For a UK accounting or finance doctorate the working answer is WRDS if your business school subscribes, LSEG Datastream through Workspace for UK and international market data, and Companies House for anything on private UK companies, with Bloomberg reserved for the variables nothing else holds. The table below compares seven sources on the criteria that matter across a four-year thesis; the shortlist and a single recommendation follow.
| Source | What it holds | UK coverage | Typical doctoral access | Reproducibility at year three | Best for |
|---|---|---|---|---|---|
| WRDS (Wharton Research Data Services) | Platform hosting Compustat, CRSP, Audit Analytics, BoardEx, I/B/E/S via LSEG, the London Share Price Database, Bureau van Dijk, TAQ, OptionMetrics, Fama-French factors and more; 350+ TB across 50+ vendors | Depends on which vendors the institution licenses; LSPD covers UK equities | Institutional subscription; individual account on the business school’s licence | High: queries saved as SAS, Python or R code against stable tables | Any archival study; the default for empirical accounting and asset pricing |
| Compustat (S&P Global Market Intelligence) | Standardised company fundamentals, North America and Global files | UK listed firms in Compustat Global, with fewer items than the North America file | Through WRDS or S&P Capital IQ | High through WRDS | Financial-statement-based accounting research |
| CRSP (now part of Morningstar) | US security prices, returns and indices back to 1925 | None; US only | Through WRDS | High | Asset pricing and event studies on US markets |
| LSEG Datastream | 35 million instruments and indicators, 120 years of history, 175 countries; equities, bonds, indices, economics, Worldscope fundamentals, I/B/E/S estimate aggregates | Strong; the standard source for UK and European market series | Through LSEG Workspace on library or business school seats, Excel add-in, APIs for Python and R | Medium: request files and code lists must be archived by you | UK and international market-based studies, macro-finance |
| Bloomberg Terminal | Real-time and historical prices, fixed income, derivatives, ESG, news, Excel add-in | Strong | A handful of terminals in the library; no remote access; export limits | Low: pulls are seat-bound and hard to re-run | Bond, derivative and ESG variables unavailable elsewhere |
| FAME (Bureau van Dijk, a Moody’s company) | Filed accounts, directors and ownership for UK and Irish companies, including private firms | Strong; the only structured source for private UK company accounts | Library subscription; also as a Bureau van Dijk vendor on WRDS where licensed | Medium: exports must be versioned | Private-firm, SME, audit and corporate governance studies |
| Companies House | Registered address, incorporation date, current and resigned officers, filed document images, charges, previous names and insolvency information for every UK company | Complete for UK-registered companies | Free, public, with bulk products and an API | High for the raw filings; the parsing is yours | Hand-collected samples, director networks, verifying FAME |
The ranked shortlist
1. WRDS — if your school has it, start here
Wharton Research Data Services is not a database but a platform: an institutional subscription unlocks the vendors the business school has licensed, and the catalogue runs to more than fifty, from S&P Global’s Compustat and Capital IQ to CRSP, Audit Analytics, BoardEx, ISS ESG, PitchBook, Preqin and NYSE TAQ. WRDS reports over 75,000 users in more than 35 countries and over 350 terabytes of data. For a doctoral candidate the decisive feature is not breadth but reproducibility: queries are written and stored as SAS, Python or R code against tables with stable names, so the sample described in chapter three can be regenerated in year four when an examiner asks why the firm count changed. Where it falls short at doctoral scale is UK coverage: the London Share Price Database is available as a vendor and Compustat Global carries UK listed firms, but the platform’s centre of gravity is US, and a thesis on AIM-listed companies or UK private firms will need Datastream and FAME beside it. Check which vendors your institution actually licenses before designing the study; the WRDS catalogue and your library’s subscription are different lists.
2. LSEG Datastream through Workspace — the UK market-data workhorse
Datastream is the source UK finance departments have used for decades, and it is now reached through LSEG Workspace, the successor to Eikon, which runs on desktop, web and mobile with Microsoft 365 integration and a Python environment. LSEG describes the database as more than 35 million instruments and indicators across 175 countries with 120 years of history, including 8.5 million active economic series, Worldscope fundamentals and I/B/E/S estimate aggregates. It suits UK and European studies where WRDS is thin: FTSE constituents through time, sterling bond yields, UK macro series, delisted UK firms. Its weakness for a thesis is discipline rather than data. Requests are built interactively in the Excel add-in, and unless you archive every request file, code list and download date, the year-one sample cannot be rebuilt in year three. Treat every pull as a dataset with a version number.
3. FAME with Companies House — the only route to private UK firms
Most UK companies are private, and neither Compustat nor Datastream sees them. FAME, from Bureau van Dijk, now a Moody’s company, structures the accounts, directors and ownership that UK and Irish companies file, and it is the standard source for doctoral work on SMEs, audit fees, going-concern opinions, family firms and director networks. Companies House is the primary record behind it: the official service provides company information such as registered address and date of incorporation, current and resigned officers, filed document images, mortgage charge data, previous company names and insolvency information, free of charge, with alerts when a company’s details change. The pairing matters for examination. FAME fields are derived, and a socio-legal examiner or an auditor on the panel will ask whether you checked a sample of FAME values against the filed accounts. Do it, and say so in the methods chapter.
