Posts Tagged Demand-Supply
Is it really feasible that there can be a single theory of ‘the market’ that encapsulates everything from tomatoes to CEOs to houses? Engineers do not think they can apply the same theory to every fluid, and similarly, it is not unreasonable to suggest markets might function very differently depending on what is being bought and sold. In this post, I’ll set out a couple of alternative interpretations of supply and demand for different markets. I developed these alternative approaches based on some well known real world observations.
A few caveats:
(1) I am not interested in deriving these schedules ‘rigorously’ from arbitrary axioms about individual behaviour. Such an approach is unnecessary as phenomena may be emergent, and it always seems to run into problems.
(2) These models should really only be interpreted as working for individual markets on a small scale, as large scale feedback effects render this type of analysis irrelevant.
(3) I am aware that there is no such thing as a demand or supply ‘curve’ in reality. Perhaps the fact that I feel the need to reduce everything to intersecting lines is a testament to how much neoclassical economics has polluted my thought. I am undecided as to whether I regard the supply-demand framework as a useful tool that can be adapted in certain circumstances, or as something that needs to be done away with entirely. I’m sure heterodox readers can give me many reasons demand-supply as a concept is not worth keeping. In any case, I regard these examples as interesting, and at the very least they are a good way to take economists on on their own turf.
(4) Finally, I apologise for my drawings, which were constructed on MS Paint.
It is an observed reality that asset prices and demand are often positively related, since in many cases an asset is purchased for no other reason than selling it later on at a higher price. Price increases can act a signal for later price increases, and the opposite is also true. Hence, we will posit a positively sloped demand curve for this model. This relationship will also be exponential: at low prices, the effects will be relatively small ,but as prices spiral , the effect will get larger and larger. Since the supply of assets is relatively inelastic (in the case of land, perfectly inelastic), the supply curve will be a steep upward sloping line.
From the diagram constructed based on the observations above, we can deduce a few interesting mechanics. First, there large number of demand-supply combinations for which there is no equilibrium. If demand is to the left of the supply schedule, supply will always exceed demand and prices will fall to zero. The most likely cause of this is simply that a company or industry is not performing well, although tight monetary policy could also do it. If the demand curve is to the right of the supply schedule, demand will always exceed supply and price will spiral upward indefinitely. This could be due to excessive credit expansion or misplaced expectations.
However, between these two extremes there is a potential ‘golden zone’ for which equilibria exist. In this zone I have drawn two potential demand curves: D2, which just touches the supply curve, and hence has one potential equilibria; and D3, the likes of which would be more common and would have two equilibria. The two lower equilbria, e1 and e2, are not stable. Any drop in price will result in excess supply that will drive prices down the schedule. An upward increase in price from e1 or e2 will result in excess demand that will continue to increase price in a spiral. In the case of e2, it will propel the market up to the one stable equilibrium: e3. If the price increases, supply will exceed demand and it will quickly fall back to e3. If it decreases, demand will exceed supply, the price will rise and the market will again tend toward e3.
The volatile behaviour displayed by most outcomes in this model is in accordance with much real world experience in the stock market, from the downward spiral of Facebook to the internet stock bubble and other frequent historical experiences. But what about e3? Is it the case that certain firms or industries exhibit relatively high, consistent returns? In fact, relatively stable share prices do exist for some firms and industries: ones that are rarely hit by volatility. Insurance, transport and many consumption goods are all stable industries. Obviously this doesn’t protect them completely: firms, industries or the market as a whole may exhibit relative tranquility before something knocks them over the precipice.
We can draw a few policy conclusions from this framework. There is scope for a central bank to reduce or increase the liquidity of the system in order to try and knock the market into the ‘golden zone.’ The effects of an FTT are indeterminant, which is actually the only conclusion I can garner from the available evidence. However, the primary conclusion I would come to – which admittedly doesn’t flow completely from the model – is that due to the level of volatility, the trading day could be be extremely limited, or trade could be cut off if prices are rising or falling too fast. This way the price spirals can be curbed while the central bank adjust liquidity, and investors and funds adjust their positions, trying to shift the demand curve back into the golden zone (though I am willing to admit the last part verges on a spurious level of precision).
