University of Jyväskylä Banking and International Finance Master's Thesis Guide: KTTS1940, 30 ECTS, JSBE and Vasara

Quick answer: what is the JYU Banking and International Finance thesis route?

The current University of Jyväskylä Banking and International Finance (BIF) programme is a 120 ECTS, two-year Master of Science in Economics and Business Administration at the Jyväskylä University School of Business and Economics (JSBE). Under the current BIFMP2024 curriculum, the thesis package has three separate components: KTTS1930 Master’s Thesis Seminar, 3 ECTS, KTTS1940 Master’s Thesis, 30 ECTS, and KTTS1945 Maturity Test, 0 ECTS. The thesis is an English-language Advanced Studies course graded 0-5. JSBE describes the thesis as a scientific research task that normally takes about five to six months, with an appointed supervisor, seminar feedback, two reviewers and formal submission through Vasara before automatic archiving in JYX.

1. Start from the exact 2024-2028 BIF thesis package

Do not begin by copying a thesis rule from another JYU programme or an older study guide. The current BIF Advanced Studies structure lists KTTS1930, KTTS1940 and KTTS1945 together with the compulsory banking, time-series, asset-pricing, international-finance and valuation courses. The 3-credit seminar and 30-credit thesis are credit-bearing, while the maturity test carries 0 ECTS. This separation matters when you plan your study schedule: seminar participation is not hidden inside the 30-credit thesis, and the maturity test remains a required completion even though it adds no credits. Your personal Sisu plan should reflect the current programme objects.

2. KTTS1940 is a scientific research project, not a long coursework essay

The official KTTS1940 description requires an independently produced thesis guided by a supervisor and reported as scientific research. Its learning outcomes move through the full research chain: defining a research problem, using relevant literature, justifying a methodology, applying methods, analysing evidence, interpreting results, drawing conclusions and explaining the contribution. A useful way to read this is that the final document is only one output of a larger research process. A polished introduction cannot compensate for an unclear research problem, an unsuitable dataset or an analysis that does not answer the stated question. Build the research logic before polishing the final wording.

3. Enter thesis work after building enough BIF disciplinary and methods background

The current KTTS1940 course page expects the majority of the compulsory BIF courses to have been completed before the thesis. That is consistent with the structure of the programme. Principles of Banking provides institutional and regulatory foundations; Applied Time Series Analysis develops financial-econometric tools; Investments and Asset Pricing develops theoretical and empirical asset-pricing reasoning; International Finance covers foreign-exchange and international-market mechanisms; and Financial Statement Analysis and Valuation supports firm-level interpretation. These courses do not dictate your thesis topic, but they are part of the intellectual base from which you are expected to formulate and defend an advanced research problem.

4. Treat the five-to-six-month JSBE estimate as a research schedule, not a writing deadline

JSBE says that a Master’s Thesis normally takes about five to six months. That period should include topic refinement, literature mapping, data-access checks, research design, possible ethics or privacy work, analysis, seminar feedback, drafting, revision and final submission. It is risky to spend most of the period collecting articles and then leave methods and writing to the last few weeks. A practical schedule is to overlap tasks: build the literature review while validating the data, write the methods section while preparing analysis scripts, and keep a decision log so that changes to variables, samples or models can later be explained rather than reconstructed from memory.

5. Use the appointed supervisor as a research-control point

JSBE states that every thesis student has an appointed supervisor and stresses continuous contact. Supervision is most useful when meetings are tied to decisions. Bring a short agenda that identifies what has changed, what evidence you have, what decision you need, and what you will do next. Important checkpoints include the research question, theoretical frame, data source, unit of analysis, sample construction, identification logic, model specification, robustness plan and interpretation boundaries. The supervisor should not be expected to discover an undocumented problem at the end. Early disclosure of uncertainty usually makes the project easier to repair.

6. KTTS1930 makes research communication part of the thesis process

The 3-credit Master’s Thesis Seminar is an Advanced Studies course completed on a Pass-Fail basis in English. It requires presentation and argumentation at different phases of the work, including a research plan, a mid-term stage and commenting on another master’s thesis. The seminar is therefore not only a presentation requirement. Use it as a structured stress test. If peers cannot identify your question, variables, design and intended contribution from a short presentation, the written thesis may also be carrying too many ideas. Peer commenting can also expose unclear assumptions before they become embedded in the final analysis.

