How to cheat an automated election – a theoretical and hypothetical discussion from Institutional Economics

In his 2007 book, The Institutional Economics of Corruption and Reform: Theory, Evidence and Policy, Johann Graf Lambsdorff outlined a theory of corruption.  Briefly, he surmises that corruption is a preventable crime if there is a sufficient incentive for one of the parties in it to renege on the “deal.”  Preventing corruption then boils down to designing incentives for cheating on an illegal contract.

Insanity is better

What does this theoretical approach have to do with poll automation?  Well, it seems to give us some comfort about the likelihood of cheating. The prevailing view of Comelec officials is that automation impedes if not prevents the cheating called dagdag-bawas for which one of its officials was supposedly recently fired.  This form of cheating involves a contract between a candidate and a Comelec insider who promises, for a consideration, to “fix” the election results.

I may be wrong, but it seems that the Comelec officials are correct.  Lambsdorff’s theory suggests that it is significantly more difficult to cheat when a machine is involved.

Consider the May elections when the counting of votes will be computerized.  How can one cheat?  According to some experts, it can be done through an “inside job.”  Someone who knows exactly how the machine works can work the voodoo magic.

Voodoo can work if there is not enough transparency in the vote count.  For example, the Comelec could decide to do the random manual audit after proclamation, and insiders could then “fix” the random audit to validate whatever results they have concocted.

But there is hope.  Any lack of transparency applies to the electorate (this is bad) as well as to a cheating candidate (this is good, as explained later below).  The cheating candidate would want to ensure that the “fix” really went in his favor.  How would he do that?  He would have to ask the insider to show him how the cheating would work.  The insider would then show him, and both parties to the fix would then have to seal their lips forever.

How probable is the scenario above?  I estimate that the probability is low.  This is because the insider would have to convince the cheating candidate that the cheating would really work.

Bear in mind that dagdag-bawas probably took a while to perfect, as it had to do with manipulating the results in between the manual vote count and the manual canvass.  In the past, the cheating candidate would know through his watchers what the poll count really was, and he can compare that with the official outcome.  My best guess is that the payment for dagdag-bawas would have been set per manipulated vote, for example P100 per vote, so that a candidate who wants to overcome a 10,000 vote margin in a close two-candidate contest would have to buy at least 5,000 manipulated votes (and pay P500,000 to the cheating insider).   This may be the best deal if the going price for “bought” votes is significantly higher than P100 per.  Of course, if the object of the contract is 1 million votes, the price goes up to P100 million.  The important thing is that the money is well spent because the cheating candidate can verify that he got what he paid for.  If he didn’t, he just might squeal on the insider side, or he wouldn’t deal again with such an insider.

But with a lack of transparency of the machine count, the cheating candidate could never be sure that he got what he would have paid for.  It would be a difficult “selling job” on the part of the insider to convince the candidate that the cheating would really take place.  (It is not impossible, since it is possible to conceive of a scenario where there is a “secret” audit for the benefit of the cheating candidate.)

What can a cheating candidate do?  He could concentrate his resources into buying votes retail.  But with the machine count, he again confronts the problem of lack of transparency.  He may have spent say, P200 per retail vote, and in a 1,000-vote precinct where he buys 50 percent, he needs to know that indeed, the 500 votes he bought went his way.  In manual elections, the poll watchers can give a rough confirmation that the votes went as bought.  In the automated elections, it seems there is no way to get such a confirmation.  The cheating candidate would have to “buy” a secret audit of the precinct results.

It is perhaps difficult to keep a manual audit secret, so there might be a search for a computerized audit.  A possible scenario is for the images originally scanned by the machine to be transmitted to the cheating candidate.  But again, the cheating candidate can never be sure that the data he might get from his insider partner represent what was really scanned.  He is just as technologically ignorant as the rest of the electorate, compared with the insider.  Cheating would involve a large element of faith on the part of the cheating candidate in the “honesty” of the insider side of the contract.  This is an interesting instance of a need for honesty among thieves.

But could “too much” transparency work in favor of a cheating candidate?  It seems that the more transparency, the better it is for an honest result.  Transparency makes the job of the insider very difficult.  And the more difficult it is, the higher the price he will charge, and the lesser would be the willingness of the cheating candidate to “buy in.”  This means that it is good to insist on the random manual audit being done before proclamation.

So there it is.  Automation is good because the very forces that would want to conspire to manipulate the results need some element of masquerade.  But that lack of transparency also makes at least one side of the criminal conspiracy reluctant to make a deal.  It is justifiable schadenfreude to enjoy the predicament of the insider-manipulators:  Transparency dooms them, whereas a lack of transparency raises the probability that there would be no demand for their services.  Either way, the likelihood of cheating is diminished.

So it seems.


Benjamin Franklin is credited with the following quotation: “If you can’t pay for a thing, don’t buy it. If you can’t get paid for it, don’t sell it.”  Thus, the discussion above makes sense if potential cheats believe in Franklin!


Homework for students of Econ 11 (Intro Economics):

Illustrate the concept of missing markets embedded in the application of Lambsdorff’s theory on corruption as discussed above.

Hint:  To get full credit, first draw the following diagram, then explain what it means.

Watson, what is all this? - Sherlock H

2 thoughts on “How to cheat an automated election – a theoretical and hypothetical discussion from Institutional Economics

  1. other than the usual, cheating in this election is not simple yet it may not require any sophisticated programs like a ‘sudo’ to do. first, it is not really that automated. it is only the counting that is being electronically processed. pen and paper work is will be done before 6:00pm on may 10 and all the machine will do when the ballot is cast, is collect all the “initiman na hugis bilog” and store it in some kind of a database. after that, math starts.

    inside the CF cards are the programs that do the said math. any programmer and even highschool studs who listen to their algebra teachers, would know whats next.

    A simple example:

    T = Total Votes
    M = My Candidate’s Votes
    N = Your Candidate’s Votes
    x = winning factor
    y = losing factor
    in other words, x>y
    provided: 0 < x+y <= 1


    M = T*x
    N = T*y

    My Candidate will surely win.. simple math, isnt it?

    more complicated formula can be written in whatever programming lingo in the CF card or maybe even just a configuration text file. this will tell the PICOS machine what do then.

    what's happening may be more complicated. i just simplified it and i know that what i think is not even unique. i believe, a lot of people have thought about this, maybe better than this. 🙂

    i have been fooled once. by my ex gf. hehe!
    so, any candidates there! my vote is for sale! $1,000,000,000.00 per candidate! hahahaha! 🙂


    1. Here’s an unproven urban-legend. The techies doing the CF cards knew the ballot was redesigned, but they were not told to change the configuration. Some perhaps thought they should tell their “bosses.” But since the latter were arrogant a**h***s, the lowly-regarded techies kept things to themselves. Ergo, 76,000 CF cards went out the door only to be recalled. Of course, the techies had a field day with overtime work fixing up the mess their super-smart bosses didn’t think about. Shows you the power of the weak. Hehe.. (do it a la David L).. Whatever the truth of this story, the fact is that Smartmatic-TIM shipped out untested CF cards. A bit like giving a buyer a brand new car without checking if the thing had gas in the tank! What a joke of a management crew there must be at Smartmatic-TIM.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s