The r-star discussion, which is to say the debate about where interest rates should be set over the longer run in order to balance the economy at full employment and stable price growth, doesn’t matter so much for Fed policy today.
So said Jerome Powell, who participated in an event hosted by Stanford on Wednesday.
I don’t think I agree. I think the location of neutral matters quite a lot for policy today. And tomorrow. And next month. And next year. It matters all the time, precisely because it’s the crux of the matter.
Powell’s comments on the equilibrium rate were meant as a reminder that the Fed’s job is to help manage the economy “through the business cycle.” The evolution of the economy’s overall, long run potential growth rate is determined by factors like productivity, immigration, trade policy and industrial policy, none of which are the Fed’s purview.
“The question of what will be the equilibrium rate going forward doesn’t really matter for policy today,” Powell ventured. “We’re really asking about once the pandemic is well and truly behind us, and we’re well into the AI investment boom and the effects of AI — what will that look like?” he went on. “It doesn’t really matter that much for getting inflation down to 2%, while keeping the economy growing and the labor market strong.”
He didn’t mean it as a dodge. Scoff as you will, but Powell’s infallibly earnest. The problem’s always the same, though. The contention (which Powell reiterated Wednesday), that the Fed doesn’t need to take a stance on the neutral rate to conduct near-term monetary policy is nonsensical. To say policy’s “restrictive” (or that it needs to be) is to assume something about the neutral rate. “Restrictive” has no meaning outside of the neutral rate discussion: “Restrictive” relative to what?
So, when the Fed asserts (as Powell routinely does) that policy’s “restrictive” (despite copious evidence to the contrary), that’s an implicit assertion that policymakers think they know where neutral is.
Of course, neutral isn’t something that can be observed. It has to be inferred or modeled. Inference is better because it relies on common sense, not overly complex econometrics: If the economy’s not decelerating for a given level of rates, then that level isn’t restrictive. Just ask Powell, who said as much in October.
But let’s say you don’t like common sense. Maybe you’re someone who prefers econometrics. Well, as discussed here last month, the New York Fed’s own models suggest Fed policy only became restrictive in Q4 of 2023.
In the March 9 weekly, I asked what it would mean if Larry Summers was correct to suggest the nominal neutral rate’s nearer to 4% than 2%. Because this issue’s central to the macro narrative, I want to share the thought experiment from that weekly with the whole audience. (If you don’t get the weeklies, I encourage you to upgrade such that you do.)
If you assume neutral’s 4.25% instead of 2.5% (the longer run dot shifted up a token 10bps to 2.6% in the March dot plot, but we can ignore that for this simple exercise), that’d mean real neutral’s 2.25% instead of the 0.73% HLW estimate from Q4. As I put it four weeks ago, the implications of that scenario for the amount of implied policy restriction are serious.
The figure above shows the effective funds rate deflated by headline PCE measured against the HLW r-star estimate and against a theoretical 2.25% real neutral rate. If nominal neutral sports a four-handle, as Summers suggested, the real funds rate’s barely restrictive today, if it’s restrictive at all.
If that’s closer to “right” than Powell’s “meaningfully restrictive” story (and I think it is), a Fed which refuses to acknowledge as much is a Fed that’ll end up letting the economy run hot in perpetuity with unknowable side effects and consequences.
It’s possible (indeed, it’s highly likely) that those side effects and consequences will alter the economy’s long run potential, which means Powell’s incorrect to suggest the Fed’s actions today are somehow distinct from the bigger picture.
The story of the US (indeed, of the global) economy isn’t a patchwork of discontinuous chapters. It’s a running narrative, and the Fed’s actions within one cycle absolutely matter for that narrative, which is to say for the next cycle, and for the one after that and so on.
Every boom-bust cycle in history can, in some respects, be traced to the price of money. And busts, if they’re large enough, can alter an economy’s potential and thereby the neutral rate. For example, Fed policy played a role in inflating the subprime bubble, the bursting of which presaged a structural shift in estimates of r-star.
Powell on Wednesday stuck to the script on rate cuts in 2024. “At some point this year,” it’ll probably be appropriate to remove some policy restriction, he said, adding the usual caveat: “We do not expect that it will be appropriate to lower our policy rate until we have greater confidence that inflation is moving sustainably down toward 2%.”
If the Fed’s underestimating neutral, they’re overestimating how restrictive their current policy settings are. That could mean the confidence they seek will be a long time coming.
Great, succinct analysis. Reminds me of your work from a few years ago (referencing Kocic) about boom-bust, leverage and volatility cycles, tail risk, etc. Not much talk lately about leverage but I doubt that’s b/c no one is playing a version of LTCM 2.0 (or 3.0, etc.).
