An Interview with Roger Farmer by Phil Armstrong: Part 4 of 4: On Modern Monetary Theory

September 23, 2021 Roger Farmer

An excerpt from Chapter 5 of Can Heterodox Economics Make a Difference, published by Edward Elgar in 2020. Here is a link to the Table Of Contents, and a link to the biographies of all those interviewed for the project.

“If the question is ‘what determines the price level?’, which I do take to be a useful question, I’m probably in bed with modern monetary theorists. My view is that beliefs determine the price level. But the reason that I’m able to make that statement and live in two worlds at the same time, is because the equilibrium models I build contain lots of interacting agents of different types.”

“There are many pluralistic views. I’m not sure why anyone would put a label on themselves, I refuse to be labelled as anything. I don’t know why anyone would want to be labelled as an ‘MMT person’ as opposed to someone who seriously considers the ideas in Modern Monetary Theory and tries to integrate them with other things that they found useful. ”

— Roger E. A. Farmer

PA: You remind me of Charles Goodhart, who also has sympathies and criticisms of various schools. I’d like to turn next to Modern Monetary Theory. You’re a very experienced economist, I’m interested to know what you’ve picked up about MMT? If you’ve not come across MMT, that tells me a lot about its outreach.

RF: It has very little outreach in terms of mainstream economics and my initial reaction to the definition is that it’s confused. [1] My work is heavily influenced by conventional neoclassical economics – that makes me mainstream. And with my mainstream hat on; I insist on constructing coherent models that are based on individual maximising behaviour.

When you start thinking about monetary theory and monetary economics in that context, what you might say immediately is, “Yes, but general equilibrium theory has no role for money”, which is partly true. But to the extent that it forces you to construct logical chains of reasoning, the arguments about endogeneity of money – that’s where I’ve met MMT – I find critiques of mainstream monetary theory to be misplaced. Of course, money is endogenous! But what does that mean?

If the question is ‘what determines the price level?’, which I do take to be a useful question, I’m probably in bed with modern monetary theorists. My view is that beliefs determine the price level. But the reason that I’m able to make that statement and live in two worlds at the same time, is because the equilibrium models I build contain lots of interacting agents of different types.As a consequence, I end up with situations where the model is not determining prices, I need something else. So there’s clearly room for a marriage of that idea with notions of endogenous money from modern monetary theorists, and I suspect that if I actually sat down with Stephanie Kelton, or you, or some other modern monetary theorist, and had a debate about that, we’d find that we had a lot in common. I’m simply more willing to use the apparatus of traditional microeconomic theory to explain those ideas.

PA: Yeah, I get the feeling, having heard you speak before, that you are open to dialogue with what I would call heterodox views. More so than most.

RF: I think that’s essential!

PA: One of the things about MMT- which leads me onto my next point – is its ‘unique selling point’ in the non-academic world: this idea that taxes drive money, the idea that the government must spend money before it can tax. For modern monetary theorists, taxes have two functions: one, to give value to state debt, because without taxes underlying the value of the currency people wouldn’t accept state money in payment for delivering services to the state. And secondly, by adjusting spending and taxes, the government can adjust aggregate demand. The government, in order to fund spending, must acquire money from the private sector and/or borrow it.

RF: I believe that this is related to the debate, in the standard theory, on the Fiscal Theory of the Price Level. Government spending, taxation and borrowing are connected by an accounting identity. In any period of time, the difference between the amount the government spends and the amount it takes in in taxes has to be funded by issuing some kind of debt. It could be money, it could be short bonds, or it could be long bonds. Treasuries spend and create liabilities. Central banks decide on the composition of those liabilities between the different categories.

There is currently a lot of angst in modern macro circles about what determines the price level. And the government accounting identity can be viewed in two ways. In one way it’s a budget constraint, which the government has, just like you and I have budget constraints. We spend, we borrow, we get income, we die, and as a consequence of eventually dying, over our lifetime we can’t spend more than we earn. That places a constraint on how much we can borrow.

Now you and I can shift resources over time by borrowing and lending, so one way of seeing the government accounting identity is that it’s just the analogue of what you and I do. Now, that’s clearly false because the government doesn’t die, and the revenues are coming in from different people from those who benefit from the spending.

One group of economists sees the government accounting identity as analogous to a household constraint. Government spending plans must be consistent with its income plans over a very long horizon. In this view, the government must make spending plans that are consistent with taxation. There is an extreme version of that argument, called Ricardian equivalence, which says that it doesn’t matter whether a given government expenditure plan is financed by debt or taxes. Different plans simply rearrange the timing of the payments, and the representative agent who must pick up the tab, is the other side of that borrowing and lending. Now, most people think that’s probably wrong.

PA: Yeah, certainly MMT-ers!

