STATE OF FLORIDA
DIVISION OF ADMINISTRATIVE HEARINGS
CHEVROLET WORLD, INC. and ) GENERAL MOTORS CORPORATION, )
)
Petitioners, )
)
vs. ) CASE NO. 86-3617
) CENTURY CHEVROLET, INC., and ) DEPARTMENT OF HIGHWAY SAFETY ) AND MOTOR VEHICLES, )
)
Respondents. )
)
RECOMMENDED ORDER
Pursuant to Notice, this cause was heard by Linda M. Rigot, the assigned Hearing Officer from the Division of Administrative Hearings, on March 16-20, 1987, in Orlando, Florida.
The following appearances were entered:
Representing Chevrolet World, Inc.: Jacqueline R. Griffin, Attorney at Law PEIRSOL, BOROUGHS, GRIMM, BENNETT & GRIFFIN
201 East Pine Street Orlando, Florida 32801
Representing General Motors Corporation: Dean Bunch, Attorney at Law
RUMBERGER, KIRK, CALDWELL, CABANISS & BURKE
Suite 900, 101 North Monroe Street
Tallahassee, Florida 32301 and
Michael J. Robinson, Attorney at Law Office of the General Counsel General Motors Corporation
New Center One Building 3031 West Grand Boulevard Detroit, Michigan 48232
Representing Century Chevrolet, Inc.:
D. Frank Davis, Attorney at Law Joseph W. Letzer, Attorney at Law BURR & FORMAN
3000 SouthTrust Tower
Birmingham, Alabama 35203
Representing Department of Highway Safety and Motor Vehicles:
No Appearance
This proceeding began when Petitioner Chevrolet World, Inc., filed an application with Respondent Department of Highway Safety and Motor Vehicles for licensure as a dealer of Chevrolet vehicles in Orlando, Florida. Upon notice of the application having been filed, Respondent Century Chevrolet, Inc., filed a letter of protest contesting the application. The matter was forwarded by the Respondent Department to the Division of Administrative Hearings with the request that a Hearing Officer be assigned to conduct a formal hearing under the terms of Section 320.642, Florida Statutes. Pursuant to the terms of that statute, the manufacturer seeking to establish the additional dealer, in this case General Motors Corporation, is a necessary party since it is assigned the burden of proof. The issue for determination herein is whether the existing Chevrolet dealers serving the Orlando area are providing inadequate representation for Chevrolet. Petitioner General Motors Corporation presented the testimony of Larry Hall; Ronald S. Rubin, Ph.D.; Marvin E. Beaupre, Jr.; and James A. Anderson. Additionally, General Motor's Exhibits numbered 1-78, 80 and
81 were admitted in evidence. As set forth in the Pre-Trial Stipulation, Petitioner Chevrolet World, Inc., adopted the case presented by Petitioner General Motors Corporation.
Respondent Century Chevrolet, Inc., presented the testimony of Marvin E. Beaupre, Jr.; John E. Larkin; William Lamb; Lyman E. Ostlund, Ph.D.; and John P. Matthews, Ph.D. Additionally, Century's Exhibits numbered 1, 2, 4, 9-11, 15-18, 20-22, and 24 were admitted in evidence; however, Century's Exhibits numbered
, 19, 23, and 25-28 for identification only were preferred after they were specifically rejected as evidence in this cause.
Respondent Department of Highway Safety and Motor Vehicles did not participate in this proceeding and, accordingly, offered no evidence.
The transcript of the hearing, consisting of eight volumes, was filed on April 20, 1987. Proposed findings of fact in the form of proposed recommended orders were filed by Petitioners and by Respondent Century Chevrolet, Inc., on May 5, 1987. Rulings on each proposed finding of fact can be found in the Appendix to this Recommended Order.
FINDINGS OF FACT
On July 31, 1986, Petitioner Chevrolet World, Inc. (hereinafter "Chevy World"), filed an application for licensure with the Department of Highway Safety and Motor Vehicles (hereinafter "Department"), seeking to become a licensed Chevrolet dealer in Orlando, Orange County, Florida. Chevy World is owned by Don Mealey, who also owns another Chevrolet dealership in Orlando known as Don Mealey Chevrolet, Inc. The new dealership would be located in a development known as "Lee Vista Center," immediately north of the Orlando International Airport.
The Jacksonville Zone for the Chevrolet Division of General Motors (hereinafter "GM"), is divided by GM into MDAs (Multiple Dealer Areas) and SDAs (Single Dealer Areas).
There are five MDAs in the Jacksonville Zone: West Palm Beach, Miami, Orlando, Jacksonville, and Tampa.
The Orlando MDA consists of all of Orange and Seminole Counties plus two census tracts in the extreme southern portion of Volusia County which are immediately adjacent to Seminole County.
There are four Chevrolet dealers currently operating within the boundaries of the Orlando MDA as established by GM: (a) Roger Holler Chevrolet in Winter Park, (b) Don Mealey Chevrolet in Orlando (owned by the owner of the applicant in this case), (c) Century Chevrolet (hereinafter "Century"), in Winter Garden, and (d) Ken Rummel Chevrolet in Sanford.
Within each MDA, GM further subdivides the area into AGSSAs (areas of geographic sales and service advantage) and places one dealer in each AGSSA. GM designates five (5) AGSSAs for the Orlando MDA. The open point is located in AGSSA 5. Theoretically, each dealer is supposed to have the greatest opportunity for sales and service in his own AGSSA when compared to other Chevrolet dealers.
Although GM lists only four Chevrolet dealers in the Orlando MDA, it is undisputed that more than those four dealers advertise in and serve the Orlando market area. For example, the Southern Bell telephone directory yellow pages for Orlando lists six Chevrolet dealers who advertise in Orlando, and this list does not include Ken Rummel's Chevrolet dealership, which GM does include in the Orlando market area.