Another market that would be expected to diverge strongly from the ‘norm’ would be the labour market. Many aspects of labour make it different to other markets: it is required to subsist; it cannot be separated from a person; it is difficult to determine the productivity of potential applicants, or even present workers; time spent in work is related to leisure time. Below I’ll outline two alternative approaches (that are also somewhat compatible).
A more comprehensive approach to labour supply, which takes into account the need for subsistence, has been provided by Robert Prasch (whose book, which inspired this post, is recommended). His approach is illustrated in the diagram below:
This can be seen as an adaption of the typical ‘backward bending‘ labour supply curve found in economics textbooks. The subsistence frontier shows the total pay required to subsist at various wage levels, whether that is interpreted as literal subsistence or a conventional and socially acceptable level of income. Obviously, the higher the wage, the fewer hours required to work to reach subsistence.
Equilibrium C is unstable as either side of it creates a wage spiral toward the next equilibrium. If the wage decreases, the labour supply will be forthcoming as workers seek subsistence level pay. At Wu they hit the limit of how long they can work and give up, settling for a ‘poverty trap’ wage at D. If the wage increases above C, wages become less necessary for subsistence and workers will withdraw their labour while remaining on the subsistence frontier. Once the subsistence frontier becomes a distant memory and wages are high enough, workers will be willing to devote more time to work in order to purchase more goods and services.
The highest equilibrium, A, is also unstable. Should the wage deviate upward, there will be a resultant upward spiral as demand always exceeds supply. Should it be reduced or capped, the wages will settle down toward a lower level. This is consistent with the observed behaviour of CEO, footballer and celebrity pay in recent decades, which shot up when marginal tax rates were significantly reduced (and a direct cap on footballer’s wages was lifted).
Clearly, a couple of evil communist interventions can be proposed based on this framework. The first is a minimum wage at or above Wc to boost the market into the desirable area. The second is steeply progressive taxation above Wa to prevent pay from spiraling Such interventions would help to enhance labour market stability, equality and reduce poverty.
I have also constructed an alternative model of the labour market. My model is shown in the diagram below and illustrates what a labour curve might look if we included the assumption of the ‘conventional’ working day/week which characterises so many people’s lives. This may be enforced legally, by social norms, or by the power of capital of labour. Whatever the cause, it is an undeniable empirical reality.
This model of labour supply could perhaps be interpreted as an elaboration or magnification of the zone between Ws and Wh on the S-shaped labour supply diagram: that is, what happens once wages reach an acceptable level. In fact, I adapted the model from one found in Arthur Lewis’ 1972 paper (p. 10) on unlimited labour, which focuses on a wage given at a conventional level. This is compatible with the above diagram.
However, I don’t wish to stretch the comparisons, as this model can also be thought of in isolation. The point is that a worker will accept work at any ‘reasonable wage’ up to a certain amount of hours – I have suggested 40, which is the norm in most developed countries. After this point the time they have normalised as leisure time is far more valuable and they expect to be paid more at an increasing rate. Eventually, they reach the physical limit of work and the effective wage must be infinite to induce labour. There are multiple equilibria if the demand curve crosses the feasible wage zone, but only one if it moves beyond it.
There are a couple of conclusions we can draw from this model. The first is that a tax on labour will not alter the amount of labour supplied if the tax + the wage is within the ‘feasible wage zone.’ In fact, empirical evidence suggests the effect of income taxation on labour supply is negligible at most (though higher for women – perhaps this analysis only applies to primary earners). Second, an increase in demand need not increase wage inflation up to a certain point, provided there are labourers willing to work within a given wage range. Past this point, an increase in demand will cause a wage-price spiral. It’s harder to verify this point empirically, although it seems consistent with the frequently observed point that inflation isn’t a problem when unemployment is high (having said that, it’s by no means the only model that predicts this).
I think a pluralistic approach such as this would be interesting. It would not be wedded to any particular model, and the task of economists would be deciding which model to apply or devising alternative models to deal with a situation. In fairness, DSGE is characterised by this approach, so it would be unfair to suggest economists do not use it at all. Nevertheless, this calls into question the rigid supply-demand framework that pervades much policy discussion of price controls, taxation and various other ‘interventions,’ and the catch-all arguments made by some economists against them.
