Switzerland or Liechtenstein? The Right Architecture for Residence, Wealth, Banking, and Family.

1. Executive decision matrix

This is not a scorecard. Each row identifies a condition that must be verified before preference can matter.

Switzerland does not win merely because it is larger. Liechtenstein does not win because it is specialized. A divided design does not win because it contains more jurisdictions. Every option must first satisfy the same minimum conditions. A failed immigration route or frozen authority cannot be offset by a lower tax number.

2. Why the conventional country comparison uses the wrong unit

“Switzerland or Liechtenstein?” sounds like a choice between two bundled products. No such bundles exist. A family can reside in Switzerland, use a Liechtenstein foundation, sign with a Swiss booking entity, engage a Liechtenstein asset manager, and hold securities through subcustodians elsewhere. Another owner may reside in Liechtenstein, operate a Swiss company, and keep strategic liquidity outside both countries. Each function has its own actor, rule, contract, and failure mode.

Eight propositions must remain separate:

1. I am legally entitled to reside there.
2. I am tax resident there under domestic and treaty rules.
3. My company is incorporated or effectively managed there.
4. A bank accepts my personal and economic profile.
5. My assets are booked, managed, or held through a particular legal entity and custody chain.
6. I am permitted to acquire and use local real estate.
7. A company, foundation, or trust is validly organized there.
8. I and my family can exercise authority and obtain liquidity when an adverse event occurs.

One does not prove the next. An immigration permit is not a final tax-residence determination. Incorporation does not deliver a bank account. An opened account is not a promise of permanent servicing. A custody agreement may not reveal the complete subcustody chain at first glance. A foundation is not proof against a spouse, heir, creditor, or foreign tax authority. Property ownership does not, by itself, create residence rights in either country.

The proper analytical unit is function × owner × event. What function should each place perform? Who legally owns, controls, or owes the asset? What changes at relocation, closing, enhanced due diligence, incapacity, death, divorce, dispute, or exit?

Breaking the problem apart appears more complex than a country table. It is actually a simplification. It prevents one immigration decision from silently becoming an untested decision about tax, banking, family governance, and the entire balance sheet.

3. Switzerland is not one tax jurisdiction

“Swiss tax” is an incomplete term for an individual planning decision. Personal income tax operates through federal, cantonal, and municipal layers. Net wealth tax is cantonal rather than federal. Allowances, rates, burden caps, inheritance and gift taxes, and partial taxation of qualifying dividends differ across cantons. Municipal choice can change the result again.

The magnitude is visible in a controlled 2025 test using the official Federal Tax Administration calculator. The same inputs were entered for seven principal municipalities: single, age 45, no children, no church tax, CHF 500,000 of taxable income for federal and cantonal purposes, and CHF 5 million of taxable net wealth. Principal cities were used deliberately; the exercise did not search for the cheapest municipality in each canton.

The permissible conclusion is narrow but important: in this non-binding 2025 model, using the disclosed inputs and principal municipalities, the calculated totals ranged from CHF 114,455 in Zug to CHF 240,568 in Lausanne. It does not prove that Zug is always the lowest-tax canton for wealthy people or that Lausanne imposes a universal 48.11% rate. Change household status, income type, balance-sheet composition, debt, municipality, religion, or tax year and the result can move.

An owner may focus immediately on the CHF 126,000-plus annual spread. This creates the next analytical trap: a real tax difference becomes a relocation decision before immigration, housing, banks, company management, family, and property have been tested. The number is real within the scenario. Its decision authority is limited.

