For Swiss R&D-intensive companies, the effective tax burden can drop to as low as 9,1 percent, with significant differences between the cantons.

For a company that concentrates a significant part of its investments in research and development, the Swiss tax geography may be very different from that suggested by the ordinary tax rates alone. In 2026, according to the BAK Taxation Index, the reference measure of the average effective tax burden passes from 13,4 to 12,2 percent for a medium research-intensive company and goes down to 9,1 percent in the case of a company with very high R&D activity.

The figure requires an important clarification, however. That 13,4 percent is not a legal rate applied indiscriminately to Swiss companies, nor does it represent how much a specific company will actually pay. It is a Effective Average Tax Rate, EATR, i.e., a modeled indicator of the actual average burden on a given investment. The BAK Taxation Index takes into account the various taxes, their interactions, the rules governing the tax base and depreciation, and assumes, in the standard model, a manufacturing company with a pre-tax return of 20 percent.

The study was conducted by BAK Economics within the framework of the BAK Taxation Index, developed in collaboration with the ZEW – Leibniz-Zentrum für Europäische WirtschaftsforschungThe 2026 update was commissioned by the Federal Tax Administration and the tax administrations, economic agencies, and promotional bodies of twelve cantons. The calculations nevertheless cover all 26 Swiss cantons, although the report publishes individual values ​​only for those participating in the project, in addition to the minimum and maximum values ​​and the GDP-weighted national average.

Three scenarios separate the average search from the edge case

The analysis does not simply assign companies a label of greater or lesser innovation. BAK constructs three investment portfolios and modifies the weight of one internally developed patentIn the first scenario, the patent represents 20 percent of the investment, while the remaining 80 percent is divided between machinery, buildings, inventories, and financial assets: this is the case assumed to be representative of a company with medium research intensity. In the second scenario, the patent rises to 60 percent, making it a very R&D-intensive enterprise.

The third scenario brings the investment in the patent to 100 percent. BAK explicitly defines it as a theoretical limiting case: it should therefore not be interpreted as the normal profile of a laboratory, a startup or a pharmaceutical company. In this configuration, the weighted average Swiss EATR, after the application of the STAF tools, drops to8 percent.

There is also a methodological subtlety. The comparison between the 60 percent and 100 percent scenarios does not start from the same EATR without incentives, because the composition of the investment and therefore its tax treatment are different. For this reason, the absolute reduction in the extreme case is 3,1 percentage points, slightly lower than the 3,2 points for the 60 percent scenario, although the final EATR is lower.

This distinction avoids a too simple interpretation of the result. The BAK Taxation Index does not demonstrate that a given tax relief automatically generates more patents, skilled employment, or productivity. Instead, it measures how certain tax rules modify the net return of investments with different intellectual property componentsThe relationship between tax incentives and actual innovation requires further evaluation.

Tax relief weighs more heavily in high-tax cantons

The instruments analysed derive from the tax reform and financing of the AVS, the STAFF, which came into force in 2020. Among the cantonal levers considered are the patent box and the additional deduction for research and development expenses. The cantons can calibrate these instruments within the federal framework, and differences in their application explain a significant portion of the regional variations.

For a company with medium research intensity, incentives reduce the EATR in almost all cantons. The exception identified by the study is Glarus, where in this scenario no relief emerges. Among the participating cantons, the greatest effect is recorded in Zurich, with a decrease of 1,8 percentage points; considering all 26 cantons, the maximum reduction is 2 points. The weighted national average drops to 12,2 percent.

When the research component increases, the ranking changes more significantly. With an investment consisting of 60 percent of an internally developed patent, the average EATR drops to 9,1 percent, 3,2 percentage points less. In Zurich, the tax relief reaches 5,4 percentage points; the national maximum is 5,9. Solothurn and Schaffhausen also significantly improve their relative positions. However, the cantons of Central Switzerland, which already had low standard taxes to begin with, remain at the top of the ranking.

This is a significant result for the policy of attracting investmentsA canton that is relatively expensive from a fiscal perspective can regain competitiveness when the company applying for a location invests heavily in intellectual property and the local system makes extensive use of STAF instruments. However, this does not mean that incentives eliminate traditional fiscal geography. BAK notes that nine cantons grant R&D relief in the maximum configuration considered, but the group of lower tax jurisdictions retains the overall advantage.

