Yet another suspect economic impact analysis

I recently picked on an analysis of the economic impact of the Indiana State Fair, so it's only fair (grin) that I pick on an economic impact study that deals with my part of the state of Indiana.  According to the press release:

"...the [Indiana University] Northwest campus’s economic impact was estimated at $105.7 million, or double its annual operating budget of $50 million. That figure includes $50.4 million in direct impact and $55.3 million in indirect and induced impact, according to the report...The Tripp Umbach report also showed that the Northwest campus directly created 638 jobs and indirectly supported or induced another 639 jobs, for a total of 1,277 jobs. The report also calculated that IU Northwest generated annual local and state sales tax revenue of nearly $5.6 million."

Where to start?  The consultant (Tripp Umbach) used a multiplier of slightly more than 2.  But according to the Bureau of Economic Analysis Regional Input-Output Modeling Series, the multiplier that the BEA has estimated for northwest Indiana for the "Education" sector is only 1.73.  (You have to pay for the data, and I have.)  And I have always thought that even the BEA estimates tend to be biased upward.  But if we use the BEA multiplier, we get an economic impact of $87.2 million.

And that is clearly biased upward.  For an economic impact analysis to be accurate, we have to be reasonably able to comclude that our measure of the initial spending stimulus is exogenous--is generated by forces outside the area we are examining.  For (many) residential universities, this may not be a bad assumption.  But for an entirely commuter campus, a university all of whose students lived in the area before they enrolled, to assume that the entire budget of the university constitutes exogenous spending is a huge leap of faith.

Assume that the IU Northwest campus vanished.  It's likely that some fraction of those students would, in fact, attend college outside the region, so their spending on higher education within the region may in fact represent some stimulus.  But it's also likely--indeed, it's almost certain--that some of the current students at IUN would not be in college, but would still be in the region.  So whatever they are spending on higher education in the region would, in all likelihood, continue to be spent in the region.  The baseline spending impact ought to be (1) state-level funding for higher education at IUN plus (2) tuition paid by students at IUN that would had IUN not existed, have been spent elsewhere.

According to what I have been able to find, state appropriations for IUN are about $17 million per year and tuition is about $30 million per year (the balance comes from other sources which I will assume are exogenous).  Let's suppose that 60% of the students at IUN would have gone elsewhere, so we can count $18 million of the tuition as, in effect, exogenous spending. 

That gives us $38 million in exogenous spending.  Using the BEA's multiplier (1.73), that would give a total impact of $65.7 million.  That's about 40% less than the Tripp Umbach number, and I would, frankly, not be surprised if even $65.7 million is more than the actual impact.

This is, in my judgment, another case of an economic impact study conducted with an eye on what the organization paying for it (Indiana University) would like to find, rather than on what is likely to be real.