Ghanaians will wake up tomorrow to a new era where the monies they transfer online and on mobile will attract a tax of 1.5% once they exceed a daily cap of GHS 100 ($13) in cumulative payments.

Because of widespread disaffection about the way the government has gone about imposing this tax, riding roughshod over many important stakeholders to bulldoze it through, it has had to make last-minute concessions in an attempt to blunt some of the anger. These concessions were mainly in the form of “exemptions”.

Payments made to a range of recipients, for a variety of purposes, and to a sender’s own accounts and/or wallets, regardless of where they are held, were exempted from the e-levy.

To enable these cross-network, industry-wide, exemption rules and processes to take effect, there was a need for some kind of central catalog of APIs and services for the nearly 400 financial services operators allowed to charge the tax to uniformly adopt for consistency.

The industry proposed a clear set of specifications that they could implement on their own and suggested the creation of a set of APIs operated by a national payments switch, such as the GhIPPS. Other commentators warned about the complexity of developing cross-network, industry-wide, charging rules to deliver fully on the exemptions. Characteristic of the government’s current attitude to stakeholders, all these warnings and admonitions were ignored.

The Ghana Revenue Authority (GRA) insisted that it had developed a solution called ELMAS that could centrally manage all these ad hoc and constantly evolving rules (example: a cumulative exemption cap of GHS 20,000 was abruptly added for corporate banking transactions). It strenuously denied that any private contractors were involved at state cost.

It was eventually discovered that not only was the ELMAS shrouded in opacity, but it was also steeped in confusion and lack of candour. Industry players insist that they detected the involvement of shady and conflicted private contractors behind the GRA front presented throughout the attempts to build the industry-wide charging and exemptions framework. “Conflicted” in the sense that the ELMAS was also pitched as a monitoring system. How then can contractors known to be active players in the industry be brought in stealthily as referees and traffic conductors were it not for cronyism and other self-serving interests?

Whatever be the motivation of the government agencies, the poor design of the whole e-levy; refusal to heed industry advice and the inputs of well-meaning critics; the unrealistic timelines; and the failure to create a truly above-board and professional framework, have all led to a situation where several critical exemptions shall not work smoothly come tomorrow.

Instead of an industry-wide framework, individual financial services operators will use their internal billing systems to apply the tax and any applicable exemptions only within the confines of their own network. Cross-operator exemptions shall be constrained.

The government’s bid to prevent mass defections from electronic payments and a reversion to cash, which would undermine both the revenue objectives of the tax and Ghana’s budding digital economy, is now under threat.

Such a sad spectacle would have been averted if good faith consultations with all relevant stakeholders had been undertaken. Instead, opacity, favouritism, aloofness and lack of candour and transparency, have once again been allowed to derail the reputation of Ghanaian government agencies for professional and effective policy design and implementation.

                                                                                                               

Every other day, one business contact or the other tells me that they don’t use WhatsApp for sensitive discussions anymore, so to continue the conversation I need to move to Signal. Since I don’t have a big problem with such requests, I haven’t bothered to reflect too much on the trend. Until this afternoon, when another request triggered me a bit, and I thought, heck, let me just write a little piece on the subject.

It is easy for folks to confuse privacy, data protection, and secrecy/anonymity when it comes to modern digital technology. All these concepts are made all the more complicated by having multiple layers of legal and technical complexity.

One simple way to get a handle on the intricacies is to be clear about the “enemy” one is seeking said privacy, protection or secrecy from or against. Is it 1) friends and neighbors, 2) the general public, 3) the government, 4) the technology developer or 5) everyone?

I will tell you upfront: privacy from everyone is simply not a realistic option if one also intends to use digital technology. These is always someone else in the mix. Even apps like Threema that try to up the privacy stakes by emphasizing anonymity cannot be completely invisible to the telecom and datacenter networks through which users access them. So, let’s discard the “everyone” case without further probing. That leaves us with four main classes of “others” against whom one may reasonably seek to assert one’s wish for privacy or secrecy.

In the messaging context, however, there is a presumed openness to friends and others with whom one is engaged in conversation. We can take that category off the list too.

The filtration done so far leaves the following “enemies” on the privacy list: the general public, the government and the technology developer.

All three apps – WhatsApp, Telegram and Signal – are sufficiently well built and maintained enough to do a reasonably good job of making sure your chats go to only the people who they were meant for. Any unintended disclosure is more likely to be due to factors, such as human carelessness, rather than the quality of the specific app. The General Public criterion is thus also easily dismissed as a serious basis of comparison.

So, finally, we get to the meat of the matter. On the few times I have tried to understand the basis of the shift from WhatsApp to Signal or Telegram, a relatively small number of respondents have cited concerns about WhatsApp’s privacy policy and the fact that Facebook’s servers are used for the transport and storage of the data. That is to say, the main concern of this class of users is the conduct of the technology developer.

Facebook’s unpalatable reputation for commercial exploitation of data without due regard to user sensitivities clearly worries some African business folk in my circles. Signal, being a privacy-obsessed non-profit, clearly wins on that score against both WhatsApp and Telegram on the Technology Provider dimension.

The vast majority of my Africa-resident business friends who ask that we shift conversations to Signal are, however, concerned primarily about government intrusion. Especially also because once the government is at the tail-end of the surveillance chain, the risks on all other dimensions multiply. For instance, a Google vendor, Mitto, was found spying for governments without the knowledge of even senior employees, a growing theme in this murky world.

Businesspeople in Africa feel under siege from shadow states, plain extortion, ruling party paranoia about their funding the opposition, and competitors with links to the intelligence services. I find the anxiety about government eavesdropping strongest in East and Southern Africa, though some pretty hairy stories have been heard in Nigeria too. Interestingly, it is precisely in the context of privacy and secrecy where government is concerned that public misconceptions abound.

For instance, there is barely any logic in moving from WhatsApp to Telegram on privacy grounds linked to malicious government or organized hacker activities. Telegram uses an opt-in (non-default) encryption model for message traffic that it refuses to disclose for independent security analysis. Determined security researchers have shown nevertheless that its cybersecurity standards are somewhat looser.

Regarding WhatsApp versus Signal, the analysis is more nuanced and also more interesting. Some users may not even be aware that both apps actually use the same open-source encryption system: the Signal Protocol, which enables end-to-end encryption and perfect forward secrecy, and thus disguises the message from non-senders and non-recipients. The two companies are located less than 30 minutes from each other in the San Francisco Bay in Northern California.

In short, both Signal and WhatsApp are within the legal jurisdiction of the American government and have similar technology philosophies. In fact, the main early financier and co-Founder of WhatsApp is the current co-leader of Signal’s owner entity, and its interim CEO.

Some claims are usually made for Signal’s approach to end-to-end encryption and its implementation of the Signal Protocol for metadata protection (hiding not just the message content but also its critical characteristics like origin, destination and timing). Some argue that as a non-profit it is somewhat less amenable to American government pressure to insert backdoors or to deliberately weaken encryption in the name of national security or law enforcement.

WhatsApp and Signal head-to-head on security. Source: Mehak (2021)

Signal enthusiasts would normally frame such distinctions as done in the above table. How each of those supposed strong suits provide protection against determined US government intrusion is highly debatable. And there is already a growing citizen movement against encryption because of things like child trafficking that are changing the terms of the debate. Even as the US Government surreptitiously buys up and hold stakes in the encryption companies themselves.

But we need not dwell too much on the details here since few Africans in the category I am discussing are worried about US government surveillance or law enforcement overreach particularly. The overwhelming majority care more about surveillance by their own governments in Africa.

There is no evidence to show that WhatsApp will be more submissive to an African government’s request for backdoors than Signal. WhatsApp has put up a fairly valiant resistance in India to government demands. The economic case for capitulation is obviously stronger in India than in Africa. The case may nevertheless be made that Signal’s small size and lack of a Facebook-like global footprint should make it more impervious. But a counterpoint can work in Facebook’s favour: its vast resources can help it implement more complex legal and political shields in places like Africa.

At any rate, an African government interested in surveillance is less likely to proceed like the US government, India or China by seeking to enter into elaborate arrangements with tech giants for backdoor design and implementation. Most African governments simply lack the technical capacity to design those kinds of regimes. They are more likely to invest in cyber-offence tools and contractors, as some of them have done already. Tellingly, Bulgaria-based Circles, a spyware vendor steadily overtaking NSO in notoriety, has a third of its government clients based in Africa.

Advanced professional hacking tools and services from the likes of Israel’s NSO and the Anglo-German Gamma Group have also been traced to African surveillance operations. In fact, it is widely believed that Uganda’s attempts to hack the Apple phones of US diplomats in Kampala using NSO’s Pegasus are what caused the abrupt switch of US posture towards NSO from tolerance to hostility.

When it comes to tools such as Pegasus, the target is the phone’s actual operating system. Any malware that takes sufficient control over the operating system of a device could also steal the private keys downloaded from the platform and render any encryption vulnerable.

We know from the Jarett Crisler case that Signal message content and metadata can be extracted by US law enforcement agencies most likely through exfiltration of encryption keys by exploiting both phone operating system and hardware vulnerabilities. Indeed, at least one Israeli company openly boasts about giving law enforcement agencies the tools to bypass Signal’s encryption. With these vendors willing to do business with any government that will pay, the risks to privacy have metastasized from the policies and conflicts of the technology provider or its privacy commitment to pure commercial jungle warfare.

In short, there are many reasons why an African businessperson may wish to switch among the big messaging apps. Keeping the government’s long nose out of one’s business affairs is, unfortunately however, increasingly less tenable as a basis for choosing among the options available.

I have received a letter from the Office of the Senior Presidential Advisor of Ghana (formerly, “Office of the Senior Minister”) purporting to be a response or rejoinder to two tweets about Ghana’s stagnant public sector reforms effort I sent out on 21st and 25th February, 2022.

For clarity, I reproduce the tweets here.

It is heartening to note that the Office of the Senior Presidential Advisor (OSPA) shares with many of us the belief that public sector reforms in Ghana remain a pressing and crucial issue, and thus require careful scrutiny.

The OSPA’s 5-page response/rejoinder to my claims of the their obsessive focus on procuring expensive cars to the neglect of other, perhaps more critical, matters is thus the least one could expect in these circumstances.

Unfortunately, the response falls short of the requisite thoroughness, accuracy and candour required to address the concerns analysts continue to have about the results to date of Ghana’s longstanding quest to reform its public sector to improve its effectiveness and contribution to total development. Before we delve into these shortcomings, it is essential, given the sheer depth of this topic, to recount the full history of Ghana’s attempt to comprehensively reform its public sector institutions from the center. Below, a compilation of the reforms and their chronology by a team led by the World Bank’s Nick Manning in 2012 is reproduced verbatim.

Non-Structural Reforms

Structural Reforms

A simpler overview that covers more recent attempts is provided below.

Source: Martin J. Williams and Liah Yecalo-Tecle (2019)

It is easy to deduce from the above that serious attempts at actual structural reforms of the Public Sector did not even begin until 1994, and this doomed attempt – the National Institutional Renewal Program (NIRP) and Civil Service Performance Improvement Project (CSPIP) – had by 2001 completely lost steam. Since then, a mix of different strategies have been pursued culminating in the current emphasis on “results” and “delivery”.

From this results and delivery emphasis, three key themes for the reform agenda emerge:

  • Increasing Personnel Productivity
  • Making Decentralisation Work and
  • Eliminating graft, corruption, poor governance & other unprofessional conduct

Even as individual initiatives, like computerisation of public finance administration and linkage of capacity building with demonstrable skills upgrades in the Civil Service, have repeatedly failed, a rinse-repeat-recycle process have kept these goals at the forefront of the reform agenda.

