Go Figure

Its one thing for a company to prove a return on investment for spending on information technology. But how about for the manufacturing industry as a whole or the nation at large? Do computers really save us time, effort, and money as promised? Not everyone is convinced. In fact, there is a whole school of academics, not to mention a few closet corporate types, who, for whatever reasons, believe information technology has contributed not one byte to the nations productivity. They believe the untold billions of dollars spent on hardware, software, and networks this decade have paid little, if any, return on their investment in measurable terms. These critics point to slowing growth rates. Writing in the New York Review of Books last March, Jeff Madrick reports that gross domestic product discounted for inflation has risen only 2.2% annually since 1990. That compares somewhat dismally with growth rates of about 2.75% during the 1970s and 2.95% in the 1980s. Madrick then rhetorically asks, "Why havent computers dramatically augmented productivity?" In his answer, he rustles up a gang of likely culprits, ranging from the belief that most computer power is superfluous and therefore fosters unnecessary applications to the possibility that many business managers dont fully understand the true costs of computerization. Madricks isnt the only Luddite voice out there. No less than Harvard University business professors have weighed in against the idea that computers make us more productive. In his book The Squandered Computer (1997, Information Economics Press), Paul A. Strassmann, former CIO at the U.S. Dept. of Defense, finds "no correlation whatsoever between expenditures for information technologies and any known measure of productivity." But what if these older measures arent the most relevant means to gauge the effect of IT on productivity, particularly when it comes to manufacturing? Thats the question raised by Jim Shepherd, vice president of research at AMR Research Inc. "I cant believe these guys cant see it," Shepherd says, speaking of the IT-productivity skeptics. Addressing the Boston IT research firms customer conference in Phoenix last May, Shepherd suggested that the naysayers were using the wrong measures. "The fact that they are using 1960s formulas for evaluating equipment and plants is the real problem," he said. "ERP [enterprise resource planning] is infrastructure -- it doesnt fit neatly into an ROI equation." He added that "traditional ROI calculations cant factor in the impact of a new business process. Companies are buying a new way of doing business." To prove his claim that IT has reenergized American manufacturing, Shepherd cited a 1989 study by the MIT Commission on Industrial Productivity, which found six reasons for U.S. slippage vis--vis other nations. These were government regulation, tax policies, shortsighted investors, workforce quality, poor educational system, and failure to invest in technology. "The only thing that has changed since then," he argued, "is investment in technology." Using a variety of measures, he demonstrated the gains manufacturing has made in recent years. "The improvement in inventory management has been staggering since 1982, due to MRP systems alone," he pointed out. His numbers -- admittedly confined to manufacturing -- were convincing. Americas percentage of shipments of goods for export, as measured in dollar value, doubled from 1977 to 1991. While manufacturing revenue growth increased just 11.5% over the period from 1975 to 1979, in the years from 1982 to 1996, when many companies installed MRP systems, manufacturing revenue growth climbed 87%. And when measured in dollar value of goods shipped per production employee, U.S. manufacturing productivity nearly doubled during the last 15 years, from just over $160,000 in 1982 to almost $300,000 in 1997. Is manufacturing a special case? I doubt it. Were it not for the banks of mainframe computers at its Concord, Calif., data center, could you imagine how many people it would take for Bank of America to process manually 20 million checks every 24 hours? Go figure.