“Introduction to Post-Keynesian Economics”, part 1: On Orthodoxy and Micro
Today starts a series of articles summarizing the key ideas of Marc Lavoie’s “Introduction to Post-Keynesian Economics”. Enjoy!
Lavoie was inspired to write this book after a visit to France. In bookstores, he found many economics books condemning “what the French call ‘la pensée unique’ (the single thought), or what the English-speaking world knows as ‘There Is No Alternative’ (TINA)”. The latter was a slogan popularized by Margaret Thatcher, claiming that free markets, free trade, and austerity were the only serious policy option. However, Marc found that these books didn’t present a coherent alternative to the neoclassical theories that supported TINA. Instead, many economists simply created more sophisticated neoclassical models, rather than starting from a fundamentally different basis. That basis, to be presented here, is post-Keynesian economics.
This book focuses on rebuking some common claims from elementary applications of neoclassical economics, such as:
- Higher demand always leads to higher prices.
- Higher real wages increase unemployment or decrease profits.
- Lower savings leads to lower investment or growth.
- Budget deficits lead to inflation and a rise in the interest rate.
As we will see, the key to economic prosperity is cooperation, not austerity and conflict.
1. Post-Keynesian Heterodoxy
1.1: Who are the post-Keynesians?
Heterodox economics is a big tent, encompassing Marxists, Sraffians, neo-Structuralists, Institutionalists, and more. Each of these schools focuses on different areas of the economy, with different core texts and sources of inspiration. Post-Keynesianism has its origins in John Maynard Keynes’ work, especially his book “The General Theory of Employment, Interest and Money”. Crucially, they interpret this work very differently than the neoclassical synthesis Keynesians (like Paul Samuelson and James Tobin) or the new Keynesians (like Gregory Mankiw, Alan Blinder, and Joseph Stiglitz). [For an in-depth post-Keynesian interpretation of “The General Theory”, I highly recommend Alex Williams’ ongoing series here]. Core post-Keynesian ideas also come from contemporaries of Keynes like Joan Robinson and Roy Harrod, as well as neo-Marxist Michał Kalecki, neo-Ricardian Piero Sraffa, and Institutionalists like Thorstein Veblen and John Kenneth Galbraith.
1.2: The characteristics of heterodox economics
To delineate heterodox and orthodox approaches, we should start with presuppositions on how economic questions should be approached. Don’t worry, I’ll explain all the big words shortly.
Epistemology is about how to obtain knowledge. Neoclassicals, following Milton Friedman, generally evaluate if a hypothesis is sound based on if it makes accurate predictions. The realist approach, meanwhile, starts with assumptions that reflect economic reality.
Ontology studies existence and being. Neoclassicals see the economy as being composed of individual, atomized agents. Institutions such as banks or companies are just intermediaries of individuals. But for heterodox economists, institutions are not just the sum of individual’s actions and preferences, and the macroeconomy is not just the sum of microeconomic results. We’ll see later that many economic decisions lead to opposite results if they are adopted by everyone instead of just an individual.
In the neoclassical account, rationality is based on perfect information (where all relevant information is known to perfect accuracy), or imperfect information (“where agents are also able to factor in the time spent searching for the optimal quantity of information”). Instead, heterodox economists think it’s regularly impossible to get crucial information. Agents instead aim for a “satisficing” outcome (compared to maximization in the neoclassical approach). They heavily rely on norms and rules of thumb, and create institutions to reduce uncertainty. Crucially, this is a perfectly sensible response to a world with such uncertainty.
Commonly, the goal of economics is defined as “the efficient allocation of scarce resources”. This reflects the neoclassical approach, where prices are generally a function of scarcity. But heterodox economists, following classical economists from Smith to Marx, start with the existence of a economic surplus, and examine how an economy maintains itself and distributes resources. Post-Keynesians especially emphasize that resources are basically never fully utilized.
Of course, a wide variety of economists with greatly varying politics can be placed in the neoclassical school. But ultimately, they think that without problems like imperfect information and externalities, flexible prices would bring the economy to an optimal state. Meanwhile, heterodox economists think markets face fundamental problems that limit their efficiency and stability. The state must position itself within markets to stabilize and regulate them.
1.3: The essential characteristics of post-Keynesian economics
Now that we’ve broadly compared heterodox and orthodox approaches, one might ask: what distinguishes post-Keynesianism? Two principles are essential, while five more are auxiliary.
