The_AT Impact_NN1 of_IO Health_NN1 Insurance_NN1 Reform_NN1 on_II State_NN1 and_CC Local_NN1 Governments_NN2 Health_NN1 insurance_NN1 reform_NN1 as_CSA currently_RR proposed_VVN in_II both_DB2 the_AT House_NN1 Tri-Committee_NP1 and_CC in_II the_AT Senate_NN1 HELP_NN1 Committee_NN1 would_VM ensure_VVI that_CST virtually_RR all_DB Americans_NN2 receive_VV0 health_NN1 insurance_NN1 ._. 
As_II part_NN1 of_IO the_AT proposed_JJ increases_NN2 in_II health_NN1 insurance_NN1 coverage_NN1 ,_, the_AT House_NN1 Tri-Committee_NP1 legislation_NN1 calls_VVZ for_IF an_AT1 expansion_NN1 in_II Medicaid_NP1 to_II all_DB individuals_NN2 under_RG 133_MC percent_NNU of_IO the_AT federal_JJ poverty_NN1 line_NN1 (_( FPL_NP1 )_) ._. 
Understandably_RR ,_, there_EX has_VHZ been_VBN some_DD concern_NN1 in_II state_NN1 capitals_NN2 surrounding_VVG this_DD1 proposal_NN1 given_VVN the_AT possible_JJ increase_NN1 in_II state_NN1 Medicaid_VVD expenditures_NN2 that_CST could_VM result_VVI ._. 
However_RR ,_, state_NN1 and_CC local_JJ governments_NN2 are_VBR already_RR spending_VVG billions_NNO2 of_IO dollars_NNU2 each_DD1 year_NNT1 providing_VVG coverage_NN1 to_II the_AT uninsured_JJ --_NN1 costs_VVZ that_CST could_VM be_VBI significantly_RR reduced_VVN as_II a_AT1 result_NN1 of_IO the_AT currently_RR proposed_JJ reforms_NN2 ._. 
Additionally_RR ,_, state_NN1 and_CC local_JJ governments_NN2 employ_VV0 more_DAR than_CSN 19_MC million_NNO individuals_NN2 ,_, and_CC their_APPGE total_JJ spending_NN1 on_II health_NN1 insurance_NN1 premiums_NN2 for_IF this_DD1 group_NN1 in_II 2007_MC was_VBDZ approximately_RR $95_NNU billion_NNO ._. 
This_DD1 group_NN1 currently_RR pays_VVZ a_AT1 "_" hidden_JJ tax_NN1 "_" in_II the_AT form_NN1 of_IO higher_JJR health_NN1 insurance_NN1 premiums_NN2 that_CST helps_VVZ to_TO cover_VVI expenses_NN2 incurred_VVN by_II the_AT uninsured_JJ ._. 
This_DD1 burden_NN1 would_VM be_VBI greatly_RR reduced_VVN as_II a_AT1 result_NN1 of_IO expansions_NN2 in_II insurance_NN1 coverage_NN1 resulting_VVG from_II health_NN1 insurance_NN1 reform_NN1 ,_, which_DDQ would_VM generate_VVI significant_JJ savings_NN2 for_IF state_NN1 and_CC local_JJ governments_NN2 ._. 
A_ZZ1 June_NP1 report_NN1 by_II the_AT Council_NN1 of_IO Economic_JJ Advisers_NN2 (_( CEA_NP1 )_) demonstrated_VVD the_AT health_NN1 and_CC overall_RR economic_JJ benefits_NN2 of_IO health_NN1 insurance_NN1 reform_NN1 (_( CEA_NP1 ,_, 2009a_FO )_) ._. 
A_AT1 subsequent_JJ study_NN1 produced_VVN by_II the_AT CEA_NN1 in_II July_NPM1 showed_VVD the_AT significant_JJ benefits_NN2 to_II small_JJ businesses_NN2 and_CC their_APPGE employees_NN2 from_II health_NN1 insurance_NN1 reform_NN1 ._. 
