{"success":1,"msg":"","color":"rgb(28, 35, 49)","title":"Now- and Backcasting Initial Claims with High-Dimensional Daily Internet Search-Volume Data<\/b>","description":"webinar","title2":"","start":"2020-11-27 13:00","end":"2020-11-27 14:00","responsable":"Carlos Vladimir Rodriguez-Caballero <\/i><\/a>","speaker":"Erik Christian Montes Sch\u00fctte (Aarhus University and CREATES)","id":"18","type":"webinar","timezone":"America\/Mexico_City","activity":"Zoom link:\r\n\r\nhttps:\/\/itam.zoom.us\/j\/99246753696?pwd=Q1c5Rm1MV0kyR2I2VmJrWFFFZWE0dz09\r\n\r\nMeeting ID: 992 4675 3696\r\nAccess Code: 475862\r\n","abstract":"We generate a sequence of now- and backcasts of weekly unemployment insurance initial\r\nclaims (UI) based on a rich trove of daily Google Trends (GT) search-volume data\r\nfor terms related to unemployment. To harness the information in a high-dimensional\r\nset of daily GT terms, we estimate predictive models using machine-learning techniques\r\nin a mixed-frequency framework. The sequence of now- and backcasts are made ten\r\ndays to one day before the release of the UI \fgure on Thursday of each week. In a\r\nsimulated out-of-sample exercise, now- and backcasts of weekly UI that incorporate the\r\ninformation in the daily GT terms substantially outperform those based on an autoregressive\r\nbenchmark model, especially since the advent of the COVID-19 crisis. The\r\nimprovements in predictive accuracy relative to the autoregressive benchmark generally\r\nincrease as the now- and backcasts include additional daily GT data, with reductions in\r\nroot mean squared error of up to approximately 50%. Variable-importance measures\r\nreveal that the GT terms become more relevant for predicting UI during the crisis,\r\nwhile partial-dependence plots indicate that linear speci\fcations are largely adequate\r\nfor capturing the predictive information in the GT terms. We are in the process of\r\ncreating a website that will provide updated, real-time now- and backcasts of UI on a\r\ndaily basis.\r\n\r\nThe paper is coauthored with David E. Rapach and Daniel Borup\r\n\r\nWorking paper: https:\/\/papers.ssrn.com\/sol3\/papers.cfm?abstract_id=3690832"}