The cantonal table makes the variety of choices visible. Ticino, Zurich, Zug, Schwyz, Solothurn, Obwalden, Bern, Aargau and Jura combine a patent box parameter of 90 percent, an additional R&D deduction of 50 percent and an overall maximum relief threshold of 70 percent. In theAppenzell Innerrhoden The three parameters are instead set at 50, 50, and 50 percent respectively. Basel City provides for a 90 percent patent box but no additional R&D deduction and limits the overall relief to 5 percent.

These percentages should not be confused with tax rates. The 90 percent patent box indicates the amount of relief applicable to qualifying patent income under the established rules; the 50 percent R&D rate is a additional deduction compared to costs already tax-deductible. Finally, the overall limit establishes how much the incentives can reduce the tax base before the cantonal brake kicks in.

A limit prevents incentives from canceling the levy

It is right there limitation of the tax reliefThe so-called "Entlastungsbegrenzung" (incentive to taxation), which particularly distinguishes the Swiss model analyzed by BAK, is designed to prevent the combination of innovation promotion tools from reducing taxable profits beyond a certain threshold. The mechanism thus creates a compromise between fiscal competitiveness, research incentives, and revenue protection.

With an average search intensity the limit only comes into play in two cantons. In the model with 60 percent of the investment attributed to the patent, it comes into operation instead eleven cantonsIn the extreme case of a 100 percent patent, it is activated in all except Uri: 25 out of 26 cantonsThis is also why the EATR does not continue to decrease proportionally to the increase in the patent component.

Recent developments in some cantons also show that competitiveness does not depend on a single lever. Compared to 2024, Basel City saw the EATR of the case with medium research intensity rise from 10,1 to 11,8 percent: BAK links the change to the combination of an increase in the maximum rate of the profit tax and a tightening of the limitation of the relief. Lucerne the opposite movement occurred, thanks to the reduction of the cantonal tax coefficient and the widening of the margin allowed for incentives.

The consequence is that simply comparing nominal tax rates can be misleading for a technology company. The location decision depends on the interaction between standard taxation, investment structure, share of internally generated intellectual property, patent box, R&D deductions, and the overall tax relief limit.

The minimum tax changes tax competition again

This cantonal architecture is now superimposed on the minimum international taxation of large multinational groupsFor companies affected by the OECD/G20 rules, a national incentive that reduces effective taxation may lose some of its value if it leads to a supplementary tax. The competitive advantage of simply reducing the tax burden becomes more difficult to maintain.

BAK, however, identifies a new room for maneuver in the package Side-by-side agreed within the OECD in 2026. The new framework allows, within conditions and limits, to consider certain qualified incentives linked to expenditure or production in a manner compatible with the minimum tax system. For activities such as research and development, this could shift the focus from purely income-dependent incentives towards instruments more closely linked to the economic substance of the investment.

A second transformation concerns the use of new tax revenues. According to BAK, Basel City, Graubünden, Lucerne, Schaffhausen, and Zug They have already introduced measures to support the economic center, including programs aimed at research and development. Part of the competition between territories is therefore shifting from the level of taxes to the ability to create favorable conditions for investment, skills, and innovative activities.

For the SwitzerlandThe result of the BAK Taxation Index 2026 is therefore not simply confirmation that research "pays less taxes." Rather, it shows how the taxation of innovative companies has become dependent on the type of investment and the decisions of the individual canton. A moderately research-intensive company obtains a relatively small advantage; when the weight of internally developed intellectual property increases, the gap can widen significantly.

Cantonal tax competition isn't disappearing: it's changing. Patent boxes, R&D deductions, limits on tax relief, and international minimum tax rules are added to standard rates. To understand which territory is truly more competitive for innovative investment, it's not enough to look at the nominal percentage of tax on profits. We need to reconstruct how value is generated, where research is carried out and which tax instruments can actually be applied.

“What do you know about Switzerland” (especially about innovation)?

EMPA, EAWAG, WSL, and PSI: Switzerland's four research institutes

CSEM, from applied research to industry: this is Swiss technological innovation.

Here are three insights that might interest you:

The journey into the future of food tech starts from the Canton of Schaffhausen
Why Switzerland attracts foreign AI, robotics, and deep tech firms
This is how Swiss research is preparing its future structure