It should be no secret that GDP is an ineffective way of measuring national output, but it does not seem to stop us from using it anyway. What is more peculiar is government’s effect on the wonderful little number. Brian and Claudia Strow find 7 issues with the figure, chief among them is “Production vs Wealth Creation” and “Government Spending”, with the former being derivable from the latter. What is worse is GDP arose only at the convenience of the federal government: “‘The Depression, and with it the growing role of government in the economy, emphasized the need for such measures and led to the development of a comprehensive set of national income accounts’ (as quoted in Landefeld et al., 2008, p. 194)” But just like every government program there are bound to be some problems, and that is where the Strows really hit home. On top of being a government construct built and maintained to allow the government easier control over the economy, they identified seven other problems with GDP.

Their first criticism is that GDP measures how much was produced, not how much utility was generated. If well-being is expressed only by how much is produced then, they argue, the Antebellum South would have had a rather high well-being score due to how much was produced. The only problem is so much of the labor force was not made better by the involuntary transactions. Come time for the slaves to be freed production took a nose dive, but former slaves were undoubtedly happier with what work they were doing. They got to choose how to make a living and contribute to the economy, thus making well-being higher than the previous status quo. They later use examples of the Soviet Union and communist Cuba as further proof that happiness very often outweighs how much stuff people have or how often they work.

Their second criticism is that true wealth creation is not tracked by GDP. Going off the utility argument, a simple lack of any measure to track wealth is present. For GDP’s sake everything made could sit on the shelf to collect dust for all of eternity and the mere fact it was made would count towards GDP. One problem, if government is the one that is making everything, GDP goes up, despite no one actually wanting certain products. Say, for example, the government hired a bunch of people to make literal cardboard pizza, as in pizza made entirely out of pizza. Also say it was their full time job to do this, so 40 hours a week, 52 weeks a year. Absolutely no one is buying the pizza, but it is counting towards GDP, yet no one bought a dime of the pizza. Though at the end of the end of the day the main stressing point is that tracking involuntary transactions are very bad for getting a firm grasp on true economic growth.

Next up is consumption. Consumption is in no way or form growth as it does not create wealth. Instead it does the opposite and takes wealth that already existed and destroys it. The planners who view consumption as important overlook the importance of saving and the value placed on current versus future consumption. Saving, though it decreases current consumption, allows for greater consumption in the future compared to what is capable at current levels. Or simply put, the pile gets bigger.

Fourth up is their criticism based on investment. The investment criticism follows similar lines to the wealth issue. Specificity is the difference with this one as what gets classified as an investment is often wrong, causing data to be skewed. Government induced investment only further adds the issue.

Government spending straight up never helps the general well-being of the economy. GDP accounting for government spending leads to the idea of governments having the ability to magically improve the economy. They point towards how “The vast majority of today’s government spending is consumption oriented”. This picture shows the trend of US government spending as a percentage of GDP, a number they are unfortunately a part of:

With all this said, I decided to do a little experiment to see if what the Strow’s had to say about GDP and government spending was true. Thanks to the generosity of Dr. Hak-Seon Lee of the Political Science Department at James Madison University, I was able to get the solid dataset Global 05 from working with him during a previous project. From that I was able to come up with a basic model, which is as follows: lngdp = education_spending (educatsp)  + defense_spending (defensesp) + public_health_expenditure (pub_health) + fdinet + percentage of urban population with access to improved drinking water sources as of 2002 (h2otot02) + income_tax_rate (income_tax) + corp_tax_rate (income_tax) + ε . What is really interesting and important about the data, and the resulting interpretation, is the fact that all these data is international, allowing one to draw conclusions on a global scale and not just local to any one economy. One quick regression later, and we get the following:

Lngdp = -4.43628 – .08185educatsp + .02763defensesp + .03131pub_health – .00537fdinet

s.e.        (1.46348)       (.02231)                (.02795)            (.09808)                 (.00444)

+ .06696h2otot02 + .05048income_tax + .04037corp_tax

s.e.        (.01116)              (.01697)                     (.03369)                                                                     R2 = 0.4868

So what exactly are we being told here? For starters the model we came up with explains roughly 48.68% of the variation in the natural log of GDP, which is not all that bad. It comes up significant overall as a model at 1% with an F-Value of 11.52. The variables of educatsp, h2otot02, and income_tax are statistically significant at 1% with t-values of -3.67, 6.00, and 2.97 respectively. Insignificant variables at 5% include defensesp, pub_health, fdinet, and corp_tax with t-values of .99, .32, -1.21, and 1.20 respectively. What all does the estimated model tell us about the effects of individual spending areas? I move first to the variables found to be individually significant to keep them at the center of attention where they belong. For starters a 10% increase in central government spending on education as a percent of GDP (educatsp increases by 10) is predicted to decrease GDP by around .81%, a fact that I am rather shocked by. At the same time though only decreasing by .8% is really quite negligible as such tiny percent changes are can very easily go unnoticed by all but the sharpest eyes. Long held notions are really thrown for a loop by the data presented. Next if I were to increase h2otot02 by 10, it is predicted GDP would increase by .67%. Lastly for the significant variables is that of income_tax. Should we increase the income tax rate by 10%, GDP is predicted to increase by .5% . Looking back it is extremely interesting how little rather substantial changes in government spending on a given area, or adjusting the tax rate, really change GDP. One would think that with the big bad government economic growth should be shooting through the roof. Free market economists, like myself, must see these results as nothing but a positive sign that there exists at least one study working in their favor.

I look at the bold line above as the whole point of this piece. We can skyrocket government spending in really any of those areas, but especially defense spending, greater public health, bringing in more foreign direct investment, or hiking up the corporate tax rate and see minimal or even no increases in GDP. Shoot, even when education spending gets increased by 10%, GDP fails to increase by a point. The Strow’s would be proud with this research, if only for using data to confirm their ideas.

At the end of the day, government has very little impact on GDP growth, assuming we continue to use GDP as a measure of economic growth. The figure fails to capture true sources of growth and measure them correctly. Spending money on defense and other areas run into the public goods problem which does not do any good in the long run of growth. Because of everything presented above, we ought be best if we abolish the figure altogether.

-The Jiffy Riddler


Government Screws up GDP? You Don’t Say!