Inheritance and gift taxation also resist national shorthand. In tax year 2025, Switzerland did not impose a federal inheritance or gift tax, but cantons maintained their own regimes. Schwyz had no cantonal or municipal inheritance or gift tax. Zug exempted several close-family categories. Zurich exempted spouses and descendants but treated parents under a different allowance structure. Basel-Stadt, Vaud, Geneva, and Ticino each used their own exemptions and mechanics. The decedent's location, relationship, asset situs, and international connecting factors can matter. Federal Tax Administration: Inheritance and gift tax 2025

Qualifying dividends require a two-layer view. Under the 2025 reference rules, 70% of dividend income from a participation of at least 10% was included at federal level; cantonal inclusion had to be at least 50% and differed in practice. The corporate rate alone therefore does not answer an owner-operator's total burden. Company profit, salary, dividend, holding percentage, and shareholder residence must be modeled together. Federal Tax Administration: Partial taxation of participation income 2025

For tax period 2025, expenditure-based taxation was not a universal flat-rate arrangement either. It was available only to qualifying foreign individuals who became fully taxable in Switzerland for the first time or after an absence of at least ten years and did not work in Switzerland. Spouses living together had to both qualify. Zurich and Basel-Stadt did not offer the regime; Schwyz, Zug, Ticino, Vaud, and Geneva used different minimums and control calculations. Ordinary rates still applied to a specially determined base.

Decision rule: lock the profile, tax year, municipality, and definitions before comparing a single number. Only then may tax enter the broader decision.

4. Liechtenstein is small; its legal architecture is not simple

The usual shorthand—foundation, trust, and low tax—collapses five different regimes: personal residence, personal tax, the operating company, the wealth or governance structure, and land acquisition.

Liechtenstein personal taxation is not one flat percentage. Taxable net wealth is converted into a statutory 4% standardized assumed return, or *Sollertrag*, which enters the personal income-tax base (Liechtenstein *Erwerbssteuer*). That is not a 4% wealth tax. The progressive national tariff applies to the resulting base. The municipality then adds an annually determined surcharge to the national tax.

For 2025, the complete official table placed all eleven municipalities at either 150% or 160%. Vaduz, Schaan, Triesen, Triesenberg, Balzers, Gamprin, Planken, and Schellenberg used 150%; Eschen, Mauren, and Ruggell used 160%. If national tax is L, a 150% surcharge produces a total of 2.5 × L, and a 160% surcharge produces 2.6 × L. The statute permits a broader 150%–250% band. The 2025 table must not be relabeled as a 2026 fact.

The familiar 12.5% corporate rate applies to taxable net income, not automatically to every measure of economic profit and not as an all-in owner burden. Registered seat or effective management can establish nexus. Participation income may qualify for exclusions subject to statutory limits and anti-abuse conditions. The generally creditable minimum tax is CHF 1,800, subject to a narrow statutory exception.

A Private Asset Structure, or PVS, is not an active founder holding company. It must not conduct economic activity, and it may hold participations only if neither the structure nor its shareholders or beneficiaries exercise actual management influence over the investee. Status, eligible assets, investor class, governing documents, and real conduct all matter. A founder cannot direct the operating company and simultaneously describe the same arrangement as passive merely because the constitutional documents use passive language.

A private-benefit foundation generally requires minimum capital of CHF, EUR, or USD 30,000. A limited liability company requires 10,000 in those currencies and a stock corporation 50,000. These are capital requirements, not total formation costs. They say nothing about fiduciary, council, administration, audit, banking, tax work, or cross-border opinions. For complex families, the largest cost is often not formation; it is sustained governance and coordination.

A Liechtenstein trust is not a separate legal person. The trustee holds assets in the trustee's own name for beneficiaries or a purpose under the trust instrument. A foundation and trust therefore differ in title, organs, information rights, continuity, and foreign classification. Both may engage registration or deposit, UBO, AML, tax, and banking duties depending on their design. “Not publicly registered” never means “anonymous” or “unreported.”

An entrepreneur can form a sophisticated foundation and still discover that the personal residence route is unresolved, the PVS regime conflicts with retained operating control, or the home jurisdiction classifies the arrangement differently. The vehicle exists. The intended personal use case does not yet work. Liechtenstein planning should therefore begin with function, not form.

5. Residence: wealth is not a universal key

Switzerland and Liechtenstein differ most sharply in personal usability.

