I've been home with a cold the past couple of days and was inspired by Simon Wren-Lewis's post on recent GDP data out of the UK to update the results (which I hadn't done with contemporary data in awhile).
In doing so, I realized that the fit to the price level didn't actually work for CPI data after 2008 -- the discrepancy is so large that it hints at the possibility of monetary regime change in the UK. However, there does not seem to be anything that the UK did that was different from the US in terms of monetary policy -- both nations conducted QE, bailed out banks, and neither devalued their currency though an exchange rate peg. Why then would the US not have a regime change while the UK did? The UK switched from an inflation target range in 1997 (and to CPI from a retail price index in 2003). Basically, this could be construed as evidence against the information transfer model requiring an ad hoc monetary phase transition to fit the data.
Any other ideas? Am I missing something?
There are no other problems besides the price level (I show the transition region is shown in gray and the two regimes in purple and blue -- the second regime uses a constant value of the information transfer index):
Here are the long (10 year) and short (3 month) term interest rates (I could only find data on monetary base reserves from 2006):
Note that interest rates appear to be unaffected by regime changes (see e.g. here). And here is the path of NGDP vs M0 (that I've described e.g. here for the US):
Update 4/2/2015: Here is the inflation rate (since I don't have seasonally adjusted data, I did a bit of LOESS smoothing to erase some of the high frequency components):
No ideas about the model, but I know some people attribute the high inflation in post-2008 Britain to the significant depreciation of Sterling around that time, e.g.ReplyDelete
Maybe this will help you.
Interesting -- thanks for the link, M.Delete
A significant depreciation could be a catalyst for a monetary regime change, but that one looks like it came from outside (exogenous shock), and not a monetary policy decision. So it would still be unexplained by the ITM ... maybe that is how it should be.
In that case, perhaps "monetary regime change" could be assessed by relative value of the currency to a basket of other currencies- that would certainly change the information value of a unit of currency.Delete
That is true, especially for countries with lots of imports and/or exports relative to GDP.Delete
Some more thoughts...ReplyDelete
Does the model still work for the Euro zone and other countries? I'm asking because the Fed's first QE program was apparently different from those conducted in Japan and the UK, as well as the latter QE programs of the Fed. Bernanke called it "credit easing" at the time:
From what I understand, the idea was to improve the capital position of banks by purchasing various assets on their balance sheets (which aren't government bonds). Other QE programs focused on government debt, I think. Anyway, what I'm trying to say is that perhaps the monetary regime changed not just in the UK but in many other countries, and that it is the US that is an exception.
That is a good point. However as the model only incorporates QE in the short term interest rates (successfully for the UK), the different mechanics of QE in Japan, EU, UK and US shouldn't have an impact on the price level -- at least according to the ITM.Delete
The model also does fine with the EU and Japan ... see e.g.:
I do have a speculative post that I am working on right now ... maybe monetary policy in the UK is identical to its monetary policy in the 1930s and 40s with a pegged interest rate. Previously that was associated with a change in the monetary regime: