (p179)

This leads Harrod to say, echoing Keynes, that central banks have entire control over the rate of interest. The markets really just have to take what they can get in this regard. If the choice is between cash being eroded by, say, a 7% rate of inflation every year and a bond yielding 3% being eroded by a 7% rate of inflation, the investor just has to make their choice and stand by it.

Of course, the money could flee the country. That is, there could be a run on the currency in question. But that seems very unlikely outside of a hyperinflation. And in a hyperinflation the dynamics will be self-limiting in two directions. (1) The rapidity of the increase in the money supply will greatly outpace any foreign outflow. (2) In a hyperinflation a currency generally loses its foreign exchange value almost completely, making it impossible for people to take their money out of the country and buy foreign assets.

Outside of a hyperinflationary collapse of an economy, people have to work and earn money and businesses have to turn a profit. This requires money to circulate within the country. Ultimately, this money has to go somewhere when it accrues as savings and if all of it left the country at once the economy would simply come to a halt. This has never happened in history, of course, and if you think it through it is truly an absurdity. In practice the markets just have to accept the fact that they might have negative yield on their hands. Ceteris paribus this will put them under very great pressure to invest the money in riskier assets within the country and this is part of the Keynes schema. Harrod writes,

What the prospect of inflation does affect is the comparative yield of bonds on the one hand and equities and real estate on the other. Equities and real estate are hedges against inflation, and, in periods when inflation is expected, the rate of interest on bonds should be higher than the yield on equities of comparable standing… But the fact that the prospect of inflation causes the yield of equities to fall relatively to the yield on bonds does not entail that it causes the yield on bonds to rise absolutely. According to Keynes it is impossible for it to have that effect. (pp179-180)

This is extremely important because the Keynesian view is very much so at odds with what many working economists will tell you. The Keynesian view will tell you that in an inflation risky assets will become more popular. That means that interest rates on these assets will fall, not rise. Interest rates on safe assets, like Treasury Bills, will remain wherever the central bank sets them. This is also what the historical data shows to be the case.

So, why don’t working economists generally accept this? Two reasons. First, is the loanable funds theory. This theory states that interest rates must increase when output increases. When confronted with the fact that the monetary authorities set the interest rate, loanable funds theorists have recourse to the soothing idea that too much demand will lead to inflation and this will lead to an automatic rise in interest rates. This myth salvages the model in which these economists have invested their intellectual capital. But it is inaccurate and at odds with reality.

Secondly, many working economists work in central banks or market institutions. The idea that inflation might lead to a rise in interest rates serves a nice mythic purpose for both. For the central bankers it acts as a taboo that reinforces their inflation fears because they believe that if they violate some sort of Divine Law then they will lose control over the situation. For the market economists it gives the illusion that their institutions — market institutions — have some modicum of control over interest rates. It also serves as an implicit threat to the authorities that if they dare to provoke inflation — which, of course, the financial markets hate beyond all else — market actors will jack up the interest rate.

But none of this is true. In reality, the authorities control the interest rate and no amount of inflation will move it beyond the boundaries in which they set it. High inflation may be an evil in its own right but let’s not fool ourselves with some sort of old time religion. Economists like Piketty would do well to give these issues a bit more thought before spooking the general public with the boogeyman of the supposed burdens of public sector debt and the supposed unsustainability, reminiscent of the doomsday warnings of the Austrian cranks, of eroding it through a healthy inflation.

I’m Pointing at the Moon, You’re Looking at My Finger: Janet Yellen on Post-Keynesian Economics

Here’s an interesting fact that I’ll bet many of you didn’t know: the current head of the Federal Reserve, Janet Yellen, wrote a short paper in 1980 examining the theories of the Post-Keynesians. You can find it here.

The paper is very clear and logically articulated. But it also manifests quite a few sicknesses of the mind that, for example, the current pluralist movement among students will almost certainly encounter in the coming months and years. Yellen cannot really think outside the confines of what she understands to be economics. So, the analysis is mainly an exercise in trying to reduce Post-Keynesian theories to their neo-Keynesian counterparts. She is not so much trying to find insights in the literature as she is trying to prove similarities between some aspects of the literature and the more mainstream neo-Keynesian literature.

This is a rather typical tendency in mainstream economists that I have pointed to before. We might call it ‘identity thinking’. That is, trying to reduce heterogeneous insights about the real-world to something that one already knows. “But we already know that”; “But we can get the same result using our model by making this and this tweak”; these are the hallmarks of identity thinking. Identity thinking is an inherently conservative mode of thought that academics should be very, very guarded against adopting. It is a mode of thought that shuns new insights — typically repressing them by pretending that they are already known or taken into account. It is an a priorist, anti-empirical and ultimately anti-scientific mode of thinking.

Another hallmark of such criticisms is the complaint that the model being examined is not “closed” or that the various “closures” used are inconsistent. I find this enormously problematic and frankly a bit irritating. If you read any of the classical macroeconomic texts there is never any interest in “closing the model”. Rather the idea is to lay out various insights, whether in algebraic or linguistic form.

Yellen, and most other mainstreamers, prefer to fetishize the model itself. They think that economics is not a discipline through which we gain insights about the real-world but rather a game in which we try to “close the model” and then other economists can discuss these “closures”. This is genuinely an attempt to turn economics from a discipline that gives its practitioners real-world insights into something more so resembling a Sudoku puzzle. Faced with this sort of rhetoric heterodox economists should not engage. Rather they should point out clearly the absurdity of what is being done and question why it is being done.

This annoys mainstream economists to no end. When faced with a challenge to the restrictive games that they play they typically cannot engage. Yellen, for example, annoyed that the authors she discusses do not do standard dynamic analysis writes at the end of her essay,

With respect to stability and disequilibrium behavior, the Post-Keynesian model could differ significantly from the standard textbook case, at least under certain conditions. Just what these conditions are however, we do not yet know, because Post-Keynesians have argued that events take place in ‘historical’ rather than ‘logical’ time and therefor have been unwilling to conduct the standard dynamic analysis. (p19)

Actually this is no longer true, although I wish it were. These days many Post-Keynesians are indeed playing this game. And in my opinion they have trapped themselves in doing so. There is a significant danger here in extricating the insights of thought experiments away from actual application in the real world and toward building toy models with little or no actual meaning or relevance. From there it is just one step away from ‘testing’ these toy models against the data using econometrics. At that stage the goose is truly cooked. Pack up and go home.