Additionally, John Larkin, the owner and president of Century and one who is familiar with the competition among dealers in Orlando, testified that there are ten Chevrolet dealerships which advertise in and serve the Orlando market. These include Bill Seidel Chevrolet in Clermont, Florida; Vanganoway Chevrolet in Eustis, Florida; Cecil Clark Chevrolet in Clermont, Florida; Fred Bonson Chevrolet in Deland, Florida; Ken Rummel Chevrolet in Sanford, Florida; Roger Holler Chevrolet in Winter Park, Florida; Don Mealey Chevrolet in Orlando, Florida; Burchfield Chevrolet in St. Cloud, Florida; Sterling Chevrolet in Kissimmee, Florida; and Century in Winter Garden, Florida. This evidence was not disputed by GM. Given this large number of dealers which impact the Orlando market, the market in Orlando for Chevrolet cars is highly competitive.
The critical substantive question in this case is whether the existing dealers are adequately serving the market. A proper definition of the market is obviously important. GM failed to prove that its MDA properly defines the relevant market area for this proceeding. It offered no explanation as to why the relevant market is not the Orlando statistical metropolitan area which encompasses a three-county area: Seminole, Orange, and Osceola Counties.
In fact, one of GM's own witnesses, Dr. Ronald Rubin, testified that the boundaries of the Orlando statistical metropolitan area would be appropriate parameters for the Orlando market area. Dr. Rubin could not state that the Orlando MDA accurately defines or describes the Orlando market or that AGSSA 5 would actually be the marketing area or community with respect to a new dealer located therein.
Given the facts that GM offered no explanation other than its own discretion why its MDA should be smaller in area than the Orlando statistical metropolitan area and that many Chevrolet dealers outside the boundaries of GM's MDA have a significant impact on Chevrolet sales in Orlando, the relevant territory or community is the Orlando statistical metropolitan area and not only the Orlando MDA.
The smaller artificially-configured market areas established by GM (Orlando MDA and AGSSA 5) are specifically rejected because they are only a portion of the market area impacted by the Chevrolet dealers in the Orlando
statistical metropolitan area and because there was no competent showing by GM that the Orlando MDA and AGSSA 5 bear scrutiny. While an MDA or AGSSA might in an appropriate case be the proper area of scrutiny, GM has not carried its burden of proving that this is such an appropriate case.
Because the relevant community or territory is the Orlando statistical metropolitan area, the sales performances and penetration rates for the Chevrolet dealers in Kissimmee, St. Cloud, Eustis, Deland, and Clermont, in addition to the four dealers located in the Orlando MDA, must all be considered in determining whether the existing dealers in the Orlando statistical metropolitan area are adequately representing Chevrolet in that market. GM presented no credible evidence as to the penetration rates in the Orlando statistical metropolitan area by all the dealers who compete there. Consequently, GM has not carried its burden of proving that the existing Chevrolet dealers are not providing adequate representation for GM in the relevant market area.
GM's evidence that there were 552 new car in-sells within the Orlando MDA by outside Chevrolet dealers supports a conclusion that the MDA is not the real or relevant market; it does not prove the propositions for which it is cited--(a) that the Chevrolet dealers within the Orlando MDA are not adequately serving the market and/or (b) that the customers are or were dissatisfied with those dealers.
There is no evidence in the record to support a conclusion that these
552 customers, or any one of them, were displeased with the Orlando MDA dealers; for example, there was no polling data or other information from those customers offered by GM which would have indicated such displeasure.
The number of in-sells into the MDA is more likely the result of dealers within the metropolitan statistical area but outside GM's artificial MDA (e.g., the Chevrolet dealer in Kissimmee, Florida) competing with MDA dealers.
Thus, the number of in-sells is probative of a conclusion that GM's artificially- drawn MDA boundaries are skewed and are not indicative of the
relevant territory or community for purposes of this proceeding, thereby casting doubt on substantially all of GM's studies and conclusions that penetration rates in the Orlando market area are disproportionately low.
Additionally, as proven by the uncontroverted testimony of Dr. Ostlund, the AGSSAs established by GM for the Orlando MDA were not drawn in accordance with GM's own established guidelines.
First, AGSSAs are supposed to be determined by and drawn in accordance with natural and man-made barriers or boundaries; however, none of the AGSSAs in the Orlando MDA are divided by or established in accordance with natural or man- made boundaries or barriers.
Second, the AGSSAs established by GM are supposed to account for and be established in accordance with shopping patterns; however, there is no analysis in the AGSSA documents for the Orlando MDA concerning non-automotive shopping patterns, and the information concerning automotive shopping patterns does not support the AGSSA lines drawn by GM.
Third, contrary to GM's own guidelines, there is no evidence that any of the AGSSA boundaries have been drawn with the factor of traffic flow taken into account.
Moreover, the evidence presented by GM clearly indicates that the statistical relationship between the residence of each dealer's customers and the AGSSA boundaries is weak. In other words, each dealer is located within an AGSSA and, according to GM's AGSSA guidelines, is expected primarily to impact the sales area within that AGSSA; however, the actual sales figures for each of the dealers reveal little if any correlation between the location of its customers and the AGSSA boundaries. Again, the appropriateness of the AGSSAs is contraindicated.
An AGSSA could, of course, in an appropriate case be a submarket appropriate for scrutiny. The burden of showing the appropriateness is on GM. In a case such as this where GM not only fails to meet its burden but the evidence shows that GM's own standards for drawing AGSSAs have not been followed, it cannot be concluded that the use of any particular AGSSA suggested by GM is appropriate. Consequently, GM's calculations and conclusions, premised upon these AGSSAs, are not supported by credible evidence.
Aside from this failure by GM to prove that it properly analyzed the relevant market area, GM failed to take other factors into account in assessing the adequacy of the existing dealers' representation. One of these factors was the closing of one of the four Orlando MDA dealers for most of 1985.
From February 1, 1985 to November 1, 1985, Chevrolet was without a dealership in Sanford, Florida, because Ken Rummel was in the process of moving the Chevrolet dealership located there to a new location.
While Mr. Anderson admitted that he was aware of Ken Rummel's closure and the probable negative effect this closure would have on market penetration efficiency, his studies did not account for this fact.
As testified by Dr. Matthews, a primary flaw in GM's use of national penetration rates to conclude that inadequacy exists in the Orlando market is the lack of any comparative analysis of the different MDAs to determine whether factors peculiar to the Jacksonville Zone or the Orlando MDA are causing the alleged shortfall in penetration rates.
