GAO Proposes a Change to Calculation of Federal Medicaid Matching Funds

When unemployment grows, so does Medicaid enrollment. States must spend more on Medicaid than they did before, but when unemployment rises, state tax revenues fall, putting the states between a rock and a hard place. They must provide benefits for more people with less money.  The problem is aggravated by the way the federal government is required to calculate each state’s federal medical assistance percentage (FMAP), the rate at which it will reimburse each state for its Medicaid expenditures. The FMAPs are calculated based on a three-year rolling average which compares the state’s per capita income to that of the nation as a whole.  The They change on the first day of the federal fiscal year, (FFY), October 1. HHS publishes them in the Federal Register between October 1 and November 30 of the year before they become effective. For example, the FMAPs effective for FFY 2012, which began October 1, 2011, were published in the fall of 2010 based on per capita income data from 2007 through 2009.

To give states fiscal relief, Congress passed legislation temporarily increasing the FMAP during the two most recent recessions. Most recently, the American Reinvestment and Recovery Act of 2009 (P.L. 111-5) (ARRA) raised the FMAP during the “recession period,” from October 1, 2008 through December 31, 2010. ARRA directed the GAO to study the relationship between FMAP, temporary changes in FMAP and states’ needs during economic downturns and to make recommendations for change.

GAO released an initial report in March, 2011 outlining the need to make the statutory formula more responsive to changing conditions, particularly during periods of economic downturn. In that report, GAO included a prototype Recently, GAO followed up with an analysis of the way its proposed prototype would have affected states’ Medicaid spending.

In previous recessions, Congress gave the states a fairly uniform increase in FMAP. But a one-size fits-all increase doesn’t address the needs of the states that need the most help, while other states receive more help than they need. In the most recent recession, the temporary increase under ARRA had three components: (1) a “hold-harmless” provision, under which no state’s FMAP could decrease during the recession period; (2) an across-the-board increase of 6.2 percent; and (3) an additional increase to states that demonstrated higher growth in unemployment.

According to GAO’s recent report, for several reasons, the temporary increase under ARRA was not as helpful to states as it could have been. First, the increase depended on Congressional action. Second, neither the hold-harmless provision nor the across-the-board increase addressed the needs of the worst-off states. Finally, the timing wasn’t right. The increases started too late and ended too soon.

Under GAO’s prototype formula, increases in FMAP would become effective automatically when certain economic statistics indicated a nationwide recession. At least 26 states would have to have growth in unemployment for two consecutive months or two months of a quarter over the same three-month period the previous year.  The increase would be based on two numbers: the percentage increase in the number of unemployed people and the percentage of decline in wages and salaries. The unemployment rate change used to calculate assistance for a given quarter is the unemployment rate for that quarter compared to the lowest unemployment rate in the previous 8 calendar
quarters.The rationale for this approach is that each 1 percent increase in unemployment leads to a 1 percent increase in state Medicaid spending because of an increase in Medicaid enrollment. Therefore, GAO reasons, there should be a corresponding decrease in the state’s share of its Medicaid expenditures.

Under GAO’s prototype,  the federal share of a state’s expenditures  would be calculated in two steps. First, the percentage of increase in the unemployment rate would be added to the state’s ordinary FMAP. For example, if the state’s ordinary FMAP was 50 percent and there was a 5 percent increase in unemployment, the state’s share would decrease by 5 percent,  from 50 to 47.5  (50 multiplied by 5 percent is 2.50). The federal share would increase to 52.5 percent.

The second step would be measured by the percentage of the decline in wages and salaries, based on research showing that a 1 percent drop in wages and salaries leads to a 1 percent cut in revenue. If the state’s wages and salaries dropped by 10 percent, the state’s share of its Medicaid expenditures would drop by 10 percent. For example, the share of  a state with a 50 percent FMAP with a 10 percent drop in wages and salaries would decrease by 10 percent of the change (50 times 10 percent = 5) to 45 percent. In any quarter, a state could receive an unemployment-based increase, a wage-based increase, both or neither.

According to GAO, if the prototype formula had been applied during the 2008 recession, assistance to the states would have begun earlier, lasted longer, and transitioned more smoothly. In the first quarter  of 2008, 41 states would have received increased assistance.  When the number of states with increased unemployment rates fell below 26, the assistance would be “triggered off”. Rather than stopping abruptly, the formula would reduce FMAP more gradually.GAO calculates that if the last quarter of assistance had been the first quarter of 2011, the average reduction in FMAP would have been 0.54 percent, with only seven states experiencing a reduction of more than 1 percentage point.

CMS did not agree with the recommendation on the grounds that determination of the growth in unemployment and declines in revenue would be difficult and that the formula was too complicated. GAO responded that the current FMAP formula is equally complicated, the federal funds would be spent more effectively without waiting for Congress to make difficult political decisions.  Because the FMAP formula is set in statutes, legislation would be required.