Research

How we built these figures

Every number on these pages comes from the MFSA's quarterly Authorised Persons Report, Q2 2024 to Q1 2026.

8 quarterly MFSA reports, Q2 2024 to Q1 2026

The Malta Financial Services Authority's Authorised Persons Reports. Q1 2026 was the latest the MFSA had published on the retrieval date.

MFSA Authorised Persons Reports used
PublisherReportEditionRetrievedHow we read itRowsPrinted totalRows add up to
MFSAAuthorised Persons ReportQ2 202410 November 2026Typed from the page image572,3912,394
MFSAAuthorised Persons ReportQ3 202410 November 2026Typed from the page image572,3592,374
MFSAAuthorised Persons ReportQ4 202410 November 2026Typed from the page image582,3802,380
MFSAAuthorised Persons ReportQ1 202510 November 2026Text layer622,4022,402
MFSAAuthorised Persons ReportQ2 202510 November 2026Text layer622,3742,374
MFSAAuthorised Persons ReportQ3 202510 November 2026Text layer622,4032,403
MFSAAuthorised Persons ReportQ4 202510 November 2026Text layer622,4132,413
MFSAAuthorised Persons ReportQ1 202610 November 2026Text layer622,4622,462
Each report is a 4-page PDF with the table on pages 2 to 4. Each row in our data keeps the report's URL, quarter, page and retrieval date.

Every count is of authorisations, and one firm can hold several

Active authorisation
An authorisation the MFSA lists as active at the end of the quarter. The report counts authorisations by type. One firm can hold more than one, and each fund, sub-fund and cell counts once.
Increase and decrease
The authorisations the report lists as added and removed in the quarter, by type.
Change over the period
The latest charted count minus the first charted count, for example Q2 2024 to Q1 2026.
Net change in a quarter
A quarter's charted count minus the previous charted count.
Payment and e-money
The MFSA's "Financial Institutions" line: firms licensed under the Financial Institutions Act, which covers payment institutions and electronic money institutions. The MFSA report gives one count for the whole category and doesn't split it by licence type.
Life insurers
The MFSA's "Life Undertakings" line: insurance undertakings authorised to write long-term (life) business. Protected cells of insurance companies are counted separately and aren't included here.
Non-life insurers
The MFSA's "Non-Life Undertakings" line: insurance undertakings authorised to write general (non-life) business, such as motor, property, liability and specialty lines. Protected cells of insurance companies are counted separately and aren't included here.
Insurance brokers
The MFSA's "Insurance Brokers Register" line: entries on the MFSA's insurance brokers register. The report also counts an insurance brokers list separately, and we keep it apart here.
Investment firms
The MFSA's "Investment Firms" line: firms holding an investment services licence, such as brokers, portfolio managers and investment advisers. Fund managers have their own categories in the report and are covered on the fund managers page.
Fund managers and administrators
The MFSA's "AIF Managers, UCITS Managers, De Minimis AIF Managers and Recognised Fund Administrators" lines: managers of alternative investment funds (AIFs), UCITS management companies, de minimis AIF managers, and recognised fund administrators. The report counts each separately. One firm can hold more than one of these authorisations, so we show each count on its own and don't add them up.
Firms and individuals authorised to provide a service
Our grouping. In the Q1 2026 report it covers: Credit Institutions; Financial Institutions; Life Undertakings; Non-Life Undertakings; Ancillary Insurance Intermediary; Insurance Agents List; Insurance Agents Register; Insurance Managers List; Insurance Managers Register; Insurance Brokers List; Insurance Brokers Register; Tied Insurance Intermediary; Exempted Custodians; Exempted Investment Managers; Registered Custodians; Registered Investment Managers; Retirement Scheme Administrator; Investment Firms; UCITS Managers; De Minimis AIF Managers; Depositories (inc. Depo Lite); AIF Managers; European Venture Capital Fund Managers; Recognised Fund Administrators; Trading Venues; Central Counterparties; Benchmark Administrator; Central Securities Depository; Crowdfunding Service Provider; Class A CSP; Class B CSP; Class C CSP; Under threshold Class A CSP; Under threshold Class B CSP; Limited Company Service Provider; Trustees and/or Mandataries and/or Administrators; Trustees of Family Trusts; Trustees authorised in terms of Article 43(8) of the Trusts and Trustees Act- Trustees and/or Mandataries; Nominees; Notary acting as Qualified Person; VFA Service Provider; Crypto-Asset Service Provider.
Funds, sub-funds, pension schemes and other vehicles
Our grouping. In the Q1 2026 report it covers: Retirement Funds; Retirement Schemes; Alternative Investment Funds; sub-funds AIFs; Retail Maltese UCITS Schemes; sub-funds UCITS; Professional Investment Funds; sub-funds PIFs; Notified Professional Investor Fund; sub-funds NPIFs; Notified Alternative Investment Funds; sub-funds for NAIFs; Recognised Incorporated Cell Companies; Recognised Private Collective Investment Schemes; Notified Securitisation Vehicle.
Cells of insurance companies and intermediaries
Our grouping. In the Q1 2026 report it covers: Cells of Undertakings; Cells of Intermediaries.
Crypto-asset issuers and white papers
Our grouping. In the Q1 2026 report it covers: Whitepaper; Issuers of other crypto assets other than ARTs or EMTs; Issuer of EMTs.

8 steps from the PDF to the number on the page

A script reads each report, and automated tests check every figure before publishing.

