Spending Without Results?

by Paul A. Strassmann

Computerworld
April 15, 1996

Computer expenditures are definitely up. The productivity gains from all that spending are definitely questionable.

The origin of the term "computer paradox" has been attributed to my statistics teacher and the Nobel Prize winner in economics, Prof. Robert Solow. Solow was searching for evidence of computer induced gains in productivity numbers, as reported in national economics statistics. He couldn't find such proof and finally gave up the search. The numbers coming from Washington weren't sufficiently reliable for anybody to figure out whether computerization or any other influence had any effect at all.

Paradox revisited

As an engineer, I prefer measurements that can be repeated, especially if I know that somebody has engaged in an organized effort to secure the data. Instead of following economic statisticians, I like to examine the results produced by accountants and auditors. Annual reports from corporations contain data about revenue, employment, profits and assets.

Collecting information about what happens to firms in a period of several years isn't difficult. Obtaining information about computer expenditures is much harder. The best numbers that are publicly available are the ones reported by computer magazines about information technology budgets.

I managed to accumulate the annual information technology budgets for 138 giant U.S. corporations; the figures were from 1988 through 1994. This includes the budgets of firms such as AT&T Corp., General Motors Corp., Rockwell International Corp. and Xerox Corp. I estimate that this sample accounts for about half of information technology spending for Fortune 1,000 corporations.

Knowing information technology budgets and the corresponding financial performance makes it possible to test whether or not there is a "computer paradox." If the growth in information technology spending is less than growth in profitability, then the paradox would be exorcised. It couldn't be raised as a challenge to the credibility of computer advocates.

The facts

I found that at a 67.4% gain, information technology budgets increased much faster than revenue (29.6%) and profit (39.7%). Only employment spending lagged, growing only 2.3%:

(in $Millions)       1988       1989       1990       1991       1992       1993       1994    Growth
Total Revenues    1,587,648  1,693,139  1,813,275  1,831,029  1,882,909  1,924,456  2,057,344   29.58%
Total Profits        85,321     84,790     73,500     53,366     64,290     91,155    119,213   39.72%
Total SG&A          256,366    275,041    308,246    335,259    346,722    350,125    363,706   41.87%

Total Employees   9,235,000  9,460,930  9,507,673  9,384,409  9,442,407  9,170,737  9,450,434    2.33%

Total IT Budgets     28,061     34,328     37,544     39,955     40,723     42,807     46,973   67.40%

These differences in growth rates don't reflect favorably on the contributions of information technology. Furthermore, the Sales, General and Administrative (SG&A) overhead-the line on financial statements, where information technology expenses are normally incorporated - has grown faster than revenue or profit. If high investments in computerization resulted in improved productivity, then SG&A should slow down as compared with revenue. That would indicate the computer payoff in terms of lower overhead expenditures, which would lead to increased profit.

It didn't happen.

Some of the following ratios explain why:

IT as a % of     1988    1989    1990    1991    1992    1993    1994
Total Revenues   1.77%   2.03%   2.07%   2.18%   2.16%   2.22%   2.28%
Total Profits   32.89%  40.49%  51.08%  74.87%  63.34%  46.96%  39.40%
Total SG&A      10.95%  12.48%  12.18%  11.92%  11.75%  12.23%  12.92%

$IT per capita  3,039   3,628   3,949   4,258   4,313   4,668   4,970

Information technology budgets have grown relative to all other indicators and are greater than profits for 56% of corporations.

Information technology per-capita spending in 1994 averaged $4,970 per employee; this cost may be the second-largest indirect expense after health, pension and other personnel benefits.

The increasing information technology cost ratios are giving many chief information officers the attention they always wished they had. There is some reason to believe that the relative importance of information technology is understated.

Local operators are becoming more inclined to acquire hardware, software and consulting services that don't show up in the information technology budget. In some companies, that makes the calculation of the productivity gains from computers confusing enough to further cloud the issue.

Based on as good data as one can get, I conclude that, for most companies, it may be too early to banish the "computer paradox" to oblivion, despite reports by leading magazines of its demise.

The only thing to do is to make sure that the rise of computer spending doesn't get ahead of the gains in profitability and productivity.

Copyright 1996 by IDG Communications, Inc., 500 Old Connecticut Path, Framingham, MA 01701.
Reprinted by permission of Computerworld

Go back up to the Strassmann, Inc. home page.