Post hawk ergo propter hawk

A rather forced pun to begin this waiting-room queue blog.

But post hoc ergo propter hoc refers to the fallacy of concluding that since after all A we observed B, A must have caused B.

This is a summary of Andrew Sentance’s response to my and Paul Krugman’s posts on why we should not opportunistically lower the inflation target to take advantage of currently low measured inflation.  Essentially, he looks back at periods when inflation was low or prices were falling, and says ‘often things weren’t all that bad’.

There’s no disputing that.

But it’s not adequate to stop there in my view.   If he wants to make progress beyond the tit for tat ‘well look at Japan’, Andrew needs to address the arguments and evidence for why mainstreamers fear deflation directly.

To itemise what I think he would need to do to mount a succesful counter, he has to:

1) Address whether he agrees that measured inflation overstates true inflation [part of the reason for the 2 per cent target] and, if not, why not.

2)  Explain whether he concurs that interest rates encountering their floor risks losing control of inflation on the downside, and that this would be a bad thing, and, if not, explain why not?  Doing this would be quite ambitious, since it would involve erecting an alternative analytical macro model that explains the data better than the one that we have currently, on which fears of deflation is based.  He’d also have to erect a model that didn’t contradict his own voting record, which often stressed the need to use interest rates to control inflation, via his urge to raise rates to stop inflation spiralling up.

3) Explain whether he concurs that much debt is set in nominal terms [if not why not] and that this risks debt-deflation spirals if the target is set too low.  If not, why not?

4) Explain whether he concurs that there is downward nominal rigidity in wages, which would lead to excess unemployment with a too-low inflation target.  And, if not, why not?

5) He has to find something that tilts the scales against the lessons learned from the crisis, which is that i) recessions are larger, and need more monetary stimulus to counter them, than we thought, and hence we need more room clear of the zero bound that a higher target would provide and ii) that equilibrium real rates are, on account of demographics, savings flowing uphill from the East, and cheaper investmen goods, likely to erode some of that ‘room’ in the future.  [It’s Andrew after all that talks of secular stagnation through the language of ‘the new normal’.]

In essence, looking back at time series charts of inflation and growth is not enough.  We need empirically validated models of the structural features in contention [DNR in wages, nominal debt, zero bound, expectations] to replay those histories and work out what inflation target would have been best then, and will be best in the future.

I’ll give Andrew a helping hand on tasks 3 and 4.  It would be possible to mount an argument like:  yeah, there’s nominal rigidity now, but this would erode pretty sharpish once people felt the pain of deflation.  Perhaps.  And perhaps the benefits would exceed the transitional pain.  But, on both counts, probably not.

 

 

Advertisements
This entry was posted in Uncategorized. Bookmark the permalink.

One Response to Post hawk ergo propter hawk

  1. “4) Explain whether he concurs that there is downward nominal rigidity in wages, which would lead to excess unemployment with a too-low inflation target. And, if not, why not?”

    So long as you have stable prices or any inflation how do you get excess UE from downward rigidity?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s