{"id":6674393,"date":"2020-07-02T21:51:42","date_gmt":"2020-07-02T21:51:42","guid":{"rendered":"http:\/\/cfe.econ.jhu.edu\/?p=6674393"},"modified":"2022-05-17T13:01:55","modified_gmt":"2022-05-17T13:01:55","slug":"inside-baseball","status":"publish","type":"post","link":"https:\/\/krieger.jhu.edu\/financial-economics\/2020\/07\/02\/inside-baseball\/","title":{"rendered":"Inside Baseball"},"content":{"rendered":"\n<p>Seasonal adjustment is one of the more obscure potential casualties of the pandemic recession. A well known problem with seasonal adjustment is that an outlier observation will lead the seasonal adjustment process to shift its belief about what is normal for that month. In cases like the coronavirus recession, this is undesirable because the unusual shocks for the last few months don\u2019t tell us anything about what is normal for this time of year. Seasonal adjusters are of course well aware of this and have the option to manually specify that the data for a given month be treated as an outlier. That then effectively freezes the seasonal factor for that month.<\/p>\n\n\n\n<p>Seasonal adjustment in the BLS establishment survey is done at the disaggregate level. But we can take the difference between total not seasonally adjusted data and total seasonally adjusted data as the implicit seasonal factor in total payroll employment. The table below shows the implicit seasonal factors for April, May and June of 2019 and 2020.<\/p>\n\n\n\n<figure class=\"wp-block-image\"><img decoding=\"async\" src=\"http:\/\/krieger.jhu.edu\/financial-economics\/wp-content\/uploads\/sites\/96\/2020\/07\/tab2.jpg\" alt=\"Implicit additive seasonal factor in total nonfarm payrolls\"\/><\/figure>\n\n\n\n<p>So we can see that the seasonal factor was actually slightly increased in April (relative to a year earlier) but was reduced in May and reduced further in June.<\/p>\n\n\n\n<p>Here is what I believe is going on. The BLS is well aware of the problem of the coronavirus distorting seasonal factors, and made a manual adjustment in April to treat most components of payrolls as having outliers, as described above. That solved the problem for April. But only for April, because the seasonal adjustment is applied to the level of employment and the level of employment is not going back to the pre-virus level for a long time. In May, the BLS treated many series as having outliers again, but not as many as before. And in June, they turned off the outliers for still more series, though many still have the outlier designation. As the BLS takes out its manual adjustments, the problem of the non-seasonal coronavirus affecting seasonal factors is now coming in. BLS can only prevent this by treating every month as an additive outlier, unless they make some other more fundamental adjustment.<\/p>\n\n\n\n<p>In my view the 2019 seasonal factors are more appropriate. This means that the gain in payrolls this morning should have been growth of only 4.655 million rather than the published 4.8 million. Of course, this is inconsequential when we think of what is happening in the economy right now. But next year, when data are presumably back to more normal ranges, the distorted seasonal factors will cause employment growth to be materially overstated.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Seasonal adjustment is one of the more obscure potential casualties of the pandemic recession. A well known problem with seasonal adjustment is that an outlier observation will lead the seasonal adjustment process to shift its belief about what is normal for that month. In cases like the coronavirus recession, this is undesirable because the unusual [&hellip;]<\/p>\n","protected":false},"author":500,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[112],"tags":[],"class_list":["post-6674393","post","type-post","status-publish","format-standard","hentry","category-analysis"],"acf":[],"_links":{"self":[{"href":"https:\/\/krieger.jhu.edu\/financial-economics\/wp-json\/wp\/v2\/posts\/6674393","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/krieger.jhu.edu\/financial-economics\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/krieger.jhu.edu\/financial-economics\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/krieger.jhu.edu\/financial-economics\/wp-json\/wp\/v2\/users\/500"}],"replies":[{"embeddable":true,"href":"https:\/\/krieger.jhu.edu\/financial-economics\/wp-json\/wp\/v2\/comments?post=6674393"}],"version-history":[{"count":5,"href":"https:\/\/krieger.jhu.edu\/financial-economics\/wp-json\/wp\/v2\/posts\/6674393\/revisions"}],"predecessor-version":[{"id":6675216,"href":"https:\/\/krieger.jhu.edu\/financial-economics\/wp-json\/wp\/v2\/posts\/6674393\/revisions\/6675216"}],"wp:attachment":[{"href":"https:\/\/krieger.jhu.edu\/financial-economics\/wp-json\/wp\/v2\/media?parent=6674393"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/krieger.jhu.edu\/financial-economics\/wp-json\/wp\/v2\/categories?post=6674393"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/krieger.jhu.edu\/financial-economics\/wp-json\/wp\/v2\/tags?post=6674393"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}