Most Popular Books by JONATHAN GRUBER

We explore the in- and out- of sample robustness of tests for consumer choice inconsistencies based on parameter restrictions in parametric models, with a focus on tests proposed by Ketcham, Kuminoff and Powers (2015). We start by arguing that non-parametric alternatives are inherently conservative with respect to detecting mistakes (and one specific test proposed by KKP is incorrect). We then consider several proposed robustness checks of parametric models and argue that they do not separately identify misspecification and choice inconsistencies. We also show that, when implemented using a comprehensive goodness of fit measure, the Keane and Wolpin (2007) test of out of sample forecasting demonstrates that a model allowing for choice inconsistencies forecasts substantially better than one that does not. Finally, we explore the robustness of our 2011 results to alternative normative assumptions.

In numerous high stakes markets skilled experts play a key role in facilitating consumer choice of complex products. New artificial intelligence (AI) technologies are increasingly being used to augment expert decisions. We study the role of technology and expertise in the market for health insurance, where consumer choices are widely known to be sub-optimal. Our analysis leverages the large-scale implementation of an AI-based decision support tool in a private Medicare exchange where consumers are randomized to skilled agents over time. We find that, prior to AI-based technology, skilled experts in this market exhibit the same type of inconsistent behavior found in previous studies of individual choices, costing consumers $1260 on average. The addition of AI-based decision support improves outcomes by $278 on average and substantially reduces heterogeneity in broker performance. Experts efficiently synthesize private information, incorporating AI-based recommendations along dimensions that are well suited to AI (e.g. total expected patient costs), but overruling AI-based recommendations along dimensions for which humans are better suited (e.g. specifics of doctor networks). As a result, switching plans, an ex-post measure of plan satisfaction, is meaningfully lower for agents making AI-based recommendations. While AI is a complement to skill on average, we find that it is a substitute across the skill distribution; lower quality agents provide better recommendations with AI than the top agents did without it. Overall productivity rises, with the introduction of decision support associated with a 21% reduction in call time for enrollment.

Although the vast majority of working individuals aged 55-64 receive health insurance coverage through their employment, many of these individuals face the prospect of losing such coverage should they retire before becoming eligible for guaranteed public coverage through Medicare at age 65. Because the expected medical expenses of this group are large and uncertain, the availability of health insurance coverage after retirement could be a key factor in the retirement decision of older workers. We examine the effect of health insurance on retirement by looking at variation in state and federal 'continuation of coverage' mandates, laws which allow individuals to continue purchasing health insurance through a previous employer for a specified number of months after leaving the firm. By allowing individuals to maintain their employer-provided health insurance after retirement, these laws decrease the cost of early retirement for those who do not have other retiree health insurance available. Using data on 55-64 year old men from the Current Population Survey, we find that one year of continuation benefits increases the probability of being retired by 1 percentage point; this represents a 5.4 percent increase in the baseline probability of being retired for this group. We also find that continuation mandates increase the likelihood of being insured after retirement.

Cities with a larger concentration of scientists have been shown to be more productive places for additional scientists to do Research and Development. At the same time, these urban areas tend to be associated with higher costs of doing research, in terms of both wages and land. While the literature on the benefits of agglomeration economies is extensive, it offers no direct evidence of how productivity gains from agglomeration compare with higher costs of production. This paper aims to shed light on the balance between local productivity and local costs in science. Using a novel dataset, we estimate place-based costs of carrying out R&D in each US metro area and assess how these place-based costs vary with the density of scientists in each area. We then compare these costs with estimates of the corresponding productivity benefits of more scientist density from Moretti (2021). Adding more scientists to a city increases both productivity and production costs, but the rise in productivity is larger than the rise in production costs. In particular, each 10% rise in the stock of scientists is associated with a 0.11% rise in costs and a 0.67% rise in productivity. This implies that firms moving from cities with a small agglomeration of scientists to cities with a large agglomeration of scientists experience productivity gains that are 6 times larger than the increase in production costs. This finding is consistent with the increased concentration of R&D activity observed over the past 30 years. However, while the productivity estimate has only modest non-linearities, the cost estimates suggest much larger non-linearities as the concentration of scientists increases. For the most concentrated R&D cities, the difference between productivity gains and cost increases is close to zero.

