Voir Dire posted a working paper yesterday that is worthy of reposting given its broad implications. The paper "University Rankings by Cost of Living Adjusted Faculty Compensation" by Terrance Jalbert, Mercedes Jalbert, and Karla Hayashi in the International Journal of Management and Marketing Research can be found at SSRN. The abstract can be viewed at either Voir Dire or SSRN, so I will save the space and not repost it here, but a bit more information can be found after the jump. It is perhaps sufficient to say that the aggregated and averaged data can be tremendously interesting to any of us that influenced by the academic job market.
The paper offers a huge contribution to to both academics interested in job pay and its effect on performance while also be a practical primer for those who are actually comparing and contrasting different positions and packages by universities (if anyone reading this is so lucky). The idea is pretty straightforward: take the salary (S) and benefits (B) offered by different universities and weight it by the adjusted cost of living (COLA). Not surprisingly, the unweighted rankings of salary are quite different than those weighted by the COLA. The authors do provide several rankings of all 574 universities and divided by the highest degree offered by each institution.
Methodologically, the COLA is the tricky part to calculate and the authors do so by creating an index based on the following:
In order to determine the relevant cost of living index for each university, the city where the university is located was identified. The identification was made based on information provided on the University’s website. Each city was searched against the Yahoo.com real estate website, neighborhood information section. This section reports, among other things, a cost of living index for U.S. cities (88).
The index is a relative ranking of each university that appears to range from an inexpensive 75 in Bluefield, West Virginia to a very expensive 377 in New York City. Generally, I am impressed with the index and am more curious about alternative measures than providing a real critique. For example, the city as the location for the school and boundaries for generating the cost of living variable can be misleading. Academics in more expensive cities may, with a non-8-5 work schedule, have more flexibility in their ability to commute and be able to live in areas that have cheaper living conditions. The commuting academic is not a foreign concept and perhaps a COLA based on a county-wide or census tract might allow for employees to substitute travel time for living cost. I am also curious how well a monthly/yearly rental rate would correlate to housing prices in a given area. Several cost of living variables are available from the census and easily exploitable (though, perhaps outdated given the current market). A proposed list of possible alternative cost of living variables are easy enough to generate: gas prices, potatoes, toothpaste, men's shirts, women's slacks, etc. (now I am going to play with this calculator as it has some odd items). I am sure the area housing prices are standard in the previous literature, but I am curious about other measures.
Regardless, the results are interesting enough to play "where's my university?" Naturally, the data are averaged over the entire university and the mileage per department may vary. As a further test, it may be possible to replicate this study using universities that are required to report employee wages. I know some public universities are required to report this to state laws – though I am not sure how universal this is.
As a final thought for now, part of the justification for the study comes from a reward for work
idea of employment – higher paid employees should be more productive.
While the risk of poor ecological inference would be present, a quick
comparison of school rankings and political science department rankings
might be revealing. However, that is a potentially socially dangerous project that I will leave to someone else.