EMI in Lithuania or Bulgaria: A Comparison

Lithuania and Bulgaria compete for the same application. Both issue an electronic money institution licence under the same European directive, both passport it across the European Economic Area, and since 1 January 2026 both do it in euro. The founder who lines the two up normally starts with capital, then the statutory timeline, then currency risk — and all three come back identical. Initial capital for an EMI is €350,000 in each, because article 4 of EMD2 fixes it. The statutory decision period is three months in each, because article 12 of PSD2 governs e-money authorisations through article 3(1) of EMD2. And currency stopped being a variable when Bulgaria adopted the euro on 1 January 2026 as the twenty-first member state at a fixed rate of 1.95583 leva, and the Bulgarian National Bank joined the Eurosystem on the same day.

What is left is everything the two directives hand back to the member state: the timetable the regulator sets for itself, the price it charges to look at a file, the cost of staff and the tax on profit, the rails the licence physically reaches, the depth of the local register and supplier market, and how a correspondent bank reads the country on the letterhead. Those are the axes on which this choice is actually decided.

What rules one of them out before the comparison starts

Four conditions settle the file without weighing merits, and it is worth testing them first because each removes one jurisdiction outright — in one case, both.

A product that needs to be licensed but does not need a passport. Lithuania issues a restricted-activity EMI licence with no minimum initial capital at all, valid solely inside Lithuania and capped at €900,000 of average outstanding e-money — its use of the small-institution waiver in Art. 9 of EMD2. Bulgaria has no equivalent: ZPUPS does not use that waiver, so every applicant, however small, goes through the full licence. A domestic pilot with modest volumes has one route into a licence, and it runs through Vilnius.

An applicant that cannot place real management in the country. This test removes both, because it is set by EU law rather than national taste: PSD2 Art. 11(3), which EMD2 Art. 3(1) applies to e-money institutions, requires the head office to sit in the same member state as the registered office and at least part of the payment-service business to be carried on there. Bulgaria writes the test into the ZPUPS licensing conditions: the registered seat and management address must coincide with the place where management is actually carried out. Lithuania requires the EMI to be a public or private limited company (AB or UAB) with its registered office in Lithuania. A letterbox licensee is available in neither.

A product with a token leg. In Lithuania the same authority licenses the EMI and the crypto-asset service provider, so an e-money token issuer and its exchange or custody arm answer to one regulator. In Bulgaria they do not: the BNB licenses the EMI and supervises e-money token issuers, while the Financial Supervision Commission is the MiCA competent authority that licenses crypto-asset service providers and receives every MiCA notification, including an EMI's. Two regulators is not a disqualification, but it is a second authorisation project, a second set of counsel and a second timetable.

A counterparty set that prices jurisdictional risk mechanically. Bulgaria remains on the FATF list of jurisdictions under increased monitoring, listed since October 2023. At its June 2026 plenary the FATF made the initial determination that Bulgaria's action plan is substantially complete, which triggers an on-site assessment before delisting. Until that concludes, any counterparty whose onboarding policy reads the list rather than the plenary language treats a Bulgarian licensee as enhanced-diligence by default. What the listing does and does not mean is set out in the FATF explainer.

The two regimes side by side

Entry conditions first. The rows where the two columns agree are the ones no shortlist should be built on.

Operating conditions are where the two separate.

What the grid actually decides

Capital carries no information here. Both regulators ask for €350,000 because a directive tells them to, and in both the figure is a floor beneath an own-funds calculation rather than the cost of entry. Bulgaria's own-funds rule is written out in ZPUPS — 2% of average outstanding e-money, plus the payment-volume scale for services unrelated to e-money, and never below the €350,000 of Art. 38. Lithuania reaches the same place through methods A to D. How those methods are built, and how the resulting figure compares across licence types, is set out in regulatory capital. A shortlist built on the capital line is a shortlist built on a constant.

The clock is where the two texts diverge while both say "three months". The Bank of Lithuania publishes what it undertakes: a decision within three months of proper and sufficiently informative documents, against a state fee of €1,463, or €1,235 for the restricted route. ZPUPS Art. 11 builds its period differently: the BNB decides within three months of receipt; if the file is incomplete the applicant is given up to two months to supply what is missing; and the BNB then has a fresh three months from the day the missing documents arrive. Its review fee is €5,000 for an EMI and €4,000 for a payment institution under Art. 70 of Ordinance No. 16 — roughly three and a half times the Lithuanian fee in the cheaper country.

