Recent Posts

Wednesday, August 1, 2018

Heterodox/Mainstream Views On Banking

David Andolfatto recently published "Reconciling Orthodox and Heterodox Views on
Money and Banking," which discusses the theoretical split between recent mainstream thinking and heterodox views on banking. From my perspective, he is addressing heterodox critiques of mainstream thinking that I am not particularly interested in; in fact, based on his criteria, I would be closer to "mainstream" than "heterodox" -- as would possibly be Hyman Minsky. I would view the problem with mainstream modelling as resulting from a blind spot regarding the business sector.

(Update: David Andolfatto reasonably asked: what is his model is missing about the business sector? My answer here is only sketched qualitatively, and based on a conjecture; I believe I would need to rebuild a similar model to prove my point. I hope to do that in the future...)

"Heterodox" versus "Mainstream"

In the introduction, Andolfatto summarises the "mainstream" view as follows:
Let me review briefly, to the best of my ability, the key elements of two broad schools of thought on the matter. The mainstream view is that modeling the operational details of banking -- the financing of assets with deposit liabilities -- can be safely ignored for questions related to the business cycle. According to this view, the degree of liquidity associated with liabilities issued by the private sector lies on a continuum, so it makes little sense to draw a sharp distinction between which of these liabilities constitute "money" and which do not (Tobin, 1963). 
Conversely, the heterodox view is summarised as:
The heterodox view is that banks are critically different from other financial market participants. In particular, while non-bank agencies wanting to investments first need to acquire the requisite funding, this is not the case for banks. In contrast to the non-bank sector, banks can create the money they lend. Consolidating the banking sector with other actors in a market for loanable funds stems from the misguided notion that banks lend out reserves that they must first acquire. Banks do not in fact lend reserves -- they lend their deposit liabilities (which are incidentally made redeemable for cash). While banks make use of reserves for settlement purposes, reserves do not constrain bank lending (Fullwiler, 2012).
As long as we do not go into theoretical paroxysms about the term "loanable funds," the theoretical propositions labelled as "mainstream" seem more plausible to me. All financial intermediaries emit liabilities that can be treated as "cash" by investors, and so the balance sheet operations end up looking similar to how the formal banking system operates. I did not arrive at this view by being indoctrinated in mainstream economics courses -- I got it from reading Hyman Minsky, who is normally viewed as "heterodox."

The problem here is one that is familiar to anyone who has interacted with economists: they tend to not agree with each other as soon as you get more than one in a room. In this case, banking seems to be a popular area for discontent.
  • People who are mad about banks for political reasons.
  • Heterodox modellers mad about "mainstream" economics for a variety of reasons, and who take heterodox models with explicit banking sector models somewhat literally.
This article is aiming at heterodox critiques that I do not agree with. For example, his policy take-away is pretty much my view:

But even if one accepts the heterodox proposition, it's not clear what the implications are for monetary policy. If financial stability is a concern, then a monetary authority must monitor economy-wide leverage, and not just the leverage emanating from the banking sector.

Proposed Model and Its Defects

He proposes a relatively simple overlapping generations DSGE model to frame the discussion. It may be that I will dig into this model if I want to include it in my upcoming book on business cycle analysis. For now, I will just accept his description of the properties of the model.

My concern is that the model is based on households, with "workers" and "investors" making investment/work/consumption optimisation decisions based on a two period life span. Even if we put aside my distaste for overlapping generations models (how long a calendar time does one period represent?), it misses the business sector.

From a mainstream perspective, the omission of the business sector can be excused on a number of theoretical grounds. The problem is that heterodox economists dispute every single one of those grounds.This makes accepting that the model captures the theoretical dispute a non-starter.

If we want to express this in terms that would make mainstream economists happy, firms should be making investment decisions based on an infinite horizon. Furthermore, there is no "utility function" associated with this maximisation decision, since most large firms are corporations, and the managers do not know the utility functions of their shareholders. (Firms with a dominant owner would be an exception, but are hardly the norm.)

