This is a long-read. The short version is as follows.

  • Two sectors of Ghana’s economy are specifically identified in the country’s recent IMF program for critical attention because of their contribution to the country’s debt debacle: energy and cocoa. The story of Tema Oil Refinery exemplifies what has gone wrong with Ghana’s energy sector.
  • Tema Oil Refinery has big problems that require hundreds of millions of dollars and sophisticated investors to fix.
  • Such money and investors will not come unless the country’s leaders are genuinely interested in an above-board transaction, and are willing to spend the taxes they have collected in TOR’s name to clear the muck, and give serious investors a fresh chance. It is clear that they are not. They are only interested in a deal that will keep their cronies positioned at vantage points.
  • So, for many years, various shady deals and shadowy investors have been paraded as saviours of TOR, of which the latest are Decimal, Legacy and Torentco.
  • None of these entities have what it takes to truly revamp TOR. The opacity, underhand dealings, shadiness and murkiness risk becoming a breeding ground for criminal and underworld activities, destroying what little reputational value remains in TOR. Already, there are fears that some of the partners proposed for the Torentco transaction may be embroiled in sanctions-busting activities.
  • An open and competitive process should be pursued to bring fresh management not beholden to any political elite factions. The unencumbered management should be empowered to run a totally above-board process to attract investors with deep pockets. Accompanying energy sector reforms, strict use of energy taxes for their intended purpose, and strict price deregulation would need to be in place to assure such investors.
  • Any shady leasing of assets to shadowy operators will lead to the compounding of TOR’s debts, continued obsolescence of TOR’s infrastructure, and an eventual reckoning in which the company’s situation becomes irredeemable and unsalvageable.

Born of Big Dreams

The Tema Oil Refinery (TOR), located in the country’s coastal industrial base, is a product of Ghana’s post-colonial industrialisation drive that saw the first government after independence engage international investors as strategic partners in a number of vital industrial projects.

The same process that brought foreign financiers and engineers to build, own and run TOR in 1963 also gave Ghana the Volta Aluminum Company (VALCO), Akosombo hydroelectric dam, and Ghacem, still among the most important anchors of Ghana’s industrialisation dream.

Osagyefo Kwame Nkrumah with ENI executives ahead of TOR’s opening.
Source: Parditey (2017)

A wave of nationalisation, however, swept the country in the 1970s, and by April 1977, TOR was fully state-owned. It is a sad testimony to the perennial decline of state capacity in Ghana that since then TOR has lurched from one crisis to another.

The Rude Awakening

From the point of nationalisation onwards, the refinery did not see any serious renovation or capacity boost until 1989. Even then it was widely recognised that capacity expansion and technological retrofitting and upgrading were essential to the ongoing viability of TOR. But then again a long list of state-run industries were in the same boat.

In 1963, when the seven year development plan was launched, several state-run industrial concerns were placed on a list for capacity expansion and upgrading by 1970 at a cost of £1.6 million ($4.5 million), or the equivalent of $45 million in today’s money. By 1992, not a single one of them had seen any serious investment for transformation.

Sample of priority factories targeted for capacity upgrades in the 1963 7-Year Ghana Development Plan.

Halting steps to redemption

A $200 million loan from Korean investors enabled the expansion of the Tema Oil Refinery (TOR) to its current 45,000 barrels per day capacity by the late 1990s.

When a new government took over from the Rawlings government in 2000, a long courtship with Korean government and private entities, especially Samsung, ensued. At stake was a fresh loan of $230 million for the long anticipated expansion and addition of modern modules to handle a wider range of crude oil types. TOR’s precious Residual Fluid Catalytic Cracking Unit (RFCC) was eventually financed by Samsung from that loan package and commissioned in November 2002. By this point, the refinery had racked up debts of over $300 million.

In 2003, the government instituted new taxes to enable the refinery pay off its debts and complete the planned expansion works.

Energy Analyst, Dramani Bukari, estimates that after these expensive technical works, plant utilisation (how much of a factory’s theoretical capacity can be practically used) at TOR ramped up to a high of 80% before beginning a disorganised descent until it hit a rock-bottom of 19% in 2009.

Tracking capacity improvements after retrofitting at TOR. Chart source: Bukari (2013)

The Imperative of Expansion & the Debt Trap

Clearly, despite absorbing tens of millions of dollars in investment for technical transformation, TOR was still struggling with viability issues, increasing the urgency for a more radical expansion program to improve unit economics. In response, officials mooted the prospect of adding 100,000 barrels per day of refining capacity at the cost of ~$200 million.

Meanwhile, the debts continued to mount. And politicians continued to interfere in the running of the organisation, only ever appointing cronies and allies to all Board and senior management positions. Management policies grew even more arbitrary. Underfunded government subsidies ensured that the same year TOR processed the most crude, 2004, it also recorded the highest losses for the decade. Samsung, tired of all this drama, withdrew its interest as a prospective strategic partner.

The bumbling continued. Despite collecting $500 million in TOR-focused taxes by 2008, the government refused to invest any significant amount of money into the facility, even as it toured the globe looking for so-called “partners”. Unwilling to fully let go of a juicy gravy train, open and competitive privatisation was never really placed on the cards.

So, by 2010, the debt had ballooned to $1.4 billion. Confronted with the prospect of total shutdown and hostile creditor actions, the government started to make some payments. GCB, a state-controlled bank and a captive TOR creditor, was on the verge of being pulled down with TOR forcing the government to arrange a $316 million repayment.

That, however, did not really stop the refinery from shutting down.

Facing a loss of public credibility, the government doubled down on efforts to address the fundamental issues.

By 2015, further payments had brought the debt down to a little under $750 million. Yet, frequent announcements of “strategic partnerships” and financing deals had yet to bear fruit. In one particularly striking example, the government announced a $900 million financing package in 2012 to tackle the crushing working capital constraints of the company.

But nothing of note happened.

God bless our hustle

Months later, in 2013, TOR was back to hustling for crude parcels around the world to keep the refinery in business.

