California nursing homes were 158 times as likely to report 99 beds as 100 beds in historical records examined by researchers studying a staffing rule. Their October paper estimates that avoiding the regulatory cutoff restricted capacity while producing little additional registered-nurse staffing among facilities that remained above it.
The National Bureau of Economic Research working paper, by Alden Cheng and Riley League, examines a requirement for facilities with at least 100 beds to provide round-the-clock registered-nurse coverage.
The result concerns the design of that size-based rule, rather than an assessment of whether nurses improve care. It shows how a requirement intended to raise service standards can also influence the scale at which businesses choose to operate.
A cutoff can change the cost of expanding
A regulatory threshold sets a point at which a different obligation applies. Depending on the rule, crossing it might mean hiring additional staff, producing more detailed reports or meeting a different operating standard.
Such thresholds can spare smaller businesses costs they would struggle to absorb. They can also make a small increase in size disproportionately expensive.
In the California setting studied, moving from 99 beds to 100 brought a requirement for 24-hour registered-nurse coverage. Registered nurses have different qualifications and responsibilities from licensed vocational nurses, and employing them cost more in the researchers’ historical data.
An operator considering one additional bed consequently had to account for a staffing obligation extending beyond the revenue from that bed. The researchers identify a concentration of facilities immediately below the cutoff, a pattern economists call bunching.
The comparison is between reported facility sizes. It is not evidence that each home physically removed a bed, or that 158 homes closed whenever one expanded.
Higher staffing above the line can be misleading
The study combines federal cost reports, inspection records and payroll-based staffing data. Its combined sample covers nearly 250,000 facility-year observations from 2000 to 2019, including more than 20,000 observations for California. A facility appearing in several years contributes several observations.
Payroll-based staffing records, available from 2017, are the authors’ principal staffing source. They examine registered-nurse hours per bed-day, which adjusts staffing for facility size.
A simple comparison could show higher staffing in homes above 100 beds and suggest the rule had worked. The difficulty is that facilities can choose their size. Homes already employing more nurses may be more willing to stay above the cutoff, while those facing greater additional costs may remain below it.
Cheng and League develop a statistical method to separate this selection from changes attributed to the regulation. They estimate that selection accounts for nearly all the observed staffing difference around the threshold.
The estimate depends on assumptions about how facility size and staffing would behave without the cutoff. The research is a working paper and has not been peer-reviewed. It is not a randomized trial assigning otherwise identical homes different rules.
The capacity cost falls outside the staffing measure
The authors estimate that homes bunching below the threshold would otherwise have averaged 113.8 beds. That is a modeled alternative to the observed rule, not an observed increase in capacity after a policy change.
Their analysis associates the capacity restrictions particularly with lower-income areas and ZIP codes with higher Hispanic population shares. Those are characteristics of areas around the facilities, rather than a direct measurement of every affected patient’s income or ethnicity.
Our examination of growth opportunities in elderly care described demand associated with aging populations. Greater demand cannot automatically produce more places when providers face staffing costs and incentives to limit their scale.
For policymakers, evaluating a size-based rule requires looking at both sides of the cutoff. Measuring staffing only among regulated facilities can miss businesses that adjust their capacity to avoid entering that group.
The paper also has a narrower measurement limit. Daily registered-nurse hours cannot establish how those hours were spread across each day. The researchers can assess total staffing responses, but cannot directly determine continuous coverage from those totals.