The conventional ‘upward sloping’ supply curve is known by everyone from an econ101 student to a professional economist. The curve posits a positive relationship between price and quantity supplied – in order to increase the quantity supplied, a higher price must be offered. What is less commonly known, however, is that an upward sloping supply curve is actually incredibly hard to justify, both in theory and in practice.
Behind the curve lies the proposition that production costs per unit increase as output increases. This makes the ‘upward slope’ a necessity: since an item costs more to produce, a higher price must be charged as production increases. Microeconomic theory posits that this relationship holds in both the short and long run. However, all signs say this can’t be true.
The Short Run
The neoclassical idea is that firms can only increase one factor in the short run, which gives birth to the increasing marginal cost argument – returns to production diminish as more and more is ‘squeezed’ out of the fixed input. Is this how production behaves?
I have previously commented on Piero Sraffa’s excellent critique of the idea of increasing marginal costs in the short run. Sraffa’s argument was two pronged, depending on how we define a ‘firm’ or ‘industry.’
Sraffa argued If we define an industry narrowly, such as a single firm, it turns out firms generally have a lot of spare capacity and can quickly employ previously idle resources to expand all inputs at once. This belies the traditional justification for decreasing returns – if we expand inputs simultaneously, we should expect roughly constant increases in output as a result.
Sraffa went on to argue that we could define an industry broadly enough that it’s reasonable to say a factor is fixed in the short run. This would be because, for a large industry, a new factor would have to be converted from other uses before it could be employed. Hence, diminishing returns may be possible. However, at this point it is no longer possible to neglect the collateral effects caused by changes in firm’s expenditure and output: the partial equilibrium method becomes contradictory, and the various curves – demand, supply, average cost – cannot move independently, which is a key assumption of the theory. So the theory as a whole is no longer appropriate.
So the idea doesn’t hold up in the short run, except in the extremely small number of cases that lie between the ‘narrow’ and ‘broad’ definitions. But what about the long run?
The Long Run
If you learn producer theory from the bottom up, one of the assumptions you start with is that inputs and outputs are infinitely divisible; in other words, they are like clay. They are also homogeneous, and available at a set price. Based on these assumptions, it is reasonable to assume that in a short run – when one factor is fixed – there may be increasing marginal costs. At this point I would defer to Sraffa’s above critique.
However, when we move to the long run, it’s incredibly hard to justify increasing MC even within the confines of the theory. Textbooks will generally assume the standard production function, which looks like this:
The downward sloping portion – for which costs fall as output rises – will generally be justified by ‘Economies of Scale (EoS).’ But what causes EoS? Bulk buying is one example, but this can’t apply because prices are taken as a given. Another is indivisibilities – if you buy a big machine it takes a while before you fully utilise it – but this quite clearly contradicts the assumption of perfectly divisible inputs. Yet another, the increasing returns to the division of labour, contradicts the assumption of homogeneous inputs.
Similarly, the upward sloping portion is then justified by ‘Diseconomies of Scale (DoS).’ Examples of this are generally few and far between – DoS is, after all, the strange proposition that firms simply become incompetent at some level – but again they tend to contradict our assumptions. One example is managerial difficulties – who is the manager if labour is homogeneous?
In fact, it turns out that few, if any, of the explanations for either EoS or DoS hold up under the available assumptions. If you increase a homogeneous perfectly divisible mass of inputs a certain amount, there is no reason to expect anything other than a constant, proportional increase in put: in other words, constant returns to scale.
Unsurprisingly, it is true that the overwhelming majority of firms report constant or falling returns to scale. Walking down a high street in a capitalist country, it’s hard to deny that firms have the available goods to accommodate an increase in demand without a rise in cost; factories are designed in a similar fashion. Furthermore, in the long run a firm is likely to respond to an increase in demand by opening up more branches, rather than simply increasing prices.
So what’s the problem? The proposition that demand determines outputs and supply determines price is logically, intuitively and empirically reasonable in any time period and for most industries. Why can’t economists just tilt their supply curve 45 degrees to the right?
