Published
Nonlinear Macroeconomic Time Series Forecasting with TSARX: An Empirical Comparison Across Three Economies
Pronósticos de series de tiempo macroeconómicas no lineales con TSARX: un estudio comparativo en tres economías
DOI:
https://doi.org/10.15446/rce.v48n3.123655Keywords:
Macroeconomic time series, Nonlinearity, Predictive ability, Seasonality, Threshold models. (en)Capacidad predictiva, Estacionalidad, Modelos con umbrales, No linealidad, Series de tiempo macroeconómicas. (es)
Downloads
The nonlinearity and seasonality of the business cycle account for the majority of short-run movements in quarterly or monthly macroeconomic time series. The multiplicative seasonal threshold autoregressive model with exogenous inputs (TSARX) combines threshold dynamics, multiplicative seasonality, and external regressor, the model is a special case of a general nonmultiplicative TAR model, but there is limited evidence on how it performs in forecasting relative to simpler models and machine-learning algorithms. We evaluate the performance out-of-sample of TSARX models using seasonally unadjusted macroeconomic time series for three economies (Colombia, the United States of America, and the United Kingdom) and three key variables (gross domestic product (GDP), unemployment rate, and inflation). For each country-variable pair, we compare the TSARX models with four competitor models: a TAR model, a linear seasonal autoregressive model (SAR), Holt-Winters exponential smoothing (ES), and long short-term memory (LSTM) neural networks. Multi-step forecasts at horizons from one to four periods ahead are produced under a rolling-window design, and accuracy is assessed using Mean Squared Error (MSE) and Diebold-Mariano tests (DM) for equal predictive ability. Across 36 series-country-horizon combinations, the TSARX achieves the lowest MSE in three cases and is often statistically indistinguishable from the best benchmark; in many situations simpler models remain di‑cult to beat. These findings show that additional nonlinear and seasonal structure does not guaranty superior forecasts and that the benefits of the TSARX models are context- and horizon-dependent.
La no linealidad y la estacionalidad del ciclo económico explican la mayor parte de los movimientos de corto plazo en las series de tiempo macroeconómicas trimestrales o mensuales. El modelo autorregresivo con umbrales y estacionalidad multiplicativa con regresores exógenos (TSARX) combina dinámica por regímenes, estacionalidad multiplicativa y variables explicativas externas, el cual es un caso especial de un modelo TAR no multiplicativo general, pero existe poca evidencia sobre su desempeño en pronósticos frente a modelos más simples y algoritmos de aprendizaje automático. En este trabajo evaluamos el desempeño fuera de muestra de los modelos TSARX utilizando series de tiempo macroeconómicas sin desestacionalizar de tres economías (Colombia, Estados Unidos de América y Reino Unido) y tres variables clave (producto interno bruto, tasa de desempleo e inflación). Para cada combinación país-variable comparamos los modelos TSARX con cuatro modelos competidores: un modelo TAR, un modelo autorregresivo estacional lineal (SAR), el suavizamiento exponencial de Holt y Winters (ES) y redes neuronales de memoria de largo y corto plazo (LSTM). Se generan pronósticos multi paso a horizontes de uno a cuatro períodos adelante bajo un esquema de ventana rodante, y la precisión se evalúa mediante el Error Cuadrático Medio (MSE) y pruebas de Diebold y Mariano (DM) de igualdad de capacidad predictiva. En 36 combinaciones serie-país-horizonte, los modelos TSARX obtienen el menor MSE en tres casos y usualmente es estadísticamente indistinguible del mejor modelo de referencia; en muchas situaciones modelos más simples siguen siendo difíciles de superar. Estos resultados muestran que añadir estructura no lineal y estacional no garantiza mejores pronósticos y que las ganancias de los modelos TSARX dependen del contexto y del horizonte.
References
Arbeláez Quintero, S. (2022), `Pronósticos en series de tiempo no lineales: aplicación del modelo tsarx y comparación con modelos para datos estacionales', Tesis de Maestría, Repositorio Universidad Nacional de Colombia. https://repositorio.unal.edu.co/handle/unal/83875
Brown, R. G. (1959), Statistical Forecasting for Inventory Control, McGraw-Hill.
Brown, R. L., Durbin, J. & Evans, J. M. (1975), `Techniques for testing the constancy of regression relationships over time', Journal of the Royal Statistical Society: Series B (Methodological) 37(2), 149-192.
Chan, K. S. (1993), `Consistency and limiting distribution of the least squares estimator of a threshold autoregressive model', The Annals of Statistics 21(1), 520-533.
Clements, M. P. & Krolzig, H.-M. (1998), `A comparison of the forecast performance of markov-switching and threshold autoregressive models of us gnp', The Econometrics Journal 1(1), C47-C75.
