Four essays on the application of nonlinear techniques to the time series in Economics

  • Autor: Salvador Ramallo Ros
  • Director/es: Máximo Camacho Alonso
  • Defensa: 30/11/2023 - Universidad de Murcia
  • Tribunal: Gabriel Pérez Quirós, Mª Luz Maté Sánchez-Val, Jesús Crespo Cuaresma
  • Calificación: Sobresaliente cum laude
  • Ver publicaciones relacionadas

Resumen

Durante años, la modelización lineal ha sido en econometría la principal aproximación para modelizar dinámicas y relaciones entre variables, debido a su carácter sencillo y accesible tanto en términos de interpretación como de computación, como a la tradicional escasez de datos, tanto en corte transversal como longitudinal. Sin embargo, la economía es habitual que no se ajuste a un comportamiento lineal, dado que se trata de un sistema complejo con múltiples variables interdependientes que interaccionan de formas muy distintas: desde la descripción de las respuestas a perturbaciones en series persistentes mediante funciones de impulso respuesta no lineales, hasta la respuesta de mercados bursátiles a las perturbaciones del precio del petróleo. El aumento reciente de la capacidad de generación y almacenamiento de dato, y la mayor capacidad de computación, han supuesto un escenario donde la modelización no lineal puede ayudar a revelar dichas relaciones. Esta tesis tiene como objetivo ilustrar las ventajas de una correcta aplicación de técnicas no lineales en cuatro problemas en la que se modelizan dinámicas de series temporales aplicadas a problemas económicos. En primer lugar, se plantea una modelización no lineal univariante, de carácter no paramétrico, a la predicción de recesiones económicas. Esta modelización viene motivada por la observación influyente que supuso la pandemia derivada del Covid-19, de magnitud económica sin precedentes, y que supuso que distintos modelos sufriesen en la estimación a partir de su incorporación. En particular, en base al caso particular del crecimiento del PIB, la serie se particiona en tramos de una longitud determinada, y se analiza la cantidad que llevarían a una recesión técnica una vez se ponderan por la probabilidad de ocurrencia de los tramos una vez se han embebido en un espacio de símbolos. Se muestra la robustez de la aproximación, con una capacidad predictiva mejor que una aproximación autorregresiva lineal o que una aproximación con régimen cambiante en base a una cadena de Markov. En segundo lugar, se analizan los determinantes del ciclo económico en España mediante la técnica de árboles de decisión basados en boosting. En particular, tras analizar la capacidad de la técnica prediciendo las recesiones fechadas por el Comité de Fechado de Ciclos, se analizan las variables que ayudan a la predicción de una correcta probabilidad de recesión, y se analizan las interacciones entre dichas variables y se analiza de forma dinámica su importancia. Variables como precios o indicadores adelantados de tendencia del PIB o venta de coches suelen ser relevantes, si bien en recesiones como la Gran Recesión variables relacionadas con sectores más afectados como la construcción ganan importancia respecto a estas. En tercer lugar, se analiza la dinámica de homicidios con arma en Estados Unidos, un problema relevante en dicho país para aseguradoras o inmobiliarias, mediante un modelo de factor dinámico con un filtro de Kalman con características no lineales para aprovechar datos con disponibilidad casi inmediata frente al retraso de los datos oficiales, de hasta casi dos años. Los resultados mejoran la capacidad predictiva de modelos lineales y de modelos basados en aprendizaje automático. Por último, se analizan las interdependencias entre once países de la Unión Monetaria y Económica europea mediante la aplicación de un modelo de panel no lineal para tratar de explicar su dinámica, mediante el que se crea un índice de conectividad con dos regímenes diferenciados. Destaca la importancia de variables como exportaciones o turismo bilaterales para la descripción de la dinámica. Por lo tanto, este trabajo ilustra en cuatro casos particulares las ventajas que supone el uso correcto de aproximaciones no lineales en problemas económicos, ya sean de carácter univariante, multivariante o incluso en problemas de datos de panel.

