What Statistics Should Libertarians Use?

I read, or at least browse, many libertarian pieces, almost on a daily basis.  I have found that many libertarians like to cite John Williams of shadowstats.com when looking at various figures such as price inflation and unemployment.

It is natural for libertarians to distrust government statistics.  I have my own concerns about certain government statistics and how accurate they are.  I certainly don't think it is out of the question that certain statistics may have their formula manipulated over time to favor the government.  I don't think the government is lying about how they are currently calculating statistics, although I am always open to evidence.  But I do think that the government will make changes in formulas to make things look better than they really are.

Since many libertarians want to shun government statistics, they look for alternatives.  I see several libertarians, or those who lean libertarian, who like to cite John Williams.  Williams is currently showing that unemployment is well over 20% in the U.S.  He is showing that GDP is negative.  He is showing that consumer price inflation is near 10% using a similar methodology used prior to 1980.

My main point of this post is to caution others in using his statistics.  Just as the government has an agenda, Williams also has an agenda.  I tend to agree with William's agenda more than the governments, but it is an agenda nonetheless.  I'm sure Williams is on our side in the fight for liberty and there are some things I agree with him on.  But his statistics are not one of them.

Just as the government is understating the CPI and unemployment, my opinion is that Williams is overstating these things.  For him to put unemployment at over 22% is kind of crazy.

Also, if you do use William's statistics, you should explain to your audience the criteria you are using.  It would be misleading to say that unemployment is 22% without explaining that this number includes part-time workers.  I think there is a big difference between someone working 30 hours per week and someone working zero hours per week.

Another thing I don't like about William's statistics is that he doesn't do a good job of explaining how he comes up with some of his figures.  In this sense, the government data is actually more open.  I would like to see a good detailed explanation on how he comes up with his numbers, particularly the CPI.

Here is my suggestion when it comes to using statistics such as the CPI.  I would use the government statistics.  As long as they don't announce a change in the formula, it is staying fairly consistent from month to month.  So while the formula may be understating consumer price inflation, we can still use the trends from the data.  If the CPI was at 2% last month and is reported at 2.5% this month, then we can see a trend that consumer prices are starting to increase at a greater rate.  And while one month doesn't make a trend, this can be useful data when looking at it over the course of several months or years.

In conclusion, I have nothing personal against John Williams and he is free to publish his own statistics all he wants.  But I would strongly caution libertarians to be careful in using his data.  I would be really careful if you are using the data to make money and investment decisions.