The REIT paper and the Chinese sector paper seem to be a little at odds, yeah? One says "add complexity via regime detection" and the other says "complexity backfires." I think a little complexity in our models is inevitable though - it does make me wonder if the real lesson is about where complexity belongs in the pipeline, not just how much.
Like everything in life, I think there is a balance. Additionally, this balance may not be the same for all asset classes or geographical sectors. For example, some assets may be more efficient than others, and require other techniques.
Strong curation. The AI-PE feedback loop paper nails a systemic risk most analysts miss by treating credit and equity silos separately. That amplification mechanism where IT spending cuts loop back into tech valuations is what made 2008 so brutal when housing contagion spread beyond mortgages. The inverse variance finding for Chinese sectors is also intresting, basically confirming that sophistication collapses when estimation error dominates signal quality.
The REIT paper and the Chinese sector paper seem to be a little at odds, yeah? One says "add complexity via regime detection" and the other says "complexity backfires." I think a little complexity in our models is inevitable though - it does make me wonder if the real lesson is about where complexity belongs in the pipeline, not just how much.
Like everything in life, I think there is a balance. Additionally, this balance may not be the same for all asset classes or geographical sectors. For example, some assets may be more efficient than others, and require other techniques.
Strong curation. The AI-PE feedback loop paper nails a systemic risk most analysts miss by treating credit and equity silos separately. That amplification mechanism where IT spending cuts loop back into tech valuations is what made 2008 so brutal when housing contagion spread beyond mortgages. The inverse variance finding for Chinese sectors is also intresting, basically confirming that sophistication collapses when estimation error dominates signal quality.