Central bank independence talk.

This is the 10 minute version of my longer post on central bank independence, for the Peterson Institute Conference on Central Bank Independence Oct 31

***

Fed independence is back in the news, with President Trump’s pressure for lower interest rates and to fire Fed governors.

Unclutch your pearls. Fed independence is not an absolute virtue. We cannot have a completely independent agency that prints money to use as it sees fit.

Fed independence is limited and constrained by accountability—the cycle of appointments and reporting to Congress—a limited mandate, and limited tools. The mandate, including “price stability” and “employment,” implies “and nothing else.” No matter how important climate change or reshoring manufacturing may seem, the Fed must ignore them. Though printing money and handing it out can powerfully stimulate the economy, and confiscating money can stop inflation cold, taxing, spending, and economic regulation must be the province of politically accountable parts of the government.

Let us admit that there is, plausibly, a problem. The Fed has greatly expanded its activities, stepping into fiscal and political territory. Quantitative Easing, by shortening the maturity structure, is fiscal policy. Mortgage backed security purchases are credit allocation. Bailouts and market support transfer money from taxpayers to over-leveraged financial institutions. The Fed expanded unwittingly, in the press of events, but it never retrenched.

Financial regulation is stuck in too big to fail, and too powerful to issue equity. The Fed waded into political issues like climate. And the Fed has presided over a banking collapse, two big bailouts, the SVB fiasco, and 10% inflation. The Fed has only offered “it’s not our fault” excuses, not analysis or reforms. Excuses imply that the Fed is powerless to prevent a recurrence.

So there is an important question: How should the terms of limited independence be reformed? We face a choice: The Fed could remain expansive, powerful, and political, but face more direct direction by elected officials, as is the Treasury. Or, the Fed could return to a narrower scope of activities consistent with its current or even greater independence. I favor the latter course. But raise the drawbridge, hoist the independence flag, and defend the status quo is not going to be tenable for long.

Why does independence matter? The rationale is weaker than we often suppose. The limits matter more.

“Insulate the Fed from political pressure” most say. Don’t let the President goose the economy ahead of elections. But governments are tempted to hand out stimulus checks, tax exemptions, regulatory favors, loan forgiveness, and other goodies ahead of elections. Yet we do not entrust these policies to independent agencies. Moreover, “political pressure” is the same as “democratic accountability” when your party is in power. If independent central bankers run amok, our elected representatives must have the power, eventually, to correct them. Independence must be limited, as it is.

Economists answer “time consistency” and “precommitment.” Precommitting to lower inflation in the future gives a better inflation-unemployment tradeoff now.

The interaction of fiscal with monetary policy is an even clearer precommitment problem. Precommitting to repay bondholders, rather than inflate away debts, makes the government more able to borrow. Historically, independent central banks have been instituted far more to precommit against debt monetization and financial repression than to forswear Phillips curve stimulus.

But other policies, such as capital taxes and sunk investments, have even clearer time-consistency and precommitment problems. Once investments are made, governments are tempted to tax capital, default on debt, or grab property, “just this once.” Knowing that, people will not build or invest. Yet we do not assign taxing and spending to an independent Treasury.

Independence alone is a weak precommitment mechanism. Independent central bankers with the government’s preferences will want to inflate every bit as much as the government does. Independence alone only works if the government appoints central bankers who are much more hawkish than the government. Such choices are not evident in appointments or Senate confirmations, to put it mildly.

Limitations on the Fed’s ability to inflate are in economic analysis more effective precommitments. Kydland and Prescott’s (1977) famous article was titled “Rules Rather Than Discretion,” not “Appoint an Independent Fed and Let it Loose.”

Independence is an important part of the package, because monetary policy cannot be reduced to rules. The government can say where to go, but not how to get there. But the constraints on Fed action remain the meat of precommitment.

Resisting fiscal inflation is going to be the issue going forward, more than Phillips curve temptations. At 100% debt to GDP, each point of real interest rate is one percent of GDP additional deficit. President Trump is already calling for lower interest costs on the debt, as after WWII. Will independence save us? No. In 2020, the Fed quickly monetized $5 trillion of new debt,^[1] and kept interest rates low for a full year. The Fed did not intend inflation, but it chose policies that over and over have had that effect. The Fed cooperated voluntarily and enthusiastically with the Administration.