Using a budget-based approach to measuring affordability, this issue brief explores whether the subsidies available through the Affordable Care Act are enough to make health insurance affordable for low-income families. Drawing from the Consumer Expenditure Survey, the authors assess how much "room" people have in their budget, after paying for other necessities, to pay for health care needs. The results show that an overwhelming majority of households have room in their budgets for the necessities, health insurance premiums, and moderate levels of out-of-pocket costs established by the Affordable Care Act. Fewer than 10 percent of families above the federal poverty level do not have the resources to pay for premiums and typical out-of-pocket costs, even with the subsidies provided by the health reform law. Affordability remains a concern for some families with high out-of-pocket spending, suggesting that this is the major risk to insurance affordability.

Concerns over the impacts of hospital strikes on patient welfare led to substantial delay in the ability of hospitals to unionize. Once allowed, hospitals unionized rapidly and now represent one of the largest union sectors of the U.S. economy. Were the original fears of harmful hospital strikes realized as a result? In this paper we analyze the effects of nurses' strikes in hospitals on patient outcomes. We utilize a unique dataset collected on nurses' strikes over the 1984 to 2004 period in New York State, and match these strikes to a restricted use hospital discharge database which provides information on treatment intensity, patient mortality and hospital readmission. Controlling for hospital specific heterogeneity, patient demographics and disease severity, the results show that nurses' strikes increase in-hospital mortality by 19.4% and 30-day readmission by 6.5% for patients admitted during a strike, with little change in patient demographics, disease severity or treatment intensity. This study provides some of the first analytical evidence on the effects of health care strikes on patients, and suggests that hospitals functioning during nurses' strikes are doing so at a lower quality of patient care.

Focal retirement ages are a central feature of Social Security programs around the world, and provide a potentially powerful tool for policy makers who are interested in reforming retirement systems to address the growing funding shortfalls. But these tools often come hand in hand with significant changes in the financial structure of Social Security that can have independent, and potentially deleterious, impacts on retirees. In this paper, we use a major reformulation of the retirement system in Finland, featuring a relabeling of retirement ages with modest and continuous changes in financial incentives allows us to separately estimate the impact of relabeling from financial incentives in driving retirement decisions. We find that relabeling is particularly powerful as a determinant of date of retirement. Both graphical evidence and estimated hazard models reveal an enormous change in retirement when individuals face a newly defined "normal retirement" age. We also present a new approach to assessing the welfare implications of induced earlier retirement: looking at the impact on return to work. We show that the marginal workers induced to retire by relabeling are much more likely to return to work over the next three years than is the typical worker. This suggests that there is a marginal increase in regret among those who respond to this change in retirement ages.

We propose an explanation for the wide variation in rates of taxation across developed economies, based on differences in labor market institutions. In "corporatist" economies, which feature centralized labor markets, taxes on labor input will be less distortionary than when labor supply is determined individually. Since the level of labor supply is set by a small group of decision-makers, these individuals will recognize the linkage between the taxes that workers pay and the benefits that they receive. Labor tax burdens are indeed higher in more corporatist nations, and non-labor taxes are lower, which is consistent with this theory. There is also some evidence that the distortionary effects of labor taxes are lower in more corporatist economies.