The next problem is that firms have no choice but to maximise nominal profits from a heterodox perspective. In simplified mainstream models where a single representative firm produces all goods in the economy, its pricing decisions directly set the price level. In heterodox theory, pricing decisions are complicated (often based on heuristics), and so we can only have a heuristic for determining future prices. Since a single firm's investment/production decisions are obviously not a factor in those heuristics, price forecasts are essentially fixed for a firm, and so maximising real profits is equivalent to maximising nominal profits.

This then runs into the feedback loop created by the Kalecki Profit Equation (primer) -- increasing investment increases profits. The profit increase is income for the business sector, which can be used to finance the investment. The heterodox argument is that this circular flow dominates the factors that Andolfatto emphasises -- the real policy rate, consumption preferences of households. Although there are real constraints on production created by a fixed stock of existing capital and a limited number of work hours, nominal profits do not face obvious constraints.

(Update: This infinite loop property is why I think we run into problems. If firms wanted to optimise profits, they should invest arbitrarily large amounts in nominal terms, and/or give arbitrarily large dividends, assuming that firm owners also want to support the profit maximisation. Under mainstream assumptions about price level setting, that would break the price level, but it might work under post-Keynesian assumptions about price-setting behaviour.)

The business cycle is driven by the animal spirits of fixed investment. The self-financing nature of investment explains why heterodox economists like myself poo-poo the notion of "loanable funds" in the context of the private sector.

In summary, although it is good to see this debate happening, the tricky bit is that the heterodox positions are a moving target from the perspective of the mainstream. As soon as one model is proposed to deal with one critique, someone else (like me) pops up with another critique...

(c) Brian Romanchuk 2018


  1. Brian, thank you for your comments. Here are my responses.

    First, I am happy we agree on policy implications -- to me, this is the main thing.

    As for your critique about the absence of a business sector, you do not mention why it is important (after all, we end up agreeing on policy implications).

    In fact, one could simply interpret my "investors" as a firm that contracts for labor, paying them in money (they acquired in a loan) or directly with claim against store inventory (the Canadian Tire coupon I mentioned).

    I do have an "animal spirits" parameter that generates business cycles. There is a also a parameter that makes banks "special."

    As for period length, I could easily have extended the analysis to N period lived agents. I kept N = 2 for simplicity. But if you want, even in this case you can think of infinitely-lived agents with short planning horizons (a "behavioral" model).

    Thanks again for your comments.

    1. Thanks. I updated the text to add a bit more context. My fear is that optimising profits in nominal terms leads to arbitrarily large nominal investment and dividends. Since that is rarely seen in the real world, we are stuck with firms following heuristics instead (which I think characterises the post-Keynesian view).

  2. Interesting, and thank you for drawing my attention to Andolfatto's paper. Like you, my views on this would be close to what he describes as the mainstream view.

    Just picking a couple of aspects that Andolfatto seems to consider mainstream: 1) that there is a spectrum of liquidity, rather than a clear distinction between money and non-money; and 2) that it is the lending itself that matters for demand, rather than its impact on the money supply. These seem to me to be fundamental aspects of the critique of monetarism by Kaldor and other post-Keynesians. I find it interesting that this approach is now considered mainstream.

    1. Yes, I think it’s hard to make generalisations about the mainstream once you get past certain theoretical bugbears (equilibrium being the main one).

  3. “But even if one accepts the heterodox proposition, it's not clear what the implications are for monetary policy.”

    The fact that there are few “implications for monetary policy” is not a weakness on the heterodox position because the heterodox crowd do not claim there ARE BIG implications for monetary policy (depending on what you mean by “monetary policy”). For example the submission to the UK’s Vickers Commission by Positive Money and co-authors (link just below), which advocates the heterodox view, does consider monetary policy: it argues that interest rate adjustments are not a good way of adjusting demand. But that idea does not stem directly from the authors’ heterodox views on banking. The link is here:

    And for another example, Warren Mosler advocates a permanent zero interest rate, i.e. he opposes interest rate adjustments, but he does not advocate the heterodox view on banking. I do not view that as a self-contradiction by Warren (though as it happens I support the heterodox view on banking).