RF: Indeed. Now, the other way of looking at this accounting identity is that it is a debt valuation equation. The government has a stream of primary surpluses that it will be running for the conceivable future. It has some existing debt in dollar terms. The price level must adjust to make sure that the value of the nominal debt is equal to the discounted present value of the surpluses. So, the sequence of accounting identities is a price level determination equation. I suspect that some of these issues arise in modern monetary theory. Perhaps you could say more about what modern monetary theorists mean by money?[2]

PA: Well, I think that what modern monetary theorists would argue is that money is credit and nothing but credit.

RF: Fine, so a treasury bill is money. Or not?

PA: Well, no – a treasury bill isn’t money, because money is simply a ledger entry. For example, what a modern monetary theorist would conceptualise is that there is no government budget constraint, it’s simply an ex-post description of what’s happened. So, in other words, if the government wants to spend, if you want to conceptualise as a model, the first thing it has to do is spend by issuing new money, i.e. new debt, and that doesn’t have any corporeal existence.

RF: You said new money and new debt in the same breath.

PA: For a modern monetary theorist, credit and debt are the same thing, made at the same time. Therefore, new money is new debt.

RF: Perfect, I’m totally on board. In the models I write down, that’s true because there are multiple equilibria and because the actions the government takes help to select one of the equilibria. Maybe I should write a paper called ‘Modern Monetary Dynamic Stochastic General Equilibrium Theory.’

PA: If the government spent money, what would you think would happen to the balance sheets and the economy, or do you not think in those terms? So if the government bought, say, a new nuclear submarine from a private defence contractor, what would be the monetary movements in that?

RF: The government has a bank account that it uses to purchase new nuclear submarines. I’m honestly bemused by the question.

PA: A modern monetary theorist, would say that if the government buys a new submarine, its balance at the Bank of England goes down. The receiving bank’s reserves would go up, which would mean there’s an internal transfer on the liabilities side of the Bank of England, there’s a reduction in the reserve account of the government, and an increase in the reserves held by the receiving bank. The asset total value of the Bank of England balance sheet remains the same.

Regarding the bank which holds the account of the company that makes the submarine, that bank’s balance sheet will expand. It will have liabilities – say it was £5 billion, and now that company has a liability at its own bank of £5 billion and on its assets side, the receiving bank’s reserves have gone up. That’s how a modern monetary theorist thinks. But you won’t be surprised to know that when I ask economists that question, they don’t answer it in those terms.

RF: How does a modern monetary theorist deal with the difference between the world we live in now and the Gold Standard?

PA: Under the Gold Standard, the government’s ability to issue debt, or to spend, is limited by its gold stocks. Because, for example, if the government runs a deficit under the Gold Standard - say it’s spent a billion pounds and took in £800 million, there would be £200 million’s worth of convertible gold currency.

RF: So, its an entirely different world.

PA: Yes. Under the Gold Standard, the interest rate was determined by the central bank to compete with the option of conversion, which is why big deficit countries under the Gold Standard would raise the bank rate.

RF: So, the question about buying a nuclear submarine would have very different implications in an economy under the Gold Standard than it would under the current standard.

PA: Yes, it would. I don’t know if you’re familiar with the employer of last resort (ELR) policy?

RF: No, I’m not.[3]

PA: Modern monetary theorists don’t believe that the main policy tool for maintenance of price stability should be interest rates, they don’t believe this approach is effective. And an ELR also maintains full employment. Under an ELR policy the state offers an unlimited number of jobs and hence, unemployment effectively falls to zero. If a person were unemployed, the government would provide one. It is, in effect, a stock control policy – in times of recession – people move into the employed labour buffer stock.

RF: I am sympathetic to both points. I agree that the state has a responsibility for maintaining full employment. But full employment is not easily defined. It is possible to have too much employment just as it is possible to have too little employment. A job for everyone who wants one is a meaningless concept for someone who thinks in terms of search.

You ask, “Do you consider the employer of last resort policy to be a viable means to ensure full employment and price stability?” I’m certain it could be a successful means of ensuring full employment, we saw that in Soviet Russia. I don’t think it was very efficient – the state allocating people to jobs is not my ideal society. Also, I don’t think that full employment, implemented by this policy, would guarantee price stability.

PA: The final question is what role do you consider that MMT might have to play in the practise of heterodox, or indeed any economics? Do you see it as something that’s so far from what economists are thinking about that you don’t even see it on the horizon?

RF: There are many pluralistic views. I’m not sure why anyone would put a label on themselves, I refuse to be labelled as anything. I don’t know why anyone would want to be labelled as an ‘MMT person’ as opposed to someone who seriously considers the ideas in Modern Monetary Theory and tries to integrate them with other things that they found useful.

PA: Well, that, in a sense, is just as good an answer as any other. That if you kind of pick apart the question, that’s really what I want to do with the questions, if you see what I mean. Well, that concludes the formal interview. Many thanks, Professor Farmer.

[1] This statement was true when I spoke to Philip in 2018. A lot has changed since then and MMT is now being more widely discussed as a policy approach to government finance.

[2] For a more complete account of my views on the Fiscal Theory of the Price Level see my recent working paper with Pawel Zabczyk (Farmer & Zabczyk, 2020).

