The Benefits to Taxpayers from Increases in Students - RAND ...
The Benefits to Taxpayers from Increases in Students - RAND ...
The Benefits to Taxpayers from Increases in Students - RAND ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
104 <strong>The</strong> <strong>Benefits</strong> <strong>to</strong> <strong>Taxpayers</strong> <strong>from</strong> <strong>Increases</strong> <strong>in</strong> <strong>Students</strong>’ Educational Atta<strong>in</strong>ment<br />
Medicaid<br />
<strong>The</strong> SIPP does not report Medicaid payments on behalf of a respondent. For each<br />
respondent, it reports whether the respondent was covered by Medicaid, the number<br />
of medical provider visits <strong>in</strong> the previous 12 months, and whether the respondent was<br />
hospitalized <strong>in</strong> the previous 12 months. We use a two-part model <strong>to</strong> estimate benefits<br />
<strong>from</strong> the Medicaid program. <strong>The</strong> first part estimates Medicaid utilization, the probability<br />
that a respondent covered by Medicaid had at least one medical provider visit<br />
<strong>in</strong> the previous 12 months. <strong>The</strong> second part estimates hospitalization, conditional on<br />
Medicaid utilization. We assume that Medicaid utilization absent hospitalization is<br />
outpatient utilization. We assume that hospitalization is <strong>in</strong>patient utilization.<br />
As the SIPP provides utilization but not benefit data, we compute per-person<br />
average <strong>in</strong>patient as well as outpatient benefits <strong>from</strong> Centers for Medicare and Medicaid<br />
Services data.<br />
As our analysis of the age profile of Medicaid participants and nonparticipants<br />
did not <strong>in</strong>dicate anyth<strong>in</strong>g remotely resembl<strong>in</strong>g a transition po<strong>in</strong>t based on age, we did<br />
not separately model Medicaid utilization for the elderly and non-elderly. Estimates are<br />
presented <strong>in</strong> Table C.5.<br />
Medicare<br />
<strong>The</strong> SIPP does not report Medicare payments on behalf of a respondent. For each<br />
respondent, it reports whether the respondent was covered by Medicare, the number<br />
of medical provider visits <strong>in</strong> the previous 12 months, and whether the respondent was<br />
hospitalized <strong>in</strong> the previous 12 months. We use a two-part model <strong>to</strong> estimate benefits<br />
<strong>from</strong> the Medicare program. <strong>The</strong> first part estimates Medicare utilization, the probability<br />
that a respondent covered by Medicare had at least one medical provider visit<br />
<strong>in</strong> the previous 12 months. <strong>The</strong> second part estimates hospitalization, conditional on<br />
Medicare utilization. We assume that Medicare utilization absent hospitalization is<br />
outpatient utilization. We assume that hospitalization is <strong>in</strong>patient utilization.<br />
As the SIPP provides utilization but not benefit data, we compute per-person<br />
average <strong>in</strong>patient as well as outpatient benefits <strong>from</strong> CMS data.<br />
As Medicare is primarily a health care <strong>in</strong>surance program for the elderly but also<br />
a disability and renal-disease program regardless of age, we modeled Medicare utilization<br />
separately for the elderly and non-elderly. An analysis of the age distribution of<br />
participants and nonparticipants also supports this choice for mere utilization (not<br />
hospitalization). Estimates are presented <strong>in</strong> Table C.6.