Tuesday, April 24. 2007
More NT curvefitting
Kasriel Recession Warning Index, combination of year-over-year percent change in real monetary base and four-quarter moving average percentage point spread between Treasury 10-yr. yield and fed funds rate, 1960-2007
To refresh your memory, the combination of a year-over-year contraction in the quarterly average of the CPI-adjusted monetary base (unadjusted bank reserves and currency held by the public) and a negative four-quarter moving average of the spread between the 10-year Treasury bond yield and the federal funds rate has signaled every recession since that of 1969. I have modestly named this combination the Kasriel Recession Warning Indicator. As shown in Chart 3, the KRWI is now sending out a recession warning as of the first quarter of this year in that the yield spread is negative and the real monetary base is contracting.
Paul Kasriel, LEI and KRWI – It’s Different This Time? (PDF)
Northern Trust Global Economic Research, 20. April 2007
Since it was fun before, let's fit another curve. Here I did a simple 1,1-GARCH fit of real GDP to the suggested indicator as a dummy variable, lagged both one and two quarters. The results were almost as good as advertised.
VARIABLE COEFFICIENT STDERROR T STAT P-VALUE(I've hidden the full table here.)
const 3.86995 0.178010 21.740 <0.00001 ***
KRWI_1 -1.36415 0.593954 -2.297 0.02163 **
KRWI_2 -1.85474 0.380043 -4.880 <0.00001 ***
While any fit off of a dummy variable will look funny, it's instructive to check:
Real GDP vs KRWI 1,1-GARCH fit, 1960-2007
Mr. Kasriel's indicator has been triggered for only one quarter, 07Q1, so per this fit he should expect continued normal GDP growth in Q1 and reduced growth but no recession in Q2. Should this indicator continue to be triggered by the end of Q2, he and I might both start worrying about the soft-landing thesis.
Comments
Display comments as
(Linear | Threaded)
Looks coincident to the onset of the (relative or actual) trough in activity. That is not what we would normally expect out of what are essentially monetary indicators. Am I eyeballing wrong?
What looks coincident, the fit, or the indicator? Both are true; I fitted against the lags only, though in retrospect, I should have done both with and without, since advance GDP is published well after you can examine the indicator.
The results there are slightly better:
http://wcw.bignose.org/images/krwifit.2.txt
http://www.bignose.org/~wcw/krwi.2.png
That naive model is, then, forecasting a drop from trend for Q1 YoY real GDP of about 1.4%. Depending entirely on what you think "trend" means, that could be anywhere from a huge disappointment to right in line with current forecasts of about 2% (QoQ annual rate).
The results there are slightly better:
http://wcw.bignose.org/images/krwifit.2.txt
http://www.bignose.org/~wcw/krwi.2.png
That naive model is, then, forecasting a drop from trend for Q1 YoY real GDP of about 1.4%. Depending entirely on what you think "trend" means, that could be anywhere from a huge disappointment to right in line with current forecasts of about 2% (QoQ annual rate).
Quicksearch
Categories
Syndicate This Blog
Blog Administration
Powered by
The best fool-proof leading recession indicator I know so far is a combination of CPI-adjusted money supply and 12-month moving average yield curve. It missed no recessions yet and never gave a false warning as well. Btw, it points to recession now, wh...
Tracked: Apr 24, 19:38
While the headline again looked good, the series I track showed, at best, a hesitant hiccup of a bounce. Orders for U.S. durable goods rose more than forecast in March, signaling business spending started to recover as the first quarter ended. O
Tracked: Apr 25, 20:38
Real GDP vs 1,1-GARCH fit on CFNAI-M3 (1Q lag), OECD CLI (1Q lag) and KRWI (unlagged), 1967-2007 (1987-2007 shown) Given the mild interest in my previous runs, both the braindead OLS vs the Conference Board's leading indicators and the 1,1-GARCH fit on
Tracked: Apr 27, 18:09