Krugman and Tobin on Banking

Paul Krugman cites a 1963 paper by James Tobin in the process of clarifying his own view of how banks should be considered as part of the broader financial system:

“All the points I’ve been trying to make about the non-specialness of banks are there. In particular, the discussion on pp. 412-413 of why the mechanics of lending don’t matter — yes, commercial banks, unlike other financial intermediaries, can make a loan simply by crediting the borrower with new deposits, but there’s no guarantee that the funds stay there — refutes, in one fell swoop, a lot of the nonsense one hears about how said mechanics of bank lending change everything about the role banks play in the economy. Banks are just another kind of financial intermediary, and the size of the banking sector — and hence the quantity of outside money — is determined by the same kinds of considerations that determine the size of, say, the mutual fund industry.”

The object of the professor’s displeasure appears to be a presumed blogosphere meme that allegedly uses the mechanics of bank lending as the rationale for characterizing a unique role for banking within the financial industry and the economy. And while it is possible that some of those he has previously characterized as banking “mystics” are conflating lending mechanics with some fundamental role for banking, others may be making separate but connected points about these two dimensions of banking quite legitimately. So we should unpack these things and then examine the effectiveness with which Professor Krugman’s channeling of Tobin (Krugman/Tobin for short) refutes the importance of either of these dimensions, whether considered separately or jointly.

First, consider the aspect of bank lending mechanics. This no doubt refers to the general statement that “loans create deposits”. So why do people make this simple point? In most cases, it is made in conjunction with a rejection of the old money multiplier idea that the central bank must supply reserves before a commercial bank can make loans. That said, there is indeed an ambiguity in that association, because the money multiplier even as an incorrectly described process itself features “loans create deposits” or at least can be interpreted that way. Accordingly, the full statement in this context should really translate to “loans create deposits without the necessity of prior reserves”. That is the point being made. And why is that point important? Because understanding how banks work is better than misrepresenting how they work, as has been done in economics textbooks over the last several generations by referring to this particular issue using the false money multiplier model. Indeed, such a view is presented in the Robert Hall paper assessed by Mathew Kline, as referenced in the Krugman post. Moreover, understanding this point is essential in even beginning to interpret the effect of the Federal Reserve’s quantitative easing process in a way that is logically correct. This point has been covered enough elsewhere for purposes here.

The Tobin paper itself refutes (in at least one part) the false causality of the textbook money multiplier, which is good to see for a paper written in 1963. So Krugman can’t be refuting anything so far, using Tobin.

Second, consider the claim that banking is unique. I can think of perhaps one blogosphere case where this issue has been tightly bound with “loans create deposits”, and that case is largely a matter of presentation style. But consider the aspect of banking uniqueness more generally. There is definitely a role for banking that is unique in finance and economics. A good description of it permeates the book Monetary Economics by Wynne Godley and Marc Lavoie. This has to do with the critical role of revolving bank loans used in the financing of business inventories. Indeed, this aspect is expressed in a formal way through the interpretation and adaptation of circuit theory within a post Keynesian context. This aspect is essential to understanding the critical role of banks in economic expansion. It is true in advanced G&L models, and it is true in the real world. In general, workers must be paid before they can buy the output of the economy. The rudiments of circuit theory help explain how that happens through bank lending and deposit creation. Economic expansion requires banks. Krugman/Tobin doesn’t touch on this essential process. If it did, it would negate its own conclusion.

Now, consider how Krugman/Tobin attempts to refute the importance of these aspects of banking – whether conflated or separate.

The main point made by Krugman in his post and in the Tobin paper to which he refers is a strangely benign one. The general point made in the Tobin paper is that banking expands in the same way that other financial intermediaries do – according to economic opportunities. But who is denying this in claiming the importance of either aspect of banking above? It is surely an obvious point to those who have unpacked the two ideas that Krugman considers jointly. Somehow, the grand conclusion of Krugman/Tobin seems to be that banks must price their loans and deposits in such a way that banking can be an economically viable enterprise, in the same way that other institutions price their assets and liabilities toward the same objective. Again, who would think otherwise, in that general sense?

