Good afternoon!
Consider a problem: a panel regression is fitted to predict corporate bond yield credit spreads. Under "corporate bond spreads" term I mean difference between yield to maturity of a corporate bond and yield to maturity of a government bond with similar duration (to be more precize, instead of government bond is used value from zero-coupon curve, so the spread I am talking about is G-spread).
The bond spread (observed daily for different corporate bonds) is target variable, and explanatory variables are some financial ratios of corporations (Net Debt to EBITDA etc.), explanatory variables are observed quarterly. My question is: how can we efficiently use these low-frequency regressors to predict high-frequency spread? I found only this link, but it's not quite informative (no links on any papers or software realizations is provided there). I am familiar with MIDAS, but the task I describe seems to be "reverse MIDAS". Any suggestions?
Thank you for any ideas in advance.
