The 2009 terminations · TFR-2026-05

The criteria: what General Motors used to pick 1,454 dealerships, and the 763 that met them and stayed open

Reference 2009 Terminations

First published 2026-08-05 · Last substantively revised 2026-08-05 · Corrected 2026-08-05

General Motors wound down 1,454 of its 5,591 dealerships in 2009, selecting them in two phases against criteria it described as objective. When the Special Inspector General for the Troubled Asset Relief Program audited that selection a year later, it found the criteria had not been applied consistently and that, at the time the decisions were made, the company had not written down why. The plainest measure is one sentence in the audit: 992 dealerships with a Dealer Performance Summary score below 80 were selected for closure, and another 763 with a score below 80 were retained.

§1Where the audit came from

The audit exists because a Senate committee chairman asked for it. On 2009-07-24 the Senate Committee on Commerce, Science, and Transportation announced that Senator Jay Rockefeller had requested that the Special Inspector General for the Troubled Asset Relief Program “review the methodology of the General Motors (GM) and Chrysler decisions to terminate more than 2,000 dealership franchises across the country.” His stated reason was not that the closings were wrong but that nobody could explain them: “There is substantial confusion, even among dealers themselves, as to how GM and Chrysler selected dealerships to terminate and what benefits, if any, they might gain by doing so.”² Congress had already taken testimony from both manufacturers and from dealers selected for termination, at a Senate Commerce hearing on 2009-06-03 and a House Energy and Commerce oversight hearing on 2009-06-12.³ The audit that followed, published 2010-07-19, opens on the senator’s question in almost his words.¹

§2The stated criteria

Setting aside the separate wind-down of Saab, Saturn and Hummer, General Motors set out to reduce its remaining network from 5,591 dealerships to 4,137, expecting normal attrition to carry it to an “ideal” network of roughly 3,300 in time.⁴ It selected in two phases, and sent every agreement at once.

Phase one, May 2009. General Motors identified 1,096 dealerships; 14 of those decisions were reversed before agreements went out and 11 dealers terminated voluntarily before the bankruptcy, leaving 1,071. Letters were dated 2009-05-14. A dealership was selected on one of two tests applied across all 5,591: a Dealer Performance Summary score of less than 70, or annual sales of fewer than 50 new vehicles in 2008.¹

Phase two, early June 2009. General Motors identified 383 more. It had by then decided to eliminate the Pontiac and GMC Medium Duty Truck brands, and 144 of the 383 sold only those brands. The remaining 239 were selected on what the company described to the auditors as a “more aggressive” set of five tests: a score of 80 or less; unprofitable in 2006, 2007 and 2008; a Retail Sales Index below 70; non-GM brands in the same facility together with a score below 100; or interference with the company’s plans for the Buick-GMC or Cadillac networks.¹

The commercial terms moved faster than any of the criteria. Agreements went to all 1,454 dealerships in the first week of June and had to be signed and returned by 2009-06-12 for a dealership to receive any compensation at all. How long that left is a question the sources answer only from the room: Senator Olympia Snowe, holding a copy of the agreement at a Senate Commerce hearing on 2009-06-03, called it “12 pages, single spaced” and said no one could “possibly make a decision within the 10 days they were required to make a decision,” to which General Motors’ president and chief executive, Fritz Henderson, replied that “almost half of the dealers signed it the day they received the wind-down agreement.”³ General Motors agreed to pay $587 million in total, on a formula built from rent, sales and late-May new-vehicle inventory, 25 percent up front and the rest on milestones.¹

§3What the criteria did not sort

The auditors ran the criteria against the company’s own data. In phase one, two terminated dealerships met neither test; in the other direction, General Motors “retained 355 (or approximately 41 percent) of the 858 dealerships that had a DPS score below 70” and retained 9 of the 394 that sold fewer than 50 new vehicles in 2008 — 364 dealerships that, on the published criteria, qualified for termination. In phase two the same pattern appears at every threshold.¹

All figures from SIGTARP-10-008, pp. 18–19. The second row is the audit’s own phrasing; its criterion elsewhere is a score below 100.

Test	Wound down	Retained
Dealer Performance Summary score below 80	992	763
Score of 100 with a non-GM brand in the same facility	269 (226 + 43)	299
Met none of the phase-two performance criteria	39	—
Missing at least one criterion in GM’s data	61	247

The last row is a different kind of failure. General Motors was missing at least one of the three inputs — the performance score, the Retail Sales Index or 2008 retail sales — for 308 dealerships, of which 61 were terminated and 247 retained. The company told the auditors the data was missing because the dealership had not provided it or was new, and that where a score was missing it looked instead at the Retail Sales Index and new vehicles sold. In its conclusions the audit puts the phase-two gap at its widest: General Motors “terminated 39 dealerships that did not meet any of the objective criteria and retained more than 1,062 dealerships that met one or more criteria for termination.”¹

