Time Series for Quants: ARIMA and GARCH

Medium·22 min read
Statistical / ML for QuantsTime SeriesARIMAGARCHVolatility Forecasting

Quick Quiz

1. For a stationary AR(1) rt=c+ϕrt1+εtr_t=c+\phi r_{t-1}+\varepsilon_t with ϕ<1|\phi|<1, what are the unconditional variance and the lag-kk autocorrelation as kk\to\infty?
2. ADF on log-prices gives statistic 1.8-1.8 (p = 0.38); on log-returns it gives 28.4-28.4 (p < 0.001). What do you conclude for model selection?
3. GARCH(1,1) on daily S&P returns: ω^=2×106\hat\omega=2\times10^{-6}, α^=0.08\hat\alpha=0.08, β^=0.90\hat\beta=0.90. Find (a) unconditional daily variance, (b) shock half-life, (c) the 10-day-ahead conditional variance forecast if σT2=4×104\sigma_T^2=4\times10^{-4}.
4. In GARCH(1,1) with α^+β^=1\hat\alpha+\hat\beta=1 (IGARCH), the unconditional variance is infinite (non-stationary in variance), yet conditional variance forecasts remain finite and usable for short-horizon forecasting.
5. GJR-GARCH adds γ1εt1<0εt12\gamma\mathbf 1_{\varepsilon_{t-1}<0}\varepsilon_{t-1}^2; equities show γ^>0\hat\gamma>0. What mechanism explains this, and which spec would you use for VaR on a long equity book?
6. BIC selects ARIMA(0,0,0) (white noise) for daily returns; a colleague prefers ARIMA(3,0,2) on lower AIC. Who is right, and what does this reveal?