My friend and colleague, Selorm Branttie, forwarded to me snippets of a Facebook thread about my recent comment on the Year of Return, in which he had been tagged and expressly asked to forward to my attention.

My inclination was to ignore the thread, for obvious reasons.

First, it is Christmas for Christ’s sake! Second, I have had to switch off Facebook because it was taking too much of my time. Facebook is designed to compel conversation. I simply don’t have hours in the day to spare for casual banter anymore; the daily hustle has made sure of that. Twitter, which I have been using more and more, is, on the other hand, a broadcast medium that allows me to disseminate random thoughts with far less expenditure of time. Lastly, I recognised the key discussants on the thread as strong partisans of the ruling party in Ghana with whom I share a bit of a history.

In the past it has been virtually impossible to have a sincere debate about anything since they seem sworn to see any dissent from positions of the ruling NPP government as motivated by “cynicism” and “ill will”. In my brief and occasional dealings with such people, sincere and mutually respectful discussion of facts and figures consistently degenerated into motive-questioning and other such base exchanges.

Frankly, I have neither the time nor ability to understand what it is that make some people so loyal to political parties to the point of losing all capacity for objectivity, but, well, we live in a complex world.

I have decided to respond to the Facebook thread for only one reason: a debate over data is always worth having in the peculiar circumstances of Ghana’s development.

Having chosen to mount what amounts to a mini-campaign about the issue of the Government’s reliance on bad data to appraise the generally successful Year of Return initiative, I feel obliged to address reactions, even if they are accompanied by insults accusing me of “cynicism” and statistical ignorance, so long as they present themselves as offering counter evidence.

The “counter evidence” presented in the said Facebook thread was along these lines:

1. In the thread, I was accused of presenting data covering all “international arrivals” instead of just “tourist arrivals”.
2. My data is thus, according to this framing, a superset. The more relevant subset of tourist arrivals is much smaller. The data that I had culled from the CEIC and IATA datasets, and which was corroborated by Ghanaian Civic Aviation Authorities disclosures (some of which are openly available on the Ghana Airports Company website) could thus, in the view of my Facebook accusers, not be relied upon since not all international visitors are tourists.
3. If the smaller tourist arrivals dataset is used, the claims by Government functionaries that $1.9 billion has been generated from the Year of Return would be justified.

This is the coherent part of the argument. How exactly the exponent of this theory jumps from these premises to satisfy the requirement of showing that $1.9 billion has been generated by the Year of Return initiative is both murky and confused.

Here is an extract from the thread so that the reader can judge for herself:

At the beginning of 2019, the Ghana Tourist Authority (GTA) projected that they were expecting 150, 000 more tourists from the African diaspora, thus making 2019 tourist number 500,000 compared to 350,000 of 2018. They estimated total spending by tourists to be about $925m.

Then in September 2019, with arrival data, the actual number was calculated at over 750,000, surpassing the 500,000. Based on the rate of arrivals, and the peaking of activities in December, they estimated 1 million tourists by year end. Note that due to the doubling of the numbers, the expected spending by tourists, also doubled to $1.9bn.

Recall that my argument was very simple:

1. In 2018, international arrivals were about 984,000.
2. In 2019, the most optimistic projection suggests 1 million arrivals.
3. The most optimistic allotment of any increase in arrivals due to the Year of Return cannot thus exceed 15,000 extra visitors in 2019.
4. Because we have a reasonably strong estimate of average spending per visitor (about $1800 in 2018), we can confidently project an additional measurable revenue gain of $30 million if we use average spending of $2000 per visitor. This of course does not account for any qualitative gains, which are almost impossible to measure in this case due to paucity of data.
5. Based on both the mean rate of arrivals for the last decade and the mean rate of projected arrivals for the decade after 2017, we have no evidence that the Year of Return has boosted arrival numbers beyond anything other than the mere increase in arrivals in 2019 over the figure recorded in 2018. And even that boost is only to the extent that the growth rate exceeds the average growth rate in visits over the last couple of years, which, sadly, it does not.

The counter-argument being canvassed in the Facebook thread is that we need to use figures preferred by the Ghana Tourism authorities and that if we do we shall record 650,000 extra or additional visitors in 2019 over the 2018 figure. Presumably, the interlocutor has no objection to the use of a spend-per-tourist figure closer to the $2000 number, which should give us $1.3 billion extra yield supposedly attributable to the Year of Return. Still, no $1.9 billion in sight though.