Here are some of the practical issues that we currently grapple with as a country because of persistent failure of public sector reforms:

  • The country spending more roughly 40% of total government revenues on paying public sector workers every year over the last decade, considerably higher than most of its peers. Yet, salary agitations remain rampant, suggesting poor distribution, high distortions and low productivity. A conclusion supported by the fact that Ghana actually employs fewer public sector workers in some crucial areas like health, education, security and agriculture than peer countries in Africa and that its public sector personnel are marginally less educated than on average.
Source: IFS (2019)
Ghana’s high wage bill as a share of government revenues compared to peers. Source: IFC (2019)
Ghana’s relative public sector compensation bill is enormous even by global standards. Source: World Bank
Source: IFS (2019)
  • Long delays in paying salaries of newly recruited staff due to poor payroll vetting procedures. Erratic pay has become so institutionalised that some rural banks are reported to be building products around it. It is not clear how an employer can maintain staff motivation when salary arrears can pile up for months on end.
  • Poor maintenance of amenities and delivery of social services in many peri-urban and rural areas as a result of the inability of local government bodies to raise local revenue.
  • Consistent poor financial and performance results at many state institutions according to perennial Audit Service indictments.
  • Continued extortion of citizens by public sector workers before delivering benefits to which citizens are entitled. Both Afrobarometer and Transparency International data shows zero improvement of citizen experience of corruption and extortion over the last decade.
Corruption Perception among the population. Source: Afrobarometer (2019)
Source: Yeboah-Assiamah et al (2016)

To do justice to the Office of the Senior Presidential Advisor’s (OSPA) rejoinder, the above context should shape our approach.

I made the following claims on Twitter:

  1. A summary of the 2020 Annual Report of the Office of the Head of Civil Service corroborates the position that the bulk of expenses made in pursuit of public sector reforms by the OSPA went to buy cars. That fact cannot be challenged as it is based on a verbatim extract from that report.
  2. A larger share of the World Bank resources given to the OSPA to spend on advancing public sector reforms in Ghana went into buying cars. The evidence for this is from the World Bank’s own reports on the project and will be presented in this essay.
  3. The largest single completed expenditure was for 7.4 million Ghana Cedis worth of cars in December 2019. The Senior Minister insists that the expense was 8.5 million GHS worth of cars and the spending year was 2020. I relied on World Bank reporting, which I will reproduce below. At any rate, this “clarification” makes no material difference to the core argument.

Before I produce the relevant extracts from the World Bank reports, it is important to remind readers to analyse the scope of activities listed by the OSPA in their “rejoinder” to determine if indeed their explanations measure up to the scope of the public sector reform challenge described thus far in this essay.

Are vehicles for personnel the most critical instruments for boosting fiscal decentralisation, eliminating corruption and extortion and boosting productivity? Is there any evidence to suggest so from any study? Can the OSPA cite any such studies? How has the public sector reform project fared as a result of this decision to focus on buying brand new cars for sixteen agencies?

Find below an overview of what the former Senior Minister committed to in exchange for funding from the public purse to transform the public sector.

Source: World Bank

Has the OSPA delivered?

  • Did 250 government agencies establish fully functional client service units by the end of last year?
  • Did Parliament of Ghana ratify the African Charter on Values & Principles and is it being applied to professionalise the public service?
  • Has the Subvented Agencies Reform Programme been successfully implemented and output measured in accordance with the targets?
  • Was there a positive improvement in macroeconomic indicators such as inflation, primary/cash deficit, exchange rate, debt sustainability etc.?

Our investigations reveal serious underperformance on each and every bar the Office has set for itself in every one of these critical areas, not to talk of complete neglect in the OSPA’s strategy of very critical matters like public sector workforce rationalisation, state enterprises (still declaring record losses) and effective fiscal decentralisation.

Even within the narrow confines of the World Bank assisted component of the OSPA’s strategy – improving specific services delivered by sixteen selected agencies across the Ghanaian Public Sector – the performance of the OSPA has been woeful.

From mainstreaming gender priorities to modernising records administration, building systems to reduce vendor payment arrears accumulation in the Civil Service to reforming pension administration, and from transforming the speed with which institutions like CHRAJ manage complaints to overhauling performance contracting at Civil Service Director level, there has so far been little evidence in the Public Sector that the strategy is even beginning to take off, much less have impact. Below is an overview of the resources committed from the World Bank loan to driving the reforms exercise between inception and 2023.

Source: World Bank

Essentially, Ghana has gone to borrow $35 million to reform the Public Sector yet little progress is evident. Still, the OSPA keeps buying cars to distribute to bureaucrats. Meanwhile, little is being done to set the project on a serious course to results.

Worse, in some cases, the targets set simply conflict with other government policy. For example, the push to reduce vehicle registration to 1 hour is based on arbitrary standards. Currently, both roadworthy certification and environmental standards compliance (for emissions tracking among others) in the vehicle registration process fall far short of acceptable standards. No resources have been devoted to improving the equipment and skills needed to enhance these and similar aspects of the vehicle registration process. Indeed, in a study by Dr. Ayetor, Dr. Ampofo and their colleagues in 2021, only 25% of Ghanaian cars in the study sample passed vehicular tests based on national and global standards. Any artificial reduction in registration time would thus only go to damage the quality of certification and compliance even further. Which then raises the question: what difference would buying cars for DVLA make in this major area, for example?

In a similar vein, the poor coverage of birth and deaths in Ghana is due to poor coordination among local government institutions and the Birth & Deaths Registry. Poor public education through strong social institutions like religious bodies and traditional authorities has also been identified as a major weakness. Based on what research then was a decision made to prioritise the buying of cars for the Births & Deaths Registry?

NITA is expected to act as the primary IT Services Provider to all government agencies. As part of the reforms effort, NITA is expected to sign service level agreements with these agencies so that they can be assured of responsive service when they need it. NITA so far refuses to invest in building these helpdesks and operationalising them for the full range of agencies that need support. What has buying cars for NITA bosses to ride around town got to do with addressing this problem?

For every single one of the 16 agencies and the targets of improvement set for them, factors other than transportation/mobility matter far more. Yet, the OSPA decided, for reasons they fail to disclose in their “rejoinder”, that cars should be the most critical input in transforming the processes of these institutions.

Not surprising then that the rating of the project continues to be unsatisfactory.

Source: World Bank

More than 80% of project indicators by my reckoning are not on track.

As is customary in government business nowadays, public accountability, which was designed into the project, has been virtually zilch.

Source: World Bank

As I have already indicated, without any serious rebuttal from the OSPA, the bulk of spending (~80%), as at end 2020, has been on cars.

Source: World Bank

The only substantive push for stakeholder engagement was money disbursed to support Civil Service Week and a so-called Results Fair, which as many would recall consisted primarily in the government announcing projects that make it look good but unrelated to public sector reform.

In short, public sector reforms in Ghana is a big deal, but the Office of the Senior Presidential Advisor is misguidedly investing in cars rather than in critical inputs required to make a serious impact on productivity, professionalism and decentralisation.

As for whether the bulk of the cars were bought in 2019 or 2020, the discrepancy is purely one of accounting treatment. The contracts to buy the cars were signed in 2019. In standard accounting, expensing happens when expenses are incurred. The OSPA can choose to focus on when the cars were actually delivered. It makes no material difference.

Source: World Bank

In view of all the above, I humbly decline the invitation of the eminent and esteemed former Senior Minister to retract and apologise for my claims. On the contrary, I choose to double down and amplify them, as I have done in this essay. Thanks to the OSPA for initiating this dialogue.

On 14th January 2022, the Pan African Payments Settlement System (PAPSS) launched in Accra. In addition to six West African central banks, mainly from the mainly Anglophone WAMZ region, Afreximbank (the lead sponsor) and the AfCFTA Secretariat, a number of important governments and corporations graced the event to signal their support.

PAPSS is considered the fourth of five key pillars supporting the effective rollout of the African Continental Free Trade Agreement (AfCFTA), a treaty signed by 54 of Africa’s 55 countries (excepting Eritrea) and ratified by 41 of them so far with the goal of creating the long awaited common market across the continent.

AfCFTA is momentous even by global standards. It is the world’s largest trade area by participating members, even if other trading blocs like RCEP and the EU have greater weight by GDP. But Africa’s integration and unity has had so many false starts that AfCFTA has been greeted by much apprehension since it was signed in 2018 and even after it went live in January 2021. Take the famous “open skies agreement” which is now two decades deep into oblivion.

Mindful of this context, many observers have been looking for “what is different this time around”, and some have fallen for tech, financial technology in particular, which is where PAPSS comes in.

PAPSS’ key promise as a saviour of AfCFTA is deliciously straightforward, not only will it save the continent $5 billion in fees that currently goes offshore, it will also help AfCFTA take trade among African countries, presently estimated at 17%, to as much as  ~22% by 2040, generating additional value of about $70 billion.

Indeed, if forces like tech push liberalisation even harder, the proportion of Africa’s trade that stays within the continent could exceed 25%. A sensible ambition when you consider that nearly 70% of European trade happens within the EU and 60% of Asian trade within Asia. PAPSS will do this by enabling African traders to pay for goods in other African countries using their own domestic currency whilst the seller gets paid in their local currency as well.

Like every big idea in Africa, PAPSS has been tried in different forms and guises over the years in Africa. A number of Africa’s major regional trading blocs have something similar. In the Southern part of the continent, SADC has a unified payment system called SADC-RTGS (previously, SIRESS) patronised by 14 of its 16 countries (with transactional value exceeding $450 billion by March 2020). The East African bloc has EAPS and both major Francophone-dominated zones – WAEMU (West Africa) and CEMAC (Central Africa) – have unified, multi-country, cross-border payments systems as well. In fact, 30% of trade in the Francophone trading blocs happen in the regional CFA currency on the regional payment networks, such as CEMAC’s SYGMA, already.

So, apart from ratcheting up the scale, could PAPSS change the game much more radically? A lot would depend on a more rigorous definition of the problem it has set out to solve and a lot of strategic wizardry. Misunderstanding PAPSS’ true opportunity would lead to disappointments and confusion.

Some have dubbed PAPSS a SWIFT killer because of certain misconceptions about the promise to “save Africa” billions of dollars of fees which currently go overseas. But this is a complete misapprehension.

SWIFT is the global behemoth that enables banks to send secure transactions to each other authorising payments from sender to beneficiary. 11,500 out of the world’s 25,000 banks, of which about 1050 are in Africa, send 42 million such messages every day to 200 countries around the world. The actual fee per message is around 4 US cents ($0.04). This is hardly the driver of the 6.3% of sending amount that senders pay on average to transfer money from country to country or the $25 to $65 senders see on the telex advice when they wire money from their bank to a recipient in another country.

Those costs are driven by intermediary banks between theirs and the recipient’s, sometimes called “correspondent banks”. Because it is unlikely that the sender and the recipient would both have accounts in the same bank, especially for an overseas transaction, the only way to transfer money in the current global banking system is for the sender’s bank to look for another bank in which both they and the recipient bank have accounts. These would typically be big global banks since it would be ridiculous for each bank to hold accounts in thousands of banks around the world.

A few hundred global and regional banks are trusted enough to serve as bankers to other banks in order to facilitate these global payments. Occasionally however one finds that there is no intermediary bank that both sender and recipient banks share in common, necessitating the search for an intermediary between the intermediaries. Now, because each bank needs to be paid for their service in the chain, costs can rack up.

For that reason, the issue with enhancing payment flows and cutting costs is not really about the need for an African financial messaging service. In fact, the biggest of the existing regional payment networks, the SADC-RTGS, actually uses SWIFT for the messaging part of the process (as does the UK’s CHAPS network and many others around the world). In fact, the SWIFT charges in the transaction fee are lower than the SADC-RTGS charges.

In many respects, the framework and architecture for all cross-border payments draw on the standard Real-Time Gross Settlement (RTGS), which dates back to 1970 when its bare contours were set by the US Fedwire system. There are about four major companies around the world trusted by central banks and other major financial system actors to build these networks: Sweden’s CMA, the UK’s Logica, Swiss-South African Perago and Montran, an American firm.

The costs of implementing RTGS networks in countries and linking them together is far from prohibitive. In 2008, the African Development Bank (AfDB) provided a grant to the four West African countries outside the Francophone CFA area that had not built an RTGS network to build one and network them together to establish a common network. The project, undertaken by Sweden’s CMA and French-Tunisian firm, BFI (eight other contractors playing minor roles), cost about $36 million.

