Skip to content 
Massachusetts runs two kinds of employer audit. Random selection draws firms from the unemployment insurance database regardless of behavior. Targeted selection follows something specific: a claim filed by someone treated as a contractor, filings that do not reconcile, a worker complaint, a referral from another agency, or a prior violation.
“We are careful, we file on time, and we still got selected. What did we do?”
“A former contractor filed for benefits. Is that all it takes?”
“Our industry gets mentioned a lot in this context. Are we simply more likely to be looked at?”
Two Tracks, Not One
The first question at the top has an answer most employers find unsatisfying and should find reassuring: possibly nothing.
Massachusetts operates employer audits on two tracks. One draws firms at random from the unemployment insurance database. The other targets specific businesses based on past violations, tips from workers, and other factors. Both are described in an analysis of Department audit data by the UMass Labor Center.
A randomly selected employer has not been accused of anything and has not appeared on anyone’s list. The selection is statistical, and the review that follows is identical to a targeted one.
That symmetry is the part worth internalizing. Being selected for no reason produces the same examination, the same records request, and the same determination as being selected for a very specific reason.
It also explains why good behavior is not a shield. An employer with spotless filings and no complaints is still inside the population random selection draws from, which means the only durable protection is being right rather than being unnoticed.
What the Data Actually Shows
There is unusually good evidence here, because researchers have analyzed the outcomes of real Department audits rather than speculating about them.
In a study of Massachusetts construction employers, audits conducted over a three-year period found misclassification at a rate of roughly one in six firms. Among purely random audits, close to 17% of employers were determined to be misclassifying at least one worker.
The second figure matters more than the first. Firms found to be misclassifying were not doing it once. The average offending employer had misclassified around eleven workers, representing close to a fifth of that employer’s workforce.
That is the shape of the risk in a single sentence: it is not rare, and where it exists it is rarely isolated. The findings relate to the construction industry specifically, and the mechanism they describe applies wherever contractor engagements are routine.
The Triggers, and What Each One Opens
| The Trigger | What the Agency Sees | What It Usually Means for Scope |
| A claim from a contractor | Someone requesting benefits with no reported wages from you | Narrow at the start, then every worker engaged on the same terms |
| Filings that do not reconcile | State wage reports against federal payroll filings and returns | The periods where the figures diverge, and the reason for the divergence |
| Zero wages with payroll activity | No reported wages beside payroll-linked expenses or coverage | Whether anyone was working and simply never appeared on a wage report |
| A worker complaint or tip | A specific account of how the work was actually performed | The arrangement described, and anyone else engaged the same way |
| A referral from another agency | Findings from a revenue, labor, or insurance examination | Whatever the other agency found, reapplied under the unemployment test |
| Measurement | Random selection needs no trigger at all | Behavior changes the odds of being selected. It does not change what a review finds once it starts |
The Claim Route
The second question gets the shortest answer in this article: yes, that is all it takes.
When someone treated as a contractor files for benefits, the agency has to decide whether the work was employment, because eligibility depends on it. Under Chapter 151A, services are treated as employment unless the business proves otherwise, so the question opens with the employer already on the back foot.
It is rarely adversarial at the outset. Most of these people left on good terms, made no complaint, and simply applied for a benefit they believed they were entitled to. The system then asks a question it is required to ask.
What makes it consequential is that the answer cannot stay personal. If the arrangement fails the ABC test, everyone engaged on those terms is exposed, and the wage reporting for those quarters comes into question with them.
The Numbers Behind Selection
- 2: the selection tracks, random and targeted.
- About 1 in 6: the construction employers found misclassifying in the audits studied.
- About 11: the workers the average offending employer had misclassified.
- Nearly a fifth: the share of that employer’s own workforce affected.
- 1: the claim it takes to open a classification question.
- 0: the accusations behind a random selection.
The Mismatch Route
Data reconciliation is the quietest trigger and the easiest to prevent, because it involves no judgment at all.
State wage reports and federal payroll filings describe the same payroll from two directions. When the two sets of figures do not agree, the discrepancy is visible without anyone investigating anything, and an inquiry follows.
The pattern that draws attention fastest is a business reporting no wages while continuing to show payroll-linked activity, whether that is coverage, payroll expenses on a return, or anything else implying people are working.
Preventing this is a bookkeeping routine rather than a strategy: reconcile the quarterly wage report against the federal filings and the ledger every quarter, and investigate a variance immediately rather than carrying it forward.
Late or missing quarterly reports belong in the same category. Where a report is not filed, the Department can estimate the liability, and an estimate built without the employer’s figures is both wrong and visible, which makes the account worth a closer look on its own.
The Referral Route
Agencies share findings, and a classification question answered in one place tends to be asked again in another.
A revenue examination that reclassifies workers for withholding purposes, a workers’ compensation audit that finds uninsured labor, or a labor investigation into wage practices can all produce information that reaches the unemployment side.
The practical consequence is that no classification problem stays contained. A business resolving one agency’s findings should assume the same facts will be examined elsewhere, and plan the response accordingly rather than treating each as unrelated.
Sequencing matters when that happens. Settling quickly with one agency to make a problem go away can produce admissions and figures that the next agency reads as a starting point, which is why the order of engagement is worth deciding deliberately.
Industry and Pattern
The third question deserves honesty. Yes, some sectors attract more attention, and the reason is that they generate more findings rather than that anyone dislikes them.
Construction is the most studied example, and the pattern is not unique to it. Cleaning and facilities, delivery and courier work, home care, salons, staffing, and professional services firms that deliver through long-term subcontractors all share the same structural feature: the people doing the work are doing the work the business sells.
That is precisely the feature the classification test examines, which is why concentration by industry is a symptom rather than a cause. The full analysis sits in the reemployment tax audit guide.
What Actually Reduces the Risk
Two things, and only one of them affects selection.
Reconciliation reduces the odds of being selected, because it removes the mismatches that flag an account. Clean quarterly filings that agree with federal returns and the general ledger simply do not generate the anomalies targeted selection looks for.
Classification determines the outcome once selected, and it is the one that matters. A business whose arrangements survive the test has nothing to find, whether it was chosen at random or for a reason, and the first days after a notice are far less stressful when the answer is already known.
Ed Parsons CPA reviews classification and reconciles payroll filings for Massachusetts employers before anyone asks, and represents businesses through examinations when they arrive, under the firm’s DUA audit help service, with the Department’s employer pages as the official reference throughout.
Where a notice has already arrived or an assessment has issued, a Business CPA Tax Resolution Case Analysis prices the exposure before anything is signed or paid.
