PhD Researchers & Doctoral Candidates
A correctly specified volatility, spillover, or asset-pricing model for a finance thesis—built to satisfy a demanding committee.
Research Methods
Financial Econometrics & Risk Analytics
Financial data breaks the assumptions ordinary regression relies on—volatility clusters, tails are fat, and relationships shift with the market regime. We use the models built for that reality: GARCH-family volatility, spillover and connectedness, wavelet and event-study methods, matched to how your series actually behaves.
GARCH family · realized volatility
Spillover · wavelet · event studies
R (rugarch) · Python · EViews
Reproducible, journal-ready
Sample output
conditional σ
returns
Overview
Models built for how markets actually move
Financial return series don't behave like the textbook. Volatility arrives in clusters—calm stretches broken by turbulent ones—tails are fatter than a normal distribution allows, and the way assets move together tightens in a crisis and loosens in calm. Fit an ordinary model to that data and the standard errors, the risk estimates, and the conclusions are all wrong in ways a finance referee will spot immediately.
So we start from the stylized facts of your series. If shocks of different signs move volatility asymmetrically, an EGARCH or GJR model rather than plain GARCH; for several series with time-varying correlation, DCC-GARCH; for transmission between markets, a spillover and connectedness framework; for relationships that differ by time horizon, wavelet coherence. The model is chosen from what the data shows, and the distributional assumptions are tested, not assumed.
Whether the work is a volatility study, a connectedness paper, an event study, or a risk model, the deliverable is the same: correct specification, the diagnostics that finance journals require, out-of-sample evaluation where forecasting is involved, and reproducible code you keep.
Who We Work With
For research on markets, risk & returns
If your data is financial—prices, returns, volatility, risk—this is the right desk to write to.
Finance & Economics Researchers
GARCH-family, connectedness, and wavelet studies done to the standard the top finance journals expect.
Energy & Commodity Researchers
Volatility and spillover analysis across energy, commodity, and financial markets—a fast-growing research area.
Research Institutes & Central-Bank Teams
Financial-stability and risk research where connectedness and systemic-risk measurement matter.
Banking & Risk Researchers
Credit-risk and financial-risk modeling for academic and applied work on institutions and portfolios.
Corporates & Industry R&D
Applied volatility and risk analytics translated from academic rigor into market and treasury decisions.
Capabilities
The full financial-econometrics toolkit
Organized by what you're measuring—volatility, transmission, or risk and returns. If your study needs a method not listed here, ask—this is the core, not the boundary.
The GARCH family & realized volatility
Conditional-volatility models for the clustering, asymmetry, and fat tails that define financial returns.
- ARCH
- GARCH
- EGARCH
- TGARCH
- GJR-GARCH
- DCC-GARCH
- GARCH-MIDAS
- Realized volatility
- Volatility forecasting
Spillovers, connectedness & wavelets
Methods for how shocks move across markets and how relationships differ by time horizon.
- Spillover analysis
- Connectedness analysis
- Wavelet analysis
- Wavelet coherence
Returns, events & risk
Event studies and pricing and risk models on market, firm, and portfolio data.
- Event studies
- Asset-pricing models
- Credit-risk models
- Financial-risk modeling
How the Analysis Works
Six steps from returns to a defensible model
A transparent sequence that starts from the stylized facts of your series and tests the distributional assumptions most papers assume. Nothing is a black box.
Steps are adapted to your data: single vs. multiple series, daily vs. high-frequency, and whether the goal is explanation, transmission, or forecasting. We confirm the model with you before estimation begins.
-
1
Explore
Examine the return series—clustering, asymmetry, fat tails, stationarity—so the model choice is driven by the data, not habit.
Checks: ARCH-LM · stationarity · normality · autocorrelation
-
2
Specify
Choose the volatility, spillover, or pricing model whose features match your data, and justify it against the alternatives.
Models: GARCH family · DCC · spillover · wavelet
-
3
Estimate
Fit the model with the right error distribution—Student-t or skewed-t for fat tails—and appropriate inference.
