The data did not load. Reload the page, or choose the data file (dividend_data.parquet).
Greenlight Starter Homes unlocks housing mobility through Move-up, Free-up: a compressed and highly efficient chain of new supply and vacancies.
A companion to the Starter Home Policy Explorer.
The explorer shows how many additional naturally occurring affordable homes (NOAH) are provided when new homes are built, both directly from new starter homes that buyers move into and from the NOAH these buyers free up when they move. This page explains the methodology for each tab and defines terms in the glossary. The numbers provided for illustration on this page reflect the United States.
The headline boxes provide an overview of the Move-up, Free-up sequence:
| Box Title | Description |
|---|---|
| Greenlight | The additional new homes in new residential subdivisions per year under Greenlight Starter Homes (Starter Home Policy Explorer) |
| Move-up | The additional new homes that are starter homes (at or below four times county median household income) |
| Free-up | The naturally occurring affordable homes (NOAH) unlocked when buyers of the additional new homes under Greenlight Starter Homes (starter-priced and higher-priced) move out of their previous homes. These consist of freed-up starter homes and freed-up rentals at or below county FMR. |
| Additional NOAH | Move-up + Free-up |
Tab 1 shows how new homes unlock naturally occurring affordable homes. Using the dropdown, users can select from the four charts below to visualize how building more starter- and higher-priced homes unlocks naturally occurring affordable homes.
Step 1. Identify 2024 new starter homebuyers (ICE, ACS)
Data: ICE 2024 first-lien purchase mortgages for new homes (reported by the lender as less than a year old). We compare the purchase price to the affordability baseline for owned homes (at or below four times American Community Survey (ACS) county median household income) to identify new starter homes.
Exclusions: Homes over $5 million, new homes under $100,000, secondary and investor purchases, co-ops, business-purpose loans, rehabbed homes, and homes not matched to a county with an ACS median income.
Result: 93,581 buyers of new starter homes.
Limitations: ICE only covers purchases with a mortgage, so cash buyers are not included.
Step 2. Identify the share of 2024 new starter home buyers who previously owned, rented, or freed no home (ICE)
Data: ICE identifies previous owners as buyers that the lender reported as a current homeowner. Of those who did not previously own a home, we assume that 75% rented and 25% did not free up a home.
Exclusions: Same as Step 1.
Result: Of 93,581 buyers, 28% previously owned and 72% did not. We therefore estimate that 54% previously rented and 18% did not free a home.
Limitations: We cannot directly observe the share of non-homeowners that rented or did not free a home from the ICE data. The assumption that 25% did not free a home is consistent with the National Association of Realtors’ estimate that 24% of first-time buyers lived with their parents, relatives, or friends (2024 Profile of Home Buyers and Sellers, Exhibit 1–20, Prior Living Arrangement of First-Time Buyers, 1989–2024).
Step 3. Identify the share of previous owners who sold a starter home vs. a higher-priced home (Arbitrage Index, ACS)
Data: The AEI Housing Center Arbitrage Index matches 2023 and 2024 purchases to the sale of the buyer's previous home through First American Melissa USPS Change-of-Address records. New homes are defined as those built in the purchase year or the year before. We determine whether a home is a starter home by comparing the sale price to the county median household income (ACS) in the purchase year. A buyer is considered to have bought and sold a starter home if both their new and previous home sold for no more than four times the median household income in their respective counties.
Exclusions: New homes under $100,000, homes over $5 million, homes built after the purchase year, and previous homes sold more than 18 months before or after the purchase.
Result: 45% of 13,075 new starter home buyers sold a starter home: 28 owners × 45% = 13 Freed Starter Homes per 100 new starter homes built.
Limitations: We assume the buyers of new starter homes in the Arbitrage Index are representative of buyers of new starter homes in the ICE data who previously owned a home.
Step 4. Identify the share of previous renters who rented at or below FMR vs. above FMR (ICE, HUD)
Data: ICE provides the monthly rent that buyers paid before purchasing, as reported by the lender. We compare this monthly rent to the affordability baseline for rentals (HUD FY 2024 two-bedroom Fair Market Rent (FMR) for the county in which they purchase their new home).
Exclusions: Rents under $200 or over $5,000 and counties with no FMR.
Result: 49% of 31,800 renters with rents reported: 54 renters × 49% = 26 Freed Rentals at or below FMR per 100 new starter homes built.
Limitations: ICE does not provide the bedroom count or the location of the previous rental. We therefore compare the rental to the two-bedroom FMR for the county in which they purchase their new home. We also assume that renters without rents reported in ICE are representative of renters with rents reported.
