The Houses That Are There and the Homes That Are Missing

On page twelve of the federal housing analysis for Puerto Rico, two sets of figures sit close together. Population falls. Households rise.

The U.S. Department of Housing and Urban Development estimated 3,183,000 people and 1,376,900 households as of July 1, 2025. The corresponding 2020 census counts were 3,285,874 and 1,340,534. In round terms, about a hundred thousand fewer people and thirty-six thousand more households. HUD links this split to a fall in the average number of people per household.[1]

If you have been imagining a housing market as a simple queue of people, this is a useful place to stop.

A home shelters a household. One person living alone occupies one home. Four people living together also occupy one home. If the four-person household becomes two two-person households, the population remains four and the demand for separate dwellings becomes two. If one of those people later leaves the island, the population falls to three while the two households may remain. Nothing mysterious has happened. We were counting people when the immediate question concerned arrangements of people.

Lives rearrange themselves. Children leave home. Couples separate. A spouse dies. Relatives who shared a dwelling find that they can no longer do so. An older person may continue living in the same house after everyone else has gone. Some changes are chosen; others are endured. A population total records the number of people at a place and time. It does not tell us how many front doors those people need.

So the claim “People are leaving; we must have plenty of housing” skips a step. There may be plenty of certain kinds of housing in certain places. There may also be a growing need for dwellings suited to smaller households, near services, at prices those households can pay. A three-bedroom house far from transport is not automatically a solution for an older person who lives alone and no longer drives. The bedrooms exist. Their existence has not completed the journey to a usable option.

The Census Bureau counted an 11.8 percent decline in Puerto Rico’s population between 2010 and 2020.[2] That substantial change describes a decade’s population outcome. Inside it are births and deaths as well as arrivals and departures. Inside the departures are many different decisions and pressures. Converting the whole decline into a count of people driven out by one industry would erase most of the inquiry before it had begun.

At the household level, the housing problem has to be assembled again from the ground up. Who needs a dwelling? Where must it be? What condition is it in? On what terms can they occupy it? And what income, savings, support, or credit can they bring to the arrangement?

The empty house across the road

You can see the roof from your window. No one appears to live there. Meanwhile you are looking for somewhere to live and finding nothing you can afford. Of course it makes you angry. It also gives you a useful question to investigate.

An empty dwelling can be waiting for a buyer. It can be between tenants, reserved for occasional use, caught in an estate, or withheld for an owner’s personal reasons. Census vacancy classifications separate several of these conditions.[3] Walking past the building gives you evidence of an apparently empty building. The deed, its condition, the owner’s intentions, and the costs of making it available remain to be discovered.

In its 2025 analysis, HUD estimated a one percent vacancy rate in Puerto Rico’s sales market and a seven percent rental vacancy rate. Separately, it reported roughly 208,500 dwellings in the Census residual “other vacant” category for 2023.[1] The rates and the count describe different groups of dwellings. They cannot be added or treated as one measure of homes ready for someone to move into.

Take five buildings on a street. The first is empty and ready to rent. The second has a sound exterior but a dangerous electrical system. The third belongs to an estate whose heirs disagree. The fourth is used by its owner during part of the year. The fifth is ready for occupancy and offered at a price the searching household cannot pay. From outside, on the right afternoon, all five can look equally empty. For the household, they present five different problems.

The first may be a straightforward match. The second requires repairs and someone willing to finance them. The third requires decisions and legal work. The fourth requires a change in the owner’s intended use. The fifth presents a gap between price and means. A policy that pays for rewiring will not resolve the heirs’ disagreement. A measure aimed at speculative withholding may miss a family unable to afford essential repairs. A rent subsidy can help with payment while leaving an unsafe electrical system unsafe.

One explanation will not serve the whole street. Someone has to inspect the wiring, someone has to settle the estate, and someone has to work out what the household can pay. “There are empty houses” gets the inquiry started. It cannot do those jobs.

