Automatic Retirement Savings Plans for Low-Income Households
Automatic Retirement Savings Plans for Low-Income Households
We analyze a new illustrative policy to create automatic retirement savings accounts for more than 56 million low-income Americans by 2030. The program is fully financed by removing the gross income adjustment for traditional 401k and similar retirement accounts without any additional contribution from households or employers. The program relies on the existing EITC administration without employer participation. After accounting for risk, individual account balances reach over $200,000 by retirement and any balances can be bequeathed upon death.
Key Points
-
Past efforts to boost retirement savings of low-income households have not led to material account balances for households upon retirement.
-
At the same time, we project that the federal government will spend about $1.25 trillion over the next decade to subsidize retirement savings in 401k and similar retirement plans that mostly benefit higher-income households.
-
Shifting this expenditure to lower-income households would have little impact on retirement savings by higher income households but would create account balances over $200,000 at retirement age for low-income households after adjusting for risk.
A rapidly aging population combined with public policies, including the Pension Protection Act of 2006, have appeared to have little impact on retirement savings for lower-income Americans. Between 1989 and 2022, households with incomes below the 80th percentile saw the average value of their retirement accounts grow slower than the value of the stock market itself, even without reinvestment of dividends.1 Retirement accounts held by households in the bottom half of the income distribution fell relative to the top half of the income distribution over the same period.2 Consistently, a leading team of researchers---whose previous work played a major role in encouraging automatic enrollment in 401(k) plans---has found that auto-enrollment has produced smaller increases in retirement savings than previously predicted.3
The level of retirement savings by low-income households could be sufficient to meet personal retirement expenses if the nation’s major old-age entitlement programs, including Social Security and Medicare, continue to pay benefits as scheduled. However, these programs also face large financial shortfalls. Put differently, boosting low-income retirement accounts could also place less pressure on the nation’s entitlement programs over time. Larger personal savings accounts could also enable low-income households to leave more assets to their heirs beyond housing assets. Greater inter-generational monetary transfers could also help reduce liquidity constraints of heirs, including for funding future education.
The illustrative retirement plan is assumed to begin in 2025. It has the following design and qualifications:
The federal government creates personal individual investment accounts at a custodian.4 Individual eligibility is based on the existing Earned Income Tax Credit criteria for administrative efficiency and some adjustments specific to retirement accounts. The individual must be between ages 25 and 64, cannot be listed as “a dependent” on another tax filer’s return, and must have investment income below $10,000 (in 2020 dollars, indexed over time using the chain CPI).5
All contributions are made by the federal government---without any contribution by the individual beneficiaries or employers. Annual federal contributions are proportional to the individual’s earned income, subject to cap and phase-out based on AGI. We present analysis for three different variations that differ only in the maximum dollar contribution per year.
-
Small: 10% contribution rate of earned income; annual maximum contribution of $2,000; contributions are phased-out at a rate of 30 percent starting at $50,000 of earned income. All dollar values indexed to Chained CPI. In 2030, we project that about 56 million individual accounts would receive deposits under this variation.
-
Medium: 10% contribution rate of earned income; annual maximum contribution of $2,250; contributions are phased-out at a rate of 30 percent starting at $50,000 of earned income. All dollar values indexed to Chained CPI. In 2030, we project that about 57 million individual accounts would receive deposits under this variation.
-
Big: 10% contribution rate of earned income; annual maximum contribution of $2,500; contributions are phased-out at a rate of 30 percent starting at $50,000 of earned income. All dollar values indexed to Chained CPI. In 2030, we project that about 58 million individual accounts would receive deposits under this variation.
Figure 1 shows that level of annual dollar contributions by earned income in 2025 for each variation:
Asset values accrue over time with investment returns and interest payments. Contributions, growth, and withdrawals are tax-exempt, much like a Health Savings Accounts but without the requirement to use the assets for health expenditures in retirement. Assets can be withdrawn by an individual upon reaching age 65. Like standard investment accounts, asset balances are fully and immediately bequeathable upon death to selected beneficiaries, whether death occurs prior or after retirement.
Importantly, as none of the contributions are made by individuals or their employers, all contributions and accrued interest are inaccessible until either reaching the minimum retirement age of 65 or upon death. Allowing for early access, including “hardship withdrawals,” would almost certainly unwind most of the accumulation in these accounts. Early withdrawals are especially likely to occur in accounts that are mostly intended to help low-income households build wealth.6
As the new retirement accounts work through the existing EITC tax administration, the associated costs are represented as negative revenue. Table 1 presents the 10-year revenue costs for the Small, Medium, and Big plan variations. Notice that the 10-year revenue losses range from about $1.1 trillion to $1.4 trillion.
| Policy | Revenue estimate (2025-2034) |
|---|---|
| Repeal retirement account exclusion | 1,243.5 |
| FRP Small | -1,138.3 |
| FRP Medium | -1,262.6 |
| FRP Big | -1,380.6 |
Table 1 also presents our 10-year revenue estimate of eliminating the above-the-line deduction for employee contributions to employer-based retirement account contributions that exist under current law. This exclusion currently allows traditional 401 (k), 403 (b), and similar employer-based plans to receive favorable tax treatment by making contributions to retirement accounts on a tax preferred basis.7 We estimate that disallowing this deduction would produce a revenue increase of about $1.2 trillion over the 10-year budget window. This revenue gain is almost the same cost as the Medium plan option. The revenue gain does not include potential additional revenue that could be achieved by eliminating tax preferences for Individual Retirement Accounts (IRA). We anticipate that some shifting of saving toward IRA’s could occur depending on rule design.
