Can You Actually Keep the House?

The address is a label only. Tax, insurance and maintenance all key off the town selected above it — so where a village has a different mailing address than its taxing town, pick the taxing town.

A HELOC with a zero balance is not a zero. If the line is open it can be drawn before closing. Freezing or closing it is a settlement term, not a formality.

Support counts as qualifying income only when it is in writing, has a receipt history, and continues far enough out. Undocumented support is income you live on but cannot borrow against.

6.66% is the Freddie Mac 30-year average for the week of July 30, 2026 — a placeholder for arithmetic, not a rate anyone has offered. Check this week's average, then use the lender's actual quote.

Most of the inputs above explain themselves. These are the ones that don't, and a few of them move the answer more than anything else on the page.

The three maintenance numbers are not the same number

This is the one that confuses people most, because three separate fields all sound like they are about the condition of the house. They do different jobs and they do not overlap.

Deferred maintenance is work the house needs right now that nobody has done. A roof at the end of its life, a failing septic, knob-and-tube wiring an insurer will refuse. It is a one-time dollar figure and it comes straight off net equity, because whoever ends up selling either fixes it or discounts for it. If you are entering a number here, use a contractor's estimate rather than a guess — this line is money taken away from the person keeping the house.

Condition % is not about repairs at all. It is a depreciation figure: what fraction of a brand-new equivalent this building represents today. The default of 72% describes ordinary, well-kept housing stock of average age. A gut renovation might be 90%. A tired antique that has not been touched since the seventies might be 55%. It only exists to estimate what the structure is worth, which the next section explains.

Maintenance reserve is the monthly figure the tool calculates for you. It is forward-looking — the money that should be set aside for the roof that will need replacing in eleven years, the furnace in six. It is not a bill anybody sends. It is the line most people leave out of a divorce budget entirely, and then it arrives all at once.

Rebuild cost, and why maintenance is not one percent of the price

The old rule of thumb says budget one percent of the purchase price each year for maintenance. In this footprint that rule breaks badly, and it breaks in opposite directions depending on where the house is.

Roofs, furnaces and siding belong to the building. Land needs nothing. So the tool estimates what the structure alone would cost to rebuild — square footage times rebuild cost per square foot, adjusted by condition % — treats the rest of the price as land, and sets both the maintenance reserve and the insurance premium against the building only.

The difference is not academic. Narragansett's median house is $850,000, but the median house there is 1,379 square feet, so the building is worth roughly $248,000 and the land carries the rest. Its maintenance reserve is about $207 a month. New London's median is $360,000 on 1,468 square feet, so almost all of the price is building — and its reserve is $220 a month. The more expensive house reserves less. Under the one-percent rule, Narragansett would have been charged $708 a month, and in a buyout negotiation that error runs one direction only: it inflates the carrying cost of exactly the shoreline houses where somebody is arguing they cannot afford to stay.

The default of $250 per square foot reflects average-quality construction in southeastern Connecticut. If you know the real figure for this house, use it.

Assessed building and land values

These two optional fields are the single best thing you can enter. The assessor has already split this exact parcel into land and building, and their split beats any estimate the tool can make from a town median. Enter both and the tool stops estimating the structure value and uses theirs — which changes the maintenance reserve, the insurance premium, and the property tax line all at once.

Enter them as they appear on the property card, without converting anything. The tool handles the assessment ratio itself: 70% in Connecticut, 60% for a primary residence in New Milford, roughly 100% in Rhode Island. Links to the right assessor database appear once a town is selected.

Fire and special district levies

A parcel can sit inside a fire, borough, shoreline or improvement district that levies its own rate on top of the town's. It is a real tax and it arrives as its own bill, which is why buyers and settlement agreements alike routinely miss it.

In Connecticut, West Haven's fire districts add roughly 7.3 to 9.0 mills, and Stonington's Masons Island district adds 1.60. In Rhode Island fire districts are widespread and bill separately from the town almost everywhere. Enter the rate here and the tax line recalculates. Leave it at zero and the tool flags on the output that you did, because a tax figure that silently omits a district is worse than no figure at all.

The assessor will tell you which district an address falls in. It takes one phone call and it can be worth several hundred dollars a year in a number two people are about to divide.

Water exposure

This changes two things. Insurance is priced differently on the shore — wind deductibles, coverage restrictions, sometimes a different carrier entirely — so the modeled premium rises from 0.50% of structure value inland to 0.75% with a water view and 1.20% on the waterfront. Maintenance rises too, by 10% and 25% respectively, because salt air is hard on roofs, siding and condensers.

There is a fourth setting that does something different from the other three. Inland with deeded water rights or beach association leaves insurance and maintenance exactly where inland leaves them, because the house itself is not any more exposed — it is not on the water, it is not taking salt, and its roof will last as long as any other inland roof. What those rights change is what the property is worth.

That distinction is easy to get wrong in the other direction. Selecting “water view” or “waterfront” because a house has beach rights would raise the modelled insurance premium by half and the maintenance reserve by a quarter, for a building that carries neither risk — and in a buyout that error runs against whoever is trying to keep the house.

