I was having lunch at a brewpub in Kokomo IN last Wednesday (Dec 16) as the multiple screens over the bar proclaimed "breaking news." Really? How is it that it is "news" that the FOMC voted to raise its short-term interest rate target to 0.25%?
Janet Yellen's press conference following the announcement reiterated that: "this is not an unanticipated policy move...[We] are not expecting to see short-term impact on financial markets." She made it clear that the decision reflected not current economic conditions, but rather the Fed's forecasts of where the economy will be in two years. That's because standard models and associated empirical results show that it takes 6-8 quarters to feel the full impact of an incremental change in monetary policy. Hence she stressed that "the Fed needs to monitor over time that [inflation] is moving as expected." She likewise emphasized "employment" as I have long done, not "unemployment."
Yellen reiterated that 2% inflation is not a ceiling but a target, and that she and others on the FOMC would prefer [PCE inflation] to be at 2% rather than near 2%. In the Q&A, she pointed out that the Fed has missed that target for 3 years, with inflation too low. Yes, oil prices have kept declining, which was a surprise, not of course limited to the Fed but including most of those involved in energy markets. That has contributed to the Fed undershooting its target, and she noted that the Fed in principle would similarly tolerate overshooting the target, if it was likewise judged due to transitory effects.
Again, none of this was "news". I was a little surprised – but only a little – that she made no reference to stock prices or other financial market metrics. She gave a good answer on the myth of "business cycles" – that expansions die of old age, to use her phrase. Instead an economy is constantly hit by shocks, positive and negative, so there is the potential in any year that such surprises could on net be negative. She gave the probability of a recession as "certainly at least 10%" but it was not "fated to happen because we've had a long expansion." She then enumerated potential upside shocks: residential construction in the US has more to surprise on the upside than on the downside, ditto oil drilling (not much room to fall further!) and (not everyone's belief) rest of world growth.
She did say a couple things that would be easy to miss. One is that the Fed is in no hurry to reduce its want reduce its balance sheet – in January 2009 it held no long-term bonds, in December 2015 it held $1.888 trillion. But the data make that clear, it's not news, it will take years to unwind. What was new to me is her statement that the Fed would likely maintain a larger balance sheet than in the past, without adding detail.
The Washington Post correspondent had the only question that caused Yellen to smile. If outside forces don't, will it be the Fed that kills the expansion? Her response: no, that's why we're watching inflation and starting to raise rates now, with low levels and small moves, so that we don't have to raise rates to high levels in large increments later.
Finally, she was asked whether their models need changing. Consistent with her repeated response that the Fed does not adhere to hard-and-fast rules, she noted that economists need humility on what their models can capture. Similarly, consistent with her discussion of whether there is such as a thing as a "business cycle" and "surprises", she noted that (my words) even good models are subject to uncertainty. Furthermore, "we do change models that are persistently not working...[but she was] not aware of a different model of inflation that would be superior to the one we employ". As my Econ 398 students can attest, that's a typical dilemma in macroeconomics: your models may not get inflation right, but that's because it's unrealistic to think we can predict the future. So model choice then hinges on ... inertia? (We read three papers on methodology, including one by Thomas Kuhn.) I'd like to hear more of Fed deliberations on that issue! What models have they judged "persistently not working" and on what basis do they judge a model superior?