The closure of Ken Rummel Chevrolet for most of 1985 illustrates that a comparative analysis of markets was not performed by Mr. Anderson or GM. The closure of one of the four Chevrolet dealers in the Orlando MDA for most of 1985 would have a significant negative impact on market penetration for that year, but GM apparently made no correction or adjustment in its analysis to take this important fact into account.
Another factor which GM did not take into account before asserting that poor penetration efficiency of the Orlando MDA dealers indicated that an additional Chevrolet dealer was needed in the Orlando market is the extensive leasing of new Chevrolets which occurs in Orlando and the heavy sales of "brass- hat" cars by the Chevrolet dealers and current-model, low-mileage fleet cars by fleet rental companies.
As defined by GM for sales and registration purposes, a fleet customer is any person or company who purchases ten or more cars or trucks of any brand anywhere in the country during a twelve-month time period.
In 1985, 29.9 percent of industry car registrations in the Jacksonville Zone were fleet registrations. The national rate was 17 percent.
For Chevrolet, 49 percent of all car registrations in the Jacksonville Zone in 1985 were fleet registrations, while nationally the rate for Chevrolet was only 21.8 percent. Stated differently, the dealers in the five MDAs in the Jacksonville Zone (3 percent of the total of all MDAs) accounted for over sixteen percent of all fleet sales in the United States. It is uncontroverted that the Orlando market has significantly higher levels of fleet sales than other parts of the country.
GM has itself concluded that the extraordinary number of passenger car and truck fleet sales into the Orlando market area has a negative effect on passenger car and truck sales and registration effectiveness. In spite of this, none of the studies presented by GM took this into consideration. GM cannot properly do what it seeks to do here, that is, compare Orlando to national or regional or so called "typical" markets while admitting this difference without accounting for the impact of these fleet sales in the market.
Additionally, many of these fleet companies who purchase new fleet cars from the dealers and manufacturers (e.g., Avis, Hertz, National, Budget, etc.) in turn sell those cars to consumers in the Orlando area. Typically, these rental cars sold by the large fleet companies are current-model, low- mileage cars, many of which are Chevrolets.
Because there is no difference in the profile of a typical new car customer and a customer who will buy a current- model, low-mileage car from a fleet company, the sale of these fleet rental cars in Orlando diminishes the market demand for new retail Chevrolets.
The reason for this deleterious effect on market demand for retail Chevrolets is that the current-model, low- mileage rental units can be sold by the rental companies for lower prices.
This fact was admitted by Jacksonville Zone Manager Hall when he testified that sales of current model rental cars by rental companies has an adverse impact on new car sales.
As further evidence that the heavy concentration of fleet sales by Chevrolet dealers and sales of current-model, low-mileage rental cars by fleet companies negatively affect new car retail penetration rates, Century introduced a letter from Don Mealey to R. J. Bresnahan (the former Jacksonville Zone Manager for Chevrolet) dated May 29, 1984. In this letter, Don Mealey, an existing dealer in the Orlando MDA and the owner of the applicant herein, delineated the problems of market penetration caused by fleet companies selling current-model, low-mileage rental cars. Mr. Mealey stated in his letter that prospective new car customers in Orlando have a "multitude of options available to them. They can shop a franchised new car dealer or they may choose from one of many outlets retailing low mileage, current models at a price considerably less than normal dealer invoice. The net effect of this second level system on the total market is incalculable." This observation by Mr. Mealey is consistent with GM's own admissions and the testimony of Century's officials during this hearing.
Mr. Mealey further highlighted in his letter the large number of Chevrolet cars operating in Orlando which are registered elsewhere but which do not appear on the sales and registration reports for Orlando. "[C]ertain major rental companies purchase or lease Chevrolet's [sic] which are registered in Dade, Broward, or other Florida Counties. These vehicles are rented in Orlando, serviced in Orlando, and most often are sold in Orlando but do not appear as new
vehicle sales or registrations. If these phantom units appeared on the R.L. Polk report, the overall Chevrolet market share would improve Again, Mr.
Mealey's observations are consistent with other evidence adduced in this case.
Despite these admissions by Petitioner GM and Mealey (an agent of Petitioner Chevy World), regarding fleet sales, Mr. Anderson denied that the large volume of fleet sales in the Orlando market has had a negative impact on the retail car penetration rates for Chevrolet dealers in the Orlando MDA. He offered no evidence in support of his conclusion. Mr. Anderson's position is unsupported by the facts in the record and belies common sense.
Mr. Anderson's and GM's failure to properly account for and to factor into their analysis the impact of fleet sales on Chevrolet penetration rates discredits their contention that the present dealers are inadequately representing the market. Initial fleet sales impact retail penetration, as do resales of current model cars. Potential new car purchasers will generally not view current-model, low-mileage rental cars sold by fleet companies as distinguishable from new cars offered for sale by a Chevrolet dealer.
Perceiving these two groups of automobiles as essentially similar, a customer in the new car market would ignore industry labelling differences between these groups of cars.
In short, as established through GM's own witnesses, the Orlando market area, given its heavy concentration of fleet sales, is significantly different from the national market viewed as a whole.
This heavy influx of fleet sales into the Orlando market area is strong evidence why penetration rates in the artificial Orlando MDA are lower than the national average.
The typical customer is most interested in low prices. To attract these customers, Century advertises for sale "brass-hat" cars, which are vehicles used previously by GM officials and sold at GM auctions.
There is no difference in the profile of a typical new car customer and a customer who will, and often does, buy a brass-hat vehicle.
Brass-hat vehicles are purchased by Century at GM auctions to serve as price leaders in advertisements and sticker sales to encourage customers to visit Century's facilities.
During 1986, Century sold an average of forty-three brass-hat vehicles per month, a large portion of which were Chevrolets.
The local Chevrolet dealers in the Orlando market buy a large number of the brass-hat vehicles sold at the auctions conducted by GM in Orlando.
Brass-hat vehicles, regardless of the model or mileage, are not treated as new cars by GM; thus, Chevrolet dealers, such as Century, do not get credit from GM for a new car sale when a customer purchases a brass-hat vehicle.
