IMANI: Moving beyond EC’s self-serving vendors posing as objective experts

In the last week and half, we have all been subjected to the lecturing of “consultants” and “experts” leading the charge for Ghana’s Electoral Commission’s (EC) wasteful push to toss millions of dollars of barely used laptops and handheld devices in order to pave the way for the EC to spend $150 million (contingency inclusive) on a brand new system. As many as two thousand of these special laptops are stated in the EC’s own budget documents as having been purchased as late as 2018.

One particular gentleman, Dr Yaw Ofori-Adjei, has been very vocal in pushing the agenda to throw away all these equipment and buy new ones.

We need to get one thing straight: everything the so-called IT experts engaged by the EC says must be discounted completely and without reservation, because they are not coming from unbiased experts proferring objective technical advice solely in the interest of Ghana or even the EC.

Dr Ofori-Adjei is a co-owner of Fairgreen Limited, an Accra based commercial equipment vendor, who has already benefitted from EC contracts as a result of the removal of the EC’s previous consultants, STL.

Fairgreen had sent proposals to the EC to provide technical support and maintenance in place of STL and accordingly been paid significant amounts of money to step in when the EC threw out STL.

No one is a champion of STL. But STL was merely a contractor and partial integrator. Upon their removal, the EC had every opportunity to work directly with the original equipment and software makers, especially HSB and Genkey, to obtain all the requisite technical support and system replenishments. Of course, this would have made Dr Ofori-Adjei redundant and threatened the considerable amounts of money his company stood to gain from technical services and support contracts to the EC, especially ahead of the June 2019 electoral exercises. He, understandably, cannot support anything that can put his livelihood and his company’s profits at risk.

Dr Ofori-Adjei and Fairgreen’s continued push for a complete replacement (note, replacement, not even overhaul) of the existing system can easily thus be seen as entirely self-serving and motivated by commercial interest.

Should the “continuous upgrading and maintenance” model be used and the middlemen, including STL, removed, in favour of dealing directly with the existing system’s developers or other 360 degree providers, Fairgreen’s self-serving advice, which is going to see Ghana pay almost twice what countries like Zimbabwe and Nigeria are paying for the same systems we are procuring, would have been rendered useless.

Dr Ofori-Adjei’s zeal for a new system ought thus to be interpreted in this light.

It is in their interest to undermine the existing system so that they can be relevant in the procurement of millions of dollars worth of new equipment and continue to justify their juicy fees. Such “IT experts” have already taken $350,000 merely for priming the existing system for electoral activities in the last year, according to the EC’s own public statements. Far more money is at stake now.

Let us not be side-tracked by self-serving vendors. Let us, instead, focus on the essentials of the matter, and these are:

1. The biometric voter system we have in place now has, roughly speaking, seven main components, each with a different lifecycle and requiring a different type, schedule and scale of periodic maintenance:

A. The databases with biometric data (fingerprints and photographs).

B. The core software that checks and authenticates the biometric data to block multiple registrations and duplicates. A.k.a (ABIS).

C. The laptops, cameras and scanners used to collect the biometric data. A.k.a BVRs.

D. The handheld devices that authenticate voters during voting and exhibition. A.k.a BVDs.

E. The administrative software that manages all of these processes. They are accessed as the administrative interface of the entire system from the EC’s various offices. A.k.a VMS.

F. The central datacenter in which the data and software are installed. I.e. the “server farm”.

G. The communications and transmissions infrastructure that the EC uses to move data around. Especially the ground satellites (VSATs).

2. To consider uprooting this whole system at once and replace everything in one go is the nearest thing to “madness” many of us have seen in the long history of wasteful conduct in the public sector.

The best way to think of this biometric voter system is to think of a bank’s branchless banking platform. Different companies make different parts and you will always be replacing different parts at a different pace. Rain can knock down one satellite. Someone can spill the tea on a laptop. A BVD can go missing. A software driver can be corrupted. etc. It is “madness” to argue that such a complex system can be managed in any other way apart from continuous upgrading and periodic overhauls.

3. The EC biometric system as a whole was audited in 2016 and equipment that was not at peak performance replaced. Over $40 million has been spent since then to keep it in top-notch shape. Whenever we add new districts we implement a fresh VMS and buy satellites etc to bring those EC district offices onboard the system. It is these improvements based on growing understanding as our capacity to manage the system matured that made the 2016 biometric elections so much more superior to the 2012 elections, where biometric system failures led to, for the first and only time, premature termination of the vote and postponement to a new day.

