Imposed versus Adopted Monetary Rules

Steven Horwitz

There’s been some recent controversy over the difference between rules and discretion in monetary policy, started, innocently enough, by a Don Boudreaux blog post and a Daniel Keuhn comment, then a full post by Daniel.  Bob Murphy weighed in as well.  I think part of the problem here is a dual meaning of the word “rule” in the context of rule-based monetary policy.  The difference here is between a rule adopted by the central bank and one imposed on it.

One kind of rule is what we might call a policy-guiding rule (or what Buchanan might call a “post-constitutional rule”), such as the Taylor Rule.  Yes, it’s a rule, but it’s a rule that informs the central bank of what sort of action it should take. So yes, the bank does not have full discretion if it is committed to that rule.  However, the rule is one self-adopted by the central bank and, as Bob points out, if that’s subject to revocation in the face of a crisis, it’s really not too much of a rule. 

Contrast that with a policy-constraining rule (or in Buchanan’s terms a “constitutional rule”).  Imagine a constitutional provision, or even one imposed by the legislature, that said the money supply can only grow at 3% per year.  Period.  Such a rule would be much more constraining than a policy-guiding rule, and far less open to interpretation and not changeable by the central bank.  It would also be, especially if constitutional, much harder to change, even in the face of a crisis.

It is the latter kind of rule that Friedman supported and it is the former kind of rule that commands the support Daniel refers to.  They are not the same kind of “rule” despite both using the same word.  The constitutional level rule is designed to tie the hands of the central bank and turn it into a conduit for a fixed growth rate.  The policy-guiding rule tells central bankers what changes in the money supply they should be making, and it subject only to the whim of those in charge of policy at any given time.  It is this last point that makes people sympathetic to Friedman unwilling to call that a “rule” in the same sense – it ultimately doesn’t constrain the central bank because it still has the discretion to change the rule.

Put differently, a policy-guiding rule is adopted by the central bank;  a policy-constraining rule is imposed on the central bank.  Their purposes are very different.

It is in this sense that Friedman opposed discretion and those in favor policy-guiding rules do not.  Friedman’s rules are not changeable by the central bank itself;  policy-guiding rules are.  The central bank has discretion to pick its “rule.”  That, Friedman thought, was the whole problem, hence his call for a rule to be imposed.

This difference is the same one Hayek talks about when he differentiates between the decentralized rules of justice and the legal order and the rules that firms might adopt internally to guide their behavior within that hierarchy.   The rule that says respect private property is a different kind of rule than the one that tells you who you must get permission from to borrow the corporate car.

So yes, everyone is talking about rules, but not all rules are the same.  The fact that critics such as Daniel don't see this should challenge us to be as clear as possible with our uses of the word "rule" and challenge them to read carefully and not assume everyone is using the word "rule" they way they would.

Comments

[Daniel Kuehn]

Thanks for linking. I probably come down less decidedly than you, but I believe we are more on the same page of the issues at hand than you seem to think: http://www.factsandotherstubbornthings.blogspot.com/2012/07/horwitz-on-monetary-rules.html

[Roger Koppl]

In a 2004 AEA piece, Alan Greenspan criticized a too-close adherence to rules such as the Taylor rule. Noting the importance of Knightian uncertainty he says, “our knowledge about many important linkages is far from complete” and we must therefore use “judgment” in monetary policy (Greenspan 2004, pp. 36-38). In Greenspan’s sense, “judgment” cannot be reduced to a simple mathematical model. “For such judgment, policymakers have needed to reach beyond models to broader, though less mathematically precise, hypotheses about how the world works” (Greenspan 2004, p. 38). The following further quotes (from pp. 38-39) shows, I think, that we are definitely talking about discretion:

“Some critics have argued that such an approach to policy is too undisciplined— judgmental, seemingly discretionary, and difficult to explain. The Federal Reserve, they conclude, should attempt to be more formal in its operations by tying its actions solely, or in the weaker paradigm, largely, to the prescriptions of a simple policy rule. Indeed, rules that relate the setting of the federal funds rate to the deviations of output and inflation from their respective targets, in some configurations, do seem to capture the broad contours of what we did over the past decade and a half. And the prescriptions of formal rules can, in fact, serve as helpful adjuncts to policy, as many of the proponents of these rules have suggested. But at crucial points, like those in our recent policy history (the stock market crash of 1987, the crises of 1997–1998, and the events that followed September 2001), simple rules will be inadequate as either descriptions or prescriptions for policy. Moreover, such rules suffer from much of the same fixed-coefficient difficulties we have with our large-scale models.”

“Inflation-targeting often originated as a fairly simple structure concentrating solely on inflation outcomes, but it has evolved into more-discretionary forms requiring complex judgments for implementation. Indeed, this evolution has gone so far that the actual practice of monetary policy by inflation-targeting central banks now closely resembles the practice of those central banks, such as the European Central Bank, the Bank of Japan, and the Federal Reserve, that have not chosen to adopt that paradigm.”

Thus, it seems to me that Daniel Keuhn may be underplaying the role of discretion in real-world monetary policy when he says, “To be honest, I doubt anyone ever did support discretionary policy.” Maybe he meant just in the academic literature. But if so, then I don’t know what he means when he says, “It's true, before Barro-Gordon and Kydland and Prescott the rules people followed were probably more ad hoc.” Maybe he meant “the rules academics advocated” rather than the rules “people followed.” Anyway, I would say that Kydland & Prescott and so on transformed the debate. I don’t think they were attacking a straw man or something. The advocates of discretion tended to speak of “stabilization policy.” I have not read his essay linked above, however, so I'd better hedge on my interpretation of Dan's views.

PS: Sorry about the long block quotes, but it would take a lot of time to trim artfully.

[Daniel Kuehn]

Roger -
Well the discussion, as it came up on FB and over at Don's blog, had more of the academic literature in mind, and my point was simply that the Keynesian literature is a rules-based literature.

I got a little broader in my post, and I do note that before "rules vs. discretion" became an issue in the academic literature the "rules" that bankers actually followed were probably more ad hoc. I'm not the best judge of whether Greenspan could be slotted under "rules-based but ad hoc" or not. We could add "implicit" or "tacit" or "unwritten" to "ad hoc" in the list of adjectives describing rules in practice. This (along with profit and loss) is the sort of rule that would govern a free banking system, I might add (which is ultimately what Steve rejects a monetary constitution in favor of, if I'm not mistaken). Private banks issuing their own money would certainly be following tacit or ad hoc rules in a lot of cases, and profit and loss would discipline those rules.

So where does discretion end and "ad hoc/implicit/tacit rules" begin? Is Lombard Street - for example - a treatise on policy rules or is it a treatise on discretion?

Not only do I not know the answer, but I think it's probably a silly question (which I suppose kind of gets me off the hook for not knowing the answer!). This obviously isn't as dichotomous as "rules vs. discretion" implies.

And what if we figure out a rule is actually a bad one? Wouldn't we want some opportunity for a regime change? We don't like rules for the sake of having rules, after all! Would this be counted as "discretion" and therefore verboten?

[Will.Luther]

I find Steve's contrast between a policy-guiding Taylor rule and a policy-constraining 3% M-growth rule a little confusing.

Can we have a policy-constraining Taylor rule? If, for example, the constitution required the Fed adhere to a Taylor rule, would that be sufficient? Or, rather, is the problem that a Taylor rule mandates an outcome whereas the 3% M-growth rule mandates an action?

In other words, are you saying (1) the Taylor rule WAS NOT a policy-constraining rule over the last decade or (2) the Taylor rule CANNOT be a policy constraining rule in practice. (Obviously, accepting 2 would imply 1...)

[Steve Horwitz]

Will,

It has NOT been a policy-constraining rule. In theory, perhaps it could.

[Jerry O'Driscoll]

The gold standard was a rule in the strong sense (policy constraining). It evolved over time, but was eventually imposed on the bank of England by statutes.

The BOE evolved operational practices to maintain the gold standard (i.e., obey the rule). Bagehot was more concerned about practices (in my terminology); he took the gold standard for granted. The gold standard was the rule.