This_DD1 report_NN1 ,_, the_AT third_MD in_II the_AT CEA_NN1 series_NN ,_, illustrates_VVZ the_AT potential_JJ benefits_NN2 of_IO health_NN1 insurance_NN1 reform_NN1 for_IF state_NN1 and_CC local_JJ government_NN1 budgets_NN2 through_II a_AT1 detailed_JJ analysis_NN1 of_IO current_JJ spending_NN1 levels_NN2 ._. 
Focusing_VVG on_II a_AT1 sample_NN1 of_IO sixteen_MC diverse_JJ states_NN2 ,_, we_PPIS2 provide_VV0 detailed_JJ case_NN1 studies_NN2 of_IO the_AT multitude_NN1 of_IO ways_NN2 that_CST state_VV0 and_CC local_JJ governments_NN2 spend_VV0 billions_NNO2 of_IO dollars_NNU2 on_II uncompensated_JJ care_NN1 ._. 
These_DD2 estimates_NN2 ,_, combined_VVN with_IW estimates_NN2 of_IO possible_JJ state_NN1 expenditures_NN2 associated_VVN with_IW reform_NN1 ,_, indicate_VV0 that_CST the_AT move_NN1 to_II greater_JJR insurance_NN1 coverage_NN1 would_VM likely_RR result_VVI in_II substantial_JJ savings_NN2 for_IF state_NN1 and_CC local_JJ governments_NN2 ._. 
Rather_II21 than_II22 harming_VVG the_AT budget_NN1 situation_NN1 of_IO the_AT states_NN2 ,_, health_NN1 insurance_NN1 reform_NN1 would_VM improve_VVI it_PPH1 ._. 
A._NP1 Scope_NN1 and_CC Methodology_NN1 of_IO the_AT Study_NN1 Determining_VVG what_DDQ states_NN2 spend_VV0 on_II uncompensated_JJ care_NN1 is_VBZ difficult_JJ ._. 
This_DD1 information_NN1 is_VBZ not_XX collected_VVN in_II one_MC1 place_NN1 or_CC in_II a_AT1 consistent_JJ form_NN1 across_II states_NN2 ._. 
To_TO gather_VVI this_DD1 information_NN1 ,_, we_PPIS2 examined_VVD publicly_RR available_JJ information_NN1 from_II each_DD1 state_NN1 government_NN1 and_CC in_II many_DA2 cases_NN2 from_II county_NN1 and_CC city_NN1 governments_NN2 ._. 
We_PPIS2 supplemented_VVD this_DD1 with_IW information_NN1 from_II federal_JJ agencies_NN2 ,_, non-profit_JJ research_NN1 organizations_NN2 ,_, and_CC other_JJ sources_NN2 ,_, all_DB of_IO which_DDQ we_PPIS2 list_VV0 in_II the_AT references_NN2 that_CST are_VBR included_VVN at_II the_AT end_NN1 of_IO each_DD1 state_NN1 summary_NN1 ._. 
Because_II21 of_II22 the_AT inherent_JJ difficulty_NN1 in_II locating_VVG comprehensive_JJ information_NN1 on_II all_DB government_NN1 spending_VVG on_II the_AT uninsured_JJ ,_, the_AT state_NN1 and_CC local_JJ government_NN1 programs_NN2 that_CST we_PPIS2 highlight_VV0 are_VBR in_II no_AT way_NN1 meant_VVD to_TO be_VBI an_AT1 exhaustive_JJ list_NN1 ._. 
Our_APPGE estimates_NN2 should_VM be_VBI considered_VVN a_AT1 plausible_JJ lower_RRR bound_VVN on_II the_AT potential_JJ cost_NN1 savings_NN2 to_TO state_VVI and_CC local_JJ governments_NN2 ._. 
It_PPH1 is_VBZ precisely_RR because_II21 of_II22 the_AT difficulty_NN1 involved_JJ in_II gathering_VVG the_AT information_NN1 that_CST we_PPIS2 begin_VV0 with_IW a_AT1 sample_NN1 of_IO states_NN2 ._. 