[3] I certainly know now as a consequence of the impact of MMT on the 2020 US presidential campaign.

An Interview with Roger Farmer by Phil Armstrong: Part 3 of 4: On Heterodox Economics

September 21, 2021 Roger Farmer

“I wrote a paper with Andreas Beyer of the ECB where we showed that New Keynesian models are unidentified. The implication of our work is that the parameters of the mainstream NK model are identified by theoretical restrictions and not by data. That may work well as a description of past data. But if the model is wrong, it will not help to predict future data nor is it a good laboratory for conducting policy experiments.”

“… as I have consistently argued since 1993 we should be embracing multiple equilibria as a way of integrating key ideas from the General Theory with General Equilibrium theory. Once you realise that economic fundamentals – technologies, preferences and endowments – are not enough to pin down the equilibrium, you open the door for psychology and sociology to come in and to close an economic model with a theory of beliefs.”

— Roger E. A. Farmer

PA: Now I’ll move on to heterodox economics, and I’m looking for your personal opinion and your perception of the environment in which you work. So, what do you personally understand by the term ‘heterodox economics’, and could you give me any examples of groups of economists that you would identify as heterodox?

RF: Well, for anyone in a major university in the UK or the US, ‘heterodox’ is an insult used to put down people who don’t agree with you. For anyone in non-mainstream universities, it’s a badge of honour, a term used to identify the fact that they really are guardians of the truth.

PA: So it’s a very loaded term, then!

RF: I don’t think it’s a very useful term. Let’s take post-Keynesians. I wrote a piece recently – it was called ‘Post Keynesian Dynamic Stochastic General Equilibrium Theory’ (Farmer, 2017) – which was an attempt to unite post-Keynesians and orthodox economists. In that article I pointed out that post-Keynesians kept alive a flame of truth from Keynes’ General Theory that had been lost by the mainstream.

The world moves on, thought moves on, and Keynes would not be a Keynesian today. There were ideas in the General Theory that were kept alive in heterodox circles. The split came in the 1950s with Samuelson, who was a dominant figure on the East Coast of the US, who rejected certain ideas of the General Theory – in particular, a central idea from the General Theory that there are multiple equilibrium unemployment rates. That idea got thrown out, and that’s the point where what we now call New Keynesians and Post Keynesians diverged.

PA: Thank you. I think you’ve accurately covered questions 11, 12 and 13, so I might ask you question 14. Do you consider yourself to be heterodox, firstly, and secondly, do you work in a university department, do you teach undergraduates? What’s your current role?

RF: That’s an interesting question for me, because I always considered myself mainstream but slightly on the edge.

PA: The fact you’re talking to me shows you must be somewhere near the edge, I think!

RF: So, what am I doing? I no longer teach undergraduates, I teach graduate students. I teach two graduate courses at Warwick. I’m also the Research Director at the National Institute of Economic Social Research in London, and as part of that, a bunch of my time is spent running a programme called ‘Rebuilding Macroeconomics’, which is an ESRC funded initiative.

After the financial crisis, the ESRC decided that macroeconomics was broken, but they weren’t sure how to fix it. So, they asked for bids from teams to allocate funds. A group of people – myself and four other people on the management team – won that bid. Whereas initially I thought that I was kind of on the edge, I soon realised otherwise.

The management team consists of me; Angus Armstrong, who’s an economist at NIESR; then we have an anthropologist, Laura Bear, from LSE, a psychologist, David Tuckett, from UCL and a complexity theorist, Doyne Farmer, from Oxford. And once I started interacting with them, I realised that my role was not to map out new territory – it was to ground the project and interactions with mainstream economists. There was a danger the project would move so far from orthodox theory that none of the mainstream economists would listen to what was being said. Having said that, I guess I must be a heterodox economist, because I’m entertaining notions of introducing psychology, sociology, anthropology, and complexity theory into mainstream economics. It’s a fascinating conversation with people from outside the discipline.

PA: That’s great. This one relates to the new consensus macro; however you want to interpret that, and the global financial crisis - do you consider that the GFC gave any evidence to contradict the NCM, the rational expectations, new Keynesian, new classical school of thought?

RF: Yes, I think so.

PA: OK, and how can they get away with it then? How can they remain in this position of ascendency if evidence of the GFC wasn’t what they expected, they were caught by surprise? I think the Queen said, “How come none of you guys saw it coming?” – how do you think they’ve managed to remain on top, all the Nobel Prize winners who’ve still got their status?

RF: Oh goodness, we need to get into the philosophy of science. Max Planck said that ‘science progresses one funeral at a time’. He meant by this that established scientists do not change their minds in the face of contradictory facts. They simply amend their theories. And eventually, established research programmes die out because they fail to attract the best new students.

The New Keynesian paradigm has simply adapted. And to use an idea from the philosopher Imre Lakatos, it’s a degenerative research agenda.[1] It’s amazing – people are taking the models they were using before the Great Recession, and they’ll say, “Well, we’re just missing a piece. We really should have had the financial market in – well, let’s just tack it on!” And now it’s there, they go back, and they say that their model works.