Free banking had a rule, which was imposed on it by customers and contracts between customers and producers (banks). Markets are complex webs of rules, evolved over time and sometimes imposed.

When you turn to a fiat money system, the idea of a rule becomes more complicated -- especially outside the confines of abstract models. As Meltzer and others have argued, any rule must come with an incentive system. Authority and responsibility must be unified in one decisionmaker.

To my knowledge, that was only done with the Reserve Bank of New Zealand and eventually abandoned in favor of a committee. Don Brash had a contract and his compnensation depended on his achieving specified (legislated) goals.

To me, what Greenspan is describing merits the term "discretion." He knows better, as can be discovered either by reading his early wrirings from his Objectivist days or his more recent statements. (A central bank is unnecessary under a gold standard.)

[Jonathan M.F. Catalán]

To build on Jerry, a great advantage of the market is that despite there being "rules," there is also room for discretion (e.g. entrepreneurial discretion). Yet, there are constraints (viz. profit and loss); this balance is one of capitalism's (i.e. the market process') great virtues.

[Aidan Walsh]

Now this is an interesting can of worms, with implications beyond monetary policy.

'Externally imposed rules'; 'coercive rules'; 'internal commitment to rules'; 'internal self-imposed rules'; 'externally imposed rules but to which there is internal commitment', 'coercive/externally imposed rules for which there is no internal commitment'; 'objectives' versus 'abstract rules'; discretion in absolute terms or 'discretion' in terms of following an exception (say a newly discovered exception by an entrepreneur) to a more general rule or 'discretion' in terms of adhering to tacit rules. All have completely different effects in terms of binding behaviour.

I think what Greenspan, above, is writing about is less about discretion and more about achieving a particular goal or objective; in effect Greenspan's goals for the US economy superseded any rules that were supposed to constrain him/Fed policy?

This emphasis away from rules towards objectives is seen clearly in some stuff I was reading about banking regulation: ‘[T]here [should be] a focus on risk not rules…. Regulators are overburdened by rules. They cannot enforce every one of these rules in every firm at every point in time.’ Baldwin, R., & Black, J. (2007). Really Responsive Regulation.

So we have monetary and regulatory authorities who do not feel that they are bound by any rules in attempts to achieve their goals (of 'saving the euro' etc) but expect us all to obey their externally imposed coercive rules in the way that they expect.

Elinor Ostrom wrote: ‘Rules provide stability of expectations, and efforts to change rules can rapidly reduce that stability’.

Ostrom described this process in relation to an irrigation system in Sri Lanka. Here old internally imposed and maintained rules that had resulted in a sustainable system were changed by expernally imposed coercive rules, in 1958, with the objective of making the system more ‘democratic’; political control of the system meant that getting elected became the goal rather than a sustainable irrigation system. Sanction by peers disappeared, water rates collection ceased and large landowners obtained special privileges. By the 1970s, the chief engineer of the system concluded, ‘there is no law now’. Ostrom, E. (1990). Governing the Commons. Cambridge: Cambridge University Press. p. 164).

[Doug]

John Taylor last night posted a long blog on the occasion of Milton Friedman's birthday. The content though is mostly historical background and analysis concerning rules versus discretion.

Sorry, I haven't had time yet to read all of the other blog links here. But I would be interested if anyone knows the context for this quote from Professor Taylor's post (what Hayek said or wrote to trigger Friedman's response):

"From the mid-1960s through the 1970s the Samuelson view was winning with practitioners putting many discretionary policies into practice.

"But Friedman remained a persistent and resolute champion of his alternative view. At one time during the 1970s, F.A. Hayek even seemed to be siding with the discretionary approach, at least in the case of monetary policy. But Milton Friedman didn’t waver. In fact he sent a letter to Hayek in 1975 saying: 'I hate to see you come out as you do here for what I believe to be one of the most fundamental violations of the rule of law that we have, namely discretionary activities of central bankers.' ...

http://johnbtaylorsblog.blogspot.com/2012/07/still-learning-from-milton-friedman.html

That is a long and substantive post on rules vs. discretion ==> I don't mean to change the topic here.

