Financial Data Analysis AI Prompts

Rigorously evaluate the statistical and economic validity of this proposed alpha signal. Signal description: {{signal_description}} Signal data: {{signal_data}} Universe: {{universe}} Look-ahead period: {{horizon}} 1. Information coefficient (IC) analysis: IC = Spearman rank correlation(signal_t, return_{t+h}) - Compute IC for each cross-section (each time period) - Mean IC: expected predictive power per period. IC > 0.05 is economically meaningful for daily signals. - IC standard deviation (ICSD): consistency of the signal - Information ratio of the signal: IC_mean / IC_std IR > 0.5: strong signal. IR > 1.0: exceptional. - % of periods with positive IC: > 55% indicates consistent directionality 2. IC decay analysis: - Compute IC at horizons h = 1, 5, 10, 21, 63, 126 trading days - Plot IC vs horizon: how quickly does predictive power decay? - The horizon where IC crosses zero defines the signal's natural holding period - Fast decay → short-term signal (high turnover). Slow decay → longer-term signal. 3. Quintile / decile portfolio analysis: - Each period: sort universe by signal into 5 (or 10) portfolios - Equal-weight each portfolio and compute forward returns - Report for each quintile: mean return, std, Sharpe, % periods positive - Key test: monotonic relationship from Q1 (low signal) to Q5 (high signal)? - Spread return: Q5 − Q1 long-short portfolio - Spread Sharpe ratio, drawdown, and turnover 4. Statistical significance testing: - t-test on mean IC: H₀: IC_mean = 0. Reject if |t| > 2.0. - Account for autocorrelation in IC series: Newey-West standard errors - Multiple testing concern: if this signal is one of many tested, apply Bonferroni or BHY correction - Bootstrap test: reshuffle signal vs returns 10,000 times and check if observed IC exceeds 95th percentile of null 5. Signal decay and overfitting checks: - In-sample vs out-of-sample IC: if in-sample IC >> out-of-sample IC, likely overfitting - Publication decay: has this signal's IC declined over time? (Sign of arbitrage) - Stability: does IC remain consistent across different market regimes? 6. Practical implementation costs: - Turnover rate of the long-short portfolio - Effective spread cost at current turnover: does signal survive round-trip transaction costs? - Break-even cost: max cost at which signal still generates positive net IC Return: IC statistics table, IC decay plot, quintile return analysis, significance tests, overfitting checks, and net-of-cost IC estimate.

Analyze the correlation structure of this multi-asset portfolio and identify instabilities. Assets: {{asset_list}} Return frequency: {{frequency}} Period: {{period}} 1. Static correlation matrix: - Compute Pearson correlation matrix - Visualize as heatmap with hierarchical clustering (assets with similar correlations grouped together) - Report the range: minimum and maximum pairwise correlations - Flag pairs with correlation > 0.9 (potential redundancy) and < -0.5 (potential hedge) 2. Robust correlation estimation: Pearson correlation is sensitive to outliers. Apply: - Spearman rank correlation: robust to outliers, captures monotonic relationships - Ledoit-Wolf shrinkage: regularized covariance matrix — critical for portfolio optimization with many assets - Minimum covariance determinant (MCD): downweights outliers automatically Compare: how much do robust estimates differ from Pearson for each pair? 3. Rolling correlation analysis: - 63-day rolling pairwise correlations for all pairs - Plot selected pairs over time - Identify correlation regime changes: periods when correlations were notably higher or lower - Crisis correlation: do correlations spike during market stress? (Diversification typically fails when needed most) 4. Principal Component Analysis (PCA): - Apply PCA to the correlation matrix - Report: variance explained by each PC (scree plot) - How many PCs explain 80% of variance? (Indicates effective dimensionality of the portfolio) - PC1 loadings: usually the 'market factor' — uniform positive loadings on all assets - PC2 onward: often sector or style tilts - Track PC1 explained variance over time: rising explained variance indicates increasing co-movement (correlation risk) 5. Instability metrics: - Correlation instability index: average change in pairwise correlations across rolling windows - Lowest-correlation period vs highest-correlation period: what drove the change? - Correlation between asset pairs during down markets vs up markets (asymmetric correlation) 6. Implications for portfolio construction: - Which correlations are most unstable? (Least reliable for diversification) - What is the maximum theoretical diversification benefit given current correlations? Return: correlation matrix heatmap, Ledoit-Wolf estimate, rolling correlation plots, PCA results, instability metrics, and portfolio construction implications.