Settlement rails are the sharpest operational difference, and the gap is narrower in law than it looks. Lithuania runs CENTROlink, a payment system operated by the central bank that gives EEA-licensed banks, EMIs and payment institutions direct access to SEPA credit transfers, instant transfers and core direct debits. Its page, updated on 20 May 2026, puts the community at more than 200 financial institutions from over 20 countries and counts more than 140 participants actively using the system — a figure that includes direct participants, the BIC holders reached through them and participants' branches.

Bulgaria has no comparable standing service for non-banks. Since 1 January 2026 the BNB has settled on the full TARGET Services, the national RINGS provisions of ZPUPS having been repealed from that date, and the BISERA rules admit BNB-licensed EMIs as participants. A licensed PI or EMI that wants into a settlement-finality system applies to the BNB under ZPUPS Art. 130a, which rules within two months of a complete application. Across the euro area the underlying right arrived at the same time for both countries: Regulation 2024/886 inserted Art. 35a into PSD2 and brought payment and e-money institutions within the Settlement Finality Directive, and after ECB Decision 2025/1148 deferred the start, non-bank payment service providers have been able to request TARGET access, including T2 and TIPS, since 6 October 2025.

Register depth decides the things that never appear in a comparison table. As of 24 September 2026 the Bank of Lithuania lists about 70 electronic money institutions with an unrestricted licence and two more with a restricted one. The BNB registers list 13 EMIs and five payment institutions on the same date. A market of about seventy carries counsel who have filed the application before, compliance officers who can be hired rather than trained, auditors who know the own-funds return, BIN sponsors and safeguarding banks that already price the licence type, and a supervisor with precedent for most questions. A market of thirteen carries none of that depth — and gives every applicant the individual attention of a regulator that has time to read the file closely, which helps a serious applicant and removes any hope of passing unnoticed.

Reputation is the last axis and the hardest to price. Lithuania's supervisory record is dense and public. In 2024 alone the Bank of Lithuania revoked nine licences and imposed 19 sanctions in a sector that numbered 76 EMIs and 43 payment institutions at the year's end. Named cases include the licence of PayrNet revoked in June 2023 after some ninety intermediaries were onboarded without proper checks, Foxpay in November 2024 and Paytend Europe with effect from 3 March 2026, the last two for serious AML failures. That record is simultaneously the discount a Vilnius letterhead carries and the evidence that the supervisor functions. Bulgaria's register is smaller and, at the margin, younger — three of its thirteen EMIs were licensed in 2026 — so there is less track record to read, and a correspondent bank's committee tends to treat that as an open question rather than a point in favour.

Choosing by profile

The licence is the same instrument in both countries, so the profile of the business, not the regime, decides which discretion matters.

A start-up wallet with a domestic pilot has an asymmetric choice. The Lithuanian restricted licence removes the capital floor entirely and costs €1,235 to file, at the price of a licence valid only in Lithuania and capped at €900,000 of average outstanding e-money, beyond which the institution has 30 days to apply for an unrestricted licence. Bulgaria offers nothing between "unlicensed" and "€350,000 plus a €5,000 review fee", which makes it the wrong place to test a product.

A B2B payments and collections business lives or dies on the account chain behind the licence, and this is where CENTROlink earns its reputation: the SEPA leg does not depend on a commercial bank's willingness to keep the relationship. The Bulgarian equivalent has to be assembled, first through a safeguarding bank and then, if direct settlement matters, through an Art. 130a application. ZPUPS Art. 45 helps at the first step: banks in Bulgaria must open and maintain EMI accounts on an objective, non-discriminatory and proportionate basis and report every refusal to the BNB with reasons. Bulgaria becomes the right answer when the customer base is domestic or regional, where a Sofia licence reads as local rather than as a passport of convenience. How the chain of accounts behind a licensed payment firm actually works is set out in correspondent banking and safeguarding and banking for licensed operators.

A card-issuing programme is the one profile where the jurisdiction matters least, because a BIN sponsor and scheme diligence set the calendar rather than the regulator.