We investigate the effect of health insurance coverage on the decision of individuals to apply for Disability Insurance (DI). Those who qualify for DI receive public insurance under Medicare, but only after a two-year waiting period. This raises concerns that many disabled are going uninsured while they wait for their Medicare coverage. Moreover, the combination of this waiting period and the uncertainty about application acceptance may deter those with health insurance on their jobs, but no alternative source of coverage, from leaving work to apply for DI.Data from the Health and Retirement Survey show that, in fact, uninsurance does not rise during the waiting period for DI benefits; reductions in own employer coverage are small, and are offset by increases in other sources of insurance. Correspondingly, we find that imperfect insurance coverage does deter DI application. Those who have an alternative source of insurance coverage (coverage from a spouse's employer or retiree coverage), are 26 to 74% more likely to apply for DI than those without such an alternative. Thus, limiting this waiting period would not increase the insurance coverage of the disabled in the U.S., but it would significantly increase applications to the DI program.

Past research has demonstrated that positive increments to the non-cognitive development of children can have long-run benefits. We test the symmetry of this contention by studying the effects of a sizeable negative shock to non-cognitive skills due to the introduction of universal child care in Quebec. We first confirm earlier findings showing reduced contemporaneous non-cognitive development following the program introduction in Quebec, with little impact on cognitive test scores. We then show these non-cognitive deficits persisted to school ages, and also that cohorts with increased child care access subsequently had worse health, lower life satisfaction, and higher crime rates later in life. The impacts on criminal activity are concentrated in boys. Our results reinforce previous evidence on the central role of non-cognitive skills for long-run success.

One of the most important economic decisions facing the elderly, and their families, is whether to live independently. A number of previous studies suggest that widows are fairly responsive to Social Security benefits in deciding whether to live independently. But these previous studies have either generally relied on differences in benefits across families or cohorts, which are potentially correlated with other determinants of living arrangements, or have used data from the distant past. We propose a new approach that relies on the large exogenous shifts in benefits generosity for cohorts born in the 1910-1921 period, and we study the impact of this change in living arrangements in the 1980s and 1990s. In this period, benefits rose quickly, due to double-indexing of the benefit formula, and then fell dramatically, as this double-indexing was corrected over a five-year period. Using these legislative changes in benefits that the living arrangements of widows are much more sensitive to Social Security income than implied by previous studies. We also find that the living arrangements of divorcees, the fastest growing group of elderly, are even more sensitive to benefit levels. Overall, our findings suggest that living arrangements are elastically demanded by non-married elderly, privacy is a normal good, and that reductions in Social Security benefits would significantly alter the living arrangements of the elderly. Our estimates imply that a 10% cut in Social Security benefits would lead more than 600,000 independent elderly households to move into shared living arrangements

We model automatic trigger policies for unemployment insurance by simulating a weekly panel of individual labor market histories, grouped by state. We reach three conclusions: (i) policies designed to trigger immediately at the onset of a recession result in benefit extensions that occur in less sick labor markets than the historical average for benefit extensions; (ii) the ad hoc extensions in the 2001 and 2007-09 recessions in total cover a similar number of additional weeks as common proposals for automatic triggers, but concentrate coverage more in weaker labor markets; (iii) compared to ex post policy, the cost of common proposals for automatic triggers is close to zero.

"There has been much debate among economists about whether nursing home quality is a public good across Medicaid and private-pay patients within a common facility. However, there has been only limited empirical work addressing this issue. Using a unique individual level panel of residents of nursing homes from seven states, we exploit both within-facility and within-patient variation in payer source and quality to examine this issue. We also test the robustness of these results across states with different Medicaid and private-pay rate differentials. Across our various identification strategies, the results generally support the idea that quality is a public good within nursing homes. That is, within a common nursing home, there is very little evidence to suggest that Medicaid-funded residents receive consistently lower quality care relative to their private-paying counterparts"--National Bureau of Economic Research web site.

In response to unsustainable growth in health care spending, there is enormous interest in reforming the payment system to "pay for quality instead of quantity." While quality measures are crucial to such reforms, they face major criticisms largely over the potential failure of risk adjustment to overcome endogeneity concerns. In this paper we implement a methodology for estimating the causal relationship between hospital quality measures and patient outcomes. To compare similar patients across hospitals in the same market, we xploit ambulance company preferences as an instrument for patient assignment. We find that a variety of measures used by insurers to measure provider quality are successful: assignment to a higher-scoring hospital results in better patient outcomes. We estimate that a two-standard deviation improvement in a composite quality measure based on existing data collected by CMS is causally associated with reductions in readmissions and mortality of roughly 15%.