The culture of hustling for crude and bits and pieces of working capital was fast becoming the defining hallmark of the TOR story. In 2011, a bizarre drama unfolded in which fake companies popped up from the undergrowth to claim ownership of crude oil delivered to TOR, forcing the intelligence services to step in.

Meanwhile, TOR’s throughput was now the lowest among peer refineries in the subregion.

Capacity benchmarking of selected petroleum refineries in the subregion. Source: Katsouris & Sayne (2013)

Turning Inwards

In 2015, another effort was made to raise domestic funding to tackle the TOR debt and capacity juggernaut, but this time in concert with other energy sector challenges. The Energy Sector Levies Act (ESLA) was passed and, in 2017, a special purpose vehicle created to transform taxes into a more effective financing mechanism for clearing legacy debts and freeing energy sector corporations like TOR to embark on structural reform. Recall that by this time, TOR’s debt was around $750 million.

The ESLA SPV raised a total of $1.6 billion in the three years following setup. The government simply expanded its general spending and devoted nothing to the structural transformation of the refinery. It did however clear a significant portion of TOR’s legacy debt, which thus dropped to $460 million.

TOR revenues, on the other hand, tumbled from a peak of ~$100 million in the last decade to less than $25 million in 2020. Accumulated losses over the period exceed a billion dollars.

Today, our sources tell us that TOR’s books have not been properly audited but provisional figures show a debt of nearly $450 million. The spate of shutdowns have intensified over the years leading to a state of near mothballing, with refining activities at a standstill for an unprecedented 26 months now. The organisation is on its fifth or so Chief Executive in five years. The center cannot seem to hold.

The Shadowy Duo: Decimal & Torentco

It is against this background that civil society activists and analysts received with dismay news of the submission to the Public Procurement Authority (PPA) of a proposal by the TOR management to lease the main assets of the refinery to a shadowy entity called Torentco Asset Management (“Torentco”) for six years. To fully appreciate the concerns of analysts over this matter, it is important to recap the key issues.

  • It has long been recognised that TOR’s working capital, refurbishment, and debt management require capital in the range of between $500 million and $1 billion to begin to get a handle on the structural viability of the refinery. And that considerable capacity expansion is essential. In fact, the last CEO at TOR to have had anything remotely approaching a normal tenure insisted that $3.5 billion was required to start from a clean slate.
  • It is also widely accepted that TOR’s simple initial design as a hydroskimming plant requires a substantial overhaul to lower conversion losses and sustain plant utilisation over the medium term. For example, TOR’s current configuration is optimised for heavy distillates hence the criticality of the RFCC. Gasoil and atmospheric residue constitute more than 60% of its output. To produce more of the products most required on the domestic market such as petrol (Premium Motor Spirit) and diesel, a robust RFCC is vital in TOR’s current configuration.
  • TOR’s adverse managerial and operational history makes it difficult to raise money on good terms, obtain feedstock (raw crude) on reasonable credit, and to respond to market dynamics in a commercially competitive manner.

For nearly two years now, the government and TOR have been promising a major deal with the features needed to tackle these serious challenges. Things seemed to have come to a head when in another of the frequent managerial reshuffles TOR has become famous for, a new CEO took the helm last year and promptly announced a strategic partnership with a little known Kenyan entity called Decimal Capital to turn the refinery around.

Decimal was clearly introduced into the mix by another short-lived CEO of TOR, who was removed in April 2020 following a civil charge by the United States government for various alleged offences. He was subsequently criminally indicted under the Foreign Corrupt Practices Act. Many analysts believe that Decimal was still being teleguided by this former CEO.

This week marks one year since TOR’s management announced to the country that the Minister of Energy has approved a strategic partnership agreement between TOR and Decimal for TOR to lease its infrastructure for the use of Decimal in order to resume stalled refining operations. Then everything went quiet. At the time, Vitol was mentioned as the commodity financier. That is to say, Vitol, a large Swiss energy trader, was to provide the raw crude for TOR to refine on flexible financial terms. Nothing was heard of that either.

It is now apparent that there are factions in the highest echelons of the Ghanaian government, each with its own preferred crony arrangement. Because at the same time the nation was being told that TOR was moving forward with Decimal, another shadowy group calling itself Torentco was also engaged in brisk negotiations with TOR to lease the very same assets supposedly being transferred to Decimal. As is typical with these crony arrangements, no information was made public about any of these developments even though TOR is a 100% state-owned enterprise under an active tax-supported bailout.

Torentco Ascending

After months of factional infighting, during which more shadowy groups, like a certain “Legacy Capital” based in Dubai and led by a certain Vladimir Palikhata (an occasional name in Russian whitecollar crime sagas) burst onto the scene, the Torentco group suddenly gained an upper hand.

On 13th June, 2023, TOR, presumably tired of dragging its feet, sent an updated proposal to lease certain of its vital assets to the Public Procurement Authority for ratification on a single-sourcing basis, the preferred procurement model when shady things are underfoot.

Torentco’s offer to TOR can be summed up as follows.

▪︎ TOR’s main productive assets will be leased to Torentco for 6 years. Excluded from the transaction are TOR’s creaky laboratory; TOR’s stake in the company (GPMS) that manages the moorings for marine vessels discharging cargo into TOR and the facilities of other bulk oil operators; TOR’s permits to use certain infrastructure it does not own but need for its work (“rights of way” and “tie-ins”); and the RFCC, a secondary plant at TOR that can extract valuable products from leftovers of products made in the primary plant, the CDU. Whilst these auxiliary facilities will not be leased out, and thus TOR will preserve the associated revenues of about $12 million, Torentco will nonetheless have access to them as and when necessary.
▪︎ Torentco is allowed to refine up to 8 million barrels of oil a year by paying $1 million every year as annual rent.
▪︎ There is also an “additional rent” amount of $1.067 million per month. The strategy behind this rent split is obscure, but Torentco argues that it is necessary to protect TOR’s lease revenue from creditors, who apparently have been trying to to attack TOR’s assets for monies owed.
▪︎ If Torentco refines more than 8 million barrels, it pays $0.5 for each extra barrel.
▪︎ So, if Torentco stretches the refinery to its full limit and refines 16.5 million barrels, it pays $17.2 million in rent.
▪︎ Torentco will invest $22 million to revamp the refinery. On top of that, it will assume responsibility for clearing provident fund arrears up to the tune of $2.5 million. To date, TOR has been struggling to pay its SSNIT and pension liabilities.
▪︎ Torentco will furthermore pay $800,000 into a reserve fund plus $0.4 per barrel refined to cover maintenance expenses, while assuming responsibility for insurance and utility payments. At maximum production limit, Torentco’s maintenance commitments will cost about $7.4 million and insurance expenses will hit about $6 million per annum.

