Standardized MAGI Conversion Methodology- General Questions

Transcription

1 Standardized MAGI Conversion Methodology- General Questions Q1. What are the reasons that a marginal (25 percentage points of FPL) method was chosen instead of the average disregard approach? A1. The marginal disregard method differs from the average disregard method originally considered in that the average disregard approach would have included all people in a given eligibility category, regardless of their net income. The marginal disregard method, in comparison, only includes individuals in the 25 percentage points of the FPL band below the net income standard. The marginal approach was selected for two important reasons. First, the average disregard approach includes people with incomes well below current net income standards, and these people also tend to have low disregards. As a result, the average disregard approach biases the conversion downward, potentially creating systematic bias in which there is a net loss of eligibility due to conversion. The marginal approach, on the other hand, focuses on people with higher net incomes who also tend to have higher disregards. As a result, the marginal disregard approach appropriately adjusts for people who are most at risk of losing eligibility due to the elimination of disregards. Second, state data can be incomplete with respect to disregards for people who have incomes well below current net income standards (e.g., people who are clearly eligible before the application of disregards), so incorporating these individuals into the calculation could skew or bias the result. More detailed information about how the marginal disregard approach was selected and the alternative methods that were considered is available in the ASPE issue brief Modified Adjusted Gross Income (MAGI) Income Conversion Methodologies. 1 Q2. How was the 25 percentage FPL point band selected for the standardized methodology? Why not a smaller or larger band? A2. The 25 percentage FPL point band was selected for two main reasons. First, the 25 percentage point band captures the majority of people who are at risk of becoming ineligible due to disregards because most eligible individuals have disregards that are less than 25 percentage points of the FPL. Second, the 25 percentage point band creates a large enough range to ensure adequate sample size for analysis. When tests were conducted on smaller marginal bands (for example ten percentage points of FPL), sample sizes were smaller with no clear improvement in accuracy. 1 Available at: 1

2 Q3. Why did CMS choose the SIPP as the data source for the standardized methodology? A3. The SIPP was selected for the Standardized MAGI Conversion Methodology for several important reasons. First, the SIPP contains data on monthly (rather than annual) income, which is consistent with the income period states use to evaluate Medicaid and CHIP eligibility. Second, the SIPP provides data needed to distinguish income sources that may be treated in different ways under current rules and under MAGI rules and to model the disregards currently available for a given state and eligibility group. Finally, the longitudinal nature of the SIPP allows for the assessment of other characteristics important for eligibility, such as pregnancy and length of disability. More detail about the reasons for choosing the SIPP is available in the ASPE Issue Brief Data Sources for Modified Adjusted Gross Income (MAGI) Conversions. 2 Q4. What are the main differences between using the SIPP and state data for income conversions? A4. There are several important differences between using SIPP and state data for income conversions. The first has to do with the population included. The SIPP analysis will be based on all individuals who are simulated to be eligible for Medicaid in a particular state, regardless of whether they are actually enrolled. In contrast, state data would use only the population of people who have actually been determined eligible for Medicaid and enrolled. There are also differences in the level of detail at which analysis can be done. This is primarily due to data constraints and feasibility of programming in the SIPP. For example, the SIPP modeling uses simplified household composition rules, while state data would reflect all state rules as the state actually applies them. The state data can also capture all disregards that the state applies, while the SIPP modeling can only capture major disregards. Finally, there are differences in the extent to which SIPP and state data can address seasonal variation. The SIPP analysis won t explicitly account for seasonal variation because it is based on a single month of data. However, weights used in the SIPP analysis account for this to a certain extent because they are representative of the population on an annualized basis. Using a full year of state data, on the other hand, could address any concerns about seasonal variation in income and/or disregards. Q5. What are the reasons that the standardized methodology using SIPP data does not take into consideration whether an asset test currently applies to an eligibility group? A5. The standardized methodology does not make an adjustment for asset tests because Section 1902(e)(14)(A) of the Social Security Act requires income conversion to account for adjustments 2 2

