The Number Factory

It’s a selling point for Implan that all of its data are customizable by users. That is, if you know that an employer moving into town pays a higher wage than the industry average, you can use that more specific number in your model and get a more accurate result—provided you know how to change corresponding variables. Similarly, if you’re dealing with the retail sector, you must realize that Implan’s data sets are all based on producer prices, so you have to adjust for retail margining yourself. If Wal-Mart sells an item for $10, for example, the company might only be keeping 10 percent of that retail price.

Lindall says the next version of Implan (due out within a year) will handle retail margining more transparently. But such user challenges also bring him back to the assertion that either an economics background or Implan training or both are needed to get a meaningful result from Implan models. “The thing about an I-O model is, if you put something in, you’re gonna get something back out,” he says.

He also expects the next iteration of Implan to be able to predict impacts between neighboring regions, something I-O models haven’t been able to do very easily. “It will provide us with better estimates of the proportion of commodities that are bought locally,” Lindall says. “That will allow people to build multiregional models so you can link different [trading] areas together.”

For now, economists who use I-O models say that they’re best applied to small regions to measure the effects of a small change; larger economies get too complex for the models to accurately represent effects.

“A substate level is when you utilize Implan or RIMS II, because they have the multipliers that will be applicable to a smaller geographic area,” says Bob Isaacson, director of communications, analysis, and research for the Minnesota Department of Employment and Economic Development. His department has used Implan to assess the impacts of the state’s Job Opportunities Building Zone (JOBZ) incentive program for companies that locate in certain rural areas of Minnesota.

But there are some who wonder whether input-output models like those generated by Implan can ever be the right tool for a discussion of economic effects, particularly if the numbers generated are used to influence the allocation of public dollars. These are not critiques of Implan per se, but of using an I-O multiplier-based model at all.

Greg Ortale, former CEO and president of Meet Minneapolis, the city’s convention and visitors association, refused to use multipliers to project the effects of any tourism event during his tenure. “It’s an unnecessary inflation of reality,” he says. “When we hosted the Super Bowl in 1992, Meet Minneapolis estimated it would be worth about $60 million in direct expenditures. That wasn’t good enough, so the Super Bowl committee decided it needed an economic impact number, and [it] came up with $150 million. The point is, it’s ridiculous. The dollars themselves are big enough to stand alone.”

Art Rolnick, senior vice president and director of research at the Federal Reserve Bank of Minneapolis, agrees. “If you took every company in Minnesota and looked at its spillover effects and went through all these multipliers, the estimate of goods and services produced would be 10 times what we actually produce,” he says. Besides, “the numbers don’t tell you what you really want to know, which is what is the public benefit? What is the cost-benefit analysis?”