tag:blogger.com,1999:blog-2098432983500045934.post1433218609362584531..comments2024-03-12T22:19:32.339-04:00Comments on The New Arthurian Economics: The Federal Funds Rate and Economic GrowthThe Arthurianhttp://www.blogger.com/profile/16501331051089400601noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-2098432983500045934.post-78264653809470168822017-03-26T12:29:13.331-04:002017-03-26T12:29:13.331-04:00Hey Jerry.
Lags, yes. That's what I was tryin...Hey Jerry. <br />Lags, yes. That's what I was trying to show on <a href="https://2.bp.blogspot.com/-PZVy6RbdBFI/WNLNlsBrp9I/AAAAAAAATxc/GXZdI5NbwbM-jhd-7WbDXfl9LCyXFoByQCLcB/s1600/Like%2BGraph%2B%252317%2Bmarkup.png" rel="nofollow">Graph #2</a> in the post before this one. I point out this sequence:<br />problematic RGDP growth in 1954 leading to<br />higher interest rates in early 1955<br />which slow growth insufficiently, apparently,<br />so interest rates rise more in late 1955 and '56<br />and we see a slower rate of RGDP growth in 1956-57.<br /><br />From this I'd estimate a lag of 6 months or less. But yes, looking at lagged values would be huge. Unfortunately, I don't have an easy way to do that yet. And if the lag varies, that's a complication.<br /><br />I think at one point you suggested calculating a variety of R-squared values to determine the optimum lag. I thought that was a good idea. Pretty complex to try to make it happen.<br /><br />For debt and GDP the highest ordinarily available data frequency is quarterly. It is hard to gauge reactions with three months between measurements. <i>Both</i> GDP and interest rates could change within a 3-month span, I wouldn't know which changed first.<br />The Arthurianhttps://www.blogger.com/profile/16501331051089400601noreply@blogger.comtag:blogger.com,1999:blog-2098432983500045934.post-73106509013109197342017-03-26T10:57:47.391-04:002017-03-26T10:57:47.391-04:00I don't know if this would show anything usefu...I don't know if this would show anything useful or not, but ...<br />If the level of interest causes growth, say, then probably there is some delay between the cause and the effect, right? It takes the bathtub a while to warm up after the kids complain and I change the temperature coming out of the faucet.<br /><br />What about shifting the time on the two axes a bit?<br />i.e. instead of a data point being "growth rate and interest rate at feb 1 2010", plot "growth rate at feb 1 2010, interest rate at jul 1 2009", or something.<br />I don't know what the right "lag" would be...I guess I would do plots at 0,3,6,9,12,15,18,20,24 months of lag and see which chart looked the best. <br />(maybe also -3,-6,-9,-12,... if you think that growth causes changes to interest rates instead, e.g. the fed watching the numbers and turning the only dial they have.)<br />I think there's a better mathematical way to do that, but I'd have to look it up and then figure out how to use it. ha.<br />Jerrynoreply@blogger.com