Skip to content

Jumping to Conclusions

October 24, 2013

(Originally posted at

Conclusions are easy to draw. I can look at any study and draw many different conclusions or policy prescriptions. That does not mean I am right or even that the data supports my argument.

Floyd Norris of the NYT Economix fell into this trap yesterday. He claimed, indirectly, that the US federal government is not bloated. I take bloated to mean too big or too intrusive. To argue this, he cites the latest job numbers.

In September, before the government shutdown, the government had 2,723,000 employees, according to the latest job report, on a seasonally adjusted basis. That is the lowest figure since 1966.

I was not aware of this interesting statistic. However, it is difficult in statistics and empirical work to find data that actually leads to the conclusion that people hope to draw.

The data do not speak for themselves. This is not “proof” whether the government is too big or not. It is not even clear evidence. Mr. Norris has to throw in two major assumptions that he does not tell the reader.

1. The first assumption is that the number of employees is what people mean by “bloated” or at least a very good approximation. There are a number of different ways to measure whether a government is too big. Why not look at amount of government spending? With the same burden of proof that Mr. Norris requires, change since 1962, this measure would lead to the opposite conclusion.

I am not arguing what the above graph actually tells readers. I do not know what is the best estimate of “bloatedness.” But let us at least be honest and not jump to conclusions with explaining why number of employees is a good approximation. If it is not a good approximation, the argument falls apart.

2. Even if the number of employees is a good approximation, the conclusion is not clear. Mr. Norris assumes that the 1962 levels were not bloated. If the federal government was bloated in 1962,  the same amount of employees today might mean it is still bloated. To tell a similar story- I was fat in high school. The fact that I was the same weight in college does not prove that I was not fat in college.

I know why people take these shortcuts. I, regretfully, take them too. But they does not mean it is a good thing. Instead, let us be honest about our assumptions and let readers know.

One Comment
  1. Anonymous permalink
    November 5, 2013 15:11

    I personally prefer to look at government spending as a proportion of potential GDP. Even if potential GDP measures might not be accurate it gives a better picture of the size of the government: in a recession both government spending goes up (unemployment benefits, food stamps etc) and GDP goes down, which is why deficit/GDP seems to shoot up after the Great Recession.

    But I fully agree with what you are saying, it’s the same as all those TV channels with the US Debt clocks showing trillions of debt going up and up: it’s just giving numbers without any context

Comments are closed.