Should you buy the rental, or the index fund?
Nobody knows what the next 15 years will return. So this runs your deal through every 15-year stretch of real Phoenix history, after taxes, in today's dollars, and shows you the whole spread of endings, not one number.
What the rental side actually models
Every month of the hold, then the sale. Not a cap-rate shortcut
Financing
Operating costs
Rental taxes & exit
Across the 407 15-year windows Phoenix-Mesa-Chandler, AZ has real data for, index funds finished ahead in 78%. How big the win or loss was depended on the start month more than on anything you control.
Same rental. Opposite lifetimes.
A percentage moves no one. So drag through every start month Phoenix has data for and watch both paths replay. The winner flips with the timing, and that's the point.
Audit this window: year by year, with the tax math
| Year | Rental cash flow | Taxable rental income | Tax effect | Gross equity home value − loan | Cash-out value if you sold this year, net of tax | Index account |
|---|
The deal is only half the math.
Everything above is a typical deal run on real history. Whether you can actually swing the down payment, and what buying it would do to your net worth, lives in your accounts. The free dashboard is open now, with no account. Get early access and a free account keeps your numbers on every device, and puts you first in line for the Safe-to-Spend app. It shows how much you can safely put toward the next down payment, and tracks your net worth as it compounds, whichever path you pick.
This page probably just told you the index wins. We're fine with that; the app works either way.
Is real estate a better investment than stocks?
Two different questions hide inside that one, and most arguments about it answer whichever one suits the arguer.
As an asset, the house loses, and it does not lose narrowly. Across the 136 years of national home-price history we hold, US home prices grew 0.5% a year after inflation. The S&P 500 with dividends reinvested grew 6.7% a year over the same months. Housing also loses to the index's price alone, before a single dividend is counted, which grew 2.6%.
As a business, it can win. A rental is bought with a bank's money, paid down by a tenant, and taxed under rules written for landlords. Those are three engines an index fund does not have, and they are strong enough to close a gap that size. Whether they actually close it depends on the deal and on when you bought, which is not something anyone can answer in the abstract.
That is what the calculator above works out, for one specific property. This section is the first question, because the second one cannot be read honestly without it.
The house as an asset, before rent, leverage or tax
Take every 30-year stretch the data supports, 1,268 of them, and compare the national home-price index against the S&P 500 total return over the identical months, both after inflation. Home prices finished ahead in none of them. The median 30-year window returned 0.5% a year for housing against 6.8% for the index.
Shorten the hold to 15 years, where one bad decade has less time to wash out, and housing wins 49 of 1,448 windows, roughly 3 percent. Time is not on the asset's side. It is on the index's.
Narrow to the modern data, where the series is a single repeat-sales construction rather than a splice, and the shape is the same: 51 years from 1975, housing at 1.3% a year after inflation against 8.2% for the index.
Three things have to be said about those figures, or they are being oversold:
- The windows overlap, heavily. 1,268 thirty-year windows drawn from 136 years of history is about 4 genuinely independent runs. Read it as a direction that never reversed, not as odds of anything.
- Everything in this series before 1953 is annual data filled in month by month, and its construction changes more than once across the span. Its own caveats open with the words "context mode, not the default", and that is how we use it: context, not the engine that runs your deal. Re-running the same test on January starts only, one start a year instead of one a month so near-identical windows are not counted twelve times over, gives 106 windows and the same answer: none.
- It is a price index. There is no rent in it, no mortgage, and no tax. There is also no property tax, insurance, maintenance or capex, and those are real money that a homeowner really pays. All four cut the same direction, so the true return on a house owned outright and left empty is worse than 0.5%, not better.
Where the gap actually comes from: leverage, cash flow and tax
That last point is the reason this section exists. If the asset itself loses by that much, then a rental that wins is not winning because it is real estate. It is winning for three reasons that have nothing to do with the house, and the calculator above splits every result into exactly those parts.
