Academic Articles Medicaid/Medicare Public Policy

Fiscal Shenanigans and Targeted Health Care Funds

Katherine Baicker and Douglas Staiger (2004) have a working paper detailing how states often expropriate federal health care funds to use in their general budget. The paper shows that while federal dollars may not always reach the intended destination, these programs can still be somewhat effective in improving health outcomes.

DSH Program

Federal Medicaid Disproportionate Share Hospital (DSH) program attempts to assists hospitals whose patients are mostly Medicaid recipients or are uninsured. The program was created in 1986 and 1989 and gives extra funds to hospitals which serve these underprivileged communities. By 1998, the program had reached $16.5 billion, which represents 9% of all Medicaid payments to suppliers. Medicaid DSH payments are determined by individual states and matched by federal grants.

An example of what some states did is the following:

  • The state would give $1 million to county hospitals. The federal government would then match this at the Federal Medical Assistance Percentage (FMAP) rate, which depends on the state’s wealth. After the state received the funds from the federal government, they levied a surcharge on the county hospital for $1 million, thus increasing hospital funding through the federal matching, but contributing nothing at the state level net of the surcharge.


(1-h)pf(DSH)+hpf(DSHIGT) – DSH(1-FMAP) + h(IGT)

pf(X) is the public health benefits from DSH payments, h is the proportion of hospitals which are publicly owned. This fact is important since state governments can only expropriate money from public hospitals. IGT stands for Intergovernmental Transfers which are the surcharge states impose on county hospitals. The first two terms represent the beneficial health outcomes at private and public hospitals respectively. DSH(1-FMAP) represents the state contribution and h(IGT) represents funds diverted back to the state.

The first order conditions are:

  • pf'(DSH-IGT)=1; The marginal benefit of net payments to hospitals will equal 1
  • pf'(DSH)=1-(FMAP/(1-h)); The marginal benefit of payments to private hospitals is set equal to the net marginal cost of these payments to the state

Which states are involved in ‘fiscal shenanigans’?

In order to find which states are involved in these fiscal shenanigans, Baicker and Staiger use three measures:

  1. An Urban Institute survey of 34 states which explicitly asks questions regarding the prevalence of using IGT to expropriate funds in the DSH program.
  2. DSH/(No. Medicaid, uninsured patients). States with a large amount of spending per poor patient is more likely to expropriate DSH funds.
  3. The percentage of DSH dollars going to public hospitals. States cannot expropriate funds from private hospitals so this statistic also measures the likelihood states using IGT.

Calculating ‘Effective’ DSH funds

Since not all of the funds actually reach the hospitals, Baicker and Staiger calculate the amount of funds that effectively reach the hospitals. They use the following regression, where the ‘capture‘ variable represent one of the three proxies listed above for the propensity to expropriate funds, and ‘X’ is the change in certain state level variables.

  • change IGT=b_0+b_1*(1-capture)*DSH + b_2*(capture)*DSH+B_3*X+e

Effect on Health Care Outcomes

Now Baicker and Staiger can how ‘effective’ DSH funds—the ones the hospitals actually receive—affect health outcomes against ‘ineffective’ DSH funds—the dollars the federal and state governments claim to have provided. While ‘effective’ and ‘ineffective’ DSH both reduced infant mortality and post-heart attack mortality, only effective DSH was statistically different from zero. Further, the point estimates for effective DSH were larger (in absolute value) than the ineffective DSH measures.


Using DSH funding, it costs $11 million dollars to save one baby’s life through reduced infant mortality and $12 million to save one adult’s life through reduced heart attack risk. These figures do not take into account that DSH funds are used to treat illnesses outside these two. Since a statistical life is often estimated to be valued between $6-$10 million, these funds may be well spent once we take into account that DSH treats other illnesses as well. Baicker and Staiger’s conclusion is that, while funds from targeted federal programs may not entirely go to their desired destination, they can still be effective means of implementing policy.