Inference: QMLE · Student-t · robust standard errors
-
4
Diagnose
Check standardized residuals for remaining ARCH effects, correct distribution, and model adequacy.
Tests: Ljung-Box · residual ARCH · sign-bias · goodness-of-fit
-
5
Validate
Where forecasting is involved, evaluate out-of-sample against benchmarks with proper loss functions and accuracy tests.
Checks: out-of-sample · loss functions · Diebold-Mariano
-
6
Report
Deliver volatility, connectedness, or event-study figures, tables, methodology, and reproducible code you keep.
Output: figures · tables · methods · R/Python code
Rigor by default
The checks a finance referee will run
Financial models fail review on the details—the wrong error distribution, residual ARCH left untested, in-sample-only forecasts. Getting them right is standard on every engagement.
Included on every project
- Tests for ARCH effects, asymmetry, and fat tails
- Appropriate error distribution (Student-t, skewed-t)
- Standardized-residual diagnostics and model adequacy
- Out-of-sample evaluation for forecasts
- Reproducible, versioned code you keep
What You Receive
Every engagement, delivered in full
Not a black-box result and a number, but a complete, documented package you can submit, defend, and reproduce.
- Clean, documented return/price dataset
- Model specification with justified distribution
- Full estimation output and diagnostics
- Conditional-volatility, spillover, or event-study figures
- Out-of-sample forecast evaluation, where relevant
- Interpretation of volatility, risk, or transmission results
- Reproducible R (rugarch) or Python code
- Journal-ready methodology and results sections
- Technical responses to methodological reviewer comments, where required
Where this fits
Part of a larger arc
Financial econometrics is strongest when the data work ahead of it is sound and the reporting after it is precise—each handled with the same care.
Research Design & Planning
Identification strategy, power, and specification decided before estimation begins.
Explore methodsStatistical & Methodological Audit
An independent check of assumptions, specification, and reproducibility before submission.
Explore auditPublication & Research Support
Methods and results reporting, journal selection, and reviewer-response support.
Explore support
FAQ
Common questions
Answers to what most researchers and project leads ask before we begin a financial-econometrics engagement.
It depends on what your data shows. If negative and positive shocks move volatility differently—the leverage effect common in equity returns—an asymmetric model such as EGARCH or GJR-GARCH fits better than plain GARCH. For several series with time-varying co-movement, DCC-GARCH; for mixed-frequency drivers, GARCH-MIDAS. We test for the features that distinguish them rather than defaulting to one.
Yes. We estimate spillover and connectedness—the Diebold-Yilmaz framework and related approaches—to quantify how shocks transmit across markets, assets, or economies, and to identify which are net transmitters versus receivers, including how that shifts over time.
Yes. We conduct event studies with appropriate estimation and event windows, correct abnormal-return models, and the parametric and non-parametric significance tests reviewers expect—handling event clustering and thin trading where they arise.
Wavelet methods decompose a relationship by both time and frequency, so you can see whether two series move together in the short run versus the long run, and how that co-movement changes across periods such as crises. Wavelet coherence is especially useful for lead-lag and time-varying relationships that a single-frequency model would miss.
Yes. We construct realized-volatility measures from intraday data and model and forecast them, handling the microstructure noise and data-cleaning issues that high-frequency work requires.
Yes, and we evaluate it honestly. Volatility and risk forecasts are compared out-of-sample against benchmarks using appropriate loss functions and tests of predictive accuracy, rather than reported only on in-sample fit.
Primarily R (rugarch, rmgarch, and related packages) and Python, with Stata or EViews where preferred. You receive versioned, commented code alongside the results and a methods section written to journal standards.
Yes, and it is the best time to involve us. Matching the volatility or risk model to your data's features—asymmetry, fat tails, frequency, and the number of series—before estimation prevents the rejections that come from a model the data never supported.
Related Expertise
Often paired with econometrics
Modeling volatility or risk?
Tell us your series and your question—we'll tell you the right volatility, spillover, or risk model, and what it takes to make it publishable.