Central Bank Independence

Fed independence is in the news again, with the on and off again disagreements between President Trump and Chairman Powell. Many economists quickly complain about a terrible threat to Fed Independence. But Fed independence is not absolute, written in stone. We do not appoint a central bank czar for life, and say “do what you want.” We certainly would not want complete independence of every other agency, from ATF to VA with SEC, EPA, FTC, HHS in between. Why is the Fed special? Is the Fed too independent? Or is it not independent enough?

Independence

The Fed is a government agency, ordained by Congress, for the government’s purposes, representing the people. The law and tradition of Fed independence acts as a precommitment. Congress knows that it and the President will be tempted to goose the economy, especially just before elections. They will also be tempted to use the printing press and lending power to subsidize politically important industries and constituencies. They understand that these are not good ideas in the long run. So, like a dieter who doesn’t keep chocolate in the house, Congress gave the Fed a good deal of independence. (This is a common, but superficial and to some extent wrong argument for Fed independence. More detail later, but let’s stick with this one for now.)

But independence in a democracy comes with a sharply limited mandate. The mandate to pay attention to price stability and employment means that the Fed should not pay attention to anything else.

Congress also gave the Fed limited tools. The Fed can only buy and sell securities and set interest rates. The Fed cannot directly print money and send it to people or businesses, nor can it confiscate money. Doing so is far more powerful for controlling inflation than moving overnight interest rates, but only a politically accountable agency can tax or spend. Similarly, labor taxes, labor regulations, and the disincentives of social programs have far more effect on employment than the overnight federal funds rate, but the Fed cannot touch them. Even within its inflation and employment mandate, the Fed is forbidden the most powerful tools.

Given the sharply limited power of the Fed’s tools, it’s strange that we assign it so much prominence for inflation and employment. Perhaps the advantages of independence outweigh the advantages of strong tools. Or perhaps, decades of Keynesian economics and Phillips curves have led us to forget the dominant microeconomic forces that drive employment, and decades of monetarism and new-Keynesian economics led us to overestimate the Fed’s ability to control it with limited tools.

The Fed also does not enjoy absolute legal independence. The chair and board members are appointed by the President, with Senate confirmation, for limited terms. The chair is reappointed every four years. The Fed must regularly report to Congress. Informal political pressure is stronger.

In all these ways, the Fed is just one point on a spectrum that includes many “independent” administrative agencies. One compilation lists 56. The CFTC, CFPB, SEC are other independent financial regulators. They have specific and limited mandates, some insulation from the immediate will of the current administration, limited tools, requirements such as the Administrative Procedures Act for rule making. But they eventually serve the pleasure of elected representatives.

By these means the government as a whole ties itself to the mast of better long-run policy. But it does not tie itself so tightly as to go down with the ship should it begin to sink.

Central banks can be too independent. If a central banker decides he or she wants to save the world from the climate crisis, remedy inequality, subsidize dying manufacturing businesses, de-bank people and businesses in disfavored industries or with unpopular political opinions, or get too cozy with Wall Street via bailouts and anti-competitive regulatory barriers, the accountable parts of government must be able to put a stop to it.

The veneration of independence also has intellectual roots in an era when people distrusted “politicians” despite their democratic accountability, and instead trusted disinterested technocrats. That vision is fading given the rampant failures of the last few decades. Faith in politicians is no stronger, but technocrats are not necessarily well intentioned, a-political, or competent either. Not every central banker even dislikes inflation. Independent technocratic agencies can descend into politicized bubbles. There is a contrary political momentum towards accountability to elected officials, who despite their warts can be tossed out.

An independent central bank is also useful to a President. No matter what the Fed does, someone will complain. If the Fed must raise rates to contain inflation, the President can say “it’s not my fault, it’s the darn independent Fed.”

The Fed’s current independence was born in the 1951 Treasury-Fed accord. From 1942, the Fed had been charged to keep the long run treasury rate below 2.5%, in order to help finance the war, and then to keep down interest costs on the debt. The accord allowed the Fed the independence to raise interest rates in order to fight inflation, a policy Chairman Martin called “leaning against the wind,” and which we would now call a Taylor rule (or Wicksellian policy). Presidents Johnson and Nixon’s pressure on the Fed is widely credited for sparking inflation in the late 1960s and early 1970s, and is the central story passed around in defense of independence when presidents want lower interest rates.

More or less independent central banks have spread around the world. A large economic literature finds they are associated with lower inflation and better overall macroeconomic performance. Like all correlations one can quibble. Governments that want to inflate don’t allow independent central banks. Governments that want fiscal probity, microeconomic reform, and stable inflation do. But I don’t seriously challenge that empirical finding.

So Fed independence, like the somewhat more restricted independence of other agencies, is a part of the general checks and balances that so well serve the US constitutional order. The Fed responds to politics, but slowly.

In sum, the hue and cry over some sacrosanct “Fed independence” is a bit silly. The Fed only has limited independence, and only should have limited independence. Whether it is too independent or not independent enough are questions one can argue. Also, the right independence for monetary policy (raising and lowering interest rates) may be different than that for asset purchases, bailouts, and other market interventions, and may be different still from that for financial regulation. For example, Andrew Levin and Christina Skinner argue that the Fed is too independent, and that financial regulation should have the same sort of oversight as other kinds of regulation.