EU/EFTA nationals benefit from the Swiss free-movement framework. Routes for employed and self-employed activity depend on duration and facts. Economically inactive EU/EFTA nationals may have a path if they demonstrate sufficient resources and comprehensive health insurance. For employed third-country applicants, admission is limited, qualification- and labor-market-tested, and generally filed by the employer. Self-employment follows a separate route not modeled here. Neither route is created by wealth or incorporation alone. Swiss State Secretariat for Migration: Working in Switzerland

Liechtenstein's physical size has direct legal consequences. EEA national residence is quota- and procedure-constrained, with part of the allocation distributed by lottery. Official materials state that at least 28 permits for employed EEA nationals and eight for non-working EEA nationals are awarded through the lottery each year across two drawings. Those numbers are not acceptance probabilities and do not represent every permit category. Swiss and UK nationals are excluded from the EEA lottery under the official application page and follow other treaty or domestic routes.

For an employed EEA or Swiss route, Liechtenstein law examines the employment relationship, work percentage or self-employed conditions, and whether reasonable cross-border commuting is available. A non-working applicant must demonstrate resources and comprehensive insurance. Third-country work admission is selective and labor-market tested; a non-working third-country route requires a special Liechtenstein interest and remains discretionary. Company formation is not a founder-visa guarantee.

Property does not solve the immigration gate. Liechtenstein's land-acquisition law uses its own residence concept based on predominant actual presence and intent to remain; for a foreign national, less than ten years of continuous authorized stay generally does not count as residence for that specific land-law test. An EEA entitlement can give equal conditions without eliminating the approval and recognized-interest analysis. A foundation wrapper does not create a property workaround.

Hard gate: no personal move receives a GO without a credible residence and work route for the actual nationality, family, and activity—regardless of the tax spread.

6. Three model families—and why a tax table is not a recommendation

The following profiles are decision instruments, not tax opinions. They use disclosed assumptions so the arithmetic can be reproduced and challenged. Swiss amounts use the nonbinding 2025 Federal Tax Administration model for the seven named principal municipalities. The Liechtenstein sensitivity combines the SteG formula consolidated July 1, 2026 with the official 2025 municipal surcharge; it is neither a historical 2025 assessment nor a verified 2026 municipal total. Vaduz's published 2025 surcharge is 150%, so the mixed-reference sensitivity is 2.5 × national tax. Deductions, debt, church tax, special regimes, foreign credits, and treaty outcomes are excluded unless stated. The two country outputs are not a same-year ranking.

Profile A: the relocating operating family

• Married, age 45, two minor children, no church tax
• CHF 350,000 taxable income for the Swiss model
• CHF 5 million taxable net wealth
• Active founder; residence and work authorization are therefore first-order constraints

The 21 live calculator runs used age 45 for both spouses and, because the interface requires exact child ages, ages 8 and 12 for the two children. The model municipalities and results are separated in the table.

The CHF 73,822 range belongs to this disclosed dataset; it is not a general canton ranking.

In Liechtenstein, the simplified base combines CHF 350,000 of taxable personal income with a 4% standardized assumed return on CHF 5 million of taxable wealth, or CHF 200,000. The resulting CHF 550,000 base produces national tax of CHF 31,105 before municipal surcharge and CHF 77,762.50 in Vaduz. Across the statutory municipal band, the simplified total ranges from CHF 77,762.50 at 150% to CHF 108,867.50 at 250%.

That figure is not directly comparable to a Swiss total until the Swiss family deductions and the Liechtenstein exclusions, allowances, nationality, work route, and exact income character have been aligned. The profile's first output is therefore not a winner. It is a sequence: permit, activity, household, tax model, housing, and only then bank and company design.

Profile B: the liquid-wealth household

• Married, age 60, no dependent children, no church tax
• CHF 900,000 of economic portfolio income
• CHF 30 million net wealth
• No assumed local employment

Using taxable income as the Swiss calculator input, the identical 2025 profile produced the following ESTV totals:

The CHF 346,325 range is material, but the displayed burden share is total modeled income and wealth tax divided by taxable income. It is not a pure income-tax rate.

