Book Navigation Links
Appendix D: Estimating The Economic Costs Of Worker Misclassification In The Massachusetts Construction Industry
Introduction
It is well understood among industry stakeholders and research scholars that worker misclassification largely occurs in the construction industry due to employers’ self-interest in evading legally-required tax contributions and other necessary expenses of legal employment. These actions are known to result in severe costs on workers and broader society. Most directly, this form of wage and tax fraud defunds critical social programs, robs workers of their legal rights to benefits and shifts much of employers’ tax burden onto the backs of workers and taxpayers at large. Indirectly, these illegal actions put downward pressure on wages for workers in the legitimate corners of the construction industry. Further, wage and tax fraud is well recognized as severely disadvantaging honest, law-abiding contracts in the bidding for new projects, driving many out of business and further quickening the “race to the bottom” when it comes to employment practices in the construction sector.
While these issues are well known to industry stakeholders and academics alike, quantifying the effects of worker misclassification encounters a familiar and challenging problem: data availability. While Appendix A was designed to estimate the incidence of worker misclassification in the Massachusetts construction industry, assessing the economic costs requires one to know how much money is exchanging hands illegally. The underground nature of this activity makes the development of a definitive answer to this question a practical impossibility. But there are available means of estimating the economic costs of wage and tax fraud in the construction industry. To those ends, this study relies on a variant of the methodology co-developed by one of the authors of this report (Dale Belman) and published in a 2019 paper commissioned by the Attorney General for the District of Columbia, whose office was similarly interested in assessing the costs of wage and tax fraud in the construction industry in their jurisdiction.28 This empirical approach was further refined in the 2020 ICERES study, and serves as the basis for the current report’s projections on the economic costs in Massachusetts’ construction industry.
Per-Worker Costs of Wage and Tax Fraud
The general framework for assessing the state-level economic costs of wage and tax fraud as advanced by these prior studies is to (a) establish the economic cost of fraud for a single worker and then (b) multiply this per-worker cost by the number of workers involved. While the number of workers affected was established in Appendix A (22,146 and 36,719), developing the per-worker cost of payroll fraud in Massachusetts’ construction industry requires a number of assumptions. The use of assumptions—both in the number of workers affected and their per-worker costs—knowingly introduces a nontrivial margin of error into the cost projections, but these are necessary given that researchers do not have direct evidence of the full extent of worker misclassification nor the amount of money that changes hands in the underground economy.
In generating the per-worker cost of wage and tax fraud, the starting assumption is that workers affected would, if paid legally, earn incomes equivalent to the 10th percentile of legal wage-and-salary employees in construction occupations in the Commonwealth.29 Drawn from a survey of legal construction employers in the state, this equates to annual earnings of $35,200.30 Using the 10th percentile of legal earnings is a conservative assumption, however there are a multitude of reasons supporting this starting point. First and foremost, it is recognized that construction workers most often affected by wage and tax fraud are in lower-skill, lower-paying jobs. Second, an analysis of the American Community Survey reveals that this number nearly matches the 25th percentile of earnings ($35,000) of all construction-industry workers in Massachusetts in 2019, a number that includes misclassified independent contractors and off-the-books workers. Further, among construction workers in the four most affected trades identified in the main report— painters, carpenters, laborers, and roofers—this value ($35,200) is between the 35th and 40th percentile of earnings in Massachusetts according to the ACS. Finally, this value approximates the median earnings ($35,000) of non-incorporated self-employed construction workers in the Commonwealth in 2019.31
While there are numerous aligning data points suggestive of $35,200 as the appropriate number, it is reminded that the actual earnings of these workers is an empirical black box. Further, given that payroll fraud occurs in all corners of the construction industry— including high-skill, high-pay jobs—the authors assess payroll fraud with a second assumption: that workers affected by payroll fraud would earn annual earnings equivalent to the 25th percentile of legal wage-and-salary workers in construction occupations. According to the Bureau of Labor Statistics, that equates to $44,960.32
In order to build the per-worker cost of wage and tax fraud—a necessary precursor to generating the aggregate economic costs—this study adheres closely to the method advanced by Dale Belman (Michigan State University) and Aaron Sojourner (University of Minnesota) in their report for the Attorney General for the District of Columbia, an empirical approach further refined in the 2020 ICERES report on payroll report. This includes the application of construction-industry data from the Employer Costs for Employee Compensation (ECEC) program administered by the Bureau of Labor Statistics; these reports offer the national average per-hour rates for all types of employees’ pay and benefits in the construction sector.33 This work is complemented with Massachusetts-specific tax and contributions schedules where possible, including insight offered by the Department of Unemployment Assistance (average UI contribution rates in construction) and the Workers’ Compensation Rating and Inspection Bureau of Massachusetts (average workers’ compensation insurance rates among construction employers).