Well, at this point a few things become apparent. The theory of the firm becomes indeterminant: one of economist’s beloved negative feedback loops that allows the economy to self-equilibrate is gone, as firm size is not limited by production costs. Hence, the marginal theory of the firm goes from explaining everything – from firm size to income distribution – to explaining very little. This also makes explicit the idea that output in the economy is driven by demand, both in the short and long run, which contradicts conventional macro theory, where demand only matters because prices are sticky.
Overall, a flat supply curve turns the conventional story told in neoclassical economics, where the economy is self-equilibrating, bar a few frictions, to one where many key variables – wages, output, firm size – go from being at the equilibrium or ‘natural’ level, into one where they are largely arbitrary. It’s easy to see why economists would resist this.
Naturally, mainstream economists have been critical of Steve Keen’s Debunking Economics. I will do a brief series within a series to try and respond to some of these criticisms. In this part, I will respond to some of the main critiques of neoclassical theory that have generated controversy: demand curves, supply curves and the Cambridge Capital Controversies. In the next post, I will respond to criticisms of Keen’s own models and his take on the LTV, as well as anything else that has attracted criticism.
Note that this post will assume prior knowledge of Keen’s arguments, so if you haven’t yet read my summaries above (or better still, Keen’s book), then do it now.
It seems there are some problems in this chapter. Keen mixes up some concepts and misquotes Mas-Collel. Having said that, he is broadly right. This is frustrating for someone on his ‘side,’ because it means mainstream economists can dismiss him when they shouldn’t.
Keen presents a quote from Mas-Colell where he assumes a benevolent dictator redistributes income prior to trade, and asserts that this assumption serves to ensure market demand curves have the same properties as individual ones. In fact, Mas-Colell is using this assumption to ensure that a welfare function, not a price relationship, will be satisfied. It remains true that a PHD textbook still assumes a benevolent dictator redistributes resources prior to trade, and subsequent economists have also used this assumption, which is not a great indicator of the state of economics. However, it was not an assumption used to overcome the Sonnenschein-Mantel-Debreu conditions.
More importantly (wonkish paragraph), it seems Keen lost some nuance in the translation of his critique to layman’s terms. He spends a lot of time talking about the Gorman polar form. This is about the existence of a representative consumer for a set of indirect utility functions (‘indirect’ because it calculates utility without using the quantities of goods consumed), but Keen makes out it is about the aggregation of preferences required for demand curves. Gorman is in many ways similar to, but not relevant to, the discussion of the aggregation of demand curves. Keen also argues that consumers having identical preferences is the same as them being one consumer, but this needn’t be the case: just because you and I have the same preferences, doesn’t make us the same person
Despite this, the competing wealth and substitution effects do create the conditions described by Keen. However, they only apply under general equilibrium – under which wealth effects are present – and not partial equilibrium – under which they are assumed away. Keen does not distinguish between the two.
In summary, Keen is correct that neoclassical economists could not rigorously ‘prove’ the existence of downward sloping demand curves. Keen himself says that it is reasonable to assume that demand will go down as price does, and classical economists were also content with this an observed empirical reality. Neoclassical economists themselves ended up having to defer to empirical reality when faced with the SMD conundrum and thus they had gained no insight beyond the classical economists, except to prove that their preferred technique – reductionism – does not work. For this reason, I interpret the SMD conditions primarily as a demonstration of the limits of reductionism (though some fellow heterodox economists might disagree).
The proposition here is pretty simple: a participant in perfect competition will have a tiny effect on price. This is small enough to ignore at the level of the individual firm, which is neoclassical economist’s main defence. However they ignore that, as Keen says, the difference is both “subtle and the size of an elephant.” Once you aggregate up a group of infinitesimally small firms making incredibly small deviations from maximising profits, you get a result that is far away from the one given by the neoclassical formula. Result? We must know the nature of the MC, MR and demand curves to know both price and quantity, just as with a monopoly. The neoclassical theory – at this level – has no reason to prefer perfect competition to a monopoly, and a supply curve cannot be derived. From what I’ve seen, the critics ignore the effect of adding up tiny mistakes, instead focusing on how tiny they are on an individual level.