Beyond the core offer, Torentco has sought to bolster its proposal by name dropping certain partners. Vitol has been mentioned again as a potential bulk buyer of the refined output. Two local engineering and construction companies have been introduced as technical contractors for the refurbishment activities: ENTTP and Litwin Engineering. Rounding up the list of proposed consortium members is a commodity supplier called Pontus.

The PPA Board convened a meeting on 15th June, 2023, to deliberate on the updated proposal. Rather than comment substantially on the Torentco offer based on its due diligence findings, the PPA simply took a leaf from the book of the biblical Pontius Pilate when he sent the condemned Christ to Herod Antipas, King of Galilee. The PPA says that it took note of what it saw as “partnership” elements in the proposed transaction and thus chose to refer the matter to the Ministry of Finance, which under Act 1039 is responsible for scrutinising Public Private Partnerships.

But there was much more to say about the transaction beyond administrative formalities.

What is wrong with the Torentco Deal?

The Torentco deal is anchored on the fact that TOR today is idle for all the reasons and factors this essay has traversed. Given that TOR has been mismanaged into a state of near comatose and its own sole shareholder – the government of Ghana – has publicly announced that most of its frontline management are thieves, Torentco reckons that no serious investor will come in.

In these circumstances, Torentco’s mastermind, Michael Darko, ably assisted by his trusted sidekick, George Antwi, and with the indicted former CEO lurking in the shadows, are completely convinced that TOR has nothing to lose if it carves out its productive assets and turn them over to a non-tainted operator to make a bit of money. Money that will be shared with TOR even as the latter continues to waffle around awaiting its salvation.

Despite the cosmetic plausibility, there is serious short-sightedness in such a proposal.

  • First, neither Mr. Darko nor Torentco ring any bells in the energy investment corridors at home and abroad. If the whole strategy hinges merely on hiving off TOR’s assets into a separate legal entity, then TOR might as well create such a subsidiary itself.
  • The issues of where working capital, a sustainable commodity supply contract, and credible offtaking arrangements will come from can only be properly addressed by a credible entity. TOR may not be one, but neither is Torentco.
  • None of the companies Torentco has mentioned, except Vitol, ring any bells either. And our sources in Vitol say no conclusive offer is on the table. ENTTP and Litwin are not the heavyweights needed to turn around a contraption as stuck in the swamps as TOR.
  • ENTTP’s Christopher Hesse-Tetteh has been mentioned in connection with some general logistics and construction works, but nothing remotely approaching the complexity and technical risk presented by TOR.
  • Litwin Engineering claims to be based in Switzerland but has zero footprint in the energy consulting and financing space except its own scanty self-attestations online.
  • The most worrying aspect of the Torentco Consortium’s credentials for this activity is its proposed source of crude oil feedstock. After all, it is inability to raise letters of credit to import crude to refine and pay off creditors that began TOR’s spiral of death. Only two entities fit the Pontus profile mentioned in Torentco’s proposals. Both are embroiled in shady oil trading situations, touching on potential fraud, sanctions busting, and even terrorism. They are Pontus Trading of Dubai and Pontus Navigation of the Marshall islands.

The last issue goes to the heart of the credibility, transparency, and above-board concerns that civil society activists and analysts have about the whole leasing transaction. Actors lacking mainstream commercial credibility more easily fall prey to the wiles of the criminal and/or unethical underworld.

The bare economics are wobbly

Some have tried to position the proposed Torentco leasing strategy as nothing more than a variant of the tolling transactions TOR has had with the likes of Total, Woodfields and Vitol over the years. Such analysis is wrongheaded.

A leasing contract involves a handover of substantial control. A tolling contract does not. Once Torentco or another party gets hold of the refinery, it retains every right to secure its own commodity supplies, refine, sell, and pay back the supplier the cost of the crude oil, pocketing the full margin. There is nothing that says that such a company must limit itself to tolling deals with well known commodity suppliers like Total and/or sell to well known traders like Vitol. It can seek to maximise its margins by obtaining crude in grey markets to refine and sell back into the same grey markets, especially in countries with porous supply chain security that cannot police even unsophisticated gold smugglers.

The sheer amount of money (crude calculations suggest between $250 million and $750 million in the TOR case depending on precise mix of refined products and plant utililisation) that can be made in today’s geopolitically fraught oil market if one can get shady oil to refine, sell and pay back over an extended credit period would be enough to corrupt multiple safeguarding institutions and overwhelm the country’s risk management capacity.

Per current proposal, there’s no scenario in which TOR comes up tops

But even if we were to take Torentco’s word for it that they intend to play within the traditional tight-margin refining space by simply cutting through TOR’s bureaucratic inertia and ringfencing the toxic legacy constraints such as debts and bad commercial reputation, the offer as it currently stand does not add up.

The actual guaranteed payments (direct and indirect) to TOR are equivalent to less than $1.5 per barrel, and thus lower than the tolling fee offers previous CEOs refused to accept. It is far lower than the $4.5 that those CEOs believe is the minimum sensible rate given the context.

In a situation where tightness in global oil supply, ahead of the electric car transition in future years, starts to provide support for $15 per barrel average refining margins in the next five years, a refinery operator would be in a good place to secure its commodity supply by whatever means. In such a world, the incentives are heavily skewed in Torentco’s favour.

Whilst the adverse supply conditions that existed in 2022 have abated somewhat, there are signs of recurrent supply trends supporting margin inflation over the medium-term.
Source: Statista

The risks, on the other hand, weigh heavily against TOR. Its debts will be compounding, since it will not receive anything close to what it needs to service them whilst its productive assets remain encumbered for 6 years. Let us be clear. The kinds of money Torentco is mentioning – $22 million upfront investment, ~$13 million in annual rent etc. – do not even begin to make a dent in TOR’s real financing needs. As recounted in the early sections of this essay, TOR requires a solution that yields many times more.

