Kevin Warsh’s call for fresh thinking at the Fed runs up against a deeper problem. The tools and concepts guiding monetary policy may no longer fit an economy reshaped by AI, shifting prices, and the fading power of interest rates.
Federal Reserve Chair Kevin Warsh has been derided as being everything from “Trump’s sock puppet” to “utterly conventional,” yet at Jackson Hole he delivered an inaugural speech that was thoughtful, amusing in parts, utterly un-Trumpish in style. His problem is that he seems to be caught in the dark web of mainstream monetary thought. Perhaps sensing as much, Warsh admirably called for the Fed to “receive the full range of ideas” on monetary policy in order to “create more reliable models.” Let’s hope he means it.
Forward Guidance
On the most controversial point in his tenure so far, Warsh made a clear case for less “forward guidance”, on the correct ground that it tends to tie the Fed’s hands, and that it sets up a feedback loop from the Fed to the markets, fostering a trading climate based on Fed statements rather than economic conditions. Fewer statements, presumably less fuel for that fire. It’s worth a try, though speculators will still bet on the policy decision, and perhaps more aggressively given the increased uncertainty about what it may be. The benefit, if there is one, will be that “analysts” who have devoted themselves to reading Fed statements like tarot cards, may have to find some other uses for their time.
Setting Interest Rates
On interest rates, Warsh dispelled a lot of nonsense with splendid candor: “We determine the path of short-term interest rates.” A brass plaque with those words might be sent to Ken Rogoff, Paul Krugman, Ricardo Caballero, and to all the Washington think tanks and lobby groups that devote themselves to prattling about the federal budget deficit. Warsh’s simple truth is a spear through the heart of the “loanable funds” doctrine, as well as its corollary, “crowding out.” The short-term interest rate is (and has been, for a century) what the Federal Reserve wants it to be.
Warsh might have added that in setting the path of short-term rates, the Fed is also the driving force, most of the time, behind long rates as well. The reason is arbitrage: long-dated bonds are a substitute for a sequence of notes and bills; the rate on long bonds combines the present short rate with the expected future path of short rates; other factors are mostly secondary. When the Fed keeps short rates low for a long time, markets become accustomed to it and the long rates follow the short rates down. Rising long-term rates – the recent global “bond rout” – reflect an expectation that short rates will be raised. It’s not complicated, and has nothing to do with CBO’s debt projections, which exist for legislative reasons and not because any financial player takes them seriously. As Robin Brooks of the Brookings Institution has written for Substack, there is no sign of a US fiscal risk premium.
There is a widespread view, shared by Brooks, Claudia Sahm, and many others, that the Federal Reserve will (and perhaps should) raise interest rates in order to “stabilize expectations” and to “show independence” from Donald Trump. The expectations argument is mystical, and almost surely wrong: one-and-done is rare in Fed funds history. Once embarked on a course of rising rates, twenty-five or fifty basis points at a time, many speculators will anticipate further increases – “squeezing the pips until they squeak”. As for independence, to raise rates merely to defy the President does not show independence; it shows insecurity, that monetary policy is being set by political appearances – exactly the wrong message for a new Chair to send.
The Dual Mandate and the Two Percent Target
Warsh acknowledged the dual mandate – full employment and “reasonable” (a word he omitted) price stability as specified in the 1978 Humphrey-Hawkins Full Employment and Balanced Growth Act. He made no mention of the obligation to account to Congress for what it does. Perhaps Congress should remind him of what the law said:
SEC. 108. (a) Section 2A of the Federal Reserve Act is amended by striking out the second and third sentences and inserting in lieu thereof the following: “In furtherance of the purposes of the Full Employment and Balanced Growth Act of 1978, the Board of Governors of the Federal Reserve System shall transmit to the Congress, not later than February 20 and July 20 of each year, independent written reports setting forth (1) a review and analysis of recent developments affecting economic trends in the Nation; (2) the objectives and plans of the Board of Governors and the Federal Open Market Committee with respect to the ranges of growth or diminution of the monetary and credit aggregates for the calendar year during which the report is transmitted, taking account of past and prospective developments in employment, unemployment, production, investment, real income, productivity, international trade and payments, and prices; and (3) the relationship of the aforesaid objectives and plans to the short-term goals set forth in the most recent Economic Report of the President pursuant to section 3(a)(2)(A) of the Employment Act of 1946 and to any short-term goals approved by the Congress.”