It’s much like Ptolemaic astronomy. When Copernicus came along, Ptolemaic astronomy continued to predict the movement of the planets better than Copernicus, initially, because followers of Copernicus were using circles instead of ellipses in the description of planetary movement. The new Keynesians are doing much the same thing. I wrote a paper with Andreas Beyer of the ECB where we showed that New Keynesian models are unidentified.[2] The implication of our work is that the parameters of the mainstream NK model are identified by theoretical restrictions and not by data. That may work well as a description of past data. But if the model is wrong, it will not help to predict future data nor is it a good laboratory for conducting policy experiments.

PA: How would you describe your relationship with new Keynesianism? You obviously have some aspects in common, like methodological individualism, rational expectations – but from hearing you speak I sense that you’re quite willing to criticise new Keynesianism. If there is such a thing as New Keynesian orthodoxy, you seem to have quite an uneasy relationship with it.

RF: Yes. I use the same methods, but I’m willing to give up on some of the assumptions. The big difference I have with New Keynesians goes back to the split that occurred in 1955 between New Keynesians and Post-Keynesians. In a sense, I am a Post-Keynesian who uses neoclassical methods. But even that description is not quite right because my research has also led me to be critical of Post-Keynesian policy prescriptions. For example, I believe an asset market intervention to support equity prices is a better cure for a big financial contraction than traditional fiscal policies.[3]

Samuelson’s new-classical agenda of combining Keynesian economics with equilibrium theory was a great idea. But he made a huge mistake by assuming that Keynesian economics was about sticky prices. Keynesian economics was never about sticky prices. My own work reconciles Keynes’ General Theory with Walrasian or temporary equilibrium theory in a different way.

There are two principle aspects of the New Keynesian model that I disagree with. Most New Keynesian economists writing before the 2008 financial crisis modelled the labour market as an auction. In an auction model, the labour market is always in equilibrium. The quantity of labour demanded is always equal to the quantity of labour supplied. The real wage and the volume of employment are determined by the intersection of the labour demand and supply curves. That’s not a good assumption for a variety of reasons, not least of which is that it jettisons the concept of involuntary unemployment. Prior to 2008, most New Keynesian economists gave up on unemployment entirely. That was a big mistake.

After the 2008 financial crisis some New Keynesian economists reintroduced unemployment into NK models using search theory. But they are doing it the wrong way. The right way is as ‘Keynesian Search Theory’, a version of search theory originally developed in my own work. This is a way of closing search models that leads to the potential for a continuum of steady state equilibrium unemployment rates. That’s one area where I have a difference with mainstream New Keynesian economists.

My second point of disagreement is with mainstream models of financial markets. Writing in 1972, Lucas threw away all of Keynesian economics and replaced it with Walrasian general equilibrium theory. But he did it in a way that made microeconomic theorists who worked in general equilibrium theory – I have in mind Frank Hahn and Ken Arrow – absolutely horrified. Instead of thinking about an equilibrium as something that the economy was tending towards, for Lucas the market was in equilibrium at all points in time. That was a massive transformation.

Ironically, I think that Lucas’ shift in viewpoint was a useful one. Moving to the ‘Lucasian’ view was the right way to go. But what Lucas knew, but glossed over, is that once you take that route you can no longer pin down equilibrium in terms of economic fundamentals. As I explain in my forthcoming encyclopaedia entry, `The Indeterminacy School in Macroeconomics’ (Farmer, 2020), there are always multiple equilibria in monetary general equilibrium models.

Not only are there multiple dynamic paths, there are also multiple steady states. Lucas was unhappy with me promoting that view argued in my book, The Macroeconomics of Self-Fulfilling Prophecies, and he wrote me a letter at the time saying, ‘why are you doing this?’.[4] For more than forty years, mainstream economists have been trying to purge multiple equilibria from their models. In contrast, as I have consistently argued since 1993 we should be embracing multiple equilibria as a way of integrating key ideas from the General Theory with General Equilibrium theory. Once you realise that economic fundamentals – technologies, preferences and endowments – are not enough to pin down the equilibrium, you open the door for psychology and sociology to come in and to close an economic model with a theory of beliefs. For me, beliefs should be modelled as fundamentals which have the same methodological status as preferences, endowments and technologies. So that, in a nutshell, is where I differ from most mainstream New Keynesian economists.

PA: And if you were to identify yourself with a school of economics, are new Keynesians, with reservations, as close as you would come?

RF: I refuse to be labelled. I’m a ‘Farmerian’!

Stay Tuned for Part 4 on Modern Monetary Theory

[1] (Lakatos & Musgrave, 1970).

[2] (Beyer & Farmer, 2008).

[3] (Farmer, 2010).

[4] For background on this interaction see Cherrier and Saïdi, (2018).