The_AT sixteen_MC states_VVZ that_CST we_PPIS2 examine_VV0 are_VBR Arkansas_NP1 ,_, California_NP1 ,_, Florida_NP1 ,_, Idaho_NP1 ,_, Indiana_NP1 ,_, Iowa_NP1 ,_, Maine_NP1 ,_, Michigan_NP1 ,_, Minnesota_NP1 ,_, Montana_NP1 ,_, Nebraska_NP1 ,_, North_ND1 Carolina_NP1 ,_, Oregon_NP1 ,_, Pennsylvania_NP1 ,_, Vermont_NP1 ,_, and_CC Wyoming_NP1 ._. 
While_CS not_XX a_AT1 random_JJ sample_NN1 ,_, this_DD1 group_NN1 covers_VVZ a_AT1 range_NN1 of_IO geographic_JJ ,_, economic_JJ ,_, and_CC demographic_JJ features_NN2 ._. 
These_DD2 states_NN2 also_RR run_VV0 the_AT gamut_NN1 from_II low_JJ to_II high_JJ uncompensated_JJ care_NN1 expenditures_NN2 ._. 
For_IF this_DD1 reason_NN1 we_PPIS2 feel_VV0 they_PPHS2 are_VBR largely_RR representative_JJ of_IO the_AT experience_NN1 of_IO the_AT states_NN2 we_PPIS2 have_VH0 not_XX yet_RR analyzed_VVN ._. 
In_II31 addition_II32 to_II33 gathering_VVG uncompensated_JJ care_NN1 expenditure_NN1 data_NN from_II a_AT1 multitude_NN1 of_IO sources_NN2 ,_, we_PPIS2 also_RR provide_VV0 estimates_NN2 of_IO how_RGQ much_DA1 states_NN2 pay_VV0 in_II higher_JJR health_NN1 care_NN1 premiums_NN2 for_IF state_NN1 employees_NN2 because_II21 of_II22 uncompensated_JJ care_NN1 ._. 
Though_CS not_XX as_RG large_JJ as_CSA some_DD of_IO the_AT direct_JJ expenditures_NN2 ,_, this_DD1 hidden_JJ tax_NN1 is_VBZ substantial_JJ ,_, especially_RR for_IF larger_JJR states_NN2 ._. 
The_AT technical_JJ appendix_NN1 provides_VVZ details_NN2 on_II the_AT methodology_NN1 that_CST we_PPIS2 use_VV0 to_TO do_VDI this_DD1 calculation_NN1 ._. 
To_TO estimate_VVI the_AT possible_JJ cost_NN1 to_TO state_VVI governments_NN2 of_IO health_NN1 insurance_NN1 reform_NN1 ,_, we_PPIS2 use_VV0 detailed_JJ statistics_NN for_IF each_DD1 state_NN1 from_II the_AT March_NPM1 2008_MC Current_JJ Population_NN1 Survey_NN1 to_TO estimate_VVI the_AT number_NN1 of_IO uninsured_JJ citizens_NN2 at_II various_JJ income_NN1 levels_NN2 ._. 
We_PPIS2 combine_VV0 these_DD2 estimates_NN2 with_IW information_NN1 on_II Medicaid_NP1 expenditures_NN2 by_II state_NN1 and_CC details_NN2 from_II the_AT proposed_JJ legislation_NN1 on_II the_AT share_NN1 to_TO be_VBI paid_VVN for_IF by_II the_AT states_NN2 ._. 
Details_NN2 of_IO how_RRQ we_PPIS2 conduct_VV0 this_DD1 analysis_NN1 are_VBR also_RR included_VVN in_II the_AT Appendix_NN1 ._. 
B._NP1 State_NN1 Spending_VVG on_II Uncompensated_JJ Care_NN1 Our_APPGE analysis_NN1 reveals_VVZ that_CST states_NN2 spend_VV0 billions_NNO2 on_II uncompensated_JJ care_NN1 in_II a_AT1 wide_JJ variety_NN1 of_IO ways_NN2 ._. 