Puerto Rico’s housing recovery programs include a Title Clearance Program intended to help resolve ownership documentation problems.[4] The existence of such a program points to a feature of housing supply that is easy to miss from a distance. A building can physically exist while the documents needed to improve, finance, or transfer it remain unresolved. The apparent shortage may contain work for surveyors, lawyers, administrators, and lenders as well as builders.

That work can be slow and unphotogenic. A completed roof makes a good picture. An estate resolved after years of uncertainty is harder to photograph. Yet both can matter to whether a family has somewhere safe to live. Public discussion tends to give construction the starring role because its output is visible. Bringing an existing dwelling into use may depend on a sequence of less visible achievements.

Count every dwelling, then ask which ones this household could live in. Remove those that are unsafe, too far from work or care, unavailable, or too costly. The list can shrink very quickly. The dwellings left out may be potential supply, but only if someone can resolve the barrier that keeps each one out.

The rent you never paid

Another figure in the HUD report opens a less familiar part of the picture. In 2023, about thirty-five percent of Puerto Rico’s renter households paid no cash rent.[1]

That category asks us to think beyond the monthly transfer normally pictured when someone says “renter.” A person may occupy a home owned by relatives. An arrangement may involve care or work. A family may permit someone to remain without requiring cash each month. The specific terms vary. The dwelling is being made available through a relationship as well as through an asset.

Imagine an adult daughter living in a small house owned by her parents. She contributes to family care and pays for some maintenance. When asked for her monthly rent, she says zero. It would be absurd to conclude that her housing costs nothing. Someone bought or built the house; someone bears repairs and other obligations; the family has chosen to forgo rent it might otherwise collect. The daughter receives real help because someone else carries part of the cost.

Now suppose ownership changes after a death. The new arrangement requires a market rent. From the daughter’s perspective, the monthly housing price has moved from no cash rent to a substantial cash payment. A published rent index may show a moderate increase over the same period. Both accounts can be accurate. She has moved from family support into the cash-rent market. That change matters far more to her budget than the average increase faced by people who were paying rent all along.

This helps explain why a community can become more vulnerable before every advertised rent has become spectacular. Informal support can weaken. A relative may need income, heirs may want to sell, or a property may become too costly to maintain. Higher potential visitor earnings can add pressure to such decisions, but family circumstances can do so without tourism. To understand the household’s loss, we need the history of the arrangement.

The same point applies to security. A low payment may coexist with uncertainty if occupancy depends on someone else’s continuing goodwill. A higher payment under a stable agreement may provide some advantages. The household evaluates both the cost and the likelihood that it can remain. A measure of rent alone can miss that choice. A count of homeowners can miss the owner who cannot afford to mend a roof.

Informal support deserves respect. It also has limits. A policy that quietly assumes families will always absorb housing need may be relying on obligations that are unevenly distributed and increasingly difficult to sustain. The relative providing the home may have a small income too. The person providing care may be giving up paid work. Housing policy enters these relationships whether it recognizes them or not, because a shortage of alternatives increases what each family member must ask of the others.

What the rental research can tell us

The Center for a New Economy examined Puerto Rico’s short-term-rental market using data from 2014 through June 2020. Its analysis found an increasingly commercialized market. Among Airbnb hosts across the study period, those with multiple listings accounted for about thirty-nine percent of hosts but sixty-nine percent of listings and seventy-nine percent of revenue. These are listing and host-account measures, rather than a register of ultimate property ownership.[5]

The spare-room story captures only part of this market. A large operator has different means and reasons to buy more property. Someone managing many listings can distribute tasks, purchase services, and evaluate properties as a portfolio. A household letting an occasional room has a different relationship to the dwelling in which it lives.

In a separate analysis of San Juan metropolitan census tracts for 2017–2019, CNE found a positive relationship between short-term-rental density and housing costs after including measured controls. The model was observational. Its estimates support concern about a connection; they do not identify every cause of an individual rent increase or measure each household’s displacement.[5]

For a resident whose landlord has converted a home to visitor use, the direct sequence may be plain. Evidence about that conversion can establish what happened to that dwelling. A claim about the average effect across a large housing market demands another kind of evidence. Both scales matter. The broader estimate does not replace the individual account, and a vivid individual account cannot tell us how common the experience is.

