Systematic Bagging Model for Market Risk ��- cross pattern learning framework for Value at Risk and Expected Shortfall estimation
Authors:
Michał Woźniak, Robert Ślepaczuk, Marcin Chlebus
University of Warsaw
10.2025, Poland
2
Value at Risk
Expected Shortfall
3
My PhD Thesis Research Landscape
�
This dissertation addresses critical gaps in market risk modeling (economic, reproducibility, methodological and modelling) through a structured, four-phase research program. Each phase builds upon the last, moving from establishing a foundational evaluation framework to proposing a novel, crisis-resilient forecasting models.
4
Research motivation
5
Ensemble methods
[source: Machine Learning Mastery]
6
Cross learning methods
7
Research purpose
8
Research questions
9
Literature review
10
Systematic Bagging Model – framework definition
11
Systematic Bagging Model – framework definition (cont’d)
12
Systematic Bagging Model – framework definition
13
Systematic Bagging Model – framework definition (cont’d)
14
Systematic Bagging Model – framework definition
15
Systematic Bagging Model – framework definition (cont’d)
16
Systematic Bagging Model – framework definition (cont’d)
17
Systematic Bagging Model – framework definition (cont’d)
18
Systematic Bagging Model – framework definition (cont’d)
Query
19
Systematic Bagging Model – framework definition (cont’d)
20
Systematic Bagging Model – framework definition (cont’d)
21
Systematic Bagging Model – framework definition (cont’d)
22
Systematic Bagging Model – framework definition (cont’d)
What Makes MASS So Fast?
The reason why MASS is so much faster than a naive approach is because MASS uses Fast Fourier Transforms (FFT) to convert the data into the frequency domain and performs what is called a “convolution”, which reduces the m operations down to log(n) operations.
23
Systematic Bagging Model – framework definition (cont’d)
24
Systematic Bagging Model – framework definition
25
Systematic Bagging Model – framework definition (cont’d)
26
Systematic Bagging Model – framework definition (cont’d)
27
Systematic Bagging Model – framework definition (cont’d)
28
Time Series Models Evaluation – best practices
29
GIFT-Eval
30
fev-bench
31
Backtesting study setting – LIGHT Benchmark
32
LIGHT Benchmark (cont’d)
Asset categories
Asset categories representants
Testing periods (number, size, horizon & volatility regimes)
VaR & ES
confidence level
Models validation techniques (regulator perspective, statistical tests, loss functions)
Light Benchmark components
33
LIGHT Benchmark (cont’d)
34
LIGHT Benchmark (cont’d)
Table: Cut-off point for 250 day out-of-sample periods (P) per each ticker (T)
35
LIGHT Benchmark (cont’d)
36
LIGHT Benchmark (cont’d)
37
LIGHT Benchmark (cont’d)
38
LIGHT Benchmark scoreboard results – VaR 1%
39
LIGHT Benchmark scoreboard results – VaR 2.5%
40
LIGHT Benchmark scoreboard results –
ES 2.5%
41
42
43
44
45
46
47
48
49
Conclusions
50
References
Basel Committee on Banking Supervision (1996a). Amendment to the capital accord to incorporate market risks.
Basel Committee on Banking Supervision (1996b). Supervisory framework for the use of back-testing in conjunction with the internal models approach to market risk capital requirements.
Basel Committee on Banking Supervision (2017). Basel iii: Finalising post-crisis reforms.
Basel Committee on Banking Supervision (2019). Minimum capital requirements for market risk.
Bates, J. M. and Granger, C. W. (1969). The combination of forecasts. Journal of the operational research society, 20(4):451–468.
Bayer, S. (2018). Combining value-at-risk forecasts using penalized quantile regressions. Econometrics and statistics, 8:56–77.
Bellman, R. and Kalaba, R. (1959). On adaptive control processes. IRE Transactions on Automatic Control, 4(2):1–9.
Bollerslev, T. (1986). Generalized autoregressive conditional heteroskedasticity. Journal of econometrics, 31(3):307–327.
Breiman, L. (1996). Bagging predictors. Machine learning, 24:123–140.
Breiman, L. (2001). Random forests. Machine learning, 45:5–32.
Buczyński, M. and Chlebus, M. (2019). Old-fashioned parametric models are still the best: a comparison of value-at-risk approaches in several volatility states. Journal of Risk Model Validation, 14(2).
Buczyński, M. and Chlebus, M. (2022). The importance of window size: a study on the required window size for optimal-quality market risk models. Journal of Risk Model Validation, 16(1).
Caiado, J., Crato, N., and Peña, D. (2006). A periodogram-based metric for time series classification. Computational Statistics & Data Analysis, 50(10):2668–2684.
Cerqueti, R., et al. (2022). Weighted score-driven fuzzy clustering of time series with a financial application. Expert Systems with Applications, 198:116752.
Christoffersen, P. F. (1998). Evaluating interval forecasts. International economic review, pages 841–862.
Clemen, R. T. and Winkler, R. L. (1986). Combining economic forecasts. Journal of Business & Economic Statistics, 4(1):39–46.
