Skip to main content

Table 6 Forecasting the conventional target variable: out-of-sample estimation results

From: On business cycle forecasting

Forecast horizon (months)a 1 3 6 9 12
Models Relative QPS
 Model with channels (a), (b), and (c) simultaneouslyb
  AR-Logit-Factor-MIDAS 0.400 (0.000) 0.507 (0.000) 0.580 (0.000) 0.867 (0.002) 1.125 (0.165)
 Models with either channel (a) or (b)
  ST-Probit-MCF 0.454 (0.000) 0.695 (0.010) 0.959 (0.635) 1.105 (0.265) 1.269 (0.005)
  AR-Probit-YS 0.795 (0.000) 0.701 (0.000) 0.696 (0.000) 0.876 (0.199) 1.072 (0.443)
 Models without any channel (a), (b) and (c)
  ST-Probit-YS-EI 0.886 (0.001) 0.895 (0.006) 0.897 (0.000) 0.980 (0.067) 1.013 (0.018)
  ST-Probit-YS 1 1 1 1 1
Models Relative LPS
 Model with channels (a), (b), and (c) simultaneously
  AR-Logit-Factor-MIDAS 0.527 (0.001) 0.598 (0.004) 0.618 (0.004) 0.896 (0.006) 1.171 (0.228)
 Models with either channel (a) or (b)
  ST-Probit-MCF 0.582 (0.013) 0.853 (0.386) 0.943 (0.646) 1.114 (0.365) 1.317 (0.010)
  AR-Probit-YS 0.731 (0.000) 0.657 (0.000) 0.616 (0.000) 0.794 (0.045) 1.027 (0.794)
 Models without any channel (a), (b) and (c)
  ST-Probit-YS-EI 0.908 (0.030) 0.885 (0.009) 0.872 (0.005) 0.975 (0.057) 1.011 (0.012)
  ST-Probit-YS 1 1 1 1 1
  1. a For each forecast horizon N (where N = 1, 3, 6, 9, 12), we evaluate the ability of each model to forecast the probability of a recession in the Nth month
  2. bChannel (a): using a flexible function form by including a lagged recession probability into a dynamic Logit or Probit model of recession forecasting; Channel (b): employing a dynamic factor model to extract common factors from many monthly or weekly economic and financial variables; Channel (c): applying MIDAS to incorporate the mixed-frequency common factors in the dynamic Logit framework