  1. Download and hash

    Each report is a 4-page PDF from mfsa.mt. We keep a copy with its byte size and SHA-256 hash in a manifest, so a later download that differs is caught before anything is read.

  2. Read the table

    The Q1 2025 to Q1 2026 reports carry their table as text. A script (scripts/sync-mfsa-authorisations.mjs) reads each row's sector, authorisation type, increase, decrease and active count from pages 2 to 4, with the page number. The Q2 2024 and Q3 2024 and Q4 2024 reports draw the table as images, so we rendered the pages and typed every row, then re-read pages 2 and 3 at a higher resolution.

  3. Give each type one key

    Type names change spelling between reports ("UCITS" and "UCITs", "ProfessionalInvestor"). The script maps each name to one stable key, and stops if it meets a name it doesn't know.

  4. Add up the rows

    For every report the script sums the increase, decrease and active columns and compares each with the printed total. The result is recorded next to the report.

  5. Group the types

    Every key falls in exactly one of 4 groups: firms and individuals that provide a service; funds, sub-funds and other vehicles; cells; and crypto-asset issuers and white papers. The groups add up to the sum of the report's rows.

  6. Build each series

    A type's series is its active count in each of the 8 quarters. A row marked with a source issue (next section) keeps its printed value in the data and the CSV, and is a gap in the chart.

  7. Compute the changes

    The change over the period is the latest charted value minus the first. The net change in a quarter is that quarter's charted value minus the previous charted value, so a gap carries the comparison over to the next charted quarter. We don't use the report's increase and decrease columns for this, for the reason the next section gives.

  8. Test before publishing

    Automated tests compare every figure quoted on these pages with the data file, check that the groups add up to each report's rows, and check that a flagged row never appears in a trend.

A second, independent reading matched every value

  • Row sums against printed totals in all 8 reports. 2 of them differ, as the next section explains.
  • A separate AI model with no access to our data read the same 8 reports again and returned the printed total, the row sum and 13 key rows for each quarter. All 120 values matched.
  • 283 quarter-on-quarter changes compared with the listed increases and decreases.
  • File hashes checked on every run of the sync script and in the test suite.

5 problems in the reports, and how we handle each

We publish the counts the way the MFSA printed them. When a value is out of line, the data keeps it with a note and the charts leave it out.

  • Q1 2025

    Rows repeat the previous quarter

    57 of the 62 rows in the Q1 2025 report match their Q4 2024 values exactly, increases and decreases included. Only the categories new in 2025 differ. Q1 2025 stays out of the trend charts.

  • Q3 2024

    A block of rows out of line

    In the insurance intermediaries block, the values sit one row away from their labels. Ancillary intermediaries show 17 (1 in the quarters either side) and the insurance brokers register shows 427 (138 and 143 either side). The block stays out of the charts.

  • Q2 2024, Q3 2024

    Rows don't add up to the printed total

    Q2 2024: the rows add up to 2,394 against a printed 2,391. Q3 2024: the rows add up to 2,374 against a printed 2,359. We chart totals only for quarters where the two agree. In Q4 2024 the decrease column adds up to 30 against a printed 29.

  • 85 of 283

    Changes that don't match the listed flows

    In 85 of the 283 cases we could test, the change in a type's active count from one report to the next doesn't equal the increases minus the decreases the later report lists. So we quote the active counts and compute changes from them.

  • 2025 and 2026

    Categories added and merged

    2025 reports add notified professional investor funds and the MiCA categories (crypto-asset service providers and issuers). The Q1 2026 report adds limited company service providers and merges 3 trustee lines into one. So a count that starts partway through the 8 quarters is a new line in the report. It doesn't mean the firms are new to Malta.

Where the 78 payment and e-money institutions come from

  1. The printed row

    The MFSA's Q1 2026 Authorised Persons Report, page 2, row "Financial Institutions" under the sector "FINANCIAL INSTITUTIONS", prints 5 in the increase column, 1 in the decrease column and 78 active authorisations.

  2. Into the data

    The sync script reads that row from the PDF text and writes it to authorisations.json with the key financial-institutions, the source URL and the retrieval date, 10 November 2026.

  3. Onto the page

    The overview and the payment and e-money page show 78. The net change for the quarter is 78 minus the Q4 2025 count of 68, so +10. The tests fail if the page and the data disagree.

  4. In the CSV

    The same row, with its page and retrieval date, in authorisations-by-quarter.csv.

The reports have no staff numbers or distinct firm counts

  • How many people work in each type of firm. The reports have no employment figures.
  • How many distinct companies hold an authorisation, because one firm can hold several.
  • Anything before Q2 2024. Earlier reports use a different layout and aren't in this series yet.

Using this research

Cite as

Subrahmanya, P. and Bonello, S. (2026). Malta's financial services sector in numbers. Aigentcy. https://www.aigentcy.com/research/malta-financial-services/methodology/ Data: MFSA Authorised Persons Reports, Q2 2024 to Q1 2026.

If you use our figures, charts or data, credit Aigentcy and link to this page.

Measure one regulated process and price the build in 2 weeks

The 2-week AI readiness assessment maps one process with the people who run it, such as onboarding, reconciliations or returns. You get the hours, the volumes and a fixed-price build estimate before you commit.

Ask your AI assistant about us

Open our summary in the assistant you use, or copy the prompt.

Read the prompt