Except that all of this seeming “analysis” is actually empty of both research value and statistical fidelity. The process of determining whether an intervention has had an “effect” is a very elementary one, taught to everyone who has ever taken the most elementary of introductory courses in statistics. It is standard “hypothesis testing”. In a situation such as this one where the means and variances are all so well known due to near-complete historical data, there is hardly any need for the kind of quibbling seen on that thread. Which brings us to the central point: data integrity and validity.

If indeed the tourist subset is about a third of the set of total arrivals, such that we had 350,000 tourists out of the 984,000 arrivals in 2018, then one can only expect 2019 to record total arrivals in excess of 3 million for the tourist component alone to hit the 1 million mark, as canvassed by our Facebook commentator. A completely absurd projection that is at variance with several Government projections.

In fact, there is no logic in holding that the arrivals data I used is the superset and that the Tourism authorities in Ghana have higher-resolution subset data that can be used to hone in on tourist numbers with better precision. Simply because such a supposition was both superfluous and easily verified to be false.

Here is the Tourism Ministry’s own set of data available from page 22 of its latest strategic plan:

This data is actually lower resolution, and includes land border arrivals that are very difficult to process due to the high translocalism of intra-African borders. My decision to focus on data that can be corroborated and triangulated using international aviation stats was driven by this very fact. That data pays as much attention as possible to pruning non-visitor noise wherever possible.

But all that is besides the point really. The most important point here is that this data does not in anyway vindicate any of the strange and statistically perverse claims made by our interlocutor on the Facebook thread. The Ministry’s tourist arrivals figure for 2016, the most recent year for which it had an actual tally is 1,202,200. Its projected figure for 2018 is 1,454,700.

Average visitor spend, according to the Ministry, for 2016, the latest year for which it had complete data, was $1,890, virtually identical to the Ministry of Finance estimate of $1,800 for 2018.

It is important to bear in mind that this average spending amount is comprehensive. To focus solely on spending on tourism activities, strictly speaking, would be to considerably shrink tourism receipt numbers since direct spending on pure tourism in Ghana is incredibly low. According to the Ghana Statistical Service’s “Trends in the Tourism Market in Ghana 2005 – 2014” report (page 15), visitor spending adjusted for inflation has actually been falling for some time:

Visitor arrivals to selected major tourist sites rose from 381,600 to 592,300 over the decade. Real revenue also rose marginally from GHȼ490,000 to GHȼ492,000 over the same period with average spend per arrival falling from 1.3 GH¢ to 0.8 GH¢. Both arrivals and real revenue grew positively in the first half of the period, at an annual average growth rate of 10.7% and 6.1% respectively, while during the second half, both categories fell every year on average by -1% and -3.3% respectively.

We are seriously talking about total earnings in the hundreds of thousands of dollars for the whole country here, and average spending in tourist sites of about a dollar and cents over the course of a full decade. One has to employ a broader definition of visitor spending for tourism revenues to make any sort of statistical sense, which points to the absurdity of trying to use a lower-resolution spec for defining “tourist arrivals”.

So, where did our commentator on Facebook find his numbers? Surely not from the Tourism Authorities, despite purporting to be filtering out tourism data. How did he construct his hypothesis that the “tourist arrivals” number used by the authorities is lower than the “international arrivals” number widely used by global observers and by the Ministry of Finance to compute tourism revenues in Ghana? Surely not from any reputable or defensible data source.

But even if we are to use the lower-resolution, looser, dataset employed by the Tourism Ministry for strategic planning, the fact remains that there is no mechanism to “extract” this strange $1.9 billion figure from any statistical manipulation known to the science.

For instance, we can use the 1.45 million arrivals projection for 2018 (as made by the Tourism Ministry in 2017 in the table reproduced above) and also use the most aggressive projection for arrivals in 2019, which is 1.5 million, and we will still only be able to come up with $100 million as the level of tourist receipts in 2019 in excess of the figure recorded in 2018. But we would then need to address the hurdle of this “increase” being lower than previous year-on-year increases. For example, the last three years has seen arrivals growth rates in the magnitude of about 10% per annum. The apparent increase in numbers in 2019 compared to 2018,  would however, if we are to benchmark against the Ministry’s data, be less than 7%, suggesting lower growth in 2019 compared to the trend over the last couple of years.