Analyze the factor exposures of this portfolio or asset using standard risk factor models. Portfolio / asset: {{portfolio}} Factor model: {{factor_model}} (Fama-French 3, Fama-French 5, Carhart 4, Barra, or custom factors) Time period: {{period}} 1. Factor model regression: Run OLS regression of excess returns on factor returns: R_i - R_f = α + β₁F₁ + β₂F₂ + ... + βₙFₙ + ε For Fama-French 3-factor: R_i - R_f = α + β_MKT(R_M - R_f) + β_SMB(SMB) + β_HML(HML) + ε Report for each factor: - Beta (exposure): with 95% confidence interval - t-statistic and p-value - Economic significance: what does a 1-unit factor shock imply for portfolio return? 2. Alpha (Jensen's alpha): - Report α with standard error and t-statistic - Annualized alpha = daily_alpha × 252 - Is alpha statistically significant (t > 2.0)? Is it economically meaningful? - Caveat: alpha depends heavily on which factors are included in the model 3. Model fit: - R² and Adjusted R²: what % of return variation is explained by the factors? - Information ratio: α / tracking_error (annualized) - Residual autocorrelation: Durbin-Watson test on residuals 4. Rolling factor exposures: - 252-day rolling betas for each factor - Plot over time: are exposures stable or do they drift significantly? - Significant beta drift may indicate strategy drift, market regime change, or reconstitution 5. Factor contribution to return: - Decompose total return into: factor contribution + alpha + unexplained - Factor contribution_i = β_i × Factor_return_i - Which factors contributed most positively and negatively over the period? 6. Residual analysis: - Is the idiosyncratic risk (residual std) large relative to systematic risk? - High idiosyncratic risk suggests security-specific risks not captured by the factor model Return: factor exposure table with CIs, alpha analysis, R², rolling beta plots, return decomposition, and residual analysis.

Profile this financial returns dataset and identify any data quality issues before analysis. Asset class: {{asset_class}} Frequency: {{frequency}} (daily, weekly, monthly) Date range: {{date_range}} 1. Basic return statistics: - Count of observations and date range coverage - Mean, median, standard deviation, min, max - Annualized return: mean_daily × 252 (or ×52 weekly, ×12 monthly) - Annualized volatility: std_daily × sqrt(252) - Skewness and excess kurtosis — financial returns typically show negative skewness and excess kurtosis (fat tails) 2. Data quality checks specific to returns: - Zero returns: flag consecutive zero returns (>3 in a row often indicates a data freeze or illiquid asset, not a flat market) - Extreme returns: flag returns beyond ±10σ — likely data errors, corporate actions, or extreme events requiring investigation - Missing dates: check against the expected trading calendar. Missing dates should be explained (holidays, halts) - Stale prices: if using prices, identical consecutive closing prices for liquid assets signal a data problem - Survivorship bias check: is this a historical dataset? Were assets included only if they survived to the present? 3. Distribution analysis: - Plot return distribution vs normal distribution overlay - Jarque-Bera test for normality: JB = n/6 × (S² + K²/4) where S=skewness, K=excess kurtosis - Report: skewness (negative is left-skewed — bad tails), kurtosis (>3 indicates fat tails) - Quantile-Quantile plot: visual check for tail behavior relative to normal 4. Autocorrelation check: - Ljung-Box test for serial autocorrelation in returns (should be near zero for efficient markets) - Ljung-Box test on squared returns (should show autocorrelation — volatility clustering is expected) - Plot ACF and PACF for returns and squared returns 5. Corporate actions and outliers: - Flag dates with |return| > 3σ as requiring investigation - For each flagged date: check if the return aligns with a known event (earnings, index rebalance, dividend) - Adjust for dividends and splits if working with raw prices Return: summary statistics table, data quality flag list, distribution plots, autocorrelation results, and a data quality verdict (suitable for analysis / needs adjustment / not suitable).

