Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
197 Introduction and Methodology To further examine DBE success, Keen Independent developed multivariate regression models to explore patterns of business success using the DBEs, state DOTs, and trade associa- tions identified as successful and a random sample of all DBEs from 17 state DBE lists. Those models estimate the effect of race, ethnicity, and gender on the probability of being a successful DBE while statistically controlling for other business characteristics. Keen Independent used probit regression models to predict the likelihood of a DBE of being identified as a âsuccessfulâ DBE by state DOT and trade associations. Independent or âexplana- toryâ variables include â¢ Number of states where the business is certified as DBE; â¢ Number of years that firm has been in business; â¢ Number of NAICS in which the DBE is certified; and â¢ Primary NAICS (using Dun & Bradstreet database). Keen Independent developed three probit regression models using DBE directories from 17 different states. These states include California, Hawaii, Iowa, Idaho, Indiana, Louisiana, Maryland, Michigan, Minnesota, New Jersey, New Mexico, Nevada, New York, Ohio, South Carolina, Texas, and Wisconsin. The study team drew a 5% random sample of DBEs (844 firm sample from 17,296 DBEs in those directories), excluding any firms already identified as âsuc- cessfulâ in those states. â¢ The model for all industries includes 1,248 observations; â¢ The model for the construction industry includes 342 observations; and â¢ The model for the professional and technical services includes 427 observations. All DBE Firms Table G-1 presents the results of the analysis for all the firms in our sample. The dependent variable is the probability of being identified as a successful firm by a state DOT or trade asso- ciation. The study team included two groups of independent variables. The first group includes the following non-binary variables: â¢ The number of states in which the DBE is certified; â¢ The number of years in operation, both in levels and squared; and â¢ The number of NAICS codes in which the firm operates (based on information in the DBE directories). A P P E N D I X G Statistical Analysis of DBE Success
198 Compendium of Successful Practices, Strategies, and Resources in the U.S. DOT Disadvantaged Business Enterprise Program The study team also controlled for industry and race/ethnicity and gender factors using a set of binary variables. â¢ For industry characteristics, the model included dummy variables for (a) agriculture, (b) con- struction, (c) manufacturing and wholesale trade, (d) transportation, (e) professional and technical services, and (f) administrative services; the omitted category is the âother servicesâ industry. â¢ Regarding the race and ethnicity of the owner, the model considered the following catego- ries: (a) African American, (b) Asian-Pacific American, (c) Hispanic, (d) majority (i.e., white males), (e) Native American, and (f) Subcontinental Asian American. The omitted category is white female owner. Results in Table G-1 show that the following variables are positively and significantly cor- related with the probably of being identified by a state DOT as âsuccessfulâ (at the 1% level): â¢ The number of states in which the DBE is certified; â¢ The number of years in operation; and â¢ The number of NAICS codes in which the firm operates. The above results suggest that firms that are certified in more than one state, certified in two or more NAICS codes, and have been in business longer are more likely to be identified as successful (by a state DOT). Each of the binary control variables for an industry exhibit positive coefficients and are sta- tistically different from zero. Thus, operating in this subset of industries increases the likelihood of success for a given firm. Regarding the ethnicity control variables, the table shows that DBEs owned by African Americans, Hispanic Americans, and Subcontinent Asian Americans are less likely to be suc- cessful compared with business owned by white females. There were a few white male-owned Variable Number of state DOTs certified 0.300 *** Years in business 0.039 *** Years in business (squared) 0.000 ** Number of certified NAICS 0.244 *** Agriculture and utilities 0.529 Construction 0.856 *** Manufacturing and retail trade 0.394 ** Transportation and warehousing 0.405 * Professional and technical services 0.322 * Administrative support 0.473 ** African American -0.610 *** Asian-Pacific American -0.036 Hispanic American -0.421 *** Majority -1.014 ** Native American 0.041 Subcontinent Asian American -0.403 ** Constant -1.617 *** *** p < 0.01, ** p < 0.05, * p < 0.1 Coefficient Table G-1. Probability of success in all industries.
Statistical Analysis of DBE Success 199 firms in the data, and they too were less likely to be identified as successful as white woman- owned DBEs (coefficient is for âmajority-owned firmsâ). Construction DBE Firms The study team developed similar models for the construction industry and for the profes- sional services industry. Table G-2 shows the results for probability of success for DBE firms with a primary NAICS code in construction. Similar to the results for all DBEs, the number of states where the DBE is certified and the number of NAICS codes in which the firm operates are positively and significantly correlated with the probability of being identified as successful. Somewhat surprisingly, the variables that consider years in operation are not statistically dif- ferent from zero. This may suggest that tenure is not relevant in success. It may be that being certified in many states and in many types of work, which occurs over time, is more important than the age of the company when explaining whether a firm is successful. This warrants further research. Regarding race, ethnicity, and gender factors, African American-, Hispanic American-, and Subcontinent Asian American-owned DBEs were less like than white women-owned construc- tion DBEs to be identified as successful. Professional and Technical Services DBE Firms Table G-3 provides the results for a similar model focusing on professional and technical services DBEs. Results show that the number of states where the DBE is certified has a positive and significant coefficient, similar to the results in the construction industry model. Unlike the construction industry model, older DBE professional services firms are more likely to be identi- fied as successful as companies in business for a shorter period of time. In terms of the race, ethnicity, and gender factors, only the variable for African American ownership was statistically significant. The model results suggest that African American-owned DBE professional and technical services firms are less likely to be identified as successful as DBEs owned by white women. Variable Number of state DOTs certified 0.219 * Years in business 0.029 Years in business (squared) 0.000 Number of certified NAICS 0.383 ** African American -0.842 *** Asian-Pacific American 0.255 Hispanic American -0.640 *** Majority -1.048 Native American 0.040 Subcontinent Asian American -0.979 ** Constant -0.521 * *** p < 0.01, ** p < 0.05, * p < 0.1 Coefficient Table G-2. Probability of success in the construction industry.
200 Compendium of Successful Practices, Strategies, and Resources in the U.S. DOT Disadvantaged Business Enterprise Program Variable Number of state DOTs certified 0.329 *** Years in business 0.055 *** Years in business (squared) -0.001 ** Number of certified NAICS 0.146 African American -0.488 *** Asian-Pacific American 0.002 Hispanic American -0.303 Majority -0.320 Native American 0.123 Subcontinent Asian American -0.053 Constant -1.577 *** *** p < 0.01, ** p < 0.05, * p < 0.1 Coefficient Table G-3. Probability of success in the professional and technical services industry.