An Interview with Roger Farmer by Phil Armstrong: Part 2 of 4: On Conventional Theories of Money

September 16, 2021 Roger Farmer

Screen Shot 2021-09-12 at 2.12.47 PM.png

“[Then] there is the question of whether central banks should intervene by controlling the yield curve. … [My] answer is yes and, I would go further. Central banks should control the price of a basket of risky assets. Policies of this kind are not currently in the remit of the Bank of England or the Fed. But they should be. I remain optimistic that my work is leading them in that direction.”

— Roger E. A. Farmer

PA: Are you familiar with the credit and state theories of money?

RF: Peripherally. I’ve been following a debate going on right now on Twitter. Some non-mainstream economists, Jo Michell at Bristol is an example, have been critical of textbook theories of money.

You gave us a set of questions we would discuss in this interview. I laughed when I saw question 7 which reads, “Do you consider the quantity of money to be determined exogenously or endogenously?”, to which my answer is “Yes.” Here’s what I mean by that.

I’m a general equilibrium theorist, in the sense that I want to conceive of the things I observe in the macroeconomy – this comes back to methodological individualism – as choices made by individuals in markets. To answer the question of exogeneity, you have to ask, first of all, what money is. That’s a hard question and there are many different ways of thinking about it. If you go back to worlds where there were commodity monies – gold and silver – it’s quite clear that the quantity of money was exogenously determined. At a point in time, the quantity of precious metals is fixed. There have always been trading mechanisms, trade credit for example, whereby other objects would serve as a medium of exchange. In today’s world where gold and silver have been replaced by fiat money provided by national governments, the distinction between exogenous and endogenous money depends very much on the policies that central banks follow.

I’ve never understood the criticism of traditional monetarist thought by advocates of modern monetary theory. Orthodox monetary theorists, and here I would include most of the New Keynesians, would agree with the statement that money is endogenous in modern fiat money systems when the central bank pegs the interest rate on short-term debt. But there is a sense in which asking if money is endogenous is the wrong question. The right question is: what determines the absolute price level? Here there is an active debate centred around the question of uniqueness, or non-uniqueness, of equilibrium in macroeconomic models. In all models where the central bank sets the interest rate, as opposed to the quantity of bank reserves, the quantity of money is endogenous.

Let me return to the question; what is money? A bank is an institution that allows many more objects to be used in exchange than in economies that don’t have banks. In that sense, private banks create money.

If you hold shares in Apple and you go into a car show room and you try to buy a new Honda by offering them shares in Apple, you’re not going to be very successful because the value of the shares in Apple fluctuates on an hourly basis and nobody really knows quite what they’re worth. Now, imagine an institution – a bank – that lends to firms and households. The bank holds liens on machines and factories to back its loans to firms and it holds liens on houses to back its loans to households. The banks acts as a guarantor of the value of these assets.

Take the example of a mortgage. You go to the bank and you borrow money for a mortgage on your house. The bank manager (at least in the past) knew who you were and had a relationship with you and realised that you were capable of repaying the mortgage. The bank manager’s guarantee turns your house into a negotiable asset that appears on the liabilities side of the bank’s balance sheet and the account that is created by lending you money can then be used in exchange. In that sense, banks create money.

Bank lending is profitable because the interest rate earned on long-term loans is higher than the rate that banks must pay on their short-term liabilities. The quantity of money that banks create is limited by the need to hold cash and other liquid assets to meet the needs of their customers. The amount of available liquid assets is limited by central bank actions that control the price of credit by buying and selling assets in the markets for short-term funds. Before the 2008 financial crisis, cash and bank reserves did not pay interest, but short-term government bonds did. In that period, there was a positive opportunity cost of holding money.

Since the financial crisis of 2008, the interest rate on short-term liabilities has been very low and, in some cases zero. And central banks began to pay interest on reserves. In that environment reserves, which are part of the monetary base, become perfect substitutes for short-term government debt and theories of endogenous money come into their own. When the opportunity cost of holding money falls to zero, monetary general equilibrium models have even less to say about the determination of the price level than they do when money is scarce.

When the interest rate is zero, money and bonds become perfect substitutes and there is a continuum of price levels and a continuum of associated values for the stock of money, all of which are equilibrium values in a monetary general equilibrium model. In those environments, the supply of money is determined by the public’s perceptions of future prices. Beliefs become fundamental in the way I described in my books, The Macroeconomics of Self-Fulfilling Prophecies (Farmer, 1993), Expectations Employment and Prices (Farmer, 2010) and Prosperity for All (Farmer, 2016).[1] The same phenomenon occurs when the central bank pays interest on reserves. Although cash is still costly to hold, cash is a shrinking component of the monetary base. When the interest rate on reserves is roughly equal to the interest rate on Treasury bills, the economy is in a liquidity trap similar to the one that characterized much of the period in the 1930s in the United States. The payment of interest on reserves puts us in a whole new world.

PA: I think your interpretation [of monetary theory] is thought-provoking…. I might go slightly away from the original script to a question that’s very interesting from my perspective: when you talk about equilibrium, do you see it as an ‘ultimate end’ of a tendency, like a point that a force that will take you towards but may not actually reach? Or is it a set point that things adjust to and what would be the adjustment speed of the pathway?