Krugman/Tobin seems to imply that those who observe “loans create deposits” are sufficiently short sighted to think that this point having been made is also the end of the process. In other words, that these loans and deposits once they’ve been created should remain on bank balance sheets forever. Who would think such a thing? It is interesting that the perception of the intellectual damage done by “loans create deposits” itself doesn’t consider the possibility that the reverse mechanics are equally obvious to those who make the initial point – i.e. that deposits are regularly used to pay down the loans that originally created them, and that in this specific context it is also true that deposits may be used to “destroy” loans. It is interesting if not amusing that those who don’t directly acknowledge what goes on in bookkeeping terms when a prospective borrower goes into a bank branch and successfully negotiates a bank loan then proceed to double up by looking away from what goes on when the same person returns to repay his loan. The idea that deposits are used to pay down loans is not only common place in real world banking, but it is also fundamental to formal circuit theory and its application within post Keynesian economics. What Tobin would not have been in a position to be aware of 50 years ago is that is that the focus of modern post Keynesianism makes this dynamic crystal clear in a rigorous accounting sense – an accounting presentation that covers a galaxy of institutional possibilities with flow of funds analyses that extend Tobin’s succinct summary of same in much more detail. However, in the hubbub of considering this overall financial flow of funds, we are led to believe by Krugman/Tobin that those who understand that loans create deposits don’t understand that deposits can destroy loans in the process of loan repayment.

In fact, Tobin’s description of this essential process is somewhat roundabout, not really identifying the necessary micro accounting that records such flows as they traverse the banking system. Tobin summarizes the macro flow of funds between different financial intermediaries in adept fashion, but without referring to the bookkeeping that must occur within the banking system when a deposit “leaves” a given bank or “leaves” the banking system in its entirety. For example, a deposit cannot “leave” the banking system without some corresponding reduction in a loan or other asset, or some corresponding increase in a non-deposit liability or equity capital – those types of entries being made within the same banking system that the deposit previously occupied. Indeed, this is an example of a required “quadruple entry” accounting constraint under such conditions. We leave the associated task of detailing the corresponding pairs of double entry accounting as an option for commenters, where perhaps it may be seized upon with enthusiastic fervor. This is not hugely difficult, and it is a good complement to that other necessary requirement for economic coherence – accurate central bank accounting.

Tobin makes the point that the banking system expands in such a way that deposits account for only 15 per cent of household wealth. Something happens relative to a counterfactual in which all new saving flows into deposits created by loans. This reflects a straightforward feature of banking system expansion. Individual loans get repaid (or default) but the system still expands. It’s analogous to the job market. The monthly flow of actual gross job creation and gross job destruction is far greater than the flow of net job creation. And similarly for international capital account flows – the gross flows exceed the net flow that matches up with the current account balance. Again, accounting plays an essential role, and such accounting is the stock/flow in trade for the broader post Keynesian object audience (implied knowingly or unknowingly) of Krugman’s criticism.

So we are led to believe that understanding the expansion process described by Tobin/Krugman dominates the relevance of the observation that loans create deposits – because in effect those who make the point that loans create deposits are assumed not to understand that deposits can and do destroy loans in the process of loan repayment. One might be forgiven for assuming that this is intended to characterize certain provocateurs of “loans create deposits” as being members of a particular savant club.

As a side issue, it is unclear what the importance today is of the point being made by these Tobin papers (mostly Tobin Brainard) with regard to their microscopic examination of the “monetary control” effect of required bank reserves or deposit ceilings in 1963. Required bank reserves are a tax in effect, working essentially through price rather than quantity, as the central bank supplies the quantity of required reserves to the banking system after the deposits that generate the requirement are created. On the more relevant pricing front, required reserves increase the overall expense associated with the economics of making a bank loan. Other expenses include such items as employee compensation, the interest rate paid on deposit funding, income tax, and the cost of equity capital. All of these things need to be taken into account in determining break even pricing for a bank loan, so that all expenses including the cost of equity capital are adequately covered. These are things that determine whether a loan that is being considered is actually made by a bank – not the quantity of immediately available bank reserves (in either required or excess form).