Table D1. A Comparison of Per-Worker Labor Costs for Legal Employers and Those Engaging in Payroll Fraud in Massachusetts, 2019 (Assuming Legal Earnings = $35,200)
| Legal Employer | Conservative Estimate | Moderate Estimate | |
|---|---|---|---|
| Value to Worker | |||
| Regular Pay | $34,329.61 | $34,329.61 | $34,329.61 |
| Overtime and Premium Pay | $870.39 | $0.00 | $0.00 |
| Fringe Benefits / Wage Premium | $5,599.27 | $5,599.27 | $0.00 |
| Subtotal (1) | $40,799.27 | $39,928.88 | $34,329.61 |
| LESS: Social Security & Medicare (EE share) (2) | $2,692.80 | $6,109.12 | $5,252.43 |
| Total – Net Value to Worker | $38,106.47 | $33,819.76 | $29,077.18 |
| Employer Contributions to Social Programs | |||
| Social Security & Medicare (ER share) | $2,692.80 | $0.00 | $0.00 |
| Unemployment Insurance | $1,105.50 | $0.00 | $0.00 |
| Workers Compensation | $1,669.18 | $0.00 | $0.00 |
| Total – ER Contributions to Social Programs (3) | $5,467.48 | $0.00 | $0.00 |
| Totals | |||
| Total Net Value to Worker (1-2) | $38,106.47 | $33,819.76 | $29,077.18 |
| Total Value to Social Insurance (2+3) | $8,160.28 | $6,109.12 | $5,252.43 |
| Total Labor Costs (1+3) | $46,266.75 | $39,928.88 | $34,329.61 |
| Differences from Legal Employer | |||
| Total Labor Cost Differential from Legal | $6,337.88 | $11,937.15 | |
| Percent Higher for Legal Employer | 15.9% | 34.8% |
Following Belman and Sojourner’s blueprint, this study starts by decomposing the pay of a worker in Massachusetts, earning $35,200 per year. But this value undercounts the sum that legal employers would have to pay to employ the worker, as this fails to capture legally-required tax contributions and optional fringe benefits. To those ends, the first column of Table D1 presents estimates of each category of employer costs so that a more complete understanding of the economic costs of payroll fraud can be estimated. These categories include:
- Regular vs. Overtime and Premium Pay: One of the defining characteristics of legal employment is that employees are entitled to overtime wage rates (i.e., time-and-a-half) if they exceed 40 hours of work in a given week; in contrast, higher rates for overtime are not required for workers misclassified as independent contractors. Further, regular employees are often granted a premium for working on holidays. The starting point of $35,200 however, does not differentiate between regular, overtime and premium pay.34 Fortunately, calculations derived from the ECEC reflect that 2.47% of construction workers’ income, on average, is derived from the overtime and premium rates on a national basis (e.g., the “half” in “time-and-a-half”).35 Applying this to the worker in question, this means that $34,329.61 was earned via regular rates with an additional $870.39 earned from overtime and premium pay.36
- Social Security and Medicare: To be conservative, this study assumes that workers’ $35,200 represents gross annual pay. The employee will then have 7.65% deducted for Social Security and Medicare, a contribution otherwise known as the Federal Insurance Contribution Act (FICA) tax; this amounts to $2,692.80 being taken out of employees’ paychecks. The employer will also be required to pay an additional $2,692.80 to cover its share of Social Security and Medicare without it showing up on the employee’s pay stub.
- Workers’ Compensation: While this rate may differs across trades and individual firms, data provided by WCRIBMA reflects that the average cost of workers’ compensation insurance for construction employers in Massachusetts in 2019 was $4.742 for every $100 of the employer’s payroll.37 For the hypothetical worker in question, this equates to $1,669.18 in workers’ compensation insurance policy costs.
- Unemployment Insurance: Legal employment in Massachusetts requires that employers contribute to the state’s UI fund; the required payment is based on a certain percentage of an employee’s first $15,000 earned via the firm. While the amount that employers must pay varies by industry, trade and firm, a representative of the Department of Unemployment Assistance confirmed with the authors that the average rate for Massachusetts construction employers was 7.37% in 2019. For the worker in question, this means that a typical employer would be required to pay $1,105.50 in UI contributions.38
- Tax-Exempt Benefit Costs: Workers’ responses on the American Community Survey do not offer insight into the dollar value associated with employer-provided, tax-exempt fringe benefit costs, which include things like health insurance and pension funding. However, calculations from the ECEC suggest that, on average, construction employers spend $17.15 on these fringe benefits for every $100 paid to the worker on a national basis. While this may be true for the average worker, benefit packages are likely to be much smaller for those workers paid an entry-level wage. As such, this study analyzes differences in the incidence of employer-sponsored health insurance in the ACS to suggest that a more appropriate fringe benefit rate for these workers should be $15.91 for every $100 paid to the worker.39 Multiplying this rate by $35,200, this implies that employers would spend $5,599.27 in insurance and pension benefits for this worker.40
Aggregating all wages, benefits, taxes and required social contributions, this employee would cost a legally-operating construction employer in Massachusetts a total of $46,266.75. Of those funds, workers would receive $38,106.47 in after-tax earnings and fringe benefits. The remaining $8,160.28 would be diverted to Social Security, Medicare, workers’ compensation and unemployment insurance programs.