Clements, M. and Hendry, D. F. (1998). Forecasting economic time series. Cambridge University Press.
Cuturi, M. and Blondel, M. (2017). Soft-dtw: a differentiable loss function for time-series. International conference on machine learning, pages 894–903.
Di Sciorio, F. (2023). Clustering analysis on hurst dynamic. Studies of Applied Economics.
Diebold, F. X. and Yılmaz, K. (2014). On the network topology of variance decompositions: Measuring the connectedness of financial firms. Journal of econometrics, 182(1):119–134.
Dowd, K. (2007). Measuring market risk. John Wiley & Sons.
D’Urso, P., De Giovanni, L., and Massari, R. (2016). Garch-based robust clustering of time series. Fuzzy Sets and Systems, 305:1–28.
D’Urso, P. and Maharaj, E. A. (2009). Autocorrelation-based fuzzy clustering of time series. Fuzzy Sets and Systems, 160(24):3565–3589.
Engle, R. F. and Manganelli, S. (2004). Caviar: Conditional autoregressive value at risk by regression quantiles. Journal of business & economic statistics, 22(4):367–381.
Fama, E. F. (1965). The behavior of stock-market prices. The Journal of Business, 38(1):34–105.
Fissler, T. and Ziegel, J. F. (2016). Higher order elicitability and osband’s principle. The Annals of Statistics, 44(4):1680.
Fissler, T., Ziegel, J. F., and Gneiting, T. (2015). Expected shortfall is jointly elicitable with value at risk-implications for backtesting. arXiv preprint arXiv:1507.00244.
Fuertes, A.-M. and Olmo, J. (2013). Optimally harnessing inter-day and intra-day information for daily value-at-risk prediction. International Journal of Forecasting, 29(1):28–42.
Garcia-Jorcano, L. and Novales, A. (2020). A dominance approach for comparing the performance of var forecasting models. Computational Statistics, 35(3):1411–1448.
Garza, A., Challu, C., and Mergenthaler-Canseco, M. (2024). Timegpt-1.
Ghalanos, A. (2020). Introduction to the rugarch package.
Glosten, L. R., Jagannathan, R., and Runkle, D. E. (1993). On the relation between the expected value and the volatility of the nominal excess return on stocks. The journal of finance, 48(5):1779–1801.
Godahewa, R., et al. (2021). Monash time series forecasting archive. arXiv preprint arXiv:2105.06643.
González-Rivera, G., Lee, T.-H., and Mishra, S. (2004). Forecasting volatility: A reality check based on option pricing, utility function, value-at-risk, and predictive likelihood. International Journal of forecasting, 20(4):629–645.
Goodfellow, I., Bengio, Y., and Courville, A. (2016). Deep learning. MIT press.
Görgen, K., Meirer, J., and Schienle, M. (2022). Predicting value at risk for cryptocurrencies with generalized random forests. Available at SSRN 4053537.
Götz, P. (2023). Quantile forecast combination in risk measure forecasting. Available at SSRN 4289578.
Graham, B., et al. (1951). Security analysis: Principles and technique. McGraw-Hill.
Granger, C. W. (1989). Invited review combining forecasts—twenty years later. Journal of forecasting, 8(3):167–173.
Halbleib, R. and Pohlmeier, W. (2012). Improving the value at risk forecasts: Theory and evidence from the financial crisis. Journal of Economic Dynamics and Control, 36(8):1212–1228.
51
References (cont’d)
Hillebrand, E. and Medeiros, M. C. (2010). The benefits of bagging for forecast models of realized volatility. Econometric Reviews, 29(5-6):571–593.
Huang, H. and Lee, T.-H. (2013). Forecasting value-at-risk using high-frequency information. Econometrics, 1(1):127–140.
James, G., et al. (2013). An introduction to statistical learning. Springer.
James, R. and Leung, J. W. Y. (2023). Forecasting financial risk using quantile random forests. Available at SSRN 4324603.
Jeon, J. and Taylor, J. W. (2013). Using caviar models with implied volatility for value-at-risk estimation. Journal of Forecasting, 32(1):62–74.
Jiang, F., Wu, W., and Peng, Z. (2017). A semi-parametric quantile regression random forest approach for evaluating muti-period value at risk. 2017 36th Chinese Control Conference (CCC), pages 5642–5646.
Jorion, P. (2007). Value at risk: the new benchmark for managing financial risk. McGraw-Hill.
Kreiss, J.-P. and Paparoditis, E. (2011). Bootstrap methods for dependent data: A review. Journal of the Korean Statistical Society, 40(4):357–378.
Kupiec, P. (1995). Techniques for verifying the accuracy of risk management. Journal of Derivatives, 3(2).
Lahiri, S. (2003). Resampling methods for dependent data. Springer.
Law, S. M. (2019). STUMPY: A Powerful and Scalable Python Library for Time Series Data Mining. The Journal of Open Source Software, 4(39):1504.
Lazar, E., Pan, J., and Wang, S. (2024). On the estimation of value-at-risk and expected shortfall at extreme levels. Journal of Commodity Markets, page 100391.