Step 5. Identify the share of previous renters who rented a unit in a 1-4 unit building vs. a 5+ unit building (American Housing Survey or AHS)
Data: AHS 2015-2023 surveys identify the previous home type and tenure of buyers of single-family homes, as well as the owner’s estimate of their new home’s value. We define a home as new if the buyer moved in within a year of when the home was built. The surveys report whether a starter-home buyer previously rented a home, mobile/manufactured home, or apartment, but not whether the apartment building had between 2–4 units or 5+ units. We therefore allocate the share of 2+ unit buildings that have 2–4 units as identified in the AHS rental stock (27.7%) to the 1–4 unit building category, which also includes homes and mobile/manufactured homes.
Exclusions: Buyers who received their homes as gifts or inheritance, buyers of non-single-family homes, buyers who moved in more than a year before or after construction, and buyers with no reported previous home type or tenure.
Result: Of new starter homebuyers who previously rented and were surveyed by the AHS, we estimate that 69% rented in 1–4-unit buildings and 31% rented in 5+ unit buildings. Applied to the 26 Freed Rentals at or below FMR per 100 new starter homes, we estimate that 18 are in 1–4- and 8 are in 5+ unit buildings.
Limitations: One national share is applied for every state and for all rentals. Starter homes are identified using the area median household income of the metropolitan area or Census division.
Step 6. Calculate the Total NOAH Unlocked per 100 Starter Homes
Result: 13 Freed Starter Homes + 26 Freed Rentals at or below FMR = 39 NOAH unlocked per 100 new starter homes.
Steps 1 to 6 for new higher-priced homes: 309,176 ICE buyers, 110,730 Arbitrage Index movers, 76,765 ICE renters, 529 AHS renters.
Result: 11 Freed Starter Homes + 12 Freed Rentals at or below FMR = 23 NOAH unlocked per 100 new higher-priced homes.
Step 1. Identify the Total New Homes Built, Starter Homes Built, and Higher-Priced Homes Built under Greenlight Starter Homes (Starter Home Policy Explorer)
Data: The Starter Home Policy Explorer calculates the actual number of homes built in new residential subdivisions over 2015-2024 (Chart 3) and the additional homes that would be built under Greenlight Starter Homes (Chart 4).
Result: 237,000 additional new starter homes and 151,000 additional new higher-priced homes per year under Greenlight Starter Homes (Chart 4).
Step 2. Apply the per-100 shares (charts 1 and 2) to calculate the Additional NOAH under Greenlight Starter Homes
Starter homes: 237,000 additional Starter Homes × 39 NOAH Unlocked per 100 = 92,100 NOAH Unlocked (30,200 Freed Starter Homes + 61,900 Freed Rentals).
Higher-priced homes: 151,000 additional Higher-Priced Homes × 23 NOAH Unlocked per 100 = 34,800 NOAH Unlocked (16,500 Freed Starter Homes + 18,300 Freed Rentals).
Result: 237,000 additional Starter Homes + 127,000 NOAH Unlocked = 364,000 Additional NOAH per Year under Greenlight Starter Homes.
Limitations: We assume that buyers of homes in new residential subdivisions over 2015–2024 are representative of recent buyers of new homes in ICE (2024) and in the Arbitrage Index (2023–2024).
Chart 3 repeats the steps for actual new homes, rather than for additional new homes under Greenlight Starter Homes.
Tab 2 shows who buys the new homes. Users can view the buyers of the chart's new starter- and higher-priced homes by one characteristic: Income, First-time buyers, Race and ethnicity, or Children.
Step 1. Identify buyers of new homes (ICE)
Data: ICE 2024 first-lien purchase mortgages for new homes (starter- and higher-priced).
Exclusions: Same as Tab 1, Step 1.
Step 2. Break down buyers of new homes by each characteristic
| Category | Description | Source | Number of buyers reporting (new starter / new higher-priced) |
|---|---|---|---|
| Income | Household income in dollars, shown in HUD income groups (not adjusted for household size) | ICE | 92,515 / 306,292 |
| First-time buyers | Whether the mortgage identifies the buyer as a first-time homebuyer | ICE | 93,581 / 309,176 |
| Race and ethnicity | The first applicant's race and ethnicity, as reported on the mortgage application | HMDA | 13,575 / 66,092 |
| Children | Dependents as reported on the mortgage (3+ count as 3) | ICE | 78,314 / 266,829 |
Limitations: We analyze data from ICE and the Home Mortgage Disclosure Act (HMDA) to determine buyer race and ethnicity. We assume our analysis is representative of all buyers.
Step 3. Scale buyer shares to chart totals
Buyer shares by characteristic are scaled to per 100 new starter homes (Chart 1), per 100 new higher-priced homes (Chart 2), per 460,000 actual new homes built (Chart 3), and per 388,000 additional new homes built under Greenlight Starter Homes (Chart 4).