7. Define one research problem before choosing an impressive method

Banking and finance offers many technically attractive methods, but the research question must come first. A defensible question identifies the phenomenon, the population or market, the relevant period, the main relationship or mechanism, and the type of inference you are trying to make. Compare “Does monetary policy affect banks?” with a more researchable question that specifies which monetary-policy change, which bank outcome, which banks, which period and which empirical design. A narrow question is not academically weak. It gives you a clear basis for choosing data, models, robustness checks and limitations.

8. Separate theoretical mechanism, empirical prediction and statistical specification

A finance thesis becomes clearer when three layers are kept distinct. First, explain the economic mechanism from theory. Second, convert that mechanism into an empirical prediction, such as an expected sign, cross-sectional difference or dynamic response. Third, choose a statistical specification capable of testing that prediction with the available data.

9. Match the unit of analysis to the claim you want to make

Before downloading data, state what one observation represents. It might be a bank-year, firm-quarter, country-month, currency-day, security-month or another defined unit. Then check whether the thesis claim is made at that same level. For example, a country-level macro variable combined with bank-level outcomes creates a multilevel structure, while daily market data create dependence patterns that differ from annual panel data.

10. Use time-series methods only when the time-series structure justifies them

KTTS4120 gives BIF students a strong foundation in financial time-series analysis, including ARMA models, ARCH/GARCH-type volatility models, VAR models and tools for non-stationary series such as Engle-Granger and Johansen procedures. These are possibilities, not compulsory thesis methods. A time-series model is useful when the research question concerns dynamics over time and the data have enough observations at a meaningful frequency. Before estimating, examine the time index, missing periods, structural changes and whether the variables appear stationary.

11. Stationarity and cointegration are research assumptions, not software buttons

Financial and macroeconomic series can trend, drift or contain persistent shocks. Running a conventional regression on non-stationary variables can produce apparently strong relationships that are not economically meaningful. If your design uses levels of persistent series, explain why the specification is valid. When cointegration is relevant, state the economic reason a long-run relationship is plausible and then use suitable diagnostics and procedures.

12. Treat volatility models as models of conditional variance, not as proof of risk in every sense

ARCH and GARCH-family models can be useful when the research question concerns changing conditional volatility in financial returns or similar series. They do not automatically measure every form of financial risk. A volatility estimate is different from credit risk, liquidity risk, tail loss, solvency risk or systemic risk. Define the risk concept before selecting the model.

13. VAR models require a clear system and disciplined interpretation

A VAR can model dynamic interactions among several time-series variables, but it does not automatically identify causal shocks. Explain why the selected variables belong in one system, how the lag structure is chosen, whether the series are stationary or cointegrated, and what ordering or identifying restrictions are used if you interpret impulse responses structurally. Forecast-error variance decompositions and impulse responses can be informative, but their meaning depends on the identification design. If the thesis only establishes predictive dynamics, say so rather than upgrading the result to a causal mechanism.

14. Cross-sectional and panel designs need their own diagnostic logic

Not every BIF thesis should be a pure time-series study. Firm, bank or country panels may be better for questions about heterogeneity across entities. The programme recommends econometric preparation that includes regression, heteroskedasticity, serial correlation, multicollinearity, endogeneity and panel-data ideas. In a panel study, explain the entity and time dimensions, fixed or random effects reasoning where relevant, clustering of uncertainty, missingness and sample selection.

15. Asset-pricing tests must distinguish model fit from economic interpretation

BIF coursework covers CAPM, multifactor models, expected returns, anomalies and empirical tests. A thesis can use this environment to study factor exposures, cross-sectional returns, portfolio strategies or market behaviour. But a statistically significant alpha does not automatically prove a persistent exploitable profit. Discuss model choice, factor construction, sample period, transaction costs where relevant, multiple testing, look-ahead risk and whether the result is in-sample or out-of-sample.

16. Banking questions need institutional definitions before estimation

For banking topics, define what type of institution, balance-sheet measure, regulation or policy channel the study addresses. “Banks” can differ by ownership, business model, size, geography and regulatory environment. Principles of Banking covers financial intermediation, competition, regulation, monetary-policy transmission and financial stability; those concepts can help frame mechanisms. If you use accounting or supervisory variables, explain definitions and comparability over time.

