One of the puzzles of the Great Recession is why the Fed did not move more aggressively to promote recovery in NGDP growth and/or inflation. Many observers, including Christina Romer and Lawrence Ball, noted that the Fed's passivity seemed to be a bit if a puzzle, given that Ben Bernanke had previously criticized the Bank of Japan on very similar grounds in a 1999 paper entitled:
During the recovery from the Great Recession, Bernanke seemed frustrated by the slow rate of growth in aggregate demand and the price level. And yet various QE programs were often ended before their objective had been achieved. When asked why the Fed did not do even more QE, Bernanke referred to vague "costs and risks" of an excessively large Fed balance sheet. At the time, I was quite dismissive of this argument, as the Fed's assets (T-bonds) are the Treasury's liability. Thus for the consolidated Federal government balance sheet, there is essentially no risk associated with a change in the price of T-bonds held by the Fed. And the Fed's mortgage backed securities issued by the GSEs had already been essentially guaranteed by the Treasury. (Joseph Gagnon discusses four possible risks, and is rightfully dismissive of all four.)
I recently came across evidence that, prior to the Great Recession, Bernanke held similar views on balance sheet. In a 2003 speech that was critical of BOJ passivity, Ben Bernanke dismissed the fear that QE could induce excessive balance sheet risk:
In short, one could make an economic case that the balance sheet of the central bank should be of marginal relevance at best to the determination of monetary policy. Rather than engage in what would probably be a heated and unproductive debate over the issue, however, I would propose instead that the Japanese government just fix the problem, thereby eliminating this concern from the BOJ's list of worries. There are many essentially costless ways to fix it. I am intrigued by a simple proposal that I understand has been suggested by the Japanese Business Federation, the Nippon Keidanren. Under this proposal the Ministry of Finance would convert the fixed interest rates of the Japanese government bonds held by the Bank of Japan into floating interest rates. This "bond conversion"--actually, a fixed-floating interest rate swap--would protect the capital position of the Bank of Japan from increases in long-term interest rates and remove much of the balance sheet risk associated with open-market operations in government securities. Moreover, the budgetary implications of this proposal would be essentially zero, since any increase in interest payments to the BOJ by the MOF arising from the bond conversion would be offset by an almost equal increase in the BOJ's payouts to the national treasury. The budgetary neutrality of the proposal is of course a consequence of the fact that, as a matter of arithmetic, any capital gains or losses in the value of government securities held by the BOJ are precisely offset by opposite changes in the net worth of the issuer of those securities, the government treasury.
The specific proposal discussed by Bernanke is less important than the final sentence, where he makes exactly the same point that I made above. The so-called costs and risks were never really about true risks to the economy; they were mostly about the risk of central bank embarrassment, of having to be bailed out by the Treasury.
This reminded me of a 2011 article by Vincent Reinhart, who worked with Ben Bernanke:
I was one of Bernanke's co-authors for an academic paper published in 2004 that did some of that criticizing. After seeing how other major central banks, including the Fed, handled similarly trying circumstances, I admit that Gov. Shirakawa has reason to feel aggrieved. In particular, the main point of contention, quantitative easing, is a policy that looks good on paper but has a flaw when implemented by a democratic central bank. . . .
Market participants have to be convinced that the central bank is committed to the policy for quantitative easing to be effective. If investors think the authorities will stop or reverse soon, then long-term yields will not move much nor will the extra reserves be used. Underappreciated in the theory (and the criticism) is that the monetary policies of major central banks, such as the BOJ, are decided by committees. Individual members do not usually see the world exactly in the same way. Because the balance of judgments may change over time, a decision at one meeting cannot presume the outcome of the democratic process at future committee meetings. As a consequence, policy statements tend to be hedged to foster compromise, making them tentative and undercutting the effectiveness of policies relying on a credible long-term commitment.
I believe these examples help us to understand why Bernanke seemed to hold back from his more aggressive recommendations for the BOJ, after joining the Fed in 2003. Before 2003, Bernanke believed that central banks had almost unlimited power to boost aggregate demand and prices, if they chose to do so. After joining the Fed, his perspective changed from what a central bank could do, to what a central banker could accomplish. This is especially important given that Bernanke moved the Fed away from Greenspan's dictatorial approach, towards a more collegial approach to decision-making.
In both of the cases discussed above the problems were not technical, they were political. The BOJ could have used forward guidance, if they were in fact committed to creating inflation. Unfortunately they were not. Every time (until 2013) the inflation rate rose up to zero (from negative territory) the BOJ raised interest rates with the intentional of preventing inflation from rising above zero. And the risks and costs of QE are entirely political; there is no actual risk to the consolidated balance sheet of the federal government.
Some of this seems to go against my longstanding view that policy decisions are better made by committee (or better yet markets), not individuals. Might we have been better off with Bernanke as dictator of monetary policy? Maybe, but of course there is also downside risk with dictators (Hitler, Mao, Stalin, Pol Pot, Kim, etc.) More to the point, Reinhart points out that the credibility problem can be overcome with a suitable policy rule:
Ill effects of this drawback of democracy can be tempered if the central bank follows a policy rule. The BOJ ultimately did so, with the promise made in 2001 to keep the policy rate at zero as long as the price level was declining. The Fed has not yet seen fit to do so.
As I just indicated, this was the wrong policy rule---one that put a zero percent ceiling on inflation. Similarly, the Fed had the wrong policy rule. As an academic, Bernanke had suggested that the BOJ engage in level targeting, promising to create inflation to offset the previous deflation. After joining the Fed, however, Bernanke found strong institutional resistance to this (level targeting) idea.
To conclude, I find no convincing evidence that Bernanke's technical views on monetary policy have changed dramatically. Prior to 2003, he believed that central banks had almost unlimited powers over nominal aggregates. After 2003 he discovered that the head of a central bank had very limited powers.