What do the words “real-time” mean? A century ago, Einstein explained that the time something occurs is not independent of the observer’s point of view. Some people use the words “real-time” as an abbreviation; they mean one thing happens at the same time as something else. That shorthand can be misleading. Einstein theorized about the speed of light, but electrons are much slower; even the quickest computers take time to process 1’s and 0’s. An automated system that purports to work in real-time is delivering its results soon after something else. What makes this time ‘real’ is merely that the delay is small enough to be considered negligible, though any decent telecoms engineer should realize that just because a delay is small per the expectations of an unscientific mind does not make it negligible relative to the task being performed. However, the worst sloppiness surrounding the use of “real-time” is not because of imprecision about when events occur, but about what those events are. People talk about the need for real-time systems without being clear about what those systems are supposedly doing in real-time. And that is a long but necessary preface to a simple analysis of the current mess occurring in Ghana, where politicians keep insisting they need a real-time system to audit the revenues of telcos, and believe it is good value to spend USD1.5mn per month to have this system. They are wrong. Ghana’s government is wasting a shocking amount of public money, and their repeated use of the words “real-time” as justification reveals they do not understand the workings of the system they are buying, nor what they hope to achieve by purchasing it.
With that in mind, let me give three simple explanations for why Ghana definitely does not need to spend USD1.5mn per month on real-time assurance of carrier transactions, and one worrying reason why it might anyway.
1. Telco Revenues Can Be Audited Much More Cheaply
A recurring theme in the hullabaloo in Ghana is that telcos may not pay all the tax they owe, and anyone opposing the implementation of a real-time system is either (1) wanting to help telcos to cheat on their taxes or (2) naive about the dangers of using a system that does not work in real-time. For the sake of argument we can avoid taking a stance on the first point, except to observe that wanting to prevent telcos from cheating their taxes does not mean you should waste money on unnecessary tests, any more than a doctor should waste money on x-raying a patient’s foot when the patient is complaining he has a runny nose. Regarding the second point, this claim that the system must work in real-time indicates that Ghana’s government suffers a fundamental misunderstanding of what it means to perform an audit in real-time. On one hand, all audits rely on ‘real-time systems’, because auditors rely on evidence of transactions created at the same time that the transaction is made. On the other hand, nobody is going to audit telco taxes in real-time because no business pays their taxes in real-time. They pay their taxes later. So “real-time” is being used to talk about information gathering, when all audit work always involves “real-time” information gathering.
Without bothering to detail what the real-time system will actually monitor, the politicians would contrast this system with reliance on CDRs, asserting that CDRs may be corrupted. They ignore that telco CDRs are no less produced in ‘real-time’ than any other data that may be produced by any system tasked with recording which network events occurred. They are correct that CDRs may be corrupted, manipulated and so forth. However, that same risk applies to any system. Systems are no more guaranteed to be accurate just because they work in real-time. Again, this just reveals a fuzzy understanding of what is supposed to occur in real-time.
But the most important point is that auditors, everywhere, already know of a perfectly good approach to capturing audit data about a business’ transactions, and this approach is much cheaper than the system Ghana wants for its telecom RA audit. Let me give three examples of the cheap, effective and universal approach to auditing business transactions.
- Buy a box of cornflakes. When you buy it, look at the weight written on the side of the box. Carry the cornflakes to some scales, without spilling any cornflakes. Weigh the cornflakes. Compare the actual weight of the cornflakes to the amount written on the box.
- Look at the water meter, and record the number on the meter. Run the water tap. Pour the water into a measuring jug. Look at the water meter again, and record the new number on the meter. Compare the amount of water in the jug to the readings taken from the water meter.
- Make a phone call. Write down when the call starts and ends. Obtain the telco’s CDRs. Look for a CDR that says a phone call was made using the phone at the time the call was made. Compare the duration per the CDR to the times written down.
Simple, simple, simple. Auditing can be both simple and effective. Why should we complicate the audit of telecoms revenues when they can be audited exactly the same way as anything else? The most common reason is that too many people want to sell fancy systems that are expensive, complicated, but excessive.
Too many so-called experts bring disgrace to our line of work by pretending the method just described is somehow inadequate. Is it not real-time? Of course it is real-time! You note what call is made at the time you are making it. But what about the risk of missing the cheating occurring somewhere else? If you are worried about that risk then keep testing. There is no magic to mathematics. If you make 100 test calls and find no issues then the likely error rate is less than 1 in 100. If you make 1,000 test calls and find no issues then the likely error rate is less than 1 in 1,000. You could make a stupidly large number of test calls without spending as much as USD1.5mn per month. But any sensible auditor would stop well before that point, because if they never found any issues, then they would conclude it is not worth wasting money on more testing.