The law is a reporting requirement. It is a requirement to report to Congress on the conduct of monetary policy, consistent with the short-term goals set out in the Act. Those goals were four percent unemployment and three percent inflation, and they applied to the entire government, not to the Federal Reserve alone. The Fed has no specific inflation mandate. Further, as Professor Stephanie Kelton reminds me, the inflation goal was subordinate to the employment goal.
Provided, That policies and programs for reducing the rate of inflation shall be designed so as not to impede achievement of the goals and timetables specified in clause (1) of this subsection for the reduction of unemployment.
In other words, Congress prohibited the use of policies designed to slow the economy or raise unemployment in the pursuit of “reasonably” stable prices, which it defined ad interim as three percent inflation as measured by the consumer price index.
Warsh therefore mis-spoke, gravely, when describing the much-noted two percent inflation target as “our mandate.” It is no such thing, not part of any statute. It’s an internal goal, which originated in New Zealand and spread during the “inflation targeting” fad of the late 1990s. Further – here I’m again in debt to Professor Kelton, who noticed a first-hand account by Larry Summers, that in 2012 Chairman Bernanke sought, and received, President Obama’s “permission” to adopt the two-percent goal. A more blatant act of non-independence and contempt for the plain letter of the laws is hard to imagine.
Nevertheless, the two-percent target was the hard core of Warsh’s speech. He declared an unswerving commitment to this goal: “our charge to keep”. It is also the weakest link in his world-view, for reasons that require an excursion into what he said, and didn’t say, about the “new economy.”
The Two-percent Target in the Age of AI
Warsh’s remarks begin with a discussion of innovation, investment, and (of course) artificial intelligence, noting that the once-fashionable views of Summers (secular stagnation), Robert J. Gordon (technological slowdown) and Ben Bernanke (the “global savings glut”) no longer apply. He does not pause to consider that perhaps they never did. Bernanke’s notion, in particular, that somehow Chinese savers (living behind capital controls) were driving down US interest rates seems especially silly in retrospect. Did they stop? No. It might have been rude to point out that all of these ideas (and many more), are bound up in a series of old ideological mistakes.
So how does Warsh think of AI? The answer may be found in his language. He describes AI as (possibly) “a new factor of production” yielding “potential for substantially higher growth” and a “sustained rise in productivity.” The mundane phrasing roots Warsh’s thinking in the frame of a textbook “production function,” based on the interaction of “capital” and “labor,” where advancing technology always enhances economic “output” and accelerates growth. Productivity is then usually defined as the ratio of output to labor input. This framing is so standard for economists of Warsh’s generation (and even more, of mine) that it generally passes beneath comment.
Yet, as a description of modern capitalism and especially of the information economy, the construct is categorically false. Technology is creative destruction and creative destruction destroys. New technologies drive older technologies to the wall. They work by rendering the functions served by previously scarce and valuable goods so abundant and cheap that they decline in economic importance. Eventually, the old goods (and associated services) disappear almost altogether from economic data. Information technology has this effect in spades, as anyone who remembers all the expensive communications paraphernalia of the analog age knows.
In an earlier era, when cars, roads, oil fields, refineries, gas stations and repair shops displaced grass-fed horses and coal-fed trains, technology also worked (with government help) to bring new economic activity into being, expanding the market and accelerating “growth”. But the present wave works in the opposite direction. It tends to erase market activity – to take paid services out of the market, by making communications and analysis fast and largely free of marginal cost. To speak of AI in terms of “productivity growth” thus betrays an obsolete mode of thought, namely the neoclassical growth dogmas of the 1950s. Creative destruction is also now trans-national. Warsh made no mention of the possibility that superior and cheaper Chinese AI may wreak it on the American versions. This is a thought that cannot be thought in terms of “production functions” – but it may happen in real life.
How then does AI affect the conceptual foundations of inflation? Paul Krugman takes the collapse of communications prices as a sign that we’ve all become rich, as though we think in terms of past spending patterns. In that case, price levels should be marked down, and real incomes should be marked up. Yet almost nobody feels that way; our budgets have moved on, to those things we still have to pay for (like food and electricity and health insurance and rent) at increasing prices. Ultimately, “inflation” as measured will reflect the prices of all that remains in the index, a basket constantly shifting, among the less-well-favored, toward essentials like food, fuel and housing and health care, precisely because they are becoming more expensive.
It is possible that with (as Warsh states) “moderate” wage settlements – meaning that workers have very little power – will translate to a declining measured average inflation rate over time, as the index lags shifts in household spending patterns toward sectors with rising prices. But there are strong opposing forces: profits are high (a fact Warsh records as positive, although profits are driven by rising prices), AI investment is energy-intensive, competing up the cost of power, and US policy seems determined to wreck the oil trade, a topic Warsh did not mention. These forces will percolate through the price structure for the near and foreseeable future, making a two-percent average elusive even if wages stay low and stable.
