Give Me Data Or Give Me Death

By Ira Seidman — part twelve of twelve in the series decentralize

There’s an opinion on every street corner, of every street, of every neighborhood in this city, and usually more than just one at that. Opinions range from the informed to the mad, from the practical to the “let’s just send it.” The worldviews which get the most respect tend to be the ones which are based on some sort of data; even then, data can be cherry-picked or just plain manipulated to tell any story we please so there’s plenty of room for doubt no matter what the argument being made is. Even so, the most trustworthy decisions are generally driven by reliable data that confirm what our experience and logic would tell us. With this I would be remiss if I concluded this series without giving you a pile of the data which inspired me to run for mayor and decentralize power.

All of the plots shown below come from a dataset that I put together using the Gallup World Poll and the CIA Factbook along with publications from United Nations Development Program, Our World In Data, the University of Groningen, and the World Population Review (please see Data Sources for links). Thirteen columns were used to predict for the corruption index in 126 countries. This index is a reflection of how corrupt a government is perceived to be by its people as measured through millions of interviews; a higher corruption score on this index actually means the government is less corrupt because the index is with respect to happiness, so a government that is perceived as less corrupt will bolster happiness (this index is referred to as Less Corruption in the plots to avoid confusion). For the full study please check out the repository associated with this project which houses both the notebooks I used to create this analysis as well as the presentation associated with it.

The number one finding from this study is that there is no smoking gun for decentralizing power. This however, as with so many of the social sciences, is not really a finding because we already knew it. If there was slam-dunk evidence that decentralizing power leads to less corruption I probably would not be running in a race where I am the only one proposing more direct forms of government during a time when public trust in politics is at historic lows, the direction of the country is always considered more negative than positive, and the approval rating of Congress oscillates between the single digits and 40%. Below are two plots that look at the relationship of over 100 countries’ autocracy-democracy polity scores on the x-axis against perceived corruption on the y-axis. The polity score is the standard measurement for how autocratic (-10) or democratic (+10) a country is and the perceived corruption index is a part of the World Happiness Report’s way of objectively assessing quality of life around the globe. The second plot excludes the most autocratic nations (less than or equal to a polity score of -8) which just include Qatar, Saudi Arabia, and the UAE; this plot is worth considering because these three countries are large outliers due to their tremendous oil wealth which can support high standards of living despite extremely centralized governments. When analyzing the effects of economic prosperity on perceived corruption it will be clear why this is so important:

In both plots there is a very small correlation between polity score and perceived corruption. When the three oil-producing autocratic countries are included there is actually a slight negative correlation in perceived corruption as countries get more democratic. While this certainly is not a call to decentralize power, it is not damning evidence of mob rule either. The correlation between decentralized government and perceived corruption is -.082 which is almost neutral as correlations are considered perfectly negative at -1 and perfectly positive at +1. The strongest correlations with a perceived decrease in corruption turned out to be median individual income (.62) among other economic indicators and the perceived freedom index (.53):

It probably doesn’t take a data scientist to figure out that people are less angry with the government when freedom and economic opportunity are abundant. What is interesting is how a country’s Gini score effects perceived corruption, or rather how it does not. Gini scores are evaluated between 0 and 100 where 0 is perfect wealth equality and 100 represents a society where all of the wealth is in one person’s hands. As Democrats, one of our biggest concerns is how unfair wealth inequality is — all the ways money can buy influence which generally is used to make more money. It turns out however, wealth inequality is very weakly correlated against perceived corruption with a correlation of just -.04. Both the autocracy-democracy polity scores and Gini scores are however substantially correlated with happiness:

As is so often the case, data can be flipped around to accommodate our case or the case can be flipped around to accommodate the data. While I probably could have found other sources that make my campaign to decentralize power look more urgent and supported by data, I would rather just share the facts that I have found and let them guide whatever discussion they can. Facts such as the overwhelming effect of decentralizing political power on a country’s happiness index or the slightly underwhelming reassurance that decentralizing power will at least not exacerbate corruption. However this data strikes you, I strongly ask that you consider the positive effects that a more decentralized government can have on our society.

Today is the Democratic Primary in New York so I do not expect this data to get me many votes, but it is critical as citizens that we vote for people whose ideas will actually bring about the change we want to see as supported by the facts. Maybe my campaign was just trying to build a society based on happiness all along, and if so, it seems like decentralizing power through referenda and online voting would be a great way to do just that. Whatever your political ideology is, hold your representatives and your own voting habits to account with better data that is both logically consistent and has historical precedent. Cut out the division, the conquering, the fear mongering, and the lies in favor of a government and society built on love, cooperation, justice, and whatever else the data tells us is in reach.

Data Sources

Our World In Data — https://ourworldindata.org/democracy
United Nations Development Program — http://hdr.undp.org/en/indicators/103006
The University of Groningen — https://www.rug.nl/ggdc/productivity/pwt/
Gallup World Poll — https://www.kaggle.com/unsdsn/world-happiness
World Population Review Median Income — https://worldpopulationreview.com/country-rankings/median-income-by-country
World Population Review Population — https://worldpopulationreview.com/
CIA Facebook — https://www.cia.gov/the-world-factbook/field/gini-index-coefficient-distribution-of-family-income/country-comparison