Source: AfDB

If there is a serious driver of cost in the African context, as far as the technology infrastructure itself is concerned, then it is primarily volumes and participants. African banking systems often have lower volumes of transactions than elsewhere, leading to a higher cost per unit transfer.

For example, take Liberia. Once the AfDB-funded RTGS platform went live, the country’s central bank needed to devise a formula to charge participants. Below is the breakdown it came up with the breakdown below.

Source: Central Bank of Liberia

It is clear from a even a cursory glance at the table that the primary cost driver here is firstly the number of participant banks sharing the capital costs, and how that will be passed through in fees to the end user. A fact clearly evidenced by the lower costs for similar services in the United Kingdom.

Source: Bank of England
Source: Bank of England (2021)

Apart from the somewhat trivial costs associated with SWIFT, its large volumes also provide a resource base that guarantees very robust infrastructure and high uptime availability.

In an analysis by the CEMAC Central Bank (BEAC) in 2017, it was shown that 80% of adverse incidents reported related to the region-specific network SYSTAC whilst transactions operated on the global SWIFT network registered only 5% of the reported incidents. Of these incidents, whilst 100% of the SWIFT ones were resolved, 60% of the SYSTAC incidents remained unresolved.

Not surprisingly, many of the major African banks have not been as enthusiastic as one would assume in promoting regional payment networks. Even the most successful regional payment networks can find this barrier daunting. Five years after launch, only 31% of banks in the SADC region had signed on to SIRESS (now SADC-RTGS) and nearly 60% of all traffic still bypassed it.

Messaging facilitation is thus not a significant source of new value to entice banks to a payment network. The real source of value, the ease and course of intermediation, is also the biggest driver of cost, partly through margins buildup and an even more critical factor, liquidity risk.

A proper understanding of the intermediation and liquidity issues also addresses another misconception around PAPSS: speed of payments.

It is not the inherent inefficiency of the SWIFT system itself that causes delays, contrary to some popular perceptions. In actual fact, the average time to settlement of 40% of all SWIFT transmissions is five minutes. Intermediary bank involvement drives this average up to 30 minutes for 50%, 6 hours for 75%, and 24 hours for 100% of all transactions on the network. Virtually, all sources of delay are due to errors during initiation, of which 34% are formatting related alone.

In 2017, SWIFT introduced a set of enhancements called the Global Payments Innovation (GPI) meant to reduce the value chain related challenges and errors that often gets blamed on it. Of the banks in Africa that have adopted it, transaction time for 70% of payments is within 5 minutes. Any extra delays are due to factors such as regulation (many African banks, for instance, run additional manual anti-money laundering and anti-fraud checks before crediting inbound remittances). The problem is that only 5% of African banks (less than 50 out of nearly 4500 GPI-adopting banks worldwide) have signed up to the strict service level commitments needed to activate GPI for their customers.

Without African banks themselves stepping up to the plate to improve intermediation, PAPSS by itself will not be able to transform the payments value chain and cut the costs, time and inconveniences that are currently in the way.

The closest thing to what PAPSS wants to create is the European Union’s TARGET system (more precisely, its TARGET2 incarnation), the principal outcome of the Single Euro Payments Area (SEPA).

If one looks at the TARGET2 price list, almost all the cost drivers relate to the risks and value of liquidity facilitation (and also error management, but that is incidental):

Source: TARGET

When one steps into a bank branch in Kampala to make a transfer of about $500, the charge for an outward EAPS transfer (made on the regional network) is $11.5 and $21 for a SWIFT payment. If the user initiates the payment online, the cost drops to less than $6 regardless of method. Anything else that is added relates to the presence or otherwise of an intermediary bank. The total expense associated with the transaction thus relates much less to the means of issuing the payment instructions (the payment platform, properly speaking) and much more to the commercial forces within the interbank network.

As the Kenyan Central Bank (CBK) noted in its recent discussion paper on the prospect of launching a Central Bank Digital Currency (CBDC), the East African Payment System has struggled primarily because of such commercial forces, in this case the burden on individual member banks to hold enough liquidity of the different currencies involved in a cross-border transaction. Indeed, it is primarily such liquidity issues that made adoption of RTGS so slow to begin with. Unlike the batch settlement method where banks net off how much they owe to each other due to the transfers that have occurred over the period, RTGS type systems require constant, or at least high frequency, settlement making smaller and weaker banks potential bottlenecks in the chain.

CBK muses in its CBDC paper whether cryptocurrency innovations can help mitigate these currency interconvertibility situation. PAPSS does not go the cryptocurrency route. Near as I can tell, the mighty balance sheet of Afreximbank is the gamechanger here. Afreximbank aspires to become the major intra-day credit provider for liquidity purposes that in many domestic RTGS systems, central banks tend to play. Would it do this for free or at a cost?

There is no doubt that the entire success of PAPSS turns on how masterfully Afreximbank can position itself as the credit facilitator of continental trade. The political economy challenges are of course formidable but the path is clear.

Even within regional monetary unions in the continent, political friction is constant. Recently, the CEMAC central bank purported to block a national switch in Cameroon for competing with the regional payment network.

 Some have suggested that PAPSS could circumvent friction by simply interconnecting the regional platforms like EAPS, SYGMA and SADC-RTGS to each other. Doing that, however, will imply a completely different business model as PAPSS would then not be able to market directly to banks.

The fact though is that 96% of all Africa-bound payments originating in the East African Community (EAC) end up within East and Southern Africa. 92% of Africa-bound payments originating in Southern Africa stay in that region. In Anglophone West Africa (WAMZ), on the other hand, less than 40% of Africa-bound payments stay within the WAMZ. Not surprisingly, enthusiasm for PAPSS is currently strongest in WAMZ, where the central banks have signed the four foundational legal instruments underpinning PAPSS, and, till date, lukewarm elsewhere.

It is not too clear that the inter-regional payments systems integration value proposition is the strongest. Right now, there is significant room for improvement within the regional payment networks themselves. PAPSS can become the powerhouse for driving collaboration among those banks that really want to cut intermediation fees and enhance the liquidity profile for cross-border trade payments, regardless of where they may be on the continent. As the SWIFT GPI saga has demonstrated, payments transformation is purely ecosystemic and value chain dictated.

Some trends point to opportunity. Between 2013 and 2017, intra-African correspondent banks increased from 230 to 260 whilst the number of foreign correspondent banks willing to do business with African banks dropped significantly. It would just be a matter of time before smaller African banks are relying more on the bigger African banks to navigate the global banking arena.

With its growing financial muscle, strong relationship with the AfCFTA Secretariat, and through deeper alliances with the AfDB and Africa’s biggest banks, Afreximbank can reduce the costs of maintaining expensive global relationships for the continent’s smaller and mid-size banks who do much of the SME banking.

The real opportunity in payments revolution is not at the RTGS layer, which is by and large a solved problem. It is in the “open banking” layer, where much smaller players can connect across simplified connection pipes to introduce life-changing services beyond moving money from point a to point b. To do that though means solving the massive issues of liquidity, identity, cross-border KYC, and currency interconvertibility, all of which are long overdue for radical innovation.

To date, regional payment networks relying on traditional tools have not been able to do this. PAPSS has the opportunity to go to the banking associations and governments and offer radical new approaches to cost-cutting within existing regulatory jurisdictions and once they bite to architect a continental model based on well functioning units. Such a move would not make it a friend of many of the traditional incumbents, but nothing ventured nothing gained.

In the 24 hours since my comment on the e-Passport controversy, more information, especially a video of the key ceremony in Montreal, has emerged to show that the situation is even crazier than initially thought.

It is clear from the video that the ICAO officials who received the Ghanaian delegation were under a completely different impression about the Public Key certificate (think of it as a very hard to forge “electronic signature” for the state of Ghana) that the Ghanaians wanted to submit to the ICAO Public Key Directorate (PKD). They clearly thought that it was meant for the other 79 governments in the PKD to be able to authenticate an electronic chip embedded in the Ghanaian Passport, not a national ID card.

The remarks of both officials who welcomed the delegation expressly relate to how Ghana’s passport would now be more secure from fraud than has been the case in the past. The occasion was meant to be celebratory. A decade ago, only two African countries were part of the PKD – Nigeria and Morocco. Ghana’s joining takes the number to an impressive 14.

Source: ICAO

As it turns out, the Ghanaians had no intention to use the certificate for the Ghanaian passport, which is still one of the few in the world without an electronic chip (one of only 9 in Africa), making it highly subject to abuse. In a famous forensic study, 46% of Ghanaian passports tested turned out to be fake.

What is even more bizarre about the whole saga is that last October the German company that built the Ghanaian solution being used for the PKD enrolment, Cryptovision, announced how proud it was that its technology was being used to turn the Ghana Card into an e-Passport. They were educated about their misconception and then compelled to correct the statement to read that the Ghana Card is now an electronic Personal Identity Card, not a ePassport.

Screenshot of Cryptovision PR statement (as of 16th February 2022)

Considering that Cryptovision is the company that built the chip solution and software, the CAmelot  platform, being used by Ghana to broadcast its national electronic signature through the ICAO PKD, it should at least know what the real arrangement with ICAO was.

Indeed, the project scope negotiated by Cryptovision’s Adam Ross for the Ghana Card in 2013 explicitly describes an electronic personal identity card rather than an e-Passport.

e-Passports, as I have tried to explain in detail in two previous articles, constitute a special category among electronic identity cards generally because they are meant to be embedded in passports. Passports have a unique capacity as travel documents already accepted across the world for human mobility purposes due to the dictates of clear international law.

As Adam Ross himself has outlined below, Ghana did not engage Cryptovision to customize its ePasslet solution for an e-Passport solution as countries like Angola and Malaysia did.

Extract from brief presented to customers by Adam Ross, a Cryptovision Executive

It is thus completely mind-boggling that Ghana, knowing all this, would still send a bevy of senior officials, including its resident Ambassador in Canada, to Montreal to attend a key ceremony, ignore the official publicity template issued by ICAO in the key ceremony dossier, and then announce to the world that ICAO has accepted the Ghana Card as an e-Passport, when its base solution is NOT configured for that purpose.

Here is how ICAO defines the eMRTD functionality:

a  Machine  Readable  Travel  Document  (MRTD)  that  contains  a  contactless  Integrated  Circuit (IC)  chip  within  which  is  stored  certain  specified  MRTD  data,  a  biometric  measure  of  the  passport  holder, and  a  security  object  to  protect  the  data  with  Public  Key  Infrastructure  cryptographic  technology,  and  that conforms  to  the  specifications  set  forth  in  ICAO  Doc  9303.

It further defines a “participant” (i.e. member state like Ghana) in the PKD scheme as:

a  Contracting  State  or  any  other  entity  issuing  or  intending  to  issue  eMRTDs  who  follow these Regulations  for  participation  in  the  ICAO  PKD.

Professor of Law, Adam Muchmore, describes the distinction between passports and other identity documents, as follows:

Passports, as prima facie evidence of nationality,’ are “normally accepted for the usual immigration and police purposes.”‘ In other words, states take daily legal action on the basis of passports issued by other states, without taking time to investigate whether the passport holder is “really” a national of the issuing state…A passport in this case is different from a national identity card, addressed only to other actors within the issuing state. [Emphasis not author’s.]

A passport’s position in international law is so fortified that no identity card however supported by the interplay of commercial and political interests in a country can usurp its role without consequence.

The principal legal specifications on the subject of electronic travel documents (or, more precisely, “machine readable travel documents (MRTDs)”) binding ICAO itself is the so-called “Doc 9303”. The 8th edition of which was released in 2021. The broader legal framework for the whole business of travel document technical specification can be found in Annex 9 to the Chicago Convention, the world’s main body of international aviation law.

In 2005, the standard to which MRTDs should conform was agreed by the ICAO member states and 2015 was set as the deadline for the phasing out of non-conforming travel documents. Doc 9303 has since been kept up to date with all essential technological evolution.