Economists have some other defences, but I interpret them as own goals. For example, there is the argument that, under perfect competition, firms are price takers by assumption. They cannot have any effect on price, by assumption. But this basically amounts to assuming the price is set by an exogenous central authority, which is odd for a model of perfect competition.
Another argument is that setting MC=MR itself is an assumption. This is a strange path to take for a theory that prides itself on internal consistency and profit maximisation. It acknowledges that MC=MR will not quite maximise profits, so amounts to the assumption that firms are not profit maximisers. There is also the similar argument that firms don’t take Keen’s problems into consideration in real life, so they don’t matter. This is a huge own goal, given most textbooks argue that it doesn’t matter what firms do in real life. I’m quite happy to acknowledge it does matter how they actually price – but that would involve abandoning the marginalist theory of the firm and using cost-plus pricing.
So, now that we have all finished discussing how many angels can dance on a pinhead (turns out it was slightly fewer than economists thought), let’s just start using more realistic theories of the firm and forget the mess that is marginalism.
Cambridge Capital Controversies
There are swathes of literature on this and I cannot hope to explore them all. The main thing I have noticed, and want to discuss, is that economists only seem to focus on capital reswitching when discussing this, and defer to empirical evidence to suggest it is negligible. I have a few problems with this:
(1) Empirical evidence is competing and some evidence suggests reswitching is more common than economists would like to think. Furthermore, it is incredibly hard to observe and therefore cannot be dismissed so easily.
(2) Most importantly, the Capital Controversies were not just, or primarily, about reswitching. Sraffa showed a number of things: demand and supply are not an adequate explanation for static resource allocation; the distribution between wages, profits and other returns must be known before prices can be calculated; factors of production cannot be said to be rewarded according to ‘marginal product’. For me these are more important, and are applicable to many models used today, such as Cobb-Douglas and other production functions, and the Solow Growth model.
With all 3 of the examples I have discussed, economists have tried to defer to empirical evidence to dismiss the problems with their causal mechanics. But generally economists do not regard empirical evidence about causal mechanics as important (the primary example being the theory of the firm), instead insisting on rigorous logical consistency. Surely, in order to be completely logically consistent, economists should at least be willing to experiment with the potential effects of SMD and reswitching in general equilibrium models and see what happens? Robert Vienneau has various discussions of this.
The common thread between these is that economists seem incredibly adept at assuming their conclusions. Of course, you can get around any critique with an appropriate assumption, but as I’ve discussed, theories are only as good as their assumptions and assumptions should not be used simply to protect core beliefs and come to palatable conclusions. Having said that, Keen’s book isn’t perfect (which is to be expected if you try and take every aspect of economics on in one book), and there are worthwhile criticisms out there. Nevertheless, Keen’s critique as a whole remains in tact, and leaves very little of what is taught on economics courses left in its wake.
P.S. Feel free to use the comments space to discuss any critiques of areas I have not covered/said I will cover.
Chapter 5 of Steve Keen’s Debunking Economics explores the marginalist theory of the firm. Keen first channels Piero Sraffa’s 1926 criticisms, then catalogues the neoclassical theory’s complete lack of real world corroboration – as noted in my title, a businessperson once referred to it as “the product of the itching imaginations of uninformed and inexperienced arm-chair theorisers.”
The neoclassical theory of the firm supposes that, in the short run, firms face increasing marginal costs – their costs per unit (average cost) increase as they produce more. This occurs because in the short run, the ‘amount’ of capital (and land) employed is fixed, so producing more involves squeezing more and more out of machines with more labour. The intersection of these increasing costs with how much they can gain from selling more, or their ‘marginal revenue’, constrains their size.
This homogeneous treatment of capital should strike many as silly. The neoclassical theory effectively supposes that, if we employ 9 people to dig a ditch with 9 spades, employment of the tenth will split the 9 spades into 10 slightly smaller, worse spades. However, if new labour is employed, new capital is – must be – employed simultaneously, whether it is bought or if it is taken from previously idle capacity. A taxi driver cannot do anything without his taxi; an office worker without a computer is also fairly useless.
