WHAT DO MALAWI ELECTIONS AND AMERICAN AFFORDABLE CARE ACT HAVE IN COMMON

I know everybody is focused on the FIFA World Cup. This is the best time to sneak in a few lines on the technical side of the Malawi elections. This story starts on the night of May 20, 2014 from the point in the elections when a polling center has finished counting votes and is ready to report them to the national tally center.

According to the United Nations Development Programme (UNDP), a software system was developed by Globe Computer Systems and partly funded by UNDP for purposes of submitting results. There was a demonstration of the system on May 14, 2014. The idea was: “… results will be manually aggregated at 4,445 polling stations and delivered to the Constituency Tally Centers in 35 District Headquarters. The data in Tamper Evidence Envelopes will be digitized and transmitted to the National Tally Centre”. My assumption is that this is a Web-based system. The transmitted data would be received by software that would update a database. Those interested in results would just do a web search and voila.

The first tally center to attempt to submit results was unsuccessful. And so was the second and third and the rest of them (I am exaggerating here). The software was not used at all. Instead the results were photocopied and sent by the old-fashioned fax machine. The Electoral Commission Chairman Maxson Mbendera, my former SCOM Chair at Chancellor College, said “Our electronic counting system has crashed, yes, and last night we migrated to our Plan B”.

Then started a long protracted electoral process that led to law suits, political rancor, and sadly one death. What went wrong?

Those of us in software development are all too familiar with these scenarios. This problem is not unique to Malawi. For the American public, a similar system failure happened at the roll-out of the Affordable Care Act in October 2013.

Modern World Wide Web (Internet) based software systems are very complex. They rely on transmission over many different types of media such as telephone lines, cellular wireless antennas, fiber-optic cables, and others. Users use these systems from many different types of computers such as smartphones, tablets, laptops, as well as desktop computers. Moreover, the software used to access Internet-based systems tends to be heterogeneous. Some users prefer Windows while others prefer MacOS. There are those who prefer Internet Explorer and those who swear by Firefox.

The most difficult aspect of Internet-based systems is to anticipate the number of users. This affects the amount of data traffic and communication connection requests. Many times such systems are tested with a few users and they work. Unfortunately such tests do not simulate real life situations. Crashing is the only language the system knows when overwhelmed. 

The answer to this problem is called Load Testing or Stress Testing. The Electoral Commission should have had a day when a high volume of requests were generated for the system. But this was probably not adequately done. What we need in the world are better and competent Software Engineers. 

Lessons from Malawi 2014 elections

The main lesson from the May 20, 2014 presidential, parliamentary, and local elections in Malawi is that there is no change. Voters in Malawi cast votes on a regional basis. There were four major candidates. Peter Mutharika of the DPP swept the Southern Region, Atupele Muluzi of UDF swept the Eastern Region, Lazarus Chakwera of MCP swept the Central Region, and Joyce Banda of PP swept the Northern Region. Parliamentary and local voting patterns were similar. The overall winner was Peter Mutharika.

Joyce Banda is an interesting case because she comes from the Eastern Region. However, her husband Richard Banda is from the Northern Region.

The Southern Region is the most populous region of Malawi. The past five multiparty presidential elections in Malawi since 1994 have been won by a candidate from the Southern Region. At this point conventional wisdom should tell us that the major presidential candidate from the Southern Region should win the elections. So why were the 2014 results surprising?

There were several factors that most pundits assumed would play a role in deciding the winner. The first was simply that the incumbent, Joyce Banda, should be favored to win. The second was that there were three major candidates, including the incumbent, from the Southern Region which would result in splitting the Southern vote. This would have two possible effects. One scenario was that the candidate from the Central Region would benefit. The second scenario was that the candidate from the Southern Region who also carried the Northern Region would carry the day. The third was the freshness of Lazarus Chakwera. For the first time in Malawi politics there was a candidate not imbued with past political controversies.

These assumed factors did not have the assumed effect, at least not to the level that would affect the outcome. The surprise was that the outcome was business as usual.

Going forward, political parties whose main following is not in the Southern Region will need to come up with new strategies. One possible strategy is to field a candidate from the Southern Region. The parties should be cautioned that just fielding a Southern candidate is not adequate. Joyce Banda is from the South but the perception was that she is from the North. The Southern Region electorate must embrace the candidate as one of their own for such a candidate to have a chance of winning. The problem with this strategy is that it condemns all non-Southern candidates from ever winning the presidency.

A second possible strategy is to form credible coalitions. A candidate from either the Northern Region or the Central Region standing on a ticket of the party that has a large following in those respective regions needs to peel off Southern support. Again parties should be cautioned that voters abandon their own regional party when they sense that the leader is betraying them by teaming up with a rival.

The third strategy is one that I hope will be fulfilled one day. That the main political parties will become truly national. My hope is that elections will be contested competitively from north to south and from east to west by multiple political parties. I hope such a day will be hastened.