Leverage is the borrowing. Twenty percent down controls five times its own value, so a 5% move in the house is a 25% move on your money. It is usually the largest single line in the split, and it is also why the worst windows on this page are as bad as they are: the multiplier does not care which way the house moves, and interest, property tax, insurance, maintenance and capex are all charged on the whole house, never on your share of it.
Cash flow is the tenant. Someone else services debt you would otherwise service yourself, and the share of each payment that retires principal is money moving into your equity instead of out of your pocket.
Tax is the code. Depreciation deducts a fraction of the building every year against income the building is still earning, and the passive-loss rules decide whether you may use that deduction now or must bank it until you sell. It is a deferral plus a rate arbitrage, not a gift, because the depreciation is recaptured at the sale at up to 25%. Every competing calculator we could open leaves this out entirely, and leaving it out always flatters the rental.
What is left over, once those three are removed, is the house on its own, owned outright and net of what it costs to hold. That is the bucket the calculator labels appreciation, and on the numbers above it is a loss far more often than not.
So the honest version of "real estate builds wealth" is that three of the four engines are financing, tenancy and tax law. Only one of them is the house, and it is the one that loses.
Cash flow versus appreciation: which one actually decides it
Landlords argue about this endlessly and almost always without numbers. The two do different jobs, and the difference is leverage.
Appreciation is leveraged. Cash flow is not. A 20% down payment multiplies appreciation by five and multiplies the rent by one, so in a strong window appreciation is the larger number by a wide margin. Every one of the best outcomes on this page is an appreciation outcome.
Cash flow decides whether you are still holding the property when that window arrives. A rental that loses money every month is one you have to feed out of your own income, often in exactly the years you can least afford to. It has a second and quieter cost here too: this comparison hands the index fund every dollar the property demands, at the moment it demands it. A property that bleeds is not only costing you money, it is buying shares for the other side.
The reverse holds when the property earns a surplus. That money is not treated as spending; it is invested in the same index and belongs to the rental path. Which means a cash-flowing rental is quietly running an index fund of its own, and part of what looks like a real-estate win is that side account.
So: appreciation decides how large the win is, and cash flow decides whether there is one. Change the rent in the calculator above and watch the share of winning windows move. You cannot change the appreciation, because this tool does not let you assume one. It comes from what actually happened.
Rental ROI versus the S&P 500: why the quoted numbers do not compare
Search this comparison and you will find rental returns quoted at 8 to 12 percent, sometimes 25 percent, set against "the S&P's 10 percent." Those figures are usually not wrong. They are just not comparable to the number sitting beside them.
Cap rate deliberately ignores your mortgage. It measures the property's return, not yours, which is useful for comparing two buildings and useless for comparing a building to a brokerage account.
Cash-on-cash is computed on the cash you put in, so the leverage multiplier is already inside it. Setting a leveraged return against an unleveraged one and declaring a winner is the single most common error in this whole argument, and it is why a rental can be made to look like it returns 25% while an index fund returns 10%.
Nearly all of them are before tax. Where one is quoted for year one, that is the year a rental looks best: the rent has not yet met a vacancy, the roof has not yet failed, and the depreciation recapture waiting at the sale is decades away and invisible.
The only comparison that settles anything is identical dollars going in at identical times, both sides liquidated under the same rules, both reported after every tax. That is what this page computes, and it is the reason the numbers here come out lower than the ones published by people who earn a percentage of rent.
Why every other calculator gives you a different answer
If you have run this comparison elsewhere and got a cleaner result, here is what usually accounts for the gap. We open the calculators that surface for this question and read them, rather than trusting how they describe themselves. Last checked 2026-08-02. Ten opened; three blocked automated access, and nothing below is claimed about those three.
