February 09, 2004
Yep. The Administration's Forecasts Are Weird Beyond Belief
Yes, it is there. Table 3-1 of the 2004 Economic Report of the President, on page 98.
To get from our current 130.1 million payroll employment number to an average number for 2004 of
132.7 131.9 [updated for employment shortfalls since the forecast was frozen last November: the administration argues that it should be allowed to lower its target since there has been bad employment news since then] million requires that payroll employment growth average 470,000 320,000 [as per lowered target] a month for the rest of this year. That's ridiculously high: nobody is forecasting growth like that.
I don't yet know whether the Bush administration High Politicians reached down into the staff forecast and manipulated the employment number, or whether some people thought they should try very hard to please their political masters even if it created forecasts that all informed economists would laugh at.
To get an idea of how silly the forecast is, consider that productivity growth has been averaging some 3.2% per year since 1995. Consider that they are forecasting fourth quarter-to-fourth quarter real GDP growth of 4% in 2004. Consider that 470,000 320,000 payroll jobs a month is an employment growth rate of
4% 3% in 2004. Consider that when employment rises, hours per worker rise as well.
Since 4% (GDP) - 3% (employment) - 1% (hours) = 0% (productivity), the administration's economic forecasters are implicitly forecasting that worker productivity will drop by 0% this year.
That's completely incomprehensible.
Posted by DeLong at February 9, 2004 12:46 PM
| | Other weblogs commenting on this post
Great news for Kerry; Snow's gonna have a blister on his tongue after 10 months of issuing retractions for those bogus numbers.
What causes that increase in hours worked?
If it is primarily because of training new people then some bright (well, maybe bright) bulb might be figuring that part of the low capacity utilization numbers means that there are plenty of trained workers ready to enter the workforce quickly.
So what cause is usually attributed?
While I would not dream of questioning you on economic matters, Brad, I can't help thinking there is something a trifle odd in this:
...consider that productivity growth has been averaging some 3.4% per year *since 2005*...
Brad writes: "The forecasters appear to have taken a dive. I'll try to find out why."
Is this in question? The forecasters have succumbed to the same pressure that George Tenet and his analyists did.
Someone in the administration demanded that the payroll numbers in November be equal to or greater than those in January 2001, lest Bush be painted with the Hoover brush. The analyists complied.
Indeed, Professor DeLong is ahead of his time.
Heh, maybe they plan to use hedonic indexing to arrive at the 470,000 a month number ;) See, you take 47,000 actual jobs per month and say that the workers are *really* ten times as good!
Brad DeLong writes:
> To get an idea of how silly the forecast is, consider that
> productivity growth has been averaging some 3.4% per year
> since 2005. Consider that they are forecasting fourth quarter-
> to-fourth quarter real GDP growth of 4% in 2004. Consider
> that 470,000 payroll jobs a month is an employment growth
> rate of 4% in 2004. Consider that when employment rises,
> hours per worker rise as well--for 4% employment growth, a
> rule of thumb is that hours will rise by 2%.
> Since 4% - 4% - 2% = -2%, the administration's economic
> forecasters are implicitly forecasting that worker productivity
> will drop by 2% this year.
Well, you are correct that the report does say what it says, but I think a more likely (if not exactly flattering) interpretation is that what the report says is just a flat-out howler. If 132.7 million was the year-end level for payroll employment rather than the calendar year average, the increase in employment is now just 2% in payrolls, or probably 1% in hours worked, so now we have:
4% - 2% - 1% = 1%
OK, so I'm not going to argue that a "70s revvial" 1% productivity growth figure is likely, but it's at least possible. So I'm suggesting that instead of being essentially corrupt, that the writers (or their editors) were being sloppy or lazy and a bit wishful. Meanwhile, I'm not an economist, but I see no reason why we couldn't project just as easily:
4% GDP growth - 0% employment growth - 0% hours worked = 4% productivity growth.
OK, so maybe I do: if hiring were that weak, consumer spending might begin to freeze up and we'd be back to explaining a 1% GDP growth using a 4% productivity growth number. But we wouldn't then need to wonder who would win the presidential election...
"Old" Foreign Policy + "New trade theory" = ?
The Economic Report of the President (as quoted by wire services) puts productivity growth at 2.1%, which it says is the 43-year average. That's 4% higher than you impute? Very odd. Total job growth (again according to wires) is 2.6 mln for 2004, which comes out to 216k/month (though we are off to a soft start). Is the report inconsistent in its claims for job growth between sections, or has the White House handed out a press release that differs from the actual report? Don't know. Is it because you are looking for "an average number for 2004 of 1.327 million" and the White House is looking at year-end. Even the generous approach (generous toward the CEA's credibility) requires the fastest acceleration in job growth since 1983.
Brad, I don't see any mention of the word "average" in the Economic Forecast. The Table 3-1 you reference simply says 132.7 million is referred to as "Level, calendar year." I assume that means they expect the level of payrolls to be 132.7 million at the end of CY 2004. Only Reuters -- which can be highly unreliable -- uses the word "average" in reference to that 132.7M figure. It would seem that they're mistaken. Am I missing something?
In Table 3-1 the "130.4" payroll figure for 2002 is not the end-of-year figure (the end-of-2002 payroll number is 130.1), it is the 2002 annual average figure.* "Level, calendar year" in CEA-speak always means "calendar year average." When they want a number to be an end-of-period number, they say so--as in the money stock tables in the back of the ERP.
*The "130.1" payroll figure for 2003 is neither the end-of-year figure (130.0) nor the annual average (129.9)--I take this to mean that the table was frozen in late December, before the last month's numbers had come out.
The 05 number is 136.3. That predicts a similarly large growth in jobs between 04 and 05. 2 years of explosive job growth? Who's hiring?
DeLong: "...the administration's economic forecasters are implicitly forecasting that worker productivity will drop by 0% this year."
What do you mean by this, exactly?
That worker productivity will be 0%, or that the productivity will not increase at all in 2004?
I assume the later, but your phrasing is hard to parse.
Bush could just barely pull it off, if he starts wars with Korea, Iran and Syria all at once and starts drafting troops like mad.
Ah. Thanks for the clarification, Brad. This is indeed astounding, then.
Umm you seem like a really hard crowd to please - you now appear to want really high productivity growth plus high job creation. But that’s only possible with extraordinarily high growth – which just aint going to happen.
So is the current forecast so weird? Have a look at 1993-1995 in
Payrolls were growing but here’s the weird thing, the productivity index reads
i.e. less than 1% growth over 2 years!
Looks to me like the Troika has its forecasts for the upturn spot on – rising payrolls and virtually static productivity growth – just like it was last time.
This whole discussion is reminding me of a remark made by Scott Adams: "Rember, the future is based on assumptions and assumptions are things you make up."
Make all you can, save all you can, give all you can.
We are never truly sure of our beliefs.
Seekers of truth invariably turn to lies.
No cause is so right that one cannot find a fool following it.