The fundamental question from here is: how much would this worker cost an employer if they were classified as an independent contractor or hired in a cash-only arrangement? This question is a bit more complicated than meets the eye. Economic theory would suggest that, in order to entice workers to forego the legally-earned benefits bestowed upon a legal employee, employers would have to pay workers extra per hour in cash; this amounts to what economists call a “wage premium.” Conversations with industry stakeholders suggests that this sometimes does happen. But certainly not always. Other times, employers are able to exploit their monopsony power in the labor market—they have the jobs that workers desperately need—and attract and hire enough workers without paying such a premium.
The wage premium paid to workers agreeing to operate in an illegal employment relationship likely differs from employer to employer and from worker to worker. For many employers and workers, there may be no such wage premium at all. Others may see a reasonable sum added to their earnings to incentivize them to work off-the-books. Unfortunately, there are no known data on this presumed value. So for the sake of offering an initial, conservative estimate of the costs of payroll fraud in the construction industry, this study follows the lead of Belman and Sojourner’s approach in their 2019 study in first assuming that workers who are employed fraudulently do earn a sizeable premium: the cash value of legal employees’ fringe benefits. This would leave the employer to save on labor costs via the (a) denial of overtime and premium pay, (b) avoiding required workers’ compensation and unemployment insurance contributions, and (c) shifting its FICA burden to employees.
Given this conservative assumption, comparing the first two columns of Table D1 projects the amount and distribution of employers’ per-worker labor costs when the firm is operating legally versus when they acting fraudulently but offering workers this sizeable wage premium. The results suggest that a construction employer that is operating legally in Massachusetts must spend $6,337.88 more on a per-worker basis than one that is operating fraudulently. Put another way, per-worker labor costs for the law-abiding firm are 15.9% higher than the one acting illegally.41
Much of this differential in labor costs is attributable to the elimination of firms’ required contributions to social insurance programs. A substantial portion of this is a shifting of the “employer share” of the Social Security and Medicare tax burden from employers and onto the backs workers. But of perhaps the most interest to state legislators may be that payroll fraud also defunds the state’s unemployment insurance fund ($1,105.50 lost per worker). In addition, there is a substantial amount of workers’ compensation insurance premiums uncollected ($1,669.18 per worker). These values are the same regardless of whether a wage premium is paid or not.
Beyond shortfalls to state and federal programs, it is notable in the second column of Table D1 that workers also lose a substantial amount even with the assumed wage premium. Including $870.39 in lost overtime and premium pay and now having to pay the employer’s share of the FICA tax burden—which is larger because taxes must paid on the wage premium whereas fringe benefits are tax-exempt—workers’ net compensation declines by $4,286.71 when working for an employer engaging in worker misclassification. In sum, even under the best situation for the worker—the presence of a substantial wage premium—both workers and taxpayers are substantially worse off.
The assumption that these workers receive a substantial wage premium for operating outside a legal employment structure does offer a conservative estimate of per-worker cost savings attributable to payroll fraud. But while a large wage premium may occur in some parts of the industry, our conversations with industry stakeholders suggests that workers more often receive little to no such premium. To those ends, the third column in Table D1 offers a less conservative assumption that workers receive no wage premium for engaging in an illegal employment structure, instead receiving only cash in the form of regular pay. Under this set of circumstances, a construction employer that is operating legally and offers fringe benefits in Massachusetts must spend $11,937.15 more on a per-worker basis than a contractor operating fraudulently. This equates to 34.8% higher per-worker labor costs for the law-abiding firms that offer benefits when compared to illegal firms that do not. 42 Given the lack of wage premium, workers in Massachusetts’ construction industry bear the brunt of this arrangement, with their net compensation being $9,029.29 less than that of a legal employee (including higher Social Security and Medicare tax liability). To be clear, however, not all of this differential represents illegality: the evasion of required social insurance contributions is illegal, but the refusal to pay fringe benefits or a wage premium is not.