Managing the outbreak of COVID-19 in India constitutes an unprecedented health emergency in one of the largest and most diverse nations in the world. On May 4, 2020, India started the process of releasing its population from a national lockdown during which extreme social distancing was implemented. We describe and simulate an adaptive control approach to exit this situation, while maintaining the epidemic under control. Adaptive control is a flexible counter-cyclical policy approach, whereby different areas release from lockdown in potentially different gradual ways, dependent on the local progre of the dis- ease. Because of these features, adaptive control requires the ability to decrease or increase social distancing in response to observed and projected dynamics of the disease outbreak. We show via simulation of a stochastic Susceptible-Infected-Recovered (SIR) model and of a synthetic intervention (SI) model that adaptive control performs at least as well as immediate and full release from lockdown starting May 4 and as full release from lockdown after a month (i.e., after May 31). The key insight is that adaptive response provides the option to increase or decrease socioeconomic activity depending on how it affects disease progression and this freedom allows it to do at least as well as most other policy alternatives. We also discuss the central challenge to any nuanced release policy, including adaptive control, specifically learning how specific policies translate into changes in contact rates and thus COVID-19's reproductive rate in real time.

Designing benefits for the growing platform workforce in the U.S. poses significant challenges. While platform workers need protection against unforeseen shocks, work that is often part time and spread across multiple platforms makes the traditional benefits model untenable. This paper reports the results from a survey of drivers and couriers working with Uber to help understand their benefits preferences. We find that there is a wide diversity across these workers in platform earnings, the share of platform earnings from Uber, the share of family earnings from platform work and the availability of benefits from other jobs. We use willingness-to-pay questions to show that workers are willing to trade off additional income for benefits; after accounting for the tax advantage of benefits, workers are roughly indifferent on average between the two. While there are some trends in valuation, such as higher valuation for pension than for health contributions, the most notable feature of the data is the wide variation across workers in their preferences across benefits types and relative to income. Workers also show a preference for benefits that can help them commit to increase savings in the future.

The population of the United States, as with the rest of the world, is aging rapidly, with the most rapid growth occurring among the age 85 and older population, those who rely most on long-term care. In this chapter, we review the delivery and financing of long-term care in the U.S. We show that the resources of most elderly in the U.S. are insufficient to finance these ongoing long-term care needs and the public sector finances the majority of long-term care spending. At the same time, informal care plays a critical role, with the elderly at every age and every disability level receiving informal care more frequently than formal care. Indeed, when properly valued, informal care accounts for more than one-third of the nearly 2 percent of U.S. GDP devoted to long-term care.

Despite the growing reliance on payroll taxation worldwide, there is limited evidence on the incidence of payroll taxes. I provide new evidence by examining the experience of Chile before and after the privatization of its Social Security system. This policy change led to a sharp exogenous reduction in the payroll tax burden on Chilean firms; the average payroll tax rate in my sample fell from 30% to 5% over this six year period. I use data from a census of manufacturing firms, which contains information on firm specific tax payments and average wages. I find strong evidence that the incidence of payroll taxation was fully on wages, with no effect on employment. A potential weakness with this approach is that some of the variation in firm-specific tax rates may be spurious, for example due to measurement error in wages. I attempt to surmount this problem by using a variety of different estimators, all of which yield consistent evidence of full shifting

