Please enable JavaScript.

Coggle requires JavaScript to display documents.

Academic Monitor - Coggle Diagram

- - - - decision makers exaggerate the probability of extreme events
- - - - Has defensive characteristics
      - Try to avoid momentum crashes
        
        Diversify
        
        Assets
        
        Signals
        
        Rebalance windows
        
        Trend breaks are increasing over time
        
        Short covering in illiquid markets explain momentum crashes
        
        Dynamic scaling
        
        Breakeout/entry filter
      - Model as an option where premium is transaction cost
      - Explained by rational inattention, underreaction
      - Works better when diversified
      - It's convex
      - Not compensated for specification risk
      - Trend following protects you from narratives
    - - International Bonds
      - Repo Markets
    - - Risk reversal premium, investors overpay for puts
    - - Carries a premium that's uncorrelated to VIX
    - - Short duration pays a premium, helps explain other premium (profitability, low risk, payout)
    - - Sentiment beta exists, positive correlation with performance until 4m windows
    - - Regionalize AW
- - - - Forward puzzle, expectation errors comoving with interest rate differential
      - Principal components work to forecast FX
      - Business cycles explain the cross section of FX returns
    - - AIAE (Agg Inv Allo to Equities) is the best out of sample valuation metric
      - conditional autocorrelation in equity markets is negative and gets even more negative around crisis and recessions
    - - Rates skewness predict bond returns, predicts survey forecast errors
      - 3m10y yield curve predicts recessions
      - Strong evidence of regime switches that correlate with business cycle
      - yield curve has momentum, post FOMC drift
      - PCA on bond changes works better than PCA on levels
      - Investors overestimate persistence of term premium and underestimate the persistence of short rate (high autocorrelation, consistent with trend)
    - - Gold/Bitcoin puzzle
      - Liquid futures - trend
    - - Loan prices add information to credit spreads
  - - - Correlation Shrinkage
      - Regularization methods
      - Economic vs statistical clustering
      - Overdispersion of eigenvalues can be shrinked
      - Equal risk contribution is hard to beat
    - - Conditional vol target fails to improve performance
    - - Calendar approach is worse than tolerance bands
      - Rebalancing premium disappears in well diversified risk balanced portfolios
      - Rebalancing = buy and hold + selling a straddle
    - - Stock bond correlations
        
        Generally positive when global monetary policy is restrictive
        
        Explained by: activity, inflation, relative yield, relative vol
      - Equity markets co crashes are more common than bond markets co crashing
- - - - Crypto
      - Semis