Influence of Per-capita GDP on the ISO OOXML Process
A report from the Digital Standards Organization ISO working group, released 13 January 2008
It is generally accepted that one of the main economic indicators of any country is its Gross Domestic Product (GDP) per capita. This indicator measures the country’s economic capacity to generate wealth, divided by the number of its citizens.
The GDP per capita is known to be positively correlated with other key country development indicators like average education levels, health care system coverage, social welfare protection, etc.1
In face of the recent and ongoing turmoil regarding the ISO DIS 29500 standard proposal also known as Microsoft OOXML , we decided to investigate whether the recent voting process reveals any asymmetry between the GDP per capita of two different pools of countries - those who voted for the approval of OOXML and those who voted against it.
We collected the most recent GDP per capita figures (adjusted to the Purchase Power Parity), available at the International Monetary Fund website  and applied a Wilcoxon rank sum test to the data (with normal approximation)2. We also added Cuba to this list, using the CIA Fact Book latest figures .
By doing this we were able to reject the hypothesis that “there is no significant difference between the average GDP per capita of countries who voted for OOXML and countries who voted against it” in favor of the alternative hypothesis: “The average GDP per capita of the countries who voted for the OOXML proposal is significantly lower than the GDP per capita of those who voted against it”. In fact, the p-value of the test is 1.01%.
Moreover, this is not a slight difference. The average GDP per capita (adjusted to PPP) of the countries who approved the proposed text amounts to 13843 USD, while the same figure for countries who voted against it amounts to 23297 USD. That represents a difference of almost 10000 USD per capita.
Countries with lower GDP per capita usually have weaker institutions representing its citizens and weaker economies that depend more strongly on foreign multinationals. For instance, the correlation between the Corruption Perceptions Index from Transparency International  and the GDP per capita is of 0.8322.
As we know Microsoft was until recently the biggest company on the planet, so it’s natural to wonder whether its lobbying power has anything to do with the the approvals by some less developed countries. Although nothing in that regard can be directly obtained from these numbers, recent news about the last minute influx of participating members3 on the ISO JTC1 working group suggests there is something unusual going on .
It should be noted that normalization processes are usually built upon consensus and controversies such as the one surrounding DIS 29500 are highly uncommon.4 If we look at the ballot results for other recently approved standards such as ISO 26300 (Open Document Format)  and ISO 32000 (Portable Document Format)  we'll find a near 100% approval rate.
The scenario for DIS 29500 clearly stands out from the usual consensus. We have shown that there is a strong correlation between per-capita GDP and voting patterns on OOXML, and other studies have shown the strong correlation between per-capita GDP and corruption. We sincerely hope that ISO acknowledges the full implications of the results of our study and moves forward to ensure transparency and rigor on the normalization activities.
- Montgomery, Douglas C. and Runger, George C., Applied Statistics and Probability for Engineers, John Wiley & Sons, 1994
- New York Times coverage - http://www.nytimes.com/2007/09/04/technology/04soft.html?fta=y and http://www.nytimes.com/2007/09/05/technology/05soft.html
- "Corrupt countries were more likely to support the OOXML document format", EFFI - http://www.effi.org/blog/kai-2007-09-05.en.html
- International Monetary Fund – World Economic Outlook Database, 2007 Edition - http://www.imf.org/external/pubs/ft/weo/2007/02/weodata/index.aspx
- The World Bank Group – 2006 World Development Indicators - http://devdata.worldbank.org/wdi2006/contents/cover.htm
- Central Intelligence Agency, The World Fact Book - https://www.cia.gov/library/publications/the-world-factbook/
- Transparency International - http://www.transparency.org/
- ISO/IEC JTC 1/SC 34 Working Group Status Report - 29/11/2007 - http://www.jtc1sc34.org/repository/0940.htm
- ISO/IEC JTC1 Directives - http://www.jtc1sc34.org/repository/0728revc.htm
- ISO/IEC 26300 Voting Summary - http://www.jtc1sc34.org/repository/0728revc.htm
- ISO/IEC 32000 Voting news report - http://www.infoworld.com/article/07/12/05/PDF-approved-as-international-standard_1.html
We were able to reject the hypothesis that “there is no significant difference between the average GDP per capita of countries who voted for OOXML and countries who voted against it” in favor of the alternative hypothesis: “The average GDP per capita of the countries who voted for the OOXML proposal is significantly lower than the GDP per capita of those who voted against it”. In fact, the p-value of the test is 1.01%.