Table D1 highlights the respective sources of these cost savings for fraudulent employers and how this leads to reduced net compensation for workers and funding for social programs in Massachusetts. For employers, avoiding legally required contributions to social insurance programs makes up a considerable portion ($5,467.48) of the cost differential, leading to substantial shortfalls in the Massachusetts unemployment insurance and workers’ compensation insurance programs, as well as Social Security and Medicare on a federal level. Further, the denial of overtime and premium pay shaves off an additional $870.39 on a perworker basis. In sum, the results of Table D1 suggest that payroll fraud allows construction employers to illegally reduce labor costs by $6,337.88 when hiring a worker at a rate equal to the 10th percentile ($35,200) of construction occupations in the legitimate labor market. These savings can expand to over $10,000 if also including the cost savings accrued by not offering fringe benefits to employees (which is not illegal).
Aggregate Costs of Wage and Tax Fraud
The results of Table D1 offer the authors’ best estimates of the minimum and maximum per-worker labor cost differential between legal employers and those who misclassify equivalent workers as independent contractors or hire workers using cash-only payments in Massachusetts. To estimate the aggregate cost of wage and tax fraud in the state’s construction industry, this study multiplies the per-worker cost differential by the number of workers directly affected by payroll fraud. As a starting point, the authors apply the lower-bound number of workers involved in Massachusetts (22,146) and continue to assume that affected workers would earn the equivalent of the 10th percentile of legal wage-and-salary employees holding construction occupations in the Commonwealth.
Table D2 presents the projected aggregate annual labor costs of these 22,146 construction workers in Massachusetts on the basis of their employment relationship. The first column presents the aggregate costs of these workers being employed legally. The second and third columns estimate the same totals but under the assumption that workers are employed fraudulently; the second column is conservative and assumes there is a wage premium equal to the cash value of fringe benefits while the third column assumes no wage premium. The results of Table D2 suggest that these 22,146 workers would have cost Massachusetts construction employers $1.02 billion if employed legally (and employees were offered insurance and pension benefits). In comparison, employers engaged in fraud paid just $884.3 million in labor costs if paying a wage premium equivalent to the value of fringe benefits, and just $760.3 million if not offering any wage premium. Ignoring the loss of fringe benefits (which are not legally required of employers), these results demonstrate that by not paying overtime ($19.3 million) and evading required taxes and social contributions ($121.1 million), construction employers in Massachusetts illegally shaved an estimated $140.4 million off their labor costs in 2019.
Of the $140.4 million pocketed by contractors and developers (through lower project costs), it is important to highlight losses to critical social programs. First, payroll fraud in the Massachusetts construction industry is projected to result in a $24.5 million shortfall in the state’s unemployment insurance fund under these assumptions. Second, these actions also led to $37.0 million in lost workers’ compensation insurance premiums. Finally, a substantial portion of employers’ savings was due to its offloading of its obligations to Social Security and Medicare onto the backs of workers; this amounts to $59.6 million in tax obligations transferred from employers to workers. This is because, under these circumstances, workers would technically be considered “self-employed” and thus responsible for both the employee’s and employer’s share of Social Security and Medicare. This increased tax responsibility comprises a large part of why workers’ net value declines so drastically due to payroll fraud.
While the results offered in Table D2 represent the direct costs associated with payroll fraud in Massachusetts’ construction industry, there are also indirect economic costs. In particular, employers’ lack of tax withholding and failure to provide employment documentation open the door to workers to not report or underreport their income to the Internal Revenue Service and the Massachusetts Department of Revenue. It is telling that, in a 2016 report, the IRS noted that only 1% of W-2 earnings were misreported on tax forms; in contrast, the agency assessed that 64% of nonfarm proprietor income—which is subject to “little to no information reporting”—is underreported on tax forms.43 As such, while the decision to fully report earnings is the responsibility of workers, the lack of employer documentation effectively opens the door for widespread income tax shortfalls.
