US workers experience large month-to-month fluctuations in pay

Over the past decade, cities and states in the US have enacted ‘fair workweek’ laws to stabilise worker schedules. This column uses administrative data on US workers’ paycheques and firms’ payrolls to document considerable monthly fluctuations in earnings. Pay instability is widespread, disproportionately hits lower‑paid hourly workers, and is largely driven by firms’ labour demand rather than worker-related determinants such as childcare demands. These fluctuations represent genuine, welfare-relevant risk that materially shapes household decisions and has attendant effects on the economy.

Over the past decade, cities and states have enacted ‘fair workweek’ laws to stabilise worker schedules. New York City’s 2017 Fair Work Week Law requires employers in the fast-food industry to provide regular schedules at the start of employment and give two weeks’ notice for any changes, with worker consent and premium pay required for modifications (Pickens and Sojourner 2025). Chicago requires employers to provide 14 days’ advance notice of employees’ schedules but provides exceptions for events outside of employer control, such as supplier delays or unexpected demand surges (City of Chicago 2019). Los Angeles, Philadelphia, San Francisco, Seattle, and other local governments have enacted or are considering similar regulations.

These policy debates turn on empirical questions: how prevalent is pay volatility, and do workers view these fluctuations as harmful? If earnings instability is widespread and unwelcome, regulations on scheduling practices could improve worker welfare. If fluctuations are rare and operationally necessary, such rules could harm business efficiency without helping workers.

In Ganong et al. (2025), we document that pay volatility is not only pervasive but also profoundly unequal – concentrated among lower-paid workers. Hourly workers, who tend to be lower-income and more financially vulnerable, face far greater earnings swings than their salaried counterparts. For the 60% of the US workforce paid by the hour, these fluctuations shape spending decisions, drive job turnover, and fundamentally alter the experience of work.

Prior research using annual data has documented earnings risk over workers’ careers (Song et al. 2015, Moffitt and Zhang 2018, McKinney and Abowd 2023, Pruitt and Turner 2020). Our study highlights the causes and consequences of monthly income fluctuations, which are invisible in the annual data.

Payroll data show workers experience considerable monthly fluctuations in earnings

We use comprehensive administrative data from workers (via paycheque deposits into Chase bank accounts) and firms (via a payroll processor) to document considerable monthly fluctuations in earnings. Monthly earnings volatility is substantial. In about three-quarters of months, workers receive a different amount of pay than they received the prior month. The median month has a change of 5%, and in one-quarter of months, the change in pay is at least 17%.

This volatility is concentrated among hourly workers. Figure 1 illustrates this heterogeneity in earnings volatility between salaried and hourly workers. For the majority of months, there is no change in earnings for salaried workers. In contrast, for hourly workers, no change in monthly earnings is a rare outcome.

Figure 1 Heterogeneity of earnings volatility

Notes: This histogram shows the distribution of the change in earnings for salaried and hourly workers with continuing employment. Because the number of paycheques varies from month to month, in this plot, the change in earnings is measured in terms of average monthly pay per paycheque. Specifically, the volatility estimate is calculated as the per-cent change in earnings in the current month from the median of the earnings in the three months prior.

These findings raise two further important questions. Why do hours move from month to month? And does this instability matter for worker welfare?

Firms play an important role in driving monthly earnings volatility

We provide evidence that worker-related determinants such as temporary unpaid leave, childcare demands, and seasonal fluctuations do not explain earnings instability; rather, firms are the key to explaining changes in worker hours. Consequently, a meaningful share of this instability is imposed on the worker and is outside their control.

The relationship between the volatility of the firm’s demand for labour – measured as the change in total firm hours – and the volatility of worker pay is shown in Figure 2. There is a positive relationship between firm labour-demand volatility and worker pay volatility; in other words, larger changes in the total amount of hours worked by all employees are associated with larger changes in individual worker pay.

Figure 2 Relationship between firm-level volatility and individual worker volatility

Notes: This figure shows the relationship between firm total-hours volatility and individual worker pay volatility. Firm total-hours volatility is measured by taking the sum of hours for all workers at the firm and computing the median absolute monthly change. Individual worker pay volatility is measured by taking the median absolute monthly change for each worker at the firm. Each dot represents the average individual worker volatility for a group of firms with similar firm-hours volatility.

Income volatility causes spending volatility

Earnings instability affects workers’ wellbeing in two ways. First, using bank-account data, we show that income volatility causes spending volatility. Spending volatility increases when workers move to higher-volatility firms (Figure 3).

Figure 3 Relationship between income and spending volatility

Notes: This figure shows the relationship between individual volatility of monthly income and monthly spending. The volatility measure is the individual’s median absolute change from the prior month. Consumption is measured as spending on non-durable goods and services, using bank-account data. Each dot represents the average consumption volatility for a group of individuals with similar income volatility.

Second, we show that hourly workers are more likely to quit high-volatility jobs, and firm volatility affects quit rates of hourly workers much more than salaried workers in the same firm.

Given that the above volatility is costly to workers, we seek to understand how much workers would pay to avoid earnings volatility. We combine the empirical estimates discussed above with standard economic models of the labour market to price how much workers dislike volatility. The median hourly worker would forgo 4%–11% of their income to attain the more stable income of a median salaried worker. Lower-income workers, who face the highest volatility, would trade an even higher fraction of their earnings for the stability of a salaried position.

Our work shows that workers dislike income volatility, but further work is necessary to understand whether firm scheduling flexibility serves essential business needs.

Bottom line: Workers face substantial monthly earnings risk that, to this point, has been missed when analysing annual data. This risk is borne primarily by relatively low-income, hourly workers and is largely driven by fluctuations in firms’ labour demand, which induces substantial costs on affected workers. This column shows that these earnings fluctuations represent genuine, welfare-relevant risk that materially shapes household decisions and has attendant effects on the economy.

Source : VOXeu