The Fallacy of Spare Capacity

One of the reasons for my interest in various competing strands of economic thought is the rich supply of fallacies they provide. Some stuff, like the idea of spare capacity, heavily emphasised by post-war Keynesians or by more modern Post-Keynesians, is completely NUTS. But, by phrasing their exposition in certain ways and using particular not-so-obscure assumptions, their approach seem fairly coherent and convincing – one of the reasons that so many economists and non-economists pick it up, and among the reasons the public believes so many mistaken things about economics (The below refutation is not to say that there is no value of Post-Keynesian Economics – PKE – in general, an approach that has many virtues; the idea of pricing and spare capacity just isn't one of them).

The idea is this: standard classroom expositions of firm behaviour and pricing argues that marginal costs are increasing the more you produce, while consumers' marginal utility (and so their willingness to pay) is falling. That implies an optimal point of production, at MR = MC, where we find our beautiful equilibrium (extended by the NAIRU approach by Friedman and Co). At that point, any increase in demand – through government stimulus or central bank expansionary policies – can further reduce the unemployment level, but only at the expense of price inflation. This constraint, according to what I call spending-is-mending doctrines, is annoying: it justifies unemployment and it restricts what governments can feasibly do (the below analysis works for disequilibrium below-potential output too)

Hence, post-war Keynesians such as Samuelson, modern-day new-Keynesian commentators such as Paul Krugman, or more coherently the Post-Keynesians, found a way around this problem: spare capacity! In his 2014 tome on PKE, Marc Lavoie, one of the most well-known Post-Keynesians gives a bunch of reasons for why firms would want to maintain some spare capacity, and not produce at the level of output imagined by our classroom example, tracing such arguments to Big Name PKEs Kaldor and Kalecki (pp. 150-152). Although framed in much-too-conspiratory language and with nothing but manufacturing in mind, some of those reasons are plausible: uncertainty of future demand and a large aversion to turning down potential customers, heterogenous fixed-proportion capital goods (a point that is conveniently forgotten in other PKE expositions...) and more: 

There are good theoretical and practical reasons, as well as good strategic and technical ones, to explain why corporations generally aim at operating much below their full capacity, in the range of constant unit direct costs. Whatever one may think of the rationality of such behaviour, one must recognize, as post-Keynesians do, that firms consistently function with larger reserves of capacity, and that any analysis of the firm must take that fact into account. (p. 154, Post-Keynesian Economics: New Foundations)

Here's the graphical exposition. Production, in the PKE world, almost always happen below Full Capacity, where marginal costs are constant: 

Why do PKEs (and the Krugmans and Samuelsons of the world) go out of their way to show – empirically and theoretically – that there are spare capacity and constant marginal costs in production below spare capacity? They brilliantly combine this with another other of their cherished critiques, in order to produce the Mother-of-All policy attacks on sound economics: price formation. 

In the PKE approach to how prices are formed, there is hardly any resort to consumers – and when so, only indirectly as intermediate tool to show a loss of market shares to competitors. Prices are simply formed by a system of “cost-plus” or “mark-up” on costs, determined by market power and whether or not you’re a “price-leader” or “price-taker” in the market. We’ll let Lavoie explain this:

it is a good approximation to assume that firms fix prices by adding a constant percentage mark-up to their normal unit costs, calculated at a conventional rate of utilization of capacity and with a target rate of return in mind. The cause of this is that prices are not designed to be market-clearing prices; rather they are reproductive prices. The purpose of prices is to ensure that efficient or average firms generate enough profits through the cycle to finance or obtain finance for the required investments. (p. 167)

Here comes the brilliancy, since PKEs ultimately want to argue that extra government spending, ramped up aggregate demand or flooding the market with cheap credit will not create an increase in prices, but only move us towards their desired goal of Full Employment:

1. There is spare capacity and constant MCs in production
2. Firms set prices at mark-up over costs

(1) + (2) = increases in aggregate demand, pushing up the capacity utilisation closer to 100% (but not above it!) therefore won't affect prices at all. Voilà, the orthodox critique of stimulus that spending leads to (price) inflation) is averted. Spend away, guys.