Table D2. Estimated Aggregate Labor Costs for Legal Employers and Those Engaging in Payroll Fraud, Massachusetts, 2019 ($ in millions) (Assuming Legal Worker Earnings = $35,200)
| Legal Employer | Conservative Estimate | Moderate Estimate | |
|---|---|---|---|
| Illegal Employment | |||
| Number of Workers | 22,146 | 22,146 | 22,146 |
| Value to Worker | |||
| Regular Pay | $760.3 | $760.3 | $760.3 |
| Overtime and Premium Pay | $19.3 | $0.00 | $0.00 |
| Fringe Benefits / Wage Premium | $124.0 | $124.0 | $0.00 |
| Subtotal (1) | $903.5 | $884.3 | $760.3 |
| LESS: Social Security & Medicare (EE share) (2) | $59.6 | $135.3 | $116.3 |
| Total – Net Value to Worker | $843.9 | $749.0 | $643.9 |
| Employer Contributions to Social Insurance | |||
| Social Security & Medicare (ER share) | $59.6 | $0.00 | $0.00 |
| Unemployment Insurance | $24.5 | $0.00 | $0.00 |
| Workers Compensation | $37.0 | $0.00 | $0.00 |
| Total – ER Contributions to Social Insurance (3) | $121.1 | $0.00 | $0.00 |
| Totals | |||
| Total Net Value to Worker (1-2) | $843.9 | $749.0 | $643.9 |
| Total Value to Social Insurance (2+3) | $180.7 | $135.3 | $116.3 |
| Total Labor Costs (1+3) | $1,024.6 | $884.3 | $760.3 |
| Differences from Legal Employer | |||
| Total Labor Cost Differential from Legal | $140.4 | $264.4 | |
| Percent Higher for Legal Employer | 15.9% | 34.8% |
The projections in Table D3 offer a range of the potential tax loss attributable to worker nonpayment and underpayment that is made possible by the lack of employment documentation provided by the employer. To be conservative, this study uses the entire range of potential income underreporting rates among self-employed construction workers that could be gleaned from the 2016 study in Public Budgeting and Finance and from IRS reports: 23.3% to 64.0%.44 This leads to a predictably wide range of potential outcomes for state and federal revenue, but the authors are compelled to adhere to this range in the absence of confirmatory data otherwise. The results of Table D3 demonstrate that Social Security and Medicare experience the most substantial expected shortfalls due to payroll fraud in the Massachusetts construction industry. Using the $35,200 income assumption for each of the 22,146 affected workers employed by Massachusetts construction employers, it is projected that these programs experience losses between $27.1 million and $86.6 million.
The bottom two rows of Table D3 also present the projected shortfalls in federal and state income taxes—using 2019 tax schedules—as a result of income underreporting made possible by the lack of employer documentation.45 Generating exact estimates, however, is practically impossible with publicly-available data given that researchers do not know which specific workers are affected by payroll fraud. As such, the authors must make some assumptions about the characteristics of these workers. First, since marital status dictates workers’ standard deduction and tax rates, this study assumes that workers engaged in payroll fraud have one child and are married at the same proportion (56.83%) as all construction workers in the authors’ analysis of the American Community Survey. In the absence of clear data on spousal income, this study assumes that all workers take the standard deduction and have no other household income. This latter assumption is extremely conservative, and suggests that the estimated resulting income tax revenue losses approximate lower-bound projections.46 With these caveats in mind, the results suggest that payroll fraud in Massachusetts’ construction industry led to federal income losses between $8.4 million and $31.7 million in 2019.
Table D3. Minimum and Maximum Estimated Tax Loss in Payroll Fraud, Massachusetts, 2019 (in $ millions) (Assuming Legal Worker Earnings = $35,200)
| Minimum | Maximum | |
|---|---|---|
| Illegal Employment | ||
| Number of Workers (Total) | 22,146 | 22,146 |
| Number of Massachusetts Residents (Est.) | 20,554 | 20,554 |
| Tax Revenue Shortfalls | ||
| Social Security & Medicare | $27.1 | $86.6 |
| Federal Income Tax (2019 tax schedule) | $8.4 | $31.7 |
| State Income Tax (2019 tax schedule) | $6.7 | $18.4 |
Estimating the corresponding loss in Massachusetts tax revenue is complicated by the fact that some of the estimated 22,146 workers affected by payroll fraud while working for Massachusetts construction employers do not live in the state and would not be responsible for Massachusetts income tax. To estimate the proportion of affected workers who are state residents, this study used data from the 2019 American Community Survey that suggests that 92.81% of construction workers employed by Massachusetts firms also live in the state. As a result, this proportion is multiplied by the number of workers affected; the results suggest 20,554 of these workers reside in Massachusetts and thus are subject to the state’s income tax.47 Using the same income and demographic assumptions above and using the 2019 state income tax schedule, Table D3 suggests that payroll fraud resulted in an estimated $6.7 million to $18.4 million shortfall in Massachusetts income tax. As a reminder, however, these estimates are lower-bound projections of income tax losses due to the conservative nature of the assumptions applied.
The first three tables of Appendix D explored payroll fraud under the most conservative set of assumptions: 22,146 workers affected and assumed earnings ($35,200) equivalent to the 10th percentile of wage-and-salary income for those in legal construction occupations. But it is reminded that this represents only one potential set of assumptions regarding the number of workers affected and their presumed earnings if they had been employed legally. To those ends, Table D4 presents the estimated aggregate costs of payroll fraud under four different scenarios based on the number of workers affected (22,146 or 36,719) and the assumed average earnings of these workers had they been employed legally: the 10th percentile ($35,200) or 25th percentile ($44,960) of wage-and-salary tradespeople in Massachusetts.