Yellen’s paper is particularly fascinating when it comes to one issue however: namely, income distribution. She fully recognises that the implication of the Kaleckian-style models that Post-Keynesians often use is that income distribution is set through the power of capitalists to control the mark-up price and hence the real wage. She also recognises that this is due to the assumption of monopoly/mark-up pricing as a matter of fact and the move away from marginalist conceptions of perfect competition and smooth factor substitution effects. But she makes nothing of this clear divergence from mainstream theory; namely, the divergence that, in contrast to the mainstream analysis, the ‘normal’ state of affairs is one in which income is distributed in line with relative social power. Again, she seems unable to get any insight from this because she is so focused on the  ins-and-outs of the model. The phrase “I’m pointing at the moon and you’re looking at my finger” comes to mind, as it so often does when addressing the inherently conservative mind of the typical mainstream economist.

When it comes to income distribution she also seems rather uninterested in the idea that different savings propensities among different income groups mean that different policies will have vastly different effects. This has massive implications but because it is not stressed in mainstream economics most policy economists seem not to appreciate this fact. Indeed, I found out recently that no government institution in the world has data that allows us to break down saving by income group and so this work had to be carried out independently by Barry Cynamon and Steven Fazzari. This means that institutions do not have information available using which you could estimate different spending multipliers for different income groups. Talk about a policy blindspot!

Yellen’s discussion of wage-led inflation is also myopic in the extreme. One thing that Post-Keynesian economics tries to show is that most inflation is likely to be wage-led. This is an empirical statement about the real world. The models just demonstrate how it might function. But Yellen is, again, totally focused on how the model might be “closed”. She writes,

In the Post-Keynesian model, higher money wage demands necessarily lead to higher prices without raising the real wage and so there is nothing to stop inflation once it occurs. Why inflation in these circumstances should not accelerate continuously I cannot understand. And since the real wage which is determined by the commodity market is only acceptable to labor by accident, it is hard to fathom why such an economy should not perpetually experience either inflation or deflation. (pp18-19)

Two things here. First of all, again notice the lack of interest in the statements made by Post-Keynesians economists as empirical statements. Again, the whole focus is on the model. And this is from the economist that now has the most influence over policy in the world! Secondly, Yellen “doesn’t get it”. But what is not to get? The model suggests that there are no price-equilibrating tendencies in real world capitalism which will usually be in a state of inflation or deflation. Well, that sounds about accurate to me! Zero inflation is the extreme exception, not the rule.

Yellen’s biases here are determined by her need to see the world as a self-equilibrating system. This is an a priori assumption on her part that is clearly not based on empirical experience of the real world and I gather that it is only after years of mainstream training that she has managed to assimilate it. These sorts of assumptions are ideology at its purest. The ideology that is fobbed off on the student is that capitalism is an inherently stable system. And once effectively indoctrinated they go into the world literally unable to comprehend an economist that says otherwise. Faced with the statement “capitalism might be inherently unstable” they reply “but how on earth do you close your model?”. Pointing at the moon; looking at my finger. Terrifying! Truly terrifying!

Most of the mainstream aren’t even aware of this. But every now and again this embarrassing fact bubbles up to the surface when the more honest and self-conscious among them engage in reflection. Thomas Piketty, for example, recently said,

“I was only too aware of the fact that I knew nothing about the world’s economic problems. To put it bluntly, the discipline of economics has to get over its childish passion for mathematics and for purely theoretical and often highly ideological speculation, at the expense of historical research and collaboration with the other social sciences.”

Indeed.

Cody Wilson and the Language of Power

I don’t usually do moral philosophy on here; except, that is, when I’m pointing out the implicit moral philosophy inherent in certain economic concepts. However, I recently came across something that I found so interesting that I felt I had to comment on it. The only way I can justify this is that I know that some of my readers — or at least one of them — gets very uncomfortable when I make criticisms from this point-of-view and what follows may be of interest in this regard. I will try to tie this back to economic by the end, I promise.

The ‘something’ in question that I found so interesting was the following video on Youtube. It is an interview with Cody Wilson; the man most responsible for distributing software packages online that allow people to create firearms using 3d printers. I suggest watching the whole video before reading the rest of this post.

Wilson’s Wikipedia page gives the impression that he is some moron anarcho-capitalist but watching the video you can clearly see that this is probably not the case at all. Rather he is quite clearly a Nietzschean of some form or other; this is confirmed by the following clip in which he, rather hilariously, suggests that Glenn Beck should read Michel Foucault. I think that Wilson’s Wikipedia page reads the way it does because it was edited by pro-gun advocates in the US, the majority of whom, as everyone knows, are on the libertarian Right.

First I should say: Wilson does support some nonsense, such as the internet fad known as Bitcoin. Bitcoin is based on a naive view of the world where the Believer convinces themselves that a system of interacting, ‘free’ people can self-regulate (usually based on some fantasy ‘non-aggression principle’). Implicit in this view is the idea that people are Naturally Good. All the turmoil in the Bitcoin market should be seen as a modern version of Robert Owen and the Utopian Socialists’ attempt to create a perfect society. What actually happens in such situations is that selfish people come in and basically start stealing and wrecking stuff and those that are naive enough to believe in the Natural Goodness of people get the short end of the stick.

Oh well. Man is not Naturally Good and requires Laws and enforced moral norms to ensure that He doesn’t do awful things. The reality of Unbridled Human Freedom is probably closer to a lynch mob than to a harmonious commune. Truth is the butt of a soldier’s rifle when the commander turns the other way. Lesson learned. Everyone go home.

Tied to this Wilson believes — and it is a belief with no real evidence — that innovation will spring up if intellectual property laws are done away with. Presumably, being an anarchist, he is also against state-funding for innovation. All he has is his belief. Again, there is no evidence for what he is saying at all and it based on some Utopian fantasy about the Infinite Potentialities of Man.

Anyway, I don’t really want to focus on that; if those particular revelations have not yet occurred to you they either will in later life or will never occur to you at all. Rather I want to focus on the way Wilson speaks in the interview. It is quite impressive. He completely avoids any of the moral questions as to whether what he has done by putting the instructions to make weapons on the internet is Right or Wrong. Rather he simply approaches the question of what he has done in a purely matter of fact way. He has, he says, completely circumvented the debate around gun control through a sort of Pure Symbolic Act. And he is correct. He has done just that. What’s more, in doing so he has given us a really interesting piece of television.

What Wilson has succeeded in doing throughout this interview is to completely avoid moral issues and instead speaks a pure language of Power. This is something that Foucault and Nietzsche mastered too and it can be read in their writings (and seen when the former appeared on television; relevant discussion starts at around the 36 minute mark, English subtitles available by clicking the ‘captions’ button on the right-hand side of the video).