3 in income under MAGI such as the loss of income disregards. In particular, the Act specifies that converted income standards shall be established that are not less than the effective income eligibility levels that applied under the state plan or a waiver of the plan on the date of enactment of the Affordable Care Act. The elimination of asset test is reflected in a different part of the statute, and is not related to the income conversion process. Q6. Are states that don t expand Medicaid still required to convert MAGI groups from net to gross standards? A6. Yes. Q7. When do states begin applying converted income standards to current Medicaid beneficiaries? A7. Income conversion is not something that happens at the level of the individual enrollee. Instead, the converted standard becomes the new way of assessing eligibility for people who apply for Medicaid or whose eligibility is being determined in Converted income standards will be used to set maximum MAGI-equivalent standards for adults in 2014 and will be used as the actual income standard in effect for children through October States will use the state plan amendment (SPA) process to identify the minimum and maximum MAGI-equivalent standards and to select the state s MAGI-based income standard for each eligibility group to which MAGI will apply in For adults for whom the Maintenance of Effort requirement expires in 2014, the selected income standard in the SPA will be anywhere between the minimum allowed and the maximum derived through the income conversion process. 42 CFR (a)(3) of the March 2012 final eligibility rule specifies when MAGI is implemented for Medicaid and CHIP beneficiaries eligible for coverage to begin on or before December 31, It specifies that MAGI will not be applied until the later of March 31, 2014 or the individual s next regularly-scheduled renewal of eligibility. At that time, states will compare their MAGI income to the MAGI-based standard specified in the State Plan Amendment for the applicable eligibility group to determine if the individual is still eligible. Q8. Is it necessary to do separate conversions for children by age group in Medicaid? A8. Yes, conversions must be done separately for each age group (e.g., <1, 1-5 and 6-18) as required in the March 2012 final eligibility rule at (c)(2). Also, if a state s current 3

4 standards are different from the pre-chip Medicaid standards as of March 31, 1997, states will need to convert the 1997 levels as well to ensure proper application of CHIP maintenance of effort requirements as well as FMAP claiming for CHIP coverage in

5 Standardized MAGI Conversion Methodology- State Data Questions Q1. For states that want to use their own data to implement the Standardized MAGI Conversion Methodology, what are the key steps in the analysis? A1. Detailed information on how to use state data to apply the standardized conversion methodology is included in the ASPE issue brief How States can Implement the Standardized Modified Adjusted Gross Income (MAGI) Conversion Methodology from State Medicaid and CHIP Data. 3 In general, there are five key steps that states will need to follow: 1. Determine which income standards for which eligibility groups (Medicaid state plan, S- CHIP state plan, or 1115 demonstration) need to be converted; 2. Determine the time period to use for extracting eligibility records; 3. Select the appropriate records to use in the analysis; 4. Pull the necessary data elements to do the analysis; and 5. Perform the income conversion calculations. More detail on each of these steps is included in the ASPE Issue Brief, and in the detailed Q&A below. Which eligibility groups need to be converted? Q2. Which eligibility groups and standards need to be converted? A2. The finalized income conversion templates that CMS is providing to states will serve as a guide for which standards need to be converted for eligibility determinations. The conversion will be based on states effective income standards for each relevant eligibility group under State plans and waivers, using the higher of the standard in effect on March 23, 2010 and the standard that will be in effect on December 31, Those standards are bolded on sheet 1 of the Excel template and are specified on the Summary Document. Income standards based on gross income will not be converted, and so are not bolded on the template. For eligibility groups that will be consolidated in 2014, most of the time states will only need to convert one standard per group. For example, if the state currently covers infants to 133% FPL through the mandatory group for poverty-level related infants and to 225% FPL through the optional group for poverty-level related infants and the same disregards are used for each group 3 Available at: 5

6 (except for any block income disregards which are included in the effective income level being used for conversion), only the higher of these two will need to be converted to determine the maximum standard for infants under the consolidated group. Q3. Do states need to convert income standards for childless adults that states believe will not be in effect in 2014 and that do not affect FMAP calculations? For example, if a state plans to reduce childless adult coverage under its 1115 demonstration from 200% FPL to 100% FPL, is it necessary to convert the 200% level? A3: States that are expanding coverage to the new adult group will need to convert their 1115 demonstration income standards for childless adults that were in effect as of 12/1/09 to ensure that appropriate FMAP claiming is possible. In addition, while states may know their intentions regarding 1115 demonstration coverage in 2014 for childless adults, states and the federal government have not started to formalize and approve those changes. As a result, CMS recommends that states consider converting all income standards that could be relevant in 2014, including their 1115 demonstrations for childless adults. CMS will be performing SIPP conversions for all categories bolded on state conversion templates, including 1115 demonstrations for childless adults, if such demonstrations use a net income standard. States should describe their income conversion approach (and information about which groups are being converted) in the MAGI Conversion Plan, which are due by April 30, 2013 for states using the MAGI Standardized Methodology with state data or an alternate methodology. CMS welcomes state questions and discussion on this issue as they undertake their conversion activities. Q4. Do states need to establish converted income standards for their 1988 and 1996 AFDC minimum standards for low-income parents? How should states convert the 1988 and 1996 minimum standards? A4. Yes. In CMS second proposed eligibility rule from January 2013, CMS proposed that the Federal minimum income standard for Section 1931 be converted. CMS explained that the proposal was aimed at protecting parents and caretaker relatives who may not have a coverage option if the state reduces its income standard to the federal minimum allowed and does not expand coverage to the new adult group. 4 The 1931 federal minimum are the state s May 1, 1988 AFDC payment standards by family size. States are also required to establish converted income standards for their 1996 AFDC income standards by family size because this is often the 4 78 FR