Moreover, the Chevrolet dealers do not receive credit with respect to
R. L. Polk registrations for the sale or registration of brass-hat vehicles.
As is the case with current-model, low-mileage rental cars sold by fleet companies, it is unlikely that, if the physical attributes of the brass-
hat vehicle are the same or substantially similar to the attributes of a new car, a potential customer will view those two classes of cars differently.
Consequently, brass-hat vehicles are competitive with new retail cars in the marketplace and decrease the sales rates for the new cars because the brass-hat vehicles are priced lower.
GM did not account for the significant levels of brass-hat vehicles sold in the Orlando market. No study was performed by GM to analyze what effects brass-hat sales had on new car sales effectiveness and penetration rates. Accordingly, GM's studies and conclusions are flawed insofar as they attempt to support a finding that the present Chevrolet dealers are providing inadequate representation in the Orlando market.
Century and other Chevrolet dealers lease a large number of new cars. The leasing of new Chevrolets in Orlando diminishes the demand for the sales of new Chevrolets because the same profile of customers is involved, and at times leasing has financial advantages over purchasing.
GM does not regard the lease of a new car as a retail sale.
Again, Mr. Anderson's studies are flawed because they did not assess the negative effects which extensive leasing (as occurs in the Orlando market) has on penetration rates for new Chevrolet cars.
In addition to the failure of GM to analyze the negative impact which Ken Rummel's closure in 1985 and the heavy concentration of leasing, fleet sales, brass-hat sales, and sales of current-model, low-mileage cars by fleet companies had on Chevrolet retail penetration rates in Orlando, the studies actually performed by GM were statistically and quantitatively suspect in many respects. Because the thrust of GM's proof focused on Mr. Anderson's five-point test, the five-point test must be analyzed in order to determine whether it is an accurate, reliable, and statistically proper analysis of the Orlando market, in general, and the adequacy of the existing dealers' representation, in particular.
First, nearly every witness agreed that the mere fact that new car registrations in a particular market area fall short of national or Zone registration rates does not indicate, without further examination, that the existing dealers are not providing adequate representation in that market area.
In determining whether the Orlando market is being adequately represented, Mr. Anderson testified that he would first ascertain whether the market penetration for Chevrolet in the Orlando MDA exceeded or fell below Zone and national retail penetration rates. Then, Mr. Anderson employs a five-point test to determine what is a reasonable level of market penetration for the particular market area. The first two steps in this five- point test entail a comparison of the demographic characteristics in terms of age and household income of the Orlando market area vis-a-vis the nation as a whole. The third step in the five- point test is an analysis of product popularity, i.e., an analysis which seeks to determine whether the particular line- make is popular in a certain market area. The fourth step is an assessment of the market to determine whether national or Zone penetration rates are being achieved or exceeded in certain defined areas within that market. Finally, the penetration rates for the areas surrounding the market in question are analyzed to determine whether national penetration rates are being achieved in these outlying areas.
This five-point test for assessing local market penetration effectiveness is flawed in several respects and thus is an unreliable and inaccurate analysis.
While GM did present demographic evidence that Orlando, compared to national statistics, was underrepresented in the 0-15 and 65-over age categories, there was no proof offered by GM that this population age data proved that Chevrolet sales potential was in any way enhanced.
There were no Chevrolet customer age profiles submitted by GM, or any quantitative analyses performed by GM indicating that Chevrolet penetration rates varied in accordance with the age of the customers. Accordingly, GM's demographic evidence concerning age distributions in Orlando vis-a-vis the nation is statistically unpersuasive and proves nothing.
The same finding is compelled with respect to the income levels in the Orlando market area. There was no evidence offered by GM that income levels relate in any way to Chevrolet retail penetration or can be analyzed as a determinant of interbrand competition.
GM's use of product popularity statistics is not persuasive because it also failed to consider line-make popularity by model. Without such additional studies it cannot be accurately concluded that the Orlando market area can be compared, in a statistically acceptable manner, to the national market in terms of product popularity.
The fourth factors analyzed by Mr. Anderson--whether certain census tracts within the Orlando MDA have retail Chevrolet penetration rates which meet or exceed the national average--is likewise a statistically unacceptable test for adequacy of representation or for the use-of national figures to demonstrate inadequacy. As established by the uncontroverted testimony of Drs. Ostlund and Matthews, this condition will undoubtedly exist in any market.
Given the small total penetration numbers present in any census tract, it is not unrealistic to expect that some census tracts might have Chevrolet penetration rates which exceed a national average.
Random penetration rates of one or two cars could mean the difference between exceeding or falling below national rates in a given census tract. Thus, the mere fact that one or more tracts have Chevrolet penetration rates above the national average does little to address the question whether there is adequate representation throughout a defined market area encompassing numerous census tracts.
Stated otherwise, it would not be unusual to find census tracts in any defined market which exceed or fall below national penetration averages. Customers living in different census tracts are not homogeneous in terms of age, income, line- make preference, etc. Thus, the fact that there may be some census tract areas within a defined market area where penetration rates exceed some national penetration average is not a proper basis to infer that the entire market area should be able to achieve or exceed the same penetration rates.
The fifth and final factor analyzed by Mr. Anderson and GM to support their assertion that the existing Chevrolet dealers in the artificial Orlando MDA are not providing adequate representation because they are not achieving national retail penetration rates is whether the market area outside the MDA but inside the Orlando metropolitan statistical area has higher penetration rates
than the rates in the MDA. GM has concluded that higher penetration rates have been achieved in the areas outside the MDA boundaries. The inference which GM seeks to draw from this fact is that the Chevrolet dealers, located within the MDA, are not making adequate sales. The flaw in this argument is that such a shortfall will ordinarily (if not always) exist where the MDA is not coextensive with the SMSA, irrespective of whether the existing dealers within the MDA are adequately representing Chevrolet in the market. The reason for the discrepancy in the penetration rates is twofold.
First, the SMSA is generally larger than the MDA and encompasses more rural and semi-rural areas. Historically, customers from rural or semi-rural areas have favored traditional domestic line-makes like Chevrolet over other line-makes.