[Roger Koppl]

What Jerry said.

Whether you are more on the side of rules or of discretion depends on one's benchmark, I suppose, but Greenspan does use the words "judgment" and "discretionary." And I think it's pretty clear that he was aiming straight at John Taylor. Note the timing. Greenspan gave his talk and the paper was published when the Fed was deviating from the Taylor rule.

I think the only coherent context for the discussion of "rules vs. discretion in monetary policy" is, well, *policy*. The context is a system with a central bank that has some influence on the money supply. The debate is probably most clear in the context of a pure fiat money, but it is there for most versions of "the" gold standard too, I think. (I think "most" probably means anything but 100% reserves.) As far as can tell, there is no question of "rules vs. discretion" in the context of free banking with no central bank.

IMHO, BTW, Friedman's point about the rule of law is important.

[Daniel Kuehn]

Aidan -

I agree it's very important to bring Ostrom into this discussion, but one thing to note is that that was an externally imposed rule on the irrigation system. That's presumably going to have very different results from a rule change that emerges out of the arguments and discussions of central bankers and monetary economists.

If anything, I think the point is a caution against insisting that these rules have to be imposed constitutionally. That might be a good idea - I'm not ruling it out. But a constitutional rule designed by a convention or the state legislatures isn't obviously going to be better than a policy rule that emerges over the years among central bankers.

[Steve Horwitz]

Daniel,

Perhaps you can clarify what an "ad hoc rule" is. From where I sit, that's a married bachelor. The whole point of a rule is to avoid ad hoc behavior. Ad hoc means, among other things, "non generalizable," which seems to me to be the exact opposite of a rule - which is meant to be applied across a variety of situations.

[Daniel Kuehn]

That's a good point. "Imprecise" is probably a better word than ad hoc, because ad hoc implies they're flying by the seat of their pants. The best example of what I mean is the fact that you have a rule like the Taylor rule, which is very precise - but the Fed was following something like the Taylor rule quite consistently since 1980 or so. That's how we got the Taylor rule, after all - studying their consistent behavior. There was some imprecisely or less defined rule governing the Fed response - probably largely based on tacit knowledge.

That sort of consistently applied tacit knowledge seems important to me, even if you don't write it down in a single growth rate or a single equation.

[Barkley Rosser]

I agree with Jerry that Greenspan was a discretion man, all the way, whatever occasional nodding or rhetoric he gave to the contrary. It is reputed that he blocked any consideration of Taylor for consideration as his successor (or for even being on the Bd of Govs) thanks to his annoyance with Taylor's constant prodding for the Fed to adopt his (Taylor's) rule, such as it is (it has some discretion built into in terms of how one determines certain parameters that go into it). As it was, there you had Taylor from about 1991 on saying "This is what the Fed has been doing since 1985, and it is a good thing (my "rule"), and they should enshrine it as a rule," with Greenspan saying "no no no" all the way.

Regarding Friedman, 1985 is a crucial year in that was when the Fed effectively began following the Taylor Rule (according to Taylor), only to stop following any rule during the past decade or so. That of course involved focusing on interest rates rather than money supply, and in the last year of his life, I gather that Friedman accepted that central banks were simply beyond following his old rule, and he came out for inflation targeting, in contrast with the Taylor Rule, as his preferred policy.

[Michael G Heller]

The important point John B Taylor rightly makes in his latest post is that discretion-vs-rules goes way beyond just central bank monetary policy.

“more stable fiscal policies based on permanent tax reforms and the automatic stabilizers... more predictable and rule-like monetary policy… the consequences of the discretionary bailouts” etc.

You might also like to consider his oped on Hayek on rules in WSJ -- http://online.wsj.com/article/SB10001424052702303674004577434774238817962.html

Taylor quoting Hayek: “freedom is dependent upon certain attributes of the law, its generality and certainty, and the restrictions it places on the *discretion* of authority”.