In short, rather than contribute to the debate in a sincere effort of enhancing policymaking in Ghana through a promotion of facts, figures, data and analysis, the Facebook thread I referred to earlier was primarily concerned with pushing a partisan agenda, and a partisan agenda alone.

From what I can see on social media, many Ghanaians are now completely tired of this inability of politically engaged people to rise above petty partisan squabbling. I share in this frustration.

Ghanaian officials are basking in glory and ecstasy following a very successful marketing and branding campaign for the “Year of Return”, an initiative of the President of Ghana marking a major timeline in the sordid history of black slavery.

So far, impressive feats of national showcasing are there for all to see: tons of positive international press, great mentions and fabulous celebrity endorsements, most of it at no cost whatsoever to Ghana.

So why is Ghanaian officialdom so keen on selling the success of the initiative on the tabletop of incoherent statistics and woolly numbers instead of better cataloguing these clear achievements? It is a very strange sight to behold. Is this sad spate of fuzzy arithmetic just another example of how as a country Ghana struggles to master data-driven policymaking or is this an isolated case of mere overexuberance?

I know that no malice is intended, but before I am pummeled to pulp for being a killjoy, let me hasten to point out that sound data is important for drawing accurate inferences.

Unfortunately, various government agencies and supporters of the Year of Return program have bandied figures such as “200,000” extra arrivals, “1.5 million” total visitors and “$1.9 billion” extra tourist spending as measurements of outcomes related to the Year of Return with zero commitment to using actual, widely available, statistical data.

This means that instead of focusing on what so spectacularly went well – the brilliant coopting of African American celebrities like Steve Harvey as informal brand ambassadors – we shall soon be luxuriating in fictitious numbers bearing no resemblance to reality.

Here is the data we do have. As at last count, 750,000 international visitors had made their way to Ghana in 2019. (https://mobile.ghanaweb.com/GhanaHomePage/NewsArchive/Year-of-Return-Ghana-Tourism-Authority-reveals-the-number-of-diasporan-visitors-815566) The Authorities are projecting total arrivals for the year to hit 1 million. This is however doubtful considering the proximity to year-end.

But even if the numbers do hit 1 million, that would only mean a tiny fraction more than the 984,250 visitors who showed up in 2018, in fact a mere 15,000 more.

According to Ministry of Finance computations, average spending per tourist was $1512 in 2014, rising to roughly $1800 in 2018. Let’s pad this to $2000; though with Cedi exchange rate depreciation outstripping inflation, foreigners should actually find Ghana about 5% cheaper than last year and might spend less. Be that as it may, the “extra spending” that could conceivably be attributed to an increase in arrivals due to the Year of Return (if the projected 1 million visitors estimate holds up) would amount to about $30 million in this scenario.

By what conceivable mechanism can a $30 million optimistic projection mutate into $1.9 billion?

As already hinted, growth in tourist numbers in 2019 may well be below the average 3% per year rate seen over the last couple of years (and certainly below the 5.1% annual growth rate trend experts have projected between 2017 and 2027). There is further grounding for such speculation in the author’s estimate of a rise in hotel and short-stay apartment rooms inventory and a fall in average room rates based on an analysis of several weeks of Booking.com (a major travel site) data.

But all this should really be beside the point since no data-conscious person would insist that every successful marketing exercise must necessarily bear fruit even whilst it was still underway. There is almost always a timelag before results materialise. The danger in elevating phantom figures to the level of truth is in the complacency they can breed. So that instead of girding our loins to build on this successful marketing exercise and translate the increased awareness about Ghana and its enduring international goodwill into tangible tourism gains, we would instead declare victory on all fronts, relying on shaky, unchallenged, numbers and then promptly relapse into business as usual.

The only reason, therefore, for sounding the alarm about these widely publicized and widely believed numbers is the wish to forestall such a bad outcome and to motivate the authorities to see their successful marketing and communications strategies as merely the foundation on which to erect a truly effective sales plan for Ghana’s tourism and investment climate potential.

And we absolutely have their back.

 

 

This week, some telecom operators came under consumer fire for sudden, unannounced, hikes in their data tariffs, prompting parliamentary intervention. At least one of the telcos involved has blamed the incidents on system glitches.

