Fun_People Archive
1 Sep
The Productivity Puzzle

Content-Type: text/plain
Mime-Version: 1.0 (NeXT Mail 3.3 v118.2)
From: Peter Langston <psl>
Date: Wed,  1 Sep 99 11:29:48 -0700
To: Fun_People
Precedence: bulk
Subject: The Productivity Puzzle

X-Lib-of-Cong-ISSN: 1098-7649
Forwarded-by: "Jack Doyle" <>

Moody, F.  1999.  The Productivity Puzzle.  Commentary.   August
31, 1999.

It's an article of faith in the high-tech industry that computers make
workers more productive. The efficiencies introduced by personal computers,
experts like Bill Gates and Esther Dyson have long assured us, will bring
about a quantum leap in productivity not seen since the invention of the
wheel. And it seemingly makes sense. Put a computer more powerful than a
1960s mainframe into the hands of every single office worker in your
company, and it stands to reason that he or she can produce reams of work
in far less time than it would have taken in pre-Information Age times.
Unfortunately, it's not true.

Nobel laureate and economist Robert Solow, in what The Economist calls "the
Solow paradox," noted a few years ago that "you can see the computer age
everywhere these days except in the productivity statistics." In fact,
economists have long puzzled over the fact that, while computers have
undoubtedly helped fuel the long economic boom, overall productivity has
remained stagnant.  (Productivity is the ratio of inputs to outputs. The
fewer man-hours you put in for a given number of cars, or soccer balls, or
corporate litigations, the higher your productivity.) This, of course,
amounts to heresy. Indeed, a recent Commerce Department report concluded
that "Value added per worker in IT-producing industries grew at an annual
average of 10.4 percent in the 1990s." Alas, like the statistic in the joke
about the man who drowned in a creek with an average depth of six inches,
this number is deceptively reassuring. It turns out that productivity growth
in computer manufacturing--a sector notorious for the prodigious
efficiencies forced on it by crushing competitive pressures--improved at
the rate of 42 percent per year from 1995 to 1999, according to a study done
by Northwestern University economist Robert Gordon. If you remove computer
manufacturing from the rest of the information economy, Gordon writes,
productivity since 1995 "has been abysmal rather than admirable."

This seems impossible. Anyone old enough to have performed the same work he
or she now does back in the precomputer age can tell you that they get far
more work done in the same amount of time, and do it better. But the numbers
don't lie. Information may in one way or another be folded, collated,
stapled and filed faster and more figuratively than ever before, but it
still is not being turned into profit any more efficiently. In fact, if
Gordon is to be believed, workers may actually be less efficient now than
in the precomputer age. Which raises the interesting question, "How come?"

Many sleepless nights spent in pondering this question have led me to the
conclusion that Microsoft's Windows is to blame. No matter how frantically
and frequently Microsoft upgrades or fixes or replaces or enhances its
operating system (exercises, it should be noted, that in themselves cost
companies huge sums spent on installing new software and retraining their
workforces), the damned thing crashes and crashes and crashes. And every
time a computer freezes up in the workplace, at least one document is lost
or damaged.

Let's assume that every information worker in the United States suffers a
computer crash, on the average, of once per day--hardly an exaggeration.
And let's say that each crash results in a minimum of two hours spent
retrieving, repairing or replacing a lost document. There are at the very
least 1 million Americans whose work on a given day depends, one way or
another, on a computer running Windows. If we assign an average value per
document of $40--a conservative estimate, considering the cost in human and
computing resources per document produced--that's 40 million bucks a day
down the tubes. Thus we can assume an aggregate annual loss in productivity
of somewhere between $10 and $15 billion. Now, in a country with a GDP of
around $8 trillion, that may not sound like much. But $10 billion here and
$15 billion there, and pretty soon you're talking real productivity loss.

The more computers we bring into the workplace, the more resources we spend
repairing system-crash damage and the more overall productivity is lost.
Until growth slows in the computer-manufacturing sector, we'll continue to
live a pipe dream, comforting ourselves with false productivity-growth
numbers. And until Windows gets replaced by a more stable operating system,
we'll still toss away billions of dollars each year in computer crashes.

prev [=] prev © 1999 Peter Langston []