Picking up where I left off last time, the next step of course, is attention to detail - namely, what role does 'economic slack' or the output gap play along with the anchoring of expectations?
Ryan Avent over at free exchange asks that if you factor in the change in central bank credibility over time, and assume that it was effective, then what happened to unemployment?
Anyway, the key here to take away is to zero in on the notion that slack in demand affects quantity shifts rather than price shifts. Essentially, "Workers expecting prices to stay flat amid falling demand will resist wage cuts, and firms expecting workers to resist wage cuts will be reluctant to cut prices. Instead firms will produce and sell less and lay off workers, turning a given shock into much more of a real output loss".
But that's not something new. To that extent, Avent draws back to an '88 paper by Mankiw, Ball and Romer on the inflation-output trade-off titled, "The New Keynesian economics and the Output-Inflation trade off". Essentially, they empirically observe the same - during high inflationary periods, firms and workers are likely to adjust prices and wages more thus increasing responsiveness to shifts in demand. In contrast, during low inflationary periods, nominal rigidity sets in.
What the authors conclude is this:
High inflation -> relatively flatter Phillips Curve (changes in nominal AD have significant effects on output). Low inflation -> steeper curve (shift in demand is reflected faster in the price level).
But back to Chapter 3.
I won't dwell much on output gaps but this is what the estimates suggest - Capacity Utilization decreased by about 5-6% since the beginning of the GFC while Unemployment gaps averaged roughly 2%. The study clearly states that what this suggests is a considerable share of the GFC unemployment increase being cyclical (the gap between current unemployment and NAIRU).
More importantly, on the anchoring of expectations angle, it turns out that although current and expected inflation are positively correlated (as you might expect), the slope is low, implying that expectations are anchored to targets rather than the current level.
The basic approach involved here is again, eerily simple. Two differences are used, the regressor is the difference in the actual inflation rate (at the time expectations are collected) and the central bank's target level. The dependent is the difference between the long-term expectation at a given time and once again, the central bank's target level rate. If you visualize this, you can see why:
1) If the regression coefficient estimate is zero, then the relationship between expectations and the actual inflation would not be significant
2) If the estimate of the intercept is zero, then expectations would be centered at the central bank's target rate.
The results are best explained in a graph of the coefficient estimates:
For the record, these are rolling (five year window) regression on a 12 advanced economy sample since 1990.
Further, evidence on the relationship between the inflation level and its responsiveness to economic slack shows that slack persists and that "the recent stability of inflation is indicative of greater anchoring of expectations and a more muted relationship between economic slack and inflation".
But these are naturally tentative observations (what about headline CPI, import-relative price inflation, lags, shocks etc?) - whether this approach holds consistently can be explored further in a more formal framework and model.
More on that soon.
Ryan Avent over at free exchange asks that if you factor in the change in central bank credibility over time, and assume that it was effective, then what happened to unemployment?
Anyway, the key here to take away is to zero in on the notion that slack in demand affects quantity shifts rather than price shifts. Essentially, "Workers expecting prices to stay flat amid falling demand will resist wage cuts, and firms expecting workers to resist wage cuts will be reluctant to cut prices. Instead firms will produce and sell less and lay off workers, turning a given shock into much more of a real output loss".
But that's not something new. To that extent, Avent draws back to an '88 paper by Mankiw, Ball and Romer on the inflation-output trade-off titled, "The New Keynesian economics and the Output-Inflation trade off". Essentially, they empirically observe the same - during high inflationary periods, firms and workers are likely to adjust prices and wages more thus increasing responsiveness to shifts in demand. In contrast, during low inflationary periods, nominal rigidity sets in.
What the authors conclude is this:
High inflation -> relatively flatter Phillips Curve (changes in nominal AD have significant effects on output). Low inflation -> steeper curve (shift in demand is reflected faster in the price level).
But back to Chapter 3.
I won't dwell much on output gaps but this is what the estimates suggest - Capacity Utilization decreased by about 5-6% since the beginning of the GFC while Unemployment gaps averaged roughly 2%. The study clearly states that what this suggests is a considerable share of the GFC unemployment increase being cyclical (the gap between current unemployment and NAIRU).
More importantly, on the anchoring of expectations angle, it turns out that although current and expected inflation are positively correlated (as you might expect), the slope is low, implying that expectations are anchored to targets rather than the current level.
The basic approach involved here is again, eerily simple. Two differences are used, the regressor is the difference in the actual inflation rate (at the time expectations are collected) and the central bank's target level. The dependent is the difference between the long-term expectation at a given time and once again, the central bank's target level rate. If you visualize this, you can see why:
1) If the regression coefficient estimate is zero, then the relationship between expectations and the actual inflation would not be significant
2) If the estimate of the intercept is zero, then expectations would be centered at the central bank's target rate.
The results are best explained in a graph of the coefficient estimates:
For the record, these are rolling (five year window) regression on a 12 advanced economy sample since 1990.
Further, evidence on the relationship between the inflation level and its responsiveness to economic slack shows that slack persists and that "the recent stability of inflation is indicative of greater anchoring of expectations and a more muted relationship between economic slack and inflation".
But these are naturally tentative observations (what about headline CPI, import-relative price inflation, lags, shocks etc?) - whether this approach holds consistently can be explored further in a more formal framework and model.
More on that soon.
No comments:
Post a Comment