To measure how policy changes affect social welfare, economists typically look at how policies affect behavior, and use a formal model to infer welfare consequences from the behavioral responses. But when different models can map the same behavior to very different welfare impacts, it becomes hard to draw firm conclusions about many policies. An excellent example of this conundrum is the taxation of addictive substances such as cigarettes. Existing empirical evidence on smoking is equally consistent with two models that have radically different welfare implications. Under the rational addiction model, cigarette taxes make time consistent smokers worse off. But, under alternative time inconsistent models, smokers are made better off by taxes, as they provide a valuable self-control device. We therefore propose an alternative approach to assessing the welfare implications of policy interventions: examining directly the impact on subjective well-being. We do so by matching information on cigarette excise taxation to separate surveys from the U.S. and Canada that contain data on self-reported happiness. And we model the differential impact of excise taxes on those predicted to be likely to be smokers, relative to others, in order to control for omitted correlations between happiness and excise taxation. We find consistent evidence in both countries that excise taxes make predicted smokers happier. This evidence suggests that the time inconsistent model of smoking is more appropriate, and that as a result welfare is improved by higher cigarette taxes

The Thai 30 Baht program was one of the largest health system reforms ever undertaken by a low-middle income country. In addition to lowering the cost of care for the previously uninsured in public facilities, it also entailed a fourfold increase in funding provided to hospitals to care for the poorest 30% of the population (who were already publicly insured). For the previously uninsured, we find that the 30 Baht program led to increased health care utilization, as well as a shift from private to public sources of care. But, we find a larger increase for the poor who were previously publicly insured, especially amongst infants and women of childbearing age. Using vital statistics records, we find that the increased access to healthcare by the publicly insured poor led to a reduction in their infant mortality of at least 6.5 per 1,000 births. This suggests significant improvements in infant mortality rates can be achieved through increased access to healthcare services for the poor and marginalized groups.

The strong negative correlation over time between smoking rates and obesity have led some to suggest that reduced smoking is increasing weight gain in the U.S.. This conclusion is supported by the findings of Chou et al. (2004), who conclude that higher cigarette prices lead to increased body weight. We investigate this issue and find no evidence that reduced smoking leads to weight gain. Using the cigarette tax rather than the cigarette price and controlling for non-linear time effects, we find a negative effect of cigarette taxes on body weight, implying that reduced smoking leads to lower body weights. Yet our results, as well as Chou et al., imply implausibly large effects of smoking on body weight. Thus, we cannot confirm that falling smoking leads in a major way to rising obesity rates in the U.S.

The Medicare hospice program is intended to provide palliative care to terminal patients, but patients with long stays in hospice are highly profitable, motivating concerns about overuse among the Alzheimer's and Dementia (ADRD) population in the rapidly growing for-profit sector. We provide the first causal estimates of the effect of for-profit hospice on patient spending using the entry of for-profit hospices over twenty years. We find hospice has saved money for Medicare by offsetting other expensive care among ADRD patients. As a result, policies limiting hospice use including revenue caps and anti-fraud lawsuits are distortionary and deter cost-saving admissions.

Previous research on unemployment insurance (UI) has focused on the costs of the program, in terms of the distorting effects of generous UI benefits on worker and firm behavior. For assessing the optimal size of an unemployment insurance program, however, it is also important to gauge the benefits of increased UI generosity, in terms of smoothing consumption across periods of joblessness. I do so through a reduced form approach which directly measures the effect of legislated variations in UI benefits on consumption changes among individuals becoming unemployed. I use annual observations on food consumption expenditures for 1968-1987 from the Panel Study of Income Dynamics, matched to information on the UI benefits for which unemployed persons were eligible in each state and year. I estimate that a 10 percentage point increase in the UI replacement rate leads to a consumption fall upon unemployment which is 2.7% smaller. Over this period, the average fall in consumption for the unemployed was 7%; my results imply that, in the absence of unemployment insurance, this fall would have been over three times as large. I also find that the positive effect of UI only extends for one period, smoothing consumption during initial job loss but having no permanent effect on consumption levels; that individuals who anticipate layoff see a smaller consumption smoothing effect; and that UI appears to somewhat crowd out other forms of public consumption insurance. Despite the substantial estimated consumption smoothing effect, however, my results imply that the optimal UI benefit level is within the range of current replacement rates only at fairly high levels of risk aversion