We use event studies to examine changes in intergenerational financial transfers and informal care within families following wealth loss, job exit, widowhood, and health shocks. We find sharp reductions in giving to adult children following negative shocks to parents' wealth and earned income, particularly in low-wealth households. Giving also decreases with some health shocks and increases following spousal death. Meanwhile, children of low-wealth households increase financial transfers to parents following adverse shocks in parental households and children of both high- and low-wealth households sharply increase their provision of informal care to parents following a wide range of adverse shocks.

Objective: To evaluate changes in dual enrollment after Affordable Care Act Medicaid expansion by VA priority group, (e.g., service connection), sex, and type of state expansion. Study Setting: Our cohort was all Veterans ages 18-64 enrolled in VA and eligible for benefits due to military service-connection or low income from 2011 to 2016; the unit of analysis was person-year. Study Design: Difference-in-difference and event-study analysis. The outcome was dual VA-Medicaid enrollment for at least 1 month annually. Medicaid expansion, VA priority status, whether a state expanded by a Section 1115 waiver, and sex were independent variables. We controlled for race, ethnicity, age, disease burden, distance to VA facilities, state, and year. Data Extraction Methods: We used data from the VA Corporate Data Warehouse (CDW) regarding age and VA Priority Group to select our cohort of VA-enrolled individuals. We then took the cohort and crossed checked it with Medicaid Analytic Extract (MAX) and T-MSIS Analytic Files (TAF) to determine Medicaid enrollment status. Principal Findings: Service-connected Veterans experienced lower dual-enrollment increases across all sex and state-waiver groups (3.44 percentage points (95% CI: 1.83, 5.05 pp) for women, 3.93 pp (2.98, 4.98) for men, 4.06 pp (2.85, 5.27) for non-waiver states, and 3.00 pp (1.58 to 4.41) for waiver states) than Veterans who enrolled in the VA due to low income (8.19 pp (5.43, 10.95) for women, 9.80 pp (7.06, 12.54) for men, 10.21 pp (7.17, 13.25) for non-waiver states, and 7.39 pp (5.28, 9.50) for waiver states). Conclusions: Medicaid expansion is associated with dual enrollment. Dual-enrollment changes are greatest in those enrolled in the VA due to low income, but do not differ by sex or expansion type. Results can help VA identify groups disproportionately likely to have potential care-coordination issues due to usage of multiple health care systems.

BACKGROUND:Measuring and assessing the relationship between inpatient nurse staffing and workload across a national health system is difficult because of challenges in systematically observing inpatient workload at the unit level. OBJECTIVE:The objective of this study was to apply a novel measure of inpatient nurse workload to estimate the relationship between inpatient nurse staffing and nurse workload at the unit level during a key nursing activity: the peak-time medication pass. METHODS:A retrospective observational study was conducted in the Veterans Health Administration, the largest employer of nurses in the United States. The sample included all patients ( n = 1,578,399 patient days) admitted to 311 non-intensive care unit inpatient acute care units in 112 hospitals in 2019 (104,588 unit days). Staffing was measured as the unit-level, nurse-to-patient ratio, and workload was measured using average time (duration) for RNs to complete the peak-time medication pass. RESULTS:We found a negative relationship between the RN-to-patient ratio and average peak-time medication pass duration after adjusting for unit-level patient volume and average patient severity of illness and other unit-level factors. This relationship was nonlinear: The marginal effect of staffing on workload decreased as staffing increased. DISCUSSION:As unit-level nurse staffing increased, average RN workload decreased. This result suggests that interventions to improve nurse staffing may have larger nonlinear effects for units with lower staffing levels. Understanding the effect of differing staffing decisions on variations in nursing workload is critical for adopting models of care that effectively use scarce staffing resources and contribute to retaining nurses in the inpatient workforce. This work provides evidence that peak-time medication pass duration is a valid process-based measure of workload and highlights the potential diminishing returns to increasing staffing.

