softwareNerd Posted June 8, 2005 Report Share Posted June 8, 2005 Every month, the press reports the latest "jobs data" in the U.S. Here (link) is the latest example. The New York Times reports that, "after the government reported that only 78,000 new jobs were created in May, the smallest monthly gain since August 2003, stocks fell as the slow-growth theme took hold for at least a day." On the west coast, this (link) L.A.Times story reported the same event thus: "The Labor Department issued its May data showing that only 78,000 nonfarm jobs were added to payrolls — less than half the amount economists had expected and a steep drop from the 274,000 created in April." The same event, was reported (link) by the Washington Post thus: "The U.S. economy created 78,000 new jobs in May, the Labor Department reported today, about half as many as economists expected and the lowest payroll growth rate since August 2003." If the above seems repetitive, bear with me while I present one more example. This (link) is from "the horse's mouth". It is a "summary" issued by the Bureau of Labor Statistics. In part, it begins: "Nonfarm employment edged up by 78,000 in May following a much larger increase in April,..." If you read any of the above, you'd reasonably conclude that (say) X new jobs were added, Y jobs were lost and that the difference (X - Y) was 78,000. Not so! I was surprised to read Gene Epstein (Barron's Austrian economist) write that the economy actually added 707,000 jobs in May. I did not trust Mr. Epstein and thought he was trying to "spin" his own version. So, I delved a little deeper into the report by the Bureau of Labor Stats. To my surprise he was right. In May, the U.S. economy added 707,000 net jobs. If you wish to check up on this, here is the link. Reminded me of Benjamin Disraeli's famous line: "lies, damn lies, and statistics". Quote Link to comment Share on other sites More sharing options...
scottkursk Posted June 10, 2005 Report Share Posted June 10, 2005 Macroeconomic data is one of those bits of voodoo that has no clear definition. As you point out, a few people can look at the exact same numbers and come up with entirely different net figures, much less a meaning behind the net figures. What happens is, each economist takes the gross numbers and then feeds them into their own models so that they can come up with their own conclusion. So, that is why there is so much difference when it comes to the numbers. You'd think of all things the actual net numbers would be able to somehow tie but the process of getting the actual gross employment numbers involves much voodoo to determine the numbers they are practically worthless. The jobs data especially has a great deal of political (read federal funding dollars) behind it so the new versus disinterested versus aged unemployed workers is hard to track. So you have multiple layers of bureaucrats who have axes to grind or careers to make, they can easily massage numbers so do this a few thousand times over again and again and you can see why the numbers are generally crap. It starts at the level of the individual clerk at the state unemployment office reporting to their boss how well they are doing their jobs, and then all the way through to our friends in Washington. (note the tone of sarcasm with the word friends) For the most part though, you can use the numbers like you would a wet finger in the wind when it comes to figuring out what the weather is going to be like. It helps tell you which way the wind is blowing at one level, but then you gotta do a whole lot of your own research about other levels, clouds, etc and guessing to come up with a decision. Quote Link to comment Share on other sites More sharing options...
softwareNerd Posted June 12, 2005 Author Report Share Posted June 12, 2005 The "78,000 new jobs" headline and the "707,000 new jobs" headline, are both based on exactly the same underlying data! However, now that I've read more about this, I understand why the figures are so different. For anyone interested in the details, here is the rough computation... 1) The data shows that 707,000 more people were working in May compared to April. 2) However, past data also shows that the month of May usually shows a large increase in people working. Given the typical seasonality surge for May, one would have expected the number of people employed in May 2005 to rise by about 630k just due to the seasonality. 3) Actually, the number of people employed rose by about 710k. So, the difference (about 80k) is reported as "new jobs created". This closer look at the data convinces me that the number being reported (78k) is not too useful. The problem comes from #2 above: how does one arrive at the expected 630K? Without looking at the link provided, using the numbers above, what would be your ballpark-guess about the change in number of people employed between May 2004 and May 2005? (Answer: The reported year-to-year increase for May is a little over 2 million !) When I get some time, I hope to delve further into the computation used. Quote Link to comment Share on other sites More sharing options...
scottkursk Posted June 13, 2005 Report Share Posted June 13, 2005 3) Actually, the number of people employed rose by about 710k. So, the difference (about 80k) is reported as "new jobs created". This closer look at the data convinces me that the number being reported (78k) is not too useful. The problem comes from #2 above: how does one arrive at the expected 630K? So what economists do is, depending on what they need to fit their economic models of what the economy should look like, they will view the economy in a couple ways. They will assume that there will always be a certian number of jobs (X) created over a certain time. So any jobs created over X, call it Y, will yield "net new jobs" it Y-X>X then = Net jobs created. So even though 710k new jobs were created 80k were considered as new since some economists work under the assumption the others would have happened anyway. Think of it like calorie expenditure. Just laying in bed will burn X number of calories per day. So just getting out of be ded and walking to the bathroom will yield a net increase in calorie consumption. Oversimplified but then so is most macro. Of course, to derive Y, you have to take out numbers of people who just quite looking for work, people who ar deported, and other nearly impossible to figure things. If you work for the administration, you'll be reporting the Y number and if you don't you will quote only the "small" number of the "net" jobs. Depends on what side of the Y curve you are being paid. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.