Figures 2-10 show potential account balances based on average annual wage earnings and the selected account Plan option. The nominal (before-inflation) annual investment return is set at 3 percent, reflecting a risk-adjusted return without expected returns to risky equities. The Appendix presents some variation around this investment return.
For each figure, the green bar indicates the annual contribution made by the government for the individual, given the individual’s earned income and the phase-out schedule. The blue line shows the total cumulative account balance including interest earned, before any withdrawals. A gray region represents a simulated year with no earned income due to voluntary or involuntary unemployment.8 A year with white (no gray or green) indicates a year where income is large enough that no contribution is made (fully phased out). The red line represents the year in which the individual turns 65. We do not include withdrawals after age 65, so in practice account balances (blue lines) could look different with withdrawals. Instead, the purpose of the blue line is to show maximum wealth accumulation if withdrawals were posted to the shown year or bequeathed upon death.
We consider three types of hypothetical savers:
-
Person A, who is 24 years old in 2025, has the lowest average annual income considered ($12,700) and receives a modest deposit almost every year.9 Over their working age years, this account balance grows to about $125,000 by age 65, depending on scenario.
-
Person B, 26 years old in 2025, is also low income but typically has more earned income ($27,900) than Person A. They receive larger deposits every year (the scenario maximum in many years). They experience a brief spell of unemployment in their late 50’s/early 60’s, therefore receiving no deposits, but the balance continues to grow even in those years due to interest. Over their working years, the account balance grows to between $150,000 and $200,000 by age 65, scenario dependent.
-
Person C, 24 years old in 2025, has the highest average annual earnings among the individuals we consider ($59,300). Early in their career, when their annual earnings are lower, they receive deposits in their account, but as they enter prime working age years, their income grows such that they are no longer eligible for annual deposits. Their account balance continues to grow during this time with interest, and in their late career as their annual earnings stagnate/decline, they receive additional deposits. This results in a cumulative balance between $100,000 and $125,000 by age 65.
Note: Gray regions indicate years with no earned income.
Note: Gray regions indicate years with no earned income.
Note: Gray regions indicate years with no earned income.
Note: Gray regions indicate years with no earned income.
Note: Gray regions indicate years with no earned income.
Note: Gray regions indicate years with no earned income.
Note: Gray regions indicate years with no earned income.
Note: Gray regions indicate years with no earned income.
Note: Gray regions indicate years with no earned income.
Table 2 reports the projected percent change in after-tax income for repealing the retirement account deduction (the pay-for) along with each policy option (Small, Medium and Big). As shown in previous analysis, the after-tax income gain for the current retirement account deduction mostly accrues to higher-income households. Those in the top 0.1 percent gain a bit less due to contribution limits under current law. In contrast, the three policy options (Small, Medium and Big) shift gains to lower-income households.
| Percent change in after-tax income, including FRP deposits (2030) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Income Quantile | |||||||||
| Policy | First quintile | Second quintile | Middle quintile | Fourth quintile | 80-90% | 90-95% | 95-99% | 99-99.9% | Top 0.1% |
| Repeal retirement account deduction | 0.0% | -0.2% | -0.4% | -0.7% | -1.3% | -1.7% | -1.9% | -1.0% | -0.2% |
| FRP Small | 4.9% | 5.2% | 2.4% | 0.8% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% |
| FRP Medium | 4.9% | 5.6% | 2.7% | 0.9% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% |
| FRP Big | 5.0% | 6.0% | 3.0% | 1.1% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% |
Table 3 reports the projected after-tax income that combines repealing the current-law deduction while adding the shown policy option. For example, Small includes the incidence associated with repealing the current retirement account deduction along with implementing the new Small illustrative plan. Since current law mostly benefits higher-income households while the new plans mostly benefit lower-income households, the combined incidence for higher-income households mostly stems from the loss of the current retirement account deduction while the combined incidence for lower-income households mostly stems from the gain from the new plan.