And, should the unthinkable happen, and the refinery find itself enmeshed in some sanctions-evading scandal because of shady suppliers or consortium members, and gets blacklisted, the burden of salvaging the tattered remains of its reputation will also fall on TOR, not Torentco.

There is no point in mincing words here. The Torentco deal as it stands now is not in TOR’s interest. It is absolutely not in Ghana’s interest. A whole load of transparency, open scrutiny, and reworking is required before it can even be taken into serious consideration.

Two days ago, the European Parliament came to an internal consensus on a draft version of the European Union (EU) Artificial Intelligence (AI) Act. Following this development, horse-trading will intensify among the European institutions to settle on a final legal text. This new phase of consensus building between the EU Parliament and the EU Council (with the EU Commission shuttling in-between) – the so-called trilogues – is an elaborate policymaking dance rivalled only by the elegant denseness of the Roman Curia’s Praedicate Evangelium.

The EU’s risk-based approach to AI regulation.
Source: EU Commission

The hope of the proponents of the EU AI Act is that it will become a global blueprint for how to realise the full benefits of AI without losing sight, and eventually guard, of its potential downside risks. The tendency of other parts of the world to follow EU thinking on how to regulate complex technical areas has led to a term: “Brussels Effect“.

Scenarios of how the EU AI Act will influence global corporate conduct.
Source: Siegmann & Anderljung (2022)

No context has experienced the Brussels Effect in so concentrated a manner as internet privacy and data protection. The EU’s General Data Protection Regulation (GDPR) is widely acknowledged to have shaped the compliance culture of companies and government agencies far beyond Europe.

Impact of the EU GDPR on the profits and sales of a sample of companies from 61 countries and 34 industries worldwide.
Source: Chen et al (2022), Oxford Martin School

The question on everyone’s mind then is, would the EU AI Act do the same?

As things stand today, many popular AI systems don’t measure up to the proposed EU standards. Source: Stanford CRFM

Reports that Parliamentarians in Ghana might initiate legislation on AI, following extensive stakeholder engagement in various African countries funded by Germany’s GIZ, outlines one obvious path through which the EU’s influence in such matters spread: donor relations and international development aid.

My more radical friends in the Civil Society movement promptly dismiss such channels as just another neocolonialist ruse. The more charitable ones reference dated theories of “normative imperialism“.

A more nuanced view of the Brussels Effect however focuses rather on its utility, the practical problems it solves for those societies around the world choosing to delegate regulatory thinking and standards setting to Europe. As pragmatists like to say, there is no genius in reinventing the wheel. If Europe has the will and wherewithal to do the heavy-lifting of tackling such complex, and protean, novel challenges such as how to tame AI for the rest of the planet, why protest?

In a preprint short essay, I explore this theme in some depth.

As is now widely acknowledged in the literature, the EU’s success in exporting and diffusing GDPR was a special case born out of exceptional circumstances, and not the sign of a general, not to talk of universal, trend.

GDPR capitalised on broad longstanding conventional wisdom about the desirability of more privacy and better data protection to build its momentum. The benefits of lax and loose privacy and data protection policies have rarely been collated into a definite counter-proposal against which GDPR had to contend. There is limited contention about the importance of good privacy and data protection rules and what they look like. The EU only needed to give heft to projects and processes scattered in bits and pieces around the world, and supported by a fairly expansive bedrock of popular and elite sentiments.

My argument, in the essay referenced above, is that emerging technologies with potentially vast benefits and a profoundly uncertain trajectory present unsurmountable barriers to the Brussels Effect.

I predict that not only will countries and companies hesitate in taking strong cues from the EU regarding how to design and respond to compliance mechanisms, but that we will also see distinct efforts to dilute the EU’s moral influence in the AI ethics & regulation domain.

Global approaches will splinter along various axes through, as India is doing, calls for genuinely multilateralist solutions, which will typically slow down any serious action as the technology races forward; and, as many others are doing, resistance movements against the EU’s signature approach of use-case classifications and proscriptions.

The Parliament of Ghana is reviewing an onlending agreement between the government of Ghana and a new Ghanaian state-owned development bank, DBG.

This agreement is a requirement for the European Investment Bank (EIB) to start releasing tranches of funds from a €170 million commitment to the DBG.

However, the decision to fund the DBG by key multilateral development banks such as the World Bank, Africa Development Bank (AfDB) and the European Investment Bank (EIB) was taken in 2019 when Ghana’s finances still had a shine. The kind of shine that billions of dollars of Eurobonds can confer.

After dragging its feet since 2020 when the EIB agreement was signed, the government of Ghana has finally dragged itself to Parliament to pursue ratification of the on-lending agreement between it (as the primary borrower) and the startup DBG (as the entity to disburse the funds).

But the world has changed enormously since 2019.

At a time when the country’s three existing development banks – Agricultural Development Bank, EXIM bank, and National Investment Bank – are widely considered to be poorly governed and heavily undercapitalised, and the finances of the country itself have crashed, the whole DBG wholesale on-lending model feels out of touch and out of place.

There are many challenges with the proposed DBG’s model that deserve very careful unpacking and in due course that will happen. But even without getting into the weeds, the very fact of expanding public debt in the name of development banking in a country struggling to restructure existing debt leaves a bad taste in the mouth.

We expect that Parliament – or at least the Opposition half of it, seeing as the government faction never performs any oversight – will subject this whole DBG program to serious scrutiny and insist on a proper national debate about the propriety of doing this at this specific time.

Bearing in mind that the World Bank is also supposed to release $250 million to the DBG at the same time as Ghana is pursuing concessional financing from the same quarters just to balance its budget. In total, the government intends to borrow nearly $800 million to execute the DBG strategy.