The projections offered by these four scenario tell a similar story, even if the dollar values change in each column of Table D4. First and foremost, the direct effects of payroll fraud are substantial no matter the case, with illegal actions of employers—not paying overtime and evading taxes and legally-required contributions—amounting to totals between $140.4 million and $286.0 million. The state unemployment insurance program is projected to have lost between $24.5 million and $40.6 million in 2019 due to payroll fraud in construction, while there were between $37.0 million and $78.3 million in uncollected workers’ compensation insurance premiums. Employers also rid themselves of substantial Social Security and Medicare obligations, amounting to $59.6 million to $126.3 million under the varying assumptions outlined in Table D4.
And this is all on top of indirect effects of payroll fraud, as the lack of employer documentation effectively makes possible widespread income underreporting to the IRS and the Massachusetts Department of Revenue. While tax estimates knowingly feature a wide margin of error, the results of Table D4 suggest that payroll fraud in Massachusetts is costing the state budget between $6.7 million and $41.3 million depending on which set of assumptions is applied. At the federal level, presumed income underreporting among Massachusetts construction workers affected by payroll fraud leads to federal income tax shortfalls of $8.4 million to $82.7 million depending on the chosen assumptions; correspondingly, Social Security and Medicare exhibit losses of $27.1 million to $183.4 million depending on the chosen assumptions.
Table D4. Estimated Aggregate Labor Costs for Legal Employers and Those Engaging in Payroll Fraud, Massachusetts, 2019 (in $ millions)
| Conservative | (22,146 workers) | Moderate | (36,719 workers) | |
|---|---|---|---|---|
| Earnings (10th vs. 25th Percentile) | ||||
| Assumed Legal Worker Earnings | $35,200 | $44,960 | $35,200 | $44,960 |
| Total Labor Costs | ||||
| Total Labor Costs If Workers Hired Legally | $1,024.6 | $1,301.9 | $1,698.9 | $2,158.7 |
| Total Labor Costs If Workers Hired Fraudulently |
Min $760.3 Max $884.3 |
Min $971.1 Max $1,129.4 |
Min $1,260.5 Max $1,466.1 |
Min $1,610.1 Max $1,872.7 |
| Direct Effects of Payroll Fraud | ||||
| Overtime and Premium Pay Not Received | $19.3 | $24.6 | $32.0 | $40.8 |
| Unemployment Insurance Fund Shortfall | $24.5 | $24.5 | $40.6 | $40.6 |
| Workers Compensation Fund Shortfall | $37.0 | $47.2 | $61.3 | $78.3 |
| Employer Share FICA onto Workers | $59.6 | $76.2 | $98.9 | $126.3 |
| Effect of Worker Income Underreporting | ||||
| Social Security & Medicare Shortfall |
Min $27.1 Max $86.6 |
Min $34.6 Max $110.6 |
Min $44.9 Max $143.6 |
Min $57.4 Max $183.4 |
| Federal Income Tax Shortfall |
Min $8.4 Max $31.7 |
Min $13.7 Max $49.9 |
Min $13.9 Max $52.5 |
Min $22.8 Max $82.7 |
| State Income Tax Shortfall |
Min $6.7 Max $18.4 |
Min $9.1 Max $24.9 |
Min $11.1 Max $30.5 |
Min $15.0 Max $41.3 |
Discussion
Appendix D has provided the authors’ preferred methodology for estimating the economic costs of wage and tax fraud in Massachusetts’ construction industry. This approach leads to cost projections suggesting that payroll fraud among Massachusetts construction employers was estimated to directly cost workers and taxpayers a minimum of $140.4 million in 2019, and potentially reaching values over $250 million. Put another way, this is effectively a public subsidy for law-breaking construction employers, general contractors, developers, construction owners, and anyone else who profits from an economic structure that is based on flagrant, pervasive and systematic violation of state and federal labor law and, in most cases, the suppression of workers’ rights.
While this empirical approach is the authors’ best attempt at projecting the costs of wage and tax fraud in the Massachusetts construction industry, it is reminded that this study is attempting to estimate the volume of money exchanging hands in the underground economy. To those ends, a “true” count is practically impossible: both employers and workers have incentives to systematically conceal payments and their involvement in fraudulent activities.
As such, the authors have had to rely on assumptions about the number of workers affected, their incomes, and other work and tax circumstances. The use of assumptions knowingly introduces a nontrivial margin of error into the cost projections, however the lack of direct evidence on many instances of wage and tax fraud leave the authors no choice but to estimate the costs of misclassification in this indirect way.
Despite considerable concerns over the margin of error in the cost estimates using this approach, the authors have reasons to believe that the direct costs of payroll fraud advanced in this study are not only reasonable, but may understate the costs. Beyond the argument that this study is likely undercounting the number of workers affected by payroll fraud (see Discussion in Appendix A), there are two primary direct costs that this study do not incorporate into its projections in order to maintain the most conservative assumptions in the absence of any data. First, legal employers must adhere to regulations imposed by the Occupational Safety and Health Administration. While this may be in the best interest of workers, it nevertheless imposes a substantial cost on legal employers that is often evaded by contractors operating fraudulently. However, since there is no known credible estimate for the cost that safe and responsible contracting imposes on legally-operating employers, it is not included in this analysis.
