Interest Rates vs. Monetary Aggregates: A (Sort of) Brief History of the Monetary Policy Implementation Debate

As I mentioned in a previous post, I will be doing a series of posts on issues surrounding monetary policy implementation. I already broached the concept of monetary policy implementation vs. strategy, an understanding of which I think is critical for delving deeper into this topic, including what’s contained within this post.

Below, I summarize Bindseil’s account of the history of the debate surrounding the optimal approach to monetary policy implementation. Many of the people and ideas contained within will be familiar to those interested in this topic, but the historical arc along which they appear may be less widely known. Bindseil relies on an extensive set of primary and secondary sources. However, these views are not necessarily unique to Bindseil. I’d imagine Post-Keynesians would give similar accounts, and likely have. I’d also emphasize at the outset that this is a 4 page summary of ~34 pages of Bindseil’s work. Bindseil obviously provides much more argumentation, citation, and quotation than I do here. Moreover, this topic is returned to throughout his ~250 page book.

Interest Rates Win in the Pre-1914 World

Bindseil traces the monetary policy implementation debate as far back as the 18^th century. There were two key questions. One – is it possible for interest rates and/or monetary aggregates to be the operational targets of monetary policy? Two – if both can be targeted, which approach is optimal?

Henry Thornton, acknowledged as one of the most advanced monetary policy theorists prior to the 20th century, published an argument in 1802 that was in favor of interest rates as operational targets. (Interestingly, he further argued that the central bank should align the interest rate with what Wicksell would eventually describe as the ‘natural rate.’)

Towards the end of the 1800s, most British bankers and academics ultimately became convinced of this view: central bank policy meant targeting interest rates, regardless of what the end strategy was. However, it wasn’t an easy road getting to this consensus (and it would be overturned in the 20^th century).

In the 1800s, Britain’s financial economy revolved around bills of exchange, which were essentially forms of short-term debt that firms would issue to finance their operations. Holders of bills of exchange could take them to banks if they at some point wanted to hold cash instead. The bank would “discount” the bills for some amount of cash less than their face value. Banks in turn could go to other banks to "rediscount" the bills if they themselves also needed cash or reserves. This activity led to a market discount rate. If the banking system as a whole was ultimately short of reserves, they could seek recourse at the Bank of England. The Bank of England's discount rate to banks was called "Bank rate." Separate from "Bank rate" discounts, the Bank of England would also participate in the bills of exchange market more broadly, directly offering investors discounts at rates competitive with the market rate of interest. The big questions were whether the BoE could unilaterally set this rate, whether 'Bank rate' led or followed the market discount rate, and if the BoE could 'make bank rate effective' (i.e., drive the market rate to the bank rate). These were very contentious issues. At first, the BoE argued it followed rather than led, not wanting to upset a market that at the time viewed such proactive interference of the Bank as distortionary and unfair (actually, not much has changed…).

Then, in the latter half of the 1800s, along came Walter Bagehot. In Bagehot’s mind, the recurrent financial crises that Britain had been experiencing throughout the 1800s were in actuality avoidable, products of a misunderstanding surrounding the nature of central banking and monetary economics. In 1873, Bagehot published Lombard Street, a book that would change the face of central banking forever. In Lombard Street, Bagehot explained that money markets are inherently unstable when left alone by central banks. This results from the combination of an inelastic supply of reserves with the elastic nature of credit, such that slight excesses or deficiencies in the quantity of reserves can lead to extreme fluctuations in interest rates. In other words, if there are insufficient reserves, rates will shoot to the sky and the payment system may fail; or if there are excess reserves, rates will plummet to 0%. This volatility would be a permanent feature of the market were it not for central banks. As such, Bagehot argued that the central bank should accommodate the market’s need for reserves at some price in order to avoid this volatility. In other words, Bagehot was saying that the central bank can and should target interest rates. The most famous takeaway from Lombard Street is that central banks should act as lenders of last resort to banks during crises, lending indiscriminately but at a high rate and against good collateral.

