By Nwaodu Lawrence Chukwuemeka
WED, JUNE 7 2017-Countries with larger government sectors tend to have more personal freedom, so says economists, (libertarian economists included) from the Calo Institute. You know Economists love to measure things, so the Human Freedom Index (HFI) from the Cato Institute is a case in point here. Its authors have assembled dozens of indicators of personal and economic freedom. They invite interested researchers to use them to explore “the complex ways in which freedom influences, and can be influenced by, political regimes, economic development, and the whole range of indicators of human well-being.”
This writeup will take a look at what we can learn from the data about the relationships among freedom, prosperity, and government. The relationships may turnout to be not quite as simple as many libertarians might think.
The Human Freedom Index consists of two parts. One is the Economic Freedom Index (EFI) from the Fraser Institute, which includes measures of the size of government, protection of property rights, sound money, freedom of international trade, and regulation. The other is Cato’s own Personal Freedom Index (PFI), which includes measures of rule of law, freedom of movement and assembly, personal safety and security, freedom of information, and freedom of personal relationships. The Cato and Fraser links provide detailed descriptions of the two indexes.
In order to explore the way freedom influences other aspects of human well-being, I will draw on a third data set, the Legathum Prosperity Index (LPI) from the Legatum Institute. The LPI includes data on nine “pillars” of prosperity, including the economy, business environment, governance, personal freedom, health, safety and security, education, social capital, and environmental quality.
The EFI and PFI cover 160 countries and the LPI 149 countries. In this post I will use the set of 143 countries for which data are available in all three indexes. Detailed methodological information are sourced from Cato, Fraser, and Legatum Institutes.
We can begin by confirming a result reported in the introduction to the Cato Human Freedom Index, namely, that economic freedom and personal freedom are closely related. Expressing both indexes on a scale of zero to ten, with ten indicating maximum freedom, a scatterplot of the two indexes looks like this:
The correlation coefficient between EFI and PFI for the 143 countries in the joint Cato-Legatum sample is 0.53—not an especially tight relationship, but statistically significant. The slope of the trend line is 0.91, meaning that each one-point increase in the EFI score is associated with a 0.91 point increase in the PFI score.
The relationship between economic and personal freedom is partly explained by the fact that both are positively associated with income. As the next chart shows, that relationship is nonlinear for both measures of freedom. The log of real GDP per capita, expressed in U.S. dollars at purchasing power parity, provides a reasonably good fit. The correlation coefficients are 0.51 for log GDP and the personal freedom index, and 0.56 for log GDP and the economic freedom index.
Using multiple regression analysis, we can recalculate the relationship between economic freedom and personal freedom in a way that controls for their common relationship to GDP. Taking GDP into account increases the correlation coefficient between the two aspects of freedom from 0.53 to 0.59, but it also reduces the slope of the relationship between PFI and EFI. Each one-point increase in EFI is now associated with a 0.61 point increase in the PFI rather than the 0.91 point increase that was estimated without including GDP. All of these results are statistically significant at a 0.01 level of confidence.
So far, so good. We have found that personal freedom and economic freedom are positively associated with each other, and that both freedom indexes are positively associated with prosperity as measured by real GDP per capita. Good libertarians should expect these results and be gratified to find them confirmed.
The previous section showed that economic and personal freedom are positively related to prosperity as measured by GDP per capita, but prosperity is more than GDP. Libertarians tend to see freedom as also conducive to other aspects of human well-being, such as education, health, and personal safety.
There are many measures of prosperity and well-being available. I hope to be able to explore several of them and their relationships to human freedom in future posts. In this introductory treatment, however, I will limit myself to the education, health, and personal security indicators from the Legatum Prosperity Index. In what follows, I will refer to the average of these three Legatum “pillars” as the education-health-safety index, or EHS, measured on a scale of 1 to 100. (By and large, the results reported below also hold for each of the three indicators considered separately, although some of the individual coefficients are not statistically significant.)
We can begin, as before, with a simple scatterplot of EHS and HFI:
As the chart shows, the relationship between the two variables is positive and strong. The correlation coefficient of EHS and HFI is 0.76. Each of the individual freedom components also correlates positively with EHS, although not quite so strongly: 0.68 for economic freedom and 0.67 for personal freedom.
Since EHS, HFI, and GDP per capita all correlate strongly with per capita GDP, we need to be cautious about interpreting the simple correlation coefficients. For example, it could be that the apparent correlation of EHS with EFI simply reflects the fact that rich countries tend to have good schools, hospitals, and police forces, but that people in rich countries that are free live no better than those in rich countries that are unfree.
We can, again, set our minds at rest by using multiple regression to sort out the individual contributions of each variable. A regression of EHS on EFI, PFI, and the log of GDP per capita yields a strikingly strong result. The overall correlation of EHS and the three variables is an impressive 0.91. Using the coefficient of determination, R2, we can interpret that result as meaning that the three variables jointly explain 83 percent of the variation in education, health, and safety among countries. The contributions of each of the individual independent variables are positive and strongly statistically significant.
It seems, then, that human freedom in both its economic and personal manifestations contributes positively to human well-being as measured by data on education, health, and personal safety—another result sure to please libertarian readers.
Things get more interesting when we dig a little deeper into the reported linkages between personal and economic freedom by breaking the Fraser Institute’s EFI down into its separate components: size of government, protection of property rights, sound money, freedom of international trade, and regulation. When we look at the simple correlations between the personal freedom index and the EFI components, we find they are all are positive, as expected, except that for the size of government (SoG), which is negative. The correlation of SoG with the personal freedom index is -0.16. Remember that for all components of the EFI, a higher value means more freedom, so the negative coefficient means that a larger government is associated with greater freedom. That is not what most libertarians would expect. Is this just an anomaly or a real statistical regularity?
As a first step toward answering this question, we need to see just what the SoG indicator really measures. SoG is itself a composite derived by averaging four subcomponents: government consumption expenditures, government transfers, marginal tax rates, and something called “government enterprise and investment” (GEI), which is Fraser’s name for the ratio of a country’s government investment to its total investment. Examining these subcomponents uncovers two problems.
One is that only the government consumption indicator is available for all countries. Data on transfers, tax rates, and government investment are missing in several cases. Where data are missing, the SoG measure is the average of the components for which there are data. This approach to handling missing data degrades the statistical power of the SoG indicator as a whole.
By analogy, suppose that we want to assess the health risks facing a city’s residents using their body mass index (BMI), their gender, and their age. To measure BMI, we need to know each person’s height and weight, but suppose we are missing the data on weight for some individuals. Rather than leaving those people out of the sample, we could estimate their weight by using the average weight for a person of a given height, age, and gender. However, that procedure would inevitably make our assessment of health risks less statistically reliable than it would be if we had had complete data for everyone in our sample.
We would be continuing this discourse in the next article. For further details call on me for in-depth discussions, busness advisory services and training – send a message via WhatsApp or SMS..
Nwaodu Lawrence Chukwuemeka (Ideas Exchange Consulting).|firstname.lastname@example.org (07066375847).|theG&BJournal Contributor