Second, import manufacturers do not have as extensive a dealer organization as traditional domestic manufacturers such as GM and have tended to concentrate their sales efforts in larger urban areas, thereby increasing the options of urban customers and decreasing the penetration rates for domestic line-makes such as Chevrolet. Consequently, the fact that Chevrolet may have achieved greater market penetration within the Orlando SMSA but outside the Orlando MDA proves little if anything about the adequacy of representation of Chevrolet by the existing dealers.
In addition to the flaws existing in each of Mr. Anderson's five factors, the entire five-point test is significantly flawed and statistically unreliable. Curiously, this is not disputed by GM. The question of-the propriety of the five-point test arose as follows. Mr. Anderson testified first, saying not that this is a proper test but only that this is the test he used, admitting he knew of no authoritative text which supported its use. Drs. Ostlund and Matthews testified the test was inappropriate to judge the question presented. Mr. Anderson never denied this conclusion, and GM by offering no rebuttal evidence on this point and chose to leave the inappropriateness of the five-point test unrebutted. Thus, no conclusion can be reached except that it is inappropriate.
Moreover, the test was not applied to the other Chevrolet MDAs in the United States to determine the accuracy and relevancy of the test.
Neither in the five-point test nor in any of his other studies did Mr. Anderson address the quantity of dealer facilities that already exist in the Orlando market, a factor which is highly important on the adequacy-of representation issue. Finally, Mr. Anderson did not testify that he had used this test consistently in his testimony in other Florida cases before the Division of Administrative Hearings.
To support its contention that an additional Chevrolet dealer is needed in Orlando, GM placed great emphasis on the fact that the population in Orlando had experienced significant growth over the years. However, the mere fact that Orlando has experienced a population growth is not evidence that an additional Chevrolet dealer is needed in the market, nor does it explain alleged low penetration rates.
In fact, rapid growth in a market indicates that many people are moving to the market area. People who have recently arrived in the market area may have less desire to purchase a new car, or less income available for a new car purchase because of moving and new dwelling expenses. Nowhere in the
testimony of, or the studies and reports prepared by, Mr. Anderson or any other GM witness is it indicated that this condition in the Orlando market was examined.
Indeed, while Mr. Anderson testified that the Orlando area has experienced a high rate of migratory growth, he admitted that he did not analyze the question whether people who recently moved into the Orlando area would more or less likely be in the market for a new car.
Moreover, the mere fact that population in the Orlando area has increased while the number of Chevrolet MDA dealers has remained the same (four) does not necessarily indicate the need for an additional dealer. Mr. Anderson's testimony was offered by GM to prove that the population growth in Orlando over the last fifty (50) years has been so large that four dealers can no longer adequately serve the increased number of customers. However, no analysis was conducted whether there exists a corresponding increase in sales and service facilities and operating investments.
Moreover, an equally plausible inference to be drawn is that there were too many Chevrolet dealers in 1930. No analysis or study was performed by GM to account for this equally plausible inference.
GM also presented evidence concerning the location of the existing dealers and the effectiveness of those dealers to attract customers residing at varying distances from the dealerships, seeking to prove that a dealer's market penetration success bears an inverse relationship with the distance a potential customer must drive to reach that dealership. However, GM's analysis is flawed in several respects. First, this analysis does not take into account the specific geographical location of a dealer in the Orlando market area, and particularly with respect to whether the dealer location is close to areas where few or no customers reside. For example, a dealer like Century is near farmland and a large lake, thereby reducing the possibility that cars will be sold to or registered by customers living in a geographical area closer to that dealership.
Furthermore, all MDA dealers will be partial servants of a geographical area because there will always be other same line-make dealers serving that market. Thus, the analysis used by GM does not address the question whether the existing dealers are adequately serving the market.
Finally, GM's position in this proceeding that the location of a dealership is not relevant leaves it little room to complain of locations.
GM .offered no persuasive evidence that any traffic patterns or congestion in the Orlando market area indicated the need for an additional Chevrolet dealer.
In fact, the evidence was to the contrary; there is no significant traffic congestion in the vicinity of the proposed new dealer point location to justify the addition of a dealer.
In any event, traffic congestion is generally not an important factor in determining whether an additional dealership is needed in a market because traffic congestion usually occurs during rush hours; typically, potential customers do not shop for new cars during rush hour.
GM presented evidence in the form of graphs and charts which allegedly established that the dealers in the Orlando MDA must service a higher number of
potential customers than dealers in other Jacksonville Zone MDA's. The inference sought to be established by GM from this data is that there are too few Chevrolet dealers in the Orlando MDA to properly service the larger number of potential customers. However, GM's evidence does not support this inference. First, this analyis is fundamentally flawed because it is premised upon there being only four Chevrolet dealers in the Orlando market area; the evidence presented in this case establishes that ten Chevrolet dealers compete heavily in the Orlando area market.
Second, the population/dealer data actually supports Century's position. According to Dr. Matthews' statistically correct computations, each dealer in the Orlando MDA services a population of 206,542; however, the average population serviced by the dealers in the entire Jacksonville Zone is 209,691. Consequently, the Orlando MDA dealers service less population than the average of the other Chevrolet dealers in the Jacksonville Zone.
This fact obviously does not compel the conclusion that a new dealer is needed in the Orlando MDA; in fact, it is strong evidence that the present dealers are adequate in number to represent the Orlando market.
Mr. Anderson's conclusions regarding registration rates per dealer in the Jacksonville Zone suffer from similar statistical flaws. After removing the Pensacola registration figures from this study and adjusting the Tampa MDA figures to accurately reflect the number of existing Chevrolet dealers, the average number of registrations per dealer point is 9,658. Thus, because the Orlando MDA Chevrolet dealers' registrations per dealer point (8,587) are below average Jacksonville Zone figures, this study also does not prove the need for an additional Chevrolet dealer in Orlando.
If another Chevrolet dealer were added in Orlando, the number of registrations per dealer in the Orlando MDA would fall to 6,869, or only 72 percent of the average registrations per dealer in the Jacksonville Zone. Therefore, the studies performed by Mr. Anderson and GM reveal that an additional dealer in Orlando is contraindicated.