As a longtime consumer rights activist, naturally my sympathies are with ordinary mobile subscribers who have had to endure the loss of value through no fault of theirs.
Ghana’s regulatory agency, the National Communications Authority (NCA), has been swift to order the telecom operators to fix their issues and compensate consumers. All well and good, but are we going to see similar speed in addressing the value for money problems that have been created by extractive and exploitative regulations over the course of the last decade?

You don’t have to take my word for it. In May 2018, the sitting Minister of Communications went to Parliament and boldly announced to a stunned chamber that according to the records available to her all the money paid to various companies in the name of revenue assurance, revenue protection, revenue optimisation etc, over the last decade – some $200 million thereabouts per my rough calculations – is good as wasted since the only thing these companies have been doing is taking data from the telcos and passing on to government in exchange for fat cheques.

According to the Minister, in the absence of what she calls “real-time monitoring”, all the previous efforts could not have detected any malfeasance on the part of the telcos and were thus unnecessary. Having served us this shocking confession from the government system, the Minister then briskly went on to justify even more spending in furtherance of similar money-wasting initiatives, such as the ongoing Kelni GVG project and its claims of using telecom traffic measurement to more accurately estimate telco revenue and thereby uncover tax fraud.

In a recent appearance on TV3’s Hot Issues, hosted by Stephen Anti, I could hardly hide my angst when Stephen, in the service of balance and objectivity, read on air a portion of a recent statement from the Minister purporting to justify the Kelni GVG contract.

Let me summarise for readers in the briefest way possible the comments I made on Hot Issues about why the Minister’s statement is dangerously wrong and ought to be retracted.

1. Telecom Signals & Traffic Volume Estimation Cannot Accurately Determine Revenue and so Cannot Uncover Tax Fraud

The genesis of all this “telecom tax revenue monitoring” business was a decision in 2008 to set a “floor price” for international inbound calls of $0.19 of which roughly 32% or $0.06 was to be paid by the telcos as a tariff to the Government (attempts to charge tax on the remaining 13 cents as “Communication Service Tax” was however defeated in the courts in 2013).

By imposing price control on this segment of the industry, a huge artificial arbitrage opportunity opened up for crooks to bypass the official channels and terminate international calls over the internet on local SIM box installations for as low as $0.04. This meant they could charge unscrupulous carriers a much lower price than the official minimum price of 19 cents and still pocket a substantial markup. “SIMbox fraud” had been born.

To fight this problem, needless as you would recall in the first place, the government engaged GVG, a Haitian company to provide a solution to monitor all inbound calls and to use random pinging methods and geolocation to identify SIMbox installations and take them down.

Whilst the problem itself was self-inflicted, there is a faint logic to the idea of monitoring call volumes to ensure that tax evasion is minimised. Given a fixed rate and charge and a fixed amount of tax to be collected per call, volume estimation becomes a sensible way of benchmarking revenue.

Unfortunately, when the decision was taken to extend revenue monitoring to domestic call traffic, as part of a wider CST imposition, and Subah came into the picture in 2012, common sense went out of the window and the principle of using volume benchmarking remained even though there is no price control in the domestic call market, and rates and charges vary considerably from period to period.

Since CST is based on a percentage of the charge value (“ad valorem” in the jargon) the revenue generated from it is tied to how the telcos “rate” (or price) each call or data session, making a linear map between call volumes and total revenues, much less tax take, completely arbitrary.

Unless the Government imposes its own pricing regime on the industry and substitutes its rating software for that of the telcos, any monitoring shall still be based on rating data supplied after the fact by the telcos, making the idea of monitoring call volumes “in real time” as a way of benchmarking revenue completely flawed and illogical.

2. But Even if Volume Fraud is an Issue, There is a Cheaper Way to Detect it

But even if, contrary to all sound technical analysis, the government still wants to find out if the telcos are lying about the volume of traffic on their networks, there is a cheaper way to go about it: using a method similar to the “random pinging” approach currently used to fight international inbound call fraud.

In a brief essay written during the heat of the initial Kelni GVG controversy, I demonstrated how with less than $100,000 a solution can be built to initiate thousands of random calls to the mobile networks so that the call records provided by the telcos can be examined for deliberate or accidental omission of these test calls. The percentage of missing test calls shall map directly to the likely percentage of undisclosed calls in the wider sample.