| Percent change in after-tax income, including FRP deposits (2030) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Income Quantile | |||||||||
| Policy bundle | First quintile | Second quintile | Middle quintile | Fourth quintile | 80-90% | 90-95% | 95-99% | 99-99.9% | Top 0.1% |
| Small | 4.9% | 5.0% | 2.0% | 0.0% | -1.3% | -1.7% | -1.9% | -1.0% | -0.2% |
| Medium | 4.9% | 5.5% | 2.3% | 0.2% | -1.3% | -1.7% | -1.9% | -1.0% | -0.2% |
| Big | 5.0% | 5.8% | 2.6% | 0.3% | -1.3% | -1.7% | -1.9% | -1.0% | -0.2% |
Table 4 reports the projected share of tax units receiving deposits and the average size of those deposits. Within each of the first four quintiles, as many as 76% of tax units receive a deposit in 2030, with average deposit amounts ranging from $1065 to $3930 depending on scenario and quantile. In contrast, tax units in the top quantile receive little.
| Share of tax units receiving FRP deposits and average size of FRP deposits by unit (2030) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Income Quantile | ||||||||||
| Policy | First quintile | Second quintile | Middle quintile | Fourth quintile | 80-90% | 90-95% | 95-99% | 99-99.9% | Top 0.1% | |
| FRP Small | Share receiving deposit | 56% | 76% | 46% | 23% | 0% | 0% | 0% | 0% | 0% |
| Average deposit amount | $1,065 | $2,375 | $3,215 | $3,270 | - | - | - | - | - | |
| FRP Medium | Share receiving deposit | 56% | 76% | 47% | 24% | 0% | 0% | 0% | 0% | 0% |
| Average deposit amount | $1,065 | $2,575 | $3,525 | $3,605 | - | - | - | - | - | |
| FRP Big | Share receiving deposit | 56% | 76% | 48% | 25% | 0% | 0% | 0% | 0% | 0% |
| Average deposit amount | $1,070 | $2,735 | $3,815 | $3,930 | - | - | - | - | - | |
Table 5 reports annual conventional revenue estimates of each new policy option (Small, Medium, Large) by year.
| Provision | 2025 | 2026 | 2027 | 2028 | 2029 | 2030 | 2031 | 2032 | 2033 | 2034 | Total, 2025-2034 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| FRP Small | 0.0 | -124.4 | -124.2 | -125.9 | -125.4 | -126.4 | -127.0 | -127.6 | -128.0 | -129.4 | -1,138.3 |
| FRP Medium | 0.0 | -137.8 | -137.6 | -139.6 | -139.1 | -140.3 | -140.9 | -141.6 | -142.0 | -143.7 | -1,262.6 |
| FRP Big | 0.0 | -150.5 | -150.4 | -152.5 | -152.1 | -153.4 | -154.1 | -154.9 | -155.4 | -157.3 | -1,380.6 |
Table 6 reports annual conventional revenue estimates for repealing the current-law retirement deduction.
| Provision | 2025 | 2026 | 2027 | 2028 | 2029 | 2030 | 2031 | 2032 | 2033 | 2034 | Total, 2025-2034 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Repeal retirement account exclusion | 20.3 | 104.1 | 123.2 | 126.8 | 132.2 | 137.2 | 142.2 | 147.3 | 152.7 | 157.5 | 1,243.5 |
These estimates do not include macroeconomic feedback effects associated with changes in the capital stock.
For Person C, the following figure shows how account balances vary by the interest rate assumption. The case of 3 percent return is the same as provided already in the main text, and it is provided here for comparison.
Note: Gray regions indicate years with no earned income.
Note: Gray regions indicate years with no earned income.
Note: Gray regions indicate years with no earned income.
This analysis was conducted by Brendan Novak under the direction of Kent Smetters who helped write this brief. Mariko Paulson prepared the brief for the website.
-
Even a less risky 60/40 stock-bond portfolio grew faster than the size of retirement accounts for most American households. https://www.federalreserve.gov/econres/scf/dataviz/scf/table/#series:Retirement_Accounts;demographic:inccat;population:3;units:median;range:1989,2022 ↩
-
Even if measured as total wealth, wealth held by the bottom half of the income distribution fell relative to the top half of the income distribution over the same period. https://www.federalreserve.gov/econres/notes/feds-notes/wealth-and-income-concentration-in-the-scf-20200928.html ↩
-
Choi, James J., David Laibson, Jordan Cammarota, Richard Lombardo, and John Beshears. Smaller than We Thought? The Effect of Automatic Savings Policies. No. w32828. National Bureau of Economic Research, 2024. ↩
-
For our purposes, we do not take a position whether these accounts are established at an integrated broker-custodian such at the Thrift Savings Plan or bid out to a private sector provider. Like TSP, we assume that the additional costs are minimal. ↩
-
This limit comes from the EITC rules which is statutorily set to $10,000 in 2020 dollars. We project that to be about $11,800 in 2025. ↩
-
PWBM is currently reviewing the related literature on early withdrawals and will issue an explainer later. ↩
-
Generally, traditional 401 (k) and similar plans allow for pre-tax contributions along with taxes paid on future withdrawals; Roth plans allow for post-tax contributions along with no taxes paid on future withdrawals. Each type of plan is subject to specific rules and qualifications. ↩
-
For any microsimulation, the exact year of unemployment will change. The shown results are representative in age-adjusted frequency. ↩
-
Average income during working years in nominal wages. At ages below 25, no contribution is made. ↩