RF: To answer that question, let me use the concept of temporary equilibrium theory.

PA: …I think it’s called a ‘traverse’, the idea of how quickly you get there, what happens on the way.

RF: Indeed. In the temporary equilibrium story, we all meet in a market every Saturday. When we meet, there must be a mechanism to determine what gets traded. That could be Walrasian market clearing, it could be sticky price equilibrium, or it could be some version of search theory, for example, ‘Keynesian Search Theory’, a term I coined in my book Prosperity for All. On Saturday we observe the trades that take place, we observe the prices that take place, and we record those trades as a list of numbers on an Excel spreadsheet. Now, if nothing in the world ever changed, our models predict that the list of numbers we record will converge to a constant list. I would call that a steady-state equilibrium. A physicist would simply call it an equilibrium.

The reason that these uses differ is that the word ‘equilibrium’ in economics is used to mean a Nash equilibrium, which is a set of plans and prices that are consistent with each other in a sense that was made precise for temporary equilibrium models by Roy Radner (Radner, 1972). In a Nash equilibrium, the values of the list of numbers in the Excel spreadsheet may be changing from one period to the next. What qualifies them as ‘equilibrium numbers’ is that at each point in time the numbers record prices at which the quantity of each good demanded is equal to the quantity of each good supplied.

PA: You’ve really answered the question about banks, so how would you consider interest rates to be determined in theory and practise? Do you think there’s a disconnect between theory and practise?

RF: Well, there’s at least one interest rate that’s set by central banks, which is the overnight rate. Then there’s a whole spectrum of interest rates that are determined in markets. Traditional monetary theory argued that central banks could control only one interest rate but, in the 2008 financial crisis, central banks appeared to exert influence over a whole range of rates. So, in that sense yes, there was a disconnect between traditional monetary theory and practice.

PA: Would you consider that the central bank could control the spectrum of rates? I’m not saying it should, but if it wanted to, would you think it a feasible thing?

RF: OK, those are two separate questions. I have never subscribed to traditional monetary theories that are set in the context of a single representative agent and in all of my work, central banks can control a whole spectrum of interest rates at different points in the yield curve.

Second; there is the question of whether central banks should intervene by controlling the yield curve. Again, my answer is yes and, I would go further. Central banks should control the price of a basket of risky assets. Policies of this kind are not currently in the remit of the Bank of England or the Fed. But they should be. I remain optimistic that my work is leading them in that direction.

[1] See my entry in the Oxford Research Encyclopedia of Economics and Finance on ‘The Indeterminacy School in Macroeconomics’ (Farmer, 2020). The Indeterminacy School is an approach to macroeconomics that was initiated at the University of Pennsylvania and at CEPREMAP in Paris in the 1980s. It has continued to evolve over the last forty years. The identifying feature of the Indeterminacy School is the willingness to embrace models with multiple equilibria to explain real world phenomena.

An Interview with Roger Farmer by Phil Armstrong: Part 1 of 4: On Economic Methodology

September 12, 2021 Roger Farmer

I claimed to be a methodological individualist – and I think that’s a useful way of thinking about the world on Saturdays – but on Wednesdays something changes. Let me explain that idea. [Roger E. A. Farmer]

In 2017 I gave a talk at the University of Greenwich where I met Phil Armstrong. Phil was finishing up a PhD in economics at the University of Southampton Solent and he asked me if I would agree to be interviewed for his dissertation on heterodox economists and to recruit some other mainstream economists to be interviewed for the book. Having recently written a piece that aimed to reunite Post-Keynesians with New-Keynesians, Phil figured that that I would be a good person for that role.

Phil and I met up again at a pub in York in 2018 where I was giving a talk at the York Festival of Ideas. We had a very pleasant and substantive interview that Phil’s daughter transcribed and that appeared in Phil’s book Can Heterodox Economics Make a Difference, published by Edward Elgar in 2020. I encourage everyone to read the book. If you have access to a university library with a subscription to Elgar Online, the complete book is available to read electronically. Even if you do not, the chapter featuring myself, and the chapter featuring Tim Congden, are open access and are free for anyone to download. Here is a link to the Table Of Contents, and a link to the biographies of all the economists interviewed for the project.

The interview is quite long but I have broken it down into four manageable parts for readers of my blog. I have made a few edits to each of the posts to make them more concise but the original interview in its entirety is available here. I plan to put out a different part each week. Here is the first part.

Part 1: On Economic Methodology

PA: Thank you for doing the interview. Which field of economics do you consider to be your specialism?

RF: Macroeconomics

PA: And within that specialism, what would you consider to be the main issues relating to your field?

RF: Oh, there are many. Much of my career has been spent on introducing beliefs as an independent driver of business cycles. I’m particularly interested in business cycles, business fluctuations, inflation, interest rates and unemployment – how those things are all related to each other – and how the fact that beliefs matter should shape our approach to economic policy.

