Does the Fed go too slow and too small: how repetitive are Fed hikes and cuts?
Or: why high serial correlation of balanced discretionary decisions is a sign you should move faster
The next Fed (Federal Open Market Committee) meetings are on September 17-18, and after a sluggish late summer dip in economic activity, commentators and investors are expecting an interest rate cut.
That cut would be a change. The Federal Reserve has held interest rates steady for about a year, and before that it progressively raised interest rates from early 2022 to mid 2023.
What should the Fed do in September? That’s tough to answer.
Maybe more interesting is the question: how should the Fed make decisions when it’s contemplating changing course?
It’s clear that for some Fed governors their calculus is different when it comes to taking an action that’s different from what they did last meeting. On August 13th, Federal Reserve Bank of Atlanta President Raphael Bostic said
he’s looking for “a little more data” before supporting a reduction in interest rates, emphasizing he wants to be sure the US central bank will not have to change course once it begins cutting.
“We want to be absolutely sure,” Bostic said Tuesday in remarks to the Conference of African American Financial Professionals in Atlanta. “It would be really bad if we started cutting rates and then had to turn around and raise them again.” [emphasis added]
Perhaps Bostic has since found the data to give him more conviction, but what’s more important is his background framework that it would be bad for the Fed to cut and then find themselves in a situation where they cut too much and need to raise rates again.
The fear of change resides in other Fed governors, as then San Francisco Fed President John Williams said in 2016: “A sudden reversal of policy could be disruptive and slow the economy in unintended ways.”
The Fed’s self-description does not discuss avoiding changing course, but you can imagine how that conclusion might echo off its words. Per the Fed,
The Federal Reserve Act mandates that the Federal Reserve conduct monetary policy "so as to promote effectively the goals of maximum employment, stable prices, and moderate long-term interest rates."
Working near the Fed and hearing again and again the goal of stable prices, the goal of stable prices could expand to a disposition towards stable actions, all part of bankers’ endless (and valuable) search to project confidence and constancy. Some of that conservatism in acknowledged in the Fed’s described principles for making monetary policy
A third principle is that the central bank should raise the policy interest rate, over time, by more than one-for-one in response to a persistent increase in inflation and lower the policy rate more than one-for-one in response to a persistent decrease in inflation. [emphasis added]
How can we see the Fed’s gradualism in its actions? We’re lucky in that the Fed - perhaps unique among large public institutions - is working against a long-lasting mandate and has a limited set of easily observable actions it can take: raise or hike interest rates, cut or ease interest rates, or keep them the same.1
Let’s focus just on cuts and hikes, as those are the more significant actions and the kind of reversals that it seems like Fed governors are concerned about.
Since 1954, Fed actions, both cuts and hikes, are more likely to lead to the same action next time: 65% of the time the past action is the same as the next action, and that’s true 61% of the time for cuts and 68% of the time for hikes. See footnote for more on methodology.2
After the repetitiveness of Fed actions fell slightly from 1990 to 2004, it’s increased again over the last two decades: it was 77% from 2015 to 2019, and it’s 66% from 2020 to 2024.
It can be tough to evaluate Fed actions near the zero lower bound. The above chart incorporates the Wu-Xia “shadow rate,” highlighted by the Atlanta Fed, when the Fed Funds rate is <= 1% to incorporate the impact of unconventional monetary policy. For a complementary perspective, below are the same calculations just applied to the Fed Funds rate, ignoring the impact of unconventional monetary policy at the lower bound.
Bounded at zero, that calculation shows lower repetitiveness of Fed cuts during the 2010 - 2019 timeframe. But that’s a very incomplete picture, missing massive monetary policy activity through unconventional means. Because of that, I think including the Wu-Xia shadow rate is the better way to evaluate the repetitiveness of Fed actions, even though the shadow rate calculations are less clear than directly observing changes to the Fed Funds rate.
But is it bad to repeat yourself?
I ask you again.
Estimating how monetary policy action flows through the economy is complicated, and gradualism could have its place. But even within monetary policy, there are suggestive clues that a pattern of repetitiveness - another way to say a bias towards slowness - may be less than ideal. Matthew Klein in FT Alphaville noted a few years back that the San Francisco Fed released a “Chair the Fed” interactive monetary policy simulation, and that the best approach was to make much larger moves than the Fed typically does.
The Fed may not be along among market participants in underreacting to economic news. Time series momentum - the tendency of a financial securities’ past returns to predict future returns - is one of the most robust (though also somewhat befuddling) findings in financial economics. And as a trader, which perhaps you can think of the Fed as, patterns of serially correlated trades (i.e., a pattern of making the same trade again and again) would suggest you should move faster and larger.