None of the ten models the tax code that decides a rental. Not depreciation, not depreciation recapture, not the passive-loss limits, not the net investment income tax, not state income tax. One goes further than the rest, handling capital gains at the sale, the mortgage interest deduction and the primary-residence exclusion, and it still leaves out every line above, which are the ones that move a rental result most. Another says in its own text that including tax would take dozens more inputs, and stops there. This is not a rounding difference: on a long hold the exit tax bill is frequently the largest single number in the whole comparison.
None of the ten replays history. Every one uses a single rate of return: a number you type in, or an average the author picked. That converts the question into arithmetic on an assumption, which is why two calculators can disagree by hundreds of thousands of dollars while both being internally consistent. Change the assumed appreciation rate from 3% to 5% and the answer flips, so the assumption is doing the work, not the model.
A large share of the surrounding analysis is published by property managers and brokerages, quoting after-tax returns for rentals well above the S&P. Those are companies whose revenue is a percentage of rent. Their numbers are marketing rather than modeling, and they are usually the ones that omit recapture.
Where we differ from a source you trust, the audit table is the place to settle it. Open any window, read the year-by-year tax math, and find the line where the two of you disagree.
How this calculator works
What this actually models
Every month of the hold, then the sale. Not a cap-rate shortcut and not an assumed rate of return. The rental path pays the mortgage down on a real amortization schedule, collects rent net of vacancy, pays property tax and insurance and maintenance and capex and management, settles a tax year every year, and then sells. The index path receives the identical schedule of dollars and invests them in the S&P 500 total-return index: the down payment and closing costs at the start, and every dollar the property later demands when its cash flow is negative.
That identical-dollars rule is the point. Neither side is handed money the other did not get, at any month, to the cent. Without it a comparison can produce any answer you like just by choosing when each side gets funded.
Both paths end in after-tax, inflation-adjusted dollars. Real dollars, not nominal, because a window that starts in 1980 ends in 1995 money and a window that starts in 2005 ends in 2020 money; putting those on one axis would make high-inflation decades look like outperformance.
The tax model
This is where most comparisons stop, and it is the part that moves the answer most. On the rental side: straight-line depreciation over 27.5 years on the building only, §469 passive-activity loss limits with the $25,000 allowance phasing out between $100,000 and $150,000 of modified AGI, suspended losses released at the sale, the §199A qualified business income deduction as an explicit user assertion, the net investment income tax, and at the exit both unrecaptured §1250 depreciation recapture (a separate bucket, capped at 25%) and long-term capital gains on the appreciation.
On the index side: dividends taxed annually in a taxable account, and long-term capital gains at the sale. Both sides get state and local income tax at your jurisdiction's rates, and both sides get the §1014 stepped-up basis if you choose to hold to death rather than sell.
Symmetry is deliberate. Taxing one path and not the other is the most common way this comparison goes wrong, and it goes wrong in both directions depending on which side the author preferred.
What we leave out, on purpose
These are exclusions, not oversights. Each one is either a different asset class, a strategy that requires a paid professional study, or a detail whose effect is smaller than the honest error bars on the inputs.
- Mid-hold §1031 exchanges. The carryover-basis mechanics are implemented and tested, but an exchange is not an input on this page. That is deliberate, so that no one can set an option the model would quietly ignore. Holding to death already captures the only durable outcome of serial exchanging.
- Cost segregation and bonus depreciation. Largely a timing shift that increases later recapture, and it requires a paid engineering study.
- The short-term-rental strategy. Material participation in short-term rentals escapes the passive-loss limits, but that is a different business, and a property renting by the night may not even be 27.5-year residential property. It is not a toggle on this model; it is a different model.
- Interest tracing. We assume purchase-money debt used entirely for the rental. Cash-out refinances and borrowing against a personal residence trace differently.
- The §163(j) business interest limitation, which the small-business exception removes for essentially every individual landlord.
- Community property step-up, which doubles the §1014 reset in nine states depending on domicile and how title is held.
- Lot-level basis selection on the index side. Aggregate basis only.
- Transaction reality: financing contingencies, tenant nonpayment beyond your vacancy rate, insurance markets repricing a whole region at once.