The “unknowable side effects and consequences” tend to appear when the Fed gets complacent. Powell brushing aside concerns about the neutral rate that defines “restrictive” and “easy” financial conditions qualifies as complacent, but who knows when/if it will have any market impact. Commodities are definitely catching a bid…
I believe Powell’s final quote is key. I believe he and the majority of the Fed are convinced that current rates are very restrictive for major segments of the economy and the majority of Americans, while the remaining segments and higher income folks are just fine and dandy, and generally untouched / unaffected by said rates, … and will disproportionally benefit and thrive even more in rate hike scenarios…thus the current conundrum… there were 10 dots for 3 cuts, 9 dots for two cuts…given recent movements of oil / gas prices, gold, silver, copper, crypto, world war, etc., it seems pretty hard to make a compelling case for any cuts in the foreseeable future if Mr. Powell is counting on inflation continuing to fall, given these circumstances … like Steve McGarrett used to say…”it doesn’t wash, Dano…”
US policy settings aren’t restrictive to the US economy (to the ‘have’ big tech, to be specific) yes, but they surely are restrictive to the global economy (i.e. the rest of the world). You can count the rest-of-world economies and businesses in the ‘have not’ category. Given the current state of the world, I can understand why Powell put the neutral rate so low on the list of things that actually matter. Scoff as you will, but Powell’s also infallibly earnest when it comes to the Have-Nots.
Sometimes we do forget that the global economy isn’t synonymous with the US economy. There are 192 other economies that look up to the US for policy direction. Sure, we should all be autonomous, but it’s simply not possible to break away from the US, which owns the currency of the world. So unless Powell wants to say f*ck you to the rest of the world, he has to take them into account too when determining whether US policy settings are restrictive or not. Whether that’s common sense or bullshit depends on who you ask. I believe it’s common sense. Even if it’s bullshit (a wise man said that it’s all bullshit). I’ll take Powell’s bullshit over Larry Summers’s bullshit everyday of the week. You can count Saturdays and Sundays too.
Over the years I was hired as an economic/financial expert to testify in court about monetary values in various legal matters. To do that required the calculation of various present values. To do this type of analysis one either needs to estimate how inflation might affect various future cash flows and use a nominal interest rate to value those flows or to eliminate these problematic numbers, estimate them without considering inflation and apply a basic real interest rate. I chose that approach because it eliminated all arguments about inflation which can’t be effectively forecast. The real rate I employed, and the one most consistent with history, was 2.5%. No opposing council ever questioned this number in more than 40 cases in which I testified.
Perhaps Powell’s responsibility in this context is to speak clearly and explicitly, as you did when honoring your oath.
Is it plausible that the neutral rate in the (AI-impacted?) future may be quite different from the neutral rate today? If so, hunting for the “long term” r-star could have less relevance to tactical decisions right now. In general, these comments seem consistent with FOMC being data-dependent not model-directed. The Fed’s experience with models during the pandemic may have been less than confidence-inspiring.
I think you are spot on about the Fed questioning the reliability of their models. Exhibit One being the impact of rates on housing supply. If you are right, that is refreshing.
But do you think that AI will meaningfully bring down inflation? This is America! Won’t the cost savings via staff “rationalization” be passed along to shareholders rather than consumers? The experience from the advent of the internet supports the notion that the impact will be patchy rather than widespread. Some goods (esp tech gear) prices fell, but service prices did not fall as much, if at all.
But what do I know?
I have a vague notion that it may depend on where the cost savings land in the income statement. If AI reduces COGS, that may have a different effect on pricing behavior than if AI reduces opex (R&D + SG&A). This is not something I’ve thought through though (yes, I typed that combination just because it looks funny).
Following up, here is my vague notion. Suppose your company is happily operating at the product price and volume that maximizes gross profit – yes, this is a classic micro-economic equilibrium hypothetical – and then along comes AI.
If AI reduces your unit COGS by say 1%, that may motivate you to test the elasticity of your end market by lowering price in hopes of higher volume and gross profit even higher than you’d get by simply pocketing the lower COGS.
If AI reduces your opex by say 5%, why would that motivate you to lower price? Your unit economics are the same, the price-COGS-elasticity-gross profit dynamics are unchanged, you’re already at maximal gross profit, so lowering price will just hurt operating profit.
Thanks JL. That’s the big unknown. The impact will probably differ from industry to industry, no? Companies with a monopoly/monopsony position will probably be inclined just to pocket the extra OP and use it to fund more buybacks. Companies in the unfortunate position of being in a more competitive sector may be forced to pass some of the benefit on to consumers if a competitor does so. We’re still in the first inning or pre-game warmups!
I think of most LLM AI applications as R&D and SG&A, but that may reflect my ignorance or lack of imagination. Anyway, right or wrong, that makes me skeptical that AI if successful will be broadly disinflationary.