    Having said that, I’d argue that actually the heterodox view does have implications for monetary policy, and as follows. As Douglas Diamond and Raghuram Rajan say in the abstract of a paper (link below), the ability of private banks to borrow short and lend long does increase liquidity (i.e. the money supply), but that comes at a heavy price, namely increased bank fragility, i.e. an increased danger of bank runs. However, as the heterodox lot point out, government and central bank can easily supply the country with whatever amount of money is needed to do the really important job: bring about full employment, and at no real cost. So what does private bank leverage achieve? Absolutely nothing!

    Incidentally I used the phrase “monetary policy” there to refer to the choice between having base money as the nation’s basic form of money and instead, using private bank issued money, which may not be correct use of the phrase “monetary policy”.

    Next, I don’t agree with David’s claim that there is a continuum between liability issuing banks and non-bank firms which also issue liabilities, and that hence it does not make sense to distinguish between the two. Our world is full in grey areas and contionuums, but that does not mean it doesn’t make sense to have specific rules for one end of a continuum. E.g. there is no sharp dividing line between very serious crimes like murder and minor offences. That is not an argument against severe punishment for murder.

    In the case of money, every central bank measures its country’s money supply. Short term liabilities of banks are counted as money, while other liabilities are not. But there are no sharp dividing lines between money and non-money.

  4. Brian, is what David Andolfatto describes as 'mainstream' views of banking actually mainstream now? I mean have they evolved recently or something? I recall the Krugman/Keene thing and don't remember Krugman's mainstream loanable funds type argument as really being all that similar to what David A. (or you maybe) seem to call mainstream now. Well maybe our economists is learning?

    1. Hello, just saw this. No longer getting notifications about comments apparently.

      What he describes as “mainstream” is definitely not monetarist. Monetarism infected mainstream views, and a lot of the heterodox/mainstream wrangling was an aversion to monetarist thinking. The question is whether everyone agrees with his assessment of what the mainstream is. There’s probably some pseudo-monetarists running around...

  5. «All financial intermediaries emit liabilities that can be treated as "cash" by investors, and so the balance sheet operations end up looking similar to how the formal banking system operates. I did not arrive at this view by being indoctrinated in mainstream economics courses -- I got it from reading Hyman Minsky, who is normally viewed as "heterodox."»

    Then you missed out a really important "detail" from H Minsky, that:

    “an economy has a number of different types of money: everyone can create money; the problem is to get it accepted.”

    And all those "moneys" have different degrees of acceptance. It is much harder to get accepted an IOU drawn on J Smith than on Citibank, surprisingly. Why? Well... H Minsky also wrote:

    “No matter how exalted a bank may have been, we all know that if assets were marked to market, the net worth of many of the giants of international banking would disappear. Nevertheless these banks are able to sell their liabilities in financial markets, because the buyers believe that they will be protected against losses by the central bank.”

    1. I don’t see those quotes as being germane to my point; the question is whether non-banks can expand balance sheets in the same manner as formal banks. Minsky described the process by how non-banks replace banks in order to bypass regulation.

  6. Banking has a mechanical problem that neither heterodox nor mainstream properly handle. Money mechanically moves between regions (an individual is a micro-region).

    Mechanical movement facilitates problems like bank runs, the Greek debt problem, and personal loan failure. Mechanical movement occurs when money from one region is loaned into a region that is unable to command the ability to recover money spent out-of-region.

    Regional inability to recover loan money (spent out of region) is invisible to the macro economy unless regional money supply and net worth data is available. A developing problem would show up as one region repeatedly increasing money supply and net worth contrasted with a second region repeatedly losing money supply and net worth.

    I think your business cycle would develop when businesses realized that they cannot continually mine monetary wealth from the declining region.


Note: Posts may be moderated, and there may be a considerable delay before they appear.

Although I welcome people who disagree with me, please be civil.

Please note that my spam comment filter appears to dislike long "anonymous" posts. I get no warning about this, and only go through my "spambox" infrequently. The best bet it to keep comments short, and if you think the spam filter struck, let me know with a short comment.