Put an end to (some) bad economic analyzes



The receiver is apparently responsible for some really good news: the end of the federal government’s regional input-output modeling system, known to friends as RIMS II. The end of a government program can be good if the program is bad. The people who work on RIMS II are good people who try to do a good job, but RIMS II users are mostly mercenary economists and careless regional analysts, and the results of the program are pretty bad.

RIMS II apparently answers an important question: If Industry X grows in my area, what impact will it have on Industry Y and Industry Z, as well as overall employment and of the region’s income? Curious minds want to know.

The problem with the input-output model is that it is based on a theory that was outdated 50 years ago, that its results are not testable, and that it is widely used to “prove” dramatically unrealistic benefits. dramatically stupid projects.

The model begins with an input-output table, which indicates that, for example, the steel industry uses the output of the iron ore industry, the electric utility industry, as well as the labor force. labor and some less important inputs. If there is an increase in steel production, it triggers an increased demand for iron ore, electricity and labor. These increases, in turn, trigger increases in the production of inputs for the iron ore industry, electric utilities, etc. Some of these ripple effects are felt inside the region, others outside. All of these side effects fuel other side effects. Finally, the total impact can be calculated.

The approach sounds cool, and technically it is. Wassily Leontief won a Nobel Prize in Economics for developing the concept. What is wrong, however, is substantial. First, the model assumes that there is no price change. For example, if there is an increase in demand for labor, the model assumes that wage rates do not change, and therefore the tightening of the labor market does not impact the changes. real employment. If you think unchanged prices are a realistic assumption, just fill your car with gasoline.

Second, the input-output model assumes fixed relationships throughout the forecast horizon. There is therefore no hydraulic fracturing that lowers the prices of natural gas. There are no new social media companies that connect consumers and businesses in different ways. Books have not been replaced by Kindle, even in part.

The third problem is that the economic impacts are not verifiable. They are shown as changes from what the economy would have done otherwise, but it’s impossible to know. If the model says that a new convention center will add 5,000 restaurant jobs, you won’t be able to test the outcome a few years later because you can’t see what the number of jobs would have been without it. convention center.

An assumption implicit in the model is that unused resources would remain unused in the absence of a project. Let’s say there is an empty lot inside the city. Economic development planners suggest building a convention center. The economic impact is calculated assuming that in the absence of the convention center, the land would remain vacant. No private sector company would see vacant land as an opportunity for new development.

With all of these biases, input-output modeling is generally a tool for pig projects. Some economists know all the problems and try to keep the models under control. Other economists believe the impacts are generally significant. Which economists are hired by the promoters of major projects? Yes, the privileged consultants are those who systematically find big impacts of the projects.

Budget cuts triggered by the receiver will end the RIMS II program. It can’t come soon enough.



Leave A Reply