No one has been able to show why this is not a vastly superior solution costwise to the $1.5 million a month ($178 million over 2 x 5 year terms) Kelni GVG monitoring contract. This is a solution that can cover all the purported four services provided by Kelni GVG at less than 5% of the total cost of the Kelni contract over its life cycle.

3. Let’s Be Clear: this Contract Implies No Savings on Previous Ones

One of the most earnest cases made by the Minister of Communications in her public statement of last month is the claim that the extant contract with Kelni GVG represents an improvement on the Afriwave and Subah ones.

In the Minister’s analysis, the Kelni GVG contract has saved the country 1.1 Million USD per month by bundling the Afriwave and Subah contracts, costing a total of $2.6 Million a month, into one contract: the Kelni GVG contract costing $1.5 Million a month.

This is plain obfuscation.

First, the Afriwave contract is very much with us, contrary to plain untruthsserved to us by officialdom. I have confirmed personally its existence and continued operation.

Considering that the interconnect clearing house (ICH) concept was justified on the same premise of blocking telco underdeclarations, one would have thought that the Minister would actually make the ICH project redundant as promised. But, no, like so much of the posturing of politicians in Ghana, it was was all mere theatrics. Since early 2019, the nearly moribund ICH has been receiving a major lease of life because of Huawei supplier credit arrangements and a significant proportion, by some estimates 25%, of call traffic now traverses the ICH. Afriwave is thus getting paid at least the $1 Million a month it last negotiated during the Subah-NCA impasse.

In effect, there has been no $1.1 Million savings. Even worse, the Government had to buy off Subah in order to bring in Kelni GVG. By our reckoning the legal costs of doing so amounted to at least $40 Million. So, rather than saving the country $1.1 Million a month, the Kelni GVG decision is costing the country (using crude amortisation over the period of the Kelni GVG contract) an extra $233,000 a year compared to the Subah-Afriwave combo.

4. No, There is No Evidence That Kelni GVG Has Detected Any Tax Fraud

The reader might be of the stock of pragmatists who believe that notwithstanding all the back and forth debate, the proof of the pudding is in the eating: has the Kelni GVG program more than a year on uncovered any major tax fraud or not?

The Minister says, yes: to the tune of $85 Million over a period of three years examined. She goes on to claim that a total of $250 million would have been saved by the end of the contract with Kelni GVG. How did she come by these numbers? She doesn’t say. Why hasn’t any telecom executive been arrested and why is no telecom company under tax fraud investigation? She doesn’t know. Has she filed a formal complaint to the Police and the GRA? No.

Some readers may recall that in 2018, as part of the effort to mobilise public support for the Kelni GVG deal, the Deputy Minister of Communications said that as a result of the real-time monitoring regime, revenue from CST shall jump to $200 million a year.

If any of what the two Ministers are saying is correct then all we need to do is chart the growth rate of CST revenue from 2008 when it was introduced (annualising the 2008 half-year outturn) to end of 2018, when according to the Minister “real-time monitoring” commenced, and then compare with the revenue recorded at end of 2019 to see if there has been a sudden jump in revenue as a result of the introduction of said “real-time monitoring”.  Seeing as the entire logic of the need to spend nearly $180 million on this endeavour is premised on the notion that only real-time monitoring can safeguard revenues we ought to see a sharp upward turn in the curve from 2018 onwards.

But what do we see? Nothing. All the evidence shows that the growth rate of CST revenue only tracks such macrostructural trends as inflation, exchange rate movements, GDP growth, subscriber growth, and growth in the CST tax rate or coverage (as in when the rate was jacked up by as much as 50% in late 2019 or when the government insisted that domestic interconnect charges should be subject to CST only to be rebuffed by the Courts a couple of years ago).

Real-time traffic monitoring has had no discernible statistical effect whatsoever on the growth rate of CST revenue and any multi-variable regression analysis should bear this out. In fact, considering the Q3 2019 increase in rates by as much as 50%, the current trend is one of underperformance, as the Minister of Finance freely admitted to Parliament during his budget speech.

So, in short?

Simple: every month that the government continues to hand over our scarce resources to Kelni GVG and tries to address clear shortfalls in CST revenue. Instead of the $200 million promised by the Deputy Minister, the country is struggling to make even $80 Million from CST this year.

Jacking up CST rates, as the Government has been doing recently however, is clearly and painfully contributing directly to the falling value of data and voice vouchers that consumers all over Ghana keep lamenting about.