So far, the governance safeguards that were demanded by the multilateral development banks are merely assurances on paper. Some of the stipulations will be easy to follow. But those ones are the least important. For example, money from the EIB to the DBG cannot be on-lent to alcoholic companies, quarries or any mining projects. Compliance in such circumstances is easily verifiable. Rules against nepotism, cronyism and disbursement to favourites are far harder to enforce. Yet, this is where state-owned banks in Ghana – development-focused or standard-commercial – typically falters. Two Ghanaian development banks collapsed in the 1990s after senior management facilitated cheque fraud by a corporate client to the tune of nearly 130 billion Cedis (old currency series). Less than 30 billion Cedis were eventually recovered.

Already, there are worrying signs that DBG operates according to the same opaque principles that all state owned enterprises (SOEs) in Ghana do. It publishes little about its strategy, advertises available roles and consulting gigs half-heartedly, and gives very limited account of how the tens of millions of GHS it says it has disbursed to SMEs were actually disbursed. Just another opaque Ghanaian SOE.

Whilst the qualifications of its initial bosses cannot be questioned, there are no indicators as yet of a corporate governance culture cut above the rest of the SOE landscape.

We hope that the EIB, AfDB and World Bank will not continue to enable public debt accumulation by Ghana with limited checks and balances and then when the country hits the rocks and is struggling to pay its creditors, everyone points the finger at someone else.

The nomination of Ajay Banga, the former Chairman of global payments giant, Mastercard, as the new President of the World Bank by the organisation’s largest shareholder, the United States, has coincided with a big debate about how to reform the Bank to strengthen its delivery as the world’s number one development champion.

The Bank’s own “evolution roadmap” sketches the contours of a new vision to tackle the urgent issues of persistent poverty, skewed prosperity & the globalisation of the planet’s biggest threats.

A major prospect outlined in the document is the injection of “additional financing” into the Bank through: “further optimizing the balance sheet, increasing the IBRD equity through various options, and increasing mechanisms for concessional funds for WBG activities to address GPGs.”

The IBRD, cited in the quote from the roadmap above, and the IDA are the two main, legally separate, public-sector entities within the World Bank Group. Broadly speaking, the IBRD caters to the higher borrowing needs of the middle income and more credit-worthy low-income countries, which despite being wealthier than the average Global South country still host the majority of the world’s poor.

To do this, the IBRD borrows from the private markets on the back of its solid credit rating (AAA) to get funds at reasonable rates. The IDA, on the other hand, uses money it gets from rich governments, as well as the capital markets, to lend primarily to the poorest countries in the world, usually in the form of soft (lower-interest) loans complemented with a generous padding of grants.

In recent months, boosting the capital of the World Bank for increased lending has gotten all the attention. All the 14 largest Multilateral Development Banks (MDBs) in the world – the World Bank plus others such as the African Development Bank  (AfDB), Asian Development Bank (ADB) and the European Investment Bank (EIB) –provide less than $170 billion annually in development financing. A little over a third of this money comes from the World Bank group. Many feel that this is far from enough.

To put the numbers into perspective, experts estimate that between now and 2025, the world ought to be spending $2.6 trillion per year, and $4.5 trillion from 2026 to 2030, just on the Net-Zero climate transition effort alone (i.e. getting the world to zero net carbon emissions). Currently, the world is on course to spend less than a trillion US dollars a year leaving a gap of more than $1.7 trillion per year between now and 2025. The private sector already outspends the MDBs by nearly 50% in sustainable development financing.

At any rate, even the modest amount spent by the MDBs is heavily dependent on borrowing in the market. Roughly 95% to 98% of the monies invested in the MDBs by rich governments operate more like guarantees for borrowing done in the world’s bond markets rather than actual cash.  

On the premise that more capital injections will increase lending capacity, the MDBs regularly undertake “general capital increases”. In 2011, for instance, the principal institutions boosted their spending power between 31% (World Bank’s IBRD) and 200% (AfDB), mostly through raising bonds. In 2018, the World Bank secured another significant capital increase of $7.5 billion in cash and $52.6 billion in guarantees from the major shareholders for the IBRD. On the back of the capital increase, the IBRD and IDA are able to disburse about $50 billion jointly these days (60% of which went to Sub-Saharan Africa) compared with $30.5 billion in 2017, before the capital increase.

The question however is whether capital increases are the most critical factor in shaping the capacity of the World Bank to respond to the economic development needs of its poorest members.

The first point to note in answering that question is that historically general capital increases have not been the main driver of an increase in the quality or quantity of development lending.

For example, the five-year average growth in IBRD commitments before the 1988 capital increase was 61%. After a near 16% jump in the year after the capital increase, growth for the subsequent 5 years averaged just 3.1% over the entire period. In a similar vein, average IBRD commitments grew by a cumulative 250% in the 4 years before the 2011 general capital increase and actually saw a drop of 12% in the four years thereafter.

A similar picture can be seen before and after the last capital increase in 2018. From a pre-capital increase high of $64 billion in 2016, commitments have grown to just a little over $70 billion today.

The second vital point is that disbursements (actual cash flowing to borrowing countries) have thus remained stagnant at ~71% between 2017 and 2022.

The combined effect of these two trends is that actual disbursements between 2016, before the capital increase, and 2022 has actually stayed flat: from $49 billion to $50 billion. If an increase in guarantees and cash of ~$60 billion did not transform IBRD lending from 2018 onwards then plans to loosen capital adequacy rules by the World Bank management, which will generate just about $4 billion extra, are unlikely to make a difference.

The technocratic consensus for the World Bank to implement another capital increase therefore ignores a serious challenge in getting funds to drive development in poor countries: increasing quality disbursements.

This author is part of a network of civil society organisations (CSOs) that frequently analyse government projects for their integrity and ESG compliance. Two decades ago, we were highly critical of the Bank’s role in a development aid system that, in our view, often permitted waste and corruption. After participating in the Bank’s Africa strategy review in 2011, CSO networks like ours began to see how domestic factors often overwhelm the bank’s systems and seriously slow disbursements.

In several African countries, the Bank’s exacting standards, including its tendency to blacklist noncompliant private contractors and alert other development banks to do the same, have led to a tendency of government officials to drag their feet and, where possible, even look for alternative sources of finance, even if more expensive.

