The Fed has decided to switch from a calendar-based monetary policy to conditional inflation targeting. In particular, Ben Bernanke stressed that the Fed wants to be able to adjust expectations for when tightening will occur "without the committee having to change dates." He also stated that accommodation removal would be gradual and based on forecast numbers. He assumes that an increase in the Fed Funds rate will only occur after unemployment goes under 6.5%. At this time, the FOMC would consider an exit strategy and this would have to be consistent with normalizing the balance sheet.
6.5% may be considered the new level of the structural unemployment, or the level from which wage inflationary pressures may build. It would be in line with estimates that one third of the increase in the unemployment rate following the crisis is due to structural factors.
One striking feature of the U.S. job market today is the cyclical stickiness of the employment to population ratio (dotted line above). The employment ratio is the product of the labor participation times one minus the unemployment rate:
Employment ratio = participation rate x (1 - unemployment rate)
Peopled forced out of the job market may reduce the unemployment rate, but not the employment ratio, which means that in spite of a statistical improvement in the job market, the whole economy is getting poorer.
Ben Bernanke was right to say that monetary policy cannot affect the long run unemployment level (the so-called verticality of the Phillips curve in the long run). But adopting an explicit unemployment target may be a vector of expectations volatility when the job market faces such a structural change.
Let's consider two scenarios:
- The employment rate remains cyclically insensitive and therefore pegged at its current level. In the meantime the pace of job market exit continues, hence the fall of the participation rate to its early 1970s level. This leads us to the consensus view of no moves from the FOMC before 2015.
- This is doubtful as such a stable employment ratio coupled with a significantly lower unemployment ratio looks highly improbable (and would be catastrophic for the U.S. economy). If we assume that the unemployment ratio creeps up at the pace observed over the last few months (which would be lower than that observed during the "jobless recovery") and that the participation rate falls at a slower pace. As can be seen in the charts below, the Fed's target could be reached as early as January 2014.
This is the "beauty" of conditional forecasting. But contrary to calendar-based communication, it comes with a cost. The Fed may have untied its hands, but exit expectations will be much more volatile.
If you expect QE4 to be efficient in capping long yields (doubtful, given that it looks more like QE2 than the balance-sheet-neutral Twist), then a good play might be on the volatility side. The implied volalility levels of the MOVE Index (basket of Treasuries futures), as well as U.S. Treasuries and early dated eurodollars, are not consistent with the VIX, or the huge uncertainty pertaining to the future pace of the decline in unemployment and the timing of the Fed exit.
For that reason I will soon be starting to consider being long U.S. Treasuries volatility.