So increasing marginal costs are unlikely to be the case with individual firms or narrowly defined industries. As Keen puts it, “engineers purposely design factories to avoid the problems economists believe force production costs to rise.” In reality, firms have excess inventories and tend to vary capital, labour and land all at once, even in the short run. They therefore face roughly constant, or falling, returns to scale.
Sraffa pointed out that it’s only really valid to treat some factor inputs as fixed if we define an industry so broadly that the factors would have to be converted from other uses. For example, if we take agriculture, and assume the country is well populated and at or close to full employment, then it’s reasonable to treat land and machinery as fixed in the short term. However, since the theory of the firm assumes that supply and demand are independent and that one ‘industry’ can be studied apart from all others, another problem appears: this situation does not lend itself well to ceteris paribus analysis. Changing wages, supply costs, and the displacement of labour from other areas will have notable impacts on the rest of the economy, such that tinkering with our curves individually cannot be deemed a proper representation of what will happen.
There are a few cases where firms or industries might fall between these two categories, but really they are the exception.
Keen cites 150 empirical surveys that found firms reporting constant or falling average costs of production. In particular he cites Eiteman and Guthrie, who found that 95% of firms out of 334 did this, whilst only 1 chose the curve that looks like the one found in textbooks. Most firms also use cost-plus pricing, rather than taking marginal considerations into account, and adopt a form of trial and error when pricing.
A flat(ish) supply curve leads us to the incredibly interesting proposition, supported by the classical economists, that supply determines price while demand determines quantity. This is, of course, a simplification ,but appears to corroborate far better with the real world than neoclassical ‘simplifications.’
In my opinion this is the strongest case against neoclassical micro as taught. Jonathan Catalan can find no objections to this section, either, and gives the story an Austrian slant. Keen says that this problem has never really been addressed by economists, but ignored, despite the clear superiority of Sraffa’s logic and the corroboration of the empirical evidence with his approach. I find it hard to believe neoclassical economists can wiggle their way out of this problem, should they ever address it.
‘Debunking Economics’, Part II: Perfect Competition, Profit Maximisation and Non-Existent Supply Curves
The second chapter of Steve Keen’s Debunking Economics explores a number of arguments: the incoherence of perfect competition; that idea that equating Marginal Cost (MC) to Marginal Revenue (MR) does not maximise profit; and, eventually, that a supply curve cannot be derived. I will offer a brief summary of each of these arguments, but I encourage further reading (Keen’s book being the obvious candidate), as this is complicated stuff and a blog post can only serve as an introduction and overview.
Keen’s first point is that, under perfect competition, the demand curve for an individual firm is not horizontal, as taught in economics textbooks, but the same as the market demand curve. Analysis of perfect competition makes the basic mistake of confusing infinitesimally small firms with firms whose size is 0 – in other words, it says that a market demand curve can be split into an infinitely small amount of flat demand curves. However, if you add up any number of flat demand curves, the result will be a flat demand curve, not a sloped one. Therefore the demand curve for any individual firm must be sloped, however shallow the slope is.
Again, it was a neoclassical economist, George Stigler, who discovered this flaw in perfect competition. Stigler’s argument is that since, by assumption, firms do not react to each other’s strategies, any change in output by a firm will change market output by the same amount, and hence affect price. This means the demand curve for an individual firm cannot be horizontal (a change in output does not affect price), but must be the same as the market demand curve (a change in output changes industry output by the same amount).
Keen’s discussion of perfect competition also notes something obvious – the use of the word ‘perfect’ is obviously value laden, despite economist’s claims that their science is value free. I’d add that whilst ‘perfect’ may have a specific definition, the broader value judgement of ‘competition is good’ is undeniably present in economics.
Marginal Cost =/= Marginal Revenue
Keen goes on to argue that the basic neoclassical theory of the firm, that firms maximise profit where Marginal Cost (the cost of producing one extra unit) equals Marginal Revenue (the revenue received from selling one extra unit) is incorrect. This is because it is vulnerable to a fallacy of composition – whilst it is rational for individual firms (at least according to neoclassical principles), it is collectively irrational for an industry, and will result in firms losing money if they all pursue it as a strategy.
