John Palmer at EclectEcon has made his own forecasts for 2009 . That's an implicit challenge to this blog, and we must not let it go unanswered. So I am now going to follow him down this foolish road, and invite you all to join me.
Rules:
Unconditional point-forecasts only, for these 5 macroeconomic variables (you can do others as well if you like).
Only those who have made a forecast will be allowed to laugh at others' forecasts in January 2010.
No prize for the best forecast (other than bragging rights).
"Best" forecast is defined as that which minimises the following loss function:
Loss = sum of absolute value of [(forecast-actual)/latest available end of 2008 actual].
Here are my forecasts (FWIW, which is not very much):
CPI, November 2009, year over year: +0.5%
Unemployment, November 2009: 7.5%
US/Can exchange rate, December 31 2009: US$ 0.90
Bank of Canada overnight rate target, December 31 2009: 0.5%
TSX, December 31 2009: 10,000
OK, let's see yours!
And I'm going to make one additional meta-prediction: the loss will decrease for later forecasts (which have the benefit of being able to see previous forecasts). In other words, I expect my (first) forecast to be the worst. (How's that for hedging your bets?)
Posted by: Nick Rowe | January 02, 2009 at 09:31 AM
Great contest! My picks are here:
http://economics.about.com/b/2009/01/02/canadian-economy-predictions-for-2009.htm
Link
Along with an invitation for others to play.
[Edited to add the click-through link - SG]
Posted by: Mike Moffatt | January 02, 2009 at 11:01 AM
In. http://stackelbergfollower.blogspot.com/2009/01/canadian-econoblogosphere-forecasting.html.
Link
[Edited to add the click-through link - SG]
Posted by: Andrew | January 02, 2009 at 11:59 AM
Wow! Battle of the Canadian Econ blogs!
Just so we have them all in one place,
Here's Mike Moffatt's:
CPI, November 2009, year over year: +0.8%
Unemployment, November 2009: 7.8%
US/Can exchange rate, December 31 2009: US$ 0.78
Bank of Canada overnight rate target, December 31 2009: 0.5%
TSX, December 31 2009: 9,400
and here's Andrew's:
CPI, November 2009, year over year: +1.1%
Unemployment, November 2009: 7.4%
US/Can exchange rate, December 31 2009: US$ 0.86
Bank of Canada overnight rate target, December 31 2009: 0.5%
TSX, December 31 2009: 11,250
Posted by: Nick Rowe | January 02, 2009 at 12:20 PM
Here are mine, although my inner econometrician is screaming "Report the predictive density! The predictive density, dammit!":
CPI inflation: 1.2%
Unemployment: 8.0%
USD/CAD exchange rate: 0.90
Overnight rate target: 0.5%
TSX: 9143.60
Posted by: Stephen Gordon | January 02, 2009 at 12:41 PM
Admit it! None of you 3 guys has a clue about the overnight rate, and you are all just following me! See what I said about the dangers of going first?
Out of interest Stephen; a thought that has been bugging me for some time: how would one write down a plausible loss function for such a contest if players did report their predictive density functions? The same thought crops up whenever I see a weather forecast with "60% probability of rain".
Posted by: Nick Rowe | January 02, 2009 at 12:56 PM
I do admit it. Once I saw that, I couldn't imagine a better scenario. The Bank is unlikely to be increasing the overnight rate target.
I think it's be easy enough to think of a linex loss function; many might think that the appropriate loss function would be asymmetric, penalising forecasts that are too optimistic. In other words, a biased forecast would be optimal.
Posted by: Stephen Gordon | January 02, 2009 at 01:14 PM
This seems like a great way to spend one of my final days off - challenge accepted. I'll post my forecast later on my site.
Posted by: brendon | January 02, 2009 at 01:34 PM
Here is what I think is in store for 2009: http://shockminuscontrol.blogspot.com/2009/01/canadian-econoblogosphere-forecasting.html
Link
CPI Inflation: 0.6%
Unemployment Rate: 7.1%
US/CAN Exchange Rate: 94 cents (with bias downward)
BoC Target: 0.75% (just to be different)
TSX: 8780
Real GDP Growth: -1.7%
That was fun guys, great idea!
[Edited to add the click-through link - SG]
Posted by: brendon | January 02, 2009 at 07:24 PM
I'm considering creating an asset that takes the form of the right to have one's prediction deleted. What would its price be today? In six months? On December 31?
Posted by: Stephen Gordon | January 02, 2009 at 07:41 PM
I might take you up on that Stephen - i'm already regretting my TSX forecast.
Posted by: brendon | January 02, 2009 at 07:47 PM
Not fond of point forecasts; here are mine:
CPI, November 2009, year over year: +1.5%
Unemployment, November 2009: 9.5%
US/Can exchange rate, December 31 2009: US$ 0.74
Bank of Canada overnight rate target, December 31 2009: 1.0%
TSX, December 31 2009: 7,500
I'd like to stake out the bear camp. If right, I'm some kind of dismal hero, if wrong, everybody else will be so happy, they will easily forget my awful forecasts.
FWIW, I'm expecting the real resource sector activity to bottom out this coming summer. That's the key assumption.
Posted by: westslope | January 03, 2009 at 12:50 AM
"FWIW, I'm expecting the real resource sector activity to bottom out this coming summer. That's the key assumption."
I'm not expecting it to bottom out, but I do think it will be weak - I based my "forecasts" on that as well. Particularly my exchange rate forecast. To have the CDN$ at .90 or so, you need oil at $65-$70. It is certainly possible for oil to be at this level (or even a whole lot higher) but I'm not convinced the global economy will be strong enough to support it.
Of course, I don't like point estimates either. My prediction for oil at the end of 2009 is that it will settle somewhere between $20 and $220. :)
Posted by: Mike Moffatt | January 03, 2009 at 08:20 AM
Now that's a hedge your bets range Mike! I reckon that there is some probability that oil could dip below US$20/barrel though I doubt it will stay there for very long.
I should add:
WTI CL, December 31, 2009: US$24/barrel
YOY real GDP growth, November 2009: -4%
Indoor pan-Canadian marijuana production (weight), YOY at year-end 2009: 30% increase
Posted by: westslope | January 03, 2009 at 12:11 PM
GDP -.01
OIL 32.00
UNEMPLOYMENT 7.2
USX .95
TSX 7,300
START OF US INFLATION
We haven't seen the real industrial shoe drop yet.
Posted by: PAT MURPHY | January 04, 2009 at 06:59 PM
I've thrown together a comparison of the Econ blog forecasts and Bay Street, check it out:
http://shockminuscontrol.blogspot.com/2009/01/consensus-forecasts-bloggers-vs-bay.html
Posted by: brendon | January 05, 2009 at 01:59 PM
CPI, November 2009, year over year: -1%
Unemployment, November 2009: 9.5%
US/Can exchange rate, December 31 2009: US$ 0.86
Bank of Canada overnight rate target, December 31 2009: 0.0%
TSX, December 31 2009: 7,400
Posted by: RP | January 17, 2009 at 08:41 PM
Good, RP's staked out the bear territory! My subjective probability distribution is somewhat skewed, and there did seem to be a non-negligible chance that the rest of us could be too bullish.
Posted by: Nick Rowe | January 18, 2009 at 06:00 AM