In this proceeding, Mr. Anderson admitted that the total registrations for the Orlando MDA might vary by nearly twenty percent from the reported sales by the Orlando MDA Chevrolet dealers.
Mr. Anderson conceded that this variation could have resulted from an error in R. L. Polk statistics.
At no time has Mr. Anderson investigated to determine whether R. L. Polk actually used statistically proper and reliable procedures to gather and verify the registration data. Additionally, Anderson testified that he has personally found R. L. Polk data to be unreliable.
Accordingly, the R. L. Polk registration reports and computations utilized by GM throughout its studies and reports have not been proven to be reliable.
GM sought to establish that Chevrolet truck penetration rates in the Orlando MDA fell below national penetration rates. From this premise partially flowed GM's argument that an additional dealer was needed to increase Chevrolet penetration rates in the Orlando market area. However, as illustrated by Dr. Ostlund, the use of truck penetration statistics is not probative of inadequacy- of the existing dealers. Commercial customers of trucks, even those purchasing
less than ten units per year (thereby not qualifying as a fleet customer by GM's standards) are a substantial portion of the truck market. Generally, these multi-unit, commercial truck customers are not concerned about the location of a dealer. Dr. Ostlund's testimony on this issue was not controverted by GM.
The last full year from which GM evaluated sales and registration data was 1985, the year during which one Orlando MDA dealer was closed. GM only used nine-month data in 1986.
However, GM's use of nine-month data for calendar year 1986 did not take into account any seasonal factors which could indicate that year-end sales figures would be significantly different. Proper statistical tools, such as regression analysis and exponential smoothing, could have been used to properly interpret the nine-month data and to extrapolate that data over a twelve-month time frame; however, no such commonly accepted statistical devices were used by Mr. Anderson or GM, thereby raising serious doubts as to the statistical accuracy of their conclusions. Moreover, no showing was made by GM that Mr. Anderson could provide competent testimony on such statistical techniques.
Furthermore, it does not appear that Mr. Anderson or GM even considered the effect which any seasonal factors had on market penetration in 1986.
There was no evidence offered by GM that seasonality patterns of Chevrolet sales were stable from year to year, particularly 1985 to 1986. Therefore, GM's failure to analyze seasonality factors renders unworthy its interpolation of 1985 data to determine annualized 1986 sales figures and penetration rates.
In many of its comparative evaluations of the Orlando market and the Orlando dealers' sales and penetration effectiveness, GM utilized data obtained from the Pensacola MDA which is not a part of the Jacksonville Zone. However, it is clear that neither Mr. Anderson nor anyone else at GM performed any studies to determine whether the Pensacola market area was comparable to the Orlando market area or other multi-dealer markets in the Jacksonville Zone. The only justification for utilizing Pensacola MDA data was that Pensacola was in the State of Florida.
Dr. Matthews, based on GM's own evidence, concluded that the inclusion of Pensacola in any of the studies was inappropriate. This was not disputed by GM.
In the absence of any showing by GM that Pensacola is a comparable market to Orlando, any conclusions or inferences made by GM which in whole or in part are based upon Pensacola MDA data are unsound.
GM offered no evidence that the sales and service facilities of the existing Chevrolet dealers in the Orlando market area are inadequate to represent Chevrolet in that market. In fact, Mr. Anderson admitted that he did not evaluate the sales and service facilities of any of the Orlando MDA Chevrolet dealers prior to concluding that they were not providing adequate representation.
GM premised its entire case on a comparison between Chevrolet penetration rates in Orlando and Chevrolet penetration rates nationwide and in the Jacksonville Zone. The flaw in GM's presentation was the complete failure
to determine whether the Orlando market is a typical market area from which conclusions as to low penetration rates and inadequacy of representation could be so simplistically drawn.
More significantly, GM presented no evidence that other manufacturers, domestic or import, were achieving or exceeding their national or Zone penetration rates in the Orlando market. The absence of any evidence concerning penetration rates for other line-makes is particularly damaging to GM's contention that the Orlando area Chevrolet dealers should be achieving higher penetration rates.
GM ranked each MDA in terms of market penetration for passenger cars and trucks. The Orlando MDA ranked 129 in passenger car sales penetration effectiveness and 130 in truck sales penetration effectiveness. The inference GM wishes to be drawn from these rankings is that the Chevrolet dealers in the Orlando MDA are ineffective in terms of market penetration and, thus, are not providing adequate representation for Chevrolet in Orlando.
However, GM's rank-ordering of the 157 MDAs in the United States in terms of market penetration is statistically unsound because there is no explanation offered for the penetration discrepancies.
There was no statistical analysis performed by Mr. Anderson to explain what factors exist in each of these markets, individually or collectively, to cause the penetration rate variations.
The evidence is undisputed that there was no showing by GM that any of these markets were comparable, and, in fact, no adjustments were made for the lack of comparability. The placement of all Florida and California dealers in the bottom portion of the rankings suggests a comparability problem not addressed by GM.
The evidence established that Century is not optimally located in the Orlando market area. Century is the only automobile dealership of any line-make in Winter Garden, Florida. All other dealerships, domestic and import, have gone out of business due to poor market demand.
In order to attract customers in the Orlando market area, Century purchases advertisements in the Orlando Sentinel and on four Orlando television stations. All told, Century spent in excess of $1.1 million in 1986 on advertising. The amount spent on advertising per retail unit by Century was the highest of any dealers in Orlando and double the Zone average. This high degree of advertising was necessitated by depressed industries and low population figures in Century's assigned area of sales responsibility.
Ultimately, there is no reason to disagree with the conclusion of GM's own evaluation done in the ordinary course of its business. As found by GM itself on its most recent evaluation of Orlando dealers, all dealers rated (Rummel was too new) were found to be doing an effective job with respect to car sales.
Even accepting GM's MDA as the market, during the last full year, Orlando MDA car sales by Chevrolet were 14.1 percent of industry, which exceeded the Zone's rate of 11.9 percent significantly and was statistically indistinguishable from the national average of 14.35 percent.