Required bank reserves are a form of “monetary control” (Tobin’s language) that have a tightening effect on monetary conditions, in the sense that they push up the interest rate at which loans can be done on economically viable terms, compared to the counterfactual. And the fact that currency may move back and forth between public holdings and bank reserve holdings as a function of interest rate levels is an intriguing analytical curiosity, and one that complicates the calculus of determining the effect of monetary controls, but one of little importance to today’s monetary system and the implementation of central bank policy. Importantly, risk weighted capital measures along with capital adequacy regulations are the modern banking system’s natural successor to these old reserve requirements. In conjunction with that, some countries at the top of their monetary game (e.g. Canada) have recognized the essential role of bank reserves as settlement balances, and when considering that along with technological advances in payments systems, have decided to eliminate required reserves. Excess reserves and their correct interpretation (pre-2008 and QE) are the only real reserve game in town.

In other words, this all seems obvious in the context of what is being analyzed more broadly by many of those at whom the professor is taking full or partial aim (it’s hard to tell who is whom in the cavalcade of assumed culprits). But dusting off Tobin from 50 years ago (admittedly two excellent papers in reference here) just isn’t absolutely necessary in this context, given an accurate understanding of how the modern banking system operates. And the mere fact that Tobin categorizes banks as one type of financial intermediary obviously suggests that there is a common element in that particular scheme – which is what Tobin describes. Great. We can all agree on that commonality and the way in which it captures the mode of setting economic objectives and the corresponding pricing behavior of financial intermediaries in a general way. But this commonality in no way logically negates the elements of uniqueness in banking or their importance.

All this being said, there are some interesting additional points to be made about Tobin-Brainard, although a detailed review of this paper must be deferred to a future post.

First, it is critical to recognize that the model simplification employed in Tobin-Brainard assumes that all government liabilities pay a zero rate of interest – currency, bank reserves, and debt. While this is no doubt done to make orthogonal points about commercial banking with more clarity, it raises enormous questions about the entire framework for “monetary control” being used in that paper. And it has significant implications for the measured effectiveness of monetary controls in that context. Said another way, the absence of a positive interest rate possibility for the fed funds target rate in such a model invites serious questions about why the primary interest rate in the economy is not considered more realistically in its standard role as a mechanism of monetary tightening and easing – in contrast to the core role that the authors employ in the case of currency used with great leverage in a zero interest rate model. That said, I see no reason why the analysis of reserve and interest rate ceiling effects wouldn’t hold in a more complex and realistic model regime of non-zero government interest rates. But with that change the story surely advances in complexity and emphasis, just as the methodical model development in Godley-Lavoie illustrates a step-wise evolution from a starting point of a simple model that bears some resemblance in construction to that of Tobin-Brainard. This consideration should be combined with the more comprehensive importance of bank capital management in determining the fuller scale of “monetary control” in today’s banking system.

Second, we should note something briefly about the contrast between the Tobin and Godley-Lavoie frameworks for banking analysis. (This is very much a personal opinion.) Tobin-Brainard employs a general equilibrium model of portfolio balancing. I think Godley-Lavoie is compatible with this approach, although Godley-Lavoie seems to dispute this to some degree. At least that is my rough interpretation of it. I think Tobin-Brainard is compatible as a model with the modern approach of banks in their strategic balance sheet management, using risk management, pricing and other comprehensive portfolio strategies. I also think that Godley-Lavoie is an enormously sophisticated and accurate view of the logic of actual bank operations as modelled. That operational view is consistent with depicting outcomes of strategic balance sheet management. I see these two different views therefore being compatible rather than conflicting.

Returning briefly and finally to the Tobin paper on money creation, it is obviously a very good one and an important one historically. But it doesn’t refute the intellectual usefulness of points made by those today who correctly observe that central banks provide required reserves after the fact, or that bank finance is essential to economic expansion. The observation that banks are like other financial institutions is obviously implicit in their categorization as one type of financial intermediary. But that this idea should somehow disprove the relevance of the observation that they are unique according to the intended context and meaning of the phrase “loans create deposits” is very strange. Banks are different in the ways described, and these are important differences for the reasons described.