Too independent or not independent enough?

Is the Fed too independent, or not independent enough? In part, one’s answer to the question will always correspond to policy preferences. If one likes what the Fed is doing, it’s either independent enough or needs more independence. If not, and the President or Congress voice policies one would prefer, one comes to the opposite conclusion.

In the fairly recent past, I notice a general congruence between Fed and Administration and Congressional desires. Perhaps the Fed is not independent enough, and gives in to political pressure. Perhaps the Fed in full independence happens to come to just the same conclusions as the President and Congress, and when things go badly was just as wrong as them.

In the runup to the financial crisis, the Congress demanded that the Fed encourage lending to a lot of people who would later have trouble repaying their mortgages. The Fed went along willingly. The Fed cooperated with Treasury on the bailouts of the financial crisis, but only the tea-party types objected. (The Dodd-Frank act did rein in the Fed’s emergency authorities, suggesting a little ex-post regret.) The Fed cooperated with a coalition of State Attorneys General in 2008, in the “robosigning” affair, commanding banks to give money to community activist groups. The Fed went all-in with Treasury on the bailouts of 2020, buying up nearly all the new treasury issues, directly buying state and local government debt, financing astounding firehoses of money. The Fed could have invoked independence to object to any of this. It did not, and indeed egged the treasury on. (Bassetto and Sargent quote an un named Federal Reserve official as saying “The best way to defend the independence of a central bank is never to exercise it.”)

That has changed, of course, with President Trump’s desire for lower interest rates. Just why the President wants the same policy that arguably stoked Bidenflation, hurt the Trump base most of all, and arguably lost Harris the election, is a good question, but that’s what he wants. And the Fed does not.

Our Fed is not completely a-political. It does listen to what Administration and Congress want. As it should, adjusting slowly and methodically.

The ECB has much more formal and informal independence. Its President and executive board members serve for one 8 year term. They are appointed by the European Council, a much less powerful institution than the US President and Senate. Being instant lame ducks, not removable for 8 years, and facing a much weaker political overseer in the EU commission and parliament than the Fed faces in Administration, Congress, and Judiciary, they can basically do whatever they want.

The ECB has vastly stretched its mandate to include climate finance and sovereign debt support, stretching its price stability mandate to the breaking point to do so. But so far the rest of the EU seems perfectly happy with this expansion so there is not much tension that lesser independence would cure.

The deeper arguments for independence

My first arguments for independence were a bit superficial. Independence for the Fed is actually more subtle than just precommitment against political short-term ism. After all, the Treasury and Congress are tempted to hand out stimulus checks, business subsidies, tax exemptions, block foreign acquisitions (US Steel) or forgive a few hundred billions of student debts in order to win elections. Shouldn’t taxes and spending also be turned over to independent commissions of technocrats, untouchable by Congress and Administration?

Well, no! But just what is the difference? What’s so special about monetary policy?

There is an important political difference. Taxing removes people’s hard-earned money by force. Spending transfers money directly to grateful beneficiaries. Nudging interest rates this way and that can at least seem less directly political. Yes, there are lots of winners and losers — borrowers, savers, workers, are all affected by interest rates and inflation. But so far, that effect is indirect enough not to require the kind of democratic accountability of direct transfers. Bailouts are more contentious, so crisis management and financial regulation are on thinner ice than monetary policy.

But the biggest difference lies two aspects of in the nature of monetary policy.

The first big issue is “time consistency.” (Governor Chris Waller’s comments at the 2025 Hoover monetary policy conference are excellent on this topic.) This goes back to a famous 1977 paper by Finn Kydland and Ed Prescott, for which they deservedly won the Nobel prize.

In this story, the tradeoff between inflation and unemployment depends on how much inflation people expect. You can only get lower unemployment by inflation in excess of what people expect. So the Fed wants people to expect low inflation. But given expectations, the Fed will want to goose the economy with a little surprise inflation. People know that, so they expect higher inflation, and the Fed is faced with a bad tradeoff. The rules of the game must tie the Fed’s hands ex ante to not do things even the Fed will want to do ex post.

This is not a case of politics, or short-termism, or a conflict between Fed and Administration. Even the most far-sighted totally benevolent government will want to inflate ex post. A big part of the argument for rules and formal or informal traditions, is to make this commitment, from the gold standard to the money growth rule to the Taylor rule.

Independence is related. If the central bankers are selected to be a bit hawkish, also committed to rules, traditions, the reputation of the central bank, and so on, and insulated from the government’s ex-post desire to inflate, the government achieves the desired precommitment.

You can see though that the central bank only needs independence if it makes discretionary decisions. If it hewed tightly to a rule (with or without government pressure), it would not need independence, other than the independence to stick to the rule. But a central bank following a purely mechanical rule is not in the cards, and rightly so. Rules work if the game is stable. But there is a sense that each event lately has been unlike the others. The financial crisis was not ordinary recession. Neither was Covid and the subsequent fiscal blowout. Huge tariffs if they come will be different still. The zero bound meant that whoops, we can’t follow the old Taylor rule. The Fed mangled many of these events, but we are humans and addressing new events is hard. A fixed rule will mangle worse.