If you think about it for more than two seconds, you realise that there's something not quite right here. If you've ever been to a bar, a gym, on a bus or train or plane, a restaurant or basically any other market production activity, you quickly realise that most of the time those establishments run with spare capacity – only to ramp up production on a Friday night, holiday season or after-work rush at the gym. Gene Epstein, in his beautiful introduction to the collection of essays by Rothbard titled Economic Controversies, brilliantly foreshadowed this idea here, and points out that Rothbard did so in criticising Samuelson's example of barbershops decades ago:

Professor Samuelson may have been missing something. Given his flexible work schedule, he may have had a habit of going for his haircut on a weekday, which would explain why he kept noticing empty chairs. Had he gone instead on Saturdays, he might have noticed that all the barber chairs were full, and that business was actually backed up. It then might have occurred to him that our hypothetical Marioni Brothers were not so dumb as to waste their money on excess capacity. The problem they actually faced as businessmen was the classic tradeoff between peaks and troughs in demand. Had they not had empty chairs during the week, they wouldn't have been able to take advantage of the glut in demand on weekends." (pp. x-xi)

Referring to Rothbard's explanation in his 1962 Man, Economy and State that intentionally planning for excess capacity must be some form of schizophrenia, Epstein continues: "It made no more sense to believe that all such businessmen would waste funds on excess as it was to believe that they would all consistently underinvest and plan on inadequate capacity" (p. xi)

Or, let's take an even more extreme example for illustration: As anyone who's ever been to Auburn, AL knows, there's a massive football stadium –Jordan–Hare Stadium – capable of seating almost 90 000 people. Most days of the week, and even entire weeks of the year (non-football season), it's rather empty; they have insane spare capacity. Calculated the way Krugman or PKEs would do, they probably operate at less than 20% capacity on average (with above  80% spare capacity) But there is no way that can approach 100% unless you invent new football seasons or add more games to each season. No monetary or fiscal stimulus in the world is gonna change that – at best, it'll fill up the seats on gameday a bit quicker, bid up the prices of the tickets or ration the seats for already filled-out games. 

This extreme example applies for virtually any other business. Spare capacity is not a market inefficiency you can manipulate and push up by government stimulus, thus lowering the unemployment level. Businessmen, as Rothbard pointed out over half-a-century ago, don't hold spare capacity for no reason: the production facilies are being used for the purpose of servicing peak demand – not to maximise some estimated capacity calculation by various economists; if the costs of keeping them "idle" is more than made up by the revenues from peak demand, there is no irrationality or waste of resources occuring. 

But not even on their own account does this spending-doesn't-cause-price-inflation-doctrine work. I remember reading such infamous notions as "Scarcity is put aside, while that of reproducibility is put to the forefront"; in PKE tradition, prices are not "an index of scarcity in general; rather, prices reflect the unit costs of producing these reproducible goods or services." (see Lavoie, p.23). Let's put aside our critiques for such nonsense for a moment. The emphasis of reproducible goods implies that some goods are not reproducible, such as raw materials. I remember Bober's Alternative Principles of Economics emphasising precisely that point. This creates a problem for their "spend-away" solution of spare capacity; if utilisation levels through an increase in demand jumps from the currently-desired level to some higher level, firms will be looking to expand capacity in order to move utilisation rates down again (remember all those reasons for why firms produce below full capacity?). Hence, they'd need more material and more capital goods, increasing the demands on their suppliers, back in the structure of production. Soon enough, the stimulus will reach raw materials, which are non-reproducible and so their prices are sensitive to increases in demand. But, since firms in the PKE-world set prices at mark-up over cost, and raw material enter into that equation as costs (directly or indirectly) this undermines the entire reasoning: even in their own framework, an increase in aggregate through government stimulus increase prices for goods on the market. 

The bottom line is this: speaking of spare capacity in the economy as a whole or for a single business is meaningless. Arguing, like Post-Keynesians do, that stimulus won't cause price inflation because we're below full capacity is completely detached from the reality and even counter to their own theoretical framework. It matters not if we're close to 100% capacity or not, what matters is how businesses deal with troughs and peaks in consumer demands. Spare capacity, as is used in economic models, is a worthless concept.