A second means by which fraudulent employers reduce labor cost that is not captured in the methodology is direct wage theft, or the explicit non-payment of wages promised by an employer. There are anecdotal reports of rampant wage theft among off-the-books workers in the construction industry, especially among the most vulnerable workers (e.g., undocumented laborers); as an example, see Part 1 of this study as well as the 2015 report of Western Massachusetts written by Tom Juravich, Essie Ablavsky and Jake Williams.48 However, while anecdotal reports are plentiful, there are no known estimates for its extent in the national or state construction industry. For the sake of generating conservative estimates, the results in this study assumed there was no wage theft among fraudulent workers. This is an egregiously conservative assumption. But if this report instead assumed that 1% of wages from fraudulent employers were not paid to workers (a rough approximation derived from the Workers Defense Project studies in Texas), the cost impact would be enormous.49 For example, an assumption of 1% wage theft paired with a $35,200 average worker income and 22,146 workers—the most conservative set of assumptions offered in this study—would suggest that wage theft would cost affected workers, on the aggregate, approximately $7 to $8 million. Under the most aggressive worker and income assumptions applied in this paper, the cost of wage theft expands to $14-$16 million. This has enormous ripple effects throughout Massachusetts, harming workers, their families, social programs needed to support them, taxpayers, and law-abiding contractors who are forced to compete against these unscrupulous, cost-shaving employers for the same projects.
Finally, the authors contend that the cost estimates developed in this study undercount the full costs of payroll fraud due to its focus on its direct effects on Massachusetts workers and taxpayers. But there are innumerous indirect effects that degrade Massachusetts communities that are not fully captured in this study. For instance, construction workers who are not eligible for UI and workers’ compensation insurance benefits may have to rely on other state-funded government programs for support during difficult times. Further, payroll fraud puts downward pressure on wages among workers employed by legitimate, law-abiding contractors who may struggle to win project bids when competing against firms who can reduce their labor costs by nearly 30% by operating illegally. State tax revenues from businesses would also be expected to decline as law-abiding (and tax-paying) firms are gradually replaced by dishonest employers (and labor brokers) who are less likely to fully report their income to the Department of Revenue. While these are important societal outcomes that result from payroll fraud, data limitations make conclusions drawn from any such analyses—if even possible—to be tenuous at best and are thus not a part of this study.
Finally, while this report offers an estimate of the dollar value of fringe benefits foregone by workers in the underground construction economy, it is reminded that the failure to offer employer-sponsored health insurance and pension benefits is entirely legal and does not, by itself, constitute payroll fraud. But the erosion of working conditions in the construction industry that results from the presence of these unethical actors will necessarily lead to a “race to the bottom” in the Massachusetts construction industry. While some subsectors may be more insulated than others on the basis of skill, licensing requirements, or other factors, the presence of these bad actors can only further exacerbate the decline in health care coverage of Massachusetts workers, leading to more reliance on public sources of funding.
28 Belman, Dale, and Aaron Sojourner. 2019. “Economic Analysis of Incentives to Fraudulently Misclassify Employees in District of Columbia Construction,” Office of the Attorney General for the District of Columbia.
29 The move to occupation rather than industry is because it is assumed that a vast majority of workers affected by payroll fraud are tradespeople; as a result, using earnings levels of construction occupations removes the influence of construction-industry architects, engineers, salespeople, and other white-collar workers who are likely less representative of the market from which these workers are drawn.
30 Data for occupational earnings for 2019 can be found at: https://www.bls.gov/oes/current/oes_ma.htm.
31 The analysis of DUA records reflects that the gross payroll of workers affected by misclassification as discovered via audit averaged far less than $35,200. While the data does not allow for additional analysis, the short-term nature of informal employment relationships in the construction industry means that it is likely that these payments were for only part-year work, thereby not precluding multiple employment opportunities in a given year.
32 Prior studies have featured only the 25th percentile as the starting point of earnings. However, Massachusetts features the third-highest earnings ($44,960) in the United States—behind only Alaska and Hawaii—at the 25th percentile for workers in construction occupations according to the BLS. As a result of the authors’ concerns that this higher rate may not be entirely representative of conditions for workers in lower-wage trades, the decision was made to rely on the more conservative 10th percentile as a starting point.
33 This study relies on ECEC rates for the construction industry as published in September 2019, available at: https://www.bls.gov/web/ecec/ececqrtn.pdf.
34 This study assumes that workers’ total of $35,200 includes all tax-eligible income payments that find their way on workers’ paychecks, including wages and salaries, paid leave, and supplemental pay.