Most_RGT obviously_RR ,_, there_EX are_VBR state_NN1 programs_NN2 to_TO cover_VVI low-income_JJ uninsured_JJ patients_NN2 ._. 
Consider_VV0 the_AT following_JJ three_MC examples_NN2 ._. 
In_II California_NP1 ,_, counties_NN2 are_VBR the_AT "_" providers_NN2 of_IO last_MD resort_NN1 "_" for_IF health_NN1 services_NN2 to_II low-income_JJ uninsured_JJ people_NN with_IW no_AT other_JJ sources_NN2 of_IO care_NN1 ._. 
In_II 2004-2005_MCMC ,_, 24_MC California_NP1 counties_NN2 spent_VVD $1.61_NNU billion_NNO providing_CS care_NN1 to_II the_AT uninsured_JJ through_II their_APPGE Medically_RR Indigent_JJ Services_NN2 Programs_NN2 ._. 
The_AT remaining_JJ 34_MC (_( primarily_RR rural_JJ )_) counties_NN2 spent_VVD $283_NNU million_NNO on_II care_NN1 to_II the_AT uninsured_JJ through_II their_APPGE County_NN1 Medical_JJ Services_NN2 Programs_NN2 during_II the_AT 2008_MC fiscal_JJ year_NNT1 ._. 
Between_II both_DB2 programs_NN2 ,_, California_NP1 spent_VVD $1.90_NNU billion_NNO ._. 
In_II Minnesota_NP1 ,_, the_AT state-funded_JJ General_JJ Assistance_NN1 Medical_JJ Care_NN1 program_NN1 provides_VVZ full_JJ health_NN1 coverage_NN1 to_II uninsured_JJ adults_NN2 up_RG21 to_RG22 75_MC percent_NNU of_IO the_AT FPL_NP1 who_PNQS are_VBR not_XX eligible_JJ for_IF federal_JJ benefits_NN2 ._. 
In_II FY_NP1 2007_MC ,_, the_AT state_NN1 spent_VVD $281_NNU million_NNO in_II payments_NN2 to_II providers_NN2 for_IF GAMC_NN1 services_NN2 ._. 
In_II Miami-Dade_NP1 County_NN1 ,_, Florida_NP1 ,_, funding_VVG for_IF uncompensated_JJ care_NN1 through_II its_APPGE public_JJ health_NN1 facilities_NN2 comes_VVZ from_II a_AT1 0.5_MC percent_NNU sales_NN tax_NN1 ._. 
In_II FY_NP1 2007-2008_MCMC ,_, revenue_NN1 from_II this_DD1 tax_NN1 amounted_VVD to_II $187_NNU million_NNO ._. 
Under_II current_JJ draft_NN1 legislation_NN1 ,_, low-income_JJ uninsured_JJ citizens_NN2 and_CC legal_JJ residents_NN2 would_VM be_VBI covered_VVN by_II Medicaid_NP1 ,_, which_DDQ would_VM be_VBI primarily_RR federally-funded_JJ ,_, greatly_RR reducing_VVG the_AT need_NN1 for_IF such_DA expenditures_NN2 by_II state_NN1 and_CC local_JJ governments_NN2 ._. 
Many_DA2 states_NN2 fund_VV0 programs_NN2 which_DDQ cover_VV0 residents_NN2 who_PNQS earn_VV0 above_II 133_MC percent_NNU of_IO the_AT federal_JJ poverty_NN1 level_NN1 ._. 
Consider_VV0 the_AT following_JJ three_MC examples_NN2 ._. 
In_II Maine_NP1 ,_, Dirigo_NP1 Health_NN1 subsidizes_VVZ health_NN1 insurance_NN1 for_IF certain_JJ individuals_NN2 up_RG21 to_RG22 300_MC percent_NNU of_IO the_AT FPL_NP1 ._. 