This profile exposes a decisive classification issue. Liechtenstein tax law excludes specified returns on assets already represented through the standardized return under Article 15(1)(a). Treating the full CHF 900,000 as taxable personal income as well would risk double counting. On the simplified exclusion branch, the 4% standardized return on CHF 30 million is CHF 1.2 million; national tax is CHF 83,105 and the Vaduz total is CHF 207,762.50. If the CHF 900,000 were instead fully taxable under the relevant facts, the proxy Vaduz total would be CHF 387,762.50. The CHF 180,000 gap is not a rounding issue. It is a legal-classification question. The branch closes only when Liechtenstein tax counsel maps each actual income stream and asset to Article 15 and the remaining personal-tax provisions in a dated written opinion; until then the output is HOLD, not a recommendation. Liechtenstein Tax Act (SteG)

A separate expenditure-based-taxation sensitivity can arise only on application and Tax Administration approval for a non-Liechtenstein national taking residence for the first time or after at least ten years away, performing no gainful activity in Liechtenstein, and living from wealth income or foreign receipts; Liechtenstein real estate remains separate. If accepted annual living expenditure were CHF 600,000, 25% would be CHF 150,000. That is a percentage of the expenditure base—not a 25% income-tax rate, residence entitlement, or selected branch in the public tool.

For this family, the bank-acceptance file, source of wealth, product restrictions, tax reporting, and residence route may dominate the municipal tax difference. A spreadsheet that silently chooses one classification has not solved the problem; it has hidden it.

Profile C: the succession and governance case

• CHF 100 million family net wealth: CHF 60 million operating-company value, CHF 25 million liquid investments, and CHF 15 million property
• CHF 50 million used only as the disclosed Swiss taxable-wealth sensitivity and as a separate possible transfer sensitivity
• Founder wants continuity but also continuing influence over an operating company
• Family members and assets may remain in more than one country

For the Swiss personal-tax sensitivity, both spouses are age 55, with CHF 1.5 million taxable income and CHF 50 million taxable wealth.

The CHF 577,251 spread is not the family's final liability: asset composition, debt, foreign-source rules, company tax, and distributions remain outside the calculator.

If a Liechtenstein structure receives a gratuitous dedication and the statutory conditions apply, a 3.5% dedication tax can be material: on CHF 50 million, the mechanical amount is CHF 1.75 million. Whether the entire amount is in scope, and how foreign jurisdictions treat the transfer and vehicle, requires case-specific analysis. A PVS can also conflict with the founder's desired operating influence. The governance problem cannot be solved by placing the word “passive” in documents while real control continues.

The family should therefore model the event chain before funding: formation, transfer, founder incapacity, beneficiary information request, divorce, creditor action, death, replacement of an office holder, bank exit, and migration. The most attractive annual tax line can be overwhelmed by one badly designed transfer or inaccessible account.

7. Banking and custody: the legal entity matters more than the flag

Private banking is not a national feature that a client buys wholesale. The contract is with a legal entity; the assets then follow specific deposit, custody, lien, subcustody, and resolution rules. A familiar group name can span several entities and countries. Booking location, relationship manager, portfolio manager, custodian, and ultimate subcustodian may be different actors.

Cash and securities require separate maps. In Switzerland, privileged deposits and the esisuisse system generally protect eligible deposits up to CHF 100,000 per client and bank, subject to the statutory and entity-specific rules. In Liechtenstein, the deposit-guarantee framework generally protects eligible deposits up to CHF 100,000 per person and bank. Liechtenstein materials describe a seven-working-day repayment target once the statutory payout event and required data are present, plus defined treatment for certain temporary high balances and small residual claims. These are not unconditional promises of immediate liquidity.

