Better the Devil You Know
August 20, 2010
By Arnaud Mares & Joachim Fels | London
This is a perplexing time for central bankers. Inflation expectations remain firmly anchored in the neighbourhood of their targets. This should be comforting, yet it is not. The variance of medium-term inflation expectations has been steadily increasing (see also: "Priced for Perfection...for Now!", The Global Monetary Analyst, November 18, 2009). So stable inflation expectations are not so much a sign of confidence as of indecisiveness across the market as to whether central banks are more likely to miss their target on the upside or the downside. This gives a new meaning to the notion that risks are ‘finely balanced'.Interestingly, central banks agree, or at least the Bank of England does. Based on its own projected probabilities of inflation two years down the road (taken from the August 2010 Inflation Report), assuming unchanged monetary policy, it believes that it is slightly more likely to miss its target than to reach it. The critical point is that it estimates inflation is about as likely to exceed 3% as it is to fall under 1%. Life was much easier in previous years. Up to 2008, the Bank of England was remarkably confident that it would meet its target without requiring much policy activism. In late 2008/early 2009, its confidence waned but it was at least secure in the belief that it if it was going to miss its target, it would be to the downside. Which policy stance to adopt in response seemed therefore straightforward at the time. Now the central bank still believes that it is likely to miss its target but, with no certainty on which side that might be, the policy response is anything but obvious.
Forced to choose between two evils. With extreme outcomes plausible rather than merely possible, central banks' reactions are likely to be shaped by their relative degree of aversion to inflation or deflation. In other words, which outcome do they most want to avoid? We believe that there are good reasons for all major central banks to fear deflation more than inflation.
The devil you know. First comes the weight of experience. There have been many successful (albeit at times painful) victories over inflation in past decades. There are only two significant episodes of sustained deflation in the past century: the US in the 1930s and Japan post-1990. Neither inspires a great deal of confidence that the means to emerge from deflation are well understood - and effective.
The zero bound is not overcome yet. In a truly deflationary environment (think Japan), central banks lose much of their ability to control their primary tool - the level of real interest rates - because nominal interest rates cannot be reduced below zero. If the risks of inflation and deflation are balanced, it is reasonable to expect that central banks would rather find themselves in a situation where their armoury remains effective. In other words, they would rather have an inflation than a deflation problem. Many suggestions have been put forward over the years to improve monetary policy effectiveness near the zero bound for nominal interest rates: expanding the monetary base (or bank reserves), purchases of long-dated public and private sector assets to push yields and spreads lower, commitment to low nominal policy rates for a protracted period, direct lending by the central bank, and currency market intervention to push the exchange rate lower and thus generate imported inflation. Most have been implemented to some degree in the recent crisis, or earlier in Japan. Neither - with the possible exception of the latter, which is obviously not available to all central banks simultaneously - genuinely overcomes the zero bound to nominal interest rates. There remains the proposal, made by Marvin Goodfriend in 2000, to impose a tax on central bank liabilities (bank reserves) to force nominal interest rates into negative territory. Independently of the merits of this proposal, it is again reasonable to expect that, given a choice, central banks would rather not be put in a situation where they have to conduct yet another monetary experiment.
Independence in inter-dependence. Central banks might be independent, but in this crisis the concept has acquired a Gaullian accent of ‘independence in inter-dependence'. The balance sheets of central banks and governments have become ever more entangled, as central banks have purchased government bonds and governments underwritten the risks assumed by central banks. Aggregating, as one should do, the balance sheet of the government and the central bank, one can see that quantitative easing results in the substitution, on the ‘whole of government' balance sheet, of long bonds for bank reserves. This has two effects: shortening of the maturity of government debt (by three years in the UK) and an increase of the sensitivity of the government's cost of funding to policy rates. Assuming that central banks were equally averse to missing their target to the upside or downside, they should prefer the outcome that is more favourable to government finances: that means inflation.
Inflation might not help a lot, but it would help at least a little. Inflation is no panacea for governments. With a substantial share of government liabilities indexed directly or indirectly to prices (in the UK, unfunded civil service pension liabilities for instance), inflation cannot be expected to fully bail out governments with overstretched balance sheets. This is not to say that some inflation would not help a little. It would certainly help a lot more than deflation (see also "Debating Debtflation", The Global Monetary Analyst, March 3, 2010).
So do central banks really prefer inflation? Past experience might not be a reliable guide to future performance, if only because, as underlined above, the current wide and symmetric dispersion of inflation expectations is a rather unusual situation. But if past inflation outcomes are in any way representative of central bank attitudes to inflation versus deflation, then one would conclude that the Fed and indeed the ECB had rather miss their targets to the upside than the downside. Both the Fed and the ECB have more frequently missed their respective inflation targets on the upside rather than on the downside.
If in doubt, do nothing. Let us now assume that central banks do not share any of the arguments listed above and that their aversion to above-target inflation is just as intense as their aversion to deflation. Then, confronted with a situation where they expect risks to be broadly evenly distributed if policy were unchanged, the most likely policy reaction would be to do...nothing. In other words, to maintain their foot firmly on the monetary accelerator, and keep real interest rates well into negative territory until inflation risks become much more skewed to the upside.
No rush for the exit. The first consequence is that one should not expect central banks to rush for the exit. And this is indeed how our country economists see it:
• In the US, Dick Berner and Dave Greenlaw expect the Fed to keep official rates on hold until 2H11. Also, the Fed's recent decision to prevent a passive tightening of policy resulting from MBS repayments by reinvesting the proceeds in Treasury bonds across the yield curve can be interpreted as buying additional insurance against a deflationary outcome.
• Likewise, in the euro area, Elga Bartsch foresees the first ECB rate hike in 2H11. Moreover, given the ongoing funding problems for banks, the ECB is unlikely to move away from providing unlimited liquidity through its refi operations anytime soon.
• And in the UK, given the dovish tone adopted by Governor King in last week's press conference, Melanie Baker has pushed back her expectation of a first rate hike to May 2011.
Inflation still the more likely of the two outcomes. If risks are symmetric ex ante, but the response of central banks is asymmetric, then the risk that deflation settles in must be less than the risk of inflation creeping up. This holds at least as long as one believes that monetary policy bears some relevance to medium-term price trends. Against this backdrop, we think that investors would be well-advised to protect themselves against an inflationary outcome to tighten.
No comments:
Post a Comment