On business cycle forecasting

Frontiers of Business Research in China

Table 5 Forecasting the conventional target variable: in-sample estimation results

Forecast horizon (months)^a	1	3	6	9	12
Models	Relative QPS
Model with channels (a), (b), and (c) simultaneously^b
AR-Logit-Factor-MIDAS	0.302 (0.000)	0.443 (0.000)	0.577 (0.000)	0.765 (0.010)	0.862 (0.085)
Models with either channel (a) or (b)
ST-Probit-MCF	0.394 (0.000)	0.539 (0.000)	0.762 (0.000)	0.941 (0.360)	1.034 (0.530)
AR-Probit-YS	0.680 (0.000)	0.708 (0.000)	0.808 (0.002)	0.945 (0.339)	1.000 (0.993)
Models without any channel (a), (b) and (c)
ST-Probit-YS-EI	0.905 (0.002)	0.908 (0.012)	0.910 (0.015)	0.986 (0.391)	0.999 (0.550)
ST-Probit-YS	1	1	1	1	1
Models	Relative LPS
Model with channels (a), (b), and (c) simultaneously
AR-Logit-Factor-MIDAS	0.319 (0.000)	0.444 (0.000)	0.606 (0.000)	0.807 (0.034)	0.893 (0.141)
Models with either channel (a) or (b)
ST-Probit-MCF	0.387 (0.000)	0.558 (0.000)	0.761 (0.001)	0.936 (0.333)	1.056 (0.277)
AR-Probit-YS	0.647 (0.000)	0.668 (0.000)	0.739 (0.000)	0.866 (0.016)	0.967 (0.302)
Models without any channel (a), (b) and (c)
ST-Probit-YS-EI	0.938 (0.059)	0.909 (0.008)	0.914 (0.025)	0.984 (0.287)	1.000 (0.912)
ST-Probit-YS	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 N^th month
^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