The Macroeconomics of Universal Health Insurance Vouchers Juergen Jung Towson University Chung Tran University of New South Wales Jul-Aug 2009 Jung and Tran (TU and UNSW) Health Vouchers 2009 1 / 29
Dysfunctional U.S. Health Care System Issues: 1 Low Coverage: 47 million in 2006 (15%) 2 High Cost: 16% of GDP in 2006 and close to 20% by 2015 Causes: 1 Market failure 2 Wrong government intervention Market Based Reform: Universal Health Insurance Vouchers (UHIV) 1 increase the number of insured individuals 2 control total health expenditure Jung and Tran (TU and UNSW) Health Vouchers 2009 2 / 29
What are Health Insurance Vouchers? Emanuel and Fuchs (2007) as well as Kotliko (2007) 1 Government issues medical vouchers to all individuals vouchers are calculated individually based on the amount of the expected health expenditures for next year keeps individual health records (like in Medicare) xes annual budget for vouchers as percentage of GDP 2 Individuals purchase health insurance from private insurance companies using the voucher 3 Participating insurance companies have to accept vouchers contracts must provide a 'base insurance' can oer additional insurance compete and monitor to keep premiums and prices for health care services low Jung and Tran (TU and UNSW) Health Vouchers 2009 3 / 29
Objectives of the Paper Develop an analytical framework to study the implications of a health insurance voucher program Our key contributions 1 A macro model with endogenous health production and health insurance choice 2 Quantify the short-run and long-run eects of introducing the voucher program Jung and Tran (TU and UNSW) Health Vouchers 2009 4 / 29
The Model: Key Features Standard stochastic overlapping generations model 1 Sectors: household, rm and government 2 Endowments: randon lifetime and ability to work 3 Markets: consumption, labor and capital New features 1 Health: a consumption and investment good 2 Health: xable, risky, and insurable 3 Private health insurance market Jung and Tran (TU and UNSW) Health Vouchers 2009 5 / 29
The Model: Preferences and Capital Accumulation Preferences: Health capital: 1 service ow from health capital u (c j, s j ) s j = s (h j ) 2 health production 3 health shocks h j = h (m j, h j 1, ε j ) Human capital: 1 accumulation 2 productivity shocks P j (ε j, ε j 1 ) = Pr (ε j ε j 1, j) e j = e (j, h, ɛ j ) for j = {1,..., J 1 } Π j (ɛ j, ɛ j 1 ) = Pr (ɛ j ɛ j 1, j) Jung and Tran (TU and UNSW) Health Vouchers 2009 6 / 29
The Model: Health Insurance and Expenditures Insurance plans: individual and group insurance Group insurance oers provided by employers: no rating and lower price Ω income (i GI,j, i GI,j 1 ) = Pr (i GI,j i GI,j 1, income) Health insurance choice: endogenous Health insurance states: in j = 1 : no insurance in j = 2 : individual health insurance in j = 3 : group health insurance Health expenditures depend on individuals' health insurance state Jung and Tran (TU and UNSW) Health Vouchers 2009 7 / 29
The Model: Worker's Program Agent state vector x j = {a j, h j 1, in j, ε j, ɛ j, i GI,j } Agents receive income (wage, interest income, accidental bequests, and social insurance) Pay taxes (payroll and progressive income tax) Agents simultaneously choose: 1 Consumption c j and asset holdings a j 2 Health expenditures m j 3 Insurance state for next period in j = {1, 2, 3} 4 If i GI,j = 1 then agents can either buy individual insurance in j = 2 or group insurance in j = 3 5 If i GI,j = 0 then agents can only buy individual insurance in j = 2 Jung and Tran (TU and UNSW) Health Vouchers 2009 8 / 29
The Model: Worker's Dynamic Programming Formulation V j (x j ) = { } max u (c j, h j ) + βπ j E εj+1,ɛ j+1,i GI,j+1 ε j,ɛ j,i GI,j [V (x j+1 )] {c j,m j, a j+1,in j+1} s.t. ( 1 + τ C ) c j + (1 + g) a j+1 + o W (m j ) + p = w j + R ( p j < w j + R ( a j + T Beq) Tax j + Tj SI + v j a j + T Beq j ) o W (m j ) Tax j 0 a j Jung and Tran (TU and UNSW) Health Vouchers 2009 9 / 29