7 state s current income limit for at least one eligibility group, or a state may lower its current limit for a group to this when maintenance of effort ends. This information will be needed for states Medicaid state plan amendments for States that currently cover a group that uses the same AFDC disregards that were in place for these groups in 1988 and 1996 can use their current data to do these conversions (e.g., if the state s 1931 group for low-income families uses the same income disregards as the state s 7/16/1996 AFDC). Similarly, if they have older data that would be relevant (e.g., if they covered a group with AFDC disregards in 2008), those data can be used for conversion. Generally, this will involve selecting records within the current group that are within 25 percentage points of FPL below the minimum standard being converted. States would then need to apply the method for converting fixed dollar thresholds which is described under Q20 below and in more detail in the ASPE issue brief How States can Implement the Standardized Modified Adjusted Gross Income (MAGI) Conversion Methodology from State Medicaid and CHIP Data. 5 If a state doesn t have current data that reflect the disregards for the AFDC minimums, it also has the option of using the SIPP results for these categories, even if it is using state data for other conversions. Q5. Are there other times when states that are using their own data for income conversion could choose to use the SIPP results for certain groups? A5. States may also choose to use the SIPP results for groups where enrollment is too low to do the conversions with much confidence about the precision of the results. However, states will need to use the same data source and method for conversions for any group that is required to be converted for both FMAP-claiming and eligibility-related purposes. For example, if a state uses state data for its parent/caretaker relatives eligibility group, it will need to do so for the March 2010/current eligibility standards as well as for the December 2009 eligibility standards. Time Period to Use for Conversion Q6. What time period should states use for their income conversion analysis? A6. In general, states should use a time period of data that will produce unbiased results. For example, if the state uses 12 months of the most recently available data, it could be reasonably certain that the results do not contain bias resulting from seasonal variations in income or 5 Available at: 7

8 disregards in their Medicaid populations. States that choose a shorter time period than 12 months will need to be able to demonstrate that the time period selected did not lead to biased results (see more about this in Q7 below). If income standards and/or disregards have changed over time, the state should be careful not to use data from time periods when different sets of rules were in place. In this case, the time period for analysis should be no longer than the number of months of data available using current rules. In some cases, this could be fewer than 12 months. There is one exception to the general rule that the most recent available data should be used. In the situation where income standards and/or allowable disregards have been reduced since March 23, 2010, states will need to use the previously applicable higher standards/disregards for conversion instead of the most current ones, and use a time period for analysis for up to 12 months prior to the rule change. This is because the converted standards set the maximum for 2014 and beyond as described in Q7 of the Standardized MAGI Conversion Methodology- General Questions. Using the previously applicable higher standards preserves the maximum flexibility for states when setting their MAGI maximum income standards. Q7. How can states show that a time period less than 12 months is unbiased? A7. One approach would be for states to extract a single month of data, and to analyze whether disregards vary depending on the month in which an individual s eligibility was most recently determined. For example, a data file for December 2012 would contain information about people who were determined eligible at various points during the year. If people whose eligibility was determined at one time of year do not have systematically different disregards compared to people whose eligibility was determined at other points in the year, then the state might conclude that using a single month of data would not lead to biased results. A second approach would be to do an analysis of disregards for single months at different points in the year e.g., January, May, September. If states find similar results, that would also support the argument that using a time period shorter than 12 months is unbiased. Finally, states that do not apply any disregards that are likely to vary seasonally could also make the argument that using a time period shorter than 12 months is unbiased. Q8. What year s FPL should be used in the income conversion analysis? A8. The FPL amounts by family size should correspond to the year of the data that is being used for the analysis. For example, if data for calendar year 2012 is being used for the income 8