As we can see in the Wilson’s interview, the language of Power is very strange when you encounter it directly in the formal Public Sphere. I do not mean by that ‘in public’ — the Foucault debate above is indeed ‘in public’ — rather I mean in a formal authoritative setting where everyone is supposed to behave and speak ‘properly’, like the BBC.

There is something disconcerting about the language of Power when it is presented in such a forum. The reason for this is because Power generally requires a cover behind which it operates in public. This is why, for example, politicians and lawyers are well-known dissemblers. Because they are engaging with Power directly and speak the language of Power, they always have to hide what they are really saying behind a sort of veil. Something similar operates on television shows like the one that Wilson is appearing on but that all breaks down when Wilson makes his appearance and the interviewer seems completely unable to get a handle on it.

Of course, in private politicians, for example, speak quite differently. They speak far more like Wilson in the interview than like the interviewer. This is not to say that either Wilson or politicians are wholly amoral. I don’t believe that Wilson is and I think most politicians do, in fact, have a moral compass (even if this may be badly oriented due to nonsense and ideology). In the case of politicians they tend to decide which way they want to lean on an issue but then understand that their main duty is to simply carry it out. That is where the language of morality falls away and the language of Power kicks in.

With Wilson the situation is more complex. He clearly does not want to consider the moral ramifications of what he has done by putting plans for weapons on the internet. If someone was shot tomorrow with one of the weapon designs he would not want to know anything about it. Rather he is interested in doing something that will, through a pure Symbolic act of Power, shake up the Power Structure as a whole. Once again, that is what the language of Power is all about. It is about changing, as it were, the objective coordinates that everyone finds themselves in. Beyond a certain point the language of Power avoids all moral considerations and simply intervenes to change the frame of the debate itself.

In principle, I think that this is very interesting. Interviews like Wilson’s really do, in a sense, give us a glimpse behind the mask. They scramble the circuits of the formal public space in which they are presented and the aesthetic effect, to me at least, is very pleasing. But the language of Power can never operate in a vacuum. Every intervention in the objective coordinates of a situation will channel more Power to one group and less to another. In undertaking any such intervention you are implicitly siding with those who gain more Power through your action. Those that will gain more Power by Wilson’s actions will be the ones who use Wilson’s weapons designs to gain Power over other people.

Who are these people? We have no way of knowing. But something tells me that they will not turn out to be anarchist liberators. Rather they will likely be people who hold Power over others through direct and very immediate and brutal acts of violence. Is Wilson responsible for their actions? Not really. But he has tipped the scales in their favour by releasing weapons designs to them. Looked at from the point of view of Power, Wilson is like a biased referee letting one side get an edge over the other. What they do when he turns his back is ultimately (and legally) their responsibility but you can’t avoid the feeling that in manipulating the game and implicitly favouring one group over another Wilson has some culpability in the outcome.

Perhaps this is where we can bend this discussion back to economics. Macroeconomics is, to a very great extent, a language of Power. (It is not a truly objective ‘science’, don’t be fooled by that guff!). When we make suggestions as to policy prescriptions a good economist should be weighing up how the action is going to distribute Wealth and Power. What people like Wilson can show us is that behind the moral facade of Rational Agents and Pareto Optimality, just as behind the language of television and politics, there resides a hidden language of Power. And every use of that language in real policy situations is an act of Power. Ignore that, and you’re a slave to your own intellectual constructions or, worse still, to intellectual constructions that you have been spoon-fed. And while I am by no means an anarchist, I do believe that people have a right and a duty to think for themselves.

A Quick Note On Hoarding and Scarcity: Applied Cambridge Economics

Yesterday and today in the comments section to my blog there was a somewhat interesting discussion about hoarding and unemployment. One commenter claimed that hoarding only caused unemployment in a monetary economy. I have heard this a lot, but I have never thought it to be remotely true.

I laid out a simple example. Here I will give another. Imagine an economy of robots with a constant population and a zero growth rate. This economy, then, just reproduces itself day after day. In order to do this oil is needed. The robots need to consume 80 barrels of oil a day to do their work (i.e. consumption) which is wholly occupied with operating the machinery, while 20 barrels are needed to use this machinery to extract the 100 barrels needed for the economy to reproduce itself (i.e. investment).

Now, what happens if I come along and hoard, say, 10 barrels of the oil? Well, the economy will experience unemployment that day. Let’s say that the 10 barrels are taken out of ‘consumption’ because the robots know that for every 1 barrel they take out of ‘investment’ they will lose 5 barrels the next day. Well, that means that 12.5% (1/8) of the robot workforce will have to go unemployed because they cannot be paid. Let’s say they ‘shutdown’ and do not operate that day**.

Another commenter correctly pointed out that this could not happen in a marginalist model. He was entirely correct but this does not point to the unreality of the situation — this scenario is undoubtedly a real one and perfectly logically coherent — but rather the inability of mainstream models to deal with such phenomena. Due to the way mainstream models are set up — in this case, the problems arise due to assumptions of substitution; the assumption that ‘capital’ is an endowed feature that does not rely on inputs; and the assumption of timelessness — they simply cannot conceive of such problems.

But these problems are real. I am no Peak Oil theorist but I think we can all agree that if 20% of the US’s oil imports were cut off tomorrow a recession and unemployment (together with inflation) would follow. Indeed, we have a natural experiment in this regard; namely, the oils shocks of the 1970s.

During the oil shocks OPEC undertook an embargo on exports of oil to the US. This was effectively the same thing as my example of hoarding. What happened? Prices rose after each shock and a recession soon followed. You can see this in the chart below (data from FRED).

As we can see, each time the OPEC countries held back oil exports to the US the economy dipped into recession. The mainstream at the time came up with all sorts of wacko explanations for these recessions (and I want to stress that oil was not the only factor) and they also came up with wacko explanations for the inflation that accompanied them. But that was because their basic models could not incorporate such events.

Thus even in a monetary economy a withdrawal of certain key resources needed for the reproduction of the system of production will create unemployment and recession. This is intuitively obvious, of course, but economists are not the most intuitive people on the planet. Rather they are led around by the nose by their models which completely structure the way they view the world.

The model that I laid out above is, by the way, a stripped down Sraffian model. Such a model could also be derived from the work of Wassily Leontief or Karl Marx in the reproduction schemas of Volume II of Das Kapital or you can find a similar framework married to Keynesian economics in Joan Robinson’s The Accumulation of Capital, but such a model cannot be derived from marginalist theory.

