"Excising" taxes from the Heritage Foundation's Economic Freedom Index
My friend Charles is fond of a certain graph that shows a negative correlation between Economic Freedom, as defined by the Heritage Foundation, and Gini index scores, which measure inequality. The argument that accompanies the graph is that free market neoliberal policies, the effects of which are ostensibly measure by the freedom index, correlate positively with equality. I wonder if Charles is unconsciously choosing which correlations to see meaning in, while ignoring other stronger correlations that don't comport with his worldview. However, there is a correlation there that is relatively unlikely to be mere coincidence. Before I go further, I merely want to propose that in general, one correlation of Gini scores should be balanced against other correlations that are similar or greater in their explanatory power. Here is the graph that he cites:
A modest but visible correlation. Ok, say I, there may be something going on with these two variables. I've reproduced his graph with only OECD countries below to make other comparisons with other variables a less arduous task.*
*Data is this blog post was compiled by me from the following websites: https://en.wikipedia.org/wiki/List_of_countries_by_income_equality
https://en.wikipedia.org/wiki/List_of_countries_by_tax_revenue_as_percentage_of_GDP
http://www.heritage.org/index/ranking
Spreadsheet data is available upon request
For OECD countries, economic freedom explains only about 1.25% of the variation in Gini scores. While there may be a correlation here, it's not exactly a strong one. Let's try with another variable to see if we can better explain why some countries are more equal than others.
It seemed to me very strange that a measure that includes tax burden as a sub-component would correlate with Gini in the opposite direction as tax burden by itself. Note: HEF gov. size scores also correlate negatively with equality, but not as strongly as the correlation of equality and total tax burden. I decided to see how the subscores of the HEF index correlate with each other. I suspected that the government size subscore is thrown into the index for ideological anti-tax reasons. And my initial findings support the idea that it does not fit well with the index's other measures:
Note that gov size is a score which rises as government gets smaller
Charles has implied numerous times that increasing taxation would be bad for the economy and therefore worsen inequality. This is shown to be more or less the opposite of what the preponderance of correlative evidence shows. I will not address Hauser's law in depth here, and hopefully ever, because I do not see it as a good use of my time. Others have pointed out that it is little more than an intentionally misleading graph that ignores large changes in total tax revenue. And as I've pointed out, if it is a law of economics, why are other countries free from the effects of this law? Also, top marginal individual and corporate tax rates do not explain even half of the variation of total tax burden across countries. Clearly, even if it were true that tax receipts have not varied in spite of marginal rates changing drastically, it wouldn't be the end of the discussion on how much we should be taxed in total.
However, I do not want to dismiss the HEF index just yet. I believe that the positive correlations between rule of law and market openness, and rule of law and regulatory efficiency show that the economic freedom index, sans the tax burden section, is measuring something important and useful. I would even go so far as to say that it supports liberalization of economies along the lines of improving these subscores, with the important caveat that this should be done with the knowledge that drastically changing total tax burden could result in much higher inequality. I plan to compare the HEF index with the World Happiness Report, to see how we should weigh each of these measures when crafting policy and imagining the society that we wish to construct.
I would like to end on a note of reconciliation. I did find that there are negative correlations, ranging from r^2 values of less than .01 to .0878, between all of the HEF index's subscores and pre-tax Gini. That is, disregarding all taxes, there is some reason to believe, as I have noted above, that the HEF is measuring something that has general applicability to building healthy, stable societies.
A modest but visible correlation. Ok, say I, there may be something going on with these two variables. I've reproduced his graph with only OECD countries below to make other comparisons with other variables a less arduous task.*
*Data is this blog post was compiled by me from the following websites: https://en.wikipedia.org/wiki/List_of_countries_by_income_equality
https://en.wikipedia.org/wiki/List_of_countries_by_tax_revenue_as_percentage_of_GDP
http://www.heritage.org/index/ranking
Spreadsheet data is available upon request
For OECD countries, economic freedom explains only about 1.25% of the variation in Gini scores. While there may be a correlation here, it's not exactly a strong one. Let's try with another variable to see if we can better explain why some countries are more equal than others.
Total size of government is actually a good indicator of inequality, but not in the direction that conservatives might like to believe.My scatter plot and line of best fit shows that there is a much stronger correlation between Gini and total tax burden, than there is between Gini and economic freedom, at least for OECD countries. Tax burden explains about 40% of the Gini, or about 32 times more Gini variation than economic freedom. This is not surprising, considering that Gini is measured after taxes and transfers... after all, that is where people live. Total size of government is actually a good indicator of inequality, but not in the direction that conservatives might like to believe.
