Project reviews
Flipping mistakes in Barcelona: how the miscalculations work
Four miscalculations recur in flipping more often than any others, and each works the same way: the error is small, while the outcome of the project changes by tens of percent.
A flipper's mistake is almost always one of four: an understated cost estimate, an underestimated permit timeline, an exit price taken from listings, and a forgotten tax. The mechanics are shared: the error is small, and what it eats is the thinnest part of the result. Below is that mechanics and what two completed DNPI projects show. There are no invented stories here; the company has no other real cases.
Which miscalculations recur most often
| Miscalculation | How it works | What checks it before the transaction |
|---|---|---|
| Estimate from a price per m² | leaves out services, structure, waste removal, access | an estimate by chapters with measurements |
| No reserve | any finding comes straight out of the result | a reserve as a budget line |
| Optimistic permit timeline | holding costs run while the works stand still | the approval regime settled before negotiating |
| Exit price from listings | a listing is an asking price, not a transaction | comparable closed transactions |
| Taxes left out of the calculation | they reduce the result at entry and at exit | a calculation with an asesor fiscal before the arras |
| No plan B | if the price falls there is no second scenario | plans A, B and C before the purchase |
Five of the six lines can be checked before the transaction, and only the exit price stays probabilistic. Hence the order of work: close out what can be verified first, then calculate the margin of safety on what cannot (mistakes in project calculations, margin of safety).
Why the estimate gets understated
Understatement comes out of the method: a figure per m² is taken from a market review, multiplied by the floor area and treated as the budget. Between the review and the property stand building services, damp, structure, access and the standard of materials, and none of those factors is in the average. The commercial portal preciom2 states itself that its prices are indicative and vary by up to ±20% (preciom2).
Next to it stands a second assumption: that renovation spending automatically turns into an increase in price. In September 2023 Gesvalt published estimates for large cities where a full renovation came to around 650 €/m² and the increase in value to around 1,050 €/m²; a fully renovated apartment nonetheless remained 7% cheaper than new-build (Gesvalt, Gesvalt). These are estimates, not a rule (how to read a renovation estimate).
What happens when the exit price is taken from listings
The asking price and the transaction price are different quantities. According to Idealista's publication of 6 October 2026, the average asking price for housing in Barcelona in September was 5,462 €/m². According to the Estadística Registral Inmobiliaria for Q2 2026, the price of registered transactions in the capital of the province of Barcelona was 5,045 €/m². The periods differ, and a negotiating discount cannot be derived from the difference between them, but the direction is visible: a transaction does not close at the listed price.
The second part of the miscalculation concerns the timeline. According to idealista/data for Q2 2026, across Spain 36% of the apartments sold spent from three months to a year on the market and 11% longer than a year. An inflated exit price does not correct itself: the property sits and accumulates holding costs (asking price against transaction price, exit strategy).
The third part, and the most expensive, is believing that price growth insures against a mistake. In table 4 of the Ministerio de Vivienda data, the average appraised value of housing in Barcelona was 3,950 €/m² in Q2 2008 and 2,385 €/m² in Q1 2014, 39.6% lower (market cycles).
How taxes drop out of the calculation
Taxes stand in two places, and usually both are forgotten. At entry the buyer pays ITP (impuesto de transmisiones patrimoniales, the transfer tax on resale housing): under the Agència Tributària de Catalunya scale from 27.06.2025 that is 10% on the share of the price up to 600,000 €, 11% up to 900,000 €, 12% up to 1.5 M€ and 13% above, and 20% for a gran tenedor, a large-scale housing holder. At exit the seller pays tax on the result and plusvalía municipal (the municipal tax on the increase in land value) on a sale for consideration (art. 106 TRLRHL). If the seller is a non-resident individual, the buyer withholds 3% of the price towards IRNR, the non-resident income tax (art. 25.2 TRLIRNR). What applies to your structure is determined by an asesor fiscal (tax on flipping profit, two methods of plusvalía, the 3% withholding).
What a worked example of an estimate error shows
A worked example, not a forecast. Assumptions on round numbers: purchase 600,000 €, ITP at 10% on the ATC scale, paperwork at 2% of the purchase price, holding costs of 12,000 € over the term, exit costs at 4% of the sale price, sale 860,000 €.
| Line | Estimate as planned | Estimate plus 20% |
|---|---|---|
| Purchase | 600,000 € | 600,000 € |
| ITP and paperwork | 72,000 € | 72,000 € |
| Renovation | 90,000 € | 108,000 € |
| Holding costs | 12,000 € | 12,000 € |
| Exit costs | 34,400 € | 34,400 € |
| Sale | 860,000 € | 860,000 € |
| Left before tax on the result | 51,600 € | 33,600 € |
The estimate rose by a fifth, and the result before tax fell by roughly a third. Add a drop of 40,000 € in the exit price and what is left goes towards zero. Inflation works separately: on INE's preliminary estimate of 29 September 2026 the annual IPC was 4.9% (INE), and a result in nominal euros is not yet a result in purchasing power (real return).
What is not visible in the two real projects
| Project | Purchase | Costs | Sale | Difference before other costs |
|---|---|---|---|---|
| Compte de Urgel, Barcelona | 410,000 € | renovation 76,230 € | 720,000 € | 233,770 € |
| Sitges, land | 509,250 € | permits and licences 20,000 € | 760,000 € | 230,750 € |
The difference equals the sale price minus the purchase and the costs listed; taxes, financing, paperwork and other costs are not included; this is not net profit and not an investor return; past results do not guarantee future ones. The project term in Sitges was 2 years.
These tables contain no taxes at entry or exit, no cost of money, no selling costs and no floor area, which means no figure per m² either. Those lines are exactly where the miscalculation happens when a project is counted by the formula "sale minus purchase minus renovation" (project costs before the split, a look at Sitges, the two projects compared).
Who flipping does not suit
Anyone without a reserve above the estimate: the first finding inside a wall becomes a loss. Anyone not ready for a "sale takes longer than a year" scenario. Anyone who counts the result as the difference between the purchase and sale prices. And anyone entering without a second and a third exit scenario (exit plans A, B and C, when a project has to be stopped).
Your own calculation can be checked line by line in the investor calculator, and the assumptions DNPI uses to calculate projects are set out on the methodology page.
Questions and answers
Which flipping mistake costs the most?
An exit price taken from listings. The cost estimate is refined before the transaction and the permit timeline follows from the approval regime, while the exit price becomes known last, and an error in it is not offset by anything.
What reserve should be built into a renovation estimate?
There is no universal standard, and I will not invent a percentage. The order of magnitude is suggested by the price source itself: the public portal states a spread of up to ±20%. The specific reserve is set from the technical survey.
Can profit be counted as sale minus purchase minus renovation?
No. That account skips ITP and paperwork at entry, holding costs over the term, exit costs, plusvalía and tax on the result.
Why is net profit not stated in the DNPI cases?
The published figures are the difference between the sale price and the costs listed. Taxes, financing and paperwork are not included in them, and calling such a difference profit would be wrong.
How is the result split between the investor and DNPI?
DNPI's fee is 50% of project profit; the other half stays with the investor. If the investor return for the whole project after the split is below the contractual 10%, DNPI pays compensation out of its own fee. This is a return for the project, not an annual rate; the terms and the limits are set by the contract for the specific project, and this is not a guarantee that capital will be preserved.