The Model: Retiree's Program Agent state vector x j = {a j, h j 1, ε j } Agents receive income (pension, interest income, accidental bequests, and social insurance) Pay taxes (progressive income tax) Forced into Medicare pay p Med j Agents simultaneously choose: 1 Consumption c j and asset holdings a j 2 Health expenditures m j Jung and Tran (TU and UNSW) Health Vouchers 2009 10 / 29
The Model: Retiree's Dynamic Programming Formulation V j (x j ) = = R { } max u (c j, h j ) + βπ j E εj+1 ε j [V j+1 (x j+1 )] {c j,m j, a j+1} s.t. c j + a j + o R (m j ) + p Med j ( a j 1 + T Beq) + R m a m j 1 + Tj Soc + Tj SI Tax j 0 a j Jung and Tran (TU and UNSW) Health Vouchers 2009 11 / 29
The Model: Firms and Insurance Companies Firms: Insurance Companies: (1 + ω) J 1 +1 = (1 + r) max {F (K, L) qk wl}, given (q, w) {K,L} J 1 j=1 j=2 µ j µ (1 + ω) J 1 +1 = (1 + r) J 1 j=1 j [ 1 {inj(xj)=2} (1 ρ) max (0, p m,insm j (x j ) γ)] dλ (x j ) ( 1 {inj(xj)=2} p (j, h) ) dλ (x j ) j=2 µ j µ j [ 1 {inj(xj)=3} (1 ρ) max (0, p m,insm j (x j ) γ)] dλ (x j ) ( 1 {inj(xj)=3} p ) dλ (x j ) Jung and Tran (TU and UNSW) Health Vouchers 2009 12 / 29
The Model: Government I Bequests: J µ j j=1 T Beq j (x) dλ j (x) = J j=1 µ j a j (x) dλ j (x) Social Security: J = J 1 j=j 1+1 µ j j=1 µ j T Soc j (x) dλ j (x) τ Soc ( we (j, h j, ɛ) 1 {inj+1 =3} p ) dλ j (x) Jung and Tran (TU and UNSW) Health Vouchers 2009 13 / 29
The Model: Government II Medicare: J = J 1 j=j 1+1 µ j j=1 µ j + J j=j 1+1 µ j (1 ρ Med ) max ( 0, m j (x) γ Med) dλ j (x) τ Med ( we (j, h j, ɛ) 1 {inj+1 =3} p ) dλ j (x) pj Med dλ j (x) Government budget is balanced: G + = J J µ j j=1 j=1 µ j T SI j (x j ) dλ (x j ) + J Tax j (x j ) dλ (x j ) + J j=1 µ j j=1 µ j v (h j (x j )) dλ (x j ) τ C c (x j ) dλ (x j ). Jung and Tran (TU and UNSW) Health Vouchers 2009 14 / 29
Calibration Preferences: Health services: u (c j, h j ) = ( c η j s1 η j 1 σ ) 1 σ s j = h j Health Production: h j = φm ξ j + (1 δ (h j )) h j 1 + ε j Markov switching probabilities between income shocks and group insurance oer states are estimated from MEPS 2004-2005 data. Human Capital: e j = e (ɛ j ) χ (h θ j 1) 1 χ for j = {1,..., J 1 }, where e (ɛ j ) are estimated eciency proles from MEPS 2004-2005 for 3 separate income quantiles β 0, β 2 < 0, β 1 > 0, χ (0, 1) and θ = 0 in benchmark version. Jung and Tran (TU and UNSW) Health Vouchers 2009 15 / 29
Calibration Baseline Parameters Demographics: Health Production: Insurance: J 1 = 9 φ = 1 γ = $26% of spending J 2 = 5 ξ = 0.35 ρ = 33% n = 1.2% δ h = [3%,..., 90%] γ Med = $90% of private deductible Preferences: Health Productivity: ρ Med = 0.25 σ = 2.5 θ =? β = 0.99 Technology: α = 0.33 δ = 10% g = 1.5% Exogenous premium growth depending on age and health Jung and Tran (TU and UNSW) Health Vouchers 2009 16 / 29
Steady States vs. Data (NO Human Capital Eect) 80 60 Insurance Coverage in % Model Voucher Data 50 Medical Spending in % of Income 40 20 20 40 60 80 Age % Insured Spending Below Deductible 80 60 40 20 40 60 80 Age Average Consumption 0.7 0.6 0.5 0.4 20 40 60 80 Age 0 20 40 60 80 Age Average Savings 1.5 1 0.5 Model Voucher 0 20 40 60 80 Age Average Health 3 2 1 0 20 40 60 80 Age Jung and Tran (TU and UNSW) Health Vouchers 2009 17 / 29
Experiment 1: NO Human Capital Eect No Human Capital Eect Benchmark 1 Vouchers Output: Y.000 101.578 Capital: K.000 104.445 Human capital: H.000.000 Med. spending: pm M/Y 12.9% 12.6% Vouchers in % of GDP 0.0% 3.5% Interest rate: R 6.0% 5.4% Wages: w.000 101.578 Consumption tax: τ C 0.050 0.085 Soc. sec. tax: τ SS 0.109 0.103 Medicare tax: τ Med 0.039 0.000 Income tax in % of GDP: 0.179 0.194 K/Y 2.656 2.731 C/Y 0.408 0.453 Jung and Tran (TU and UNSW) Health Vouchers 2009 18 / 29
Key Mechanism: Savings Eect Income eect Young generation Old generation no premium payments: no payroll tax: Substitution eect Price of c Price of m increase in τ c : % coverage: Jung and Tran (TU and UNSW) Health Vouchers 2009 19 / 29