DANE (2015), `Metodología general: Gran encuesta integrada de hogares (geih)', Documento metodológico.
DANE (2016), `Metodología general: Indicador de seguimiento a la economía (ise)', Documento metodológico.
De Gooijer, J. G. & Vidiella-i Anguera, A. (2003), `Nonlinear stochastic inflation modelling using seasetars', Insurance: Mathematics and Economics 32(1), 27-36.
Diebold, F. X. & Mariano, R. S. (1995), `Comparing predictive accuracy', Journal of Business & Economic Statistics 13(3), 253-263.
Dixon, M. F., Halperin, I. & Bilokon, P. (2020), Machine Learning in Finance: From Theory to Practice, Springer, Cham.
Edgerton, D. & Wells, C. (1994), `Critical values for the cusumsq statistic in medium and large sized samples', Oxford Bulletin of Economics and Statistics 56(3), 355-365.
Franses, P. H. & van Dijk, D. (2000), Nonlinear Time Series Models in Empirical Finance, Cambridge University Press, Cambridge.
Ghysels, E., Osborn, D. R. & Rodrigues, P. M. M. (2006), Forecasting seasonal time series, in G. Elliott, C. W. J. Granger & A. Timmermann, eds, `Handbook of Economic Forecasting', Vol. 1, Elsevier, pp. 659-711.
González, J. & Nieto, F. H. (2020), `Bayesian analysis of multiplicative seasonal threshold autoregressive processes', Revista Colombiana de Estadística 43(2), 251-285.
Hansen, B. E. (2011), `Threshold autoregression in economics', Statistics and Its Interface 4, 123-127.
Hochreiter, S. & Schmidhuber, J. (1997), `Long short-term memory', Neural Computation 9(8), 1735-1780.
Holt, C. C. (1957), Forecasting seasonals and trends by exponentially weighted averages, in `O ce of Naval Research Memorandum', Carnegie Institute of Technology, Pittsburgh.
Hyndman, R. J. & Athanasopoulos, G. (2021), Forecasting: Principles and Practice, 3rd edn, OTexts, Melbourne. https://otexts.com/fpp3/
Hyndman, R. J. & Koehler, A. B. (2006), `Another look at measures of forecast accuracy', International Journal of Forecasting 22(4), 679-688.
Lee, Y.-M. & Wang, K.-M. (2012), `Searching for a better proxy for business cycles: With supports using us data', Applied Economics 44(11), 1433-1442.
Makridakis, S., Spiliotis, E. & Assimakopoulos, V. (2020), `The m4 competition: 100,000 time series and 61 forecasting methods', International Journal of Forecasting 36(1), 54-74.
Milas, C., Rothman, P. & van Dijk, D. (2006), Nonlinear Time Series Analysis of Business Cycles, Elsevier, Amsterdam.
Montgomery, A. L., Zarnowitz, V., Tsay, R. S. & Tiao, G. C. (1998), `Forecasting the u.s. unemployment rate', Journal of the American Statistical Association 93(442), 478-493.
Tashman, L. J. (2000), `Out-of-sample tests of forecasting accuracy: An analysis and review', International Journal of Forecasting 16(4), 437-450.
Tong, H. (1990), Non-linear Time Series: A Dynamical System Approach, Vol. 6 of Oxford Statistical Science Series, Oxford University Press, Oxford.
Tong, H. & Lim, K. S. (1980), `Threshold autoregression, limit cycles, and cyclical data', Journal of the Royal Statistical Society: Series B (Methodological) 42(3), 245-292.
Tsay, R. S. (1989), `Testing and modeling threshold autoregressive processes', Journal of the Royal Statistical Society: Series B (Methodological) 51(1), 231-246.
Tsay, R. S. (1998), `Testing and modeling multivariate threshold models', Journal of the American Statistical Association 93(443), 1188-1202.
Vaca, P. A. (2018), Analysis of the forecasting performance of the threshold autoregressive model, Master's thesis, Universidad Nacional de Colombia.
Winters, P. R. (1960), `Forecasting sales by exponentially weighted moving averages', Management Science 6(3), 324-342.
Yilmaz, F. M. & Arabaci, O. (2021), `Should deep learning models be in high demand, or should they simply be a very hot topic? a comprehensive study for exchange rate forecasting', Computational Economics 57(1), 217-245.
Zivot, E. & Wang, J. (2006), Modeling Financial Time Series with S-PLUS, 2nd edn, Springer, New York, NY.
How to Cite
APA
ACM
ACS
ABNT
Chicago
Harvard
IEEE
MLA
Turabian
Vancouver
Download Citation
License

This work is licensed under a Creative Commons Attribution 4.0 International License.
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).