The neoclassical profit maximising formula only focuses on the effect of changes in a firm’s own output on revenue, ignoring the impact of changes in industry. Whilst perfect competition assumes firms do not change their output in response to one another, the above result shows that industry output will change by the same amount as a firm’s output. If I’ve interpreted Keen correctly, the reduction in price resulting from this industry change (assumed away by neoclassical theory) is missing from neoclassical formula, so revenue will be lower than predicted, and equating MC and MR will yield a loss.
Is there a supply curve?
It is well known that a supply curve can only be derived for a perfectly competitive market – if firms are price takers. However, once firms have some market power, price cannot be taken as a given, because if they produce more (less), the price will fall (rise), meaning MR and demand diverge. Once this happens, MC will equal MR and not price; a change in price will not cause a firm to move smoothly along its MC curve (which is the supply curve), but instead will depend on MC, MR and demand.
As Keen notes, this explains why economists have been so keen to cling to perfect competition, with its blatant lack of real world corroboration and its seeming incoherence. However, Keen’s own arguments suggest that even under perfect competition, firms have some impact on industry output, and MC cannot be equated to price without making a loss on some sales. Therefore, unless individual firms behave irrationally – something that is obviously contrary to core tenets of economic theory – a supply curve cannot be derived as taught in economics textbooks.
I had a hard time getting my head around this but eventually became convinced of Keen’s arguments. Once you take into account a perfectly competitive firm’s own impact on industry output, the standard analysis of MC = MR breaks down and all the problems with deriving a supply curve, previously assumed away by economic theory via perfect competition, return. This bears something of a resemblance to the problems with demand curves, which were well-known but assumed away by Hicksian demand functions, only to return once you introduced more than one consumer.
The general impression Keen has given me so far is that economic theory is disturbingly aware of its own flaws.
The first substantive chapterof Steve Keen’s book Debunking Economics explores the idea that the demand curve for a market does not necessarily slope downwards, as is the norm in economics textbooks. Instead, once you move past the individual level, then according to neoclassical principles a demand curve can have any shape at all, as long as it doesn’t double back or intersect itself.
This idea was first expounded by the (neoclassical) economist William Gorman in 1953. Gorman, on discovering this, went to great lengths to introduce assumptions that nullified the result, but these assumptions effectively amounted to assuming there was only one consumer, what Keen calls a ‘proof by contradiction:’ starting from an analysis of market demand curves, but eventually having to assume the demand curve is not for a market but for an individual. Keen does an impressive job of communicating this argument, which is obscurely and abstractly stated in Gorman’s paper.
Keen starts with the observation that price changes can have many different impacts, due to the interlinked nature of markets. Demand curves depend on the condition that income remains constant, so that we can study changes in price independent of other effects. However, if we consider somebody buying a ‘basket’ of goods, an increase in the price of one good might make another good unaffordable. The resultant boost in income (from not buying the second good) may well increase demand for the good whose price went up. The various interactions between goods depending on whether they are necessities, luxuries and so forth can create some interesting looking demand curves.
Economists are well aware of this problem, and have of course managed to assume it away! This is known as the Hicksian compensated demand function. I won’t go into too much detail here, but this involves adjusting income until utility is the same level as before, then allowing for adjustments in income. This allows us to separate out the ‘income effect’ (the one listed above) and ‘substitution effect’ (the ‘pure’ effect of price on demand).
The problem is that once you introduce more than one party to the economy, it is impossible to separate out changes in income from changes in demand, as one person’s spending is another’s income. This interplay between spending and income brings back all the problems glossed over by the Hicksian demand function. This is well documented, and known as the Sonnenschein–Mantel–Debreu theorem.
Naturally, economists have managed to assume this away, too. But here their assumptions jump the shark. Here is Gorman in his 1953 paper:
The necessary and sufficient condition quoted above is intuitively reasonable. It says, in effect, that an extra unit of purchasing power should be spent in the same way no matter to whom it is given.
