NYTimes.com Article: Don't Count on It

["Steven G. Brant" <trimtab@sprynet.com>]

For those of us who cringe at the way statistics are used by politicians in
Washington, DC (I can't speak for other countries), this OpEd piece from The
New York Times's Paul Krugman is a sorely needed bit of truth telling.

Perhaps there's something we students of Dr. Deming in the USA should
communicate to our elected leaders about this variation on the old
expression "Lies, Damn Lies, and Statistics!"  We might also think about
communicating to these people about the critical importance of making
mistakes (as contrasted with the common attitude that mistakes are "bad".)

Steve Brant

---------------------------------------------
The New York Times, September 5, 2001

Don't Count on It

By PAUL KRUGMAN


In the 1970's and 1980's one often read about remote mountain
villages where people stayed healthy and vigorous into their 90's,
with some individuals reaching extraordinary ages. How did people
in these simple, traditional societies achieve such longevity?

 The answer, it turned out, was that people in simple, traditional
societies aren't very good at counting. There were some quite old
people in those villages, but they weren't as old as they thought
they were. Of course, one reason people were vague about their ages
was that in their societies nothing important hinged on getting the
numbers right.

 In one respect, the story of our vaunted "new economy" is similar.
In the last few years of the 1990's the data just kept getting
better, culminating in an estimated 5 percent growth rate for real
G.D.P. in 2000 — not an unusual growth rate for a poor country
playing catch-up, or for an economy recovering from a recession,
but amazing for an advanced economy late in an economic expansion.
How did we do it?

 The answer, it seems, is that we weren't very good at counting.
Last month's revision by government statisticians marked down U.S.
growth rates in each of the last three years, with the biggest
markdown for 2000. At 4.1 percent growth was good, but not as good
as we thought it was.

 But there's one difference between us and the villagers: In our
society numbers matter a great deal, and fuzzy math can have nasty
consequences.

 It's clear in retrospect that we shouldn't have taken the original
estimates of economic growth too seriously. Anyone who has looked
at how economic statistics are constructed knows that they are
based as much on educated guesswork as on hard facts. And the
guesswork gets more speculative as the economy's center of gravity
shifts away from solid, physical products to more ethereal stuff;
the biggest item in last month's revision involved a reduction in
estimates of business investment in software.

 Nonetheless, people who should have known better treated the
estimated growth rates as solid data, and in particular took to
heart their apparent message that the best was yet to come.
According to the figures, the U.S. economy wasn't just growing
fast, it was growing at an accelerating pace. And if that
continued, all of our worries about the long run — about the
burdens of providing pensions and health care to an aging
population — would simply melt away in the face of a red-hot
economy.

 Chief among the what-me-worriers was none other than Alan
Greenspan. His crucial January testimony in favor of big tax cuts
began with a paean of praise for the productivity revolution, "the
key factor driving the cumulative upward revisions in the budget
picture," and cited with approval long-term projections of an
"on-budget surplus under baseline assumptions well past 2030."
Seven months later the productivity estimates had been revised
downward, and the on-budget surplus was gone.

 Mr. Greenspan is still optimistic about long-run growth, as are
many other people; they argue that a productivity revolution driven
by information technology has decades to run. I would respectfully
submit, however, that these optimists have no idea what they are
talking about. I don't mean that they're necessarily wrong; I mean
that history has repeatedly made fools of people who try to predict
future technological developments, let alone the implications of
those developments for long-term economic growth.

 And most of the mistakes have been in the same direction. Since
the 1960's, futurists have consistently overestimated the future
rate of technological progress and economic growth. (Rent "2001: A
Space Odyssey" or read Herman Kahn's "The Year 2000" if you don't
believe me.) The only major upside surprise was the productivity
surge from 1995 to 2000 — and that, it turns out, was partly a
figment of our statistical imagination.

 So how should our society conduct itself now that the recent past
is not what it used to be? Any villager could tell you the answer:
Hope for good news but don't count on it. That is, don't commit
yourself to tax cuts and spending promises you may not be able to
afford.

 But of course we have already made those commitments. So the
question becomes, will our leaders ever admit that they made a
mistake, that we may not be able to afford the full tax cut after
all? And the answer is obvious: We should live so long.

http://www.nytimes.com/2001/09/05/opinion/05KRUG.html?ex=1000706768&ei=1&en=
2fe70753ab5dd5a9

Copyright 2001 The New York Times Company