MSU Consolidation Study Seriously Flawed

A new Michigan State University report on school consolidation appears to contain a substantial amount of plagiarized material, including whole paragraphs that seem to have originated from other sources, the Mackinac Center reported Aug. 18. Diligent reviewers shouldn't stop there, however, because the study's methodology is also deeply flawed. Even if one believes that all districts would save money through consolidation, the conclusion that they could save $612 million is wildly exaggerated.

The report, authored by MSU's Education Policy Center senior scholar Sharif M. Shakrani, takes the findings and mathematical model from a 2001 study of 12 rural New York state school district consolidations between 1985 and 1997, and applies them uniformly to all 551 Michigan districts, regardless of size or location.

The authors of the New York study acknowledged that they only examined rural districts, and even in these they could find no savings from consolidating larger districts: "We find no support for the use of state tax dollars to encourage consolidation among districts with 1,500 or more pupils."

This limitation should have eliminated about half of Michigan school districts from the MSU study's calculations. Instead, most of the 10 counties for which it calculated savings contain large urban districts, including Kent, Washtenaw, Muskegon, Saginaw, Kalamazoo and Genesee.

Beyond this dubious geographical misapplication, the MSU researcher also appears to have incorrectly applied some of the specific New York findings. For example, those earlier scholars found that consolidations saved 8 percent in overall "operating" costs, defined as everything except capital outlays. However, in the MSU report this overall savings rate is applied to a more specific "operations and maintenance" category.

This matters because the MSU author also tries to break down potential savings from consolidating services, not just districts, and incorrectly applies the same 8 percent savings prediction to this narrower goal. The result is exaggerated projections of savings for both consolidating districts and consolidating services. Actually, the 2001 study does not seem to predict cost savings at all for operations and maintenance.

The MSU report also misrepresents the data by claiming not to include charter schools in the consolidation, but including them in its cost savings calculations. The MSU report itself explains why doing so is a mistake: "(C)harter schools are not considered for potential inclusion in school district consolidation since they are intended to function independently. ..."

But the data used to generate the report's headline estimate of $612 million in potential savings comes from the National Center for Education Statistics, which includes charter school spending. This mistake further inflates purported savings.

Furthermore, the MSU study does not consider capital costs, which the New York scholars found actually increased after consolidation. Again, the result of excluding these costs is exaggerated savings compared to those that would result from a proper application of the earlier study model's details.

There are other problems. The original 2001 study projects consolidation savings over a 12-year period, but the MSU report appears to compress these into just three years.* Also, most of the figures Shakrani used to calculate potential Michigan savings come from one table in the New York study listing the projected "economies of scale effects" of combining two 1,500-student rural districts. Yet these are the very districts the earlier authors concluded should not be expected to produce consolidation savings, because other rising costs offset them.

In sum, the MSU report on school consolidation, and the widely reported potential cost saving figures it generated, suffers from two irreparable flaws: The first is simplistically applying 20-year-old data from a narrow class of school districts to current spending trends for the entire state of Michigan. Compounding the methodological error, procedures from the New York study were misapplied and data was misrepresented, significantly inflating the potential cost savings claims. These flaws invalidate the product as a credible resource for any discussion about how to generate better results at lower costs in the Michigan public school system.

* It is possible that the MSU author reran the New York study's statistical model using Michigan-specific data to find the savings over a three-year period. However, the report does not provide any results from a regression if one was used, nor does it mention anything about rerunning the statistical model. It appears that author simply multiplied the predicted savings rates (although incorrectly in at least one case) by total 2009 spending for certain school services in Michigan.

Permission to reprint this blog post in whole or in part is hereby granted, provided that the author (or authors) and the Mackinac Center for Public Policy are properly cited.