2022 Fiscal Year Final Research Report
Asset Prices and Investment Theories Based on Statistical Learning
| Project/Area Number |
20K01587
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 07030:Economic statistics-related
|
|---|
| Research Institution | Hitotsubashi University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2020-04-01 – 2023-03-31
|
| Keywords | 低ベータ・アノマリー / 信用デフォルト・スワップ / ハミルトンニアン・モンテ・カルロ / 分散リスクプレミアム / 歪度リスクプレミアム / 確率的分散 / 自己励起的ジャンプ / ベイズ推定 |
|---|
| Outline of Final Research Achievements |
Based on a paper pointing out that the cause of the low beta anomaly is the credit risk of individual firms, we conducted a study to extract credit information from corporate CDS (Credit Default Swap). Assuming several models of hazard rates, we analyzed CDS of Japanese individual firms and found a model with a high degree of fit. We were able to show that the statistical estimation method, that is, Bayes estimation combined with the numerical solver of ordinary differential equations works when volatility indices and options, for which no analytical solution is available, are used as observables. In the estimation of the dynamic structure of financial time series using statistical learning theory, we estimated Eurozone sovereign bonds with a cointegration structure and solved the Bellman equation to obtain the optimal investment.
|
| Free Research Field |
計量的ファイナンス
|
| Academic Significance and Societal Importance of the Research Achievements |
金融時系列の動的構造として動的ベータ、信用リスク、確率分散、自己・相互励起ジャンプ、共和分構造をもつ確率金利などを研究し、統計的学習の一例としてHamiltonian Monte Carlo法を用いたBayes推定と金融時系列を特徴付ける常微分方程式の数値解法を組み合わせることで統計的推論が可能であることを明らかにした点で学術的意義があると考えられる。金融資産のリスク管理や最適投資問題に対しても応用の可能性を示すことができた点で社会的意義があるものといえよう。
|