During the same year (1985) the so-called AGSSA 5 out-performed the MDA by posting a 21 percent share of total. In fact, total penetration by GM in AGSSA 5 during that year was better than in 3 of the 4 other AGSSAs. Chevrolet retail share in AGSSA 5 during the last full year for which statistics were presented was nearly the same as the Zone average of 11.5 percent.
Because the evidence is uncontroverted that dealerships are not appropriately added solely based on truck penetration, (this being a small percentage of overall sales and numbers of dealers and location being relatively unimportant), GM's own evidence of better than adequate car sales by the Orlando MDA Chevrolet dealers for the last full year of dealership operations, along with GM's own admitted findings that all the Orlando MDA Chevrolet dealers were effective with respect to car sales during the 1985 ratings, compels the conclusion that GM's case was simply not proved.
CONCLUSIONS OF LAW
The Division of Administrative Hearings has jurisdiction over the subject matter hereof and the parties hereto. Section 120.57(1), Florida Statutes.
Section 320.642, Florida Statutes, establishes the standard for granting or denying an application for a motor vehicle dealer license and provides as follows:
The department shall deny an application for a motor vehicle dealer license in any community or territory where the licensee's presently licensed franchised motor vehicle dealer or dealers have complied with licensee's agreements and are providing adequate representation in the community or territory for such licensee. The burden of proof in showing inadequate representation shall be on the licensee.
The parties have stipulated that Chevrolet's present dealers in the Orlando market have complied with their franchise agreements. Therefore, the sole issue is whether those dealers are providing adequate representation in the community or territory involved. Petitioners have failed to meet their burden of proof.
First, GM's studies and analyses were based upon factually unsupported and therefore improperly defined market areas: the Orlando MDA and AGSSA 5. The Orlando MDA considers only four of the ten Chevrolet dealers impacting the Orlando market. Additionally, GM's own guidelines for drawing AGSSA boundaries were not complied with in regard to AGSSA 5. Lastly, Century's witnesses testified that Don Mealey who owns both an existing dealership in the Orlando MDA and the applicant herein alone sells forty percent of the vehicles sold in AGSSA 5, and AGSSA 5 is in the middle when viewed in relation to the other AGSSAs in the Orlando MDA, which alone disposes of GM's contention that there is inadequate representation in AGSSA 5. GM offered no explanation for its failure to use the statistical metropolitan area by which its data is collected or for using a smaller market area for Orlando. The evidence in this case establishes that for purposes of Section 320.642, Florida Statutes, the relevant territory or community under consideration Is the Orlando statistical
metropolitan area. Consequently, GM's failure to adequately assess the existing dealers' performance in this broader, more appropriate sales area undermines the legal efficacy of its contentions.
Second, GM's analysis of market penetration and adequacy of representation failed to take several important factors peculiar to the Orlando market and the existing dealers into account. For example, no adjustments were made to GM's market penetration analysis to account for Ken Rummel Chevrolet's closure during 1985; the annualized 1986 data does not factor in seasonal fluctuations; and the heavy concentration of fleet sales in Orlando, the large number of sales of brass-hat vehicles (which GM considers new for financing but used for statistics as opposed to demonstrators which are new for statistics) and leasing in Orlando which do not appear in penetration statistics, and the large "secondary market" which has developed in Orlando due to sales of current- model, low mileage rental cars to customers who would otherwise enter the new car market were not considered.
Third, the studies performed by GM to prove that an additional dealer is needed in the Orlando area, particularly Mr. Anderson's novel five-point test, are statistically and quantitatively invalid and unreliable. No explanation was offered by GM as to how the factors in Anderson's five-point test lead to the conclusion that it is inadequately represented or the conclusion that it is appropriate to compare Orlando to the Zone or the nation rather than a comparable market to determine adequacy of representation.
Lastly, Anderson's test does not allow for the consideration of any factors which might be influencing a market under consideration.
Based upon the foregoing Findings of Fact and Conclusions of Law, it is, therefore, A
RECOMMENDED THAT a Final Order be entered denying the application of Chevrolet World, Inc., for licensure as a dealer of Chevrolet vehicles in Orlando, Florida.
DONE and RECOMMENDED this 5th day of June, 1987, in Tallahassee, Florida.
LINDA M RIGOT
Hearing Officer
Division of Administrative Hearings The Oakland Building
2009 Apalachee Parkway
Tallahassee, Florida 32399-1550
(904) 488-9675
Filed with the Clerk of the Division of Administrative Hearings this 5th day of June, 1987.
ENDNOTES
1/ The MDA boundaries, as drawn by GM, are also defective because there is no sound basis for the exclusion of the Chevrolet dealer located in Kissimmee,
Florida, which is located far closer to AGSSA 5 and the proposed new point site than Ken Rummel Chevrolet in Sanford, which is included by GM in the Orlando MDA.
2/ Mr. Anderson did admit that Chevrolet retail penetration rates would be enhanced, in comparison with Zone and national rates, if the Orlando MDA boundaries were to be drawn so as to be coextensive with the Orlando statistical metropolitan area.
3/ Mr. Anderson conceded that his studies did not take into account the significant negative impact on new car sales which would result from heavy sales of cars by fleet companies in the Orlando market area. Furthermore, Mr.
Anderson performed no studies seeking to determine whether many of the fleet cars used in the Orlando MDA were actually registered in Miami where many of the fleet companies maintain their headquarters. Finally, Mr. Anderson admitted that he performed no linear regression analyses of the Orlando market area using fleet registration or sales as the independent variable and market penetration for retail cars as the dependent variable.
4/ Sales of demonstrator automobiles are considered by GM to be new car sales, whereas a brass-hat vehicle or a current-model, low-mileage rental car is not considered by GM to be the sale of a new car, even though the same models and mileage are involved. Essentially, there is no difference in the sales.
Additionally, brass-hat vehicles qualify for new car financing through GMAC. There is no plausible reason why brass-hat cars should be treated any differently, in terms of sales and registration rates, than demonstrator cars.
5/ Mr. Anderson could not identify any authoritative text or treatise that supported his five-point test for determining whether it is appropriate to compare national or Zone penetration rates with the penetration rates in the Orlando market.