OBJECTIVE To measure key characteristics of the Veterans Health Administration's (VHA) Community Care (CC) referral network for screening colonoscopy and identify market and institutional factors associated with network size. DATA SOURCES VHA electronic health records, CC claim data, and National Plan and Provider Enumeration System. STUDY DESIGN In this retrospective cross-sectional study, we measure the size of the VHA's CC referral networks over time and by VHA parent facility (n = 137). We used a multivariable linear regression to identify factors associated with network size at the market-year level. Network size was measured as the number of physicians who performed at least one VHA-purchased screening colonoscopy per 1000 enrollees at baseline. DATA EXTRACTION Data were extracted for all Veterans (n = 102,119) who underwent a screening colonoscopy purchased by the VHA from a non-VHA physician from 2018 to 2021. PRINCIPAL FINDINGS From 2018 to 2021, median network volume of screening colonoscopies per 1000 enrollees grew from 1.6 (IQR: 0.6, 4.6) to 3.6 (IQR: 1.6, 6.6). The median network size grew from 0.63 (IQR: 0.30, 1.26) to 0.92 (IQR: 0.57, 1.63). Finally, the median procedures per physician increased from 2.5 (IQR: 1.6, 4.2) to 3.2 (IQR: 2.4, 4.7). After adjusting for baseline market characteristics, volume of screening colonoscopies was positively related to network size (β = 0.15, 95% CI: [0.10, 0.20]), negatively related to procedures per physician (β = -0.12, 95% CI: [-0.18, -0.05]), and positively associated with the percent of rural enrollees (β = 0.01, 95% CI: [0.00, 0.01]). CONCLUSIONS VHA facilities with a higher volume of VHA-purchased screening colonoscopies and more rural enrollees had more non-VHA physicians providing care. Geographic variation in referral networks may also explain differences in the effects of the MISSION Act on access to care and patient outcomes.

OBJECTIVE To estimate whether those enrolled in the Veterans Health Administration (VHA) were less likely to use VHA-delivered colorectal cancer screening colonoscopies after the MISSION Act. DATA SOURCES AND STUDY SETTING Secondary data were collected on VHA-enrolled Veterans from FY2017-FY2021. STUDY DESIGN This retrospective cross-sectional study measured the volume and share of screening colonoscopies that were VHA-delivered over time and by drive time eligibility-defined as living more than 60 min away from the nearest VHA specialty-care clinic. We used a multivariable logistic regression to adjust for patient and facility factors. DATA EXTRACTION Data were extracted for VHA enrollees (n = 773,766) who underwent a screening colonoscopy either performed or purchased by the VHA from FY2017-FY2021. PRINCIPAL FINDINGS In the 9 months after the implementation of the MISSION Act, and before the onset of the Covid-19 pandemic, the average monthly VHA-share of screening colonoscopies decreased by 3 percentage points (pp; 95% confidence interval [CI] = [-4 to -2 pp]) for the non-drive time eligible group and it decreased by 16 pp (95% CI = [-22 to -9 pp]) for the drive time eligible group. The total number of screening colonoscopies did not significantly change in either group during this time period. After adjusting for patient characteristics, a linear time trend, and parent facility fixed effects, implementation of the MISSION Act was associated with a reduction in the probability of a VHA-delivered screening colonoscopy (average marginal effect [AME]: -2.5 pp; 95% CI = [-5.1 to 0.0 pp]) for the non-drive time eligible group. The drive time eligible group (AME: -9.4 pp; 95% CI = [-13.2 to -5.5 pp]) experienced a larger change. CONCLUSIONS The VHA-share of screening colonoscopies among VHA enrollees fell in the 9 months immediately after the passage of the MISSION Act. This decline was larger for VHA enrollees who were targeted for eligibility due to a longer drive time. These results suggest that the MISSION Act led to more VHA-purchased care among targeted VHA enrollees, though it is unclear whether total utilization increased.

Dialysis care in the United States is expensive and is mostly paid for by Medicare. To reduce the cost of providing dialysis services, the federal government has relied on a law that designates Medicare as a secondary payer in the first 30 months of dialysis. During this period, private health insurers are the primary payer and pay for the majority of dialysis-related costs. Private health insurers often pay substantially higher prices for dialysis care than does Medicare, possibly due to highly concentrated dialysis-provider markets. A perspective by Boumil and Curfmin in this journal discusses how a recent ruling by the US Supreme Court may limit Medicare's role as a secondary payer, potentially altering the economic relationship between dialysis providers and private insurers. Boumil and Curfmin discuss how these changes may ultimately promote competition in dialysis-provider markets and lower dialysis-related costs paid by private health insurers. We compare this viewpoint to responses and concerns voiced by other stakeholders in the kidney-care community and outline additional ways in which the Supreme Court's ruling may affect competition in dialysis markets and prices paid for dialysis by private insurers.

Mixed-motive negotiations are strategically, socially, and motivationally complex. Theorists argue that multiparty negotiations exacerbate these complexities leading to inefficient outcomes. In contrast, we propose that multiparty negotiators respond to the complexities of their negotiation context by using strategy differently than dyadic negotiators and that one reason why is that they arrive at the negotiation table with weaker fixed pie perceptions than dyadic negotiators. Results across three studies showed that multiparty negotiators had weaker fixed pie perceptions compared to dyadic negotiators. Multiparty negotiators also engaged in more integrative strategic behavior and used more complex complementary and structural sequences of behavior than dyadic negotiators. These differences in use of strategy helped multiparty negotiators achieve Pareto efficient outcomes equal to those of dyads. To obtain a conservative estimate of the effect size of context on fixed pie perceptions, we conducted an internal meta-analysis of our file drawer.

