You are here: Social Security Administration > Office of Retirement Policy > Projection Methodology

Projection Methodology

The Office of Retirement Policy uses the Modeling Income in the Near Term, Version 7 (MINT7) microsimulation model to make all projections related to beneficiaries on this website. SSA's Office of the Chief Actuary makes Trust Fund solvency projections. MINT7 is built from Social Security's administrative records matched to the Survey of Income and Program Participation (SIPP), a detailed Census survey. It projects 21st century retirement income, marital trends, Social Security benefits, income, and poverty. The model builds the projections from the economic, demographic, and programmatic assumptions in the 2012 Trustees Report. These projections allow researchers and policy analysts to study future retirement conditions such as income and poverty, as well as the effects of policy changes.

For more information on the model methodology, see the Urban Institute's MINT7 research paper.

Definitions—Measures and Labels Used In Policy Option Projection Tables

Scheduled and Payable Benefits

Scheduled Benefits are those that current law promises. Payable Benefits are those that projected Trust Fund balances can finance. Payable benefits are an across-the-board reduction from Scheduled Benefits in a given year starting with the Trust Fund exhaustion in 2033; therefore, Payable Benefits and Scheduled Benefits are the same in 2030. The SSA Actuaries estimate that the across the board reduction under a Payable baseline will be 23.5 percent in 2050 and 24.6 percent in 2070. These reductions, which come from the 2012 Trustees Report assumptions, would apply to the final benefit amounts, not the average indexed monthly earnings (AIME) or the primary insurance amount (PIA). Both the Scheduled Benefits tables and the Payable Benefits tables compare the respective baseline to a policy option built on top of Scheduled Benefits.

Table Subgroups

Sex: Female or Male.

Ethnicity/Race: We list Hispanics first; the rest of the groups are non-Hispanic.

Country of Birth: We differentiate between U.S. and foreign.

Age: Our population includes those aged 60 and older because it is the earliest eligibility age for any aged benefits under current law.

Marital Status: Refers only to the marital status in the year of analysis. An individual's marital status can change in the future and may have been different in the past.

Education: Is the reported number of years of education.

  • Graduate means >16 years of education,
  • Bachelor means 16 years of education,
  • Associate means 14–15 years of education,
  • High school means 12–13 years of education, and
  • Less Than 12 Years means <12 years of education.

Official Poverty: This indicates whether the person is in a household that has income above or below the official poverty line under current law (Scheduled Benefits). The household income used for the official poverty measure differs from the household income used elsewhere in our results because the official measure counts asset income differently. The official poverty measure of asset income includes dividend income, interest income, and rental income as reported on income tax returns. In contrast, the asset income value used for the household income measure is based on the annuitized wealth a household has each year from things like 401(k) accounts and personal savings. This annuitized value is considered the asset income for households. The household income measure generally produces a substantially higher asset income value than the asset income value used for official poverty calculations.

Household Income Quintile: Represents an individual's annual household income, including:

  • household earnings;
  • asset income, which includes income from defined contribution plans;
  • defined benefit pensions;
  • means-tested income;
  • non-means-tested income;
  • Social Security (scheduled benefits);
  • SSI; and
  • non-spousal co-residents' income.

We calculate the quintiles for each year for all elderly beneficiaries aged 60 or older. For example, in 2030, we tabulate the household income of all elderly beneficiaries aged 60 or older, determine the dollar thresholds for each quintile, and assign each beneficiary to the appropriate quintile. The dollar ranges shown are in annual real 2015 dollars.

Current Law Benefit Type:

  • Retired worker: receives only a retired worker benefit on his or her earnings record.
  • Widow(er): receives a survivor benefit (may or may not receive a lower worker benefit from his or her own earnings record, also known as dually entitled).
  • Spouse: receives a spousal benefit (may or may not receive a lower worker benefit from his or her own earnings record, also known as dually entitled).
  • Disabled worker: receives a disabled worker benefit on his or her earnings record and is under the Full Retirement Age (FRA). Disabled workers convert to retired workers at FRA.
  • None: no benefit in the analysis year, but may start benefits in a later year or may never receive a benefit. See the Never Beneficiaries fact sheet for more information.
Affected Threshold for Those With Lower or Higher Benefits/Income

To be considered affected, the degree of difference from Scheduled or Payable Benefits must be equal to or greater than 1%, or equal to or less than -1%. We consider those with differences between -1% and 1% to be unaffected.

The amount of the difference required for an individual to be considered affected can change the analytical results to a great degree. Beneficiaries affected by rule changes but who will still receive the same benefit amount are not considered affected in our projections.

Median Percent Change vs. Scheduled/Payable

We produce the median percent difference by calculating the percent differences for each individual, ordering this range, and then using the middle value when there is an odd number of individuals or the average of the two middle values when there is an even number of individuals.

Sample Size Restrictions

If the sample size is less than 20 for any subgroup, we omit the data for the category in which the subset appears (for example, Current Law Benefit Type or Age) for confidentiality reasons.

Help reading the tables

Tables 1 & 2. Individual Benefit Changes/Household Income Changes

This is the median individual percent change in benefits (or household income) for all and affected beneficiaries. The tables also show the percent with lower or higher benefits (or household income).

