Building a calculation model forces you to answer questions that conversation tends to sidestep. And it puts a number where, as a rule, impressions circulate.
Over recent months we built a tool that compares the tax burden of the same business in Italy and in San Marino, year by year. It was not calibrated on assumptions: it reproduces, to the cent, completed transactions from our own files, and before it was allowed to work it had to pass more than five hundred checks.
The benefit we expected was to have numbers ready. What arrived was different: building a model forces you to answer questions that conversation tends to sidestep. You cannot program an "it depends". You have to say what it depends on, and by how much.
Four answers deserve to be shared, because they disprove things that are repeated with confidence about these transactions — and that, without the figures in hand, are easy to take on trust.
In the usual account, dividends flowing up from Italy to a San Marino holding company bear no withholding tax, provided the treaty requirements are met. That is correct — but not from day one.
The treaty, in the text as amended in 2012, requires the shareholding to have been held for at least twelve months at the date of the distribution resolution. A holding company incorporated for the transaction does not have those twelve months when the resolution on the first financial year is passed: that distribution falls into the residual case and bears fifteen per cent.
It is not a permanent obstacle, and it is purely a matter of calendar. But it falls in the year in which the structure costs most and yields least, and in our calculations, with a shareholder who remains tax resident in Italy, the first financial year requires between one hundred and two hundred and fifty thousand euros of additional profit compared with the five-year average for the transaction to stand up.
Anyone presenting the saving from year one is showing a number that will not materialise.
This is the most counterintuitive result, and the one we checked most often before accepting it.
The scenario everyone imagines first is the simplest: the business stays in Italy, the shareholding moves under a San Marino holding company, and only the route of the dividend changes. With a shareholder who remains tax resident in Italy, this scenario does not break even at any level of profit. We checked up to five million.
The reason lies in the last step, from the company to the individual. A shareholder resident in Italy pays a final twenty-six per cent tax on foreign dividends, with no credit for what has already been paid upstream. The San Marino route therefore adds the San Marino levy to the Italian one, against a flat twenty-six per cent on the domestic route. No increase in profit overturns arithmetic of this kind.
A conclusion worth keeping follows: a San Marino holding company serves to retain profits, not to distribute them. We also checked the comparison with the obvious alternative — an Italian holding company — on the upward step alone: the San Marino advantage exists, but it is a fraction of a point. It does not repay the costs of a foreign structure, if that is the only objective.
Where the San Marino choice may have different reasons — the holding company's own income, capital gains, exit — those are assessments a model of this kind does not price, and we make no claim to settle them with a calculation.
The question we receive most often is what the size threshold is: from what turnover it becomes worthwhile. It is the wrong question, for two reasons.
The first is that turnover has nothing to do with it: the tax advantage is measured on profit, not on revenue. A company with ten million in turnover and fifty thousand euros of profit does not have a tax problem — it has a margin problem, and relocating it does not solve that.
The second is that, at equal profit, what is moved matters far more. In our calculations, moving the business with a genuinely new operation halves the profit needed to break even compared with the same business simply continued: the difference is in the order of more than three hundred thousand euros of required profit. In threshold terms, that one condition is worth as much as the entire structure.
And a figure we keep in mind whenever we discuss it: each employee moves the break-even point by roughly two hundred and eighty thousand euros of profit. Not because San Marino staff cost more than Italian staff — quite the opposite. But because it is a fixed cost that must be covered before the tax advantage begins to count, and the structure always weighs on the new side.
On this we are explicit, because the numbers are.
In our calculations, the individual's tax residence weighs more than any choice about the corporate form. Break-even thresholds change by an order of magnitude depending on whether the shareholder remains tax rooted in Italy or not, and the scenario in point 2 — the one that never breaks even — changes in nature.
This is not a recommendation: it is the observation that the decisive variable sits upstream of the corporate choices, and that assessing it means examining a concrete personal situation, with its family, professional and financial constraints. It is not a matter for a model, and it is the first thing to put on the table when that is the declared objective.
Three of these four answers are less favourable than the version in circulation. The fourth shifts the discussion onto ground many prefer not to address at the outset.
These numbers do not catch us by surprise: we knew the direction from experience. The difference is having measured it — being able to put it on the table beforehand, not once the structure has already been set up. This is where you see who approaches these transactions with the right tools and who works on impressions.
We write them for a practical reason: anyone who notices these points after setting up a structure has spent money badly. We would rather they discovered them first, even when the answer is that the transaction is not worthwhile.
And because the risk in our profession is not saying no to someone who could have made it work: it is saying yes to someone who could not.
If you are weighing a transaction of this kind, a confidential conversation commits you to nothing.
Get in touchThis note presents calculation results on stated assumptions and does not constitute advice on a specific situation. Every transaction has elements of its own that alter its conclusions, and the figures cited vary as the assumptions vary.
Nicola Ciavatta
United Consulting S.r.l. — Republic of San Marino