PA: Quite a wide range there! How would you describe the underlying axioms or principles of your view?

RF: How long do you have?

PA: As long as it takes, within reason…

RF: Oh, goodness. I think that, to a large extent, I’d say methodological individualism, with lots of caveats. So, let me tell you how I think about that, and this is going to be a discursive discussion. If you go back to ‘what is methodological individualism?’ I take it to be the notion that a human being pops into the world at the age of eighteen with a preference ordering that makes him or her capable of making choices over every conceivable option that they might ever face in their life. That view of a human being arrived into economics sometime in the 19th century when Walras started to formulate the notion that markets work well.

There were other schools in England – but the Lausanne School in Switzerland – specifically Walras, developed general equilibrium theory, followed by Pareto who introduced the notion of what it means for an outcome to be good in some sense. What Walras and Pareto were doing at that time was formalising Adam Smith’s notion of the invisible hand. In order to do that, they needed a concept of choice – if markets work well – what does that mean? What it means to Pareto is the very narrow sense that market economies are not wasting things. Now in order to develop that idea, you have to have – first of all – a concept of what people want. That’s where the notion of a human being as a preference ordering enters economics.

There’s another notion of a human being that predates Walras and can be found in the work of Adam Smith. Smith wrote ‘The Wealth of Nations’, but he also wrote ‘The Theory of Moral Sentiments’, which contains a view of human nature that is the furthest thing from the notion of a selfish human being that you can get. The selfish human being is very useful for thinking about economic liberty, and in particular, the efficiency of markets. However, it’s not at all useful for thinking about another component of the role of individuals in society, and that’s political liberty. The development of general equilibrium theory by the Lausanne School is the point where economics breaks off from the other social sciences.

I have a view, now somewhat unfashionable, that John Stuart Mill’s essay ‘on liberty’ should be on the reading list of every high school and university. In Mill’s essay, it’s impossible to think of a human being in the same way that a person is pictured in Walras and Pareto because Mill’s conception of liberty is that you and I can change our minds. When we have a conversation – and we’re having one now – and you tell me, “Roger, I’m a modern monetary theorist for these reasons”, and I say to you, “Philip, that’s nonsense, because…”, and I give you an argument ... and perhaps the argument changes your mind; or maybe not. Maybe you respond with an opinion that causes me to change my mind and, as a consequence of that conversation, I go out into the world and do something differently. I make a choice that was not the same one I would have made before we had the conversation.

If you come from the view of homoeconomicus developed by Walras and Pareto, the only way that you can conceive of what just passed in our conversation is that one of us acquired information that was not previously in our information set; information that helped us to make the choices that we would otherwise not have made. Now that, I think, is not a very useful way of thinking about conversations. Conversations, that is, human interactions, change our preference orderings.

I claimed to be a methodological individualist – and I think that’s a useful way of thinking about the world on Saturdays – but on Wednesdays something changes. Let me explain that idea.

The framework I use to think about macroeconomics goes back to John Hicks. Hicks wrote ‘Value and Capital’ where he develops a concept that today we call temporary equilibrium theory.[1] People meet on Saturdays. They bring some goods to market. They make choices, they have beliefs about what they think is going to unfold in the future and there’s some mechanism by which they trade. In temporary equilibrium theory that mechanism is the Walrasian concept of market clearing. People go away and they come back next Saturday, and they trade again.

Almost all macroeconomics can be thought of in that framework, whether you’re a heterodox economist, a classical economist, or any other flavour of economist. Different schools of macroeconomic thought have their own views about how people trade on Saturdays. If you’re an economist, all you need to develop a theory of the macroeconomy is a conception of time and a model of how economic transactions occur. Other social scientists don’t see the world in the same way.

A sociologist or a political scientist is not going to be comfortable with Hick’s notion of temporary equilibrium theory as a complete description of economic interaction in markets because they have a different view of the nature of human beings. The economist’s view of human beings and the sociologist’s view are perfectly consistent with each other. They describe different aspects of human interaction in social structures.

The way to integrate the economist’s and the sociologist’s concepts of human beings is to think about sequences of interlaced interactions. On Saturday we all go to market and trade with each other; given a fixed set of preferences. Then, on Wednesdays, we read a newspaper, or we read a book, or we have conversations. You open your iPad and you go on the internet – the social interaction that ensues changes the preference ordering that you take to the market on the following Saturday. Now, economists have been very opposed to thinking about changing preferences, simply because once you take on that point of view you lose the ability to make value-free judgements about market allocations. The preferences you have might be changing. Economists want to believe that preferences are fixed because they want to make value-free statements about what is – and what is not – a good social outcome.

PA: Well, some answers will be longer than others, that’s absolutely fine. In your work, to what extent would you say that history has a role to play?