Table 3. Poverty

Poverty rate shows the percent of the population that is under the official poverty threshold. The change is the percentage point change in the poverty rate, not the percent change in the poverty rate.

CORRECT interpretation: the poverty rate would drop by 1.4 percentage points.
INCORRECT interpretation: the poverty rate declined by 1.4%.

Table 4. Beneficiary Characteristics

Read Table 4 vertically because it shows how the column population is distributed in each of the various subgroups. The total for each population column is 100%, because the percentages for each subgroup add to 100%.

CORRECT interpretation: X% of affected beneficiaries are female, Y% of affected beneficiaries are male
INCORRECT interpretation: X% of females are affected, Y% of males are affected

Table 5. Beneficiary Status

Read this table horizontally for each subgroup. It starts with current law beneficiaries and shows changes to that population due to the option, finishing with the option beneficiary population. All populations are in thousands, so a population of 73,810 is 73,810,000 or 73.8 million people.

  • Current Law Beneficiaries: Current law Social Security beneficiaries aged 60 and older.
  • Changes in beneficiary status under proposed option covers three columns:
    • Added: Number of beneficiaries added under the option who were not beneficiaries under current law. A plus sign (+) shows additions to the beneficiary population.
    • Removed: Number of beneficiaries who have a benefit under current law, but do not have a benefit under the option. A minus sign (-) shows losses to the beneficiary population.
    • Net Change: This is the difference between the current law and option beneficiary populations. A plus or minus sign is used depending on whether the total change is an addition or subtraction to the current law beneficiary population.
  • Policy Option Beneficiaries: Beneficiary population under the option.

Note: Due to rounding, the columns may not add up perfectly.

Benefit Rules Modeled

There are many Social Security program rules and they are often complicated. The MINT model includes most of the basic rules, but does not account for them all. We based the scheduled and payable Social Security benefit projections on the fullest set of current law benefit rules that the MINT projections can support, including:

  • Average Indexed Monthly Earnings (AIME)
  • Primary Insurance Amount (PIA)
  • Special minimum PIA
  • Windfall Elimination Provision (WEP)
  • Cost of Living Adjustment (COLA)
  • Early retirement reduction factors and delayed retirement credits (DRCs)
  • Spousal and divorced spousal benefits
  • Surviving spouse and divorced surviving spousal benefits
  • Independent entitlement for divorced spousal benefits
  • Widow's limit
  • Child benefits
  • Child-in-care spousal benefits
  • Government Pension Offset (GPO)
  • Family maximum
  • Retirement earnings test (RET)
  • Lump sum death payment (LSDP)

Social Security benefit rules not included in the MINT model:

  • Benefit rules from before 1979, such as earlier PIA formulas or special benefits that no longer exist.
  • Monthly benefit adjustments: MINT only has annual earnings and income amounts, so monthly benefit rules, such as the earnings test, are calculated on an annual basis.
  • Complicated claiming situations: MINT projects a single claiming age and does not distinguish between claiming survivor, worker, or spousal benefits. It also does not allow for sophisticated claiming strategies (file and suspend, withdrawal of application, etc.).
  • Disability administration and adjudication: MINT does not have information on the type of medically disabling condition, if the initial claim was denied, how long it takes to appeal and receive benefits, etc. In our projections, disability benefits are paid from the onset date until death or upon converting to retired worker benefits. While the MINT model projects up to three disability spells, our benefit calculator is not structured to handle them.
  • Disability return-to-work rules including the trial work period, extended period of eligibility, and Continuing Disability Reviews (CDRs).
  • Return to work initiatives (Ticket to Work, Benefit Offset National Demonstration (BOND), Work Incentives and Planning Assistance (WIPA), etc.)
  • Dependent Parents: MINT does not include information on parents.
  • Disabled Adult Children (DACs): MINT does not include the children's disability status or the earnings records needed to estimate eligibility (children are ineligible if they can receive their own disabled worker benefit).
  • Taxation of benefits: Although the MINT model has income tax projections, the benefits and income shown for baselines and policy options are pre-tax.
  • Representative Payees: MINT does not have any information needed to model beneficiaries' representative payees.

MINT projects Supplemental Security Income (SSI)-related information. SSI benefit changes due to policy options are included as part of the household income calculation. However, the SSI projections are less robust than the Social Security benefits simply because SSI is more complicated and only part of the SSI rules are included in MINT.

  • SSI State Supplements: Since SIPP is a nationally representative survey, the state information does not support estimating individual state supplement amounts.

Notes, Caveats, and Limitations

The benefit amounts shown are pre-tax, and no Medicare premiums are subtracted from the benefits.

Child benefits: MINT does not know which parent, if either, the child is living with, if the parents are not married when the child is eligible for child benefits. However, MINT assumes a sole parent if unmarried at the time the child is born, or the mother if the parents divorce after the child is born.

The Windfall Elimination Provision (WEP) and Government Pension Offset (GPO) are rules that reduce Social Security benefits when there is a pension from noncovered employment. MINT does not have the value of those noncovered pensions. Instead, we impute the value of the noncovered pension based on what proportion of the individual's career earnings came from noncovered jobs.

A small percentage of beneficiaries will receive a benefit reduction under a policy option intended to increase benefits and vice versa. These counterintuitive results happen because of interactions with program rules such as the retirement earnings test and dual-entitlement.