‘Local software must demonstrate THAT they cannot be manipulated and have a proven track record’
I was very much involved with the banking industry when it moved from hand posted ledgers to basic analog accounting machines (NCR 32) which operated on gears to very basic accounts system where magnetic strip behind the ledger card was used (NCR 299), like what we have on debit/credit cards. The next stage of automation took us to an actual software that was the basis of an operating system. But we always depended on international companies for software. I don’t think that software companies existed at that time in Bangladesh so the thought of using a local software company did not cross our minds. This was, of course, in the early 80s.
I think at first we saw individuals who on a personal basis marketed software. Local companies started to develop basic software which were used in very elementary situations. In fact, some banks did use local software but those weren’t the core platform. They used software for non- core peripheral processes. The core banking software, for all banks, that had automated systems were imported. The four major government banks just recently got automated by; I think the Indian company Polaris. But until 2000 only some branches of these banks were automated with sometimes different software for each cluster. So, it was almost like four, five or six different softwares were working in tandem in one bank. The consolidation was in most cases done manually.
I think the basic issue of using local software is not competence. I am absolutely certain that the local software writers have the competence and the ability to put together a platform that will work as a core banking system. But that is not the issue at all. The issue is of track record of performance over time. Because banking requires such a digitized operating system that if you go wrong somewhere inside, unravelling it, finding out and fixing it becomes very difficult. And of course now there is the more global priority of data security and integrity. For financial institutions this is fundamental.
So, the robustness that is required, not just for a year or two, but for years, and the ability to keep a proper audit trail, let’s just say, there are things which are at risk of going wrong. These are the things that bankers worry about, and they want to know these things for sure. If you ask a banker to look at your software and explain the wonderful features it has, he will ask where else it has been used. So, what he is trying to do is mitigate his risk. His risk is that what if things go wrong. And that risk is not mitigated by doing a demonstration. The real question is how long had somebody used it without a problem and if things went wrong I want to see that you can recover or fix that. And if that gets proven I don’t see any reason why Bangladeshi developers cannot provide the CBS. So, to implement a big system software like CBS the most important thing is track record and credibility.
Another thing is that banks constantly have to work with national and international institutions and regulators. They have to be sure about the integrity of the numbers. Software produce that number. How would you demonstrate to them and create that confidence in them that what you are giving to them has integrity? If you answer my question correctly, then you have got it. But what it is that will convince regulators, your international partner or international stake holders? International regulators nowadays are very careful about ensuring institutions they engage with whether have a holistic and rigorous approach to data security or not, this means the core banking platform must have international credibility.
Provided that the integrity of the software is ensured and I think local banks should move towards much less costly local solutions. But you have to remember that these aren’t theoretical matters. There have been instances where software was manipulated. Suppose you could manipulate the system and make interest rate zero on the day when the rate will be calculated; what happens then? These are big risks.
Local software must demonstrate that their products cannot be manipulated. The integrity of the output has to be ensured. The robustness of your system has to be ensured. And this is not a one-year thing. As a buyer you have to know that there is road ahead.
There was this story of the CEO of a major US automaker, who when asked why the price of cars did not come down over years as computer prices had he reportedly said ‘do you want your car to work like a computer? Computers get stuck and frozen. Sometimes the command or click remains unresponded and you have click more than once. Do you want your car to respond like that when you step on the brake?’ That is what software should be. You have to be able to trust your software instinctively.
Someone asked me that how would I define trust. I thought about it and tried to break down ‘trust’ into a process, after which I arrived at a definition. My definition of trust is; having consecutive positive outcome from a causal relationship that allows you to take intuitive decision. When you trust something it has to work every time you ask for its service. When you get in your car, what happens? I turn on the ignition and the car comes to life and I can drive. It happens time after time without aberration. So, you end up trusting the car. If I now need to go to Mirpur I will decide without thinking that I will take my car. Why would I decide that? Because, I trust it. Why do I trust it? Because it works every time.
I attended a seminar in Sweden where I was asked to speak on why people are losing trust on long established institutions around the globe. Examples that were cited were religion, marriage, political systems and so on. Is this new situation good or bad for the future of humanity? It is in this context I evolved the definition of trust I gave above. My own view on trust and institutions is that it is an innately human nature to question and therefore challenge everything. If you cannot challenge, that means progress will stop and we will all embrace status quo and not move forward. Nothing should be completely trusted or mistrusted. You should, as a human being, have that ability to take the decision whether or not to trust something or someone. The entire concept of human progress relies on this. So, trust is important. Otherwise societies do not evolve. This is the essence of democracy.