This time-consistency problem pervades government. Ex-post, even the most benevolent government wants to grab capital, impose rent controls, default or inflate away its debt, and then promise never to do it again. If it does, nobody invests in capital, builds apartments, or buys debt. So the government needs a way to commit ex ante not to do things it prefers ex post. Property rights, limited government, etc. are all about solving the time-consistency problem.

But Kydland and Prescott’s time-consistency hinges on the underlying view of monetary policy, that the inflation-output tradeoff is made worse by higher expected inflation. If the Phillips curve were static, as in the 1960 view, it would not matter. If the Phillips curve proves to be a Phillips cloud and not an exploitable tradeoff, then this time consistency issue vanishes. We’re back to the much weaker political self-control arguments that also apply to fiscal policy and other agencies.

The second big issue is a broader vision of the nature of monetary policy. Though the legal and institutional wisdom of setting up central banks with limited mandates, limited tools (often excluding treasury debt rather than mandating it, to stop deficit finance), and some degree of political independence goes back centuries, the current academic and intellectual case for independence stemmed from Milton Friedman’s great victory. Friedman persuaded the profession and later everyone else that central banks are the overwhelming masters of inflation. Friedman overturned the Keynesian consensus that inflation was a matter of “wage price spirals,” union wage setting rules, fiscal stimulus past the aggregate supply curve and so forth. Friedman said no, central banks control inflation by setting the money supply. The “new-Keynesian” views that currently dominate academia and central bank research likewise attribute inflation mainly to central bank actions. In the verbal tradition, there may be “shocks” that move inflation around temporarily, but central banks can swiftly offset them. This presumption that central banks control inflation is so strong that many articles on time consistency, independence, and related issues — such as Kydland and Prescott — simply assume that central banks directly control inflation!

But it’s no longer so obvious that central banks do tightly control inflation. In my view, they can have a negative effect on output by interfering with the banking and credit system, but the evidence that they can boost inflation with low interest rates in a functioning economy is much weaker. This is the “pushing on a string” or “money is oil” view of the macroeconomy — too little oil seizes the engine, but 8 quarts instead of 4 won’t make you spurt down the freeway. Whereas Friedman and Schwartz convincingly showed that central bank mistakes lay behind fluctuations up to 1960, the more recent literature struggles to explain much at all variance of inflation or output to central bank actions. Maybe they are just optimally running things, much as a well driven car tracks down the middle of the lane even in a windstorm. Maybe not.

Obviously fiscal theory of the price level and, to me, the 2021-2023 episode screams that fiscal policy also determines inflation. Even conventional views now chalk up much inflation to “supply shocks” or “demand shocks,” implicitly saying that the Fed couldn’t do much about it. The “nominal anchor” in conventional (non FTPL) macro is no longer the money supply, but a somewhat nebulous hope for “anchored” expectations. The Fed and ECB don't pretend to control actual inflation or the price level, hoping only to guide it back to a 2% forecast over a medium-term forecast horizon.

In sum, much of the contemporary case for independence relies on the worldview that the Fed tightly controls inflation, either by money supply or interest rates, and that it faces an exploitable inflation-unemployment tradeoff that gets worse with expected inflation. That world view seems a lot weaker than it was in Kydland and Prescott’s time.

Bottom line

I’ve left you with a good one-hand other-hand too-long economists’ essay. In my view, the US institutional framework is about right. The Fed faces just enough democratic accountability and enough institutional independence. It remains one of the most a-political and competent Washington agencies. Yes it bends to politics, but we are a democracy after all. When the Administration announced a “whole of government” climate crusade, the Powell Fed pushed back just as much as it should have, and accommodated also about what it should have. (This judgement is a political one not an endorsement of the policy.) I disagree with a lot of Fed decisions, but I don’t think more independence or less will help much. Checks and balances can only slow down the will of the voters. The ECB strikes me as a bit too independent, especially given the weakness of its political overseers. Fixed terms and a permanent bureaucracy can go off on “independent” projects far beyond what they should do alone.

We talk too much about the power of the President to influence or fire Fed officials, and raising and lowering short term interest rates. There are deeper unexamined questions in the rules of the game. Should the Fed’s mandate and tool limitation be changed? I would prefer a strict price level mandate. Should the Fed’s tools be expanded or limited? Should QE, yield curve control, crisis lending, asset price supports, etc. be limited by Congress? Should the Fed be forced to buy only Treasurys and not mortgage backed securities? Or should it have to buy other assets? Historically many central banks were forbidden to buy Treasury securities, to forestall inflationary finance like we saw in 2021-2023. Should it hold a large portfolio of assets at all, or create reserves as the ECB initially did by lending to banks against collateral? Should financcal regulation be separated? How should financial regulation be reformed? (Are we going to be living with Dodd-Frank rules for 100 years?) Should the Congress force the Fed to allow narrow banks to rise? What about crypto? Oh no, maybe we shouldn’t open this can of worms, at least in public view : )