- Any claim about the future. This tool reports what history did. It does not forecast.
Where the home-price history comes from, and what it misses
Home prices are the FHFA House Price Index for your metro. That is a repeat-sales index, meaning it tracks the same houses selling more than once rather than comparing this year's sales mix to last year's. It is the right construction for this question, and it has a known limitation: the index is built from mortgages bought or guaranteed by Fannie Mae and Freddie Mac, so it under-represents cash purchases, jumbo loans and the subprime market. In practice the very top and the very bottom of a market are thinner in this data than they are in reality.
Coverage also varies a lot by metro. Some series begin in 1975; others start in the late 1980s. A handful have interior gaps where the data stops and resumes, and a window is only ever drawn wholly inside one unbroken stretch, so a metro with a gap supports fewer windows than its calendar span suggests. Whichever series answered your question, and the exact span it covers, is stamped on the page beside the result.
Rents are seeded from the Zillow Observed Rent Index and property values from the Zillow Home Value Index; property-tax rates come from the ACS five-year survey at the county level; state average homeowners premiums come from a published 2026 industry survey. Every seeded value is editable, and every one shows where it came from.
Why the win rate is a lean, not a probability
The windows overlap. A 15-year hold starting in January 1990 and one starting in February 1990 share fourteen years and eleven months of the same history, so they are not independent observations; they are one long history sliced many ways. Four hundred windows are not four hundred trials.
That is why the page shows, beside every 'percent of windows' figure, roughly how many genuinely non-overlapping runs the data supports. It is usually a small number. Read the win rate as the direction history leaned, not as the odds of anything.
One more honest limit on the stock side: the S&P total-return series is spliced from more than one source at its recent end, and that splice carries a small residual uncertainty, bounded at well under one percent over a thirty-year window, in the direction of understating the index. It does not change any conclusion on this page, but it is the reason we do not quote the stock column to the dollar.
Questions people ask
Why would I deal with tenants and repairs when an index fund returns 10% a year?
Because 10% a year is an average nobody actually earned. It is the long-run mean of a distribution whose individual outcomes range enormously depending on when you started. This calculator does not use an average at all. It replays your deal across every start month in the price history of your metro, and reports the whole spread: what the typical window did, what the best and worst did, and how often each side won. The honest answer to 'which is better' is a distribution, not a rate.
I compared my down payment growing against the house appreciating and the rental looked great. What am I missing?
Two things, and they point in opposite directions. You are comparing the growth of your down payment against appreciation on the whole house, which is the leverage effect and is real, but you have probably not subtracted the cost of the borrowed money, and you may not have counted the rent at all. A complete comparison needs the mortgage interest, taxes, insurance, maintenance, capex, vacancy and management on one side, the rent coming in, and the tax bill at the sale. That combination is what this page computes.
In what situations does the rental actually beat the index fund?
Consistently: a high rent-to-price ratio at purchase, a long hold, a low-tax state, and dying with it rather than selling it. Rent yield is the single biggest lever: it is the part of the return that does not depend on appreciation. A long hold amortizes the 6-9% round-trip transaction cost across more years. And holding to death replaces the entire deferred tax bill with a stepped-up basis under §1014, which is the one outcome the tax code genuinely rewards. Change the horizon and the exit strategy on this page and you can watch each of those move the win rate.
What makes more millionaires, real estate or the stock market?
Real estate has a real claim on this: a great deal of genuine wealth was built in it. The asset's returns are not the reason. In the national home-price history this page draws on, houses did not beat the S&P 500 total return in a single 30-year window on record, and lost to the index's price alone before any dividends. Those windows overlap heavily, so read that as a direction that never reversed rather than as odds. What real estate has that an index fund does not is leverage an ordinary buyer can actually obtain, a tenant servicing the debt, a tax code written for landlords, and a mortgage payment that forces saving whether you meant to or not. Those are the mechanisms, and the section above separates each one so you can see how much of a result is the house and how much is the financing.