Recently, researchers analysed more than 400000 contracts awarded under MDB projects between 2000 and 2019 to the tune of over $850 billion in nearly a 180 countries. World Bank projects scored much higher than regional MDBs for their use of controls to avoid corruption and collusion.

Our own experience in Ghana and elsewhere in Africa testifies to this. Attempts to access documentation on a recent $750 million facility from Afreximbank to Ghana have been blocked at all levels by both the Finance Ministry and the Parliament for six months now and counting. Such a thing would be unthinkable for a World Bank facility.

The high transparency requirements, strict procurement rules, and elaborate monitoring and evaluation yardsticks grate against unresponsive bureaucratic and governance cultures bolstered by the recent expansion of African sovereign access to the “ask no questions” private bond markets.

In Ghana, recent reviews of major World Bank public sector reform projects expose strong barriers in transcending this disbursement challenge. What is more, little has changed over the many decades of World Bank lending to the country suggesting a metaphor of state dyslexia.

A much more interesting case study to drive home the point comes, however, from South Africa. Experts frequently argue that the low interest rates charged on IBRD loans make the World Bank a potentially powerful mediator between African countries with massive capital needs, like South Africa, and the open markets. Yet, the disbursement rate for the entire South African IBRD portfolio is a paltry 16.89%.

This is despite very strong alignment between the IBRD’s strategy and South Africa’s most pressing infrastructure needs. The country’s power crisis has been described by its President as a “state of disaster”. The power shortfall was long anticipated and problems unlocking private capital were exactly of the sort designed to be fixed by the IBRD’s financial intermediation.

So, a project to lend $3.75 billion to the main state-owned energy utility, Eskom, for a power plant at Medupi was put on the table in 2009. After nearly two years of preparation, the project launched. The first year and half went well and nearly half of the committed funds were disbursed. Then the ill-fated second phase of the Zuma presidency started to bear down on government operations. Medupi became trapped in a loop of project management chaos and confusion. Despite five restructurings, full disbursement could not be achieved.

More tellingly, it is the green/climate financing component of the project, with its greater need for management sophistication, that suffered the most. Nearly 13 years after the project was initially mulled, it had to wrap up with nearly $600 million undisbursed, of which $408 million were for clean energy interventions, including a critical grid-scale energy storage solution. Those funds have been rolled over into a new scheme but reports indicate that problems persist. $100 million slated for capacity reforms had to be cancelled outright.

Despite taking more than double the estimated time to complete, project challenges were never fully resolved. The Medupi plant’s throughput (or “availability factor”) hovers below 58% compared to the international standard of 92% for equivalent installations. Not surprisingly, as project risks escalated, the Bank’s risk aversion followed suit, hobbling delivery of ancillary projects.

In short, more cash for the World Bank through yet another general capital increase would do little to address the serious blocks at country level preventing the effective absorption of already committed funds. What is urgently needed is a frank conversation about policy integrity followed by a concerted effort to unify the standards regime for accessing money across different sources so that countries are more incentivised to comply with strong rules.

Whilst the World Bank touts a recent improvement in ratings of its projects, its independent evaluation group notes that Completion & Learning Reviews, the gold-standard monitoring instruments, have fallen to a record low. In fact, in 2021, only two were conducted, versus 99 in 2011.

Through this limited evaluation aperture, historically challenged African regions continue to record very low performance: only 49% of World Bank projects were rated as moderately satisfactory or higher.

Tragic though it may seem, the ongoing shutout of several African countries from  private bond markets forcing a reversion to the MDBs also offers a narrow but compelling opportunity to think through and agree on a unified standards regime, especially because private lenders are better attuned to the need for fiscal discipline at country level now more than ever.

Given the MDB’s heavy reliance on private lenders for their resources, the fact that the latter have been competing for the same loan opportunities in the same attractive developing countries, and in some ways driving down standards, makes harmonisation to close the ESG arbitrage gap in development finance markets even more imperative. Especially now that debt relief campaigners are accusing private lenders of reckless lending and insisting on punitive debt cancellations; it is amply clear that relying solely on risk premia in debt pricing in private markets won’t cut it. Capacity building to boost project performance standards at country level is critical.

It is good to highlight the urgent need for more development capital overall but such calls ring hollow when they don’t also address all the cash being left on the table, because of the Disbursement Crisis, or wasted because of weak absorption capacity and a compromised ability to manage project integrity in some of the world’s neediest countries.

This author’s personal acquaintance with Ajay Banga’s commitment to strong execution over rhetoric suggests that he is well equipped to tackle this challenge. It is unclear though whether he has the risk appetite to go against the technocratic consensus on capital increases.

Regular readers of this page were puzzled when after weeks of projections by analysts of a final participation rate of between 60% and 65% in Ghana’s domestic debt restructuring exercise (DDE), the government triumphantly announced a “more than 80%” figure this afternoon.

We shared our understanding on Twitter: the government was determined to announce a high participation rate and would thus adjust any necessary definitions to suit the objective, principally by altering the participation rate equation. But why is the participation rate even important in the first place?

The only reason why a government, or indeed any debtor, would seek to restructure their debt is because times are so hard for them that they cannot pay according to the original terms. A need therefore arises to reduce the amount they are obliged to pay from time to time, whether in interest, principal, or both.

Thus, for a debt restructuring exercise, two things matter above all else: the average haircut amount and the participation rate. The participation rate, crudely speaking, is the ratio of A) the face value of the restructured debt (so, in this case, the new bonds) to B) the face value of the original, unpaid, debt (or outstanding principal). “A” is the numerator and “B” is the denominator, yielding a basic equation: A/B. The participation rate together with the haircut amount thus determines the overall debt relief, the amount of money saved by the debtor, in this case the government, as a result of not having to service its debt according to the original, more burdensome, terms going forward.

When setting out to restructure debt, a debtor is always conscious of the entirety of debt that can be restructured. This is what is referred to as eligible debt. Think of, say, an individual seeking to sort out their personal debt after losing their job. The bank may be willing to refinance the mortgage; a friend could be persuaded to forgive a wedding loan; but a leasing company may refuse any adjustment to the terms of the car loan. In such a situation, the mortgage and wedding loan constitute the eligible debt for that individual’s restructuring effort.