9 conversion calculations, then the FPL amounts for 2012 should be used. For example, when expressing disregards as percentages of FPL, the amounts used for each FPL by family size should correlate with the FPL guidelines from Selecting Appropriate Records Q9. What is the correct income conversion starting point for eligibility categories that currently use block of income disregards? A9. If a state has a block income disregard in place for an eligibility group (for example, a block of income disregard between the AFDC payment standard and 110% FPL for the 1931 group) this disregard should be added to the net income standard to calculate the effective income level ), which is the income threshold that is converted for MAGI (e.g., 110% FPL in this example). The state should use 110% FPL as the net income standard, and calculate the converted standard by adding back in the average value of any other disregards to that standard within the relevant 25% FPL marginal band of individuals (i.e., from 85% to 110% FPL). The income conversion templates being finalized by CMS with the states take block of income disregards into account in determining the starting point for income conversion, which are bolded on sheet 1 of the Excel template. Q10. Can case-level data be used for the analysis? A10. In general, income conversion needs to be done at the individual level, because family income and disregards may vary for different individuals in the eligibility unit. However, it is possible to use case-level data for conversions if the case includes: ONLY people who are enrolled in the same eligibility group ONLY people whose income is counted together to determine Medicaid eligibility, and who have the same disregards. If cases are used, case-level records should be weighted by the number of people in each case. 9

10 Q11. What are some examples of when case-level data would be problematic to use for income conversion? A11. One common example of when it would be problematic to use case-level data is when the case or family includes people who are enrolled in different eligibility groups e.g., multiple children who are enrolled in different Medicaid or CHIP eligibility groups, or a parent whose income is counted for his spouse s eligibility under section 1931, but whose eligibility is determined under a disability category. Another example would be when the eligibility unit is different for people within the case---for example, in families with stepparents where stepparent income does not count toward the stepchild s eligibility, but would count toward eligibility for the spouse/ biological parent; or when sibling income is not counted for sibling eligibility. Q12. How should states that maintain their eligibility data at the case level that can t be used for income conversion (see Q11 above) apply the Standardized MAGI Conversion Methodology? A12. These states will need a way to properly allocate income and disregards to individuals. One potential method for accomplishing this would be to use a sample of records and do a detailed review to ensure that the correct income and disregards are applied for each sampled individual. More information about sampling methodologies is included in Q13 below. Q13. Within each eligibility group, should states that are applying the standardized method use all individuals whose net incomes are within 25 percentage points FPL of the current net income standard to do the conversion, or is it permissible to use a sample? A13. States may choose to use all individuals whose incomes fall within 25 percentage points of FPL of the current net income standard (or narrower band as may be applicable), 6 or may choose to use a sample (for example, one reason to use a sample might be if the state currently stores data at the case level and needs to allocate information to the individual level). In developing a sampling plan, states that choose a sampling approach should evaluate approaches used by the state for other purposes (such as PERM) with respect to sampling criteria, error rates, precision requirements, and confidence levels. Once a sampling plan has been established, states may pull the sample using the OIG s free statistical software package RAT-STATS. 7 6 A narrower band would be used if the current net standard is below 25% FPL. 7 Available at: stats/index.asp 10

11 Q14. In implementing conversion, should states use data from all records in the marginal 25 percent of the FPL band, including applicants, redeterminees, and established beneficiaries? Won t including established beneficiaries lead to double counting, particularly if we use data from multiple months? A.14. States should use data from all applicants, redeterminees, and established beneficiaries present in each month used in the analysis, even if this means some individuals are counted multiple times. For example, if the state is using data from January through December 2012 for its conversions, an individual who enrolled in March would appear 10 times and an individual who enrolled in October would appear 3 times. Limiting the sample to applicants only, or selecting applicants and redeterminees in each month while ignoring established beneficiaries, could bias the sample by including too many applicants relative to other enrollees. If applicants have disregards that are systematically different from other enrollees (for example, if applicants tend to have higher incomes than established beneficiaries), average disregard calculations could be inaccurate. Performing Calculations Q15. If a state wants to use the Standardized MAGI Conversion Methodology with its own data, what data elements will it need to use? A15. In general, states will need 1) information on net income of each person and the size of the Medicaid eligibility unit to establish which enrollees fall within the 25 percentage point band below the current net income standard; and 2) data on the total amount of disregards for each individual within the 25 percentage point band if this is not stored as a data element in the state s system, this can be calculated by adding up individual disregards, or as the difference between gross income and net income. Q16. Which disregards are included in the conversions using the SIPP? A16. Although the income conversions that are being done with the SIPP data include the major disregards that states currently use to determine Medicaid eligibility, the modeling does not include every disregard that applies in every state. The SIPP does not collect information on every income category or disregard that an individual state may use to determine Medicaid eligibility, and feasibility of programming every disregard that a state may use also factored into the final decision about which disregards to include in the SIPP modeling. 11

