This November will find millions of Americans at the polls to determine who will be the face of our nation for the next four years. This election year, more than years in recent memory, will feature the economy at the local, state, and national levels at the forefront of all discussion, and will ultimately be one of the most important factors that lead to the election of our next President and a multitude of Congressmen.Eloquent rhetoric will no doubt spill forth from the mouths of the would-be candidates, but does what they say have a basis in statistical fact, or are the numbers skewed to suit their needs? I will be providing a secondary analysis on the data put forth by the Bureau of Labor Statistics, Shadow Stats (operated by economist John Williams), and critiquing defense spending and the data given by more mainstream news outlets. I will also be conducting a case study on political advertising.
As I’m writing this, an advertisement for Barack Obama has just aired on the television, condemning the fiscal policies employed by Mitt Romney during his tenure as Governor of Massachusetts. The commercial uses visual aids like bar graphs and pie charts to compare the state of the economy before Romney as well as under his leadership. The advertisement cites web sites such as the Bureau of Labor Statistics as the source of their data. Can you trust the data that you read though? Surely, a government entity like the Bureau of Labor Statistics would be above the kind of corruption that cynical Americans believe to be commonplace in the daily activities of government officials, right? The truth is that we have altered different statistical models at different times in our history to skew the numbers so that they don’t look as bad as they really are.
How does the Bureau of Labor Statistics gather its data? The short answer is that they use surveys, and because they use surveys a great deal, there is an inherent flaw in employing this method of data collection. First, people are not completely honest about how much they earn. It can be a source of pride or shame for some people, so disclosing that kind of information is a sensitive subject. Second, when you do a survey, your sample size has to be highly randomized to get an accurate sampling of the population. Third, your sample size has to be large enough to extrapolate and make a generalization regarding the remaining population. The Bureau of Labor Statistics acknowledges their payroll survey’s confidence interval is suspect due to the admitted lack of randomization and built-in bias. Recently, the BLS has adjusted its bias factor. Instead of merely adding 150,000 jobs per month to account for new job creation and no job loss, the BLS have now accounted for that possibility and numbers have ranged from -321,000 to +270,000 in past years (Williams, 2004).
In order to better understand unemployment in the United States and the formula used in the past and the one used now, it would be beneficial to understand the concept of what exactly a discouraged worker is. A discouraged worker is any physically able person with the will to work, but has given up searching for a job because of the perception that there are none to be had. As much as I like President Clinton, it was during his tenure that he removed these people from the unemployment formula. Clinton removed all discouraged workers who had been out of work for more than a year and did not include those people in the data at all. Not only did Clinton shore up his unemployment numbers by omitting these unemployed people, but he also lowered the survey sample size from 60,000 to 50,000 or a 16.7 percent change. Before George W. Bush took office, the survey sample size had reverted to its original sample numbers, presumably to make Clinton look as though he were the one responsible for higher employment numbers and less poverty, leaving Bush to take the blame for economic downturn.
Please don’t get the impression that I’m trying to throw the Bureau of Labor Statistics under the bus, I’m not. They have the monumental task of accounting for statistical data in a manner in which they are directed to collect it. The only problem with the BLS is that they don’t have any way of collecting the kind of data that can be representative of the entire nation, often through no fault of their own. Perhaps the majority of the population isn’t ready to know the real unemployment numbers; maybe the outlook is just too bleak. Sometimes, ignorance is bliss.