The specification defines an ePassport as:

Commonly  used  name for an  eMRP. See Electronic Machine Readable Passport (eMRP).

An eMRP, on the other hand, is defined as:

A TD3 size MRTD conforming  to the specifications of Doc  9303-4  that additionally incorporates a  contactless integrated  circuit including  the capability of biometric identification of  the holder. Commonly referred  to as “ePassport”.

What the term “TD3” means can be found in Part 4 of Doc 9303.

Parts 5 and 6 of the same technical document describe other machine-readable official travel documents specified in various bodies of international law, especially those dealing with crew, refugees and human rights.

Section 2 of Part 4, relating to the description of ePassports, describes the form factor (i.e. how an ePassport should be presented) as follows:

The MRP shall take the form of a book consisting of a cover and a minimum of eight pages and shall include a data page onto which the issuing State or organization enters the personal data relating to the holder of the document and data concerning the issuance and validity of the MRP. After issuance, no additional pages shall be added to the MRP.

Section 2.2 goes further to specify the dimensions of the booklet:

The nominal dimensions shall be as specified in ISO/IEC 7810: 2019 (except thickness) for the TD3 size MRTD, i.e.:  125.00 mm (4.921 in) wide by 88.00 mm (3.465 in) high

Part 4 also presents the all-important machine readable zone in the drawing below:

Source: ICAO

And furthermore specifies the biodata page as in the diagram below.

Source: ICAO

Even clearly itemising which zones are mandatory.

Source: ICAO

It then presents several examples of conforming pages of a compliant e-Passport, such as this one:

Source: ICAO

Such is the serious technicality of what is at stake here. It is evident from all the preceding that ICAO indeed had no authority to go contrary to its own enabling regulations and authorize the Ghana Card, which simply does not conform to the specification (contrary to several assertions by government of Ghana spokespersons) to be used as an e-Passport.

Even more emphatically, the government’s continued insistence that the Ghana Card is a e-Passport is both incorrect and embarrassing. It puts ICAO in a tight spot since the government refuses to accept the former’s informal clarification. The official posture of the government is to doggedly hold on to its position despite clear evidence of confusion in the aviation community. In this strange conduct, the government is aided by pliant public actors who can’t or won’t call it out.

Yet the PKD as structured requires clarity as to which documents precisely a member is applying its certificate to.

Extract from an official ICAO Presentation

The live verification protocol of the PKD is a tight security system based on rules and standard operating procedures.

Extract from an ICAO Presentation

The overarching principles for the establishment of the PKD are set out in the procedures as follows:

  1. The ICAO PKD has been established to promote a globally interoperable ePassport validation scheme for electronic travel documents to support ICAO’s strategic objectives to improve aviation security and improve the efficiency of civil aviation. 

2. The benefits of ePassport validation are collective, cumulative and universal. 

3. The objective of the ICAO PKD is to support validation of all ePassports that are widely accepted for travel and identity verification purposes by ICAO Contracting States.

Extract from ICAO PKD Procedures Manual

No country in the world that has joined the Scheme has ever thrown into question the conformance of the procedures and state practice, on the one hand, with the ICAO PKD enabling legal documents, on the other hand, in the way we have witnessed these last few days.

The time for wordplay and needless argumentation to end is now. The government of Ghana should formally retract its ePassport claims for the Ghana Card and immediately take steps to transition the indomitable Ghana Passport to a full electronic passport (with an ICAO-compliant chip) so that the country can fully benefit from the fees it is paying to be part of the ICAO PKD. A serious Minister of Foreign Affairs would make a clear statement on this matter, in their capacity as the seniormost official responsible for the sanctity of the Ghana Passport.

And to the extent that international law makes the question of nationality of a passport holder moot, it is also time for the government of Ghana to rescind all the rules made by the National Communications Authority, the Bank of Ghana, and other overzealous agencies barring holders of the Ghana passport from accessing telecommunications and banking services in Ghana. No country in the world has done this. Ever! Ghana is not that special.

We know that the investors behind the country’s heavily outsourced Ghana Card intends to meet their target of $1.2 billion of revenue in a few years. But it should not be at the expense of the country’s international image and the rights of its citizens.

The first time I read Amartya Sen’s The Argumentative Indian, I laughed and said to myself, “the eminent economist has probably never met a Ghanaian in his life!”

My reaction was tongue in cheek, however, since Professor Sen’s main thesis is not that Indians just love arguing for arguing sake, but rather that the subcontinent’s heritage of intellectual pluralism is foundational to its unique democratic project.

Ghana’s “argumentative tradition” is more literal; in Ghana, everything is debatable!

On 6th November 2021, I appeared on Citi TV’s Big Issues program, where the topic of discussion was Ghana’s “digitalization agenda”, championed by the country’s Vice President, a subject I subsequently wrote about here. In the studio was the Executive Secretary of the Vice President, who also doubles as a Chief in both the Sefwi and Mamprusi areas of Ghana.

A lively debate ensued about a flurry of PR statements from the Vice Presidency touting the national ID card (the “Ghana Card”) as a “e-Passport” due to Ghana’s decision to join the ICAO PKD Network. During the discussion, it became apparent that unnecessary confusion was being created about how the development was a benefit conferred on Ghana because of the Ghana Card, seeing as that misconception is completely untrue. The Ghana Card has no role to play. It is not the card that has been “recognised” but Ghana’s “public key”, which the lay reader can think of as a “highly secure electronic signature”. That key can be embedded in any ID document at all. After the show, I wrote my article and considered the matter closed.

Until I woke up a few days ago to an official Ghana News Agency report splashed across Ghana’s media networks that ICAO has “recognized the Ghana Card as an e-Passport”. The subsequent rebuttal of this bizarre and incoherent claim by ICAO itself did nothing to quieten the controversy. Debate continues on social media even now, with government affiliates still pushing the narrative of the “Ghana Card having become an e-Passport” and opposition activists having a field day with the ICAO statement.

But what makes this a uniquely Ghanaian contrived controversy?

Because, first of all, no country in Africa that has joined the ICAO PKD network has succeeded in transforming it into a high-stakes, high-octane, political jamboree. See the list below.

List of African PKD Participants

Nigeria has been a member of the PKD network for more than a decade, but I don’t recall seeing a single episode of government-opposition tussle over what this means. In each of the African countries that have chosen to publish their public keys in the ICAO directory and comply with the smart chip standards that make verification of the public key possible by the other 79 countries that are part of the process, the whole endeavour has been treated as an obscure technical development.

Joining the ICAO PKD simply means that other countries in the network can confirm that the “electronic signature” on documents issued by a joining government are genuine. This process itself has got nothing to do with e-Passport functionalities per se, which are about encoding the principal information of the traveler (including biometrics) in electronic format and the acceptance of the document in which the electronic information is embedded. India has been a participant of the PKD since 2009 but it is only now about to add e-Passport functionality to all its passports (some Indian officials have had e-Passports since 2008).

Uganda has been part of the PKD for nearly three years now but is only now about to transition from machine-readable biometric passports to a full e-passport. Indeed, Ghana initiated that process in 2014 but could not conclude the contracting to complete. Of course, Ghana had not joined the ICAO PKD when it was contemplating this transition in 2014 but one does not need to be part of the ICAO PKD to issue e-Passports.

In fact, there are countries like South Africa and Kenya with advanced digital travel identity systems in Africa that are not yet participants in the PKD system. Some countries like Taiwan that have migrated from machine-readable to the e-Passport model cannot directly participate in the ICAO process because of geopolitics, even though Hong Kong and Macao, autonomous regions of China are both participants.

So, this is a very simple matter. In the past, passports could only be read visually. The border agent/immigration officer would process the traveler’s entry by typing their details into host/receiving country’s system. The world eventually moved on to the use of machine readable passports whereby the immigration officer just scanned the passport and the computer automatically extracted the information.

In 1998, Malaysia pioneered the use of biometric passports, the next step in the evolution. The biometric signature of the passport holder was transformed into a format that could be read by scanners at many immigration counters around the world. For this to be possible, an 8 kilobyte integrated circuit card (or “chip”) was embedded in the passport to hold the traveler’s biometric data.

The problem was that, for full security, it was important for this biometric data to be encrypted else a criminal could cut their own chip and embed false data. But encrypting the data also meant that some method had to be found for the destination country’s immigration officers to decrypt and read the data on the card. This is where the ICAO PKD came in handy.

In 2004, ICAO published a specification of how all countries should store and secure information in the passport on the chip, and Belgium became the first to comply, and one of the most enthusiastic upgraders in the world. In 2007, ICAO also established a secure channel, the PKD, through which countries can decrypt the secure information in the e-Passport chip.

Source: Kumar & Srinivasan (2012)

Note that all this while we have been talking about the passport that everyone is familiar with, that same precious booklet that for some is the symbol of true freedom. We have not talked about smartcards. In fact, there is no expectation that a country will use a smartcard for any of this stuff. A “e-passport” is a functionality not a device.

Of the 150 countries that issue passports with some e-passport capability, virtually all of them integrate that capability into the passport booklet itself, primarily because that is the one that all countries accept and is currently universally compatible with the global visa regime and the ICAO e-Passport definition. Thus, countries have focused their investments on incorporating e-passport functionality into their passport booklets. In fact, some countries like the United States deny their visa waiver programs to nationals of countries that do not do this.

Notwithstanding the fact that many e-Passport issuing countries are not part of the ICAO PKD, commercial vendors have found a means to embed the capability to read most of these passports in scanners and software that they market to governments around the world.

Ghana is, in fact, one of the few countries in the world still issuing passports with no chips. Passports that therefore cannot be securely validated by commercially available software or through the ICAO PKD as the map below shows.

Source: Read ID (2021)

With more and more countries embedding e-passport capabilities into their passports, making them readable in more and more airports, there is a steady race towards a period where visas will also be “written” to chips in passports issued by other governments. But this will in no way lead to an abandoning of passports for smartcards because electronics can and do fail. Hence, visual and optical reading will for a long time remain necessary. Passport booklets are not going anywhere soon, only the chips in them will evolve.

Image credit: Infineon

So where does the Ghana Card comes in? The Ghana Card as I have stated elsewhere is a major distraction from the critical task of securing the Ghana passport! Ghana is a member of the rapidly declining group of countries whose passports still have no chips (as clearly evidenced in the map above.)

Ghana’s laggard status here is really problematic considering the continued abuse of the Ghana passport by identity criminals and the associated risk to national security. And considering the unfavourable attention the country has courted internationally. Yet, because of powerful vendors behind the Ghana Card and the fact that it is a massive, highly commercialized, Public-Private Partnership (PPP), there is a crass attempt to divert resources to it at the expense of the beleaguered passport, something that a serious Ministry of Foreign Affairs wouldn’t and shouldn’t tolerate.

Of course, there are countries that have embedded chips in their national ID cards like Ghana has done. And some of those chips are ICAO PKD-compliant. But as can easily be seen by analysing the ICAO PKD key ceremony, it is not ICAO PKD-compliance that confers the e-Passport functionality. Joining the PKD, at the cost of less than $30,000 a year, simply means that any ID document that Ghana electronically signs can be verified by other members of the PKD. But the overwhelming majority of countries apply the PKD system to their normal passport booklets instead of national identity smartcards (like the Ghana Card).

To activate a e-Passport, the chip should be incorporated in a document that is already accepted as a passport by a wide range of states. Interestingly, even among the 82 PKD states only a little more than half actually connect to the system to validate ID documents operationally. In short, the vast majority of states have border systems set up to verify passport documents by conventional means.

Right now, more than 90% of the countries in the world where Ghanaians may want to travel accept only the Ghana passport for the obvious reason that they are not party to the ECOWAS Biometric Identity Card (ENBIC) accord (which itself is still not fully operational since Senegal pioneered compliance in 2016). So the decision to prioritise the Ghana Card over the Ghana Passport for smart chip functionality is quite frankly bizarre.