Even when stated in this form, this is obviously not a reasonable proposition. If you give Peter money, he’s not going to spend it in the same way as Paul. It’s really that simple. However, Gorman is slightly disingenuous with this statement, as the real conditions are laid bare later on in the paper:
The older work of Allen and Bowley was based on the assumption that the classical Engel curves for difference individuals at the same prices were parallel straight lines, but this has been rejected in the more recent work of Houthakker in favour of a doubly logarithmic form. However, the earlier assumption fits the data remarkably well.
Keen points out that since all utility functions pass through (0,0) (zero consumption yields zero utility), and all parallel straight lines that pass through the same point are the same line, this amounts to an assumption that consumers are all exactly the same – effectively, that there is only one consumer.
In other words: the only way we can make a market demand curve the same as an individual demand curve, is if we assume that it is an individual demand curve. There is an obvious logical problem with this approach.
The funny thing about this chapter is that Keen concludes that, generally speaking, “there are reasonable grounds to expect that…demand will rise as price falls.” So why go to all the trouble?
The first reason is to show that ‘more is different’ – theories should not necessarily be built up from individual behaviour. Keen notes that the assumptions that all consumers are the same is only defensible if one analyses from the classical perspective of different classes, something the neoclassicists were trying to avoid.
The second reason is to show that, despite economist’s assertions to the contrary, it cannot be proved that a market economy will necessarily maximise social welfare, as, even on their own terms, it is logically possible to have multiple equilibria, some of which are more socially desirable.
The third reason is that the SMD conditions also establish that it is impossible to measure these things independent of the distribution of income, highlighting more general problems with neoclassical ceteris paribus analysis.
I myself experienced some cognitive dissonance when Keen came to that conclusion (I don’t know what it is about studying economics that creates this), but the fact is that this kind of logical inconsistency must be exposed, and, contra Friedman, this does not depend on whether the conclusions are completely sound or not. As I noted in my opening, these problems are all well documented by neoclassical economists themselves. So how can they excuse ignoring them?
If somebody presented you with a static snapshot of weather patterns, it would be clear that the model was fairly useless; as it didn’t capture dynamics, it would have nothing to tell you about the weather. Similar problems apply to neoclassical models: once you attempt to incorporate dynamic events, they don’t just become ‘wrong’, they become completely irrelevant.
I posted recently about how CA is irrelevant for developing countries, but I’d like to expand: it is completely irrelevant for arguments about protectionism. The problem is that if you take into account the effects of tariffs on the productivity of the industries they are aimed at, it has nothing to say – not just for developing countries, but for any industry whatsoever. CA assumes that every country has an innate productive capacity in each industry that does not change over time. If you froze the world, CA might be a persuasive argument for free trade, but in a dynamic economy it completely irrelevant.
Say the price of a necessity goes up due to a supply shortage. Modelling this as a simple ‘price increase’ suggests that demand would go down. However, if people are expecting the supply shortage to continue or worsen, then isn’t it more probable that demand will go up? Mainstream economists might have an answer: the price increase can be modelled as a movement of the supply curve, whilst the new information about the supply shortage can be modelled as a movement of the demand curve:
(D1 to D3, S1 to S2)
Problem solved. Except this movement of the demand curve leads to a higher price, which in turn would cause people to alter their expectations of the shortage, leading to another movement, and so forth. This may be a highly specific example, but it touches on a central Sraffian criticism of these models, which is that the curves cannot move independently; a change in one creates ripple effects that violate ceteris paribus. Thus, taking a picture of the state of them at any one time tells you as much as a photo of a moving train tells you its velocity.
Both ‘curves’ are partially derived from expectations – one from expectations of returns on investments, and one from expectations of future needs for liquidity. Therefore, a similar criticism to Demand-Supply applies – movement of one curve alters expectations and so affects the other. This creates a feedback loop that simply cannot be captured by two intersecting curves. At any one moment, the diagram might be said to be ‘right’ (putting aside other objections), but this doesn’t mean it is useful.
I expect economists won’t appreciate a whistle-stop tour of their models that claims to have debunked them, but at the same time I expect they’d agree that the above ‘weather’ example would so obviously flawed that it would not need to be refuted formally. In order to avoid special pleading, economists will have to argue that the economy is at or close to equilibrium, rather than a dynamic system. I do hope nobody claims this after 2008 (or the recurrent crises for centuries before that).