What lesson should we learn from this? Simple. Different models tell us different things about different aspects of the economy. Even if you don’t think that marginalist models are complete garbage you must understand that they only produce results based on the assumptions that they make. If the initial assumptions rule out, for example, the effects that shortages of a commodity that is needed to reproduce the system of production lead to, then the model-user will not be able to understand such phenomena. They will literally suffer from a myopia that is caused by their own rigid adherence to the models that they have been told are “superior” by those who mark their exams and give them promotions.

If that is not, at the very least, a good case for economic pluralism, I don’t know what is!

_________

** Actually, it’s a little more complicated than this. If the robots were rational they would recognise that the entire workforce was required to operate the machinery. Thus, by removing some of the workforce by ‘shutting them down’ they would not get the maximum amount of oil even if they kept investment constant. Thus, they would have to balance this properly. I think that the end result comes out at about 1 barrel of investment-oil forgone and 9  barrels of consumption-oil forgone. Thus around 11.25% (9/80) of the workforce go unemployed. This will lead to a shortfall of 5 barrels the next day that will have to be distributed and so on and so on until the economy returns to full employment. But I laid it out above as I did to keep things simple. The overarching point is the same.

Update: I actually ran through the simulation properly in the comments. What we actually get is a permanent unemployment equilibrium. Here I quote my comment from below:

In the period in question — call it Period t — demand is 100 (but 10 are essentially disposed of). The inputs needed to cater for this demand come from Period t-1. These were produced in a full employment economy. The robots then decide how to distribute these 100 inputs, minus the 10 extracted by the hoarder (imagine these disappear), in line with how best to produce inputs in Period t+1. The best way to do this is to “shut down” 19 of the robots and have the rest work.

In Period t-1: 80 units of labour (at subsistence wage) and 20 units of (circulating) capital produce 100 units of output. The labour:capital ratio here is 4:1. For every 4 units of labour, 1 unit of capital is used. Together each “coupling” produces 5 units of output, so 100 units.

In Period t: 10 units disappear (hoarded). Leaving the economy with 90 units. These are divided up in line with the previous ratio (4:1). So, that means that 72 units of labour are thrown together with 18 units of circulating capital (again, 4:1, the optimal ratio). The rest of the labour force remain unemployed because they are in “shutdown” and cannot be reactivated without more output. This system again reproduces itself. It produces 90 units of oil.

Now that I actually work through it what we get is a permanent unemployment equilibrium ala Keynes in the General Theory. Wow! Very interesting!

Is the Speculative or the Precautionary Demand for Money More Important in Real World Capital Markets?

In Keynes’ General Theory is is famously stated that the demand for money relies on three distinct functions. These are: the transactions demand for money; the precautionary demand for money; and the speculative demand for money. Or, more formally:

M = Mt + Mp + Ms

In that work Keynes — as he regularly did in his monetary theories — laid rather a lot of emphasis on the speculative demand for money and not a great deal of emphasis on the precautionary demand for money. In chapter 13 of his General Theory he wrote,

It may illustrate the argument to point out that, if the liquidity-preferences due to the transactions-motive and the precautionary-motive are assumed to absorb a quantity of cash which is not very sensitive to changes in the rate of interest as such and apart from its reactions on the level of income, so that the total quantity of money, less this quantity, is available for satisfying liquidity-preferences due to the speculative-motive, the rate of interest and the price of bonds have to be fixed at the level at which the desire on the part of certain individuals to hold cash (because at that level they feel “bearish” of the future of bonds) is exactly equal to the amount of cash available for the speculative-motive. Thus each increase in the quantity of money must raise the price of bonds sufficiently to exceed the expectations of some “bull” and so influence him to sell his bond for cash and join the “bear” brigade. If, however, there is a negligible demand for cash from the speculative-motive except for a short transitional interval, an increase in the quantity of money will have to lower the rate of interest almost forthwith, in whatever degree is necessary to raise employment and the wage-unit sufficiently to cause the additional cash to be absorbed by the transactions-motive and the precautionary-motive. (My Emphasis)

I have quoted that passage at length because readers of Keynes will note that it ties in with his earlier work A Treatise on Money through its mention of ‘bulls’ and ‘bears’. In that work too Keynes viewed the money markets as inherently speculative in nature. As can be seen from the highlighted part of the above quote he thought that the other components of the demand for money — namely, Mt and Mp — were not subject to changes in the rate of interest. Indeed, Keynes seems to leave aside these aspects of money demand and focus instead on the speculative motive.

Roy Harrod spends quite a good deal of his book Money arguing against this. Harrod claims that the precautionary demand for money is often far more important than the speculative demand. He also argues that the precautionary demand is just as sensitive to changes in the environment as the speculative demand. In the book he notes two distinct actions in the money market that are undertaken in line with the precautionary motive and have substantial effects on various aspects of these markets. These are: covering and hedging.

Harrod’s main example is taken from a situation when the exchange rate is expected to shift but I think that we could also find purely domestic instances of the same phenomena. Covering is the action taken by businesses — typically large businesses — that they should be able to ‘cover’ all their known future commitments as soon as they are entered into. These businesses then seek advice from professional financial advisers, often economists, who keep an eye on developing trends. If, for example, the financial advisers think that a currency which their clients receive in exchange for exports is going to decline in value they may advise companies to sell this currency in the future market.

Hedging is very similar. It has to do with the fact that companies and individuals hold financial assets from various different countries. If, for example, foreigners own a lot of sterling assets and their financial advisers believe that the sterling might decline they will likely hedge against this decline . They would do this, again, by selling the sterling in the forward market in equal amount to the amount of sterling assets that they hold. That way, if the sterling does decline the profit that they make on the forward sale will balance out with the losses they make on their assets.

Again, Harrod’s examples have to do with currencies but similar processes take place with respect to domestic issues such as inflation. Harrod argues that these processes make up far more of financial market activity than simple speculation. Of course, both speculation and precaution are both due to the existence of Keynesian or Knightian uncertainty in these markets but Harrod and Keynes are perfectly correct to delineate between them. But I do want to emphasise that the reason for their existence is identical to that of speculation: the future is not a mirror of the past and so many financial transactions are taken, not with given measurable numerical probabilities in mind (when these are given they are faked to give management the illusion of control), but rather in the face of true uncertainty.

This is not to say that Harrod thinks that speculation never plays a large role. He gives the example of Britain after the war when the Chancellor of the Exchequer Dr. Dalton insisted on holding interest rates down despite the likely upsurge of inflationary pressure. Many in the financial markets held money back from the stock market in the expectation that Dalton would be forced to raise interest rates, the market would decline and they could pick up stocks at bargain prices. Nevertheless, Harrod sees the precautionary motive as more important generally speaking than the speculative motive.

