My colleagues and I at Good Jobs First were excited at the publication of the New York Times series on the “United States of Subsidies,” since it brings a great deal of attention to a problem — corporate abuse of economic development assistance — that we have been working on for more than a decade.
We were also pleased to see the online
that the Times posted to go along with the articles. We had provided a copy of the master spreadsheet for our
database to Louise Story, the author of the series, and she made extensive use of it. Although the Times obtained some information from other sources, it appears that about 98 percent of their company-specific listings come from Subsidy Tracker. (
SEE UPDATE BELOW.
Now that we have had a few days to examine the Times database, we see that there are some flaws in the way the paper used our data.
First, a few words on Subsidy Tracker. In recent years, a growing number of states began to put company-specific information on at least some of their economic development awards—grants, tax credits, tax abatements, etc.—online. This was often in response to the subsidy accountability movement that we and our allies have built.
The problem was that these disclosures usually happened via hard-to-find reports and web pages that were often difficult to search even when you did locate them. Good Jobs First decided to collect all these disclosures and combine them into one national search tool. We introduced Subsidy Tracker in December 2010 with 43,000 listings from 124 subsidy programs in 27 states.
Over the following two years, we have expanded that to the current total of 247,000 listings from 409 programs in all 50 states and the District of Columbia. That expansion was not due entirely to wider official online transparency. Using open records requests, we also obtained unpublished data on scores of additional programs (the total is currently 89). By posting this information to Subsidy Tracker, we became, in effect, the original online disclosure source for these programs.
In recent months we’ve begun applying this approach to city and county subsidy programs, which are far behind their state counterparts in terms of online disclosure.
Despite all this effort, we recognized that we still could not claim to have captured anywhere near all the subsidy awards that have been made across the country. Not only did we still lack many programs, we also have irregular numbers of years of data among programs. That’s why we have not yet built into Subsidy Tracker a feature that enables instant aggregation of all the awards going to a particular company.
Along with the remaining gaps in the data, there is much that needs to be done with regard to the listings we do have to allow accurate aggregation. This includes the standardization of the variations in company names in our source materials, linking of parent and subsidiary companies, and accounting for mergers and acquisitions. There’s also the problem that some states report subsidy amounts for single years and others for multiple ones. These challenges are all part of our future work plan.
After getting our raw data, the Times did not consult with us on exactly how it would be used. We thus had no opportunity to warn the paper against the perils of aggregation. Specifically, we were not aware of the paper’s plans to create what it calls its $100 Million Club.
It is with this listing that the pitfalls in the Times approach become most apparent. The companies that receive the largest subsidies often get them in the form of packages negotiated with state and local officials. These packages usually consist of awards from various programs and may also involve project-specific awards outside established programs. Some of these pieces of packages are not included in state disclosure channels. It is part of our plan to research packages through other means and add them to Subsidy Tracker as a separate category. We’ve already begun the process in the Key Deals section of the state pages of the
section of the Good Jobs First website.
The Times supplemented the roughly 154,000 entries it took from Subsidy Tracker with about 2,000 listings the paper obtained on its own or from an expensive subscription service produced by a company called Investment Consulting Associates. This enabled the inclusion of entries that were gleaned from press releases but had not yet been reported in the official program lists we rely on for Subsidy Tracker.
Yet the $100 Million Club still ends up with numerous instances in which the totals understate the true amount the big subsidy grabbers have received.
For example, the Times lists a total of $338 million for Boeing, including $218 from South Carolina. Yet it has been
that the package Boeing got by locating a new Dreamliner assembly line in the Charleston area could be worth some $900 million.
Apple is said to have received a total of $119 million, yet the Times fails to include more than
in subsidies the company got for a data center in North Carolina.
The Times $100 Million Club also misses some major recipients entirely, including Volkswagen, which got more than
in connection with an assembly plant in Tennessee, and ThyssenKrupp, which got more than
in subsidies for a steel mill in Alabama.
And these only include deals dating back to 2007, which is the period the Times used in compiling its $100 Million Club. The larger Times database seriously understates the size of major deals that took place earlier. For example, it lists only $19.3 million for GlobalFoundries in New York State, even though the company took over a
deal originally offered to Advanced Micro Devices (which isn’t listed at all).
We applaud the Times for the great reporting that went into its United States of Subsidies articles, but the paper fell short when it came to the compilations featured in its database. Good Jobs First will continue to build our Subsidy Tracker tool and in the future will create what we hope will be a more accurate and complete version of a $100 Million Club.
(reposted from the
Dirt Diggers Digest)
After this blog was posted, Louise Story contacted us with some concerns. She confirmed that 98 percent of the entries (a total of 152,729) in the Times database came from Subsidy Tracker, but she says the number of entries that came from other sources was actually 3,844 rather than the 2,000 we estimated. She added that in dollar terms, a subject we did not address in the blog, Tracker entries account for 67.3 percent of the total in the Times database. However, we cannot verify that number because the Times has not given us its underlying spreadsheet.
Story also believes that the blog should have mentioned the fact that she contacted me several weeks ago to say that the articles and database would be published soon and in effect told me about her aggregation plans. She did indeed contact me but gave the impression that her work was completed, meaning that an effort to suggest any changes in methodology would have been moot at that point.