For Switzerland, esisuisse currently describes customer payout as taking several weeks; the present seven-working-day step concerns transfer to the liquidator, while the customer-level seven-day target is scheduled for 2028 after payment instructions. Neither system should be modeled as day-zero liquidity. esisuisse: Swiss deposit insurance · EAS Liechtenstein: Deposit protection

Securities held in custody are generally client assets rather than bank deposits. That distinction is valuable, but “segregated” is not the same as “instantly transferable under every event.” Title, omnibus arrangements, liens, securities lending, settlement status, fund gates, sanctions controls, and the subcustody chain can affect access. Liechtenstein's investor-compensation framework may cover eligible claims of nonprofessional investors up to CHF 30,000. It is not a blanket additional insurance layer on every portfolio loss.

Before onboarding, request written answers to the following:

A resilient design usually separates emergency cash, operating cash, and long-horizon custody. It may use more than one genuinely independent counterparty, but duplication inside one group or one subcustody chain is not true diversification.

The consequence is not theoretical. A family may have CHF 20 million in an intact custody account and still miss a CHF 4 million closing in five days because an enhanced review blocks the transfer, the nominal second custodian uses the same subcustodian, and only CHF 500,000 is immediately free. Ownership survives; negotiating power does not. The test therefore runs at 24 hours, five days, and 30 days.

8. Companies, foundations, and trusts: function before legal form

An operating company should be located where management, people, customers, financing, governance, and banking can actually work. Switzerland offers cantonal choice and a broader operating ecosystem. Liechtenstein offers EEA connectivity and a 12.5% ordinary tax rate on taxable net income, but the rate does not neutralize effective-management, substance, withholding, shareholder, or foreign anti-deferral rules.

A family-governance vehicle serves a different purpose. A Liechtenstein foundation is a legal person with foundation assets and organs. A Liechtenstein trust is a legal relationship in which the trustee holds title. The choice changes who owns, who controls, who receives information, how succession operates, and how foreign courts and tax authorities may classify the arrangement.

Swiss family foundations have a narrow statutory purpose. Switzerland may recognize foreign trusts under the Hague Trust Convention and conflict rules, but recognition of the legal relationship does not automatically answer tax attribution, forced-heirship, matrimonial, creditor, or reporting questions. The same is true in reverse for a Liechtenstein vehicle connected to a Swiss resident.

The correct question is not “Which vehicle protects assets best?” It is: what lawful governance task must continue, under which governing law, with which real decision makers, and what happens when another jurisdiction applies its own mandatory rules?

9. The family-event test

Structure diagrams are normally drawn on a calm day. Families discover their quality on a bad one. Each proposed architecture should be tested against at least these events:

Assume the founder is suddenly incapacitated on a morning when payroll and a property-completion payment are due. The family holds the assets, but the bank does not accept the domestic power of attorney, the foundation council requires two signatures, and the second signatory is traveling. The legal structure still exists. The authority chain does not execute. That is why the event test below asks who can instruct each bank and body, under which document, and with which fallback.

1. Founder incapacity without a valid bank mandate.
2. Sudden death before ownership and beneficiary records are aligned.
3. Divorce in a jurisdiction different from the vehicle's governing law.
4. A forced-heirship or compulsory-share claim.
5. A creditor challenge to an earlier transfer.
6. A tax-residence change by founder, beneficiary, or decision maker.
7. Replacement, resignation, or deadlock of trustees or foundation council members.
8. A beneficiary information demand.
9. A bank exit or freeze during enhanced due diligence.
10. A sanction, nationality, or residency change affecting serviceability.
11. A sale of the operating company.
12. A desire to distribute, lend, or pledge assets.
13. A dispute over an investment or protector direction.
14. Termination, migration, or redomiciliation of the structure.

There is no universal two-year safe harbor that makes a transfer immune. Avoidance, clawback, forced-heirship, matrimonial, insolvency, and criminal-law rules use different tests and connecting factors. A lawful structure can create continuity and disciplined authority; it cannot honestly be marketed as divorce-proof, creditor-proof, tax-proof, or sanction-proof.

10. Transparency: private is not invisible

Both jurisdictions operate inside modern AML, tax-cooperation, and beneficial-ownership systems. Banks, fiduciaries, corporate-service providers, auditors, and authorities may require extensive information even when a public register exposes less.