Li, Z. (2022). Honour thesis: A joint value at risk and expected shortfall combination framework and its applications in the cryptocurrency market. arXiv preprint arXiv:2202.10918.
Liang, Y., et al. (2024). Foundation models for time series analysis: A tutorial and survey.
Lis, S. and Chlebus, M. (2023). Combining forecasts? keep it simple. Central European Economic Journal, 10(57):343–370.
Longerstaey, J. and Spencer, M. (1996). Riskmetricstm—technical document.
Maharaj, E. A., D’Urso, P., and Caiado, J. (2019). Time series clustering and classification. Chapman and Hall/CRC.
Mattera, R. and Otto, P. (2024). Network log-arch models for forecasting stock market volatility. International Journal of Forecasting.
McAleer, M., Jimenez-Martin, J.-A., and Perez Amaral, T. (2010). Has the basel ii accord encouraged risk management during the 2008-09 financial crisis? Available at SSRN 1397239.
McAleer, M., Jimenez-Martin, J.-A., and Perez-Amaral, T. (2013a). Gfc-robust risk management strategies under the basel accord. International Review of Economics & Finance, 27:97–111.
McAleer, M., Jiménez-Martín, J.-Á., and Pérez-Amaral, T. (2013b). Has the basel accord improved risk management during the global financial crisis? The North American Journal of Economics and Finance, 26:250–265.
McNeil, A. J. and Frey, R. (2000). Estimation of tail-related risk measures for heteroscedastic financial time series: an extreme value approach. Journal of empirical finance, 7(3-4):271–300.
Nadarajah, S., Zhang, B., and Chan, S. (2014). Estimation methods for expected shortfall. Quantitative Finance, 14(2):271–291.
Nolde, N. and Ziegel, J. F. (2017). Elicitability and backtesting: Perspectives for banking regulation.
Opschoor, A., Van Dijk, D., and van der Wel, M. (2017). Combining density forecasts using focused scoring rules. Journal of Applied Econometrics, 32(7):1298–1313.
Pascual, L., Romo, J., and Ruiz, E. (2006). Bootstrap prediction for returns and volatilities in garch models. Computational Statistics & Data Analysis, 50(9):2293–2312.
Petropoulos, F., Hyndman, R. J., and Bergmeir, C. (2018). Exploring the sources of uncertainty: Why does bagging for time series forecasting work? European Journal of Operational Research, 268(2):545–554.
Politis, D. N. and Romano, J. P. (1994). The stationary bootstrap. Journal of the American Statistical association, 89(428):1303–1313.
Rasul, K., et al. (2024). Lag-llama: Towards foundation models for probabilistic time series forecasting.
Rockafellar, R. T. and Uryasev, S. (2002). Conditional value-at-risk for general loss distributions. Journal of banking & finance, 26(7):1443–1471.
Sarykalin, S., Serraino, G., and Uryasev, S. (2008). Value-at-risk vs. conditional value-at-risk in risk management and optimization. State-of-the-art decision-making tools in the information-intensive age, pages 270–294.
Semenoglou, A.-A., et al. (2021). Investigating the accuracy of cross-learning time series forecasting methods. International Journal of Forecasting, 37(3):1072–1084.
Storti, G. and Wang, C. (2022). Nonparametric expected shortfall forecasting incorporating weighted quantiles. International Journal of Forecasting, 38(1):224–239.
Tan, S. Y., et al. (2020). Granger causality based hierarchical time series clustering for state estimation. IFAC-PapersOnLine, 53(2):524–529.
Tavenard, R., et al. (2020). Tslearn, a machine learning toolkit for time series data. Journal of Machine Learning Research, 21(118):1–6.
Taylor, J. W. (2020). Forecast combinations for value at risk and expected shortfall. International Journal of Forecasting, 36(2):428–441.
Timmermann, A. (2006). Forecast combinations. Handbook of economic forecasting, 1:135–196.
Tumminello, M., Lillo, F., and Mantegna, R. N. (2010). Correlation, hierarchies, and networks in financial markets. Journal of economic behavior & organization, 75(1):40–58.
Vogel, R. M. and Shallcross, A. L. (1996). The moving blocks bootstrap versus parametric time series models. Water resources research, 32(6):1875–1882.
Wang, X., et al. (2023). Forecast combinations: An over 50-year review. International Journal of Forecasting, 39(4):1518–1547.
Wolpert, D. H. (1992). Stacked generalization. Neural networks, 5(2):241–259.
Woźniak, M. and Chlebus, M. (2021). HCR & HCR-GARCH-Novel Statistical Learning Models for Value at Risk Estimation.
Woźniak, M. and Ślepaczuk, R. (2023). Light benchmark-comprehensive backtesting framework for market risk models comparison.
Yeh, C.-C. M., et al. (2016). Matrix profile i: all pairs similarity joins for time series: a unifying view that includes motifs, discords and shapelets. 2016 IEEE 16th international conference on data mining (ICDM), pages 1317–1322.
Zhong, S. and Mueen, A. (2024). Mass: distance profile of a query over a time series. Data Mining and Knowledge Discovery, pages 1–27.