With Lombard Street, Bagehot temporarily won the argument for what some call the ‘bank rate school,’ or the school of thought that views interest rates as the preferred, if not fundamental, operational target for central banks. In addition to coherent theory, the Bank of England’s implementation of Bagehot’s recommendations gave his argument concrete evidence.

Quantity Theorists Take Over for Most of the 20^th Century

In opposition to the bank rate school arose a new class of quantity theorists who viewed monetary aggregates as the preferred operational target of central banks.

Bindseil traces the resurgence of the quantity view to the synergistic effect of two academic contributions. The first is Irving Fisher’s popularization of the quantity theory of money (QTM) in his 1911 paper The Purchasing Power of Money. The QTM is often summarized by the equation MV=PY. The second was C.A. Phillips’ 1920 ‘discovery’ of a relationship between the quantities of bank reserves and the quantity of bank deposits, referred to as the ‘money multiplier.’ Philips’ work and others’ worked magically together as they seemed to link the QTM with monetary policy implementation. The reasoning was that since the CB can affect the quantity of reserves via open market operations, they can therefore influence credit creation by impacting the level of excess reserves, which would then stimulate more lending. This line of thinking would eventually be labeled under the ‘reserve position doctrine.’ Any student of economics is aware of the money multiplier story. Indeed, it has dominated economics textbooks for the past 70 years around the world. Bindseil notes that Mishkin (2004) still devotes forty pages to it.

Of course, as will be argued, the balance of the debate has shifted in recent years. Indeed, Bindseil, Post-Keynesians, and recent mainstream research reject this view of the so-called money multiplier. I found this anecdotal observation from Bindseil interesting:

“This is in marked contrast to central bank practice. At least since joining central banking in 1994, the author has never heard any reference to the money multiplier in internal discussions on policy decisions or in public announcements from central banks, in particular with regards to open market operations.”

In any case, Bindseil notes that there was much focus on monetary aggregates long before the development of monetarism in the 1950s and 1960s. Bindseil cites many primary sources from academics and the central bank to support this notion. Surprisingly, Bindseil identifies Keynes as one important contributing factor the success of quantity theorists. Bindseil notes that Keynes attacked the bank rate school in the second volume of his Treatise on Money (1930) for their focus on interest rates as opposed to money aggregates. Keynes called this the ‘orthodox doctrine in England.’ Specifically, he criticized the idea that the ultimate purpose of open market operations, which alters monetary aggregates, is to effectuate the central bank’s interest rate policy. Keynes believed OMOs had independent and significant impacts on the economy. Bindseil states that Keynes was also the first to argue that changes of reserve requirements should be used as a monetary policy instrument, an idea that came to dominate monetary policy theory until the 1970s.

Then came Milton Friedman and the monetarists. Friedman argued that OMOs alone are a sufficient tool for monetary policy implementation, and that standing facilities and reserve requirements could be done away with. Friedman in general gave little attention to interest rates, and in select papers, Bindseil notes that Friedman argued that interest rates as operational targets are neither possible nor desirable.

Bindseil quotes several passages from Friedman along these lines and identifies what he sees as flaws. He also notes that a Friedman-like view of monetary policy implementation is still prevalent today, referring again to Mishkin’s popular textbook but also to Allan Meltzer’s history of the Fed. Bindseil quotes several relevant passages from both authors, and highlights several errors in his eyes, including the conflation of monetary policy strategy with monetary policy implementation (this comes up a lot).

Poole’s 1970 Error that Sends Monetary Economists Down a Rabbit Hole

Bindseil argues that Poole’s (1970) model of optimal monetary policy instrument choice is what likely drove economists’ belief in the ability to control monetary quantities in the 1970s and 1980s. As I’ve detailed above, prior to Poole, the monetary policy implementation debate mainly placed interest rates in opposition to quantities. Poole’s supposed major breakthrough in his 1970 paper was that either target could be suitable but depended on “the stochastic structure of the economy, and that a combination policy was likely to be the best approach.” Bindseil reviews Poole’s model and logic, acknowledging its simplicity and elegance, and explains that monetary economists would in the following decades look to develop more mathematically sophisticated variants of Poole’s model to help in the instrument-choice problem of monetary policy.

