Thursday, October 11, 2012

John Williams on Lies, Damned Lies and the 7.8% Unemployment Rate

To provide a sense of how seasonal adjustment factors play an important role in the data released by the Bureau of Labor statistics, last week I gave an example of how using seasonal adjustment factors, I could manipulate the CPI.

In a recent interview with J.T Long at The Gold Report, John Williams, of provides an detailed report on what the BLS tells us about the seasonal adjustments for the unemployment figures and what they don't. This is a must read: Author John Williams wonders if politics are at play behind the latest jobs report, which shows 114,000 new U.S. jobs since September and a 0.3% drop in unemployment since August. Investors need to know how seasonal factors and month-to-month volatility affect the Bureau of Labor Statistics' reports. In this exclusive interview with The Gold Report , Williams explains why he doubts that we are in a recovery. The take-away? Look at the unadjusted figures before you sell your gold.

The Gold Report: John, as Mark Twain famously quipped, "There are three kinds of lies: lies, damned lies and statistics." The Bureau of Labor Statistics (BLS) just came out with new jobs numbers that show the country added 114,000 jobs since September and the unemployment rate dropped to 7.8%, down from 8.1% in August. On, you argue that the numbers are wrong and pointed to politics as a possible reason for the incorrect figures. Are unemployment statistics being manipulated and if so how?

John Williams: I normally put out a commentary on the numbers, and, in this one, I raised the possibility of politics as a factor. The problem is very serious misreporting of the numbers and the result is what appears to be a bogus unemployment rate. The BLS reported a drop in the unemployment rate from 8.1% to 7.8%, three-tenths of a percentage point, which runs counter to what is being experienced in the marketplace.

What few people realize is that the headline unemployment rate is calculated each month using a unique set of seasonal adjustments. The August unemployment rate, which was 8.1%, was calculated using what BLS calls a "concurrent seasonal factor adjustment." Each month the agency recalculates the series to adjust for regular seasonal patterns tied to the school year or holiday shopping season or whatever is considered relevant. The next month, it does the same thing using another set of seasonal factors. Rather than publish a number that's consistent with the prior month's estimate, it recalculates everything, including the previous month, but it doesn't publish the revised number from the previous month.

The assumption is that the monthly recalculations don't make much difference over time, but they do. The depth and the protraction of the current severe economic downturn have thrown off the annual seasonal-factor adjustments. The result is very volatile seasonal factors month-to-month. That means the new calculations for the September number may have resulted in a very significant revision to the August number. Again, though, the BLS doesn't publish that, so the headline August-to-September 2012 change in the unemployment rate is not consistent and not comparable. Last December, when the BLS put the seasonal adjustments on a consistent basis for the year, as it does once per year, the November 2011 unemployment rate had just been reported as showing four-tenths of a percentage point drop – an unusually large monthly decline that never took place. When revised to a consistent basis, the drop in headline November unemployment revised to two-tenths of a percent. That is a big change. I think something like that happened here.

The BLS knows what the actual number is. It has an actual estimate for August, which is consistent with September, but it doesn't publish it because it says it "doesn't want to confuse data users." But it is putting out numbers that have no meaning month-to-month. One month before the election and a month after Federal Reserve Chairman Ben Bernanke announced Quantitative Easing (QE) 3, is not a time to have inaccurate numbers. The BLS should publish the consistent numbers now.

TGR: You have said that BLS has been using this recalculation method for years. Do you feel that this month the numbers were more skewed than usual because of the political timing?

JW: Because there is no transparency in the calculation and reporting process, it leaves open the possibility of manipulation. What has happened here, though, is that in the wake of the economic collapse, the seasonal factors have been heavily distorted and are not stable on a month-to-month basis. Where the concept originally might not have made that much of a difference, it does make a big difference now. I suspect that is why we woke up to such a screwy unemployment rate this time around.

The 114,000 jobs growth in the payroll survey (which reflects the number of payroll jobs, counting multiple jobholders more than once) also is suspect and subject to concurrent-seasonal-factor adjustments. There, however, the BLS publishes revised estimates for the two prior months that are on a consistent basis with the headline number. Nonetheless, jobs in even earlier months are not re-reported, although they too are recalculated each month, with the effect that jobs reported in earlier periods can be moved into present reporting, boosting the current numbers, without the related earlier changes being revised in the published historical numbers. Nonetheless, the purported 114,000 jobs gain was not statistically significant.

From the household survey, which gives us the unemployment rate and counts the number of people who are employed (multiple-job holders are counted but once), the headline gain in employment was 873,000, the largest seasonally-adjusted monthly increase since Ronald Reagan's first-term. That number clearly is nonsense and again suggests there is a severe problem with the seasonal factors.

TGR: Do you think the unemployment rate was manipulated on purpose or did the bad economy just make the reporting more confusing?

JW: It could have been manipulated. I do not know and do not have direct evidence of current political massaging of the data. I know for certain that there have been direct political manipulations by different administrations, since the days of President Lyndon Johnson, involving various data sets that have included the gross domestic product (GDP), the trade numbers and the employment and unemployment numbers.

From what I've seen of the Obama administration, the reporting has been reasonably clean. Nonetheless, at best, the administration is using seriously flawed data, and the reporting and calculation process has the potential for manipulation. The timing of the announcement of such a big downside swing in unemployment certainly is a fortuitous circumstance for the administration's political needs.

Main Street U.S.A., however, has a much better sense on the economic reality than do the government's economic statisticians. If the headline unemployment rate is not as advertised, a goodly portion of the public will not buy it. Past experience has shown gimmicked reporting often backfiring on the manipulators.

TGR: What is the correct unemployment rate? What would be a reliable data set?

JW: I don't know of one. The unemployment rate comes out of government surveying and data manipulation, and the base number is wrong. What are good in theory are the un-adjusted numbers, although unemployment definitions still suffer. Those don't get revised for the seasonal factors. But there you have regular annual patterns of economic activity, so you'll see the unemployment rate go up and down as it follows the normal flow of annual business activity through the various seasons. Even so, it makes some sense to look at that unadjusted series over time. The average person doesn't think of himself or herself as employed on a seasonally adjusted basis, but a lot of people, according to the government, are so employed.
Read the rest here.

No comments:

Post a Comment