Limitations: For every group, we assume that 75% of buyers who did not previously own rented and 25% did not free up a home, usually because they lived with parents or other close relatives.
Tab 3 shows the price distributions of the homes that buyers free up. The first histogram shows the price of the homes that previous owners sold, while the second histogram shows the rent previous renters paid. The green bars represent homes priced at or below the affordability baseline, while the gray bars signify those above the affordability baseline. Using the “Show” dropdown, users can instead view the household income of the buyers who free up these homes, colored by whether the home they freed up was at or below the affordability baseline.
Step 1. Identify the price of the homes that previous owners sold (Arbitrage Index, ACS)
Data: The Arbitrage Index identifies the sale price of the previous home of 2023 and 2024 buyers of new homes, which we compare to the affordability baseline for owned homes (four times ACS county median household income in the purchase year).
Exclusions: Same as Tab 1, Step 3.
Result: The histogram shows the sale prices of the previous homes of 13,075 buyers of new starter homes and 110,730 buyers of new higher-priced homes.
Limitations: Same as Tab 1, Step 3.
Step 2. Identify the rent that previous renters paid (ICE, HUD)
Data: ICE provides the monthly rent that buyers paid before purchasing, as reported by the lender, which we compare to the affordability baseline for rentals (HUD FY 2024 two-bedroom FMR for the county of their new home).
Exclusions: Same as Tab 1, Step 4.
Result: The histogram shows the rents of 31,800 renters who bought new starter homes and 76,765 who bought new higher-priced homes.
Limitations: Same as Tab 1, Step 4.
Step 3. Identify the household income of previous owners and renters (ICE, Arbitrage Index)
Data: For previous renters, household income is stated on the mortgage for the new home (ICE). For previous owners, we analyze data from ICE and the Arbitrage Index. The green bars show buyers who sold a starter home or rented at or below FMR.
Exclusions: Same as Steps 1 and 2.
Result: We identified 3,101 of 13,075 (24%) previous owners who bought new starter homes and 30,325 of 110,730 (27%) previous owners who bought new higher-priced homes.
Limitations: We assume the incomes of previous owners in the Arbitrage Index are representative of the incomes of all previous owners.
Tab 4 shows who moves into the Additional NOAH. This includes those who move into the Unlocked NOAH – the next owners of Freed Starter Homes and the next renters of Freed Rentals at or below FMR – and the buyers of the new starter homes. Users can view these households by income, race and ethnicity, or children.
Step 1. Identify the next owners of Freed Starter Homes (Arbitrage Index, HMDA, ICE)
Data: We analyze data from the Arbitrage Index (2023-2024), ICE, and HMDA.
Exclusions: Same as Tab 1, Step 3.
Result: The chart shows the next owners of 29,125 Freed Starter Homes: 5,870 sold by buyers of new starter homes and 23,255 by buyers of new higher-priced homes.
Limitations: Same as Tab 1, Step 3.
Step 2. Break down next owners by each characteristic
| Category | Description | Source | Number of next owners reporting |
|---|---|---|---|
| Income | Household income in dollars, shown in HUD income groups for the freed home's county (in the purchase year) | HMDA | 11,840 |
| Race and ethnicity | The first applicant's race and ethnicity, as reported on the mortgage application | HMDA | 10,463 |
| Children | Dependents reported on the next owner's mortgage | ICE | 7,082 |
Limitations: We assume the next owners that match are representative of all next owners.
Step 3. Identify the next renters of Freed Rentals at or below FMR (ICE, ACS)
Data: As there is no direct source showing who moves into a freed rental, we weight renters in the ACS to align their county composition using $100 rent bands derived from the freed rentals in ICE. Each ACS renter keeps its household weight, scaled within its county and rent band. We split Census areas (PUMAs) that span several counties across those counties by population.
Exclusions: ACS renters who moved in more than two years before the survey, paid under $200 or over FMR, or have no gross rent.
Result: 317,280 ACS renters are used to represent the next renters of 15,493 Freed Rentals at or below FMR from buyers of new starter homes in the ICE data and 25,734 from buyers of new higher-priced homes in ICE. In each county and $100 rent band, the weights of the ACS renters sum to the number of Freed Rentals in ICE. The characteristics of the next renters therefore reflect the location of the Freed Rentals and their rent price. 98% of Freed Rentals at or below FMR from buyers of new starter homes in ICE correspond with at least one ACS renter in the same county and rent band.
Limitations: We assume next renters are representative of recent ACS renters in the same county and rent band. We also assume each freed rental is in the county of the new home. Freed rentals with no ACS renter in the same county and rent band are excluded.
