Delegates’ voting behaviour: Elusive, but just for Now

 

 

Greenwell Matchaya, PhD

One of the most frustrating things for any investor is being unable to know the possible returns of his/her investment with some level of certainty. Similarly, for a political investor/politician it can be frustrating to devout so much effort in a competition where the voter behaviour is unknown, because doing so would easily lead to wastage of resources or in other words, would limit gains per dollar invested.

While voter behaviour is complex and often elusive, when the same or similar voters vote for candidates over a number of times, it is easier to demystify their behaviour. Using known techniques of analysis often taught in elementary statistics, econometrics and mathematics classes, it is possible to study the patterns and derive generalizations or even “theorems” around such behaviour. With advances in machine learning and AI in general, and a good understanding of physiology, sociology, one can torture data or even create some and predict future behaviour. Such predictions can inform political investment decisions and for those who can access such analytics and use them, it would be possible to increase rates of return on political investments, under certain assumptions.

Thus, I would say with certainty that where data for past voting behaviours are available, agile and sophisticated politicians can tilt competitions in their favour as they would understand their voter behaviour and indeed their own locational voter equations.   For instance given the tools at hand with many, it is trivial to predict outcomes of presidential and parliamentary elections give several behavioural assumptions of politicians and political parties. This is possible because the data generated by voters since democratic voting begun in 1994 are available and although there is attrition as people change locations, populations and their structure changes, certain parameters in those populations, which help explain behavioural formation are immutable in shorter horizons and even if they change , such change is capable of being modelled. This is good because one can use the understanding of such patterns and voter behavioural equations to move pieces, cobble alliances, or choose running mates, or even know who to work with etc. Nevertheless, one can also simply use those to explore business deals with predicted winners etc.

Unfortunately, for now, the delegate voting behaviour is still a puzzle. This is not because delegates are weird of strange, rather, this emanates from the fact that voting by delegates has become popular just recently and each political party appears to subtly differ from another in terms of how such delegates are given power or identified. Equations to characterize their behaviour are not yet clear which makes investment in such party level competitions not as less risky. Thus, it is not surprising that many casualties of delegates voting in DPP, MCP and more recently UTM were caught off-guard. They did not expect the results simply because they did not have sufficient ground to realistically predict  their odds, for lack of past data on voting behaviour.

The good news is that just as it is now possible to predict future elections (eg 2025 elections) outcomes with precision once teams are formed or ex-ante, under certain assumptions, the delegate voter behaviour will not be elusive for too long from now. The databases being built by the elections of individual political parties will become more valuable as they receive more data from subsequent elections. Soon, it will be possible to predict odds of success given certain assumptions. Such abilities to prospect the future with reasonable precision will save many politicians from time and unexpected financial losses.

Sometimes such predictions can be right but one may still suffer implications if they do not interpret the results well. For example, you can predict a win by a person or a party etc., but that is different from you as an individual, winning. This point will be clarified in future issues. Therefore, bottom line is that if your time and resources are so important, please respect data before you can spend much, unless you are spending someone’s time and resources of course.

 

Feedback to greenwellmatchaya@yahoo.com

Disclaimer: These views are those of the author and do not represent views of anyone or any institutions associated with the author directly or indirectly.