The hard question concerns the history we cannot directly see. What would rents in the same neighborhood have done if short-term rentals had grown less, while other relevant conditions developed as they did? We observe the history that occurred. The comparison has to be built from evidence. A neighborhood becoming more desirable may attract visitors and higher-income residents at the same time. A new transport connection may affect both rents and bookings. Measuring what can be measured helps, but the remaining differences still matter.

That is why the study’s method matters. Does it separate the effect of visitor rentals from other changes happening at the same time? Or could those other changes explain some of the result? Studies can answer different parts of this question. We can use what they show while keeping track of what they leave open. We do not have to choose between knowing everything and learning nothing.

The count of listings requires care too. If a study collects every listing that appeared over several years, the total includes entries that were active at different times. Researchers also have to check whether changed accounts or multiple platforms have produced duplicates. An inventory of continuously available whole homes is a different object from an archive of advertisements. A responsible registry would help connect the online business to the actual premises and their authorized use.

None of this makes a household’s difficulty disappear into methodology. It improves the chance that a response reaches the source of the difficulty. If most relevant conversions are concentrated in a few streets, an islandwide average can obscure where action is urgent. If the largest operators account for a disproportionate share of activity, rules designed around occasional hosts may have missed the business they are regulating. Better measurement changes what can be done.

A rising sale price can produce four quite different reactions around one family table.

The owner sees a larger asset. The adult child sees a more distant purchase. The relative who earns commissions on sales sees work. The tenant wonders what the landlord will expect at the next renewal. They are discussing the same market with different positions inside it.

Even the owner may be less fortunate than the price suggests. Selling a house at a higher price is useful if the owner wants to move somewhere cheaper or has another secure place to live. If every suitable replacement nearby has risen too, the gain may be difficult to use without changing the household’s life. The house can be worth more while the owner remains short of money for repairs. The price the house might fetch will not pay Tuesday’s bill until the owner finds a way to use it.

For a prospective purchaser, financing adds another layer. The purchase price, down payment, interest rate, insurance, and maintenance obligations all affect the monthly burden. Two buyers offering the same price may face quite different costs if one brings cash and the other needs expensive credit. Equal treatment at the point of sale can coexist with unequal ability to reach that point.

Then there is the statistic everyone is reading. A median sale price describes the middle transaction among the homes sold in the relevant sample and period. Suppose five modest houses sell one year, while the following year’s transactions contain several expensive waterfront homes. The median can rise even if each particular house has changed little in value. A repeat-sales measure asks a different question by following properties over time. Asking prices, completed sale prices, assessed values, and rents also describe different things. Read the label before making the claim.

For renters, an average rent among existing tenants can lag the price facing someone searching today. A household that has remained under an older agreement may pay less than a new entrant. Someone forced to move can therefore encounter a sudden increase that is hard to see in an average of all occupied homes. If we want to know what a young adult or displaced tenant can afford now, the prices and conditions of available homes deserve direct attention.

Wages complicate the picture again. A rent can remain unchanged and become less affordable after a loss of work. A rent can rise while becoming more affordable for a household whose income rises faster. A tourism business can raise pay and still struggle to recruit if the workers it needs face housing costs beyond the reach of the new wage. Affordability is a relationship between resources and costs, experienced by particular households rather than by an average person who rents an average fraction of an average home.

That average person is wonderfully easy to house. Unfortunately, nobody has met them.

Enough certainty to make a choice

Tax incentives are often brought into this argument as if their name settles the matter. The U.S. Government Accountability Office’s 2025 examination of Puerto Rico’s resident-investor and export-service incentives found mixed evidence on overall economic costs and benefits and difficulty isolating their effects.[6] That finding leaves work to do. It does not settle whether the programs were worth their cost, or how much they contributed to housing pressure.