17. International-finance questions need currency, market and timing discipline

International Finance covers exchange-rate markets, international assets, parity relationships, hedging, diversification and open-economy mechanisms. When a thesis crosses currencies or countries, specify the base and quote currency, conversion convention, data timestamp and market frequency. Check whether holidays and time zones create mismatched observations. If returns or prices are converted, document the formula.

18. Macroprudential and financial-stability topics require careful policy inference

Optional BIF coursework addresses systemic risk, contagion, early-warning indicators, macroprudential policy and European financial data. If your thesis enters this area, separate prediction from policy evaluation. An indicator that rises before stress episodes may be useful for classification without proving that changing the indicator would prevent a crisis. Policy variables can be introduced in response to existing risk, creating reverse-causality problems.

19. Distinguish statistical significance, economic magnitude and causal interpretation

These are three separate questions. Statistical significance concerns uncertainty under a model. Economic magnitude concerns whether the estimated effect is large enough to matter in the context of banking or financial markets. Causal interpretation asks whether the design supports the claim that changing one factor would change another.

20. Pre-specify a robustness plan that tests your main vulnerability

Robustness should not mean trying many models until one gives the preferred result. Identify the assumption most likely to change your conclusion and design checks around it. Examples include alternative variable definitions, different sample windows, lag choices, treatment of outliers, alternative standard errors, different factor specifications or excluding a period dominated by an unusual shock. Explain why each check is informative.

21. Build a reproducible data trail from raw source to final table

Keep raw data separate from transformed data. Record the source, download date, variable definitions, filters, joins, currency conversions and transformations. Use scripts or documented steps so that the final tables and figures can be recreated. If a commercial source imposes licensing restrictions, do not place restricted raw data in a public repository merely for reproducibility.

22. Use the BIF research-data-management requirement early

The current degree structure includes the one-credit LIBEDATA Introduction to Research Data Management course. Its English course guidance is designed around preparing a data management plan for the thesis and recommends doing the work with the thesis seminar before data collection or handling. Use the DMP to decide what data you will use, who can access it, how it will be documented and stored, what agreements or licences apply, what happens after the thesis, and whether any data can be opened. Treat the DMP as a working research document, not as an administrative form completed after the analysis.

23. Personal data can exist even in a finance thesis

A finance topic may appear non-personal because it uses firms, banks or markets, but some projects include interviews, surveys, employee-level records, entrepreneur data, social-media information or identifiable expert responses. JYU instructs students to assume that information collected from or about identifiable living people may be personal data. In an independently conducted student thesis, the student is normally the data controller. Plan the purpose, data minimisation, information to participants, retention and secure storage before collection.

24. Ethical review is conditional, but research ethics are always relevant

JYU follows national research-integrity and human-sciences ethical principles. A thesis does not automatically require prior ethical review merely because people are involved. However, when a TENK review criterion is met, review must be sought before the research begins. Discuss the need with the supervisor during planning.

25. Build the literature review around the research task, not around a list of authors

The JSBE assessment matrix rewards a theory section that is linked to the research task, uses relevant sources critically and handles concepts consistently. Organise the literature by mechanisms, debates, models or empirical findings that directly help answer your question. Compare studies that disagree, explain differences in samples or methods, and identify the gap your design can realistically address. JYU’s general citation guidance requires proper acknowledgement when quoting or paraphrasing and emphasises consistent referencing.

26. Write to the JSBE assessment criteria from the beginning

The official JSBE matrix uses grades 1-5 and identifies five main thesis criteria: research task, theory, methods and data, results and conclusions, and research report. The matrix also considers mastery of the research process, including initiative, responsibility, feedback and scheduling. Use these criteria as a design checklist. Ask whether the research task is clearly limited, whether theory supports it, whether data and methods fit it, whether the results actually answer it, and whether conclusions stay within the evidence.

27. Use the JSBE template and current faculty instructions as formatting controls

The central JYU Master’s thesis page links the current JSBE Master Thesis template and JSBE-specific instructions. Use the current template rather than reproducing a layout from an old thesis found in JYX. A previous thesis can be helpful for seeing how another student presented a study, but it is not the authority for current formatting, submission or assessment rules. Check heading structure, abstract placement, tables, figures, references and accessibility against the current template and current instructions.