12 The disregards that are being included in the SIPP income conversion analysis include earned income, unearned income, child support paid, child support received, interest income, dividend income, SSDI income, children s income (earned and unearned), and dependent care expenses. The model uses state-specific rules for each of these disregards. Q17. Are states required to use all disregards or is it allowable to use only the disregards that are being applied in the analysis using the SIPP data? A17. For some states, the fact that state data include complete information on all disregards used to determine eligibility is an attractive reason to use state data in the conversions. In addition, states may find it easier to use data on total disregards (or gross income minus net income) in the analysis rather than adding up individual disregards. For these reasons, states using their own data may want to use all disregards rather than a subset in performing their analysis. States will not be required to include all disregards, but at a minimum must include the disregards that will be included in the SIPP conversions (see explanation in Q16 above). Since the SIPP data include the major disregards that are currently used, we expect that a decision to use all disregards or only the disregards being used in the SIPP analysis will not have much impact on income conversion results from state data. Q18. In performing the income conversion calculations, how should states handle income disregards that are time-limited? A18. Separate conversions for applicants and beneficiaries are permitted in cases where there are temporary earned income disregards (e.g. disregarding one-third of earned income for 1931 groups) that vary for applicants and beneficiaries. For other types of time-limited disregards, the analysis does not need to distinguish between enrollees receiving time-limited disregards and enrollees whose time-limited disregards have expired. By calculating the average disregard amount using the actual distribution of both individuals receiving time-limited disregards and individuals who are no longer eligible for these disregards, states will appropriately adjust for the probability of receiving a time-limited disregard. Q19. How should states handle income conversion in situations where income disregards are different for applicants vs. beneficiaries already enrolled? A19. Where different disregards apply for applicants vs. beneficiaries, separate calculations of the average disregard amount for individuals within the relevant income range for conversion should be performed. States would use the same procedure as outlined in the ASPE issue brief 12

13 How States can Implement the Standardized Modified Adjusted Gross Income (MAGI) Conversion Methodology from State Medicaid and CHIP Data 8. However, for groups with different disregards, they would calculate one converted standard using data from applicants, and a separate converted standard using data from established beneficiaries who are no longer eligible for applicant disregards. The process ultimately would lead to two converted standards, one that would be used when individuals initially apply, and a second that would be used at the next redetermination after the time-limited disregard expires. Q20. How should states apply the income conversion methods to eligibility groups where the income standards are defined as fixed dollar amounts rather than in terms of FPL? A20. The ASPE issue brief How States can Implement the Standardized Modified Adjusted Gross Income (MAGI) Conversion Methodology from State Medicaid and CHIP Data 9 provides detailed instructions and examples that illustrate how to handle conversions of standards that are expressed in fixed dollar limits rather than FPL. To perform conversions for these eligibility groups, states should first translate the fixed dollar income limits into FPL terms for each family size. States should then define a 25 percentage point band for each unit size. This will be the starting point for income conversion. Next, states should calculate average disregards for people in the relevant bands as a percent of FPL, and add this amount to each family size unit s FPL standard. The resulting FPL-based standards should then be converted back to fixed dollar amounts by family size. Q21. Why did the recommended approach for converting fixed dollar standards change? A21: The previous approach defined the marginal band based on the weighted average FPL across unit sizes, and added the average disregard amount to the weighted average FPL for all unit sizes combined. The problem with this approach was that it could result in a converted standard below the current standard for some family sizes. The new approach defines the marginal band separately by family size. This has two important advantages. One, it never results in a converted standard that is lower than the current standard. Two, this method better maintains existing variation in state standards across family size categories. 8 Available at: (revised April 2013) 9 Available at: (revised April 2013) 13

14 Q22. How will the 5% disregard required under Section 2002 of the Affordable Care Act be applied? A22. The 5% disregard will be applied after income conversion, when determining eligibility for an individual. In the January 2013 notice of proposed rulemaking 10 CMS proposed that the 5% disregard be applied only to the highest threshold under a MAGI-based income group for which an individual could be eligible. For example, in a state that does not expand coverage to the new adult group, if the income standard for eligibility under section 1931 in a state were 90 percent of the FPL and a parent with 95 percent of the FPL who met the categorical requirements for coverage applied, the 5 percent disregard would apply to that parent resulting in eligibility for the section 1931 category. However, if the state does expand coverage to the new adult group, the 5 percent disregard will only be applied to the new adult group (individuals between 133 and 138 percent of the FPL). In such situations, the 5 percent disregard would not apply to parents and caretaker relatives between 90 and 95 percent of the FPL, so such individuals would be eligible for coverage in the adult group. CMS will consider comments received in response to the NPRM in finalizing the policy around application of the 5% disregard. 10 January 22, 2013 Notice of Proposed Rulemaking, available at /pdf/ pdf 14