Households’ uncertainty about Medicare policy

The AP- GfK Poll was conducted Nov. 18-22, 2010, by GfK Roper Public Affairs and Corporate Communications. It involved landline and cell phone interviews with 1000 adults nationwide, and has a margin of sampling error of plus or minus 4.3% points.

For instance, http://budget.house.gov/news/documentsingle.aspx?DocumentID=292036.

Those policy mixes are consistent with some of the options considered in Attanasio, Kitao, and Violante (2010). In addition, we assume that households expect a lower cut in benefits relative to the increase in tax rates following a survey by Pew Research Center in 2011. That survey finds that while Americans see a need for changes to the entitlement programs, they decisively support keeping Social Security and Medicare benefits as they are now.

The Alternative Fiscal Scenario incorporates several changes to current law that are widely expected to occur or that would modify some provisions of law that might be difficult to sustain for a long period.

We assume that young workers are fully covered by insurance, as in Scholz, Seshadri, and Khitatrakun (2006).

Households are allowed to borrow up to the value of a^min, which represents the present value of the worst possible future labor income stream when the household supplies the maximum feasible working hours.

We use a 2-point for the health status, good and bad; a 126-point non-uniform grid for the asset space, with more points around zero asset holdings; an 8-point grid for the average indexed yearly earnings following Nishiyama and Smetters (2007); a 3-point grid for policy uncertainty consistently with Table 6.

Source: Staff estimates from the Congressional Budget Office.

Storesletten, Telmer, and Yaron (1999) estimate a process for exogenous labor income, modeling both a transitory and a persistent component. Other studies, such as French (2005) and Nishiyama and Smetters (2007), estimate a wage process accounting for an endogenous labor supply, but they do not model a transitory component. Even though in our paper labor supply decision is endogenous, we choose to use the estimates of the wage process obtained by Storesletten, Telmer, and Yaron (1999) to account also for transitory innovations. The reason behind our choice is two-fold. First, Carroll (1997) highlights the importance of transitory shocks in explaining the observed household propensity to consume and save. Second, a well-behaved profile for consumption and savings is particularly relevant for our work since the saving motive is one of the main drivers of our results.

See, for instance, Cocco, Gomes, and Maenhout (2005).

The space of average yearly historical earnings is divided into two subsets (high and low), each consisting of 4 grid points. We assign a household to the high labor income class if its average yearly historical earnings belongs to higher grid points or to the low labor income class otherwise.

Health care cost is projected to grow faster than nominal GDP over time (CBO, Long-Term Budget Outlook 2012), affecting the price of medical services relative to consumption. However, the growth in the price of medical services in excess of GDP growth is expected to be larger than 2% yearly only after 2030. Hence, the impact of that cost growth on our results in the short run appears to be moderate as policy uncertainty in the model is resolved by 2026.

In our model, the Medicare system does not cover medical expenses for individuals younger than 65 since we do not model disability or severe terminal diseases. This assumption does not bias significantly our results because, as we mentioned earlier, those categories are a small percentage of the covered population.

Medicare Part C represents an alternative way of obtaining the benefits as defined in Part A and B and enrollees pay the same premium as Part B. For this reason, we do not consider this as a separate structure.

In our baseline, we do not consider any supplemental health insurance. Since households have a large set of options for supplemental insurance, such as Medicare Advantage, employer-sponsored, Medigap, and other public/private coverages, there is a large heterogeneity of coverage rates and premiums. Allowing households to endogenously choose also supplemental health insurance will become computationally cumbersome. We perform a simulation with a version of the model that accounts for a simplified scheme for supplemental health insurance and the results are available from the authors upon request.

Effective marginal tax rates computed by CBO, Tax Analysis Division Staff.

Effective marginal tax rates provided by CBO, Tax Analysis Division Staff.

We use 2010 tax rates, since in 2011 there was a temporary cut in payroll tax rates.

We acknowledge that grandfathering could be modeled in different ways, but there is a large uncertainty about the possible alternatives that the government may decide to put in place. For that reason, we chose a computationally viable alternative.

To reflect expectations that mandatory outlays are more difficult to modify, we carry out an experiment with a lower probability associated to a cut in Medicare benefits. The results are available from the authors upon request.

The average tax rate is 7.6% in the baseline. For the single rates see Figure 6.

Those policy mixes are consistent with some of the options considered in Attanasio, Kitao, and Violante (2010).

The experiments in the two environments with certainty and uncertainty over Medicare reforms assume exogenous changes in taxes or benefits. In a general equilibrium framework, though, households will work more and save more in the presence of policy uncertainty and presumably taxes or benefits will have to change by less than in an environment which features policy certainty, to achieve the same effect on the government budget. Moreover, even keeping tax or benefit changes equal in the two environments, higher savings and labor supply in response to policy uncertainty would yield higher tax revenues, a lower interest rate, and, ultimately, a lower debt-to-GDP ratio. A lower debt-to-GDP ratio in turn would weaken the hazard of reform. In sum, the general equilibrium effects could moderate the impact of policy uncertainty on households’ welfare.

Since the effects of policy uncertainty become weaker as time approaches the resolution date, we choose to focus on the first two years of our simulation results.