These_DD2 subsidies_NN2 are_VBR financed_VVN by_II an_AT1 earmarked_JJ assessment_NN1 on_II health_NN1 insurance_NN1 and_CC self-insured_JJ companies_NN2 and_CC drawing_VVG on_II the_AT state_NN1 treasury_NN1 's_GE cash_NN1 pool_NN1 ._. 
In_II 2008_MC ,_, Dirigo_NP1 had_VHD subsidy_NN1 costs_NN2 of_IO $41.5_NNU million_NNO and_CC operating_VVG costs_NN2 of_IO $2.8_NNU million_NNO ._. 
In_II Pennsylvania_NP1 ,_, the_AT adultBasic_JJ program_NN1 provides_VVZ subsidized_JJ basic_JJ health_NN1 insurance_NN1 to_II legal_JJ residents_NN2 with_IW incomes_NN2 up_RG21 to_RG22 200_MC percent_NNU of_IO the_AT FPL_NP1 ._. 
In_II 2008_MC ,_, the_AT program_NN1 cost_VVD $172_NNU million_NNO ._. 
Due_II21 to_II22 high_JJ demand_NN1 and_CC budget_NN1 constraints_NN2 ,_, the_AT program_NN1 is_VBZ limited_VVN in_II size_NN1 and_CC there_EX is_VBZ a_AT1 substantial_JJ waiting_NN1 list_NN1 for_IF the_AT subsidized_JJ coverage_NN1 ._. 
In_II Vermont_NP1 ,_, uninsured_JJ citizens_NN2 who_PNQS are_VBR not_XX eligible_JJ for_IF Medicaid_NP1 or_CC other_JJ state_NN1 programs_NN2 and_CC do_VD0 not_XX have_VHI reliable_JJ access_NN1 to_II an_AT1 employer-sponsored_JJ plan_NN1 can_VM enroll_VVI in_II a_AT1 "_" Catamount_NN1 Health_NN1 "_" plan_NN1 ,_, and_CC may_VM receive_VVI state-funded_JJ premium_JJ assistance_NN1 if_CS they_PPHS2 meet_VV0 certain_JJ qualifications_NN2 ._. 
In_II state_NN1 fiscal_JJ year_NNT1 2008_MC ,_, Vermont_NP1 paid_VVD a_AT1 net_JJ amount_NN1 of_IO $10.2_NNU million_NNO in_II state_NN1 funds_NN2 for_IF Catamount_NN1 Health_NN1 enrollees_NN2 ._. 
Under_II current_JJ proposals_NN2 for_IF reform_NN1 ,_, these_DD2 individuals_NN2 would_VM be_VBI eligible_JJ for_IF subsidized_JJ health_NN1 insurance_NN1 through_II the_AT national_JJ health_NN1 insurance_NN1 exchange_NN1 ,_, at_II no_AT cost_NN1 to_II the_AT state_NN1 ._. 
Finally_RR ,_, providing_VVG uncompensated_JJ care_NN1 to_II the_AT uninsured_JJ imposes_VVZ a_AT1 "_" hidden_JJ tax_NN1 "_" on_II health_NN1 insurance_NN1 premiums_NN2 for_IF the_AT insured_JJ ._. 
This_DD1 tax_NN1 increases_VVZ premiums_NN2 for_IF all_DB employers_NN2 ,_, including_II state_NN1 and_CC local_JJ governments_NN2 and_CC their_APPGE 19.4_MC million_NNO employees_NN2 (_( 16.5_MC million_NNO as_II measured_JJ by_II "_" full-time_JJ equivalents_NN2 "_" )_) ._. 
By_II greatly_RR reducing_VVG uncompensated_JJ care_NN1 ,_, health_NN1 insurance_NN1 reform_NN1 would_VM reduce_VVI this_DD1 hidden_JJ tax_NN1 ._. 