Conduct a comprehensive tail risk analysis for this return series. Portfolio or asset: {{portfolio}} Return series: {{returns}} 1. Empirical tail analysis: - Left tail: distribution of returns below the 5th and 1st percentile - Right tail: distribution of returns above the 95th and 99th percentile - Tail asymmetry: is the left tail heavier than the right? (Typical for equity strategies) - Comparison to normal: at the 1% quantile, how does the empirical loss compare to the normal distribution prediction? 2. Extreme Value Theory (EVT) for tail estimation: Peaks Over Threshold (POT) method: - Choose threshold u at the 95th percentile of losses - Fit Generalized Pareto Distribution (GPD) to exceedances: F(x) = 1 - (1 + ξx/σ)^(-1/ξ) - Report: shape parameter ξ (> 0 = heavy tail, = 0 = exponential, < 0 = bounded tail), scale σ - ξ > 0.5 indicates very heavy tails — normal-based risk measures severely underestimate risk - Use GPD to estimate VaR and CVaR at extreme quantiles (99.9%) beyond the data 3. Maximum Drawdown analysis: - Maximum drawdown (MDD): largest peak-to-trough decline - Average drawdown - Drawdown duration distribution: how long do drawdowns last? - Recovery time distribution: how long does it take to recover to prior peak? - Calmar ratio: annualized return / |MDD| - Pain index: integral of drawdown curve over time 4. Tail correlation (co-tail risk): - For a multi-asset portfolio: does the portfolio tail loss exceed what uncorrelated risks would imply? - Tail dependence coefficient: probability that both assets suffer extreme losses simultaneously - Clayton copula for lower tail dependence: captures asymmetric dependence in down markets 5. Stress test scenarios: Apply historical stress scenarios: - 2008 financial crisis (Sept–Nov 2008) - COVID crash (Feb–Mar 2020) - 2020 interest rate spike (Q1 2022) - Dot-com crash (2000–2002) For each: what was the portfolio loss? How does it compare to VaR predictions? 6. Reporting: - At what loss level does your risk model break down? (Where does the normal approximation stop being conservative?) - What tail risk is not captured by standard VaR? Return: empirical tail analysis, GPD parameter estimates, drawdown metrics, tail correlation analysis, stress test results, and risk model limitation statement.

Analyze volatility regimes in this return series and build a regime classification model. Asset / index: {{asset}} Return series: {{returns}} 1. Realized volatility estimation methods: Compare these estimators and explain when each is appropriate: - Close-to-close: std(log returns) × sqrt(252). Simple but uses only end-of-day prices. - Parkinson: uses daily high-low range. More efficient than close-to-close. - Garman-Klass: uses OHLC prices. More efficient than Parkinson. - Yang-Zhang: handles overnight gaps. Best all-around estimator for daily OHLC. - Rolling window choice: 21-day (1 month), 63-day (1 quarter), 252-day (1 year) — each captures different features 2. GARCH volatility modeling: Fit a GARCH(1,1) model: σ²_t = ω + α ε²_{t-1} + β σ²_{t-1} - Report: ω, α, β, and their standard errors - Persistence: α + β. If > 0.99, volatility shocks are very long-lived. - Half-life of volatility shock: ln(0.5) / ln(α + β) - Likelihood ratio test: GARCH vs constant variance (ARCH test) - Plot conditional volatility over time 3. Regime detection: Method A — Hidden Markov Model (HMM): - Fit a 2-state Gaussian HMM to the returns - State 1 typically: low volatility, higher mean (bull) - State 2 typically: high volatility, lower/negative mean (bear) - Report: state means, state variances, transition probability matrix - Plot: smoothed state probabilities over time Method B — Threshold-based regime classification: - Low vol: rolling 21-day vol < 33rd percentile of historical vol - Medium vol: 33rd–67th percentile - High vol: > 67th percentile - Simpler, more transparent, but not probabilistic 4. Regime statistics: For each regime, report: - Mean return (annualized) - Volatility (annualized) - Sharpe ratio - Average duration (how long do regimes last?) - Transition frequency 5. Practical implications: - Does the current period appear to be in a high-vol regime? - How should portfolio risk management differ across regimes? Return: volatility estimator comparison, GARCH results, HMM regime probabilities, regime statistics table, and current regime assessment.