Key Mechanism: Savings Eect Replacing Medicare by Vouchers results in income and substitution eects. Removing insurance premimum increases income (income eect) payroll tax increases income (income eect) while consumption tax increases price of consumption (substitution eect). savings and physical capital K aects wage and interest rates increases household income (G.E. income eect) These increase the demand for health care services Net result: total health care expenditure increases, but as fraction of GDP health expenditure decreases Jung and Tran (TU and UNSW) Health Vouchers 2009 20 / 29
Steady States vs. Data (WITH Human Capital Eect) Insurance Coverage in % Medical Spending in % of Income 80 60 Model 40 Voucher Data 20 20 40 60 80 Age % Insured Spending Below Deductible 80 60 40 20 40 60 80 Age Average Consumption 0.7 0.6 0.5 0.4 20 40 60 80 Age 50 0 20 40 60 80 Age Average Savings 1.5 1 0.5 Model Voucher 0 20 40 60 80 Age Average Health 3 2 1 0 20 40 60 80 Age Jung and Tran (TU and UNSW) Health Vouchers 2009 21 / 29
Experiment 2: WITH Human Capital Eect No Human Capital Eect Human Capital Eect Benchmark 1 Vouchers Benchmark 2 Vouchers Output: Y.000 101.578.000 102.610 Capital: K.000 104.445.000 106.656 Human capital: H.000.000.000.401 Med. spending: pm M/Y 12.9% 12.6% 14.9% 14.7% Vouchers in % of GDP 0.0% 3.5% 0.0% 4.2% Interest rate: R 6.0% 5.4% 5.5% 5.0% Wages: w.000 101.578.000 102.199 Consumption tax: τ C 0.050 0.085 0.065 0.103 Soc. sec. tax: τ SS 0.109 0.103 0.109 0.104 Medicare tax: τ Med 0.039 0.000 0.045 0.000 Income tax in % of GDP: 0.179 0.194 0.175 0.195 K/Y 2.656 2.731 2.783 2.893 C/Y 0.408 0.453 0.377 0.431 Jung and Tran (TU and UNSW) Health Vouchers 2009 22 / 29
Key Mechanism: Human Capital Eects Savings eect Human capital eect Vouchers induce households to spend more on health (moral hazard). health and therefore human capital depending on whether health is productive increaes wage and interest rates, household income and again the demand for health care services Result: the demand for health care, but as fraction of GDP health expenditure decreases Jung and Tran (TU and UNSW) Health Vouchers 2009 23 / 29
Transitions: NO Human Capital Eect 102 Output 110 Capital 101 105 Consumption 120 110 Human Capital 101 99 Wages 102 % Medical Expenditure 98 96 Interest 6 5.5 5 Consumption Tax 10 101 % 8 6 Jung and Tran (TU and UNSW) Health Vouchers 2009 24 / 29
Welfare Analysis: NO Human Capital Eect 5 Losers Compensating Consumption per Lifetime Consumption (in %) % 0 5 10 15 Winners Old Regime Agents New Regime Agents 20 25 15 10 5 0 5 10 15 Generation 5 Compensating Consumption per GDP (in %) Losses % 0 5 Gains 10 0 5 10 15 20 25 30 Jung and Tran (TU and UNSW) Health Vouchers 2009 25 / 29
Transitions WITH human capital eect 104 Output 110 Capital 102 105 Consumption 120 110 Human Capital.5 Medical Expenditure 98 96 Interest 5.5 % 5 Wages 104 102 % 4.5 Consumption Tax 15 10 5 Jung and Tran (TU and UNSW) Health Vouchers 2009 26 / 29
Welfare Analysis: WITH Human Capital Eect % 0 10 20 30 Losers Winners Compensating Consumption per Lifetime Consumption (in %) Old Regime Agents New Regime Agents 15 10 5 0 5 10 15 Generation 5 0 5 Compensating Consumption per GDP (in %) Losses Gains % 10 15 20 0 5 10 15 20 25 30 Jung and Tran (TU and UNSW) Health Vouchers 2009 27 / 29
Conclusion Health vouchers seem promising in being able to sustainably nance health care expenditures while providing full health insurance coverage to the entire U.S. population The decrease in health care expendiure as fraction of GDP is primarily due to a general equilibrium savings eect The human capital eect is potentially important Welfare gain Jung and Tran (TU and UNSW) Health Vouchers 2009 28 / 29
Extensions Empirical structurally estimate health production parameters φ, ξ, δ(h) and health shock process Modelling the supply of health care services m and prices p m insurance rm competition and its eect on price of health care services and insurance premiums Issues privatization of public health insurance programs nancing health costs in an aging economy Jung and Tran (TU and UNSW) Health Vouchers 2009 29 / 29