Forecasting large covariance matrix with high-frequency data: A factor correlation matrix approach
We propose a factor correlation matrix approach to forecast large covariance matrix of asset returns using high-frequency data. We apply shrinkage method to estimate large correlation matrix and adopt principal component method to model the underlying latent factors. A vector autoregressive model is...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | text |
| Language: | English |
| Published: |
Institutional Knowledge at Singapore Management University
2018
|
| Subjects: | |
| Online Access: | https://ink.library.smu.edu.sg/soe_research/2270 https://ink.library.smu.edu.sg/context/soe_research/article/3269/viewcontent/FCM_20181025.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Singapore Management University |
| Language: | English |
| Summary: | We propose a factor correlation matrix approach to forecast large covariance matrix of asset returns using high-frequency data. We apply shrinkage method to estimate large correlation matrix and adopt principal component method to model the underlying latent factors. A vector autoregressive model is used to forecast the latent factors and hence the large correlation matrix. The realized variances are separately forecasted using the Heterogeneous Autoregressive model. The forecasted variances and correlations are then combined to forecast large covariance matrix. We conduct Monte Carlo studies to compare the finite sample performance of several methods of forecasting large covariance matrix. Our proposed method is found to perform better in reporting smaller forecast errors. Empirical application to a portfolio of 100 NYSE and NASDAQ stocks shows that our method provides lower out-of-sample realized variance in selecting global minimum variance portfolio. It also provides higher information ratio for Markowitz portfolios. |
|---|