Even within the West African region, the smart card authorization systems needed to read the ECOWAS Biometric Identity Card format do not exist in several airports. Most airlines are equipped to read the MRZ in the standard passport (that is why you see the check-in agents swiping the edge of the biodata page when checking you in) and not smartcard chips, nor are they allowed to access the ICAO PKD LDAP as yet (except in pilot form). Airlines around the world have security systems geared to the conventional MRZ in passport booklets.

Whilst the Ghana government has the sovereign right to tell airlines to just look at the smart card and allow people to board, there are security implications. For example, in some countries like the United States, which have laws like the REAL ID Act, airlines must confirm that international travelers are in possession of a passport or specifically authorised documents.

The key issue here is that whilst the passport remains the most critical travel document in the world, the government of Ghana, for totally unclear reasons, has decided to neglect its security, and is instead prioritising a domestic ID card. Its agents and spokespersons are on social media arguing over everything else apart from the essential question of why.

Talk about the Argumentative Ghanaian!

Ghana’s recent downgrading by Moody’s and the country’s angry response has triggered a flurry of commentary about the broader issue of bias in the ratings issued for African countries by three large Western corporations.

I recently wrote a piece that sought to look at the matter comprehensively. But judging by the feedback I have received it is the “bias” issue that most readers find interesting. So, I have updated that section of the essay with the following additional thoughts.

Whilst there is no denying the fact that African and Latin American countries have experienced more downgrades during this crisis, it is not clear how that fact alone can justify a charge of “systematic bias” when such bias is not aligned with the usual incentives of the agencies widely believed to favour upgrades.

Economist Hippolyte Fofack suggests that this could be due to an overactive sense of “reputational preservation”. Or, put another way, the agencies are overreacting in order not to miss a default and thus lose the confidence of investors. Perhaps, the sense is that in Africa it is safer to be wrong pessimistically rather than optimistically. Considering though that the rating agencies are super keen to grow their business in Africa and any country downgrade tends to result in automatic downgrades of existing and prospective corporate clients whose credit is tied to their host countries, such a logic is hard to fully unpack.

To be truly objective, we must also consider reasons other than bias why Africa and Latin America suffer such an unfavourable skew of downgrades. Historically, Latin America has accounted for a significant proportion of all global sovereign debt defaults and restructuring episodes. So much so that some of the countries there, like Argentina, have come to be viewed as serial defaulters (10 or so at last count). Between 1861 to 1920, 41 of the 58 defaults were in Latin America. Later in the century, African sovereigns joined the melee. In the 1980s “debt crisis” sparked by Mexico, Africa accounted for 34 of the 70 defaults, whilst Latin America followed with 29.

Usual suspects | The Economist
Source: Economist.com
Jonathan Fortun on Twitter: "Last week, we saw #Argentina's ninth default  in history. The #Latam region has been prone to defaults, especially during  the 80s https://t.co/77jEOn0iMR" / Twitter
Citing: Jonathan Fortun (2020)
Sovereign defaults from 1824 to 2004. Source: Borenzstein & Panizza (2008) | Figures quoted in the main text also draws on this work.

In short, there is a historical basis to suggest that African governments are risky business. Take the case of Mali, for instance, which lost its S&P rating in 2008, and then proceeded to default in 2012, thereby validating the action retrospectively.

To compound the historical default rate problem, African sovereigns have also historically accounted for a small proportion of syndicated loan and bond issuances worldwide. The small sample size of African issuance concentrates the effects of default and reduces the amount of information available for effective risk analysis, whilst also promoting overfitting to narrow risk models. In brief, the history of past defaults exert a greater Bayesian force on futuristic risk predictions when dealing with smaller data sets in view of the wider bounds of uncertainty.

Africa’s limited issuance compounds the effect of prior defaults on perception due to wide uncertainties. Chart credit: Presbitero et al (2016)

Even if we accept the arguments holding that there is a systemic bias against African governments unrelated to fundamentals, such as those by Fofack and the Africa Peer Review Mechanism (APRM), which a few years ago set up a program to monitor the effect of credit rating actions on African countries, the resulting analysis should merely establish a uniform hurdle for the whole of Africa. What then are we to make of Ghana’s specific deteriorating circumstances?

The above factoid highlights the scale of the Ghana “special case” and makes a compelling pitch for the Fitch and Moody’s downgrades, for instance. It warns that Ghana may simply not be the right poster child for a case of rating agency bias, however prevalent such a bias. For the full treatment of this argument, please read the main essay.

This afternoon, Evans Mensah, News Editor and Head of the Political Desk at Joy, a subsidiary of Ghanaian media giant, Multimedia, reached out to discuss the topic he was planning for his PM Express show tonight. He wanted to look at the Ghanaian government’s aggressive response to a decision by rating agency Moody’s to downgrade its bonds deeper into junk territory (Caa1).

This is something of a wonkish subject, and quite challenging to format right for television. I was nevertheless intrigued by the challenge because of the government’s over the top reaction. It added spice.

One Deputy Minister of Finance has called the Moody’s rating action propaganda. The same Minister also insisted that the rating action was the result of a grand conspiracy to force the country into an IMF program.

Even the official Ministry of Finance response to the downgrade takes the unusual step of singling out Moody’s Primary Analyst for technical incompetence, calling into question the overall professionalism of the agency’s operation. As follows:

We are very concerned that Ms. Villa may not properly understand and evaluate Ghana’s deepening credit story since obtaining our first credit rating back in 2003. She also has not visited the country since assuming the role and as such this downgrade at this critical time was based entirely on a desktop exercise, virtual discussions and what we believe to be the omission of critical data provided.

Extract from Ghanaian Ministry of Finance’s response to Moody’s Rating Report

I decided to join Evans, a government’s spokesperson, and the main Opposition Party’s parliamentary lead on finance to dig into this contentious affair. Our discussion revolved around two main questions:

  • Was the rating action based on an accurate assessment of the Ghanaian economic situation?
  • Was the rating action indeed biased against Ghana as asserted by the Ministry of Finance?

It is instructive to note that whilst Fitch and Moody’s have both downgraded Ghana in their recent actions, S&P chose to leave their rating unchanged for now. For that reason, the government has been full of praise for S& P. The inference we are being invited to draw is that whilst Fitch and, particularly, Moody’s are biased and inaccurate assessors, S&P gets it.

Before delving into the details of the bias and accuracy claims, a sentence or two can be spared on whether this is all a fuss over nothing.

In richer countries, some are already suggesting that downgrades (yes, rich, northern, countries get downgraded too) don’t matter all that much anymore. In fact, they have been saying this for quite a while now.

In less wealthy countries, on the other hand, such nonchalance is not affordable. Positive sovereign credit ratings have been linked with everything from helping boost domestic financial systems to tackling the ever so pernicious perception risk premia Africa continues to grapple with. Because most investors cannot develop specialised knowledge about the dozens of small economies around the world whose high-yielding investment opportunities (such as debt packaged as bonds) they are increasingly learning to like, they rely on credit ratings agencies, like the Big 3 (Moody’s, S&P, and Fitch), to give them the condensed version. Many investment committees will normally use sovereign ratings as a quick filter of which frontier market’s debt to include in their portfolio.

Hence, notwithstanding the the Financial Stability Board of the G20 (the group of the world’s 19 largest economies plus the European Union) having warned against such “mechanistic” or unsophisticated reliance on ratings, and though many of the top investment houses use ratings only as part of a composite measure when deciding to invest, sovereign ratings still have an outsized, strangely linearised, role in determining if a small, frontier, economy can borrow in the highly liquid international capital markets or not.

This is despite most ratings reports being littered with disclaimers like:

MOODY’S PUBLICATIONS MAY ALSO INCLUDE QUANTITATIVE MODEL-BASED ESTIMATES OF CREDIT RISK AND RELATED OPINIONS OR COMMENTARY PUBLISHED BY MOODY’S ANALYTICS, INC. CREDIT RATINGS AND MOODY’S PUBLICATIONS DO NOT CONSTITUTE OR PROVIDE INVESTMENT OR FINANCIAL ADVICE, AND CREDIT RATINGS AND MOODY’S PUBLICATIONS ARE NOT AND DO NOT PROVIDE RECOMMENDATIONS TO PURCHASE, SELL, OR HOLD PARTICULAR SECURITIES. NEITHER CREDIT RATINGS NOR MOODY’S PUBLICATIONS COMMENT ON THE SUITABILITY OF AN INVESTMENT FOR ANY PARTICULAR INVESTOR. MOODY’S ISSUES ITS CREDIT RATINGS AND PUBLISHES MOODY’S PUBLICATIONS WITH THE EXPECTATION AND UNDERSTANDING THAT EACH INVESTOR WILL, WITH DUE CARE, MAKE ITS OWN STUDY AND EVALUATION OF EACH SECURITY THAT IS UNDER CONSIDERATION FOR PURCHASE, HOLDING, OR SALE.

And

MOODY’S CREDIT RATINGS AND MOODY’S PUBLICATIONS ARE NOT INTENDED FOR USE BY RETAIL INVESTORS AND IT WOULD BE RECKLESS AND INAPPROPRIATE FOR RETAIL INVESTORS TO USE MOODY’S CREDIT RATINGS OR MOODY’S PUBLICATIONS WHEN MAKING AN INVESTMENT DECISION. IF IN DOUBT YOU SHOULD CONTACT YOUR FINANCIAL OR OTHER PROFESSIONAL ADVISER.

Yet, it is widely accepted that many investors lazily use ratings for more than they should and that, sometimes, poor regulatory design end up even giving ratings an oppressive force of law. Sovereign Credit Ratings are also firmly recognized as part of a triumvirate including sovereign bond spreads and nonresident holdings of general government debt. Said triumvirate dictates whether a country can retain access to international capital markets.

With these general background matters out of the way, let us now look at the central matters of accuracy and bias.

Accuracy

The Ministry of Finance’s press release takes issue with Moody’s calculations of the country’s prospective Debt to GDP, forthcoming balance of payments liabilities, and the impact of announced fiscal consolidation measures.

S&P, by contrast, seemingly aligned with the government’s perception of these matters leading to conclusions such as the following:

We project that government revenue including grants will increase to 17.2% of GDP by year-end 2022 versus 14% pre-pandemic, reflecting the introduction of new taxes, including this year’s pending electronic transactions levy.

And:

If the planned fiscal measures bear fruit, we project that net general government debt will stabilize at 77% of GDP this year, before gradually declining to 75% in 2025.

whilst Moody’s, on the other hand, doubled down on the view that:

Moody’s estimates that government debt ended 2021 at 80% of GDP while interest payments alone consumed half of government revenue that year (positioning Ghana with the second largest ratio among Moody’s rated sovereigns)

And

The government has announced a 20% cut in primary spending, equivalent to a 4% cut on a year-on-year basis or 16% in real terms, to compensate for any shortcoming in the government’s revenue measures package. Such an unprecedented fiscal tightening will be socially, economically, and politically challenging to implement.

Also:

Ultimately, Moody’s expects that a higher interest bill in 2022 and 2023 will offset the improvement in the government’s primary balance, thereby maintaining double-digit fiscal deficits (in cash terms) with a concomitant increase in the government’s debt burden.

On the ESG (Environmental, Social & Governance) front, where the government protests loudly about the lack of quantitative data in the Moody’s report, the rating agency asserts as follows:

Moody’s has lowered the governance issuer profile score to reflect domestic revenue mobilisation challenges and significant constraints on fiscal policy effectiveness reflected by very weak debt affordability. The authorities have undertaken some institutional reforms on the revenue and competitiveness front, which will take some time to produce results.

Instructively, both S&P and Moody’s disagree with certain data points and numerical projections preferred by the government. S&P, for example, believes the country’s cash deficit for 2022 shall be 9.4% instead of the government’s predicted 7.9%. It also insists that the cost of servicing the debt is 47% of all state revenue and not the government’s claimed 38%. It furthermore pegs GDP growth for 2022 at a more realistic 4.2% instead of the government’s bullish 5.8% estimate (2022 budget).