James M Buchanan took it further than any of you guys -- no discretion, never, nowhere, nohow. It’s very broad political economy, which is the point I tried to make in my post posted 27 July (beat Don to it!):

http://www.project-syndicate.org/blog/very-likely-indiscretions-of-discretionary-regulation

For example no one here mentioned macroprudential policy. I know it's Milton's b'day and everything, but his perspective, however laudable, could be limited on cross-disciplinary questions.

[Jerry O'Driscoll]

Would anyone care to translate "macroprudential policy" into English?

A couple of years ago I participated in a gathering of experts on the financial crisis. "Systemic risk" kept being dropped until someone requested a definition. Numerous participants offered their definitions; no two agreed.

[Roger Koppl]

Why Jerry, the distinction between "macroprudential" and "microprudential" is every bit as clear and coherent as that between "macroeconomic" and "microeconomic." Gheez, *any*body knows that! :-)

There is a JEP article on that probably expresses something like the modal view of such things. I think this particular article is open access.

http://pubs.aeaweb.org.proxy.libraries.smu.edu/doi/pdfplus/10.1257/jep.25.1.3

As you know, "time-varying capital requirements" come up a lot under the label "macroprudential." Another one that comes up a lot is in the Hanson et al. JEP paper: regulating the "shadow banking system."

[Roger Koppl]

BTW, I am unhappy with the term “shadow banking system,” which suggests something irregular, unsavory, and dangerous. The term seems to have been coined by Paul McCulley who defined it disparagingly as “the whole alphabet soup of levered up non-bank investment conduits, vehicles, and structures” (McCulley 2007, p. 2). The Urban dictionary (www.urbandictionary.com) defines “alphabet soup” as “A various concoction of illegal drugs, taken concurrently, the effects of which cannot be anticipated.” Presumably a “leveraged up” shadow bank is like a hyped-up drug user high on an alphabet soup. McCulley complains that the shadow banking system is “vulnerable to runs” because deposits are uninsured and the system is not “backstopped by access to the Fed’s discount window” (2007, p. 2). McCulley wanted the shadow banks to come out of the shadows and into the regulatory fold of the Federal Reserve System. Thus, the very word we use to identify many non-bank financial intermediaries suggests that they are dangerous, perhaps irrational, and must be controlled. The use of such language is argument by insinuation.

[Jerry O'Driscoll]

Thanks, Roger.

I take the "shadow" terminology to mean "out of sight," but not necessarily in a perjorative sense. It certainly has been understudied, and it was at the heart of the financial crisis.

The Fed seems to have engaged in benign neglect of it. I asked a senior Fed official if anyone in the system had been studying it before the financial crisis. He just shook his head no.

[Jerry O'Driscoll]

To follow up on my last comment. Some have noted that there wasn't much money growth in the Greenspan era. But take a look at velocity.
http://research.stlouisfed.org/fred2/series/M1V?cid=32242

I suspect the upswing in M1 velocity, and its subsquent collapse, mirrors the growth and collapse of shadow banking. Some who comment here think the collapse in velocity argues for yet more monetary stimulus. If movements in velocity reflect the financing of asset bubbles, however, I find the policy recommendation unpersuasive.

Do proponents of monetary stimulus crave another bubble? Do they, like Fed officials, confuse bubbles with prosperity? If a trend is unsustainable, it is not a guide to policy.

Velocity is back to 1980s levels. That wasn't such a bad time, was it? The prosperity of the 1980s was real. So, too, I think were the early Clinton years. By the end of the 1990s, marked by the inauguration of the Greenspan Put, the economy began to run on fumes.

[Leonardo IHC]

The Taylor rule is a formalisation of a "discretionary" behaviour aimed to manage the trade-off between inflation and GDP growth (or inflation expectation vs output gap, or unemployment... the same staff in the end). A full discretionary Central Bank with a dual mandate would simply act by balancing inflation and GDP perspectives, quite the same as the Taylor Rule tells; the Taylor Rule was actully "found" as a regression of central rates on inflation and GDP, then this statistical result was taken as a recipe for Central Banking. As a constraint, then, is actually a bit "loose" (accept the rule to act as if you were discretionary doing what a consciuous discretionary central banks would).