A crypto on- and off-ramp needs a fiat leg and a token leg, and in the EU that means an EMI alongside a MiCA CASP authorisation. In Lithuania both come from the Bank of Lithuania. In Bulgaria the EMI and e-money token supervision sits with the BNB while crypto-asset services go to the Financial Supervision Commission, so the project carries two regulators, two filings and two timetables. A narrower route exists in both countries: an EMI may add custody and transfer of its own e-money tokens by notifying its competent authority at least 40 working days ahead under MiCA Art. 60(4) — in Bulgaria the FSC — while any wider crypto-asset service needs CASP authorisation. Either route also brings the institution within the EU sanctions rule on who may own a crypto-asset service provider. The CASP route itself is covered in the MiCA CASP licence guide, the token-issuer rules in issuing ARTs and EMTs, and the wider token map in the map of crypto licences.

Cost and time

Published figures cover only the regulator's own price list. Everything else — local team, counsel, the compliance and ICT stack, the audited accounts, the safeguarding bank — is priced by the market rather than by either authority.

The arithmetic runs against the intuition that a cheaper country produces a cheaper licence. The BNB charges more than three times the Lithuanian review fee, and both regimes require the head office and part of the business in the licensing state, so substance is a fixed cost in either. Where Bulgaria is genuinely cheaper is after authorisation: Eurostat's 2025 labour-cost data put an hour in finance and insurance at €16.1 against €28.7 in Lithuania; profit is taxed at 10% against 17%; and a dividend paid to a foreign company bears 5% Bulgarian withholding against 17% in Lithuania, the latter before the participation exemption or treaty relief. Those savings are real but recurring rather than visible at the filing.

Where the choice goes wrong

Six failures account for most of the wasted filings.

Comparing the capital line. It is the same number in both countries and in every other EEA state, set by the directive. A comparison that turns on it has not started yet.

Reading three months as a delivery date. In both countries the period is the regulator's maximum for a decision on a file it treats as complete, and in Bulgaria the period restarts once. Neither figure is an estimate of how long a real application takes.

Quoting leva for Bulgaria. Since 1 January 2026 Art. 38 of ZPUPS states the initial capital in euro — €350,000 — and material that still quotes a leva figure predates the changeover.

Treating euro-area entry as a change to the licence. It changed the settlement environment and the currency in which the thresholds are expressed. It did not alter the authorisation, the documents, the safeguarding duty or the passport, all of which came from PSD2 and EMD2 and were already in force.

Assuming the licence opens central-bank settlement. It does not, in either country. Lithuania has a service to join; Bulgaria has an application to file under Art. 130a with continuing conditions attached. In both, the safeguarding account still sits at a commercial bank.

Assuming agents can carry the product. In both countries an EMI may distribute and redeem e-money through persons acting on its behalf and provide payment services through agents, but Art. 3(5) of EMD2 forbids issuing e-money through agents. Where an agent network is the whole business model, the structure has to be checked against that limit before the jurisdiction is chosen — the mechanics are in payment agents and passporting, and the case for not holding a licence at all is in business under someone else's regulation.

One shared horizon closes the comparison. As proposed by the Commission, PSD3 would repeal EMD2, fold the EMI regime into the payment institution regime and have existing EMIs reassessed and authorised as payment institutions. Both files remain unadopted, awaiting the Council's first-reading position, so each country's current statute continues to apply — and when the reform lands it will land on both at once, which is one more reason the choice cannot be made on the regime text. The expected changes are analysed in PSD3 and the PSR; the wider licence landscape sits in the fintech hub.

The EU sanctions perimeter

EU sanctions against Russia never decide between Vilnius and Sofia, because they sit in a regulation that applies identically in both. What they decide is the customer base and, for a product with a crypto leg, who may own and run the institution. Three provisions of Regulation 833/2014 carry the logic.

The service ban is narrower than it sounds. The Commission's guidance confirms that executing payment transactions is not covered, that an EU or third-country company owned or controlled by Russians remains an eligible customer unless it is used to circumvent the ban, and that existing accounts need not be closed, although the prohibited services must stop.

The ownership rule reaches only the crypto side, but an EMI crosses into it easily. Custody or transfer of its own e-money tokens under MiCA Art. 60(4) is a crypto-asset service, so an EMI that adds it — or holds a CASP licence alongside the e-money licence — may not have Russian nationals or residents of Russia among its owners, controllers or board members unless they fall within the Art. 5b(3) exemption. A pure e-money business is outside the rule; there the ownership filter is the qualifying-holding assessment, in which either regulator can oppose an acquirer whose influence would harm sound and prudent management. The wider route map is in sanctions.