I think that Harrod is indeed correct. The financial architecture is taken up a very good deal of the time with precautionary rather than speculative ebbs and flows of capital. Indeed, this is how the capital markets generally advertise themselves; they portray themselves not as speculators, but as insulating their clients from the uncertainties of the real world. This, of course, is propaganda and a great deal of speculation does take place. But I think that Harrod is basically correct in that the bulk of the activity that causes interest rates and other financial variables to respond to uncertainty in line with Keynes’ liquidity preference theory are due to the precautionary rather than the speculative motive. He is also correct to point out that talk of ‘speculation’ any time, for example, exchange rates become unstable might be misleading. He writes,

The overemphasis of speculation, as against precaution, has been instanced in an entirely different field, by the frequent reference to ‘speculation’ against sterling or dollar in recent periods. Doubtless there was some speculation in these cases, but this was probably of minor importance compared with the vast movement of funds, due to the precautionary motive. (p173)

I too, like Keynes, am somewhat guilty of this oversight. In my working theory of assets prices I focus almost entirely on speculation. I very rarely mention that the price dynamics that I describe might be due to precaution in the face of uncertainty rather than speculation pure and simple. But this is really just a question of emphasis. As I noted above, both the speculative and the precautionary demand for money are motivated by the existence of uncertainty and so any framework that can deal with one can deal with the other. Both also give rise to self-fulfilling dynamics in which, when the markets begin to believe something, this something has a very good chance of coming true simply because this belief exists. This is what Soros calls ‘reflexivity’ and it is a key component of Keynes’ financial theories. That said, when I do revise my theory of asset pricing (which is currently in somewhat ill health) in the coming months and years I will give far more consideration to the precautionary as opposed to the speculative motive.

How Do Changes in Interest Rates Affect the Level of Economic Activity?

Roy Harrod has some rather interesting opinions on the effectiveness of interest rates. As readers of this blog know I am rather skeptical of using interest rates to steer the economy. Basically this is because I think that using them on their own will only result in ever-diminishing returns. Steve Randy Waldman discussed this with reference to the work of Michal Kalecki here, but similar arguments can be found in Joan Robinson’s Introduction to the Theory of Employment from 1937.

Another problem with using interest rates to steer economic activity is that they can, as Kaldor pointed out, lead to substantial instability in expectations and thus diminish investment in productive plant and machinery. I discussed this at length here but I will quote Kaldor once more because it is instructive. In his paper Monetary Policy, Economic Stability and Growth he wrote:

If bond prices were subject to vast and rapid fluctuations [due to the central bank manipulating the interest rate to steer the economy], the speculative risks involved in long-term loans of any kind would be very much greater than they are now [i.e. in the Keynesian era], and the average price for parting with liquidity would be considerably higher. The capital market would become far more speculative, and would function far less efficiently as an instrument for allocating savings – new issues would be more difficult to launch, and long-run considerations of profitability would play a subordinate role in the allocation of funds. As Keynes said, when the capital investment of a country “becomes a by-product of the activity of a casino, the job is likely to be ill-done”.

All that aside there is still the question as to how monetary policy actually has its effects and I think that Harrod’s views deserve some consideration on this point. He runs through a few thought experiments to show that a change in the interest rate shouldn’t actually effect the real level of investment all that much. He then turns to ask if monetary policy can prove effective in increasing or decreasing the level of economic activity. He comes up with two ways in which it is effective.

The first is simply on housing construction. Harrod is entirely correct about this. Housing construction is a very large component of economic activity and the interest rate has a substantial effect on this. The same is true of auto loans. But the second reason he gives is even more interesting. Indeed, it seems to be a progenitor to the ‘credit rationing’ arguments that started to penetrate the literature in the 1980s and 1990s. But I think that it is slightly more sophisticated than the rationing theories in that it explicitly makes a point about the institutional structure of the capital markets that such theories miss. I will quote Harrod at length here from his book Money.

It must be remembered that the capital market is for most people an imperfect one. There are bits of the capital market which function in the way of perfect markets, where at any one moment there is a going price at which the individual (other than some giants, like the government broker) can satisfy his needs. Such are the gilt-edged end of the stock exchange [i.e. stocks for state-backed, nationalised industries] and the discount market. But most borrowing for capital outlay is not done in these markets. There are all the various channels for borrowing, the commercial banks themselves, the market for new issues, financial syndicates, insurance companies and, above all, trade credit, which plays a vital role. In these various markets it is not a question of just taking out as much money as one wants at the going price. It is a question of negotiation. When the aggregate money supply is reduced, a would-be borrower may find it more difficult, and even impossible, to raise money through his accustomed channels; and conversely. It is essentially the imperfection of the capital market that makes monetary policy a powerful weapon. (pp64-65)

I think that Harrod is fundamentally correct on this issue. Raising interest rates does actually mean that, for a lot of borrowers, their interest rate doesn’t simply rise. Rather they cannot access the market for funds at all. They are shut out. And Harrod is also right to point out that there are many corners of the capital market in which prices simply do not arise; lenders simply say to borrowers “no thanks!”.

This does not, however, do anything to alleviate my concerns about using monetary policy as the main tool for macroeconomic stabilisation. I still maintain that this is inherently problematic. But it does show one channel through which interest rate changes might affect the real economy.

Is There Really No Labour Market in Keynesian Theory?

It was recently said by James Galbraith that a model that includes a labour market is “anti-Keynesian”. This caused a little bit of fuss but I thought that it was a very good first approximation of what Keynes’ economics in the General Theory is all about.

We can give this a little more precision by examining what Keynes thought to be the orthodox theory he was attacking and what he thought his new theory to be. A good place to draw this argument from is Athanasios Asimakopulos’ seminal book Keynes’ General Theory and Accumulation. In this book Asimakopulos examines the lectures that Keynes undertook prior to the publication of the General Theory, draft chapters and other sources, to better understand what the ideas in the book were all about.