One problem can have several contributing mechanisms, and the mechanisms can reinforce each other. Damage may reduce usable supply. Difficulty financing repairs can prolong the loss. A small number of additional high-budget buyers can intensify competition in a thin local market. Visitor earnings can make conversions attractive. Low wages can prevent residents from bidding successfully even when a dwelling is physically available. Find out which pressures matter here, then choose a response that reaches them.

Suppose a district has plenty of sound, legally clear apartments that have shifted from year-round letting to continuous visitor use. Restrictions or incentives aimed at that use may affect residential availability. Suppose instead that its empty dwellings are severely damaged and their owners cannot fund repairs. A rule about bookings leaves the main barrier in place. In another district, new jobs may be arriving faster than homes can be built. There the construction pipeline and infrastructure may deserve priority. Similar symptoms can call for different treatment.

They also help us check a claim of success. Five hundred approved homes are not yet five hundred completed homes. Completed homes are not necessarily homes available to the households most in need. A repaired building may enter the visitor market if the funding arrangement leaves its use unrestricted. An affordable unit may become unaffordable after a temporary condition expires. Each stage changes what the public can reasonably claim to have achieved.

For the household, the test is direct. Is there a place I can safely occupy, on terms I can sustain, in time for the life I have to lead? Public reporting should be capable of following that question through the system. How many dwellings moved into usable condition? Who gained access? How durable is the arrangement? Which households are still waiting? The answers may be less impressive than a single large total, but they are closer to the actual work.

There is room for urgency here without inventing certainty. A municipality can improve its property records while researchers investigate causal effects. A program can address documented repair barriers while the debate about islandwide investment continues. Existing protections can be enforced where violations are established. A town can test a proposal against local evidence and change it as results arrive. No family should have to wait for a theory that explains every home on the island.

The empty house across the road remains a useful image, provided we finish the question it begins. Why is it empty? What would it take to make it a home? Who would live there afterward? Those questions lead to ownership, public authority, finance, and the distribution of gains. They also lead back to the fear with which this book began: that a place can become more valuable while a resident’s claim to a life within it becomes more fragile.

A house has walls. A home also requires a workable claim on time: enough confidence that you can keep living there to arrange the rest of your life around it. The next part of our inquiry follows the rules that grant, price, and sometimes weaken that claim.

Sources and notes

1. U.S. Department of Housing and Urban Development, Comprehensive Housing Market Analysis: Puerto Rico, as of July 1, 2025, pp. 3, 11–12, 22, and definitions on pp. 25–26. The 2025 population and household figures are analyst estimates; the 2020 figures are census counts. Market vacancy rates and the residual Census vacancy category have different denominators and meanings. The no-cash-rent figure uses 2023 ACS data.
2. U.S. Census Bureau, 2020 Census apportionment results, April 26, 2021. Puerto Rico’s population declined from 3,725,789 in 2010 to 3,285,874 in 2020.
3. U.S. Census Bureau, Housing Vacancy Survey definitions and explanations, particularly the distinctions among units for rent, for sale, held for occasional use, and other vacant units. The five-building street is illustrative, not an inventory from a Puerto Rican neighborhood.
4. Puerto Rico Department of Housing, Title Clearance Program. Cited for program purpose; individual eligibility and completed outcomes are separate questions.
5. Raúl Santiago-Bartolomei, Deepak Lamba-Nieves, Enrique Figueroa Grillasca, and Ysatis A. Santiago Venegas, The Impact of Short-Term Rentals in Puerto Rico: 2014–2020, Center for a New Economy, December 2022, pp. 5–8 and 12–15. The housing-cost analysis uses pooled observational regressions for San Juan metropolitan tracts, 2017–2019. This chapter deliberately avoids treating the report’s headline percentage shorthand as an islandwide causal coefficient.
6. U.S. Government Accountability Office, Puerto Rico: IRS Should Improve Oversight of Taxpayers Claiming Exemption from Federal Taxes, GAO-26-107225, published December 8, 2025, publicly released December 12, 2025.