We start with the principle of effective demand, which states that “the production of goods adjusts itself to the demand for goods. … The economy is therefore demand-determined, and not constrained by supply”. This notably includes the supply of savings, which is determined by investment rather than vice-versa. Marxists and new Keynesians, for example, may agree with this in the short run. But post-Keynesians think it holds in the long run too.
Next is dynamic historical time. Drawing from Robinson, post-Keynesians focus on how the economy adjusts to changes. Instead of changes being immediate and easily reversible, “Time is irreversible: once a decision is made and implemented, it cannot be reversed, except perhaps at a great cost”. Given this, the economy in the long run is NOT independent of the short run. Changes today have permanent effects in the future.
Now, onto auxiliary features. First is the view that flexible prices aren’t always good. More stable prices can help plan future business and investment, reducing uncertainty. And lowering wages during a recession, for example, would only worsen the situation by decreasing effective demand even more, thereby increasing the relative debt burden of firms.
Second is a focus on monetary production. Contracts are in terms of money, and households own financial assets, rather than directly owning the physical assets of companies. The role of banks is crucial, since only they can create the money needed to start production. How much they loan, and the interest rates they charge, largely depend on growth of the overall economy.
Third, we have the notion of fundamental uncertainty, closely tied to Keynes’ work on probability, as well as economist Frank Knight. Under fundamental uncertainty, “it is impossible to calculate either the probabilities of an event occurring or possible outcomes”. Confidence (or “animal spirits”, as Keynes said) is therefore crucial to production and investment. Some neoclassicals might respond that this would make the effects of any economic policy unknowable. But actually, we can predict real results from this concept: people hesitate to act because they lack information, and they aren’t likely to greatly change their behavior because of new information.
Fourth is a microeconomic approach which we’ll explore in chapter 2. And last is a commitment to pluralism. Since the economy is multi-faceted, different models may apply to different situations.
1.4: The various strands of post-Keynesian theory
A natural result of a realist approach is to draw from several different theories, as different parts of the economy simply act differently. So, post-Keynesian theory “is far from being a homogeneous approach”. But three main strands stand out.
First are the fundamentalists, such as Paul Davidson and Hyman Minsky. Mainly drawing from Keynes, they emphasize fundamental uncertainty, money, and financial instability. Notably, they consider neoclassical economics to just be a special case of their more general approach.
Next are the Sraffians. Through Sraffa, they are inspired by classical economists like Smith, Ricardo, and Marx. Their focus is on relative prices, the choice of production techniques, input-output analysis, and technical questions of value. The latter is important as the Sraffian approach counters both the neoclassical theory of distribution, and the Marxist labor theory of value.
Lastly, we have the Kaleckian approach. They are influenced by Kaldor and the Institutionalists, as well as by Marx through Kalecki. Kaleckians see their approach as more applicable to the real economy, rather than the more general approach of the fundamentalists.
Some post-Keynesians may reject attempting to integrate all three of these approaches. Yet, they share much overlap, such as a focus on effective demand or shared theories of growth. This book, while touching on all approaches, will mainly focus on the Kaleckian approach. It presents a clear alternative to TINA, while being relatively easy to grasp, and fitting for empirical research.
2. Heterodox Microeconomics
Post-Keynesianism disagrees with 2 core elements of neoclassical microeconomic theory. First is marginalism, the idea that price is related to marginal utility (the usefulness of one additional good). Second is decreasing marginal utility: that each additional unit of a good is slightly less useful.
2.1 Consumer choice theory
The origin of post-Keynesian micro theory is in marketing and psychology. Rather than starting from abstract assumptions, it starts from observations about actors in the real economy. Post-Keynesianism makes a distinction between needs (which can be ranked and prioritized) and wants (preferences within a category). For example, you might need a drink more than you need to eat, but want ginger ale more than you want water. Given that basis, we can now discuss 7 principles of post-Keynesian consumer choice theory.
- Procedural rationality means consumers will tend to follow habits and find shortcuts in making consumption decisions. Instead of maximizing utility, as in neoclassical approaches, they instead satisfice.
- Wants are satiable. Beyond a threshold, more of a good won’t bring additional satisfaction. Compare this to neoclassical models, where more of a good always brings somewhat more utility.
