Nnamdi's Notes

Search

SearchSearch
      • A Local Projections Approach to Difference-in-Differences Event Studies
      • A Neural Phillips Curve and a Deep Output Gap
      • agency
      • An Exploration of Trend-Cycle Decomposition Methodologies in Simulated Data
      • Autoregressive models
      • Ben Franklin's Autoencoder
      • Beveridge-Nelson Decomposition
      • Boosting: Why You Can Use the Hp Filter
      • Career success and career capital are cointegrated
      • Careers are a random walk
      • Change leads to insight far more often than insight leads to change
      • Correlation between prices and quantities reveals main market driver
      • Diffusion models
      • Don't detrend random walks with linear trends
      • Don't rely on inspiration to write
      • Dynamic covariate balancing: Estimating treatment effects over time with potential local projections
      • Dynamic factor model
      • Dynamic Factor Models, Factor-Augmented Vector Autoregressions, and Structural Vector Autoregressions in Macroeconomics
      • Essays in Macroeconometrics
      • Exponentials drive asymmetry
      • FAQ: How do I extract the output gap?
      • Fat tails preclude ergodicity
      • First impression are high variance
      • Focus on the residuals
      • For the effectual reasoner, residuals are inputs rather than outputs
      • Forecast Error Variance Decompositions with Local Projections
      • Forecasting levels of log variables in vector autoregressions
      • Generalized Random Forests
      • Get leverage on the fixed cost of pain
      • Goals should be binomial
      • Good ideas are testable ideas
      • Good representations reduce sample complexity of downstream tasks
      • Granular IV
      • Granularity
      • Growth and Ideas
      • Hamilton filter
      • Highly likely events do not yield much information
      • How Big Is the Random Walk in GNP?
      • How Competitive is the Stock Market? Theory, Evidence from Portfolios, and Implications for the Rise of Passive Investing
      • How the Wealth Was Won: Factor Shares as Market Fundamentals
      • Ideas generate increasing returns because they are nonrivalrous
      • Identification and Estimation of Dynamic Causal Effects in Macroeconomics Using External Instruments
      • If you're not sure something is getting better, it's not
      • Impulse Response Estimation by Smooth Local Projections
      • In the long-run, all costs are variable
      • Inference in Heavy-Tailed Nonstationary Multivariate Time Series
      • Information is relative
      • Information is surprisal
      • Interaction generates non-normality
      • Invest in companies like you invest in your career
      • Iterative tinkering enhances effective IQ
      • Just do more
      • Labor relationships have debt-like features
      • Large-dimensional Dynamic Factor Models: Estimation of Impulse–Response Functions with I ( 1 ) cointegrated factors
      • Lasso Inference for High-Dimensional Time Series
      • Learned representations are more important than what you do with them
      • Linear Algebra Done Right
      • Linearly separable problems are easy to solve
      • Local projections
      • Local projections vs. VARs
      • Local Projections vs. VARs: Lessons From Thousands of DGPs
      • Look for high talent and high agency
      • Mathematical Foundations of Machine Learning
      • Maximize the entropy of your information sources
      • Maximize your output of testable ideas
      • Model Selection for Local Projections Instrumental Variable Methods - Empirical Application to Government Spending Multipliers.
      • Model-Free Impulse Responses
      • Moving average model
      • Network egress is a lock-in tactic for public cloud providers
      • Not all spurious correlation is random
      • Observe before you analyze
      • Origins of Stock Market Fluctuations
      • Perfect competition is not optimal under increasing returns to scale
      • Planning your time is like choosing to dance on beat
      • Pre-train your representations before learning judgement
      • Prediction is compression, compression is expression
      • Principal component analysis
      • Principal Component Analysis for Nonstationary Series
      • Principal components estimation and identification of static factors
      • Random walk
      • residuals
      • self-supervised learning
      • Short and Variable Lags
      • Sparse Trend Estimation
      • Subset selection for vector autoregressive processes using Lasso
      • Systematic shifts in scaling behavior based on organizational strategy in universities
      • Tariffs and Growth: Heterogeneous Effects by Economic Structure
      • The Adaptive Lasso and Its Oracle Properties
      • The Bayesian Lasso
      • The Disappointing Recovery of Output after 2009
      • The Liquidity Channel of Fiscal Policy
      • Under fat tails, look for quick reasons to say yes
      • Under fat tails, outliers drive movements in the mean
      • Under fat tails, unbiasedness is overrated
      • Use TK liberally
      • variance
      • Vector autoregression
      • VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learning
      • Volatility is information
      • When in doubt, don't difference
      • When unskilled, complicate the game and add randomness
      • With the right representation, judgement is cheap

    Dynamic factor model

    DFMs are VARs that operate over unobserved factors rather than observed variables. The unobserved factors are derived via PCA. Idiosyncratic deviations from the common component are modeled via autoregressive models.