Asimakopulos draws on on draft chapter by Keynes to show how Keynes understood the orthodox theory. Asimakopulos writes,

With production conditions conducive to the existence of the competitive markets implicitly assumed by Keynes, the size and number of co-operative production units would be determined by the net marginal product of labour and the marginal disutility of labour. If the marginal product of labour was greater than its marginal disutility, then more labour would be employed in existing production units and/or new units would be established, until equality was obtained. In such an economy total output and employment, as well as the real-wage rate, would be determined in the labour market. All the employment, and thus output decisions, taken on the basis of conditions in the labour market, are then automatically validated in product markets by the payment of output shares to the factors of production. Keynes surmised that these special labour market conditions would also be fulfilled in an economy where production units are owned and operated by a class of entrepreneurs, as long as total expenditure is always sufficient to purchase, at expected prices, whatever total output is produced. He calls this second type of economy ‘a neutral entrepreneur economy, or a neutral economy for short.’ There is no explicit reference to a neutral economy in The General Theory, but it appears to underlie his view of classical theory. ‘The classical theory assumes, in other words, that the aggregate demand price (or proceeds) always accommodates itself to the aggregate supply price’. (pp19-20 — Emphasis Original)

Keynes’ conception of a ‘neutral economy’ is, of course, the familiar labour market of marginalist theory where the supply of labour is determined by the marginal disutility that workers associate with working and the demand for labour is determined by the marginal productivity of workers.

But Keynes saw his own theory as examining an entirely different system. Keynes did not believe that this ‘neutral economy’ was a good representative of the real world. Asimakopulos goes on to lay out Keynes’ own views clearly.

In contrast to the co-operative and neutral economies there is what ‘we will call a money-wage or entrepreneur economy‘. Here, production and employment decisions depend on the expectations of money proceeds relative to variable costs, and ‘it is in an entrepreneur economy that we actually live today’. It was this economy that provided the setting for the General Theory, and it is here that employment is determined by independent aggregate demand and supply functions for output, and not by conditions in the labour market. (p20)

Thus perhaps a better way of formulating Galbraith’s comment is to say that: any model in which the labour market is the key to determining the level of employment is an anti-Keynesian model. This is why models that assume wage and price rigidities are not true Keynesian models.

In his book Asimakopulos shows clearly that Keynes had ceased to focus on price changes after his Treatise on Money and turned to a wholly different conception of how the economy functioned that effectively ignored the labour market as a determinate of the level of employment. He also shows that Keynes got into a terrible muddle about all this because he was still working on effective marginalist microfoundations. But if you read the material it is clear that the argument Keynes was making was definitively that the labour market does not play a role in the determination of employment.

Update: I have just come across a fantastic quote in Asimakopulos’s book that shows beyond a shadow of a doubt that Keynes rejected the idea that the labour market determines the level of unemployment. In an early draft of the General Theory Keynes writes. “we may well discover empirically a correlation between employment and real wages. But this will occur, not because the one causes the other, but because they are both consequences of the same cause”. That cause, of course, is the level of effective demand.

A Brief History of the Bank of England’s Endogenous Money Policies: An Ode to Roy Harrod

Roy Harrod, usually remembered today for his part in the development of the Harrod-Domar growth model was also, so far as I can see, the most sophisticated monetary economist among the early Post-Keynesians. His book Money, designed as a sort of textbook put together over the years using his lecture notes, is a testament to how a course on monetary economics should be taught.

Harrod should probably be credited, for example, with the first truly institutional description of endogenous money theory — a theory that the Bank of England has now come to endorse. In his book he discusses how the British banks of the time lend to one another in the open market — this was done by scrambling for ‘call money’ when they found that their books didn’t balance. Call money is so called because you literally pick up the phone and call a variety of lenders to try and raise the money needed to meet the reserve requirements in place at any given moment in time.

Harrod was well aware that “loans create deposits” — indeed he uses the phrase quite a few times in the book — and that banks then seek to raise the money to meet the reserve requirements after, not before, the loans are made. If they cannot fill the gap in the market for call money they turn to the Bank of England. Here I will quote from Harrod at length to show just how ahead of his time he was.

If, as a result of these operations, including of a calling in of call money, they [i.e. the banks] find themselves unable to balance their books, they can resort to the Bank of England, and rediscount bills with it. Here the Bank of England operates in its role as lender of last resort. In normal conditions the discount market has to borrow from the Bank of England at Bank Rate [i.e. the British equivalent to the Fed Funds rate]. This is above the market rate on bills, and thus during the period of such borrowing the Discount Houses [i.e. effectively, the banks] find themselves making a loss. They will have lent money on bills at one rate and have had to borrow from the Bank of England at a higher rate, commonly called the penal rate. It is accordingly highly expedient for them to get out of debt to the Bank of England as quickly as possible. They must therefore firm up their own rates, so as to discourage borrowers and encourage lenders.  The Bank Rate thus has a powerful effect on open market interest rates. (pp51-52)

Clearly the Bank of England sets the Bank Rate and allows the quantity of money to float. The Bank Rate and the lending rate gravitate toward each other because when banks extend sufficient loans that they are forced to borrow at the penal rate from the central bank they will quickly be incentivised to raise their own rates to squeeze off lending. Harrod is quite clear that this is how the process of money creation works. He continues,

When the market borrows from the Bank, this has the effect of increasing the money supply (deposits and notes) in the country, so long as this borrowing is outstanding. (p52)**

Harrod goes on to note that this system is different to the one in the US at the time (i.e. the post-war era). In the US the Federal Reserve was far less concerned with penalising banks. Although the mechanics of the monetary system were very similar, the institutional structure of the British system (in the post-war era) was far more geared toward penalising banks that borrowed from the central bank.

This was because of the peculiar situation in Britain at this time. In the post-war years, the British were very self-conscious about their balance of payments. Any time the balance of payments began to deteriorate the British authorities would try to bring the economy to a halt. This became known as the economic policy of ‘stop-start’ and it greatly hampered the ability of Keynesian policymakers to keep economic growth high at the time. Thus, the institutional structure of the Bank of England came to reflect the need for a central bank that could quickly ‘squeeze’ the market for funds when the balance of payments started to deteriorate.

The US, on the other hand, were rather cavalier about their balance of payments position in this era. After all, the Bretton Woods system at the time was based on the dollar and they issued the dollar. Indeed, many countries were more than happy when the US ran balance of payments deficits as this meant an outflow of dollars which developing countries could use to expand economic activity.

After the demise of the Bretton Woods system and the rise of Thatcher, the British began to care less and less about their balance of payments position. This is because they came to find — rather accidentally, it should be added — that capital inflows into the City of London were usually (but not always) sufficient to maintain the value of the sterling. And so, the era of stop-start economic policy came to an end. Since then the structure of the Bank of England has come to more so resemble that of the Federal Reserve in the post-war era; right up to their recent endorsement of endogenous money theory this year.