How long do I have to hold before it is worth it?
Long enough for appreciation and rent to clear the round-trip transaction cost, which is roughly 2% to buy and 6-7% to sell, on the full value of the property, not on your equity. On a leveraged deal that cost is a large multiple of your down payment. Shorten the horizon on this page and the rental's win rate falls sharply, which is that cost showing up.
Depreciation shelters the rental income. Don't the tax breaks make real estate win?
Depreciation is a deferral plus a rate arbitrage, not forgiveness. You deduct roughly 1/27.5 of the building's value each year against ordinary income, then pay it back as unrecaptured §1250 gain when you sell, at a rate capped at 25%. If your ordinary rate is above 25% you genuinely win the difference, and if it is below 25% you can lose on the trade. This calculator computes both halves (the annual deduction and the recapture at the sale) instead of counting only the first.
I never claimed depreciation on my rental. Do I still owe recapture when I sell?
Yes. The rule is 'allowed or allowable': recapture is computed on the depreciation you were entitled to take, whether or not you took it. Skipping the deduction does not avoid the tax, it just means you paid full price for it. This model assumes depreciation is always taken, which is both correct under that rule and the conservative assumption.
How much does depreciation recapture actually cost when I sell?
It is taxed as a separate bucket from your capital gain, at your ordinary rate capped at 25%. Most comparisons blend one capital-gains rate across the entire gain and understate the bill, because the recapture layer is taxed higher than the appreciation layer. This engine splits them and taxes each correctly, then reports the total exit tax as its own line so you can see what the sale cost you.
Why can't I deduct my rental loss against my W-2 income?
Section 469 treats rental real estate as passive. There is a $25,000 allowance for active participants, but it phases out between $100,000 and $150,000 of modified AGI and is gone above that, which is exactly the income range of most people running this comparison. Disallowed losses are not lost; they suspend and release when you sell. This calculator tracks the suspension and the release rather than assuming the deduction lands in the year it occurs.
Doesn't the stock side get taxed too? Isn't this rigged if you only tax the rental?
It gets taxed, and symmetrically. In a taxable account the index path pays tax on dividends every year and capital-gains tax when it is sold, at the same federal, state and NIIT rates the rental faces, with the same §1014 step-up if you hold to death. You can also switch the index side to a tax-advantaged account to see the comparison without that drag. Taxing one side and not the other is the most common way these comparisons go wrong, in both directions.
How much does my state's income tax change the answer?
More than most comparisons admit, and it usually favors the index. Rental income is ordinary income at the state level every year, while a buy-and-hold index position mostly defers until sale. This engine applies state and local income tax on both paths (local income taxes are modeled for the eleven states that have them), which is why the same deal can flip between a no-income-tax state and a high-tax one. One stated limitation: state capital-gains treatment is modeled at the ordinary rate, correct for the roughly 41 states that tax long-term gains as ordinary income and conservative against the rental elsewhere.
Can I just 1031 exchange forever and never pay the tax?
An exchange defers; it does not forgive. It also requires you to keep buying replacement property, and it carries your old basis forward, so the deferred recapture and gain follow you. The only durable escape is dying with the property, which resets the basis under §1014, and that outcome is directly modeled here as the hold-to-death exit. Mid-hold exchanges are deliberately not an input on this page rather than being an option the model would quietly ignore.
With 20% down I'm leveraged 5x, so 5% appreciation is a 25% return. How do stocks compete?
The multiplier is real, and it applies to the losses identically. It also arrives with costs: you pay interest on the borrowed money, and property tax, insurance, maintenance and capex are charged on the full value of the house, not on your share of it. A 5x multiplier on appreciation net of a mortgage is a very different number from 5x gross. The worst windows on this page are worst precisely because leverage worked in reverse.
Isn't comparing a leveraged property to an unleveraged index fund a mathematical trick?