RF: History is important – but so is the history of thought, so is mathematics and so is statistics. I’ve been asked many times if economics is a science, and my answer to that question is that it’s definitely a science, but it’s not an experimental science. Macroeconomics, in particular, is a little bit like sitting in the late nineteenth century, assembling a group of chemists, giving them an unknown substance and saying, ‘What is that? Find out what it is. But you can only conduct three experiments a century, and you can’t read the research notes of your predecessors.’ The history of thought is analogous to the research notes of our predecessors. The three experiments I refer to are big natural events like the Great Depression or the stagflation of the 1970s, and I would count the Great Recession we’ve just been through as the third big experiment in the last century.[2] These big natural experiments shake up the way we think.

History is important because it represents our data. It is to a macroeconomist what a record of the heavens is to an astronomer. Neither astronomers nor macroeconomists can conduct controlled experiments. Nature conducts those experiments for us.

Mathematics is also an important skill because it ensures that our thinking is logically consistent. I think it was Marshall who wrote a letter to Bowley, he of the Edgeworth-Bowley Box; do you know that letter?

PA: I don’t know the letter – I know about the Edgeworth Bowley Box, but not the letter.

RF: Marshall’s letter is a statement to the young Bowley about the uses of mathematics, and roughly speaking, he says, “Figure out a good problem, a good idea, and formalise it with mathematics.” And then – I may be getting the order wrong here – “… translate it back into English and throw away the mathematics. Next look for some examples, and if you can’t find any good examples, throw away the English.”

As a profession we’re not very good at following Marshall’s advice. I use a lot of mathematics in my own work. As far as possible I try to bury it in the appendix, but there are just some parts where you can’t do without mathematics. There’s good economic writing that uses mathematics, and there’s bad economic writing that uses mathematics. An example of good economic writing is anything Ken Arrow ever wrote; the mathematics is there, but it’s explained in words, it’s not superfluous, and it’s key to everything he writes. For bad uses of mathematics in economic writing, read 90% of any modern economics journal.

PA: Didn’t Marshall put most of his maths in the footnotes?

RF: Yes, I think that’s right.

Stay tuned for Part 2 on Monetary Theory

[1] (Hicks, 1939).

[2] I am editing this transcript in March of 2020 and I would add a fourth natural experiment; coronavirus. The effects of that event are playing out as I write, and it is certain to be transformative for macroeconomics in ways that will help us sort between alternative theories.

Rebuilding Macroeconomic Theory

January 29, 2021 Roger Farmer

Image is from Chapter 22 of Dynamic Macroeconomics by George Alogoskoufis.

Martin Sandbu has written a nice piece in the FT about the latest issue of the Oxford Review of Economic Policy, which is dedicated to an idea I have been promoting for more than three decades. Multiple Equilibria matter. For David Vines and Samuel Wills, editors of the volume, this is a new revelation. For some of us labouring on the coal front, it has been obvious for a long long time. And Martin is correct to point out that the ship of research is finally beginning to change course, albeit thirty years after the initial ideas were first floated. Recent books that feature the Indeterminacy School are Dynamic Macroeconomics from George Alogoskoufis and A History of Macroeconomics by Michel De Vroey. Beatrice Cherrier and Aurélian Saïdi wrote a very nice survey piece linked here) and I recently published an encyclopaedia article on the history of The Indeterminacy School in Macroeconomics.

For a glimpse at the truly revolutionary implications of adopting the multiple equilibria approach see my piece “The Importance of Beliefs in Shaping Macroeconomic Outcomes” in the Vines-Wills volume (ungated preprint here). The Indeterminacy School in Macroeconomics is indeed revolutionary. And while I welcome David and Sam to the party, we don't, IMHO, need to rename the revolution after David’s mentor, James Meade.

The Indeterminacy School, as I have argued for sometime, has the potential to do for Keynesian economics what the RBC school did for classical economics. And don’t be fooled by the claim of ‘New’ Keynesian economists that economics has already ‘been there’ and ‘done that’. NK economics is neither new, nor Keynesian. It is a beautiful recreation of the verbal theory of business cycles that Pigou wrote about in 1923.

What’s missing in almost all NK economics, with some rare exceptions, is the possibility of being permanently stuck in a high unemployment equilibrium. Why has the mainstream ignored the existence of multiple equilibria for more than three decades? The answer, is that it was deeply subversive to the rational expectations agenda. When I published this piece in 1991 , Bob Lucas sent me a personal letter (quoted below in Aurelian and Cherrier’s) survey pushing back against the indeterminacy agenda.

“I don't see any reason to imagine that the fact that equilibrium is indeterminate on a 'big' set of parameters values in a certain class of theoretical models suggests that equilibrium is likely to be indeterminate in a 'lot' of real-world markets. Do you think that God is drawing pieces of reality at random from some probability space on a Kehoe-Levine parameter space?”

There are great possibilities for pushing these ideas in new directions and for uniting them with Post Keynesian economics. If you are a grad student interested in pursuing a thesis that explores multiple equilibria, you can get some ideas from my second year lectures in the University of Warwick MRes programme linked here.

Roger Farmer's Economic Window

Part 1: On Economic Methodology