by Tom Kando
Abstract:
This article attempts to show that country SIZE (population) , as an independent variable, can predict quality of life. That is, smaller countries enjoy a better quality of life than larger countries. The dependent variable - quality of life - is operationalized through three indicators: per capita GDP, the murder rate, and life expectancy. It is shown that smaller countries indeed enjoy higher per capita income, lower murder rates, and longer life expectancy. Correlations between the three dependent variables are also examined: As expected, the relationship between per capita GDP and life expectancy is positive, and the relationship between the murder rate and life expectancy is negative. However, the relationship between per capita GDP and the murder rate turned out to be POSITIVE, which came as a surprise.
This study is largely descriptive, not explanatory. While I offer a few explanations, my aim is not to provide a detailed causal analysis. The relationships I examine are quite possibly spurious. They are certainly part of a much more complex set of variables, including political, cultural and geographical factors. However, these data offer a global view of how four major variables interact.
Methodological Issues: Mean vs. Median and Individual vs. Household
_
The Mean ( X ) and the Median are two statistical measures of central tendency. That is, they both summarize a characteristic of an entire group, for example its INCOME. The Mean is obtained by summing all individual incomes and dividing by the number of individuals:
___
X = jX
The Median is the number that is halfway into a set. To find the median, the data should first be arranged in order from least to greatest. The Median is the 50th percentile:
Median = 1 (n+1)th value
2
“Average” and “per capita” can be used synonymously with “mean.” The simplest way to compute the per capita/average/mean income of a country is to divide the country’s GDP (Gross Domestic Product) by its population. For example, in 2014 the United States’ GDP was $17,420,000,000,000 (seventeen and a half trillion). When you divide this by the American population - 322,583,000, three hundred twenty two and a half million - you get $54,100 as the annual income of the average American.
However, the Median is a much better indication of the income and standard of living in a given country. Why? Because it is not skewed by abnormalities in the extreme ends, for example the presence of a small number of billionaires. Consider the following example, a class of 100 college students: Let’s say that they make from $10,000 to $40,000 annually, with quite a few starving students among them. Their average income is $18,500. Now add to this class a 101st student - say the child of a Silicon Valley billionaire tycoon, with an annual allowance of a million dollars. Suddenly the average income of the class goes up to over $28,000.
When you include just one or a few very rich people in the mean/per capita/average income of a population, you create the impression of a much richer population than is actually the case.
This is what is going on in the United States: Annual per capita income in America is, as just mentioned, $54,100. This suggests that your typical four-member family of two parents and two children lives on $216,400 per year. Of course this is absolutely not so. Everybody knows by now that America’s income distribution is very unfair, and becoming more so every year. Other countries also suffer from economic injustice. In Latin America and in many other parts of the Third World, inequality is even worse than it is in the US. In much of Europe it is less pronounced.
Because the Median reflects a truer picture of MOST people’s income in any given country, I went looking for those data. I found a couple of sources. One is a 2013 Gallup study and another one is a 2012 list of OECD countries: Median Household Income .
Unfortunately, (1) these studies present Median HOUSEHOLD income, not individual income, and (2) they only list thirty to thirty-five countries - primarily the advanced countries of the Western World.
Even so, one thing immediately jumped at me: Median HOUSEHOLD income in the United States is listed as $43,585 in 2006-2012 (Gallup) and even less than that in the OECD study. And this is not even per capita income! A household is defined as comprising 2.58 people. Therefore, according to these sources, median INDIVIDUAL US income would have been $16,893 in 2006-2012.
The authors warn that these studies vastly underreport the US amounts, compared for example to census figures. They argue that a truer figure is $58,997 between 2006 and 2010.
In addition, we should extrapolate this by multiplying the Gallup figure by 1.16 for inflation, to $68,437 for 2014.This means that the median for individuals in 2014 is therefore $70,796 divided by 2.58 = 26,530.
So you can see that no matter how you slice it, MEDIAN American income is much lower than MEAN/PER CAPITA income - perhaps only HALF as high. This reveals how skewed America’s income distribution is, with most people struggling, while a few are extremely wealthy.
It would have been far preferable to compare the world’s countries’ MEDIANS rather than their MEANS. However, I could find no comprehensive list of Medians. The Gallup and OECD sources only list the thirty to thirty-five highly developed countries in the world. So I had te resort to comparing per capita GDP after all. For now, this will have to do, as a rough approximation of the standard of living in various countries. I ended up using the following source: List of countries by GDP (PPP, i.e. Purchasing Power Parity per capita). Of the three options at this sight, I used the CIA data, because theirs is the most comprehensive list of “countries” - two hundred and thirty. These include a variety of jurisdictions, “principalities,” “dependent territories,” etc. (Places such as Hong Kong, Macau, Greenland, the Isle of Man, etc.). Without quibbling, whatever jurisdiction is included in the CIA list I accepted as a unit of analysis, i.e. as a “country.” For most of these countries and jurisdictions, the data are for 2014.
Other Variables:
For my independent variable - Population - I used the following source: List of Countries and Dependencies by Population. The figures are generally for 2013.
For Murder Rates, I used List of Countries by Intentional Homicide Rate. For a significant number of jurisdictions, other sources had to be used.
For Life expectancy, the following sources were used, by and large: List if Countries by Life Expectancy, Geohive and Life Expectancy, Country.
The use of Chi Square, a non-parametric test of statistical significance:
There is no need to apply any test of significance in this study, be it parametric or non-parametric, because I analyze the entire world population of countries, and there is no larger universe to which my findings can be generalized. Whatever relationships my study documents are REAL. Nevertheless, I did Chi Square tests on my six sub-hypotheses in order to gauge the STRENGTH of each hypothesized relationship. Partial and multiple correlation coefficients or regression analysis would be far in excess of the main purpose of this article, which is to illustrate empirically a few widely suspected but rarely demonstrated relationships ON A WORLDWIDE BASIS.
The Descriptive Data
(To be continued)
© Tom Kando 2015
leave comment here