It seemed to me very strange that a measure that includes tax burden as a sub-component would correlate with Gini in the opposite direction as tax burden by itself. Note: HEF gov. size scores also correlate negatively with equality, but not as strongly as the correlation of equality and total tax burden. I decided to see how the subscores of the HEF index correlate with each other. I suspected that the government size subscore is thrown into the index for ideological anti-tax reasons. And my initial findings support the idea that it does not fit well with the index's other measures:
Note that gov size is a score which rises as government gets smaller
If Economic Freedom is a "thing" that has real effects in the world, it seems that these effects are entirely or almost entirely separate from the effects of tax burden.The negative correlations within the HEF index's subscores (in red), though small, are with government size and regulatory efficiency, and government size and rule of law. There is a small positive correlation between government size scores and market openness. However, taken as a whole, the other indices correlate rather well with one another, while the tax/government size index does not. The construct validity of economic freedom as a whole seems to suffer when taxes are added to the picture. That is to say, if Economic Freedom is a "thing" that has real effects in the world, it seems that these effects are entirely or almost entirely separate from the effects of tax burden. To reiterate, other subscores of economic freedom correlate positively with equality, while government size scores correlate negatively, and total tax burden scores even more so. I've included the "total tax burden" to show how the total tax burden actually correlates positively with the HEF index sans gov. size. This further casts into question the premise that taxes fit in with the rest of the HEF index.
Charles has implied numerous times that increasing taxation would be bad for the economy and therefore worsen inequality. This is shown to be more or less the opposite of what the preponderance of correlative evidence shows. I will not address Hauser's law in depth here, and hopefully ever, because I do not see it as a good use of my time. Others have pointed out that it is little more than an intentionally misleading graph that ignores large changes in total tax revenue. And as I've pointed out, if it is a law of economics, why are other countries free from the effects of this law? Also, top marginal individual and corporate tax rates do not explain even half of the variation of total tax burden across countries. Clearly, even if it were true that tax receipts have not varied in spite of marginal rates changing drastically, it wouldn't be the end of the discussion on how much we should be taxed in total.
However, I do not want to dismiss the HEF index just yet. I believe that the positive correlations between rule of law and market openness, and rule of law and regulatory efficiency show that the economic freedom index, sans the tax burden section, is measuring something important and useful. I would even go so far as to say that it supports liberalization of economies along the lines of improving these subscores, with the important caveat that this should be done with the knowledge that drastically changing total tax burden could result in much higher inequality. I plan to compare the HEF index with the World Happiness Report, to see how we should weigh each of these measures when crafting policy and imagining the society that we wish to construct.
I would like to end on a note of reconciliation. I did find that there are negative correlations, ranging from r^2 values of less than .01 to .0878, between all of the HEF index's subscores and pre-tax Gini. That is, disregarding all taxes, there is some reason to believe, as I have noted above, that the HEF is measuring something that has general applicability to building healthy, stable societies.
For the moment I will limit my own comments to what you have written about Hauser's Law. As far as I know, no one - not even Kurt Hauser - has suggested Hauser's law is a law in the same sense as supply & demand or marginal utility. It is an empirical rule that has held true since about 1945. Before that the government revenue share of GDP was considerably less than 19%. However, after World War II government revenue has stubbornly stuck around the Hauser limit. Moreover, it is not a limit you can evade simply by increasing tax rates or changing the tax structure, as you can verify from the graph at URL https://i0.wp.com/www.adividedworld.com/wp-content/uploads/2017/03/Hausers_Law-1.png. If you can not find the magic variable(s) to change that would change the Hauser limit, all you would succeed in doing by increasing tax rates would be to reduce GDP growth. If you were to increase tax rates enough to fund the additional welfare income programs you suggested, I would be immensely surprised if you did not actually cause GDP to fall. If that happened and the Hauser's limit remained in effect, you would find your tax policy incredibly counterproductive as tax revenues fell, even though you had increased tax rates. People would simply change their behavior to have less taxable income. The large corporations would just direct their investments overseas (Norway? Ireland?) to avoid income taxes in the United States.
ReplyDeleteOne natural thought about Hauser's law is that it is something culturally dependent with different countries having different values of the Hauser limit. Clearly, countries with more extensive social welfare programs would have to have a higher limit if they have any limit at all. I found support for this idea in a post on The Economist webpage, entitled "Is there a limit to revenue-raising?" by Buttonwood, with URL http://www.economist.com/blogs/buttonwood/2014/10/tax-policy-and-economy. I went to the World Bank to find total tax revenues as a percent of GDP for various nations, but they gave values for the United States that were much too small (around 12%) and time varying. That lead me to recheck the data for the United States. Usefully, the Brookings Institution has republished historical tables from the Office of Management and Budget, which did indeed display Hauser's law. You can find the data of interest at the URL http://www.taxpolicycenter.org/statistics/source-revenue-share-gdp. If you look at the total receipts column (third from the end), you will discover the only significant fluctuations from the 19% limit are fluctuations downward.
Before you decide to attempt to raise taxes to fund more welfare programs, you had better find the right social variables to change to push Hauser's limit upwards. Otherwise, all you will succeed in doing is to create economic and social disaster. Also, you (and everyone else) should start worrying about the implications of the graph at the URL https://i2.wp.com/www.adividedworld.com/wp-content/uploads/2016/12/REV_Entitle.jpg.
By the way, is there a way to more succinctly post a figure with a comment, or to use HTML tags to more concisely refer to a link?