{"id":{"repo_id":"ku","oai_identifier":"oai:kuscholarworks.ku.edu:1808/39491"},"canonical_url":"https://search.dev.ndltd.org/etd/ku/oai:kuscholarworks.ku.edu:1808/39491","repository":{"repo_id":"ku","name":"University of Kansas","base_url":"https://kuscholarworks.ku.edu/server/oai/request"},"display":{"title":"Conditional Asset Pricing Models via Machine Learnings for the Chinese Stock Market","abstract":"Motivated by the institutional features and distinctive return dynamics of the Chinese stock market, this dissertation develops a unified conditional asset pricing framework that integrates time-varying risk exposures, distributional asymmetry, and high-dimensional information. The analysis addresses three interrelated challenges in empirical asset pricing: state-dependent factor loadings, cross-sectional dependence driven by latent common shocks, and nonlinear relationships between returns and conditioning information in high-dimensional environments.","abstract_html":"Motivated by the institutional features and distinctive return dynamics of the Chinese stock market, this dissertation develops a unified conditional asset pricing framework that integrates time-varying risk exposures, distributional asymmetry, and high-dimensional information. The analysis addresses three interrelated challenges in empirical asset pricing: state-dependent factor loadings, cross-sectional dependence driven by latent common shocks, and nonlinear relationships between returns and conditioning information in high-dimensional environments.","abstract_has_math":false,"creators":["Jin, Jingwei"],"institution":"University of Kansas","degree_name":"Ph.D.","degree_level":null,"degree_discipline":"Economics","degree_department":null,"school":null,"contributors":[],"advisors":["Cai, Zongwu"],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026-05-31","date_published":"2026-05-31","updated_at":"2026-07-24T02:47:27Z","subjects":["Chinese stock market","conditional asset pricing","conditional CAPM","conditional quantile model","high-dimensional factors","machine learning"],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier.other","label":"Dc Identifier Other","values":["https://www.proquest.com/LegacyDocView/DISSNUM/32696511"],"render_values":[{"text":"https://www.proquest.com/LegacyDocView/DISSNUM/32696511","href":"https://www.proquest.com/LegacyDocView/DISSNUM/32696511","code":true}]}]},"links":{"outbound_url":"https://hdl.handle.net/1808/39491","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Cai, Zongwu"]},{"key":"dc:creator","label":"Author","values":["Jin, Jingwei"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2026-07-15T22:54:12Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2026-07-15T22:54:12Z"]},{"key":"dc:date.issued","label":"Date","values":["2026-05-31"]},{"key":"dc:publisher","label":"Institution","values":["University of Kansas"]},{"key":"dc:type","label":"Dc Type","values":["Dissertation"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Economics"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Chinese stock market","conditional asset pricing","conditional CAPM","conditional quantile model","high-dimensional factors","machine learning"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.other","label":"Dc Identifier Other","values":["https://www.proquest.com/LegacyDocView/DISSNUM/32696511"]},{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/1808/39491"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Motivated by the institutional features and distinctive return dynamics of the Chinese stock market, this dissertation develops a unified conditional asset pricing framework that integrates time-varying risk exposures, distributional asymmetry, and high-dimensional information. The analysis addresses three interrelated challenges in empirical asset pricing: state-dependent factor loadings, cross-sectional dependence driven by latent common shocks, and nonlinear relationships between returns and conditioning information in high-dimensional environments.","The first chapter proposes a conditional factor-augmented CAPM in which factor loadings vary systematically with macroeconomic and financial instruments, and the error structure incorporates unobserved common factors to capture cross-sectional dependence. Empirical evidence supports time-varying and instrument-dependent risk exposures, with conditional homogeneity across assets not rejected. Among the conditioning instruments, inflation, book-to-market ratio, and net equity expansion exhibit the strongest predictive content. Machine learning methods, particularly neural networks, substantially improve out-of-sample performance relative to linear benchmarks at both the stock and portfolio levels.","The second chapter shifts the focus from conditional means to conditional distributions by introducing a conditional quantile framework based on the Fama–French five-factor (FF5) structure. This approach uncovers