Data and Definitions

We examine Texas construction workers’ and their families’ utilization of the five largest means-tested safety net programs for which data are available: Medicaid; Children’s Health Insurance Program (CHIP); basic household income assistance under Temporary Aid for Needy Families (TANF); Earned Income Tax Credit (EITC); and Supplemental Nutrition Assistance Program (SNAP). Responsibility for funding the health programs is shared by Texas and the federal government. We include only the cash assistance portion of TANF, and this program too receives funding from both Texas and the federal government. While there is a state-level EITC program, in this analysis we include only the federal EITC. The federal government alone funds SNAP. We analyze only programs that function as income supplements, omitting job-training, housing cost assistance, educational, and other programs that indirectly assist low-income families.

To calculate the numbers of working families who participated in safety net programs, we restrict the sample to those who work 27 or more weeks per year and 10 or more hours per week in all industries in Texas. We exclude workers who live in institutional group quarters. To identify construction workers, we further use the 1990 Census Bureau industrial code All Construction (60), and the 2010 Census Bureau occupation codes from First-Line Supervisors of Construction Trades and Extraction Workers (6200) to Construction Workers, n.e.c. (6765), and we include W2 workers and the not-incorporated self-employed but exclude the incorporated self-employed.

Results

Table 1 shows the annual enrollment in safety net programs of construction workers and their families in Texas between 2015 and 2019. We estimate that 46% of construction working families are enrolled in at least one program, significantly more than the 31% of all working families. Construction working families have higher enrollment than all working families in every program except TANF, where both groups have a negligible enrollment.

Overall, construction working families are almost 50% more likely than all working families to participate in one or more means-tested safety net programs in Texas. These families are almost 60% more likely to be enrolled in Children’s Medicaid, 54% more likely to be enrolled in EITC, and 50% more likely to be enrolled in SNAP.

Table 2 presents the combined annual expenditures by the State of Texas and the federal government on safety net program usage by construction working families and all working families, again averaged over the years 2015-2019. In total, $1.9 billion is spent on safety net program utilization annually by construction working families in Texas.

Table 3 presents the health insurance coverage status of construction workers and all workers in Texas. More than half (55%) of construction workers lack insurance coverage. The rate at which construction workers lack health insurance is two and half time the rate for all workers in Texas (20%).

Determining the full cost of uninsurance in Texas, let alone the cost for uninsured construction workers, is beyond our scope. But in addition to causing hardship for uninsured construction workers, uninsurance creates significant expenses for states, counties, and the federal government. The Kaiser Family Foundation reports that in the years 2015-2017, uncompensated health care costs for the uninsured nationwide averaged $42.4 billion per year, with the public picking up around 80% of these costs.[10]

Discussion

The low wages and exploitative practices in the construction industry, both in Texas and nationally, cause profound hardship for workers and their families. It also costs the public. When employers misclassify their workers or pay them under the table, they are defunding and defrauding government programs, including workers’ compensation, Social Security, and Medicare. Overall, misclassification is estimated to cost state and federal coffers at least $3,000 annually for every worker that is misclassified.[11] Researchers have estimated that 37% of Texas construction workers are misclassified, at a cost to the state of $1.2 billion in payroll taxes annually.[12] And, as found in this analysis, low-road employment practices cause above-average utilization of safety net programs by construction working families in Texas. Importantly, the numbers provided in this analysis do not fully reflect the deprivation among this workforce. Undocumented immigrants currently comprise 13% of the construction workforce nationwide (compared to 5% of the overall US workforce),[13] and with rare exceptions they are ineligible for state and federal assistance.[14] Their working conditions, among the worst in the industry, are not reflected in this analysis.

In most states, general contractors and subcontractors are not liable for—and in fact benefit from—payroll fraud found further “down the chain” of subcontractors; these practices continue “with or without the knowledge, assistance or willful ignorance of the owners, developers, general contractors, or construction managers.”[15] In the absence of legislation preventing these practices, construction workers in Texas should expect to continue to be exploited and cheated, and lawful contractors should expect to find it more and more difficult to compete in the industry.

Appendix: Methods

We mainly rely on four sources of data: the US Census Bureau’s American Community Survey (ACS), the March Supplement of the US Bureau of Labor Statistics Current Population Survey (CPS), the US Bureau of Labor Statistics Occupational Employment Statistics (OES), and administrative data from the Medicaid, CHIP, TANF, EITC, and SNAP programs. Medicaid figures exclude aged, blind, and disabled enrollees. The ACS surveys a large number of respondents and asks them about their work history, income, and family structure. The March Supplement of CPS, also known as the Annual Demographic Supplement, asks respondents about receipts of cash and noncash transfer payments during the past year and includes questions about the programs we examine in this analysis.

Survey databases like the ACS and CPS frequently have safety net program utilization counts that differ from program administrative data. We adjusted the CPS so that its program utilization estimates match the program administrative data. The CPS does not provide a large enough sample size to accurately estimate program utilization for construction workers at the state or county levels. The ACS does have sufficient sample size for this analysis but lacks specific questions about program utilization, and its occupational employment counts differ from more accurate data like the OES. On the other hand, while the OES has accurate employment counts for wage workers, it does not include independent contractors. To overcome these issues, we built a model using CPS data to predict program utilization based on income, demographics, and family structure. We then used that model to impute program utilization onto the ACS data. We calculated the ratio of wage workers to non-incorporated self-employed workers based on the ACS and used it to adjust the OES data for non-incorporated self-employed workers, and then adjusted the employment counts in the ACS to match the adjusted OES data. Finally, we used that imputed and adjusted ACS data to analyze safety net program utilization in families of construction workers.

For a detailed explanation of methodology, please see Appendix A: Methodology from Fast Food, Poverty Wages: The Public Cost of Low-Wage Jobs in the Fast-Food Industry.[16]