The issue here clearly is thus hardly one of S&P deciding to credit statistics and projections issued by the government whereas Moody’s refuses to follow suit. It is more about differing stresses and emphases, leading ultimately to different interpretations of contentious estimates, a very legitimate outcome of this type of analytical judgement exercised by different teams. In addition, Moody’s decided to be more cautious than S&P about whether the government can cut expenditure and raise as much revenue as it says it can but that caution arises from pessimism, or even scepticism, about projections that S&P also shares (as indicated above).

At any rate, rating agencies can have substantial methodological differences. Take for instance the all-important issue of “defaults”, the issue at the heart of all credit rating, the sovereign types included. Whilst Moody’s focuses on both the probability and severity of defaults, S&P, according to NYU Stern professor, Aswath Damodaran, focuses primarily on probability alone.

In such a context, it is hard to understand the government’s position that Moody’s did not consider critical data submitted or that it inaccurately analysed said submitted data.

To buttress its claim, the government goes hard on the competence of the Moody’s primary analyst. They question her familiarity with Ghana and dismisses her work as “desktop”. By so doing, the impression is created as if the Moody’s rating outcome is entirely the doing of one busybody through whom a set of arbitrary decisions has crystallised out of near-ignorance.

But every generic account of the rating process uses a chart like this that I picked up from a publication:

Source: S&P (McGraw Hill Publishing)

Like many of its ilk, the latest, now controversial, rating by Moody’s of Ghana’s credit status was solicited and paid for by Ghana. It went through a series of quality checks, including a full vote by a rating committee constituted by multiple individuals, some of whom are very familiar with Ghana. Ghana’s appeal was not assessed by the primary analyst but by the appeals subcommittee. Many of these steps are controls implemented since accusations very similar to what Ghana is now making were brought against the agencies by regulators in Europe nearly a decade ago.

On the charge of conclusions reached without quantitative analysis, Ghana assumes that the rating report finally issued to the ratee/issuer is the full dossier that backs the primary analyst’s recommendation. But this is not correct. Moody’s ESG analysis is nowadays, contrary to Ghana’s claims, backed by quantitative models.

Below is a generic overview of key parameters and variables that go into a rating decision.

And regarding the specific claim by the government that Moody’s ESG analytics contain no quantitative data, the below summary suggests otherwise.

But something being “quantitative” does not mean that it will not reflect varying “analytical judgement” or span a range of materiality. At the core of the ratings exercise is the stochastic attempt to affix hard numbers to fudgy predictions about the likelihood that when a debt comes due the borrower may/can or may not/cannot pay.

Thus, Ghana’s B- rating by S&P, for instance, reflects a belief, very crudely put, that in the course of the next 3 to 5 years, a default rate of between ~10 and ~16 percent is likely. That “10%” or “16%” is not anymore “solid” than the gut instinct of a shrewd investor, if one really decides to stretch the ontology of these exercises to breaking point. And, yet, by assigning numbers to certain measures and gauges, the work of many others in the chain of finance becomes easier as they now have manipulable inputs to use in many different permutations in many complex models.

But even a default is not the end of the chain. Other models, also linked to these same rating numbers, exist to tell the analyst the likely proportion of the debt at stake that is likely to be recovered during a default.

Source: Moody’s

Over time, these models are refined by empirical experience. Because defaults do happen, and when they do, everyone is afforded the opportunity to trace back and judge for themselves the degree to which these attempts at prophesy can be depended upon.

It bears emphasising, as related in the introductory discussion, that no serious investor is expected to use Sovereign Credit Ratings alone in their decision making. The existence of other measures of creditworthiness for state borrowers serves the additional role of providing proxies to check the accuracy of the analyses of rating analysts. One of these proxies operate by looking at how investors themselves are gauging the credit of a sovereign borrower on the open markets.

Prior to the latest downgrades by Fitch and Moody’s, there had been a steady stream of gloomy news suggesting clearly that investors were already pricing in higher repayment and default risks.

The appropriate response from the government, noting how Ghana had been catapulted into the global leagues of default prospects, would have been to address investor sentiment clearly with a budget focused on freeing up fiscal space, and with super realistic targets.

Instead, the government presented this:

A budget promising wild revenue expansion and even wilder expenditure growth. To sustain the credibility of the proposed ~43% primary revenue expansion (in one year), the government introduced a highly unpopular tax on all digital financial transactions without any plan for consensus in a hung parliament.

The markets went agog. Investors started deserting government of Ghana bonds as if they bore the plague.

It is hard to imagine what rational observers would have made of it if all this frenzy had attracted no action at all from any of the rating agencies, recalling how often they are berated for rating sovereigns too infrequently. Would that have helped the integrity of the rating process if investors, at whom ratings are targeted, are themselves giving strong hints that they fear a default yet rating agencies remain aloof?

This question provokes much of the discussion in the second half of this essay. But before we get into it, the statement below made on 24th March 2015 by Ghana’s Vice President, then in opposition, at Central University, near Accra, on the occasion of an earlier Moody’s rating change resonates with the discussion so far.

Vice President Bawumia, in a lecture at Central University near Accra on 24th March 2015

Bias

In its official response to the Moody’s downgrade, the government of Ghana alleges “institutional” and “anti-Africa” bias in the former’s conduct. In a separate communication, it praises S&P for being more progressive. The accusation here appears to be that some animus, unrelated to the mission of giving investors objective information about a borrower’s risk of default, is what is driving Moody’s behavior. There is a hint also that perhaps Moody’s has perverse incentives to downgrade rather than to upgrade.

This is curious since S&P, despite having issued, proportionately, the fewest downgrades worldwide, compared to the other two major agencies, has been the most enthusiastic downgrade-issuer in Africa at the peak of the pandemic.

Clearly, then, the government’s chastisement of Moody’s and praise for S&P cannot be the result of its concern about “anti-Africa” bias per se. It must be confined strictly to the specific Moody’s downgrade related to Ghana.

Regarding the issue of “pro-cyclicality” (the tendency of rating agencies to issue downgrades during a crisis thus possibly prolonging the same crisis responsible for the downgrade in the first place), Ghana is certainly far from being the only complainant. The problem though is that rating agencies are damned if they do and damned if they don’t. The biggest complaint against them has been that they have been too enthusiastic in issuing upgrades, a phenomenon known as “rating inflation“! According to one specie of this charge, the agencies’ reckless overoptimism in dishing out top ratings (or “investment grade” ratings) was exactly what caused the subprime mortgage crisis.

Indeed, the best framing of the perverse incentives and conflicts of interests in the rating industry is one where issuing downgrades actually loses the agencies money!

Whilst there is no denying the fact that African and Latin American countries have experienced more downgrades in this crisis, it is not clear how that could be the result of “systematic bias” when such bias is not aligned with the usual incentives of the agencies.

Economist Hippolyte Fofack suggests that this could be due to an overactive sense of “reputational preservation”. Or, put another way, the agencies are overreacting in order not to miss a default and thus lose the confidence of investors. Perhaps, the sense is that in Africa it is safer to be wrong pessimistically rather than optimistically. Considering though that the rating agencies are super keen to grow their business in Africa and any country downgrade tends to result in automatic downgrades of existing and prospective corporate clients whose credit is tied to their host countries, such a logic is hard to fully unpack.

To be truly objective, we must also consider reasons other than bias why Africa and Latin America suffer such an unfavourable skew of downgrades. Historically, Latin America has accounted for a significant majority of all global sovereign debt defaults and restructuring episodes. So much so that some of the countries there, like Argentina, have come to be perceived as serial defaulters (10 or so at last count). Between 1861 to 1920, 41 of the 58 defaults were in Latin America. Later in the century, African sovereigns joined the melee. In the 1980s “debt crisis” sparked by Mexico, Africa accounted for 34 of the 70 defaults, whilst Latin America followed with 29.

Usual suspects | The Economist
Source: Economist.com
Jonathan Fortun on Twitter: "Last week, we saw #Argentina's ninth default  in history. The #Latam region has been prone to defaults, especially during  the 80s https://t.co/77jEOn0iMR" / Twitter
Citing: Jonathan Fortun (2020)
Sovereign defaults from 1824 to 2004. Source: Borenzstein & Panizza (2008) | Figures quoted in the main text also draws on this work.

In short, there is a historical basis to suggest that African governments are risky business. Take the case of Mali, for instance, which lost its S&P rating in 2008, and then proceeded to default in 2012, thereby validating the action retrospectively.

To compound the historical default rate problem, African sovereigns have also historically accounted for a small proportion of syndicated loan and bond issuances worldwide. The small sample size of African issuance concentrates the effects of default and reduces the amount of information available for effective risk analysis, whilst also promoting overfitting to narrow risk models. In brief, the history of past defaults exert a greater Bayesian force on futuristic risk predictions when dealing with smaller data sets in view of the wider bounds of uncertainty.

Africa’s limited issuance compounds the effect of prior defaults on perception due to wide uncertainties. Chart credit: Presbitero et al (2016)

Even if we accept the arguments of systemic bias unrelated to fundamentals by Fofack and the Africa Peer Review Mechanism (APRM), which a few years ago set up a program to monitor the effect of credit rating actions on African countries, they should merely establish a uniform hurdle for the whole of Africa. What then are we to make of Ghana’s specific deteriorating circumstances?

The above factoid highlights the scale of the Ghana “special case” and makes a compelling pitch for the Fitch downgrade, for instance. It warns that Ghana may simply not be the right poster child for a case of rating agency bias.

Below is a list of countries, including four from Africa, that inhabited Moody’s B3 band alongside Ghana until the latter’s downgrade.

•Belarus

•Bosnia

•Cape Verde

•Moldova

•Mongolia

•Nicaragua

•Niger

•Pakistan

•St. Vincent & the Grenadines

•Ukraine

•Togo

•Swaziland

•Tajikistan

A quick check at the World Bank’s data store immediately underlines the revenue crisis that Moody’s so overwhelmingly emphasised:

Debt Servicing Costs as a share of total Revenues. Source: World Bank

Surely, with such a wide gap, can an argument not be made that whatever anti-Africa residual bias may be at play, Ghana’s situation deserves its own separate analysis?

Even Ghana’s current neighbourhood, the Caa1 peer group, does not offer a reassuring contrast.

Debt Servicing Costs as a share of total Revenues. Source: World Bank

All of these countries have severe crisis-induced challenges. COVID-19 has been less damaging in Africa, as Ghana itself stridently points out, so the pandemic has limited explanatory power in accounting for this strange divergence from peers, whereby Ghana pays so much more for paying interest and principal on its debt.

Even those countries in rating bands below Caa1 on the Moody’s scale do not have the same degree of debt service burden as Ghana’s.

Could something other than bias account for Ghana’s fiscal congestion? There is a good candidate.

The pace at which the country’s debt portfolio transformed to “predominantly commercial” was much too fast for the pace of the country’s development. Ghana seems to have “graduated from aid” way faster than its institutions and other capacities could catch up, saddling it with expensive, non-concessional, debt that its sub-Saharan lower middle income peers have yet to grapple with to the same degree.

To probe this a bit further, consider the two graphs below.

Is a debt crisis looming in Africa?
Debt to GDP in 2019 for selected African countries (Ghana’s ratio has significantly appreciated since 2019) Source: Brookings
The G20 External Interest Payments Moratorium Will Partly Ease African  Sovereign Debt Service Burdens | S&P Global Ratings

Compare Cape Verde and Mozambique to Ghana, and it becomes crystal clear how the debt composition can have an outsized impact on debt service burden. Whilst the former two countries borrow considerably from multilateral and bilateral sources, Ghana is increasingly being locked into pure commercial debt (also referred to as “external private” below) as a consequence of foreign policy choices (not investing enough attention into economic relations) and sheer fiscal desperation.

Cabo Verde: Staff Report for the 2018 Article IV Consultation—Debt  Sustainability Analysis in: IMF Staff Country Reports Volume 2018 Issue 104  (2018)
Cape Verde National Debt Portfolio Composition. Source: IMF
Who's Next to Fall? Moody's Points to Mozambique - The China Africa Project
Mozambique: National Debt Portfolio Composition. Source: China Africa Project
Ghana’s external financing was overwhelmingly concessional until it issued its first Eurobond in 2007.
By 2018, the shift to non-concessional had taken root in Ghana’s public debt portfolio. Source: Ministry of Finance.
By 2019, the dominance of commercial debt had become entrenched. Source: Ministry of Finance.