Table_NN1 1_MC1 shows_VVZ our_APPGE estimates_NN2 of_IO the_AT amount_NN1 spent_VVN in_II each_DD1 of_IO our_APPGE sixteen_MC states_NN2 on_II uncompensated_JJ care_NN1 and_CC the_AT hidden_JJ tax_NN1 on_II the_AT health_NN1 insurance_NN1 provided_VVN to_TO state_VVI employees_NN2 each_DD1 year_NNT1 ._. 
There_EX is_VBZ substantial_JJ variation_NN1 across_II states_NN2 ,_, most_RGT obviously_RR because_CS states_NN2 vary_VV0 greatly_RR in_II size_NN1 and_CC thus_RR in_II the_AT number_NN1 of_IO uninsured_JJ ._. 
But_CCB importantly_RR ,_, in_II each_DD1 case_NN1 ,_, the_AT estimates_NN2 are_VBR large_JJ ._. 
Summing_VVG the_AT sixteen_MC states_NN2 together_RL ,_, we_PPIS2 estimate_VV0 that_CST they_PPHS2 spend_VV0 at_RR21 least_RR22 $4.2_NNU billion_NNO on_II uncompensated_JJ care_NN1 per_II year_NNT1 ._. 
As_CSA described_VVN above_RL ,_, it_PPH1 is_VBZ simply_RR impossible_JJ to_TO track_VVI down_RP every_AT1 state_NN1 and_CC local_JJ program_NN1 that_CST contributes_VVZ to_II covering_VVG the_AT uninsured_JJ ._. 
As_II a_AT1 result_NN1 ,_, true_JJ expenditures_NN2 on_II uncompensated_JJ care_NN1 are_VBR surely_RR substantially_RR larger_JJR than_CSN our_APPGE estimates_NN2 ._. 
Therefore_RR ,_, health_NN1 insurance_NN1 reform_NN1 that_CST greatly_RR reduces_VVZ uncompensated_JJ care_NN1 would_VM reduce_VVI costs_NN2 to_II the_AT states_NN2 by_II more_DAR than_CSN the_AT amount_NN1 that_CST we_PPIS2 identify_VV0 ._. 
This_DD1 is_VBZ true_JJ even_RR taking_VVG into_II account_NN1 the_AT fact_NN1 that_CST some_DD uncompensated_JJ care_NN1 would_VM remain_VVI following_VVG reform_NN1 ._. 
One_MC1 way_NN1 to_TO quantify_VVI the_AT degree_NN1 to_II which_DDQ our_APPGE state-by-state_JJ estimates_NN2 of_IO uncompensated_JJ care_NN1 could_VM be_VBI too_RG low_JJ is_VBZ to_TO compare_VVI it_PPH1 to_II estimates_NN2 using_VVG different_JJ approaches_NN2 ._. 
Hadley_NP1 et_RA21 al_RA22 ._. 
(_( 2008_MC )_) use_VV0 individual-level_JJ data_NN from_II the_AT Medical_JJ Expenditure_NN1 Panel_NN1 Survey_NN1 to_TO form_VVI an_AT1 estimate_NN1 of_IO uncompensated_JJ care_NN1 for_IF the_AT United_NP1 States_NP1 as_II a_AT1 whole_NN1 ._. 
They_PPHS2 estimate_VV0 that_DD1 state_NN1 and_CC local_JJ governments_NN2 spent_VVD $15.9_NNU billion_NNO on_II care_NN1 for_IF the_AT uninsured_JJ during_II the_AT 2008_MC calendar_NN1 year_NNT1 ._. 
We_PPIS2 estimate_VV0 that_CST the_AT hidden_JJ tax_NN1 on_II the_AT insurance_NN1 policies_NN2 provided_VVN to_TO state_VVI employees_NN2 adds_VVZ another_DD1 $1.6_NNU billion_NNO to_II costs_NN2 for_IF state_NN1 and_CC local_JJ governments_NN2 for_IF the_AT country_NN1 as_II a_AT1 whole_NN1 ._. 