While there has been broad interest in the direct effects of major life events on older households that experience them, little attention has been paid to the intergenerational transmission of those effects — how negative shocks in parents’ households affect the outcomes of their adult children — or to the role that grown children play in helping their parents recover from adverse events. We use regression and event study approaches to examine within-family changes in monetary transfers and informal care following wealth loss, involuntary job displacement, spousal death, and health shocks in retirement-aged households. We find that giving to adult children is responsive to changes in parents’ wealth and earned income. We document large reductions in the likelihood of making financial transfers to children following wealth loss and job displacement, particularly in households with low accumulated wealth. We also find that parents increase their transfers following spousal death and reduce them with the onset of disability or poor health. We find that upstream transfers are also responsive to life events — children, particularly those with low-wealth parents, increase their financial transfers and in-kind assistance following adverse shocks in their parents’ households.

I evaluate the degree of misclassification error when identifying job stayers in the Current Population Survey using the method proposed by Card and Hyslop (1997). I find that even though their method misclassifies approximately one-third of hourly workers, these errors do not substantially impact estimates of downward nominal wage rigidity, a common application. However, I find that misclassification error influences the results in studies that require estimating the share of the population that are job stayers and changers and their wage changes, such as studies of wage adjustment, and develop a correction for it.

The Supplemental Nutrition Assistance Program (SNAP) provides food assistance to nearly 44 million Americans each year. I document a substantial increase in the program’s ability to stimulate food consumption from 1990 to 2010, as measured by the marginal propensity to consume food (MPCf) out of SNAP. I provide the first evidence for a mechanism driving this increase: the transition from paper coupons to Electronic Benefit Transfer (EBT) cards. Using plausibly exogenous variation over states and time I estimate that the introduction of EBT doubles the MPCf out of SNAP and accounts for 25 percent of its observed increase. ∗The University of Arizona, Department of Economics, 1130 E Helen St., McClelland Hall Rm 401, Tucson, AZ 85721 (email:eckcs1@email.arizona.edu). I would like to thank Martin Dufwenberg, Price Fishback, Hidehiko Ichimura, Michael Kuhn, Ashley Langer, Jessamyn Schaller, Gary Solon, Tiemen Woutersen, and seminar participants at The University of Arizona for their excellent comments and suggestions. Many of the 43 million Americans living in poverty struggle to consume enough food. For children who grow up in poverty, malnutrition is associated with lower childhood and adult health (Cook et al., 2004). To address this problem, Congress created the Supplemental Nutrition Assistance Program (SNAP; formerly called the food stamp program) in 1964 to help people living at or below the poverty line gain access to a secure source of food. SNAP distributes money that can only be used on eligible food items at qualified retailers. Today, SNAP is the second largest in-kind transfer program in the United States and distributes $66.5 billion in food assistance to 44 million people (USDA, 2017). By providing in-kind, rather than cash, benefits policymakers hope that SNAP recipients will consume more food than they would with an equivalent cash transfer (Currie and Gahvari, 2008). However, there is a long-standing debate over whether in-kind transfers, and SNAP benefits in particular, are better at steering consumption towards certain types of goods. Standard demand theory predicts that inframarginal households—those that would spend more on food than they receive in SNAP benefits—will treat SNAP benefits and cash as perfect substitutes (Southworth, 1945). This result implies that, for the majority of SNAP households, SNAP benefits and cash should lead to similar increases in food consumption. Despite the clear theoretical predictions, the empirical evidence on how much SNAP increases food consumption is mixed. Most studies have focused on estimating the marginal propensity to consume food (MPCf) out of SNAP to measure the effect of SNAP benefits on total food consumption. The MPCf out of SNAP indicates how much food expenditures rise in response to a $1 increase in SNAP benefits. As shown in Figure 1, from the beginning of the program until the early 1990s, the evidence suggests that the MPCf out of SNAP was near .1, which is in line with most estimates of the MPCf out of cash (Moffitt, 1989; Schanzenbach, 2002; Hoynes and Schanzenbach, 2009). In other words, inframarginal consumers treated a cash transfer and SNAP benefits in the same way. However, a new wave of studies focusing on consumers in the late 2000s and early 2010s finds the opposite result: SNAP induces much more food consumption than an equivalent cash transfer (Collins et al., 2016; Beatty and Tuttle, 2015; Bruich, 2014; Hastings and Shapiro, 2018). Methodological differences between the two sets of studies are insufficient to explain the increase in estimated MPCfs over time. In each sample period researchers have used both experiments (Moffitt, 1989; Schanzenbach, 2002; Collins et al., 2016) and exogenous policy changes (Hoynes and Schanzenbach, 2009; Beatty and Tuttle, 2015; Bruich, 2014; Hastings and Shapiro, 2018) to identify the causal effects of SNAP on food consumption. Moreover, 76 percent of SNAP households are inframarginal in my sample and this proportion is relatively stable over the sample period.