In August a friend sent me an appraisal. A warehouse of about a thousand square metres in an industrial area, the owner selling, a potential rent above €40,000 a year. He and I looked at it with the same eager eyes: bricks, rent, yield. But I opened a spreadsheet to build myself a business case, and…
I came out with a contentious answer and a method I liked very much. I’m sharing it because the question belongs to everyone: an income property or a government bond? The short answer is that the question is wrong. The long answer is that the hidden costs, the ones nobody usually counts, and not being able to sell in a day, all weigh, and all have a number.
How much of a €42,000 rent is actually left?
The market rent, verified against comparable listings in the area, was €42,000 a year: on a €360,000 price that’s 11.7% gross, the number that seduces everyone. Then the waterfall begins.
Subtract expected vacancy (9%, eleven empty months over a decade), subtract the expected loss from a tenant defaulting, and you’re at €37,426 of effective revenue. Subtract IMU, the Italian property tax, which between the industrial building and its buildable land runs €12,000 a year and alone eats 29% of the rent. Subtract insurance including the now-mandatory catastrophe cover, maintenance, management, the registration tax on the lease. Operating income: €9,739. Then corporate taxes, and what lands in the company’s account is €7,668.
Eighteen percent of the figure written on the lease. The rest evaporates along the way, and not one item is optional.
What does it cost to get in and out?
Buying a commercial property in Italy costs 7% on top of the price, between transfer taxes, notary and due diligence: €25,400. Selling it costs up to 8% of value, between the negotiation discount and the agent’s fee: a little over €35,000. Over ten years that friction alone is worth about 1.6 points of annual return. Before we’ve even mentioned tenants, roofs or floods. And at the sale there’s a sting: the depreciation you deducted has lowered the tax basis, so a capital gain is taxed even if you sold below what you paid.
What does a government bond return to a company?
On 10 August 2026, when I ran the numbers, the 2036 BTP yielded 3.96% gross. On 16 September it yields 4.41%: forty-five basis points in five weeks. An Italian company doesn’t get the 12.5% flat rate individuals enjoy: it pays full corporate tax at 24%, plus a stamp duty on the securities account. With coupons reinvested prudently, the net was 2.77% in August and 3.07% on 16 September: from here on I use that one. No tenants, no capex, no Saturday mornings on the phone.
The comparison is on equal terms: same capital, ten years, everything net, exit included. At a €360,000 purchase price the warehouse makes 0.94%, the bond 3.07%. It isn’t a contest. And the bond isn’t risk-free, I’ve stress-tested it myself: it’s a different risk, one that doesn’t call on a Saturday.
What does not being able to sell in a day cost?
But a property shouldn’t match the bond: it should beat it by a margin that pays for everything the bond doesn’t have. I tried to break that margin down, instead of saying “bricks yield more” the way people do at the bar, and to write it in euros, because in euros it hurts more.
Illiquidity is worth 1 point: selling a warehouse takes eight to twelve months, and anything can happen in the meantime. On €385,400 of capital (the €360,000 price plus notary, land registry and the rest) that’s €3,854 a year, almost €40,000 over the decade, just for not being able to get out in a day. The other risks make 2 points: rent that arrives only if the tenant is alive and solvent, management time that isn’t free, the roof to redo or extraordinary works and unpredictable events (here in Romagna we know something about that, between floods and tornadoes).
Total: 3 percentage points, €11,562 a year. The warehouse has to return 6.07% net to deserve the capital.
