tag:blogger.com,1999:blog-3634628224045926034.post8813221207303293839..comments2023-08-29T01:27:13.772-07:00Comments on Magpie's Asymmetric Warfare: Open Letter to Young Leftists.Magpiehttp://www.blogger.com/profile/07528637318288802178noreply@blogger.comBlogger10125tag:blogger.com,1999:blog-3634628224045926034.post-24593352033843376462016-10-31T23:33:08.827-07:002016-10-31T23:33:08.827-07:00"That procedure would yield a line, no matter...<i>"That procedure would yield a line, no matter what data, of course. So it cannot provide evidence against what I said; it is basically assuming what is to be proved."</i><br /><br />Come on, Calgacus, the procedure I used is exactly the opposite: the data do not suggest a better trendline. In fact, one doesn't need to be an expert statistician to see that the data look like a linear trend.<br /><br />----------<br /><br />But this problem has an easy resolution.<br /><br />You believe my procedure is flawed. Fair enough.<br /><br />Prove it. No need to write volumes for that or bet about what other people would say. Present your alternative.<br /><br />As none of us is a professional statistician, let's stick to the basics. Again, download the data and use any spreadsheet you may have. <br /><br />Assuming you are using a garden-variety spreadsheet: they have plenty options. The one I use (LibreOffice Calc version 5.1.2.2) has:<br /><br />(1) Linear<br />(2) Polynomial<br />(3) Logarithmic<br />(4) Exponential<br />(5) Power<br />(6) Moving average<br /><br />You can specify for the spreadsheet to present the equation and its coefficient of determination (R^2). Do it.<br /><br />We are interested in a model where the COE share of GDP is a function of time; we also need a comparison statistic (the easiest is R^2, which probably every single spreadsheet calculates). These two criteria rule (6) out. But, other than linear, you still have 4 options.<br /><br />Pick the linear trend line and write down its results (equation and R^2). Do the same for your own alternative and next time you reply, instead of arguing this and that, (i) explain what made you think of that model (for instance, which diagnostic technique, like plotting on log scale, or whatever, you used) and (ii) present both its equation and coefficient of determination, so as to compare them with those of the linear trend.<br /><br />Something simple, like a paragraph to explain why you chose that model, maybe some links, and a table (like the example below), will do:<br /><br />Your alternative | linear<br />------------------+------------<br />your R^2 | linear R^2<br />your model | linear model<br /><br />It should take you less time to do that than it took me to write this, but I'll give you one day. Sounds good?Magpiehttps://www.blogger.com/profile/07528637318288802178noreply@blogger.comtag:blogger.com,1999:blog-3634628224045926034.post-26328475209687153572016-10-31T23:25:21.322-07:002016-10-31T23:25:21.322-07:00This comment has been removed by the author.Magpiehttps://www.blogger.com/profile/07528637318288802178noreply@blogger.comtag:blogger.com,1999:blog-3634628224045926034.post-26630466116420176032016-10-31T20:19:31.758-07:002016-10-31T20:19:31.758-07:00Calgacus, download the data and adjust a linear tr...<i>Calgacus, download the data and adjust a linear trendline and tell me what it says. :-)</i><br /><br />That procedure would yield a line, no matter what data, of course. So it cannot provide evidence against what I said; it is basically assuming what is to be proved.<br /><br /><i>Why would one object not including the first one -- as you do -- but say nothing about the last three ones?</i><br /><br />Not sure of your exact meaning, but show the graph to anyone, without labelling it or saying what it is, and I assert they would say what I did: the graph shows that something changed (around 1973). The graph agrees with the "Trentes Glorieuses" followed by a distinct neoliberal era, not one period without major change.<br /><br />I'm not saying nothing about the last 3, just saying what many have said, that the neoliberal era recoveries have become weaker and much more angled towards the top. This means a decline in wages after 1973, as we both agree.Calgacushttps://www.blogger.com/profile/06031818010224747000noreply@blogger.comtag:blogger.com,1999:blog-3634628224045926034.post-8716646193693299072016-10-30T00:43:24.318-07:002016-10-30T00:43:24.318-07:00Hi Calgacus,
Thanks for your comment.
