The passage and implementation of the One Big Beautiful Bill Act (Public Law 119-21) will create significant changes in the way states operate public benefit programs, including Medicaid and SNAP. The new law requires certain Medicaid enrollees to demonstrate that they are working or otherwise engaged in their community as a condition of eligibility in the program. Additionally, the law expands a preexisting work requirement for certain SNAP enrollees, and subjects states to significant new penalties in the form of forced cost-sharing if the calculated payment error rate in their SNAP program exceeds certain thresholds.

One policy idea being considered to ease implementation is to increase the cadence of reporting of wage data to state unemployment systems from the current quarterly standard to a per-pay-period frequency (Garden-Monheit et al 2026). The policy would increase the amount of up-to-date data available to benefit programs for eligibility processing.

One benefit of the policy is that it would reduce administrative burden, by reducing the need for applicants and enrollees to provide documentation to receive and maintain benefits. In addition to reducing paperwork burdens on beneficiaries, the use of more frequent earnings data has the potential to improve benefit accuracy by better targeting benefits to eligible recipients and ending receipt for recipients who have become ineligible. Finally, it would also reduce states’ need to buy income data from expensive private data brokers.

However, the move to more frequent reporting of earnings means that benefit administrators will need to adapt to the reality of earnings instability when making decisions regarding benefit administration. Our ongoing research documents widespread earnings instability (Ganong et al 2025) for American workers. Because of earnings instability, more frequent earnings data can increase the frequency of improper redeterminations.

If a state receives information which suggests that a person makes too much to be eligible for Medicaid, then they need to initiate a redetermination. Receiving one unusually large paycheck or one month in which earnings are unusually high could be used to initiate this process. As part of that redetermination, the state will request information from the individual. If the individual doesn’t respond to that request for information, they are procedurally terminated from the program—even if they are in fact eligible for the program. An improper redetermination occurs when a state agency receives earnings data that leads the agency to initiate a request for information for someone who is actually eligible. The initiation of such improper redeterminations negatively affects program recipients both by creating additional administrative burden and also by lowering the actual Medicaid enrollment of eligible citizens who are unable to overcome those burdens to establish their eligibility (Meyerson, Espeseth, and Dague 2026).

In this memo, we define a household as eligible when their income over a six-month horizon is low enough that they would be eligible for Medicaid; concretely, this means that the household’s income is below 138% of the poverty line. We focus on Medicaid eligibility, but we note that similar issues arise for other benefit programs such as SNAP. The calculations in this memo rely on a synthetic dataset built from the Survey of Income and Program Participation (SIPP) plus administrative payroll data.

This memo shows that more frequent earnings data may lead to more improper redeterminations through two specific channels. In each case, we describe how alternative implementation choices can reduce the rate of improper redeterminations. Through these implementation choices, policymakers can obtain the upside benefits of a pay-period reporting cadence—i.e., reduced paperwork burdens, increased program integrity, and lower costs of program administration—while mitigating the downside risk of improper redeterminations.

Going forward, states will reevaluate Medicaid recipients every six months. One way in principle that a state could use more frequent earnings data is to disqualify any worker whose paycheck covering two weeks of work if multiplied by 13 would be higher than the Medicaid income limit for that time period. Figure 1 shows a fictional earnings history for a worker who is eligible for Medicaid. This worker would be improperly redetermined five times if a single paycheck can be considered information of a change in circumstances for an individual’s Medicaid eligibility and would be improperly redetermined twice if two consecutive paychecks can be considered information of a change in circumstances for an individual’s Medicaid eligibility. However, if we were to increase the number of consecutive pay periods required to three pay periods (or more), no improper redetermination would occur.

We find that improper redeterminations may occur frequently in the absence of a a floor on the number of paychecks that trigger a change report. For example, suppose that a state decides to redetermine the eligibility of a worker who has two consecutive paychecks with average income above the FPL threshold. Over the course of a six-month time period—which is the usual time horizon used to verify Medicaid eligibility—a benefit administrator who is evaluating whether a recipient remains eligible will file a change report for about 30 percent of recipients. With the benefit of hindsight, if they had instead waited for six months of data to assess eligibility, they would find that more than a third of those flagged for redetermination remained eligible.

Figure 2 shows how different time horizons affect the proper and improper redetermination rates. One scenario shown is the two paychecks scenario described above, in which 12.9% of recipients receive an improper redetermination but are in fact eligible for the program. The figure shows that the percentage of recipients who are eligible but are flagged for redetermination decreases dramatically as the number of consecutive periods required increases. For example, if we require six consecutive pay periods before triggering a change report, the number of eligible people for whom a change report is incorrectly filed drops to only 1.0 percent.

This figure therefore shows a tradeoff in benefit accuracy. Waiting for data from additional paychecks reduces the rate of improper redeterminations, but at the same time, there is a longer delay before proper redeterminations occur.

Another challenge for using more frequent pay period data is that Medicaid program eligibility is usually measured using monthly income, but monthly income has artificial volatility for workers who are paid fortnightly—as is the case for the plurality of U.S. workers. This challenge arises when a state agency uses a one-month time window to determine Medicaid eligibility. Figure 3 illustrates an example of how this issue can arise. In this illustrative example, the worker receives exactly $750 in every single paycheck. This amount of income puts them below the eligibility threshold when looking at any individual paycheck and also when looking across a six-month time horizon. However, the worker receives three paychecks in April (April 1, April 15, and April 29) for a total of $2,250 per month, which would result in an improper redetermination.

To measure how frequently the situation described above arises, we calculate monthly income using two different methods. In the first method, the benefit administrator computes monthly income using just check date; we call this “naive income.” In the second method, the benefit administrator collects the pay period start and end dates and assigns income to months by spreading it equally within each pay period; we call this “true income” because it correctly assigns income to the month it was earned (rather than the day when it happened to be paid to the worker). Eligibility is determined using the month with the highest earnings under each method over the course of a 12-month period. We draw a sample of workers who are eligible for Medicaid using true income.

Figure 4 shows that many workers will be improperly redetermined using check date. The figure shows a scatter plot where each dot is one household in the SIPP. Using check dates inflates the highest month of pay compared to pay period dates by artificially inflating volatility. The figure shows this pattern because most of the households in the figure have higher naive monthly income than they do true monthly income. The eligible workers most likely to be improperly redetermined are those who are doing the most work—those with true incomes which are between 70% and 99% of the Medicaid eligibility threshold. Summarizing the figure quantifies the size of this effect: it shows that 1-in-5 Medicaid eligibles will be improperly redetermined if naive income is used.