General Motors did offer reasons, and they belong on the page. It told the auditors the inconsistencies were attributable “primarily to a desire to maintain coverage in certain rural areas where they have a competitive advantage over import auto companies that are not typically located in rural areas,” and that others were kept because they were recently appointed, were key wholesale parts dealers, or were minority- or woman-owned. The audit records the explanation and, in the same sentence, what cuts against it: “although ultimately close to half of all of the GM dealerships identified for termination were in rural areas.”¹

§4The part that could not be checked

A criterion applied unevenly can still be defended case by case, if the cases were written down. They were not. The audit’s finding on this is one sentence and it is the load-bearing sentence of the whole report:

During the time these decisions were made, GM did not document why some dealerships meeting the criteria were retained while others were wound-down.

SIGTARP-10-008, p. 19

To answer the auditors’ questions, General Motors officials “had to contact various regional or field representatives over several weeks to obtain their reconstruction of the impetus for decisions made several months prior.” The consequence, in the audit’s conclusions: there was “little or no documentation of the decision-making process to terminate or retain dealerships with similar profiles, making it impossible in many cases for SIGTARP to determine the causes of deviations from the supposedly objective criteria.”¹

§5The score itself

The Dealer Performance Summary score had been in use since 2002. It is the sum of four weighted category scores — the Retail Sales Index at 50 percent, the Customer Satisfaction Index at 30 percent, capitalization at 10 percent, profitability at 10 percent — each a ratio of actual to expected performance. The expectations are averages: General Motors calculated expected sales from “a segment-adjusted state average” and the expected customer-satisfaction score from “a regional average.” A score of 100 was average; below 70 was a poor performer that would not be retained. So half the weight of the number that decided a dealership’s future was its sales measured against a state-level expectation, and a further 30 percent a satisfaction score measured against a regional one.¹

Dealerships could see their own score on the same website they used to order vehicles, and the auditors confirmed that they could — then counted how many had. “SIGTARP found that only 26.1 percent of terminated dealerships viewed their DPS score on the website in 2008, and 47.5 percent did so in 2009.”¹ The number that selected them was published to them, and in the year the selection was built on, roughly three in four had never looked at it.

§6The appeal, and what it was not

General Motors set up an appeals process after announcing the closures and entering bankruptcy. It opened 2009-06-04 and closed 2009-08-07. Dealers who appealed still had to sign and return their wind-down agreements by 2009-06-12, so the appeal ran after the decision it challenged had been accepted in writing. The company created an appeals review team and an Executive Review Committee but, in the audit’s words, “did not establish criteria for the review or for the reversal of wind-down decisions,” and the review itself “was based on a second look at the same data used in the original wind-down decisions.” It received 935 appeals from dealerships facing complete wind-down and granted 64 reversals; counting partial wind-downs, 1,316 appeals and 86 reversals, 22 of them partial.¹

It gave no guidance on what to submit. The auditors reviewed 323 appeal packages and found dealers had sent whatever they judged relevant, from updated financial statements to letters from their towns. The audit reproduces one, from a dealership whose closure was reversed:

We have not heard back from anybody. We have just moved into a brand new dealership 04/14/2009. We do not understand this letter. We would like to appeal this. Please look at our investment. We have moved to the corner of two major highways and invested over 2 million dollars. We feel you might not be aware of our new dealership since it was addressed to our old address and name.

Excerpt from a dealer appeal package, quoted in SIGTARP-10-008, p. 20. The audit does not identify the dealership.

The audit’s next line is that General Motors reversed that dealership’s closure and did not document why. Asked afterwards, the company again gathered “undocumented recollections” from field representatives; the reasons supplied were rural market coverage, recent facility upgrades, corrections of erroneous score data, legal advice and leadership review. “Without proper documentation from GM, SIGTARP could not validate the reasoning or consistency of appeal decisions.”¹

§7Chrysler, and the finding that runs the other way

Chrysler published no thresholds. It ran a market-by-market review built around Project Genesis, its pre-bankruptcy plan to have every dealer sell all three brands by 2014, using an analysis by the consultancy Urban Science across 1,712 markets. Its primary tests were whether the location was one the company was targeting, which brands the dealership offered, its new-vehicle sales volume and its Minimum Sales Responsibility score. It terminated 789 dealerships, provided them no financial assistance, and offered no appeal at all. The audit puts the whole of it “within 22 days” — a phrase it repeats and never derives; the interval between the letters of 2009-05-14 and the effective date of 2009-06-10 in its own timeline is 27.¹ The two routes, and the dates on each, are set out at two companies, two legal routes, seven weeks.