28. Follow JSBE’s AI rules, especially the abstract restriction

JSBE states that generative AI is generally permitted, but it must be used responsibly and transparently. AI should not replace your own thinking, reading or final text production, and outputs should be checked against original sources. Report the tool and purpose of use as required. The most important thesis-specific restriction is that AI is not allowed in the master’s thesis abstract when the abstract serves as the maturity test.

29. Treat Turnitin as a pre-submission control, not a last-minute obstacle

Before final Vasara submission, JYU instructs students to complete the Turnitin check according to the supervisor’s guidance. Use similarity feedback to examine quotation, paraphrasing and source attribution rather than focusing only on a percentage. A similarity score is not itself a plagiarism verdict, and a low score does not prove that every source has been used correctly. Resolve citation problems while revisions are still possible.

30. Vasara, Sisu, PDF/A and JYX form one final submission chain

Before submitting, confirm with the supervisor that the thesis is final, make sure you are registered for both the thesis and maturity-test implementations in Sisu, complete the required Turnitin process and convert the thesis to an accessible PDF/A. Vasara is used for submission, evaluation and archiving across JYU faculties. The evaluated thesis is automatically transferred to JYX. JYU also warns that confidential background material must not remain in the submitted public thesis file.

31. Understand the examiner and grading chain

The degree regulations require two examiners. The supervisor-in-charge must have a doctoral degree and be employed by JYU; at least one examiner must hold a doctoral degree, and the second examiner may be the supervisor. JSBE separately states that two teachers or other field experts review the thesis and normally provide their statement within one month after final submission for review. You have an opportunity to respond to the reviewers’ report before the final grade.

32. Use the interruption right carefully

JYU’s degree regulations allow a student to interrupt the thesis-review process before the final grade is given. That safeguard should not be misunderstood as permission to submit an unfinished thesis casually. If review is interrupted and the work is later resubmitted, the regulations require meaningful reworking under supervision rather than a simple unchanged resubmission. The better strategy is to use supervisor feedback and seminar review before final submission.

33. The maturity-test route depends on prior language demonstration

KTTS1945 is a 0-credit Pass-Fail requirement. The current course information states that if the required language skills were already demonstrated in the bachelor’s-degree maturity test, the master’s thesis abstract serves as the maturity test. Students whose language route differs may have another maturity-test form. Therefore do not copy a friend’s process without checking your own prior degree and language status.

34. Co-written theses are possible, but individual contribution must remain assessable

JYU degree regulations allow pair work and the JSBE provides separate co-written thesis guidance. Collaboration does not remove individual academic responsibility. The contribution of each student must be independently assessable, and the pair should agree early on working practices, ownership, use of research material and what happens if collaboration ends. If you are considering a co-written BIF thesis, use the current JSBE co-writing document and discuss the arrangement with the supervisor before building a joint dataset or manuscript structure.

35. Keep confidential material outside the public thesis document

Master’s theses are public academic works and are stored in JYX. This has practical consequences for industry or bank-related projects. A company can provide restricted background information, but confidential business material, personal data or contract-protected content should not simply be inserted into the public thesis file. Agree early on what may be published, what can be aggregated or anonymised, and what must remain in separate working material.

36. Final quality check: can every chapter answer one assessment question?

Before submission, read the thesis from the examiner’s perspective. Can the introduction state a precise research task? Does the theory section explain why the question matters and what mechanisms are expected? Does the methods section justify the data and model?

37. What should you recheck immediately before submission?

Recheck the current KTTS1940, KTTS1930 and KTTS1945 pages, the JSBE Master’s thesis instructions, the current JSBE template, the JSBE AI principles, the Vasara submission page and your Sisu registrations. Confirm supervisor readiness, Turnitin completion, maturity-test route, PDF/A accessibility and any data/privacy obligations. This guide is verified against the current 2024-2028 BIF curriculum and JYU instructions available on 11 September 2026. It deliberately does not claim a universal BIF thesis word count, page count, citation style or mandatory statistical model where the official sources do not establish one.