Table 2:
Results: Demand and Marginal Cost
| Beneficiaries not eligible for low-income subsidies |
||||||
|---|---|---|---|---|---|---|
| Risk type 1 | Risk type 2 | Risk type 3 | Risk type 4 | Risk type 5 | LIS | |
| (1) | (2) | (3) | (4) | (5) | (6) | |
| Panel A: Parameters of the utility function | ||||||
| Premium, $000 | ||||||
| Mean | 3.16 | 3.25 | 2.46 | 2.56 | 2.56 | −7.93 |
| (0.35) | (0.61) | (0.33) | (0.38) | (0.39) | (0.56) | |
| Standard deviation | 0.24 | 0.50 | 0.11 | 0.33 | −0.07 | - |
| (0.50) | (0.39) | (0.62) | (0.67) | (0.72) | - | |
| Standard deviation, inner option | 0.67 | 6.92 | −0.20 | 0.37 | 0.83 | - |
| (0.55) | (0.73) | (0.35) | (0.48) | (0.33) | - | |
| Annual deductible, $000 | −10.61 | −11.36 | −8.00 | −7.97 | −7.92 | - |
| (2.29) | (1.61) | (1.27) | (1.56) | (1.48) | - | |
| Indicator for having any coverage in the gap | ||||||
| Mean | 3.83 | 3.75 | 1.82 | 3.38 | 4.62 | - |
| (2.08) | (1.32) | (1.12) | (0.86) | (1.14) | - | |
| Standard deviation | 2.05 | −0.33 | 1.97 | −0.19 | 0.87 | - |
| (0.85) | (0.98) | (0.53) | (0.65) | (0.72) | - | |
| Number of most common drugs covered | 47.87 | 34.03 | 40.85 | 43.92 | 60.00 | 1.05 |
| (17.57) | (11.11) | (14.11) | (10.85) | (8.68) | (7.39) | |
| Measure of pharmacy network breadth | 0.36 | 0.35 | 0.24 | 0.31 | 0.28 | 0.10 |
| (0.02) | (0.03) | (0.01) | (0.10) | (0.04) | (0.07) | |
| Number of years the plan is on the market | 1.09 | 1.24 | 1.07 | 1.05 | 0.92 | 0.48 |
| (0.19) | (0.12) | (0.10) | (0.13) | (0.12) | (0.04) | |
| F-statistic first stage across all risk types | 245 | 23 | ||||
|
Panel B: Marginal cost projection for non-LIS market Annual deductible, $000 |
−0.19 | −0.23 | −0.29 | −0.42 | −0.52 | |
| (0.04) | (0.05) | (0.06) | (0.09) | (0.11) | ||
| Number of most common drugs covered | −0.78 | −0.30 | −0.11 | −2.11 | −2.83 | |
| (2.21) | (2.69) | (3.42) | (4.95) | (6.17) | ||
| Indicator for having any coverage in the gap | 0.30 | 0.36 | 0.46 | 0.68 | 0.83 | |
| (0.01) | (0.01) | (0.01) | (0.02) | (0.03) | ||
| Measure of pharmacy network breadth | −0.098 | −0.14 | −0.18 | −0.28 | −0.34 | |
| (0.04) | (0.05) | (0.07) | (0.10) | (0.12) | ||
| Number of years the plan is on the market | 0.03 | 0.04 | 0.04 | 0.06 | 0.08 | |
| (0.00) | (0.00) | (0.00) | (0.01) | (0.01) | ||
| Mean dependent variable (inverted MC, $000) | 0.67 | 0.80 | 1.00 | 1.46 | 1.81 | |
| Standard deviation dependent variable | 0.16 | 0.20 | 0.25 | 0.36 | 0.44 | |
| R-squared | 0.77 | 0.76 | 0.76 | 0.76 | 0.76 | |
| N | 756 | 756 | 756 | 756 | 756 | |
The table reports parameter estimates for demand models as described in Section 3.1 (Panel A). The pa-rameters are reported separately for risk types 1 to 5 of regular consumers (columns 1–5) and consumerseligible for low income subsidies (column 6). We report standard errors in parentheses. All models include,but do not report: a constant, fixed effects for parent organizations, year fixed effects, geographic marketfixed effects,a dummy for an enhanced plan, the number of drugs on the plan’s formulary, and the number ofdrugs placed in Tiers 1–2 of the formulary. F-statistics are reported for the first stage regressions of regular (column 1) and LIS (column 6) premiums on price instruments as described in Section 4. Panel B reportsthe results of a hedonic regression of marginal costs - for plans not distorted by LIS random assignmentincentives - estimated via the inversion of the first order conditions on plan characteristics in 2010. Theregression includes the same plan characteristics as the demand model for regular consumers.