Harrod is also quick to note that the Bank of England stands behind the market for government debt at all times and effectively sets the interest rate on this debt through its operations. This is as true today as it was in Harrod’s time but is not spoken about very often. Harrod’s presentation is top-class in this regard in that he highlights that the rules put in place for bidding in the primary market for government debt operated seamlessly with central bank operations to set the effective interest rate on government debt. He writes,

Discount Houses are under an implicit obligation to take up Treasury Bills at issue each Friday. If the rise in interest rates has not achieved an attraction of outside funds into the market, nor deterred borrowing [by the government], the Discount Houses may be in a difficulty. The Government cannot trim down at short notice the amount of Treasury Bills it asks the market to take up. In such circumstances the Discount Houses will be unable to balance their books without resort to the Bank of England. If they go to the front door, they will lose money [i.e. they will be penalised, as we saw above]. They may have to put up with this for a week or two. But if the market remains inelastic and the Discount Houses have to go and borrow at the Bank week after week, the situation becomes intolerable. The Bank may seek to ease it by allowing borrowing at the ‘back door’. To the extent that it does this, or indeed to the extent that there is borrowing at all, the purpose of the original squeeze will be frustrated. The Bank will have reduced the money supply by Open Market operations and have had to replenish it again by ‘back door’ lending. (p56)

What does this mean? Well, basically that the central bank is under an implicit obligation to stabilise the market for government debt. If the government borrows and this puts upward pressure on interest rates, this does not lead to ‘crowding out’, as the textbooks say. Rather it leads to the central bank stepping in to put a ceiling on interest rates through ‘back door’ lending.

This also means that government borrowing and spending can trump any attempt by the central bank to control the level of economic activity. If the central bank tries to squeeze interest rates to curtail economic activity but the government insists on running deficits, the central bank will be under the obligation to undo their tight monetary policies by effectively providing the funds to the government.

Anyway, Harrod’s book is a fascinating read and, although long out of print, is far better than the textbooks used in monetary economics courses today. After picking it up I am convinced that Harrod should be recognised as one of the eminent monetary economists of the post-war era.

___________

** It should be noted that in the book Harrod does seem to think, as the Bank of England continues to maintain today, that through such interest rate manipulations the central bank can “control the money supply”. This is probably false, but that I have discussed that particular topic elsewhere.

 

Exploring Inequality: Real Wages and Productivity Growth

I recently had an argument with a few New Keynesian types in the comments section of Lars Syll’s blog. I won’t get into the nuances here as they are not very interesting. Basically the New Keynesians were trying to defend the idea that, in the long-run, wages are indeed flexible. The argument went nowhere partially, I think, because they misunderstood what the phrase “wages are flexible in the long-run” means.

To me that means that wages are both flexible in the sense that they will clear labour markets in the long-run (i.e. in the long-run there will be full employment so long as the correct level of interest rates is maintained in line with some Taylor Rule, as New Keynesian theory states) and it also means that wages adjust in line with the marginal productivity of labour.

You cannot really prove either of the above (dubious) assumptions with recourse to data. They are strictly a priori assertions, articles of the faith and any attempt to prove them empirically will run into unavoidable problems. Anyway, one of my interlocutors suggested that you could (and then added the caveat that you probably couldn’t) by looking at real hourly wages and real hourly output per worker.

Anyway, I decided to crunch the numbers to see what we would find. After some messing around with nominal wage growth rates my interlocutor pointed me toward real wage growth measure. The results do not say much for the idea that hourly work is compensated in line with hourly productivity. I broke the growth rates down into 8 year periods (because the data starts in 1965 and, as everyone knows, real wages started to stagnate in 1973) and here is what I found:

Or, to give you the same data in graph form:

As we can see there is certainly no linear relationship here. In 1965-1973 wages and productivity chugged along nicely but after this productivity continued rising while wages stagnated. Only during the Clinton boom era (here included in the 1997-2005 aggregation) did wages actually start growing again but these were far slower relative to productivity than in the 1965-1973 era.

Marginalists can avoid this data in any number of different ways by engaging in sophistical arguments. But the fact of the matter stands: trying to explain wages by appealing to productivity measures is extremely problematic. In a world where worker bargaining power is strong we may well see wage gains keep up with productivity but this is not due to some sort of market distributing resources in line with marginal productivities. Rather it is due to unions — probably being advised by economists who are tracking productivity growth — pushing for their living standards to keep pace.

Many economists today, in the post-Piketty world, are interested in inequality. (Economists are generally shallow types with watery convictions who want to bandwagon on every trend that comes their way). But they are also unwilling and unable to drop the stupid assumptions that were spoon-fed to them in university. They will try to bend these assumptions in all manner of different ways but at the end of the day they are just not up to discussing inequality in any rational or, dare I say, productive manner.

Update: This post was, as I indicated, partially a response to someone on Lars Syll’s blog. His name is Pontus and I had already told him quite clearly that he needed to dis-aggregate the data to get an idea of how wages and productivity are related. What followed was a rather nice illustration of bad data analysis. In the original comment I wrote:

Your regression results are spurious because you used too long a time period. Divide it up into six 8 year periods and you will get different results for each time period. This will suggest that you do not find a linear relationship and the relationship between wages and productivity is to be explained by some basket of confounding variables (to find out which ones, close Eviews and open a history book).

I thought that this was pretty clear. I quite clearly stated that if he divided up the data into six different time periods he would “get different results for each time period”. Pontus instead divided the data up into six time periods and ran a regression on these six datapoints.

In effect, he lessened the number of observations and then re-ran a regression he had already run. Why did he do this? What did he think he was proving? I’m not altogether sure. From the comments below it appears that he thought I was trying to “trick” him into getting worse results through using less datapoints. (That wouldn’t be much of a trick if you ask me).

So, what did I mean? Well, why not quote Keynes here in his critique of Tinbergen’s econometric study?

Put broadly, the most important condition is that the environment in all relevant respects, other than the fluctuations in those factors of which we take particular account, should be uniform and homogeneous over a period of time. We cannot be sure that such conditions will persist in the future, even if we find them in the past. But if we find them in the past, we have at any rate some basis for an inductive argument. The first step, therefore, is to break up the period under examination into a series of sub-periods, with a view to discovering whether the results of applying our method to the various sub-periods taken separately are reasonably uniform. (Professor Tinbergen’s Method, pp566-567)

So, how is this quote relevant to our present discussion? Well, the original argument was whether there was some sort of law-like relationship between wage growth and productivity growth. This is what a regression study should try to prove (or disprove). But, as Keynes implies, you cannot simply pile the data into a blender and then smugly read off the R-squared. Well, you can get away with this in the journals but they are void of any real empirics.