One local wrinkle worth knowing: a number of shoreline beach associations in Connecticut are separate taxing districts with their own mill rate on top of the town’s, and several of Old Lyme’s beach communities are organised that way. If the parcel sits in one, that rate belongs in the district field, and the annual dues belong in grounds and services. The assessor will confirm both.

Note that flood insurance is separate and starts at zero deliberately. A homeowners policy never covers flood, there is no honest default figure, and on a coastal parcel a policy can run into four figures a year. Check the address on the FEMA map and get a quote.

The existing loan: rate, months remaining, and loan type

The tool asks for months remaining rather than years because that is what appears on a mortgage statement, and because the deferred-sale path amortizes the existing loan forward to work out what is still owed at the eventual sale.

Loan type is worth more attention than it looks. FHA, VA and USDA loans are generally assumable. If the existing loan carries a rate from a few years ago and the servicer permits assumption with release of liability, the spouse keeping the house can hold onto that payment instead of refinancing at today's rate. On the tool's defaults that is a difference of nearly $56,000 over five years — larger than the equity gap the page opens with.

Two catches, both of which the output states. Assumption produces no cash, so the buyout still has to come from a second lien or another asset. And release of liability is the part people forget: without it, the departing spouse remains on the debt no matter what the decree says.

Which equity the buyout is calculated on

This dropdown decides whether the buyout figure is a share of gross equity (value minus payoff) or net equity (after the costs of an eventual sale). It defaults to gross, because that is what settlement agreements overwhelmingly do — and the whole point of the first section is to show what that convention costs the person keeping the house.

Switch it to net and you can see the same deal priced the other way. Neither is right or wrong as a matter of law. It is a negotiated term, and it is easier to negotiate when both sides can see the size of it.

The one assumption that drives the five-year comparison

Share of gross income each can direct to housing and savings. This is the most consequential field on the page and the least obvious, so it is worth understanding before you trust the three-path comparison.

Comparing "keep the house" against "sell and rent" requires a common yardstick, because the two paths spend different amounts on housing every month. So the tool gives each spouse a monthly budget — that percentage of their gross income — and charges housing against it. Whatever housing does not consume gets invested at the stated return. Whatever it overruns comes out of savings, and the output warns you when a path runs the account negative.

The default is 35%. That is a planning figure, not a rule. Someone with no other debt and low expenses might genuinely direct 45%; someone supporting two households on one income might have 25% and no more. Move it and the ranking of the three paths can change, which is not a flaw in the model — it is the actual answer. Whether keeping the house is the right call depends on how much room the person keeping it actually has.

Appreciation, investment return, and rent inflation

These three set the terms of the comparison and none of them is knowable. Appreciation at 3% is roughly the long-run average and says nothing about the next five years in a specific town. Investment return at 6% is a conventional balanced-portfolio assumption. Rent inflation at 3% is a guess.

Use them as a sensitivity test rather than a forecast. Run the comparison at 1.5% appreciation and again at 4.5%. If the same path wins both times, the decision is robust and you can stop arguing about it. If the ranking flips, then the disagreement in the room is not really about the house — it is about what people believe the market will do, and that is worth naming out loud.

Support, and why the checkbox matters more than the amount

Support received is added to the recipient's monthly budget in the five-year projection either way. The checkbox controls something different: whether it counts as qualifying income for the refinance.

Lenders generally require support to be court-ordered in writing, to have a documented history of actual receipt, and to be scheduled to continue at least three years past closing. Support that fails any of those is money you live on but cannot borrow against. Since the continuance term is something that gets drafted into the agreement, this is worth raising with the attorney before the loan application rather than after — the output shows what the debt-to-income ratio would be if it counted, which is usually the argument for getting the language right.

Cost basis, the capital gains rate, and the NIIT

Original purchase price plus capital improvements is the cost basis. Without it the tool refuses to compute the tax module, because a capital gains estimate built on a guessed basis is worse than no estimate. Missing improvement records overstate the gain and therefore the tax — a new roof, an addition, a finished basement all raise basis, and the receipts are worth finding.

The federal rate defaults to 15%, which covers most filers; high earners pay 20% and some pay nothing. The net investment income tax is an additional 3.8% that applies above certain income thresholds. State tax is added automatically at Connecticut's or Rhode Island's top rate. All of it is an estimate and all of it belongs with a CPA or a CDFA — the tool's job is to show you there is a number worth asking about, not to compute your return.

What this tool does not do

It does not value the house. Every figure scales off the value you enter, and the tool says on the output which basis you told it that value came from. It does not know about pensions, retirement accounts, business interests, or any other marital asset that might offset the house in a settlement. It does not know your jurisdiction's approach to property division, or this judge's, or this mediator's.

It prices one asset carefully so that the people dividing it are arguing about the same numbers. That is the whole job.

Optional. Nothing is sent unless you fill this in and press send. Email is not a confidential or privileged channel — do not include anything you would not want read by someone else. The property address and file reference are included if you entered them; clear those fields first if you would rather they were not.