Over the past two decades, the United States has witnessed a spectacular increase in interior immigration enforcement. Since the inception of the U.S. Department of Homeland Security (DHS) after 9/11, the budget of Immigration Customs Enforcement (ICE)–the branch in charge of enforcing immigration law in the interior of the United States—has more than doubled. Its 2020 budget of $8.4 billion is almost equivalent to the $9.7 billion appropriated to the Federal Bureau of Investigation (FBI), far exceeding the $2.3 billion appropriated to the U.S. Drug Enforcement Administration (DEA) (Council, 2020). A broad literature has documented and examined the many ramifications of intensified immigration enforcement on undocumented immigrants, their families, and the communities in which they reside (e.g., Miles and Cox, 2014; Alsan and Yang, 2018; Hines and Peri, 2019; Wang and Kaushal, 2019; Amuedo-Dorantes et al., 2022; East and Velásquez, 2022). In this paper, we focus on yet another consequence of intensified immigration enforcement on a growing population segment frequently relying on low-education domestic workers, many of whom are undocumented, for help with basic daily tasks–namely, American elderly. We examine how the implementation of Secure Communities, an interior immigration enforcement initiative responsible for the vast majority of the 200% increase in deportations from 2003 through the 2010s (East, 2020), led to increased institutionalization of older Americans through its impacts on the labor supply of workers in the household services market.

The importance of assessing the impact of immigration policy on the elderly cannot be overstated. The U.S. population is aging. Over the next 40 years, the U.S. population is expected to grow 25%. However, the number of people aged 65 and above is expected to grow by 92%, and the number of people aged 85 and above by an astonishing 198% (Vespa et al., 2018). The 2019 American Community Survey shows that within the elderly population, the likelihood of residing in an institutional setting quickly rises with age. While only 2.4% of individuals 65+ years old live in an institution, a total of 8.3% of 85+ years old do (ACS, 2019). In addition to the out-of-pocket costs paid by individuals and private long-term care insurance providers, Medicaid and Medicare spent roughly 93 billion dollars on institutional care in 2020 –a sum that is projected to reach 153 billion per year by 2030 (CMS, 2020). Yet, responses to surveys conducted before the COVID-19 pandemic suggest that the elderly overwhelmingly prefer to age in place, either in their own homes or in the homes of loved ones (Binette and Vasold, 2018). These preferences are likely to have exacerbated following the pandemic given the overrepresentation of nursing home residents and staff among COVID-19 fatalities.

There is causal evidence that, in places with more immigrants, U.S.-born elderly are less likely to live in institutionalized settings (Butcher et al., 2021). This is presumably because immigrants are over-represented in household services sectors, such as housekeeping and home health aides, that the elderly needing assistance with independent living typically rely upon to age in place. We dig deeper into this finding to learn about the role of immigration policy in shaping the institutionalization rates of the elderly.

To that end, we focus on one critical policy, namely, the Secure Communities (SC) program. SC is an interior immigration enforcement initiative implemented in phases across U.S. counties between 2008 and 2014. Under SC, information gathered by local police after an arrest, including fingerprints, is entered into an electronic system connecting the Federal Bureau of Investigation (FBI) and Immigration Customs Enforcement (ICE) databases. Undocumented immigrants are flagged by the system, likely leading ICE to issue a detainer request asking the law enforcement agency to hold the individual for as long as 48 hours after a release date to provide ICE agents enough time to decide whether to take the individual into federal custody to initiate removal proceedings.

In theory, SC might not have had a discernible impact on the living arrangements of elderly Americans since most deported immigrants during that period were men (Carroll, 2014), whereas practically all domestic workers providing in-home support for the elderly are women. Nevertheless, SC may have impacted the market for household services –a market in which almost one in five workers are undocumented (Svajlenka, 2020). Specifically, SC could have reduced the number of individuals willing to work as housekeepers, home health workers and personal care aides due to deportation fears based on their own undocumented status or that of family members.

In a study showing that SC led to increases in labor supply of college-educated U.S. born mothers, East and Velásquez (2022) present evidence of SC-induced decreases in the number of hours worked by housekeepers and childcare workers. The lower labor supply in those occupations could stem from increased out-migration from, or decreased in-migration to, counties with a tougher stance on immigration enforcement. Alternatively, their reduced labor supply could arise from the decision of individuals that feel threatened by the policy, either directly or indirectly, to decrease their work hours to circumvent police encounters. In both instances, the adoption of SC would curtail the availability of housekeepers, home health aides, and personal care attendants. Elderly Americans who may have been able to age in place with such assistance may then find it increasingly necessary to rely on institutional long-term care.

Focusing on data from the American Community Survey (ACS) from 2006 through 2012, we examine how a tougher stance on immigration affected institutionalization rates of elderly Americans by exploiting the temporal and geographic variation in the implementation of SC. We find evidence of SC raising the institutionalization likelihood of elderly Americans by 0.26 percentage points (6.8 percent) –an impact that did not predate the policy's adoption and was immediately evident during the first two years following its implementation.1

Heterogeneity analyses reveal that the policy's impact was larger among Whites and college-educated elderly, who are more likely to afford domestic help. In addition, it is those with only moderate levels of disability that were most affected by the policy. Among those residing in a nursing home, the policy adoption was associated with lower disability levels –a finding suggestive of elderly Americans with relatively minor disabilities gravitating towards institutionalized care as domestic help became harder to find. Finally, as we would expect given the demographic characteristics of the undocumented workers directly targeted by the policy, SC had a larger impact on the institutionalization of elderly Americans in localities with a greater share of less educated Hispanics in key health care and housekeeping occupations in the home health and private household services sectors.

To further understand the likely mechanisms at play, we examine the impact of SC on the labor supply of housekeepers, home health workers, and personal care workers. We find evidence of reductions in the number of workers providing such services, although consistent with the findings by East and Velásquez (2022), it is hours of work that were more meaningfully curtailed. Importantly, the labor supply reduction was mostly driven by less-educated Hispanic workers from Mexico and Central America, who are also the ones most directly impacted by SC.

The remainder of this paper is organized as follows. Section 2 provides background information on Secure Communities and discusses how the policy may have impacted the living arrangement choices of elderly Americans. The data and empirical modeling are discussed in Section 3. Section 4 presents our main findings, along with a series of identification checks and heterogeneous analyses. Likely mechanisms are discussed in Section 5, and Section 6 concludes the study.