4. Compustat and CRSP — essential for US-market theses, partial for UK ones
Compustat from S&P Global Market Intelligence is the standardised fundamentals file that empirical accounting is built on; CRSP, now part of Morningstar, holds US security prices and returns back to 1925. Together, matched through the CRSP–Compustat link on WRDS, they are the infrastructure of the US-based literature your thesis will be positioned against. For a UK-focused doctorate they are supporting, not primary: Compustat Global carries UK listed firms with a narrower set of items than the North America file, and CRSP has no UK coverage at all. The decision is about the literature you want to speak to. A thesis replicating a US design on UK data will use Compustat Global and Datastream and defend the differences; a thesis contributing to the US literature will use these two and treat the UK as an out-of-sample test.
5. Bloomberg — for the variable nothing else holds
The terminal is unmatched for fixed income, derivatives, ESG scores and news, and most business schools hold a small number of seats. As a thesis foundation it is the weakest option here: access is seat-bound, downloads are subject to limits, and a pull made at a terminal in year one is close to impossible to reproduce exactly in year three. Use it for the specific variables your design needs and nothing else can supply, document the field codes and dates, and build the rest of the sample elsewhere.
6. Audit Analytics, BoardEx and I/B/E/S — the specialist layers
Three vendors deserve naming because they define sub-fields. Audit Analytics, now under Ideagen, is the source for auditor changes, fees, restatements and internal-control opinions; BoardEx for director and executive biographies and board interlocks; I/B/E/S, distributed by LSEG, for analyst forecasts and the earnings-surprise literature. All three are reached through WRDS where the institution licenses them, and all three are US-heavy with UK coverage that must be checked firm by firm before the research questions are fixed. The variables table of a management doctorate should name the vendor and the field for every measure, and the same rule applies in accounting and finance.
The recommendation
Build the thesis on WRDS if your institution subscribes, take UK and international market series from Datastream through Workspace, and go to FAME and Companies House for private firms. Use Bloomberg for individual variables only. If your institution does not hold WRDS — and many UK business schools outside the largest do not — Datastream plus FAME will carry a UK-focused thesis, and the decision about whether to add a WRDS-based US comparison should be taken before the confirmation of registration, not after. Whatever the mix, the examinable point is the same: every variable in the thesis traceable to a named source, a named field, a download date and a stored query. Our comparison of R, Python, SPSS and Stata covers the tooling that keeps those queries reproducible.
Access, cost and what to check before you design the study
None of these sources is bought by the candidate. They are held on library or business school licences, and the gap between what WRDS lists and what your institution has paid for is where doctoral projects stall. Before fixing the research design, obtain in writing from the library which vendors are licensed and for what date range, whether the licence permits the thesis to be deposited with the data appended, and whether access survives a change of institution or a writing-up year. Candidates on a self-funded route should ask the last question twice. The same discipline that our piece on medical and epidemiology data sources describes for health data — application timelines and data-sharing agreements — applies here in milder form: the licence terms decide what the thesis may publish.
Reproducibility is the criterion examiners now apply
The methodological turn in empirical finance has reached the viva. Examiners ask for the sample construction table — firm-years at each filter — and increasingly for the code. A study whose sample cannot be rebuilt because the Datastream request was never saved, or because the Bloomberg pull was made once at a terminal, is exposed at exactly the moment it cannot be repaired. Save every request file, record vendor, field, date and licence for each variable, and keep the code that joins them under version control from the first month.
Keeping that record is a writing problem as much as a data problem. Draft your data and methods chapter in Tesify: keep the sample construction table, the variable definitions with their vendor and field codes, and the chapter text in one workspace, so that the numbers in chapter three and the numbers in chapter five come from the same record. Before submission, the Tesify plagiarism checker lets you check your own manuscript for unattributed overlap with the papers you replicate — a check to run on yourself before the examiners run theirs. The data licences, the queries and the results stay yours; the tool keeps the record consistent.
Frequently asked questions
Do I need WRDS for a UK finance PhD?
Not necessarily. A UK-focused thesis can be built on Datastream through LSEG Workspace for market data and FAME with Companies House for company data. WRDS becomes essential when the thesis positions itself against the US literature or needs Compustat, CRSP, Audit Analytics or BoardEx.
What is the difference between Compustat and Datastream?
Compustat is a standardised company fundamentals database from S&P Global, strongest for North America; Datastream is LSEG’s time-series database of more than 35 million instruments across 175 countries, strongest for market data and for UK and European coverage, with Worldscope supplying the fundamentals.
Is CRSP available for UK stocks?
No. CRSP covers US securities. The UK equivalent for long-run share prices is the London Share Price Database, available through WRDS where licensed, and Datastream for current and historical UK equity series.
Where do I get data on private UK companies?
FAME structures the accounts, directors and ownership that UK and Irish companies file, and Companies House provides the underlying filings free, including document images, officers, charges and insolvency information.
Can I use Bloomberg for my whole thesis sample?
It is inadvisable. Terminal access is seat-bound and downloads are limited, so a sample built at a terminal is hard to reproduce when examiners ask. Use Bloomberg for specific variables and build the sample on WRDS or Datastream.
How do I make my financial data reproducible for the viva?
Store every query as code or a saved request file, record the vendor, field code, download date and licence for each variable, keep a sample construction table showing firm-years lost at each filter, and version the code that merges the sources.
Does the university pay for these databases?
Yes. WRDS, Datastream, Bloomberg and FAME are held on institutional licences through the library or business school. Confirm in writing which vendors and date ranges are licensed before you fix the research design.
Can I deposit my dataset with the thesis?
Only if the licence permits it. Most vendor licences prohibit redistributing raw data, so theses deposit the code and the sample construction table and describe how to rebuild the dataset from the licensed source. Companies House data is public and can be deposited.