>n Sat, 7 Feb 2004, Doug Henwood wrote:
>> The late-2001 titanium
>> Powerbook G4 I'm typing this on seemed pretty snazzy two years ago.
>> Now it's seeming a little constrained & poky. So that has to be
>> accounted for somehow. But it's far more apparent if I'm editing a
>> large Photoshop doc than if I'm just typing an email. What's the
>> "real" increase in output between a late-2001 vintage machine and
>> today's? The BLS PC price index says it's 56% more. Is it really?
>
>So let me get this straight: we're going to bump up the values of
>computers this year because last year's model seems "a little poky"
>compared to this year's? Is there any actual evidence at all
>that this marginally faster speed reading email and producing
>photoshop documents actually led to any actual increase in the
>production of goods and services? To be a little frivolous,
>couldn't be it like that onion article: "network down,
>employee productivity rises 200% because nobody's
>using their computers to browse Ebay?"
>
>I know economists don't worry about whether their utilitarian
>models actually predict complex human social behavior, but
>this hedonic pricing stuff is just goofy to me. Am I
>missing something?
Like I've said, I think something like the adjustment is justified. Today's computers are a lot more powerful than those from five years ago. I can keep Eudora, IE, iTunes, Photoshop, Illustrator, InDesign, Word, Excel, SoundStudio, and Acrobat running all at once now and shift instantaneously between them, often carrying data or images. A decade ago, I'd have had to quit one program to start another. So something dramatic has happened to my "real" output over that decade. But the techniques the BLS and BEA use assume that physical specs alone are accurate proxies for the increase in real output, which seems wildly overdone to me.
Here's the official story - the BLS's PC's and peripherals component of the CPI. Columns show the value of the index (which started in Dec 97 at 100), the yearly change, and the "real" value of 1,000 1997 computer dollars over time:
index yty chg "real" $ 12/97 100.0 1,000 12/98 64.2 -35.8% 1,558 12/99 47.2 -26.5% 2,119 12/00 36.5 -22.7% 2,740 12/01 25.3 -30.7% 3,953 12/02 19.7 -22.1% 5,076 12/03 16.2 -17.8% 6,173
Has the real output of PCs grown over sixfold? Hard to accept that, but that's what the hedonic formula says. What's the real impact on real users? That'd be much harder to measure than the physical descriptors (RAM, processor speed, storage capacity, etc.) that go into the hedonic model, but conceptually, that's what it should be.
Doug