Every hidden cost, with its number
| What the appraisal doesn’t write | Euros a year | Weight |
|---|---|---|
| Expected vacancy and a defaulting tenant | 4,574 | 11% of the rent |
| Property tax (IMU) | 12,000 | 29% of the rent |
| Insurance, maintenance, management, registration | 15,687 | 37% of the rent |
| Corporate taxes (IRES and IRAP) | 2,071 | 5% of the rent |
| Extraordinary works (65,000 over the decade) | 6,500 | 1.7 points |
| Getting in at closing and out at the table (60,725) | 6,073 | 1.6 points |
| Tax on the capital gain at the sale (26,440) | 2,644 | 0.7 points |
| Not being able to sell in a day | 3,854 | 1 point |
| The other four risks | 7,708 | 2 points |
The first four rows leave the account every year and add up to €34,332, 82% of the rent; the other five leave once, or never, spread here over ten years as a simple average, and weigh almost seven points. The first seven have a cell in the model: some with a source and a date, others with an estimate declared within its range (cadastral income, insurance premium, vacancy, works pending a survey). The last two are the premium I demand, item by item.
At what price does the property beat the bond?
Here a property’s return isn’t a property of the property: it’s a function of the price you pay. With September’s bond, at €274,000 the warehouse matches the government bond. At €191,000 it beats it with the full premium. At €360,000, the price under discussion, it returns less than a third. And the exit is already generous: at the €450,000 market value, with the purchase discount recovered and taxed.
I also ran the perfect world: zero vacancy, a tenant with a bank guarantee, property tax at the minimum, no works, appreciation of 1.5% a year. In that world, at €360,000, the warehouse returns 5.78%. In August that was, within a hundredth of a point, the bond plus the premium: the price on the table was the right price for the perfect property, and nobody buys the perfect property. In September even that isn’t enough, because the bar has risen to 6.07.
Redoing the math in September: in five weeks the warehouse hasn’t changed by a single brick, but its break-even price has dropped from €285,000 to €274,000, because the alternative went up. Those rooting for bricks will say rates rise with inflation and the rent is indexed. Half true: at 75% of the index, and older Italian warehouses have been flat or falling for years.
My mistake, before his
I do due diligence for a living. The first version of my model had a €60,000 rent and a 10% net target. The market said 42,000 and the math said 0.94. I was in love with the warehouse before doing the numbers, like the clients I lecture about it. Bricks you can touch, bonds you can’t, and the Italian brain has a genetic weakness for things it can touch. How I built the model with AI is in the case on the AI page.
So, does a warehouse beat a government bond? Yes, if you pay the price that discounts all its flaws. No, if you pay the price that assumes it has none. Between the two answers there were more than a hundred and sixty thousand euros, and no seller gives them away: you find them only by doing the math to the end, everything net, as with net debt when you buy a company, and then waiting.
Because the real surprise, in the end, was that waiting cost nothing: the capital parked in bonds earned almost 3% net while the sale ran its course. For once, time was on the side of whoever had done the homework. And it has never been this easy to do it.
And all of this on the good price: the owner was asking €500,000.
The case study (PDF, 9 pages, in Italian) and the ten-year model (Excel, 1,151 formulas, in Italian) are downloadable, anonymised, frozen at 10 August: for the day’s numbers just change the bond-yield cell. If you open them and find a mistake, write to me.
Sources: BTP 1 July 2036, 3.80% coupon (IT0005706285), gross yield to maturity as published by Borsa Italiana on 10 August 2026 (3.96%, price 99.22) and 16 September 2026 (4.41%, on the 95.52 reference price; last 95.86); BTP-Bund spread at 89 basis points, QuiFinanza, 16 September 2026; comparable rental listings in the area, August 2026; Agenzia delle Entrate, OMI H2 2025; Colliers, logistics Q4 2025 (vacancy); ABI-Cerved, outlook 2025-27 (default risk); Law 213/2023 and Ministerial Decree 18/2025 (catastrophe cover); Law 199/2025, 2026 Budget Law, art. 1(42-43), rewriting art. 86(4) TUIR (capital-gain instalments). IRAP is computed without the flat deduction of art. 11(4-bis) of Legislative Decree 446/97, worth a few hundred euros a year in the building’s favour. This is not financial or tax advice: it is a single case with its assumptions stated, and the rates need checking with your own accountant on your own building.