The audit’s finding on Chrysler’s criteria is the reverse of its finding on General Motors’, and both halves have to be read together: “Perhaps not surprisingly in light of the case-by-case nature of the process, SIGTARP did not identify any instances in which Chrysler’s termination decision varied from its stated, albeit subjective selection criteria.”¹ The company with the objective-looking tests departed from them; the company that never claimed objectivity did not. That is a finding about internal consistency and nothing else. It is not a finding that Chrysler’s dealers fared better, and the audit’s conclusions say the opposite in the same breath: “Chrysler’s process did not even include an opportunity for dealerships to appeal the termination decision.” A criterion outside the published list is visible too: in one market the audit sets out, the terminated dealership with the highest Minimum Sales Responsibility score in the group — 442 percent, on 486 new vehicles sold in 2008 — was cut because it was “blocking the addition of Jeep franchises in three other sales areas.”¹

§8What the audit concluded

The audit’s judgment is directed less at the criteria than at the speed and at Treasury’s decision not to watch:

Although the restructuring of GM and Chrysler inevitably required an overall reduction in their own workforces (and the termination of a certain number of poorly performing dealerships), it is not at all clear that the greatly accelerated pace of the dealership closings during one of the most severe economic downturns in our Nation’s history was either necessary for the sake of the companies’ economic survival or prudent for the sake of the Nation’s economic recovery.

SIGTARP-10-008, p. 29

On oversight it is narrower and sharper: “In the absence of effective oversight, GM purportedly employed objective criteria but then deviated from such criteria, making termination decisions with little or no transparency and making a review of many of these decisions impossible.”¹ Deferring to the manufacturers on what the criteria should be was, the audit says, “certainly understandable”; the decision not to monitor how they were applied is what it calls “far more questionable.”

Three further findings bear on whether the criteria were doing the work claimed for them. The acceleration “was not done with any explicit cost savings to the manufacturers in mind,” and the estimates produced afterwards did not survive contact with each other: Chrysler put the saving at $45,500 per terminated dealership, General Motors at $1.1 million, a spread the audit says “casts doubt on their credibility.” Once arbitration became law, General Motors announced on 2010-03-05 that it would send letters of intent to 666 dealers that had filed claims and Chrysler announced on 2010-03-26 that it would offer 50. And Ford, restructuring without federal ownership, “is closing dealerships at a rate similar to that in GM’s original restructuring plan which was rejected by Treasury.”¹

§9Considerations

What the audit establishes is narrow and documented: that General Motors’ published criteria were not applied uniformly, in both directions and at every threshold the auditors tested; that the company did not record its reasons at the time and reconstructed them from staff recollection months later; that it was missing at least one scoring input for 308 dealerships; and that its appeals process had no written criteria for review or reversal and no documented reasons for either.

What it does not establish, and what this page therefore does not say, is that any particular dealership was closed for any particular improper reason. The missing documentation is what made it “impossible in many cases” to determine why the criteria were departed from, and an absence of records supports no inference about which way a specific decision went. Two further limits are the audit’s own: it records “broad consensus that GM and Chrysler generally needed to decrease the number of their dealerships,” its objection being to the pace rather than the reduction, and it records that the experts it consulted disagreed with each other about where and how fast the cuts should fall.¹

§10Still open

Between the terminations and the audit, Congress wrote the criteria into a statute. Section 747 of the Consolidated Appropriations Act, 2010 (P.L. 111-117, 123 Stat. 3219), signed 2009-12-16, gave each covered dealership a right to binding arbitration and imposed a disclosure duty on the manufacturers first. Within 30 days of enactment — by 2010-01-15 — a covered manufacturer had to provide each covered dealership “a summary of the terms and the rights accorded under this section to a covered dealership and the specific criteria pursuant to which such dealer was terminated, was not renewed, or was not assumed and assigned to a covered manufacturer.” Section 747(d) then made “the covered dealership’s performance in relation to the criteria used by the covered manufacturer to terminate, not renew, not assume or not assign the covered dealership’s franchise agreement” the sixth of seven factors an arbitrator had to weigh.⁵

So a statute required each manufacturer to state the specific criteria behind each termination, and made performance against those criteria a statutory factor, at a point when a federal audit had already found that one of the two had not documented why it applied its criteria as it did in a large number of cases. Whether criteria reconstructed after the fact can satisfy a duty to state them is a question this record raises and does not answer, and it is the factual ground under the arbitrations — 1,575 cases filed out of 2,789 eligible dealerships, on the count the American Arbitration Association reported to Congress in November 2010 — and under the litigation that followed them.⁶

What is not open is the mechanism. The election window under Section 747 closed 40 days after enactment, on 2010-01-25, and the section creates no right exercisable now. A 2009 criteria letter is a historical record in 2026, not a live claim.

▪ Ceased · GM wind-down selection and appeals · 2009-05-14 → 2009-08-07
▪ Expired · § 747 arbitration election window · closed 2010-01-25

Corrections to this page are logged at /corrections/ and can be reported to editor@hometownautodealers.org.