The useful distinction is between public accessibility and regulated disclosure. A structure may not publish every beneficiary or internal document to the world while still being fully reportable to the relevant intermediaries and authorities. Automatic exchange of information, AML source-of-funds and source-of-wealth review, sanctions screening, and tax filings remain separate layers.

This is operationally positive when designed well: consistent records shorten reviews. A family should maintain a living evidence pack containing ownership, control, tax residence, wealth origin, transaction purpose, audited or reliable financial data, governance decisions, and powers. Contradictory narratives across bank, tax, immigration, and structure files are a preventable risk.

11. When Switzerland wins, when Liechtenstein wins—and when neither does

The split architecture is often the most interesting and the most abused. It works only when every additional entity, adviser, account, and jurisdiction has a defined purpose and owner. If the design cannot explain what risk a component reduces and how that improvement will be verified, the component is decoration.

12. Five-year total cost of ownership

Headline tax is only one row. Use a five-year model:

\[ TCO_5 = 5(T_p + T_c + B + G + A + H) + F + M + E + R \]

where T_p is annual personal tax, T_c recurring company or structure tax, B banking and custody cost, G governance cost, A administration/advice, H incremental housing and travel cost, F formation and funding cost, M move cost, E expected event or exit cost, and R a risk reserve for items that cannot responsibly be expressed as a point estimate.

The formula is deliberately simple. The important controls are not hidden mathematical sophistication but input provenance and scenario ranges. Use low, base, and high values; separate one-time from recurring amounts; state currency and tax year; and do not assign a false probability to an unresolved permit or legal classification. A failed hard gate is not a large cost. It is NO-GO.

13. Embedded Decision Lab

Run the decision in this order:

1. Admissibility: nationality, family, activity, permit, work right, and property restrictions.
2. Tax characterization: residence, income type, wealth base, company management, transfer, and foreign classification.
3. Operational fit: people, customers, housing, travel, schools, advisers, bank acceptance, and evidence burden.
4. Custody resilience: legal entities, deposits, title, subcustody, liens, transfer, emergency liquidity, and substitute provider.
5. Family execution: authority after incapacity, death, divorce, dispute, relocation, and beneficiary change.
6. Five-year economics: comparable tax year and profile, all recurring and one-time costs, range rather than one magic number.
7. Reversibility: time, friction, tax, consent, and documentation needed to unwind or migrate.

The public Decision Lab outputs only HOLD or REVIEW-READY; it never issues GO, ranks the countries, or averages gates into a score. Blank costs, Profile B's unresolved Article 15 branch, and Profile C's non-comparable Liechtenstein base remain HOLD. Even REVIEW-READY means only that inputs were marked complete: evidence, confirmer, date, expiry, and responsible professional approval still sit outside the public tool. In the professional implementation, a NO-GO on lawful residence cannot be averaged away by good banking.

14. The correct implementation sequence

This order prevents the expensive error of forming and funding first and discovering later that the founder cannot live, work, control, report, or bank as assumed.

15. Final judgment

Switzerland is usually the broader operating and residence platform. Liechtenstein is usually the narrower specialist jurisdiction. That distinction is more useful than declaring one country superior.

The strongest result may be Switzerland-led, Liechtenstein-led, split, or neither. The answer changes with nationality, activity, family, wealth composition, control, income character, banks, and foreign-law connections. The disciplined decision maker therefore does three things: refuses incomparable tax numbers, maps every legal entity and authority, and tests the structure on the day something goes wrong.

The goal is not to collect jurisdictions. It is to create lawful optionality that remains executable under scrutiny.

Method and evidence boundary

This analysis uses official 2025 tax calculators and tables where stated, primary legal and regulatory materials, and entity-specific questions where public evidence cannot close the issue. It is strategic information, not individual legal, tax, investment, or immigration advice. Calculations are simplified, non-binding, and exclude facts not expressly stated. 2025 values are not described as 2026 rates. Unresolved foreign classification, bank acceptance, subcustody, permit, marital, succession, creditor, and treaty outcomes remain professional-advice gates rather than inferred answers.