However, Bindseil then proceeds to point out several flaws in Poole’s paper, and notes that the eventual academic return to interest rates as operational targets in the 1990s actually happened without reference to Poole. Bindseil argues the flaws in Poole (1970) stem from a conflation between the concepts of instruments, operational targets, and intermediate targets. As a result, Bindseil says that Poole fails to distinguish between short-term and long-term interest rates; reserve market quantities and monetary aggregates; and macroeconomic shocks and reserve market shocks. Ultimately, the problem is that Poole’s model simply assumes that central banks can control monetary aggregates on a day-to-day basis, an operational possibility that does not exist according to Bindseil, many central bankers, and academics around the world. (Later in the book, Bindseil does acknowledge and explain that in some sense, central banks can control reserve aggregates under an excess reserve policy. However, Bindseil’s discussion of Poole and Poole’s paper do not assume this type of regime.)

Nonetheless, monetarists finally got their way, when from 1979-1982, the Fed explicitly attempted to target monetary quantities. Theory finally got put into practice. As it turns out, the experiment is widely regarded by most as an abysmal failure. Interest rates were extremely volatile, and even the quantities were as well. I am told the success of this experiment is still debated in some circles, although it does seem there is a general consensus that the experiment simply did not work. After all, the Fed abandoned it. It seemed that intermediate or final targets were the only candidate roles left for monetary aggregates to play. Bindseil puts it as such: “… a clear sequence remains: first the observation of news on the appropriate monetary policy stance (for example, news on monetary quantities), and then the decision to adjust the operational target, namely, the short-term interest rate, such as to bring the (intermediate) target variable back on track.”

Bindseil observes the lasting influence of Poole (1970), by for exampling noting the emphasis that Walsh’s 2003 textbook places on it. All in all, Bindseil concludes that the quantity view of monetary policy implementation lacks coherent micro-foundations and especially a model of bank behavior in the interbank market. He also believes the empirical research to be flawed, “mixing macroeconomic and microeconomic times series and shocks of different significance and frequency.”

The Interest Rate Returns

A trickle of work in the 1980s starts to re-emphasize interest rates as monetary policy targets. Bindseil cites McCallum, Barro, Judd and Motley, and Taylor. Then in the late 1990s, it becomes apparent that the new approach to modeling monetary policy implementation revolves around the steering of interest rates. Bindseil cites Hamilton; Bartolini, Bertola, and Prati; Woodford; Ayuso and Repullo. Bindseil also highlights Woodford’s comments in Interest and Prices regarding the significant errors academics have made in the past with respect to modeling monetary policy implementation. (I was delighted to see this, as I recognized this here as well.)

Ironically, some of these modern approaches borrow from a paper Poole wrote in 1968, which proposed a simple but effective model that seemed to be discarded after Poole (1970). In general, Bindseil argues that in contrast to the old ‘Bank rate’ lit, the literature of today models much more convincingly the way in which central bank’s control interest rates, with the help of modern mathematical tools. Nonetheless, we’ve essentially returned to a consensus view that was held more than a century ago, having taken a detour through most of the 1900s.

Bindseil emphasizes that understanding the microeconomics of the reserve market is crucial to understanding modern day-to-day monetary policy implementation. He identifies only a sparse collection of academic research in this area throughout the 1900s. It wasn’t until the 1990s that this line of research built a strong corpus of work. This may be in part why it took so long for the interest rate view to again dominate academia. Indeed, Bindseil asserts that the new micro literature on this topic has unambiguously “demystified” the differences between the various instruments of monetary policy implementation (e.g., OMOs, standing facilities, reserve requirements) and has made clear that they all work together to achieve a target interest rate.

And so here we stand.

Upcoming

In future posts, I hope to explain some of the monetary policy implementation models that Bindseil ends his history with. I’ll reconcile these with explanations I’ve seen from Post-Keynesians and others that share similar views. At some point, I’d also like to dive into the quantity-oriented models and share the more substantive critiques of these models. I expect, however, that much of this will be possible by focusing on the insights of interest-rate oriented models.