6/ Exhibits 58-60, relating to 1986 age and household income distribution for AGSSA 5, the Orlando MDA, and the United States as a whole, offer little if any relevance to the issue of adequacy of representation. There was no persuasive evidence offered by GM that these age and income distributions either helped or hindered penetration rates for Chevrolet in AGSSA 5 or the entire Orlando MDA. GM did not explain how these individual statistics prove that Chevrolet should be achieving higher penetration rates.
7 Because Pensacola is not in the Jacksonville Zone, its figures should not be utilized. Also, because Tampa has six active dealers (not seven), the population per approved dealer point in the Tampa MDA is 235,123. These proper adjustments to the data significantly alter the conclusions to be drawn therefrom: the dealers in the Orlando MDA service a smaller customer base than the average Jacksonville Zone dealer.
8/ The same analysis should be applied to Mr. Anderson's studies, reflected in Exhibits 66 and 67, concerning truck and combined car/truck registrations per Chevrolet dealers in the Jacksonville Zone. After again removing Pensacola from the analysis and adjusting the Tampa figures to accurately reflect only six dealers, the average number of combined car/truck registrations per dealer point is 12,934. Thus, the Orlando MDA dealers' registrations per dealer point (12,615) is below the Zone average. An additional dealer in Orlando would drive the Orlando average down to 10,092, less than eighty percent of the Zone average.
APPENDIX TO RECOMMENDED ORDER, CASE NO. 86-3617
Petitioners' Proposed Findings of Fact
It should be noted that although many of Petitioners' proposed findings of fact initially appear accurate since they contain truisms, definitions and accurate data, they have been rejected as being subordinate, not supported, irrelevant, etc. The rejection, and reason therefor, is required due to the findings set forth in this Recommended Order that GM's data is incomplete, dubious in the context in which it was offered, is always true and therefore proves nothing, is only true where the market is accurately defined, and, most of all, because GM failed to prove any connection between the data and the issue of whether GM is receiving adequate representation and further failed to show that comparison of the Orlando market to the Zone and the nation is a reasonable or appropriate standard. Also, many of the proposed findings are in support of GM's argument that Century did not prove some alternative methodology for analyzing retail penetration as it relates to adequacy of representation. Since GM has the burden of proof statutorily, Century need not advance an alternative methodology; it is only necessary that Century show GM's computations and conclusions to be defective.
Specific rulings on Petitioners' findings of fact are as follows: 1-4. Adopted either verbatim or in substance.
5-9. Subordinate.
Not supported by credible evidence.
Adopted either verbatim or in substance. 12-14. Subordinate.
Irrelevant.
Contrary to the weight of the credible evidence. 17-34. Subordinate.
Not supported by credible evidence.
Subordinate.
Subordinate.
Not supported by credible evidence.
Not supported by credible evidence. 40-43. Subordinate.
Contrary to the weight of the credible evidence.
Irrelevant.
Adopted either verbatim or in substance.
Not supported by credible evidence. 48-52. Subordinate.
Not supported by credible evidence.
Not supported by credible evidence. 55-57. Subordinate.
Adopted either verbatim or inn substance.
Contrary to the weight of the credible evidence.
Contrary to the weight of the credible evidence.
Not supported by credible evidence. 62-64. Subordinate.
Contrary to the weight of the credible evidence.
Adopted either verbatim or in substance.
Subordinate.
Not supported by credible evidence.
Subordinate.
Not supported by credible evidence.
71-73. Subordinate.
74-79. Not supported by credible evidence.
Irrelevant.
Adopted either verbatim or in substance.
Irrelevant.
Contrary to the weight of the credible evidence.
Irrelevant.
Irrelevant.
Not supported by credible evidence.
Subordinate.
Subordinate.
Contrary to the weight of the credible evidence.
Irrelevant.
Irrelevant.
Subordinate.
Not supported by credible evidence.
Adopted either verbatim or in substance.
Subordinate.
Not supported by credible evidence.
Subordinate.
Subordinate.
Not supported by credible evidence.
Adopted either verbatim or in substance.
Adopted either verbatim or in substance.
Subordinate.
Subordinate.
Contrary.
105-112. Subordinate.
113-121. Irrelevant.
122. Adopted either verbatim or in substance.
Respondent Century Chevrolet, Inc.'s, Proposed Findings of Fact
Specific rulings on Century's proposed findings of fact are as follows:
1. Adopted either verbatim or in substance. 2-7. Subordinate.
8-38. Adopted either verbatim or in substance.
39. Not supported by credible evidence.
40-116. Adopted either verbatim or in substance.
117. Irrelevant.
118-121. Adopted either verbatim or in substance.
COPIES FURNISHED:
Jacqueline R. Griffin, Esquire PEIRSOL, BOROUGHS, GRIMM, BENNETT
& GRIFFIN
201 East Pine Street Orlando, Florida 32801
Dean Bunch, Esquire
RUMBERGER, KIRK, CALDWELL, CABANISS & BURKE
Suite 900, 101 North Monroe Street
Tallahassee, Florida 32301
Michael J. Robinson, Esquire Office of the General Counsel General Motors Corporation New Center One Building
3031 West Grand Boulevard Detroit, Michigan 48232
D. Frank Davis, Esquire Joseph W. Letzer, Esquire BURR & FORMAN
3000 SouthTrust Tower
Birmingham, Alabama 35203
Michael J. Alderman, Esquire Assistant General Counsel Department of Highway Safety
and Motor Vehicles Neil Kirkman Building
Tallahassee, Florida 32301
Leonard R. Mellon Executive Director
Department of Highway Safety and Motor Vehicles
Neil Kirkman Building Tallahassee, Florida 32399-0500
| Issue Date | Proceedings |
|---|---|
| Jun. 05, 1987 | Recommended Order (hearing held , 2013). CASE CLOSED. |
| Issue Date | Document | Summary |
|---|---|---|
| Jun. 21, 1990 | Agency Final Order | |
| Jun. 05, 1987 | Recommended Order | Five prong test used by General Motors to show inadequate representation by existing dealers was invalid, unreliable, and failed in intended purpose. |
