| 研究課題/領域番号 |
19K13671
|
|---|
| 研究機関 | 慶應義塾大学 |
|---|
研究代表者 |
クリネ シモン 慶應義塾大学, 経済学部(三田), 特任講師 (70814797)
|
|---|
| 研究期間 (年度) |
2019-04-01 – 2021-03-31
|
| キーワード | High-Frequency data / Microstructure noise / Statistical inference / Noise-Robust estimation / Parametric noise / Plug-In estimation |
|---|
| 研究実績の概要 |
Of the three proposed projects, the first two are on their way to publication. The first one, based on an empirical application of microstructure noise regression and which was already quite advanced at the beginning of FY2019, is accepted in the Journal of Business and Economic Statistics [1]. The second one, on theoretical properties of noise-robust estimators, is in revision for the Annals of the Institute of Statistical Mathematics. The original approach using LASSO estimation was replaced by a more direct plug-in method.
I have also attended several conferences, in particular: EcoSta2019 (Taichung); CMStats (London).
[1] Clinet, S. and Potiron, Y. Disentangling sources of high frequency market microstructure noise, Journal of Business and Economic Statistics (To Appear)
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
Since two projects are on a good way for publication, things are going smoothly although there are questions left to answer and investigate. The third project remains for now completely open, and the second project needs to be carefully revised for publication.
In particular for project 2, there are a certain number of difficulties that have arisen in the theoretical method that need to be dealt with, but the way to solve them seems clear for now. In particular, the LASSO approach proposed to solve the second problem presented difficulties that suggested to change our estimation method. The new robust plug-in method is now more efficient and well-understood.
|
| 今後の研究の推進方策 |
For now I am working with my collaborator on the revision of Project 2. This is a substantial revision which should take a few months.
Next, I will start the investigation of the third project and investigate how it can be designed efficiently in the light of the first two projects. This will require frequent skype meetings with my collaborator Mathieu Rosenbaum.
I will also investigate the theoretical and practical implications of the first two projects on statistical inference for high-frequency data. Indeed, the general method I have used for high-frequency estimation may also work for other closely related cases, such as high-frequency cointegration estimation. A confirmation of this potential would be a strong point in favor of the method, and I think this is worth investigation.
|
| 次年度使用額が生じた理由 |
For the upcoming fiscal year, I plan to present as much as possible the fruitful results of FY2019 previously described (two projects on their way to completion). I therefore plan to give several talks in international conferences and workshops by the end of FY2020, such as Bernoulli, Econometrics Society, and Bachelier conferences. I will also try to meet my collaborator Mathieu Rosenbaum in Paris at least once to make progress in our common project, and I plan to invite Professor Frederic Abergel from Paris as well for consultation.
Finally, although I did not have to use research assistants during FY2019 and managed to work only with my collaborator Yoann Potiron, I may need to hire a research assistant for coding tasks (with R or Python) related to large-scale statistical inference.
|