As proof that this is not merely a feature of Ghana’s lower middle income and higher GDP situation, Kenya’s debt composition is reproduced below as well.

Kenya: National Debt Portfolio Composition. Source: Central Bank of Kenya

Clearly, whatever grievances Ghana may have against the rating agencies, the country is still best served looking at more fundamental issues like its growing addiction to expensive commercial debt, well above its economic capacity to service.

Even if the growing calls for “Africa based” continental rating agencies and local regulations on ratings (cf. along these lines) were to be heeded, investors would still choose which ratings agencies to believe, forcing the matter out of the hands of governments. One even wonders how much commitment African governments and corporations will show were such entities to emerge, considering that only 6 countries have met some of the key milestones on the APRM’s roadmap for improving Africa’s sovereign rating situation.

It is up to Ghana’s leaders to decide when to move beyond the sense of persecution to focus on the deep structural issues of efficient spending and realistic tax mobilisation from an increasingly overburdened population.

From a serious policymaking point of view, the question as to whether “Ghana should go to the IMF” is not a complete, or particularly meaningful, one.

The IMF has many product lines, designed for different purposes, so the more appropriate questions are:

  • Should Ghana go for a/some IMF product(s)/program(s)?
  • If yes, which?
  • What are the costs and benefits?

In actual fact, two of the key functions of the IMF result in some pretty automatic programs for all 190 countries in the world that have signed the Bretton Woods Agreement.

It helps to recall that as the second world war was drawing to a close, and the increasingly triumphant American-led bloc was designing how they were going to govern the world they were about to inherit, economic “viruses” like hyperinflation and shortage of goods, both in the earlier interwar period and during the war, had already shown clearly how contagious they could be if countries do not work to stem their spread. Just like COVID-19.

Even a casual look at the four main functions of the IMF immediately points out which two should be considered “baseline” for all participating countries:

  • Surveillance of the international monetary system
  • Monitoring of members’ economic and financial policies
  • Provision of Fund resources to member countries in need
  • Delivery of technical assistance and financial services.

All countries in good standing and in “continuing communion” with the IMF cooperate with it on the first two functions. That fact can be translated to mean that every one of the 190 countries have a program of some sort with the IMF. The famous Article IV Consultations” that see the IMF hold bilateral discussions with member state governments almost every year is legally grounded in the surveillance and monitoring powers granted to the Fund under the 1944 Agreement.

So, presumably, when sensitivities emerge about whether Ghana should do business with the IMF, the issue is about the last two functions: borrowing money and/or being subjected to something akin to “technical supervision” of how the government of Ghana conducts its business of governing the economy. But even in respect of these two areas, as I have already stated above, the Fund is like a restaurant with a menu. There are varied dishes and a bit of room for some customization. Every program is designed on the back of a bespoke agreement. Much depends on a country’s negotiating skill.

It is a bit sad, for all the foregoing reasons, for the Finance Ministry to simply bundle everything IMF into the “bailout” category. By making the government’s IMF posture look like a patriotic avoidance of beggary and prostration before a foreign overlord, the Finance Ministry has completely distorted a proper discussion about “what Ghana can use the IMF for” strategically.

Such conduct follows in a line of worrying missteps in the ongoing management of Ghana’s revenue crisis. Sound analysis requires robust debate about all the options available to Ghana in these trying times. The “echo chamber” style of governance, often neglecting facts and data, is estranging the policy community in Accra, who no longer finds a counterpart in the government for genuinely robust policy engagement.

IMANI’s Professor Bopkin is something of a scholar of Ghana-IMF relations. He has helpfully produced the chart and table below to show that Ghana is the 5th most enthusiastic IMF customer in Africa. The country has approached the Fund 16 times since it joined in 1957. Since then, usually after every political transition or onset of major reforms, Ghana “goes” to the IMF.

Source: Godfred Bopkin

A deeper dive into the details shows that the aversion of some economic managers for IMF “conditionality” (terms or conditions associated with some of the IMF products) generally result from the type of products Ghana usually qualifies for. Ghana’s IMF engagements have often been characterised by long-term (3 to 5 years) and intrusive arrangements. But those are not the only offerings in the Fund’s stables.

Below is a quick overview of the most popular IMF products.

Source: IMF

Ghana’s practice over the years has been to wait until it is almost desperate before approaching the IMF. It therefore finds quite often that it only qualifies for programs like the Enhanced Credit Facility (ECF) and Extended Fund Facility (EFF) that deserve some of their reputation for reducing political flexibility. IMF programs requiring “structural reforms” of the economy do certainly signal to the market that a country is in a bit of a patch, a factor that can, in certain circumstances, affect borrowing on the commercial market. But there is a lot of nuance.

Kenya, for instance, is actually consuming three IMF products concurrently, all of the “bailout” variety – the EFF, ECF and the RCF. Through these facilities, it has secured nearly $2.2 billion in the last two years alone (twice what the e-Levy is expected to yield in the Finance Ministry’s most optimistic forecasts).

Ghana decided in 2020, like Nigeria, to take advantage of a global COVID-19 response mechanism and go for just an RCF, which is a fairly sweet-deal kind of bailout package with virtually no conditions. It received 738 million SDR, a type of IMF “currency” or unit of account (or roughly $1 billion), whilst Nigeria, because of its bigger economy, bagged 2.45 billion SDR. In both cases, the “program” expired soon after the money hit the account. Behaving just like a normal commercial loan, where the bank generally leaves the borrower to run their business as they see fit.

The above suggests that Ghana has no problem taking IMF money per se. But, like Nigeria, its twin on quite a number of fiscal measures nowadays, it does not appreciate any schoolmaster type tutelage. Which is fine, but a number of strategic issues arise, almost none of which has been properly debated.

Why is the Finance Ministry not going for a Rapid Finance Instrument (RFI) product like Namibia (2020), Senegal (2020), Bahamas (2020), and Costa Rica (2020)? Surely, these are not basket-case countries? The RFI product is also light-touch, with virtually no conditionality. South Africa successfully negotiated a $4.3 billion RFI facility in July 2020. Once the money hit the account, that was it, just like a normal commercial loan. The IMF actually do call these “outright loans”, and packages them with no conditions other than the standard interest and repayment terms.

There is a very important reason why this is a crucial point. IMF money is cheap. Many of its products are interest rate free until the borrowing country exceeds 187.5% of its quota, at which point the rate hits 2%.

Some products like the RCF are not only interest free regardless of amount, they also come with a grace period, typically 5 years during which the country pays nothing.

Compare that with a tranche in Ghana’s most recent Eurobond issuance in 2021. A 7-year $1 billion bond that attracted interest of 7.75%. In very crude terms, Ghana “loses” $57.5 million a year by borrowing a billion dollars in the commercial markets instead of from the IMF. Not chicken feed.

So, if IMF money is cheap, why can there be any concern at all? Well, as you may recall from above, there is also the issue of “national ego” and the more pragmatic concern about “market signaling”. Let’s talk about the latter.

Some observers erroneously believe that an IMF program somehow blocks a country from borrowing on the international markets until the program is complete. This is of course completely inaccurate. Kenya successfully issued a $1 billion Eurobond in June 2021 at a fairly decent rate of 6.3% just two months after securing a $2.34 billion IMF facility. In some ways, getting an IMF facility can actually improve the market’s perception of a country’s creditworthiness. Ghana tested this in 2015 when after enrolling in an IMF program in April it approached the market in October of the same year with a Eurobond issuance. Being at the height of the dumsor crisis, it didn’t get a good rate, but at least it sold.

The signaling issue is thus not open and shut. And, of course, if Ghana believes that national pride prevents it from appearing as if it needs an IMF stamp to be able to access the international capital markets, why does it not try for the big leagues?

There are a number of IMF programs that only the “big boys” go for. They include the Flexible Credit Line (FCL) and the Precautionary and Liquidity Line (PLL). In November 2021, Mexico went for a $50 billion FCL. Chile had earlier collected nearly $24 billion in May 2020. Panama went for a $2.7 billion PLL in the same timeframe. If Ghana genuinely believes that its fundamentals are solid, as the Finance Ministry has said, and that it cannot be seen, in the remotest, to be experiencing a balance of payments situation, then why not test the country’s mettle by going for a PLL or FCL, both of which are only granted, like a revolving overdraft, to countries with sound economic fundamentals so long as they continue to meet the eligibility criteria? Unless, of course, things are not that rosy, and the Ministry is afraid that Ghana may not be eligible for anything other than plain-vanilla balance of payments support.

The last issue to consider is the “competition” between e-levy and an IMF program. Some commentators are promoting a “return” to the IMF as an alternative to going ahead with the e-Levy. This only makes sense if, as some suspect, the government’s intention of passing the e-levy is to convince the market to become more receptive to another Eurobond issuance. If that is indeed the case, then given the political costs of the e-Levy strategy, an IMF program would make more sense, especially when you reflect on the earlier point about how countries have often sequenced their Eurobond campaigns right after entering an IMF program.

If, on the other hand, the government is genuine in its rhetoric about cutting down borrowing, from whichever source, however cheap or well-priced, and getting Ghanaians to fund the greater part of their own government, regardless of the political and economic repercussions, then of course the e-Levy is a superior idea. After all, cheap as they may be, IMF facilities (except the relatively smaller “Catastrophe containment” grants) still need to be paid back.

Somehow, many in the policy community are not too convinced that the government’s recent actions are not merely a show for investors to reopen the doors to the sweet nirvana of Eurobond issuances. Early indications of the government’s domestic borrowing plans do not really corroborate the notion of significantly cutting its overall borrowing appetite.

Whatever be the fact, there is one certainty: neither an IMF program nor the e-Levy can truly address the fundamental issues that have led Ghana into the current ditch: weak productivity in the private sector exacerbated by a high cost of doing business, which in turn is compounded by wasteful spending habits in a fast-bloating state sector.

As I have argued elsewhere, the Ghanaian government has over the last decade become overbearingly expensive to run, with too many price-inflated projects competing for the attention of the politically connected. This has led to a neglect of quality policymaking that can actually boost efficient private sector output. Without the latter, GDP growth will not translate into jobs, innovation, and taxes.

In a way, talk about both e-Levy and IMF-return are distractions from the most critical issues at stake. Even if, depending on the real intentions of the government, one of the two is bound to be superior to the other.

The recent debate over the proposed e-levy tax has put the spotlight very intensely on the subject of taxpaying attitudes in Ghana.

Ghanaian politicians regularly chastise their fellow citizens for being entitled windbags who complain constantly about every minute failing of the government in providing services and amenities. Lamentably, in the view of the politicians, these same Ghanaians slink down their holes when asked to pay their fair share of tax for the government to obtain the resources needed to meet all their noisy demands.

One former Minister is categorical: Ghanaians just don’t like paying taxes. An activist of the ruling party implicates “dirty politicking” in this attitude of rampant tax dislike and evasion. According to the country’s serving Finance Minister, only 8.2% of working Ghanaians pay tax.

As far as some politicians would have it, a vast proportion of the population is literally made up of “tax thieves” blatantly stealing from the state that which morally and legally belongs to it.

Yet many Ghanaians disagree with the politicians about this alleged aversion to tax.

72% of the country’s residents tell AfroBarometer (2019) that they will gladly pay more tax to support the government to deliver on its mandate.

Source: AfroBarometer (2019)

Ghanaians express a far greater willingness to back Government’s tax-levying power than citizens of most other African countries according to AfroBarometer surveys. Even if 10 years of watching politicians pour money down the drain has dampened that enthusiasm a bit (from 90% down to 79%).

Source: AfroBarometer (2020). Image Credit: Isbell & Olan’g (2021)

Of course, there is no country on Earth where more vigilance and diligence in tax collection would not yield more government revenue.