Further, differences in the consumption basket between different types and classes of household make the average inflation rate a remote abstraction. What matters to each household is the prices of the items they happen to consume. Existing housing is a special problem: for renters it’s a rising cost, but for many American households it’s an asset, which they own. Actual rents are a big burden; the rising “imputed rent” that homeowners “pay” to themselves isn’t – it’s an increase in wealth, like an increase in the Dow Jones. Are lower prices for existing houses per se desirable? If so, why not also target lower prices for corporate stocks?
In a world of structural shifts, diverse and changing spending patterns and rapid innovation, the two-percent “consumption expenditure” target is to a degree, merely arbitrary. It is also definitely hard to hit, and it is not a problem AI will solve. It is, in short, as outdated or as meaningless (take your pick) – as Bernanke’s “savings glut” or Summers’ “secular stagnation.” Since it’s not rooted in law, the Fed can choose to drop it, as a mistake that should not have been made.
Higher Interest Rates Have Lost Their Traction Against “Inflation”
What is certain, is that a higher “path of short-term interest rates” will not bring inflation down. In Paul Volcker’s day, the consumer economy and business investment ran on limited credit. When the Fed raised interest rates to twenty percent in 1981, manufacturing employment, investment, and wages and purchasing power collapsed, and the dollar soared, making imports cheap. But manufacturing is no longer the key sector, and unions no longer drive wage bargains. There is no link from higher rates to slower growth, higher unemployment, weaker wage settlements and slower price increases.
Today, the federal debt is about four times larger, relative to GDP, as it was then. Raising rates thus pumps money into economic activity, which is mostly services and not interest-elastic, and into assets, where price increases bleed into the PCE. Also, if housing construction falls, rental rates go up. And if consumers cut back on cars and durables, a larger share of the decline falls on imports, mitigating the effect on domestic output. Jerome Powell raised the Federal Funds rate by over 500 basis points. The economy did not slow down, and there was no effect of the policy, direct or otherwise, on inflation – which was largely driven up by supply issues and oil prices, and then faded (to a degree) on its own.
Thus the one tool that the Fed controls – the “path of short-term interest rates” – is stuck in a sandpit, from which Kevin Warsh cannot dig it out. Neither tough words nor higher rates will have any serious effect on the rate of change of the price level. Entirely different measures, including peace with Iran, detente with China and Russia, and strategic controls over prices and profits, would be required. These are not within the Federal Reserve’s authority.
Decisions Based on “Data”: A Commitment to Short-termism
Warsh has been heavily praised by press and pundits for his firm stand on the two-percent fantasm – excuse me, goal. This is generally taken to mean that interest rates will rise, sooner or later. Warsh says that exactly when will depend on “the data.” Unpacking this point brings us to an unpleasant question: What data?
If decisions have not yet been taken, the only data that can influence them, relative to what is now known, is the monthly or quarterly economic news. By committing himself to “the data”, Warsh has locked the Fed into an extreme short-termism, based on the news. But preliminary economic data are by their nature ephemeral. Revisions often tell a different story. And even if accurate, short-term data can be transient; what pops up one month can go back down the next.
This is the opposite of how sound policy should be made. It is not how either Paul Volcker or Alan Greenspan operated, little though I approved of either one. Volcker set out to restructure the whole economy by breaking the synchronized and interdependent wage bargains known as the “Treaty of Detroit” and thereby trade union power. (He did this by destroying the industries for which the unionized workers worked.) Greenspan kept his powder dry and did not choke off the Internet boom as unemployment fell, despite the NAIRU bunkum prevalent in the 1990s. Both resisted the news, based on their larger objectives, their structural analysis and their political convictions.
Warsh has declared a similar conviction about his goal. He has staked his reputation on it. He lacks the means to achieve it. New “evidence” will not help.
Conclusion: What Should He Do?
So, what should Chairman Warsh now do? He cannot cut interest rates; President Trump has foreclosed that possibility, merely by wanting it to happen. But raising rates, especially if based on the “news”, is pointless and counterproductive.
So there’s really only one proper course of action: Do nothing. The situation, as Warsh said, is new and uncertain. Wait and see. Show spine by resisting pressure from every angle. And meanwhile, seek out the new ideas, the sharp critiques, and the original minds who might kick the Federal Reserve out of its ideological rut. It might again become – as it was in the 1930s and 1940s – an exciting place where unorthodox thought can flourish.
Well, one can dream.