"The ...Hi Calgacus,<br /><br />Thanks for your comment.<br /><br /><i>"The chart above doesn't show that.<br /><br />"Constant share of about 50% followed by continual decline when les Trentes Glorieuses ended, at the oil crises, the 70s is clearly a much better description"</i><br /><br />Calgacus, download the data and adjust a linear trendline and tell me what it says. :-)<br /><br />You can see the data has a random component and there are ups and downs. There is a number of recoveries lasting more than a year or two (we are currently going through one):<br /><br />1. 1948-1953<br />2. 1965-1970<br />3. 1996-2000<br />4. 2006-2008<br />5. 2013-2015<br /><br />You can see that in the chart. Why would one object not including the first one -- as you do -- but say nothing about the last three ones?Magpiehttps://www.blogger.com/profile/07528637318288802178noreply@blogger.comtag:blogger.com,1999:blog-3634628224045926034.post-85909257401630646192016-10-29T14:38:38.042-07:002016-10-29T14:38:38.042-07:00Since 1950 the share of gross domestic income corr...<i>Since 1950 the share of gross domestic income corresponding to wages has been falling in the US, the richest, most advanced capitalist economy.</i> <br /><br />The chart above doesn't show that. <br /><br />Constant share of about 50% followed by continual decline when les Trentes Glorieuses ended, at the oil crises, the 70s is clearly a much better description.Calgacushttps://www.blogger.com/profile/06031818010224747000noreply@blogger.comtag:blogger.com,1999:blog-3634628224045926034.post-30773241322482731162016-10-02T16:43:02.439-07:002016-10-02T16:43:02.439-07:00Oh my god.... it's true that Brad Delong is in...Oh my god.... it's true that Brad Delong is in the bible...Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3634628224045926034.post-60698658785755187252016-09-30T23:53:43.592-07:002016-09-30T23:53:43.592-07:00I'd add that workers have been losing the clas...<i>I'd add that workers have been losing the class war for far more than 66 years; by my count it's been about 5,000 years.</i><br /><br />I know what you mean, Jim. Those 66 years of loss I show are just as far as the data go.<br /><br />It seems as I age I'm growing more optimistic, paying less attention to the negative bits.<br /><br />Now, that's a disturbing thought. Magpiehttps://www.blogger.com/profile/07528637318288802178noreply@blogger.comtag:blogger.com,1999:blog-3634628224045926034.post-2085630405559864812016-09-30T15:41:01.228-07:002016-09-30T15:41:01.228-07:00Well said. Keynesianism is the left pole of respec...Well said. Keynesianism is the left pole of respectable leftism and it has done tremendous harm. Society's problem becomes not one of concentrated and hostile minority power, but merely a bit of insufficient demand that can be stimulated with a wee bit more government spending. I'd add that workers have been losing the class war for far more than 66 years; by my count it's been about 5,000 years. <br /><br />Jim<br />https://commentsongpe.wordpress.comAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-3634628224045926034.post-6629568158329895422016-09-29T16:51:51.977-07:002016-09-29T16:51:51.977-07:00Thanks, Birdwatcher.
Your friend, the -- 0 ---,
...Thanks, Birdwatcher.<br /><br />Your friend, the -- 0 ---,<br /><br />:-)Magpiehttps://www.blogger.com/profile/07528637318288802178noreply@blogger.comtag:blogger.com,1999:blog-3634628224045926034.post-37685729975901737632016-09-29T13:53:35.138-07:002016-09-29T13:53:35.138-07:00good post,swooping bird
- the oogood post,swooping bird<br />- the ooAnonymousnoreply@blogger.com