Friday, July 6, 2012

Unemployment reports

I finished my statistics course three weeks ago and got a B+ in my class and learned how reports can be skewed and how to make a good analysis of data. I am reading a lot about unemployment today because June's numbers came out, I want to make a few points on how their analysis is invalid.
The unemployment report every month is merely a first draft. It takes three months to get the full picture as the forms that people sign for employment continue to come in. Assuming the economy is failing when you see the unemployment rate stays the same is just false. You need to see the historical trends when you look at the data, and a few things become clear on how irresponsible people are being.

1. Look at the trends for that month. There is one large trend in unemployment that people always leave out, according to the Bureau of Labor Statistics, unemployment increases and decreases twice every year except for one year in the past 20 years this has been the case, that year was 2008 when our housing market went belly up and the stock market crashed. The peaks are reliably in June and January, right before summer hiring and right after Christmas shopping. The low points are reliably October and April. We are still seeing this. gives you the raw information.

2. Skew is common in reports. People will look at the short-term, and show charts of just Obama's presidency. When you spread the graph out it makes our job growth look more flat, and you have no comparison with previous downturns, like the one that lasted across the majority of Ronald Reagan's first term. If you look at the long-term trends you can compare the slope with previous downturns which helps understand if this is unusual.

3. Without long-term trends you can't see the comparative slope difference between different downturns. Once you have this perspective you can see that using seasonally adjusted data to make it easier to find the low and high points that this recovery is going faster than the recession at the beginning of George W. Bush's presidency. It took 29 months (2 years, 5 months) for unemployment to drop 2% in these past two years, however, during the recession that followed the 2000 bust it took 72 months (3 years, 4 months) for unemployment to drop 2%. The 1992 bust took 16 months (1 year, 4 months) to drop 2%. The Reagan bust of 1982 took 10 months to drop 2% which was correlated with crashing oil prices. So, looking at this we can see this is an average recession in terms of recovery, and there is nothing to really panic about as long as we don't get the easy access to credit that brought the housing bubble. The thing that is truly slowing the economy down is that no one is focusing on how to get unemployment to drop and immediately start panicking. The media is saying it is our "new normal" with no evidence to back it up, but a lot of evidence on the unemployment graph to support that this is just a normal recession because they do not compare it to previous recessions, because that creates hope.

When people see these reports people get frustrated and people spend and hire less, there is a gigantic mood over the American public right now, and the blame has to go to the media. The brokers on the stock market start selling everything when the report isn't as good as they want and this doesn't help the economy. We have seen no major improvement in infrastructure (aka job growth) because the stimulus bill was weak, and people just are not hiring right now at the desired rate, but at the historical rate in the past.

Those are the current statistical errors that the media has been committing over and over again over the past few years.

No comments:

Post a Comment