Rather you have to reorganise the data in line with your specific argument and then lay it out. That is what I did above. I said clearly that it was well-known that the wage growth:productivity growth ratio was far close together in the period 1965-1973 than it was afterwards. This suggested that something else (a confounding series of variables, to use statistics-speak) was causing their divergence in the other time periods. What’s more this something else must have been having enormous effects in different time periods because the two rates diverged so sharply. You can see this extremely clearly in the graph below that reproduces the first graph but puts in another indicator that shows the divergence of the wage rate and the productivity rate from one another.

As we can see, in the periods 1965-1973 the rates only diverged by about 1% while in 1973-1981 and 1997-2005 they diverged by 2.5-3.0%.

In economics if we want to make a meaningful causal argument we must be able to establish, as Keynes said, whether “various sub-periods taken separately are reasonably uniform”. If they are we can make a strong causal argument about the relationship (in this case that wages are set largely in line with productivity). But if they are not we can tell pretty much instantly that we are missing a key part of the story.

Anyway, the above confusions speak loudly to the problems with econometrics that Syll and I often highlight on our blogs. Most economists are not very good at establishing causal arguments and use advanced statistical methods as a crutch on which to prop up their flimsy arguments.

Update II: There were some problems with the new graph. This is a sometimes counterintuitive argument to make with data. All is fixed now. For real this time.

Update III: One final note before I’m done with this. What was the correlation that Pontus was picking up in his original regression? Why is it that wages do move to some extent with productivity? The answer to that is rather simple. It is because both wages and a certain measure of productivity are very strongly correlated to another variable. No prizes for guessing what: yep that’s right… Real GDP.

There was some divergence after 1980 but this is due to the ensuing relatively high unemployment years and we shall not discuss this here.

Why Sraffa’s Theory Does Not Contain a Labour Theory of Value

In my last post I tried to argue that, for a number of reason, once we make additions to Sraffa’s theory to make it comprehensive enough to confront the real world any potential interpretation of the theory in line with the labour theory of value falls apart. In that post, however, I never dealt wit why some people have sought out a labour theory of value in Sraffa’s work.

The reason for this seems to start in chapter III of Productions of Commodities By Means of Commodities. In this chapter Sraffa considers what impact variations in wage-profit distributions will have on the prices of goods in any given economy. I will first try to lay out the basic argument in as simple terms as I can.

Assume that all income goes to wages in a given economy. Now assume that different industries have different or heterogeneous inputs. Thus, while some need quite a lot of labour relative to other inputs, some need very little labour relative to other inputs. Now assume that we lower the wage rate by increasing the profit share. We must also, for simplicity’s sake, assume that the profit share is homogenous between industries. (I argued against this for other reasons in the last post but I am happy to concede on this assumption here for the sake of argument; note however that my criticisms in the last post with regard to the idea that firms that undertake superior innovations in marketing can add a higher mark-up does away with the labour theory of value anyway).

So, what happens when the wage rate is lowered to accommodate profits? Quite simply that those industries that have more labour intensive production will have a greater ‘saving’ than those industries that have less labour intensive production. This is a very intuitively simple argument. If Industry A has ten times the labour inputs that Industry B has then it will accrue ten times the cost-savings relative to Industry B from any reduction in wages. What’s more we have assumed a homogenous profit-rate so both Industry A and Industry B will have to pay the same profit-costs (in Post-Keynesian language, they will have to add the same mark-up). These two movements put pressure on Industry B to raise its prices relative to Industry A.

But the story doesn’t end there. In order to understand how relative prices are affected we also have to conceptualise what might be referred to as ‘knock-on effects’. You see, in Sraffa’s system, inputs are the product of other inputs. So, we have to take into account the fact that as relative prices change in line with different labour intensities these prices will be passed on to other industries. To conceptualise this imagine that there is an industry, call it Industry L, that has nothing but labour costs. Industry L will obviously see the greatest relative price decreases if wage rates are brought down. But it follows that those industries that more heavily use the inputs of Industry L will see more substantial reductions in their relative prices through this channel than industries that use these inputs less heavily. As the reader can appreciate, this all gets rather complicated rather quickly. The sharp-eyed reader will also note that at its essence this is the basis of Sraffa’s famous ‘reswitching’ argument.

Implicit in this argument is that labour is the variable input. This, I think, is why some people interpret Sraffa’s theory as a labour theory of value. They assume that since in Sraffa’s argument, changes in the wage-profit distribution is that which causes relative price changes, then what we are talking about is a labour theory of value proper. This is simply incorrect.

Although Sraffa does use the example of wage-profit distribution changes to discuss price changes we could equally discuss any number of other inputs. A good example might be energy. Since we must assume that almost all modern industries use energy this makes energy just as important as labour. So, we could equally explore the effects that changes in the price of energy due to, say, geological scarcity, have on prices.

A labour theory of value adherent might counter this by saying: “No, you misunderstand. The labour theory of value is concerned with distribution between human beings. When we change the distribution between wages and profits we change the relative amount that different classes of human beings receive of the social product. When we discuss energy price increases we are not discussing distribution in the sense that it is absurd to say that more income is distributed to ‘energy’ — a non-conscious, non-social entity.”

But this is a fallacious argument and is based on the nonrecognition of the demand side of the problem; one which is not dealt with in Sraffa’s book. In fact, we must assume that the different social classes consume commodities that contain different degrees of each input. In the case of energy (or food!) wage-earners will consume more of the energy-intensive commodities (heating oil, petrol etc.) out of their relative income share than their capitalist counterparts (for some empirical justification of this see the following post). Thus, as a price increase in energy inputs makes its way through the system, an effective redistribution is effected between wage earners and profiteers.

Clearly then, changes in any input price will affect the distribution of income (sharp-eyed readers will note here the essence of the Sonnenschein-Mantel-Debreu theorem). Once we take the Sraffian theory out of its vacuum and add a demand-side wherein different income groups consume different ‘baskets’ of goods (note, we need make no silly marginalist assumptions when doing this) then a change in any input cost will have redistributionary effects. This clearly shows that not only is the wage-profit relationship not the exclusive determinate of prices in a simple Sraffian economy, but it also shows that the wage-profit relationship is not the exclusive determination of income in that economy once we add a demand side — one which is implicit in Sraffa’s work anyway, which is clear when he discusses luxury goods in the second chapter.

So, even without taking into account the criticisms of the labour theory of value that I put forward in the last post — criticisms that rely on adding assumptions to the Sraffa model that generate greater realism — we can still say that the labour theory of value is false. Wages are just an input like any other — they are simply the price of labour. When the price of labour changes it affects relative prices and the distribution of income; but the same is true if the price of any other input changes.

This is not to say that wages are not probably the most important variable to consider when trying to understand income distribution. They probably are. But that has all of nothing to do with the labour theory of value.