Showing posts with label Mixed-Frquency. Show all posts
Showing posts with label Mixed-Frquency. Show all posts
Saturday, March 23, 2019
Big Data in Dynamic Predictive Modeling
Our Journal of Econometrics issue, Big Data in Dynamic Predictive Econometric Modeling, is now in press. It is partly based on a Penn conference, generously supported by Penn's Warren Center for Network and Data Sciences, University of Chicago's Stevanovich Center for Financial Mathematics, and Penn's Institute for Economic Research. The intro is here and the paper list is here.
Sunday, January 27, 2019
Mixed-Frequency Big Data
Monday, April 9, 2018
An Art Market Return Index
Rare and collectible goods, from fine art to fine wine, have many interesting and special aspects. Some are shared and some are idiosyncratic.
From the vantage point of alternative investments (among other things), it would be useful to have high-frequency indices for those asset markets, just as we do for traditional "financial" asset markets like equities.
Along those lines, in "Monthly Art Market Returns" Bocart, Ghysels, and Hafner develop a high-frequency measurement approach, despite the fact that art sales generally occur very infrequently. Effectively they develop a mixed-frequency repeat-sales model, which captures the correlation between art prices and other liquid asset prices that are observed much more frequently. They use the model to extract a monthly art market return index, as well as sub-indices for contemporary art, impressionist art, etc.
Quite fascinating and refreshingly novel.
From the vantage point of alternative investments (among other things), it would be useful to have high-frequency indices for those asset markets, just as we do for traditional "financial" asset markets like equities.
Along those lines, in "Monthly Art Market Returns" Bocart, Ghysels, and Hafner develop a high-frequency measurement approach, despite the fact that art sales generally occur very infrequently. Effectively they develop a mixed-frequency repeat-sales model, which captures the correlation between art prices and other liquid asset prices that are observed much more frequently. They use the model to extract a monthly art market return index, as well as sub-indices for contemporary art, impressionist art, etc.
Quite fascinating and refreshingly novel.
Sunday, November 26, 2017
Modeling With Mixed-Frequency Data
Here's a bit more related to the FRB St. Louis conference.
The fully-correct approach to mixed-frequency time-series modeling is: (1) write out the state-space system at the highest available data frequency or higher (e.g., even if your highest frequency is weekly, you might want to write the system daily to account for different numbers of days in different months), (2) appropriately treat most of the lower-frequency data as missing and handle it optimally using the appropriate filter (e.g., the Kalman filter in the linear-Gaussian case). My favorite example (no surprise) is here.
Until recently, however, the prescription above was limited in practice to low-dimensional linear-Gaussian environments, and even there it can be tedious to implement if one insists on MLE. Hence the well-deserved popularity of the MIDAS approach to approximating the prescription, recently also in high-dimensional environments.
But now the sands are shifting. Recent work enables exact Bayesian posterior mixed-frequency analysis even in high-dimensional structural models. I've known Schorfheide-Song (2015, JBES; 2013 working paper version here) for a long time, but I never fully appreciated the breakthrough that it represents -- that is, how straightforward exact mixed-frequency estimation is becoming -- until I saw the stimulating Justiniano presentation at FRBSL (older 2016 version here). And now it's working its way into important substantive applications, as in Schorfheide-Song-Yaron (2017, forthcoming in Econometrica).
The fully-correct approach to mixed-frequency time-series modeling is: (1) write out the state-space system at the highest available data frequency or higher (e.g., even if your highest frequency is weekly, you might want to write the system daily to account for different numbers of days in different months), (2) appropriately treat most of the lower-frequency data as missing and handle it optimally using the appropriate filter (e.g., the Kalman filter in the linear-Gaussian case). My favorite example (no surprise) is here.
Until recently, however, the prescription above was limited in practice to low-dimensional linear-Gaussian environments, and even there it can be tedious to implement if one insists on MLE. Hence the well-deserved popularity of the MIDAS approach to approximating the prescription, recently also in high-dimensional environments.
But now the sands are shifting. Recent work enables exact Bayesian posterior mixed-frequency analysis even in high-dimensional structural models. I've known Schorfheide-Song (2015, JBES; 2013 working paper version here) for a long time, but I never fully appreciated the breakthrough that it represents -- that is, how straightforward exact mixed-frequency estimation is becoming -- until I saw the stimulating Justiniano presentation at FRBSL (older 2016 version here). And now it's working its way into important substantive applications, as in Schorfheide-Song-Yaron (2017, forthcoming in Econometrica).
Subscribe to:
Posts (Atom)