It is a fair objection and it is why this page is built the way it is. Both paths receive an identical schedule of dollars (the down payment and closing costs at the start, and every dollar you have to feed a negative-cash-flow property later), so neither side gets money the other did not. Then the gap is decomposed into leverage, cash flow, tax and appreciation, so you can see how much of the result is the borrowing itself rather than the asset.
I have cash. Buy a rental outright, or put 20% down and the rest in index funds?
That is exactly the comparison this engine runs, because the rental path is two assets, not one: the property, plus a side account holding every dollar of positive after-tax cash flow it throws off. Lowering the down payment moves money from the property into the market and raises both the upside and the downside. Change the down payment on this page and the spread widens or narrows in front of you.
What should I budget for vacancy, capex and maintenance? Is 1% of value enough?
Maintenance and capex are separate things and both are lumpy. Maintenance is the ongoing repair spend; capex is the roof, the HVAC, the water heater: costs that arrive rarely and large, which is why they have to be reserved monthly rather than noticed annually. This page defaults to 1% of value for maintenance and 0.5% for capex, with vacancy from Census data for your metro, and every one of them is editable. Under-reserving these is the single most common reason an amateur pro-forma looks better than the property.
Does a property that passes the 1% rule beat the index?
The 1% rule (monthly rent at or above 1% of purchase price) and its stricter 2% cousin are screening shortcuts, not returns. They tell you a deal is worth underwriting; they do not tell you what you end up with. Neither one knows your mortgage rate, your property tax rate, your insurance premium, the age of the roof, your state's income tax, or the depreciation recapture waiting at the sale, and those are what decide the outcome. Rent yield genuinely is the strongest single lever in this comparison, so a property clearing 1% starts a long way ahead of one that does not. Put its real numbers into this page and you get the distribution of outcomes instead of a pass or a fail.
Isn't cap rate the rental's version of the stock market's return?
No, and treating it that way is an apples-to-oranges error that flatters the rental. Cap rate is unlevered net operating income divided by price: it excludes financing, excludes appreciation and excludes tax. A stock market return is total return after nothing has been removed at all. Comparing a 6% cap rate to a 10% market return compares two quantities that do not measure the same thing.
Couldn't I just withdraw from my brokerage each month and get the same income?
Yes, and that is the cleanest way to see why 'it pays me monthly' is not by itself an advantage. Selling 4% of a portfolio a year and collecting 4% in rent are economically comparable; the differences that matter are the tax treatment, the transaction costs, and whether the underlying asset holds up. This calculator compares terminal after-tax wealth precisely so that cash flow is counted without being mistaken for a return.
Real estate is less volatile than stocks, so isn't it lower risk?
Measured volatility is lower partly because house prices are measured from infrequent transactions and appraisals, which smooths the series; the price does not update daily just because the value moved. Underneath that, a rental is one undiversified asset, in one county, usually bought with leverage, and it cannot be sold in an afternoon. The 2007-2011 windows on this page show what that combination does when it goes wrong.
It's called passive income, but how many hours a year does a rental actually take?
We do not put a number on your hours, because it varies too much for a default to be honest. Reported experience runs from a few hours a month to twenty or forty hours around a single turnover. What this calculator does instead is let you price the alternative: property management is set to 8% of collected rent by default, which is roughly what it costs to buy those hours back. Set it to zero and the deal improves by exactly the amount you are working for.
Where does this calculator's answer come from, and can I check it?
Home prices come from the FHFA House Price Index for your metro, which begins between 1975 and the late 1980s depending on the market; stock returns come from the Shiller S&P 500 total-return series; rents and property values are seeded from Zillow ZORI and ZHVI, property-tax rates from the ACS five-year survey, and tax parameters from the 2026 federal, state and local schedules. Every window on this page can be opened year by year: the audit table shows each tax year's cash flow, taxable income, tax and cash-out value, so the headline is checkable rather than asserted.