35 This number is estimated by adding up all tax-eligible income payments made to workers; from the ECEC, this would include the categories of wages and salaries, paid leave, supplemental pay. That sum for September 2019 was $31.14 per hour. Of that, $0.77 per hour was deemed to be from overtime and premium pay. Dividing $0.77 by $31.14 yields 2.47%.
36 This calculation assumes that workers affected by payroll fraud are on the job for the same number of weeks and weekly hours as regular employees. While data on this is unclear—since This was investigated in the 2020 ICERES study, which concluded th. For more, see Ormiston, Russell, Dale Belman and Mark Erlich. 2020. “An Empirical Methodology to Estimate the Incidence and Costs of Payroll Fraud in the Construction Industry.”
37 For more, see: https://www.wcribma.org/mass/IndustryInformation/RateFiling/2020/WCRIBMA_Filing/Filing_2020.pdf.
38 This calculation assumes that workers remain with their respective employer throughout the year and that all of their earnings are paid fraudulently (either via 1099-MISC documentation or by cash payment). These assumptions are suspected to undercount the amount lost by the UI system. Given high levels of worker turnover in the construction industry, it is likely that many workers would be employed by multiple employers in a given year. In that case, each employer would have to contribute to the state UI fund up to the first $15,000 earned by the worker for that company. Nevertheless, in the absence of other data, the authors adhere to the assumption of a single employer in order to provide conservative estimates of UI shortfalls.
39 The rate of $17.15 is deflated by comparing the average rate of employer-sponsored health insurance across the entire industry (61.75%) against the rate for those who earn between $30,000 and $40,000 (57.28%); empirically, the calculation is 17.15*0.5728/0.6175=15.91. To be fair, using the industry-average ratio of $17.15 to calculate the fringe benefit packages of lower-income workers would also have offered validity: the smaller benefit packages would be a product of working with a smaller base income. But this study was compelled to deflate the rate of fringe benefits at the lower-income range to account for the lack of union employers in this income range and for the sake of generating conservative empirical estimates.
40 This study ignores potential mandatory health insurance costs such as the opt-out of the Affordable Care Act. A vast majority of construction firms do not employ the 50+ employees that would make it legally obligated to comply with the law.
41 With a $35,200 assumption, this study estimates that the per-worker labor costs for legal employers is 15.9% higher than firms who operate fraudulently and pay a wage premium to workers who operate as misclassified independent contractors or who work in a cash-only relationship. This is calculated using the earnings of workers in these illegal employment relationships as the denominator. This is not the same as the percent of cash savings using the legal employers’ labor costs as the basis of analysis; using that as the denominator, the estimated differential is 13.7%.
42 When using the labor costs of legal firms as the denominator (instead of those of the fraudulent employers), the cost difference without wage premiums is estimated to be 25.8%. These numbers straddle the industry’s long-held 30% rule of thumb when it comes to estimating the cost savings attributable to payroll fraud.
43 For more, see: Internal Revenue Service. 2016. “Federal Tax Compliance Research: Tax Gap Estimates for Tax Years 2008-2010.” IRS Publication 1415.
44 The lowest rate (23.3%) in this expanded range was not included in the calculation of the incidence of payroll fraud because it produced estimates of the number of workers affected that were so low as to be contradicted by a preponderance of other studies on the issue. That said, the authors include these lower rates here in the absence of confirmatory data otherwise on income underreporting. For more discussion of how these rates were generated, see: Ormiston, Russell, Dale Belman and Mark Erlich. 2020. “An Empirical Methodology to Estimate the Incidence and Costs of Payroll Fraud in the Construction Industry”; Alm, James, and Brian Erard. 2016. “Using Public Information to Estimate Self-Employment Earnings of Informal Suppliers,” Public Budgeting & Finance, 36(1), 22-46.
45 Per-worker tax estimates derived from looking at breakdowns from the Tax Foundation: https://taxfoundation.org/2019-tax-brackets and https://taxfoundation.org/state-individual-income-tax-rates-and-brackets-for-2020/.
46 One counterbalance to the conservative nature of estimated federal income tax losses is that this study does not consider potential earned income tax credit (EITC) benefits.
47 This assumes that workers affected by payroll fraud follow the same residence and commuting patterns as regular employees. While it could be argued that workers affected by payroll fraud may be more likely to be local—thus suggesting that the number of in-state workers used in this study could be an underestimate—the absence of any data on this issue compels the authors to rely on the overall industry average for all workers.
48 For more, see; Juravich, Tom, Essie Ablavsky, and Jake Williams. 2015. “The Epidemic of Wage Theft in Residential Construction in Massachusetts,” UMass-Amherst Working Paper Series.
49 For more, see: Workers Defense Project. 2013. “Building a Better Texas: Construction Conditions in the Lone Star State”; Workers Defense Project. 2009. “Building Austin, Building Injustice”.