When Ghana announced its debt restructuring exercise on 5th December 2022, it pegged the eligible debt being treated in the announced DDE program at roughly 137 billion GHS ($11 billion).

When things started to get rough and the government was forced to amend the offer to entice more creditors, it amended the eligible debt to roughly 130 billion GHS ($10.4 billion).

After a series of extensions, belated concessions, and veiled threats failed to break the front of holdouts smarting from the bizarre and incomprehensible failure of the government to engage creditors well ahead of the formal launch of the restructuring exercise, panic began to set in.

Finance Ministry Mandarins hunkered down with the expensive suits from Lazard Frères, the government’s high end consultants in this exercise. Over copious beverages and after much brainstorming, punctuated by the occasional exasperated stutter, a plan started to form. Ghostly smiles etched on lip corners as its simplicity and sheer elegance became clear.

The plan was that the government will choose whatever eligible debt number it wants when it wants to present the results of the DDE and do so in anyway it wants. A sizzle of self-congratulatory buzz went round the room as the coup sunk in. No one leapt up to do a boogie-woogie, but the waltzing tunes from the adjoining antechamber became very audible very suddenly, stirring up some graceful head-bobbing.

Analysts who have been tracking every step of the DDE meanwhile kept issuing frantic updates to their audiences about the inevitability of a low participation rate, and thus a likely lackluster debt relief outturn. How naïve. They completely misread the entrails. When a tenacious government used to getting its way teams up with expensive, highly experienced, international consultants, a poor result is impossible!

Thus it was that later today, after the earlier press release, the government issued another press release. In it, the “over 80%” figure had crystalised into an “85%” participation rate.

On the second page, the abracadabra was presented in all its glory: the government has selected eligible debt of roughly 97 billion GHS ($7.6 billion) for the purposes of calculating the participation rate.

Talk of magical mathematics!

Image Source: Dexter’s Lab Wiki

Knowing that a few nosey people might poke around the numbers, a perfunctory sop of an explanation was thrown their way. Essentially, the government has changed its mind about what debt was eligible to be restructured. At the tail-end of the exercise. End of story.

But even amidst the cursoriness, some analysts’ antenna went up. One terse justification of the lastminute reduction in eligible debt is a claim that some investors have converted their bonds into treasury bills. This could only have been done with the permission of the government. But why?

A theory doing the rounds is that it may be related to some swaps in which the government, most likely through the Bank of Ghana, is entangled. According to this theory, such derivatives are intertwined with a chunk of bonds such that any attempt to extend the maturity of those bonds would trigger serial defaults.

Whatever be the actual facts of the matter, and trust that they will be unveiled in due course, regular readers can rest assured that the debt relief analysis presented earlier still holds in its entirety.

First, using a more expansive bracket of government marketable securities (such as what was indicated in the debt exchange memoranda) as the denominator, one obtains a participation rate of between:

83 billion GHS (new bonds)/137 billion GHS (original eligible debt) = 60%


83 billion GHS (new bonds)/130 billion GHS (in-program amended eligible debt) = 63%.

So, analysts’ projections of a rate between 60% and 65% is far more robust than the government’s preferred 85%.

Does it matter?

Well, review the government’s own conduct following the last deadline of the exercise. One day to the final deadline, a junior Minister announced a participation rate of 50%. Once this was reported by the Press, friendly journalists were immediately briefed undercover to start circulating a new number of 70% plus. The junior Minister was promptly instructed to stop commenting further. Government affiliates in the media then took over the narrative. This morning, the number became “over 80%”. It has finally alighted at 85%. Clearly, there would be no need going through all these hoops if the narrative didn’t matter to the government.

The universal and enthusiastic reporting of the “over 80%” and “85%” participation rates in the mainstream press without any of the nuances above is clear evidence that the participation narrative matters greatly to the government and its key audiences.

For more specialised observers however (the ones most likely to pay attention to the kind of detail often explored in these pages), a more accurate number matters for its sharper reflection on the evolving fiscal picture.

The government’s PR successes are fundamentally irrelevant to such people.

Only the consequential analysis of the likely debt relief attainable from the just ended exercise matters. Keen observers are more likely to be concerned about the fiscal hole of over $2 billion in the government’s affairs that we insist can only be tackled by a sincere, independently supervised, expenditure “rationalisation” program because the current national budget doesn’t go far enough.

Of course, some secondary concerns also emerge when one considers the sequence of events so far. Some of the government’s explanations, self-serving though they are, also point to some disorganisation in the fixed income market. We have earlier warned about the potential error rates of settlement due to the rush, for instance in relation to unprocessed withdrawal requests by late holdouts. Some disgruntled people exposed to the debt exchange through the actions of fiduciaries, such as banks, have been baring their teeth. In these circumstances, litigation risks still persist, notwithstanding the formal end to the program.

The government itself concedes the need to tie the loose ends of the exercise gingerly, and has extended the settlement date to 21st February (to the ire of holdouts demanding a cure by 17th February of temporary defaults on the old bonds). It has unilaterally amended the exchange agreement to give it room to pursue additional exchanges and introduce mopping up instruments, among other imminent creative responses to the underwhelming performance of the debt exchange program.

In short, the government’s magic does not extend far beyond the spinning and yarning antics currently on display.

[This is a developing story. Updates may be made to this note as more information emerges.]

Regular readers would recall a recent tweet about whispers floating around Accra’s political alleys about the role of fixers and go-betweens in Ghana’s Vice President’s vaunted “Gold for Oil” initiative.

A bit earlier, we explored the merits and prospects of the whole Gold for Oil program in some detail.

When the above tweet was made, the Chamber of Bullion Traders, a gold merchandising lobby group in Ghana, was aggressively lobbying and briefing about their grievances over the Gold for Oil program and the harm it was causing to their interests.

The government’s decision to hoover up large quantities of gold from the small scale mining sector to support the gold and oil barter agenda could, according to their analysts, spell the end for some of their members and set in train a slow spiraling death for the entire private gold exporting sector.