- Needs are separable into different categories. Taking into account comparisons of all goods on the market would be hard. Instead, consumers allocate a certain amount of money to each category. What they buy in that category then depends on relatives prices of those goods, rather than the relative prices of all goods. Studies back this up: “the price elasticities for major subgroups are extremely weak (between –0.003 and –0.072 according to one such study) and cross-price elasticities close to zero (in fact, inferior to 0.02, 30 out of 36 times; see Eichner, 1987, ch. 7)”.
- The subordination of needs, which is similar to Maslow’s hierarchy of needs. Income goes to one type of need until it is satisfied, then remaining money goes to the next level, and so on. Changes in relative prices within a level only change consumption of goods in that level.
- As someone’s income grows, their level on the pyramid of needs increases, with the top level being moral and environmental needs.
- Pulling from Galbraith, post-Keynesians believe needs are determined by marketing and imitation of others.
- Consumer’s choices are not independent of order, but depend on past experiences.
It’s important to note an asymmetry of the relationship between price and demand in post-Keynesian consumer choice theory. If a price of an essential good increases, then demand for luxury goods decreases. This is in line with neoclassical economics. But, contra to neoclassicism, if the price of a luxury good increases, the demand for essential goods doesn’t change.
2.2 Oligopolistic markets and objectives of firms
Post-Keynesian firms operate in oligopolistic markets, where a few mega-corporations dominate smaller firms. The former are price setters: they make the prices, they aren’t forced to take them by the market. Smaller firms are price takers, having to follow the prices set by big companies. Prices are not set to clear markets and equalize supply and demand. For simplicity, we’ll assume firm owners and managers have the same goal: survival of the firm in the long run. This isn’t always the heterodox approach: Galbraith, for example, instead emphasizes the conflicts between them.
In basic neoclassical theory, firms maximize profit and face increasing marginal costs. This second assumption is crucial: without it, the profit-maximizing point is no longer where marginal revenue equals marginal costs.
The post-Keynesian approach is different. Surveys show that increasing marginal costs are rare, and that management actually has many objectives. All of these objectives require the firm to survive. Survival, in turn, is aided by economic and political power. The main way firms acquire this power in the marketplace is through a high market share. So, as a simplification, we can say that firms maximize growth rather than profit.
Yet, we know that firms still care about profits. The reason why is simple: growth requires investment, in R&D or just to build new plants and the like. So, firms need to have money left over after paying their costs. Crucially, profits also allow firms to borrow from banks. But they cannot borrow totally freely: as Kalecki’s principle of increasing risk notes, all else being equal, banks will be less willing to lend to a firm who is more indebted relative to retained earnings. Firms can prove their creditworthiness and decrease borrowing costs by increasing retained earnings. In short, “profits are the solution to the financial requirements that constrain the growth-maximizing objective of the firm.”
To examine the relationship between profit and growth explicitly, we need to think of two frontiers. First is the expansion frontier, which shows the maximum growth rate that can be achieved with a given profit rate. In other words, if investment is done as efficiently as possible. At first, the slope of this function is increasing, as new technology is deployed to decrease costs. But at a point, the slope is decreasing, as the firm expands into new and uncertain markets. [Notably, the same basic analysis applies if the slope is instead constant after a certain point, as Kalecki appeared to think.]
Second is the finance frontier, showing the minimum profit rate needed to obtain a given level of growth. Banks lend at a multiple of retained earnings, which we’ll call ρ. Therefore, the maximum amount of investment is given by the equation below:
Where I is investment, P is profit, and i is the interest rate on capital K. Note that both i and ρ depend on bank confidence and the macroeconomy. Next we’ll define two more variables: the profit rate, r = P/K. And the growth rate of capital, g= I/K. By dividing both sides of the equation above by K, and some simple algebra, we obtain the equation for the finance frontier:
The two frontiers are combined below:
Being growth maximizing, firms will want to reach point G, not the profit-maximizing point R. Note that all firms (in the medium run) must exist in the area between, or on, the frontiers. It’s not possible to invest more efficiently than the expansion frontier, and if a firm drops below the finance frontier, it would not be able to make its interest payments, and would default.
2.3: The Shape of Cost Curves
The post-Keynesian approach to costs collows the Leontief production function, in which there is a fixed quantity of labor and capital required to produce a good. These quantities are determined by technology. The neoclassical approach instead uses Cobb-Douglass functions, where labor and capital can be substituted for each other.