    References

    @stockDynamicFactorModels2016

    @kirkerWhatDrivesCore

    ATSA19 Lecture 8: Introduction to Dynamic Factor Analysis

    Metran

    Large dynamic factor models, forecasting, and nowcasting

    Dynamic Factor Analysis

    eonline

    Backlinks

    • The Disappointing Recovery of Output after 2009
    • Model Selection for Local Projections Instrumental Variable Methods - Empirical Application to Government Spending Multipliers.
    • Dynamic Factor Models, Factor-Augmented Vector Autoregressions, and Structural Vector Autoregressions in Macroeconomics
    • Identification and Estimation of Dynamic Causal Effects in Macroeconomics Using External Instruments
        • A Local Projections Approach to Difference-in-Differences Event Studies
        • A Neural Phillips Curve and a Deep Output Gap
        • agency
        • An Exploration of Trend-Cycle Decomposition Methodologies in Simulated Data
        • Autoregressive models
        • Ben Franklin's Autoencoder
        • Beveridge-Nelson Decomposition
        • Boosting: Why You Can Use the Hp Filter
        • Career success and career capital are cointegrated
        • Careers are a random walk
        • Change leads to insight far more often than insight leads to change
        • Correlation between prices and quantities reveals main market driver
        • Diffusion models
        • Don't detrend random walks with linear trends
        • Don't rely on inspiration to write
        • Dynamic covariate balancing: Estimating treatment effects over time with potential local projections
        • Dynamic factor model
        • Dynamic Factor Models, Factor-Augmented Vector Autoregressions, and Structural Vector Autoregressions in Macroeconomics
        • Essays in Macroeconometrics
        • Exponentials drive asymmetry
        • FAQ: How do I extract the output gap?
        • Fat tails preclude ergodicity
        • First impression are high variance
        • Focus on the residuals
        • For the effectual reasoner, residuals are inputs rather than outputs
        • Forecast Error Variance Decompositions with Local Projections
        • Forecasting levels of log variables in vector autoregressions
        • Generalized Random Forests
        • Get leverage on the fixed cost of pain
        • Goals should be binomial
        • Good ideas are testable ideas
        • Good representations reduce sample complexity of downstream tasks
        • Granular IV
        • Granularity
        • Growth and Ideas
        • Hamilton filter
        • Highly likely events do not yield much information
        • How Big Is the Random Walk in GNP?
        • How Competitive is the Stock Market? Theory, Evidence from Portfolios, and Implications for the Rise of Passive Investing
        • How the Wealth Was Won: Factor Shares as Market Fundamentals
        • Ideas generate increasing returns because they are nonrivalrous
        • Identification and Estimation of Dynamic Causal Effects in Macroeconomics Using External Instruments
        • If you're not sure something is getting better, it's not
        • Impulse Response Estimation by Smooth Local Projections
        • In the long-run, all costs are variable
        • Inference in Heavy-Tailed Nonstationary Multivariate Time Series
        • Information is relative
        • Information is surprisal
        • Interaction generates non-normality
        • Invest in companies like you invest in your career
        • Iterative tinkering enhances effective IQ
        • Just do more
        • Labor relationships have debt-like features
        • Large-dimensional Dynamic Factor Models: Estimation of Impulse–Response Functions with I ( 1 ) cointegrated factors
        • Lasso Inference for High-Dimensional Time Series
        • Learned representations are more important than what you do with them
        • Linear Algebra Done Right
        • Linearly separable problems are easy to solve
        • Local projections
        • Local projections vs. VARs
        • Local Projections vs. VARs: Lessons From Thousands of DGPs
        • Look for high talent and high agency
        • Mathematical Foundations of Machine Learning
        • Maximize the entropy of your information sources
        • Maximize your output of testable ideas
        • Model Selection for Local Projections Instrumental Variable Methods - Empirical Application to Government Spending Multipliers.
        • Model-Free Impulse Responses
        • Moving average model
        • Network egress is a lock-in tactic for public cloud providers
        • Not all spurious correlation is random
        • Observe before you analyze
        • Origins of Stock Market Fluctuations
        • Perfect competition is not optimal under increasing returns to scale
        • Planning your time is like choosing to dance on beat
        • Pre-train your representations before learning judgement
        • Prediction is compression, compression is expression
        • Principal component analysis
        • Principal Component Analysis for Nonstationary Series
        • Principal components estimation and identification of static factors
        • Random walk
        • residuals
        • self-supervised learning
        • Short and Variable Lags
        • Sparse Trend Estimation
        • Subset selection for vector autoregressive processes using Lasso
        • Systematic shifts in scaling behavior based on organizational strategy in universities
        • Tariffs and Growth: Heterogeneous Effects by Economic Structure
        • The Adaptive Lasso and Its Oracle Properties
        • The Bayesian Lasso
        • The Disappointing Recovery of Output after 2009
        • The Liquidity Channel of Fiscal Policy
        • Under fat tails, look for quick reasons to say yes
        • Under fat tails, outliers drive movements in the mean
        • Under fat tails, unbiasedness is overrated
        • Use TK liberally
        • variance
        • Vector autoregression
        • VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learning
        • Volatility is information
        • When in doubt, don't difference
        • When unskilled, complicate the game and add randomness
        • With the right representation, judgement is cheap

      • whoisnnamdi.com
      • X
      • LinkedIn
      • GitHub