WP_Query Object ( [query] => Array ( [post_type] => publicaciones [meta_key] => _wpcf_belongs_tesis_id [meta_value] => 3807 ) [query_vars] => Array ( [post_type] => publicaciones [meta_key] => _wpcf_belongs_tesis_id [meta_value] => 3807 [error] => [m] => [p] => 0 [post_parent] => [subpost] => [subpost_id] => [attachment] => [attachment_id] => 0 [name] => [pagename] => [page_id] => 0 [second] => [minute] => [hour] => [day] => 0 [monthnum] => 0 [year] => 0 [w] => 0 [category_name] => [tag] => [cat] => [tag_id] => [author] => [author_name] => [feed] => [tb] => [paged] => 0 [preview] => [s] => [sentence] => [title] => [fields] => [menu_order] => [embed] => [category__in] => Array ( ) [category__not_in] => Array ( ) [category__and] => Array ( ) [post__in] => Array ( ) [post__not_in] => Array ( ) [post_name__in] => Array ( ) [tag__in] => Array ( ) [tag__not_in] => Array ( ) [tag__and] => Array ( ) [tag_slug__in] => Array ( ) [tag_slug__and] => Array ( ) [post_parent__in] => Array ( ) [post_parent__not_in] => Array ( ) [author__in] => Array ( ) [author__not_in] => Array ( ) [ignore_sticky_posts] => [suppress_filters] => [cache_results] => 1 [update_post_term_cache] => 1 [lazy_load_term_meta] => 1 [update_post_meta_cache] => 1 [posts_per_page] => 10 [nopaging] => [comments_per_page] => 50 [no_found_rows] => [order] => DESC ) [tax_query] => WP_Tax_Query Object ( [queries] => Array ( ) [relation] => AND [table_aliases:protected] => Array ( ) [queried_terms] => Array ( ) [primary_table] => wp_posts [primary_id_column] => ID ) [meta_query] => WP_Meta_Query Object ( [queries] => Array ( [0] => Array ( [key] => _wpcf_belongs_tesis_id [value] => 3807 ) [relation] => OR ) [relation] => AND [meta_table] => wp_postmeta [meta_id_column] => post_id [primary_table] => wp_posts [primary_id_column] => ID [table_aliases:protected] => Array ( [0] => wp_postmeta ) [clauses:protected] => Array ( [wp_postmeta] => Array ( [key] => _wpcf_belongs_tesis_id [value] => 3807 [compare] => = [compare_key] => = [alias] => wp_postmeta [cast] => CHAR ) ) [has_or_relation:protected] => ) [date_query] => [request] => SELECT SQL_CALC_FOUND_ROWS wp_posts.ID FROM wp_posts INNER JOIN wp_postmeta ON ( wp_posts.ID = wp_postmeta.post_id ) WHERE 1=1 AND ( ( wp_postmeta.meta_key = '_wpcf_belongs_tesis_id' AND wp_postmeta.meta_value = '3807' ) ) AND ((wp_posts.post_type = 'publicaciones' AND (wp_posts.post_status = 'publish'))) GROUP BY wp_posts.ID ORDER BY wp_posts.post_date DESC LIMIT 0, 10 [posts] => Array ( [0] => WP_Post Object ( [ID] => 3793 [post_author] => 19 [post_date] => 2023-12-12 12:35:55 [post_date_gmt] => 2023-12-12 11:35:55 [post_content] => Abstract Purpose: Research on temporal dynamics of crime in the United States is growing. Yet, mathematical tools to reliably predict homicides with firearm are still lacking, due to delays in the release of official data lagging up to almost two years. This study takes a critical step in this direction by establishing a reliable statistical tool to predict homicides with firearm at a monthly resolution, combining official data and easy-to-access explanatory variables. Method: We propose a dynamic factor model to predict homicides with firearm from 1999 to 2020 using official monthly data released yearly by the Centers for Disease Control and Prevention, provisional quarterly data from the same agencies, media output from newspapers, and crowdsourced information from the Guns Violence Archive. Results: Statistical findings demonstrate that the dynamic factor model outperforms state-of-the-art techniques (AI and classical autoregressive models). The dynamic factor model offers improved ability to backcast, nowcast, and forecast homicides with firearm, and can anticipate sudden changes in the time-series. Conclusions: By decomposing the time-series of homicides with firearm on common and idiosyncratic components, the dynamic factor model successfully captures their complex time-evolution. This approach offers a vantage point to policymakers and practitioners, allowing for timely predictions, otherwise unfeasible. [post_title] => A dynamic factor model to predict homicides with firearm in the United States [post_excerpt] => [post_status] => publish [comment_status] => closed [ping_status] => closed [post_password] => [post_name] => a-dynamic-factor-model-to-predict-homicides-with-firearm-in-the-united-states [to_ping] => [pinged] => [post_modified] => 2023-12-22 12:52:53 [post_modified_gmt] => 2023-12-22 11:52:53 [post_content_filtered] => [post_parent] => 0 [guid] => http://doctoradodecide.com/?post_type=publicaciones&p=3793 [menu_order] => 0 [post_type] => publicaciones [post_mime_type] => [comment_count] => 0 [filter] => raw ) ) [post_count] => 1 [current_post] => -1 [in_the_loop] => [post] => WP_Post Object ( [ID] => 3793 [post_author] => 19 [post_date] => 2023-12-12 12:35:55 [post_date_gmt] => 2023-12-12 11:35:55 [post_content] => Abstract Purpose: Research on temporal dynamics of crime in the United States is growing. Yet, mathematical tools to reliably predict homicides with firearm are still lacking, due to delays in the release of official data lagging up to almost two years. This study takes a critical step in this direction by establishing a reliable statistical tool to predict homicides with firearm at a monthly resolution, combining official data and easy-to-access explanatory variables. Method: We propose a dynamic factor model to predict homicides with firearm from 1999 to 2020 using official monthly data released yearly by the Centers for Disease Control and Prevention, provisional quarterly data from the same agencies, media output from newspapers, and crowdsourced information from the Guns Violence Archive. Results: Statistical findings demonstrate that the dynamic factor model outperforms state-of-the-art techniques (AI and classical autoregressive models). The dynamic factor model offers improved ability to backcast, nowcast, and forecast homicides with firearm, and can anticipate sudden changes in the time-series. Conclusions: By decomposing the time-series of homicides with firearm on common and idiosyncratic components, the dynamic factor model successfully captures their complex time-evolution. This approach offers a vantage point to policymakers and practitioners, allowing for timely predictions, otherwise unfeasible. [post_title] => A dynamic factor model to predict homicides with firearm in the United States [post_excerpt] => [post_status] => publish [comment_status] => closed [ping_status] => closed [post_password] => [post_name] => a-dynamic-factor-model-to-predict-homicides-with-firearm-in-the-united-states [to_ping] => [pinged] => [post_modified] => 2023-12-22 12:52:53 [post_modified_gmt] => 2023-12-22 11:52:53 [post_content_filtered] => [post_parent] => 0 [guid] => http://doctoradodecide.com/?post_type=publicaciones&p=3793 [menu_order] => 0 [post_type] => publicaciones [post_mime_type] => [comment_count] => 0 [filter] => raw ) [comment_count] => 0 [current_comment] => -1 [found_posts] => 1 [max_num_pages] => 1 [max_num_comment_pages] => 0 [is_single] => [is_preview] => [is_page] => [is_archive] => 1 [is_date] => [is_year] => [is_month] => [is_day] => [is_time] => [is_author] => [is_category] => [is_tag] => [is_tax] => [is_search] => [is_feed] => [is_comment_feed] => [is_trackback] => [is_home] => [is_privacy_policy] => [is_404] => [is_embed] => [is_paged] => [is_admin] => [is_attachment] => [is_singular] => [is_robots] => [is_favicon] => [is_posts_page] => [is_post_type_archive] => 1 [query_vars_hash:WP_Query:private] => c37461c311c12c931ebb3afb2ca28d77 [query_vars_changed:WP_Query:private] => [thumbnails_cached] => [allow_query_attachment_by_filename:protected] => [stopwords:WP_Query:private] => [compat_fields:WP_Query:private] => Array ( [0] => query_vars_hash [1] => query_vars_changed ) [compat_methods:WP_Query:private] => Array ( [0] => init_query_flags [1] => parse_tax_query ) )

Publicaciones relacionadas

Array ( [_edit_lock] => Array ( [0] => 1720167527:19 ) [_edit_last] => Array ( [0] => 19 ) [_wpcf-autor-sort-order] => Array ( [0] => a:1:{i:0;i:27604;} ) [wpcf-autor] => Array ( [0] => Salvador Ramallo Ros ) [wpcf-director] => Array ( [0] => Máximo Camacho Alonso ) [wpcf-fecha-defensa] => Array ( [0] => 1701302400 ) [wpcf-lugar] => Array ( [0] => Universidad de Murcia ) [wpcf-tribunal] => Array ( [0] => Gabriel Pérez Quirós Mª Luz Maté Sánchez-Val Jesús Crespo Cuaresma ) [wpcf-calificacion] => Array ( [0] => 10 ) )