Even in fabulously high-solidarity and egalitarian Scandinavia, Matthew Yglesias estimates that the top 1% of income earners (i.e. rich people) evade 30% of true tax due.

Source: Matthew Yglesias (2017)

Every government therefore can be expected to try to do more to prevent tax evasion and to minimise unethical tax avoidance. Where we have a problem is when attempts are made to suggest that the situation in Ghana with regard to taxation is particularly dire because Ghanaians are exceptionally unscrupulous and anti-tax. I disagree simply for the reason that the data simply doesn’t support the claims.

Seriously the Finance Minister ought to fire some of his research assistants. Throughout this e-levy debate, they have failed to furnish him with sound data and advice. The claim, made in the international press no less, that only 8.2% of working Ghanaians pay tax is extremely distressing. There is no grounding in fact for this belief.

According to the Ghana Revenue Authority, which reports to the Finance Ministry, 6.6 million Ghanaians file taxes. Based on the provisional results of the latest census in Ghana (2020, postponed to 2021), the true labour force (economically active citizens) of Ghana is just a little under ten million. Thus, depending on whom you believe, the number of tax paying individuals is either a high of 57% of the population (not adjusting for the “economically active but currently unemployed” to allow for easier subsequent comparisons across countries), according to the Tax Authorities. Even if we were to use the Finance Ministry’s own curious compliance data, the resulting figure would be 20.5%. Not 8.2%.

Source: Ghana Statistical Service (2021)

The Finance Ministry does have a point, however, when it complains that the amount of money made in the economy that goes to the government (approximated by the metric known as “tax revenue/GDP”) is smaller than in more developed countries around the world. I have addressed that issue more fully here.

There are many caveats to bear in mind when comparing how much other governments are able to squeeze out of their economy, compared to Ghana’s, however.

First, the historical evidence does not suggest that the government is able to squeeze more out of the economy when Ghanaians suddenly become upright and scrupulous in between epochs in history. It suggests, rather, that how much government takes in taxes reflects shifts in the relation between the state and business.

In the 1960s when the state run large parts of the economy, and the private sector was young, it was natural for the citizens to pay less tax. As Ghana liberalised, and private sector activity grew, more tax was paid. Then the Acheampong government came in and nationalised large swathes of the economy (for example, the government took 55% in major mines and factories hitherto majority owned by the private sector). State-control took the tax-to-GDP ratio to its lowest in 1981 – 1982, in the heat of the last attempt at a socialist revolution. Tax to GDP continued to climb as the emphasis shifted to private sector empowerment, until the discovery of Eurobonds in 2007, the onset of oil production in 2010, and the start of the gold and cocoa bull super-cycle (in both price and production terms) that is still underway.

Source: World Bank, as cited by Matthew Ocran (2018)

Second, as I have hinted elsewhere, the considerable boom in oil, gold and cocoa prices and production over the last decade, backed by copious quantities of eurobonds, have considerably shifted the momentum in the economy towards the government-controlled center, to the detriment somewhat of the private sector. Without considerable private sector productivity growth, it is hard for taxes, versus other forms of revenue like royalties, to grow in line with or faster than GDP.

Third, the structural issues in the economy are also reflected by the sheer number of persons in various sectors where income is highly irregular or is taxed in ways making personal and corporate/enterprise taxation completely incongruous. For instance, 75% of Ghanaians farm, fish, “buy and sell” or engage in low-level artisanship. Farmers sell food with high implied non-state subsidies created by a massive differential between farmgate and urban retail prices. Cash crop farmers must actually sell at the government-dictated price so that the state can take a margin. These are all implicit taxes.

Source: Ghana Statistical Service (2021)

Furthermore, due to high domestic borrowing, central government financing (money printing, among others), and other “money emission” factors, coupled with a tendency to inflate away the value of the money it has borrowed from citizens, the government of Ghana enjoys high rates of seigniorage and inflation tax. As noted by many commentators in this sub-discipline, inflation tax has a habit of flipping at an inflection point and yielding negative returns. The chickens may have simply come home to roost in Ghana.

Not surprising then that two West African countries whose governments consistently lament their low tax to GDP ratios, Ghana and Nigeria, have also historically benefited the most from inflation tax.

Source: George Anang (2020)

In short, letting the data do the talking reveals the startling truth, in deep contrast with the standard commentary in Ghanaian policy circles, that Ghana actually outperforms many peers and superiors when it comes to registering people to tax them.

Author’s calculations from public data.

A person holding brief for the Finance Ministry is likely at this juncture to push back. Mobilising people into the tax net does not by itself mean that people are still not cheating the treasury. True, but that is not a peculiar Ghanaian issue. At any rate, politicians have consistently been twisting the truth by making it look like the “tax net” in Ghana is not “capturing” millions of shady characters with sackfuls of cash under their mattresses. What the above clearly shows is that over the last decade or so we have brought millions into that tax net. At least, they are visible.

We can now tackle the question of whether beyond the structural reasons (inflation, natural resource dependency, debt-fueled-but-low-productivity-growth etc.) I have provided above and elsewhere, there is strong reason to believe that under-declaration is particularly rampant in Ghana.

Let me get this out of the way: there is no reason to doubt that there is considerable under-declaration. As a brilliant friend of mine at a top consultancy likes to remind me, the Ghana Revenue Authority meets more of their stretched targets than they miss. Somehow, in some years, they manage to find the money lurking somewhere in the economy regardless how aggressive the target is set. The question we need to ask though is: how widespread is this practice, really?

If it turns out that we have a few ultra-wealthy cabals doing most of the under-declaration, as opposed to the general masses of the long-suffering people of Ghana, then the key thesis of this essay holds: there is no evidence that Ghanaians generally don’t like taxes (anymore than the average human), aren’t paying taxes or are particularly unscrupulous about taxes. It is entirely up to the political class to rein in their cronies in the cabals, currently enjoying political protection, hoarding the wealth and refusing to chip into the national kitty.

To make progress along that line of inquiry, it is important to unpack the recent decline of revenues-to-GDP from 22% in 2005 (per Matthew Ocran’s calculations) to roughly 14% today (beyond the structural factors I have hinted at above, that is). What really in the tax basket has been leaking out?

A 2021 report by the UK’s Institute of Fiscal Studies (IFS) reveals much. (As a side note, it is curious that the IFS in its home base of Britain is regularly counted on to produce independent revenue forecasts for proposed taxes, like our e-levy here; but in Ghana, the government routinely refuses to encourage such rigorous tax analysis and forecasts before introducing its tax measures.)

Let me quote the IFS:

Tax collections on imported goods have become far less important in the revenue mix, though they remain significant: 30% of overall tax revenues were collected on imported goods in 2019 (including VAT on imported products), compared with 54% in 2000. The contribution of import duties specifically to total tax revenue declined from 18% in 2000 to a low of 12% in 2019.

And:

Much of the growth in Ghana’s GRA-collected tax revenues since 2000 has come from increased corporate and personal income tax and VAT and similar taxes, though revenue growth from the latter two has stagnated more recently. These taxes made up over 70% of total collections in 2019 – up from 57% in 2000.

Trends in tax revenue composition from 2000 to 2019. Source: IFS (2021)

In short, whilst Ghanaians have been paying lots more in income taxes on both their personal and business income, the ports have been leaking.

There are two main ways to account for those leakages. First, graft. The Vice President has described the corruption at the Port as “unbearable”. Second, tariff policy. The average tariff in Ghana is about 92.5% of the value of the item being imported. Essentially, on average, importers pay almost the equivalent of the price of the good being imported to the government as tax.

The high tariffs are a clear incentive to bribe one’s way through if one has the means, so the two main causes are interconnected. But given the implied import-substitution and infant-industry protection bent in current Ghanaian trade policy, a significant reduction in tariffs is unpalatable to almost everyone in the policy establishment today.

Source: World Trade Organisation

Below is a graph showing how Ghana compares with other countries its policymakers and other commentators like to compare the country to. Note the presence of both liberal and “developmental-statist” economies on the list. It is clear that imports have become so expensive over time that the growth in their volumes has been dropping, leading to a drop in revenue from import duties and other port-levied taxes. The trend is furthermore being exacerbated by bribery and corruption.

Data Source: World Trade Organisation

The truth is: when you look at the part of the revenue mix obtained purely through GRA collections (as opposed to the broader measures of “revenue” used elsewhere in this essay), what we have is a stagnation due to a declining trend in import tax collection, not fully offset by a growth in personal and corporate income tax collection!

Source: Institute of Fiscal Studies (2021)

The alarmism by politicians, targeted at the conduct of citizens, is completely unwarranted and the media must stop enabling it and start asking hard questions.

Of course, politicians will still push back. They will say things like: people are doing many side jobs and not declaring income so the rise in income tax revenue is purely as a result of the burden falling on a smaller and smaller subset of taxpayers.

Whilst there is no doubt about there being a skewed burden that must be corrected over time, this has become something of a universal burden brought about by a growing concentration of wealth.

In the US, for instance, the top 20% pay nearly 70% of the taxes.

US tax burden distribution. Credit: Tax Policy Center (2020); as cited by the Peter G. Peterson Foundation

In Singapore, the top 20% of tax payers account for 91% of total tax revenue. Only 7.8% of Malaysia’s 1.25 million registered enterprises pay any tax at all.

So, if a Finance Ministry Mandarin is tempted to argue that Ghana is yet to see any significant benefit from boosting registration of taxpayers, they should temper their sorrow by checking with India where aggressive taxpayer mobilisation saw filings grow to 57.8 million. Yet, those actually owing the Indian government dropped to 14.6 million (coupled with massive numbers of Indians now entitled to tax rebates/credits). To repeat: over a period of 8 years, the number of persons registered for tax but who don’t have enough income to pay any tax to the government increased by 300% in India.

Times of India (2020)

It is not a one-way street. The more government pries into people’s incomes, the more likely some of them are going to realise that they don’t actually get enough help from the government and therefore it is the government that in fact owes them money.

In 2019, 37% of even American taxpayers assessed as owing tax didn’t have the money to pay. In recent years, Indonesia has seen spells where less than 1 million citizens are able to pay tax out of a population of over 270 million.

The last rebuttal to any Finance Ministry Mandarin pressing the old charge of ruling a nation of tax delinquents is wastage.

I am not even talking about the incongruity of tax aggressiveness in a context of high tax burden in a country where many are struggling to make ends meet and therefore only a shrinking number of people shall increasingly be responsible for most of the tax paid, as is the case elsewhere, and how therefore any wastage of taxes may impact tax morale. I am talking specifically about why the government feels that Ghanaians pay far less tax than they can afford when the data says otherwise. And when Ghanaians are indeed subject to some of the highest tariffs and consumption taxes in the world. Just check out comparative VAT rates below.

Source: African Tax Administration Forum (2020)

Or compare aviation taxes in Ghana with what pertains in its rival “aviation hub” aspirants in the region. Ghana charges double the rate in Rwanda and Kenya, and nearly four times the Ethiopian amount. Yet, it insists that it is on course to eclipse them as a continental aviation nexus. Talk of hubris.

Airport Passenger Charges in selected African countries. Source: AirInsight (2021)

No I don’t mean any of that at all. What I mean by “waste” here is the sheer cost of doing government in Ghana. Ghanaian state-subsidised football administrators regularly requests, and often receives very close, to three times the amount their Nigerian and Kenyan counterparts get, and ten times what their Malawian and Gambian compatriots receive.

When Ghana builds bridges or airports, a careful examination of the costs (including borrowing costs) would inevitably show that, pound for pound, they tend to be twice or thrice more expensive. This is the case when you compare Kotoka Terminal 3 in Accra with the bigger 3rd terminal at Bole International in Ethiopia; or the planned Volta River Bridge in Ghana with the Makupa bridge in Kenya.

In short, the government of Ghana has been harbouring this mistaken notion of Ghanaians being inveterate, incorrigible, tax thieves because it feels that it is being squeezed. However the “squeezing” is entirely the doing of the state, and the victim is actually the people.