A few definitions before we continue. Unit cost (UC) is simply the average cost needed produce 1 unit of a good. Unit direct cost (UDC) is unit cost minus overhead, such as administration. Lastly, marginal cost (MC) is the extra cost needed to produce 1 more unit of a good. Note the distinction between average and marginal costs.
We’ll be using a few stylized facts in this section, depicted in the graph below with output on the x-axis and cost on the y-axis.
Firms are made up of different plants or locations, which are in turn made of up of different segments (like assembly lines). Each of these segments has a practical capacity (also known as the engineer-rated capacity). The level of full capacity for a firm is q_fc, the sum of practical capacity across all segments. Past this point, marginal costs increase because of overtime pay, machines breaking from overuse, and/or workers becoming overworked. q_th is the level of theoretical capacity, which is the maximum level of output possible.
Empirically, firms’ capacity is between 70% and 85% of full capacity. This isn’t some accident of mismanagement: they prefer to have spare capacity. To explain why, recall fundamental uncertainty. Firms just don’t know exactly what demand will look like in the future. Excess capacity, and stored inventory, allow companies to react quickly to an increase in demand without changing prices. Building a new factory, for example, could expand capacity too. But that would take a comparatively long amount of time. In that time, shortages could occur, and the firm could lose customers as they shift to other products.
2.4 Price setting
“All post-Keynesian models rely on cost-plus pricing”, where a costing margin is added on top of some measure of cost to obtain the final price. Prices are administered: determined well in advance of products coming to market. There are 3 main models of administered pricing, each with a different method of determining the costs or markup.
First is the oldest and simplest method, still used by some small and medium-sized businesses, called markup pricing. This was first laid out by Kalecki, and is often used in post-Keynesian macroeconomic models. The equation is below, with θ being the gross costing margin, which covers overhead.
P = (1 + θ)(UDC)
Next is normal or full-cost pricing, from Hall and Hitch’s study of firm’s pricing strategies in 1939. This is a more realistic method, common among large firms for at least 100 years. In the equation below, NUC stands for normal unit cost (the unit cost at a normal level of capacity utilization) and Θ is the net costing margin.
P = (1 + Θ)(NUC)
Lastly we have a method generally preferred by large and medium-sized firms: target-return pricing, as discussed by Lanzillotti in 1958. This is the most complicated, as it requires knowing the value of a firm’s capital. The equation is the same as for normal-cost pricing, but Θ is set to achieve a specific target rate of return on the firm’s capital.
2.5: The determinants of the costing margin
It may be hard to believe that firms freely change their prices when unit costs change. But it’s important to note that these pricing methods are based on costs in normal times. Firms prefer to absorb short term changes in costs by lowering the profit margin while keeping prices constant. This keeps buyers loyal: a surprise change in prices could mess with buyers’ plans.
As noted before, different firms tend to set the same prices. Yet less efficient firms have higher unit costs, and therefore need a lower costing margin to keep prices competitive. This lower margin, in turn, means those firms have less profit, and less money to invest. That means lower growth: other firms in their industry will outpace them. The inefficient firm will shrink relative to competitors, and eventually fail.
It’s reasonable to ask at this point what exactly determines the costing margin. Different traditions offer different explanations.
Which one of these is correct? Well, all of them are probably true to some extent. In fact, we can combine these explanations with the frontiers from earlier. Lower bargaining power of capitalists or lower monopoly power, for example, shifts the expansion frontier downward. We can augment the finance frontier equation like this:
With i_n being the trend interest rate (as managed by the central bank), and g_s being the growth rate of the firm or industry. Thus, these explanations are compatible with each other.
2.6: Consequences for macro theory
The micro discussion above carries important implications for macroeconomics. For example, in post-Keynesian theory, an increase in aggregate demand generally does NOT lead to an increase in prices. Exceptions to this exist for raw materials and agriculture, which act more like the neoclassical model. In those cases, post-Keynesians support buffer stock programs to ensure higher prices don’t cause inflation elsewhere through higher costs. Also, higher demand does lead to higher prices at sustained full capacity: so, avoiding supply bottlenecks is key.
Chapters 3 and 4, discussing the macroeconomic monetary circuit, effective demand, and the labor market, are covered here.