4. Repeat: we have spent over $40 million since 2016 based on the recommendations of an objective system audit to get the existing system into top shape and the quality of the elections we have had since then can testify to that. All the CODEO and EU reports tell us that the 2016 elections and biometric operations were smoother than the 2012 version and that the 2019 district level elections and biometric operations were smoother than the 2016 version. For instance, reports indicate that 33% of polling stations recorded failure of BVDs in 2012. This dropped to less than 7% in 2016 and went even lower to less than 5% in the December 2019 elections. This is the context in which a false rejection rate of just 0.64% was recorded in the 2019 District elections. You may compare this rate to the 3% false rejections figure recorded in a study of Indonesia eKTP, one of the world’s largest and most sophisticated biometric systems. India’s famed and acclaimed Aadhar system has a reported average false rejection rate as high as 5% (see: https://scroll.in/article/822764/chhattisgarhs-way-of-dealing-with-aadhaar-when-fingerprints-fail-take-photos). So how can a system that can be shown to be objectively getting better and has proven just a month ago to be more accurate than some of the world’s best, be described as “obsolete”?

More worrisome than even the plain disingenuity is the recklessness. Freshly implemented biometric systems tend to have more teething problems. In many countries, this pattern of brand new biometric installations facing challenges that only get fully resolved with time and maturity has been consistent. In 2011-2012 in Nigeria, the new biometric system became an embarrassing spectacle when it rejected former President Obasanjo and thousands of Nigerians in the initial phase of the rollout. The recent situation in Afghanistan where the vendor, Dermalog, had to be flown in last minute to fix major glitches, with the announcement of results taking more than 6 weeks, is a very cautious tale for Ghanaians.

5. We have comparable biometric systems in Ghana that have also relied on infrastructure built over a period of time, undergoing periodic upgrades when necessary to maintain peak performance. The biometric passport system is an example. It dates from 2010. At this point in 2020, a person can have a new passport issued without having their biometrics taken all over again simply because new vendors have been building on top of the existing system, upgrading and overhauling different parts of it at different times.
Only a vendor with their own interest to serve will counsel throwing away $60 million worth of systems (accounting for depreciation) to start from scratch. Systems that are used so infrequently that we still have thousands of BVDs that have been used only once, and in some cases never at all. Consider the case of the 2018 BVRs that have only been used for a few weeks in 2019 during the limited registration, for example. The idea that all Ghanaians of voting age should go and queue under the hot sun at the cost of $70 million to supply their data all over again is sheer wickedness. Even more preposterous is the argument that registering only the 1.2 million voters who have recently come of age or who couldn’t register in June 2019 will cost more than registering all 17 million voters afresh. How possible? We know that the EC intends to spend as much as $70 million on a fresh registration. We know that the top-up or “limited” registrations (as opposed to full registration) of 2014 and 2016 cost about $10.6 million and $20 million respectively. So where from the argument that a limited registration would cost more than a full one?

6. All the savings numbers the EC has been throwing at the public are unreliable because they come from self-interested vendors, as already explained. We have checked what other countries have been paying for similar systems and found that all the numbers given by the EC are heavily inflated. If the EC does a proper competitive tender to just replace the components that need replacing they will end up spending just $15 million this year to refresh the system. And not the nearly $150m it will cost them to replace all the physical infrastructure and to register all eligible voters all over again. For instance, whilst the EC is proposing to spend $3000 per BVR, Zimbabwe has been able to procure virtually identical systems at $1300 just last year. In fact, as far back as 2011, Nigeria was spending $1440 per device and in 2005, the DRC was paying $2500. The story of electronics is that prices tend to fall over time.

7. Let us not forget that it is the same EC IT folks being blamed for the mess Dr Ofori-Adjei claim the existing system is currently in that are now being positioned to procure and oversee a totally new system. We are entrusting our fate into the hands of people Dr Ofori-Adjei claims couldn’t even arrange basic warranties and hoping that they can implement a brand new, superior, system, including the negotiation of new legal agreements and the fresh calibration of devices between now and April 2020.

If they have bungled the management and maintenance of the existing system so badly, why do we have confidence in them to use just 3 months to replace it and do a better job on a completely different  platform?

As they say in Nigeria, Ghanaians must “shine their eyes”.