Resisting fiscal inflation is much harder than resisting Phillips-curve temptations. The decision to inflate vs. enact spending cuts or tax hikes, or let a crisis burn, is intensely political. It’s going to require a much stronger anti-inflation mandate from Congress, clearly signaling the elected branches’ desire. That mandate has to include Congress’ commitment to repay debts without inflation. Yet sometimes inflation is the least bad option. We would not want to lose WWII on the altar of price stability. Congress must stand ready to declare and take responsibility for the rare exceptions.

When the doctrine of independence developed, monetary policy had few fiscal implications, so it could be technocratic and a bit a-political. That’s not our world any more.

Crisis intervention poses a similar precommitment problem. Bailouts are the only way to stop a run. Yet if the government predictably bails out creditors, then people have little incentive to avoid risk, and the government is forced to intervene more often.

Again, independence alone won’t work. The Fed will not voluntarily abstain from intervening in a crisis just to contain moral hazard next time around. People will only believe the Fed (and Treasury) won’t intervene if they can’t intervene, and don’t need to intervene. And once a large enough crisis has started, not intervening is unwise. 1933 was pretty bad.

Again, rules and mandates have to be the centerpiece. Clean up (at last) the over-leveraged financial system so the Fed doesn’t need to intervene so often. Add serious limits so that the Fed cannot intervene, without explicit congressional declaration that the rules are suspended. Independence still helps: The Administration is likely to want to bail out too often.

We should focus our analytical and historical efforts on mandates and limits: Gold, with good and bad parts; the mandate to hold down interest costs, after WW II and coming soon. What does “price stability” mean? Was it wise for inflation targeters to drop employment and add numerical accountability? A financial stability mandate is growing. What does it mean? How is it limited? What limits on the Fed’s tools should there be? Independence is only icing on the cake.

Bottom line: I think the Fed should remain about as independent as it is. But independence must be accompanied by a more explicit limited mandate, limited tools, and stronger accountability. Only this package will address the Fed’s expansion into fiscal and political territory, while maintaining and enhancing the precommitments needed for successful monetary policy and financial regulation. “Congress is dysfunctional,” you may say. But especially in this room, we should not rush heedlessly to Autocracy or Aristocracy in place of checks and balances.

Update:

Conference videos here.

The conference was great. If you have time, watch the videos. I especially liked Tom Hoenig’s talk at the very end, which was an eloquent summary of many ways that the Fed went wrong. Dan Tarullo on Stable coins and the follow up also sparked some commentary. If stable coins can count uninsured deposits as assets, can the Fed avoid bailing out stable coins or uninsured deposits?

I found the legal panel with Carola Binder, Lev Menand, Paul Tucker, and Stephen Vladeck particularly informative because I don’t know enough about law. Stepping back from the Fed to Humphrey’s Executor issues in general, I was interested to learn a bit how the SEC FTC and related agencies work. They have slow turnover with dedicated slots to the party out of power, and a tradition that the Chair resigns. They were reportedly not as partisan back in the day, but now are quickly reversing themselves with each change of administration/congress.

My question, which I still wrestle with, is the right amount of check and balance for an administrative agency, which after all makes laws (regulations have force of law), investigates, prosecutes, judges, and penalizes all under one roof. Such agencies if totally indpendent would naturally tend to get captured, expand, or run off the rails, as Republicans now seem to feel many did. There has to be some mechanism by which the voters can throw the rascals out. On the other hand, they should not be run by simple majority and election losers lose everything; burn one rule book and institute another over night. The key to democracy is you have to be able to lose elections, and big changes only happen when opinions solidify. The independent agency with Senate confirmation and slow-ish turnover of a board, is a reasonable if not perfect device to give some check and balance to administrative agencies. It’s arguably not in the Constitution, but the Founders never expected administrative agencies. I’m interested in what works first and legalisms second. The panelists reported that the agencies used to be more continuous and technocratic and are now more partisan and becoming Tit for Tat. They did not opine that agencies should turn over more slowly, as the Fed does, but that seems an interesting implication.

At the end of the day, we’re talking about just how fast board members turn over with a change of power, which is hard to get all that excited about. As Larry Summers emphasized, there are a hundred other things going wrong (though we might not agree on which ones!) I was interested that most people felt the current arrangements about optimal. Maybe that’s just lack of imagination.

---

^[1] https://fred.stlouisfed.org/series/RESPPANWW