2. Ghana’s Obsession with Auditing Telco Revenues Stems from a Bad Decision that Hurts the Economy
Ghana is one of a number of African countries that has adhered to the WTO Reference Paper on Telecommunications. Nevertheless, in June 2010, a surcharge on international inbound traffic that had been legislated in December 2009 came into effect, raising the termination rate from an average of USD 0.11-0.13 to USD 0.19, meaning an increase of the termination charge of 46-73%. At the same time, traffic decreased 48% from 2009-2011 (FCC data and USTR 2013). Out of the new termination rate (USD 0.19), the Ghanaian Government takes USD 0.06 and operators get the remaining USD 0.13.
Data from the FCC reveals that the average payment per minute (i.e. proxy for termination charge) of a call originating in Ghana and terminating in the United States was USD 0.02 in 2011. In the opposite direction, the average payment per minute that United States carriers paid Ghanaian carriers was around USD 0.16. An increase in the termination rate can be expected to decrease incoming traffic if it substantially increases retail rates. For instance, the Office of the United States Trade Representative (USTR) in its 2013 report states that “according to FCC data, in 2009, the United States sent over 300 million minutes of traffic to Ghana. In 2011, however, the number of minutes was less than 170 million, a decline of over 48 percent.” Additionally, this report noted the concern “about Ghana’s obligations under the GATS Annex on Telecommunications and GATS Reference Paper.” The USTR stresses that it “will continue to engage with Ghana to seek removal of the mandated rate increase.”
Those words are included in a report written for the OECD, the international body that reviews world trade. What they were telling Ghana is so simple that even a schoolchild can understand it: if you put prices up, then you sell less. The government of Ghana is obsessed with auditing telecoms taxes because it is trying to squeeze the maximum amount of money from telecom services, but its efforts are counterproductive.
One result of artificially high prices is increased fraud, because neither Ghanians nor anyone else is willing to pay much higher prices just because of government taxes. Fraud reduces the amount of tax collected, which is why the Ghanian government is obsessed with fighting it, partly through the detailed audit of telcos.
The other result of high Ghanian prices is the overall damage to the economy, as Ghanians are less likely to do business, and Ghanians overseas are less likely to call home. That damage to the economy also reduces the amount of tax collected, though the connection is harder to measure. Nevertheless, many African governments have realized that high taxes do more harm than good. They would rather stimulate the overall economy. A stronger economy yields stronger tax returns, whilst Ghana is stifling economic growth by over-taxing an enabler of growth.
3. Voice Revenues Are Dying Anyway
The internet is killing voice revenues, and so it will also kill taxes imposed on voice services. Increasing prices will just hasten the transition to OTT services that can escape the taxes of the Ghanian government. Put into this context, trying to make money from an expensive multi-year audit contract is just a reckless gamble anyway. Ghana’s government does not know how much revenue it will gain from its expensive audit, nor can they predict how rapidly those tax revenues will decline in the next few years, irrespective of their audit.
But there is one counter-argument. Ghana’s government keep insisting the new audit represents good value because it will also cover mobile money transactions. Mobile money revenues will grow, instead of being subject to the same decline as is occurring to voice. Ghana’s government is right to expect mobile money transactions will represent a significant and growing proportion of the total economy. However, if over-taxing voice calls is damaging to economic growth, imagine the danger of over-taxing every transaction that occurs. On one level it may seem like a dream way for governments to make money. On the other hand, a system with the potential to take a cut of every purchase is a potential nightmare for taxpayers. At the very least, Ghana’s government should be honest if this is their intention. However, Ghana’s citizens would be safer if their phones were not being turned into spies that report on every penny they spend.
The revenue assurance discipline still lacks the maturity to take moral responsibility for the often shockingly bad advice given by so-called ‘experts’. Many will look at Ghana’s example, and salivate at the prospect of systems where taxpayers will be forced to pay USD1.5mn per month in exchange for some questionable benefits. We should be concerned that such systems could enable the worst kind of monopolistic rent-seeking behavior, where ‘assurance’ just means a guarantee that we can bleed ordinary people at every opportunity.
Ghana’s taxpayers are being told it is in their interests to spend USD1.5mn per month on a system that, in all likelihood, will find no under-declaration of telco revenues, just as previous audit systems failed to deliver promised riches. However, Ghana’s citizens may also be paying for a system that could soon be used to increase the taxes that every Ghanian pays. Previously the focus was on taxing inbound international calls; Ghanians may not have cared about these taxes because they felt that the taxes are only paid by foreigners, though Ghanians living abroad would be paying them too. The new audit system will presumably encourage taxes to be levied on transactions that take place wholly within Ghana. That is a bad deal for them, and once the yoke has been imposed via every phone in every pocket, they may never be able to remove it from their shoulders.