Thus_RR ,_, these_DD2 estimates_NN2 suggest_VV0 that_DD1 state_NN1 and_CC local_JJ governments_NN2 spend_VV0 a_AT1 total_NN1 of_IO $17.5_NNU billion_NNO nationally_RR on_II uncompensated_JJ care_NN1 ._. 
Our_APPGE estimate_NN1 based_VVN on_II detailed_JJ analysis_NN1 of_IO state_NN1 programs_NN2 is_VBZ $4.2_NNU billion_NNO for_IF sixteen_MC states_NN2 ._. 
These_DD2 sixteen_MC states_NN2 include_VV0 38.1_MC percent_NNU of_IO the_AT total_JJ population_NN1 of_IO the_AT United_NP1 States_NP1 ._. 
If_CS one_PN1 scaled_VVD up_RP our_APPGE estimate_NN1 to_TO be_VBI an_AT1 estimate_NN1 for_IF the_AT entire_JJ country_NN1 (_( by_II multiplying_VVG by_II 100/38.1_MF )_) ,_, the_AT resulting_JJ number_NN1 is_VBZ $11.0_NNU billion_NNO ._. 
This_DD1 suggests_VVZ that_CST our_APPGE direct_JJ identification_NN1 of_IO expenditures_NN2 on_II uncompensated_JJ care_NN1 is_VBZ indeed_RR a_AT1 lower_JJR bound_VVN by_II a_AT1 significant_JJ margin_NN1 ._. 
C._NP1 Bottom_JJ Line_NN1 for_IF State_NN1 Governments_NN2 Table_NN1 1_MC1 also_RR shows_VVZ our_APPGE estimates_NN2 of_IO the_AT costs_NN2 to_II the_AT states_NN2 of_IO expanding_JJ coverage_NN1 ._. 
For_IF uninsured_JJ citizens_NN2 and_CC legal_JJ permanent_JJ residents_NN2 above_II 133_MC percent_NNU of_IO the_AT federal_JJ poverty_NN1 level_NN1 ,_, current_JJ proposals_NN2 call_VV0 for_IF the_AT creation_NN1 of_IO an_AT1 insurance_NN1 exchange_NN1 with_IW a_AT1 sliding-scale_JJ subsidy_NN1 ._. 
These_DD2 subsidies_NN2 would_VM be_VBI paid_VVN for_IF entirely_RR by_II the_AT federal_JJ government_NN1 ._. 
Therefore_RR ,_, they_PPHS2 would_VM add_VVI nothing_PN1 to_TO state_VVI expenditures_NN2 ._. 
Current_JJ proposals_NN2 call_VV0 for_IF uninsured_JJ citizens_NN2 with_IW incomes_NN2 less_DAR than_CSN 133_MC percent_NNU of_IO the_AT federal_JJ poverty_NN1 level_NN1 to_TO be_VBI covered_VVN by_II Medicaid_NP1 ._. 
Under_II existing_JJ proposals_NN2 ,_, the_AT federal_JJ government_NN1 would_VM pay_VVI 100_MC percent_NNU of_IO the_AT cost_NN1 of_IO this_DD1 addition_NN1 to_II Medicaid_NP1 for_IF the_AT first_MD three_MC years_NNT2 and_CC State_VV0 governments_NN2 would_VM pay_VVI none_PN ._. 
After_II that_DD1 ,_, the_AT federal_JJ government_NN1 would_VM pay_VVI 90_MC percent_NNU and_CC the_AT State_NN1 governments_NN2 would_VM pay_VVI 10_MC percent_NNU ._. 
As_CSA with_IW the_AT current_JJ Medicaid_JJ program_NN1 ,_, only_JJ citizens_NN2 and_CC legal_JJ permanent_JJ residents_NN2 would_VM be_VBI covered_VVN ._. 
Table_NN1 1_MC1 shows_VVZ our_APPGE estimate_NN1 of_IO the_AT cost_NN1 of_IO increased_JJ Medicaid_JJ coverage_NN1 for_IF each_DD1 of_IO the_AT states_NN2 we_PPIS2 analyze_VV0 and_CC for_IF the_AT sum_NN1 of_IO the_AT sixteen_MC ._. 