But I’m intrigued to what extent the repetitiveness of Fed actions reflects general cultural values of gradualism and consistency in decision-making that we should adjust.
Lots of language picks up this positive association between slow, gradual change and success and wisdom. Slow and steady we win the race by going one step at a time. Maybe you’re not convinced yet: wise not to rush to judgment.
And other phrases we use stack the deck against drastic actions: Don’t panic is definitionally correct advice in so far as panicking - sudden unreasoning terror often accompanied by mass flight, per Merriam Webster - is never right. Accusations of flip-flopping against rival politicians still hold weight among politicos, though voters may be less focused on that. And of course consistency can be deeply good, in our habits, in our values, in our commitments to family, friends, and neighbors, in how we treat coworkers and customers.
But discretionary institutional decisions made under uncertainty should less likely have a consistent pattern to their judgments. In institutional decision-making, we can draw some differences in the type of decision-making patterns. The most influential modern formulation is Jeff Bezos’ framing of two types of decisions:
One common pitfall for large organizations – one that hurts speed and inventiveness – is “one-size-fits-all” decision making.
Some decisions are consequential and irreversible or nearly irreversible – one-way doors – and these decisions must be made methodically, carefully, slowly, with great deliberation and consultation. If you walk through and don’t like what you see on the other side, you can’t get back to where you were before. We can call these Type 1 decisions. But most decisions aren’t like that – they are changeable, reversible – they’re two-way doors. If you’ve made a suboptimal Type 2 decision, you don’t have to live with the consequences for that long. You can reopen the door and go back through. Type 2 decisions can and should be made quickly by high judgment individuals or small groups.
A less frequently drawn out corollary is that Type 2 decisions should often be reversed. That’s sometimes grasped when evaluating pilots and the (tough to internalize) self-awareness that many pilots should fail (though the value from pilots collectively should also be very high!). A low failure rate is a sign of either risk aversion in launching or rigorously evaluating pilot outcomes. Another way to say this is that repeated Type 2 decisions should be uncorrelated to each other, after adjusting for the background benefits and risks of the options. And a pitter patter of repeated institutional Type 2 decisions (“yes, let’s make the same choice again!”) may reflect a failure to see clearly the (new) situation and react fully to it.
Reading Bostic’s quote again “It would be really bad if we started cutting rates and then had to turn around and raise them again” and you can hear him thinking of the decision in front of the Fed as mostly a Type 1 one-way door, or at least a door he really hopes is one-way. But while the Fed has often been repetitive in its decision-making, it also changes course, and perhaps it should be more comfortable changing course, and going faster, too.
Within these three actions, there are nuances: Fed statements are carefully scrutinized for further guidance, and through unconventional monetary policy like quantitative easing the Fed can use a broader toolkit. Still, viewing Fed monetary policy actions through a {tighten, ease, no change} triumvirate captures their essential action set.
Data: uses FRED’s federal funds effective rate, supplemented with Wu-Xia’s shadow federal funds data on the Atlanta Fed where it’s available. The shadow federal funds rate is an attempt to estimate the stance of monetary policy from unconventional monetary policy actions and is especially useful in capturing the impact of those actions when interest rates are at the zero bound. Consequently, in the first chart here I flip to using the shadow rate whenever the Fed Funds rate is <= 1% and when there’s shadow rate data available (as there is from 1990 on).
Calculations: If the rate decreases from the prior month, it’s a cut; if it increases, it’s a hike; if it’s the same, it’s ignored. Any time there’s a hike, we can then calculate whether the next action turned out to be a hike or a cut, and we can do likewise for every cut. Above links take you to a view only version of the Google sheet and calculations.
The Fed does not only care about inflation. It also cares about the stability of the banking system, and the greater financial system as a whole. Sudden changes to interest rates can be extremely destabilizing to banks. Witness the bank failures as a result of the quite gradual interest rate hike recently, which gutted the value of long-term debt on their books.
Any long-term contract or instrument that has exposure to interest rates is strongly affected by the actions of the Fed, and sudden unpredictable moves would make them much riskier. Since this class includes systemically important stuff like mortgages, the Fed is forced to go slowly and predictably. So, while it may be true that from a pure inflation fighting perspective quick moves and rapid changes in direction would be more effective, from a financial stability perspective they are terrible.
Because of this, I don't think the Fed's tendency to repeat itself is some form of mental bias. It's very careful and deliberate policy, and arguably the correct policy.