No discretionary Central Banks would ever spontaneously act in line with a Friedmanian rule, where no degrees of discretion is permitted. This is a rule in strict sense.

[Roger Koppl]

I wish we wouldn’t say things like “That's how we got the Taylor rule, after all - studying [the Fed’s] consistent behavior” (Daniel Kuehn) and “the Taylor Rule was actully ‘found’ as a regression of central rates on inflation and GDP, then this statistical result was taken as a recipe for Central Banking” (Leonardo IHC). Have the people who say such things actually, you know, *read* Taylor’s famous 1993 paper? The Taylor rule emerged from the theoretical and empirical literature. Using a rational-expectations model, Taylor ran many simulations to see what sorts of rules seemed to work better in different countries among the G-7. The upshot was his famous rule. He then found that this theoretically derived rule fit the “recent” behavior of one central bank, the Fed, pretty well. It fit the 1987-1992 period well, due account being taken of the “significant deviation” associated with the cash of October 1987. That's just not the same thing as seeing what the Fed did and going "cool rule, dudes."

[Leonardo IHC]

... which means that we take as a "rule" a behaviour consistent with discretional balancing the inflation-GDP trade-off (discretional in the sense that no one really knows what the "social" weights of GDP and inflation are, unless you assume perfect concentrated knowledge).
In different terms, just to stress the point: I put some doses of sugar in my coffe; someone leads a research which concludes that a model with one dose and half of suger fits very good my behaviour; then I decide to follow the 1 and 1/2 doses of sugar in my coffee. As the rule mimicks my discretional behaviour, is it "really" a constraining rule for me? Moreover, if it is not such constraining, can we really call it "rule"?
That is why I consider Taylor Rule not as a Rule in strict terms, so that it takes no real monetary discipline; it just formalises and old inflationistic central behaviour. A Rule has to force you to do something you won't naturally do, in this case not exploiting the power to vary money supply for the "sake" of people and GDP.

Just my humble economics-amateur opinion.

[Roger Koppl]

I don't think I understand your last comment, Leonardo. In terms of your analogy it's more like this: Someone does a lot of research, looking into the health of different people who put different amounts of sugar in their coffee, and so on. After tons of study they decide that 1.5 is the right amount. They see that just lately, though not sooner unfortunatey, you've been using 1.5 doses. Noting this fact they say, "Good, keep using that much; stick to that rule." Such a rule does NOT "mimic [your] discretional behaviour."

[Daniel Kuehn]

Roger - thanks for the correction.

The last sentence of your post doesn't capture what I was implying, of course. It's not just a matter of "cool rule, dude". I'm definitely aware of its roots in the theoretical literature, and my understanding was it emerged by observing the behavior of the Fed in a period it did well, to parametrize a rule consistent with the theoretical literature. So, needless to say, not quite "cool rule dude".

I did not know it was generated by looking at a much broader sample of bank behavior and that the match with the Fed was more ex post. So thanks for the clarification.

[Leonardo IHC]

@Koppl
I try explain my point the last time just to have a talk with you, as my opinion is anyway not so fundamental for economic research and it will make you just waste time.
My example says that you can have researchers who conclude about a prescription under some hypotheses (a certain monetary policy rule under balancing inflation and growth or 1.5 sugar doses under human biology and tastes), then the guys who can be interested in this research (Central Banks, and I) say "hey, this model fits right my behaviour, so I can say I am following that rule!" and all people around (market operators, and those who care of my health) say "yeah, now you are following the rule, you are not discretionally acting, so we are happy", but those interested guys (Governors and I) just keep acting the same as in full discretion, so the difference is all in the perception of the people around but not in the drivers of (Governors' or mine) actions.

A central bank creates 4% monetary inflation to push GDP 1%, so goes inflating - and an Austrian knows what it really implies. Whether it is because of an "individual" estimation about pro's and con's or follows from a rule which weighs pro's and con's in the same way, does it really makes a difference?
Taylor rule, in my view, looks like window dressing in terms of policy, so I cannot consider it as a real rule.