Again_RT ,_, the_AT estimated_JJ cost_NN1 varies_VVZ substantially_RR across_II states_NN2 because_CS both_DB2 the_AT number_NN1 of_IO people_NN who_PNQS would_VM be_VBI covered_VVN and_CC the_AT Medicaid_NP1 costs_VVZ per_II person_NN1 vary_VV0 substantially_RR across_II states_NN2 ._. 
The_AT total_JJ cost_NN1 of_IO coverage_NN1 expansion_NN1 in_II the_AT sixteen_MC states_NN2 we_PPIS2 analyze_VV0 is_VBZ $11.4_NNU billion_NNO (_( in_II 2007_MC dollars_NNU2 )_) ._. 
In_II current_JJ versions_NN2 of_IO the_AT draft_NN1 legislation_NN1 ,_, states_NN2 would_VM be_VBI required_VVN to_TO pay_VVI zero_NN1 under_II the_AT 100_MC percent_NNU federal_JJ matching_JJ rates_NN2 for_IF the_AT first_MD three_MC years_NNT2 ._. 
Under_II the_AT 90_MC percent_NNU matching_JJ rate_NN1 after_II three_MC years_NNT2 ,_, the_AT amount_NN1 would_VM be_VBI $1.1_NNU billion_NNO per_II year_NNT1 ._. 
For_IF the_AT sixteen_MC states_NN2 we_PPIS2 analyze_VV0 taken_VVN together_RL ,_, the_AT total_JJ net_JJ saving_NN1 is_VBZ $4.2_NNU billion_NNO per_II year_NNT1 for_IF each_DD1 of_IO the_AT first_MD three_MC years_NNT2 when_RRQ the_AT federal_JJ government_NN1 is_VBZ paying_VVG for_IF all_DB of_IO the_AT expansion_NN1 of_IO Medicaid_NP1 ._. 
Importantly_RR ,_, even_CS21 when_CS22 the_AT federal_JJ matching_JJ rate_NN1 is_VBZ reduced_VVN to_II 90_MC percent_NNU ,_, the_AT saving_NN1 to_TO state_VVI governments_NN2 from_II health_NN1 insurance_NN1 reform_NN1 is_VBZ substantial_JJ ._. 
We_PPIS2 estimate_VV0 that_CST the_AT sixteen_MC states_NN2 we_PPIS2 analyze_VV0 would_VM save_VVI $3.0_NNU billion_NNO per_II year_NNT1 with_IW the_AT 90_MC percent_NNU match_NN1 ,_, with_IW the_AT savings_NN2 more_RRR than_CSN offsetting_VVG the_AT additional_JJ Medicaid_NP1 costs_VVZ in_II every_AT1 one_MC1 of_IO the_AT sixteen_MC states_NN2 ._. 
Thus_RR ,_, health_NN1 insurance_NN1 reform_NN1 ,_, far_RR from_II harming_VVG state_NN1 budgets_NN2 ,_, would_VM likely_RR improve_VVI them_PPHO2 substantially_RR ._. 
In_RR21 addition_RR22 ,_, further_JJR savings_NN2 may_VM come_VVI from_II the_AT Children_NN2 's_GE Health_NN1 Insurance_NN1 Program_NN1 (_( CHIP_NN1 )_) ._. 
In_II FY_NP1 2008_MC ,_, the_AT sixteen_MC states_VVZ that_CST we_PPIS2 analyze_VV0 spent_VVN $1.3_NNU billion_NNO on_II CHIP_NN1 coverage_NN1 for_IF low-income_JJ children_NN2 ,_, with_IW the_AT federal_JJ government_NN1 paying_VVG an_AT1 additional_JJ $2.7_NNU billion_NNO on_II CHIP_NN1 in_II these_DD2 same_DA states_NN2 (_( Kaiser_NNB Family_NP1 Foundation_NN1 ,_, 2009_MC )_) ._. 