Step 4. Break down next renters by each characteristic
| Category | Description | Source | Number of next renters reporting |
|---|---|---|---|
| Income | Household income in dollars, shown in HUD income groups | ACS | 317,280 |
| Race and ethnicity | The head of household's race and ethnicity | ACS | 317,280 |
| Children | People under 18 in the household | ACS | 317,280 |
| Rent affordable for (rentals) | The lowest HUD income group that can pay rent plus utilities at 30% of income | ACS | 317,280 |
Step 5. Scale next owner and next renter shares to chart totals
Next owner and next renter shares by characteristic are scaled to the Unlocked NOAH per 100 new starter homes (Chart 1) and per 100 new higher-priced homes (Chart 2). Charts 3 and 4, which show the Additional NOAH from new starter homes and Unlocked NOAH, combine the characteristics of the buyers of new starter homes and the next owners of the Freed Starter Homes and next renters of the Freed Rentals at or below FMR.
A scatter plot compares each state's starter home share of additional new homes with its Additional NOAH per 100 additional new homes under Greenlight Starter Homes. By default, 17 key states are shown (the 9 with the most starter home potential and the rest of the 13 most populous states), while all states can be selected.
Below it, the map colors each state by one measure, with a ranking and table. States fall into five color classes, chosen so that states in a class are as close in value as possible. Gray states are not shown. The “Show” dropdown includes 75 statistics from the rest of the explorer by state, grouped by tab.
A state is flagged as a "Small sample" (†) if it has fewer than:
300 ICE 2024 mortgages for new starter homes;
100 Arbitrage Index buyers of new starter homes who sold a previous home; or
100 Freed Rentals at or below FMR from buyers of new starter homes.
| Source | Description | Used for |
|---|---|---|
| Starter Home Policy Explorer | AEI Housing Center estimates of how many new homes were built in new residential subdivisions over 2015–2024, and how many additional new homes would have been built under Greenlight Starter Homes | - Actual new homes built per year - Additional new homes built per year under Greenlight Starter Homes |
| ICE | ICE Mortgage Technology's 2024 mortgage records | - The price that 2024 buyers of new homes (starter- and higher-priced) paid. - Whether buyers of new homes previously owned, and the rent that previous renters paid. - The income, first-time buyer status, and dependents for buyers of new homes. - Whether the next owners of Freed Starter Homes have dependents. |
| Arbitrage Index | AEI Housing Center analysis of home moves, using data from First American Melissa USPS Change-of-Address records | - Whether buyers of new homes sold a starter home or a higher-priced home. - Who the next owners of Freed Starter Homes are. |
| HMDA (Home Mortgage Disclosure Act) | Mortgage applications | - The race and ethnicity of buyers of new homes. - The race, ethnicity, and income of the next owners of Freed Starter Homes. |
| AHS (American Housing Survey) | Census Bureau survey, 2015–2023 (biennial) | - The share of previous rentals in 1–4 unit vs. 5+ unit buildings |
| ACS (American Community Survey) | 2020–2024 households; 5-year county median household incomes (2024 for 2024 moves, and 2023 for 2023 moves) | - County median household income - The income, race and ethnicity, and dependents of the next renters of Freed Rentals at or below FMR. |
| HUD Fair Market Rents | Two-bedroom FMR by county, FY 2024 | Rentals at or below FMR |
Move-up, Free-up: Buyers move into new homes (Move-up) and free up the homes they leave (Free-up).
NOAH: Naturally occurring affordable homes (starter homes and rentals at or below FMR).
Unlocked NOAH: The NOAH that buyers of new homes free when they move.
Additional NOAH: New Starter Homes Built plus the Unlocked NOAH.
No Home Freed: The 25% of buyers who did not previously own a home who are assumed to have freed no home.
Affordability baseline: For owned homes, at or below four times the county median household income. For rentals, at or below the 2-bedroom county FMR.
Starter home: Homes priced at or below the affordability baseline – four times its county's median household income (ACS 5-year: 2024 for ICE buyers, the purchase year for Arbitrage Index moves). A higher-priced home is above it.
FMR: HUD's FY 2024 two-bedroom Fair Market Rent for the county.
Freed Starter Homes: Homes the buyers sold at or below four times county median household income.
Freed Rentals at or below FMR: Rentals the buyers moved out of, at or below two-bedroom county FMR.
Greenlight Starter Homes: A projection in which states adopt a 1,000 sq. ft. maximum minimum lot size for new residential subdivisions.
HUD income groups: Household income as a percent of county median household income: very low-income under 50%, low-income 50-80%, moderate-income 80-120%, above moderate-income over 120%.
Next owners, next renters: The households that move into Freed Starter Homes and Freed Rentals at or below FMR.
Geographies: The US and 46 states, plus the District of Columbia. Alaska, Connecticut, Hawaii and Vermont are not shown, as these states have no buyers of new starter homes in the ICE data (2024) or in the Arbitrage Index (2023–2024).