pronounced tail asymmetry in small-cap stocks, consistent with the presence of shell value and speculative trading. Unlike traditional mean-based models, the quantile specification captures asymmetric downside and upside risks and provides distribution-sensitive signals relevant for portfolio construction.","The third chapter extends the quantile framework to a high-dimensional setting by incorporating a rich set of stock-level characteristics. The expanded information set enhances the identification and quantification of tail asymmetry, with amplification effects concentrated primarily in the upper tail. While consistent with the baseline quantile findings regarding small-cap exposure, the high-dimensional model yields sharper cross-sectional ranking signals and clearer portfolio-relevant implications.","Taken together, the three models offer complementary perspectives on average return dynamics, tail behavior, and investment decision-making. By combining econometric structure with modern machine learning methods, this dissertation contributes to the econometric modeling of conditional asset pricing by jointly addressing time variation, distributional heterogeneity, and latent cross-sectional dependence in high-dimensional panel data."]},{"key":"dc:title","label":"Title","values":["Conditional Asset Pricing Models via Machine Learnings for the Chinese Stock Market"]}]}],"canonical_facts":{"dc:contributor.advisor":["Cai, Zongwu"],"dc:creator":["Jin, Jingwei"],"dc:date.accessioned":["2026-07-15T22:54:12Z"],"dc:date.available":["2026-07-15T22:54:12Z"],"dc:date.issued":["2026-05-31"],"dc:description.abstract":["Motivated by the institutional features and distinctive return dynamics of the Chinese stock market, this dissertation develops a unified conditional asset pricing framework that integrates time-varying risk exposures, distributional asymmetry, and high-dimensional information. The analysis addresses three interrelated challenges in empirical asset pricing: state-dependent factor loadings, cross-sectional dependence driven by latent common shocks, and nonlinear relationships between returns and conditioning information in high-dimensional environments.","The first chapter proposes a conditional factor-augmented CAPM in which factor loadings vary systematically with macroeconomic and financial instruments, and the error structure incorporates unobserved common factors to capture cross-sectional dependence. Empirical evidence supports time-varying and instrument-dependent risk exposures, with conditional homogeneity across assets not rejected. Among the conditioning instruments, inflation, book-to-market ratio, and net equity expansion exhibit the strongest predictive content. Machine learning methods, particularly neural networks, substantially improve out-of-sample performance relative to linear benchmarks at both the stock and portfolio levels.","The second chapter shifts the focus from conditional means to conditional distributions by introducing a conditional quantile framework based on the Fama–French five-factor (FF5) structure. This approach uncovers pronounced tail asymmetry in small-cap stocks, consistent with the presence of shell value and speculative trading. Unlike traditional mean-based models, the quantile specification captures asymmetric downside and upside risks and provides distribution-sensitive signals relevant for portfolio construction.","The third chapter extends the quantile framework to a high-dimensional setting by incorporating a rich set of stock-level characteristics. The expanded information set enhances the identification and quantification of tail asymmetry, with amplification effects concentrated primarily in the upper tail. While consistent with the baseline quantile findings regarding small-cap exposure, the high-dimensional model yields sharper cross-sectional ranking signals and clearer portfolio-relevant implications.","Taken together, the three models offer complementary perspectives on average return dynamics, tail behavior, and investment decision-making. By combining econometric structure with modern machine learning methods, this dissertation contributes to the econometric modeling of conditional asset pricing by jointly addressing time variation, distributional heterogeneity, and latent cross-sectional dependence in high-dimensional panel data."],"dc:identifier.other":["https://www.proquest.com/LegacyDocView/DISSNUM/32696511"],"dc:identifier.uri":["https://hdl.handle.net/1808/39491"],"dc:language.iso":["en"],"dc:publisher":["University of Kansas"],"dc:subject":["Chinese stock market","conditional asset pricing","conditional CAPM","conditional quantile model","high-dimensional factors","machine learning"],"dc:title":["Conditional Asset Pricing Models via Machine Learnings for the Chinese Stock Market"],"dc:type":["Dissertation"],"thesis:degree_discipline":["Economics"],"thesis:degree_name":["Ph.D."]},"updated_at":"2026-07-24T02:47:27Z"}