It would seem that all the briefings and counter-briefings have roused anti-government media activists who have this weekend drawn first blood.

WADR, a Washington DC based non-mainstream media outlet strongly opposed to the current Ghanaian government has made a number of allegations corroborating some of the briefings swirling in the undergrowth of Ghanaian political chatter. Their allegations are backed by invoices and trade documents featuring actors who have been mentioned in confidential briefings as architects of the Gold for Oil program.

At the heart of the fascinating schema, as per the briefings, are two highly secretive and super-discreet investment bankers and dealmakers who also double as on-off financial advisors to former British Prime Minister, Tony Blair, himself a favourite Advisor to Ghana’s Vice President.

The two men – Reyhaan Aboo and Prashant Francis – are principals of London based Portman Partners and UAE domiciled Alphastream. They, especially Mr. Francis, were key enablers of Prime Minister Blair’s now dissolved, and once highly scrutinised, Firerush Ventures.

Alphastream’s key operatives are JP Morgan alumni who are well connected with the American investment bank’s powerful Ghanaian network, with members ranging from fund industry gurus to a former junior Finance Minister.

Alphastream, with primary investors in a UAE Sovereign Wealth Fund, brands itself as a multi-commodity royalty streaming company. The same business the ill-fated Agyapa entity attempted to enter.

According to the briefings, now corroborated by documents leaked to WADR, the Bank of Ghana has signed a secret agreement to sell $1.1 billion worth of gold and an unspecified quantity of silver through Alphastream to obtain dollars for procuring refined petroleum products.

It is absolutely unclear what specific value Alphastream adds in a simple gold export transaction but theories, mostly unsavoury, abound.

Regular readers would recall our recent tweet that Russia’s Litasco, under pressure from the EU’s expanding sanctions regime following the Ukraine invasion, has set up in the UAE and is exploring creative trades. Litasco is reported to have brought in the first consignment of gasoil blends from the Baltic Port of Vystosk under the Gold for Oil program.

Given Litasco’s reputation for “crude oil for refined products” deals, most notably in Nigeria, its involvement in Ghana’s barter program was unsurprising. Yet, impeccable sources insisted that they had not taken gold for payment.

The shadowy intrigues surrounding the very first transaction cast a harsh light on the opacity, lack of forthrightness, poor policy grounding, parliamentary sidelining, and zero accountability beclouding the entire Gold for Oil program.

Such murkiness could only deepen the clouds of suspicion and provide more fodder for the whispering campaign. One mystery above all drove the intense chatter: how had Litasco been paid?

The leaks to WADR shed light on this strand of the affair and ties some of the loose ends of earlier analysis. The money no doubt came from StoneX Commodities DMCC, part of the $1.7 billion a year trading wing of the StoneX Group, whose precious metals unit, especially the Girish Surendran trading desk, has been particularly aggressive of late in exploring multi-sided transactional solutions.

According to WADR, 502 kilograms of gold, out of an Alphastream quota of 18 tons, plus an unspecified quantity of silver have been freighted on Emirates airlines to StoneX’s refinery in the UAE.

Based on proforma invoice pricing, the total value of the shipments should amount to about $33 million. Roughly equivalent to the approximately $32.8 million that 41000 metric tons of Rotterdam gasoline blend would have cost on CIF basis using the relevant Argus pricing benchmarks.

January Pricing Benchmarks for White Products. Source: Argus

In short, as many people have said in the past, the whole Gold for Oil program simply entails diverting gold from Ghanaian private sector players and selling through politically connected middlemen for dollars in Dubai. Then paying a Russian trading firm that already buys crude from Ghana and, finally, netting off in spot settlements. Far from the revolutionary gameplan of recent government narratives. At first approximation, the scope for backhand fees and hidden premia is, unfortunately, massive.

Had this convoluted ring-a-ring-a-roses been equivalent to a conventional oil trade intermediated by the usual Ghanaian Bulk Distribution Companies (BDCs), we could have put the whole thing down to the Ghanaian politician’s need for populist dramatics. But the multiple layers of related parties playing middlemen for a cut and the thick cloud of opacity ring serious alarm bells.

Some of the claims of WADR are hard to follow and the organisation’s key activists use unconventional scales and conversion rates (examples: suggesting that “129 pounds = 1kg” and that 1 pound of local gold costs 5000 GHS, both of which are erroneous). But the general concerns about shadowy brokers have been raised by others far more steeped in the commodities industry, and fears of potential losses for Ghana are reasonably grounded.

The slow rate of gold mobilisation, no doubt funded by cash printing with inflationary potential, and the multiple hoops of exchange rate and credit exposures, all come together to suggest a tottering wreck of a scheme that may be too clever by half. There is, after these revelations, no rational basis for holding on to a belief in the scheme strengthening the Cedi or leading to lower fuel prices at the pump.

Before further leaks and more confidential briefings ensue, we strongly urge the government of Ghana to publish every piece of information relating to the various roles played by the Bank of Ghana, Precious Minerals Marketing Corporation, Litasco, StoneX, Alphastream, and the secret local gold aggregator muscling out the established gold traders and exporters in this strange procurement scheme.

Why should every policy of government become a magnet for drama and controversy?

Word on the street is that Ghana’s drama-filled debt restructuring/exchange program (“DDE”) saw between 60% and 65% of all eligible extant marketable government securities (simply, “bonds”) tendered in by their holders in exchange for new bonds offering lower average interest and longer average repayment tenures. Investors accepted losses of between 19% and 47% instead of the 55% to 88% (depending on inflation & discount rate assumptions) they would have suffered under the government’s original December 5th 2022 plan.

Whilst the outcome is perfectly in line with analysts’ expectations, the government had held out hopes of hitting its 80% non-binding target, and postponed the program 5 times to increase chances of doing so. Deniable leaks from government sources to selected media in Accra pegging participation at 70% are considered less credible, and at any rate don’t change much by way of effect.

After a week that saw pensioners, some in wheelchairs, accost Finance Ministry officials and a former Chief Justice declare the entire exercise a complete illegality, the DDE architects can breathe a sigh of relief even if the participation rate and debt relief outcomes, the worst in modern world history, are not exactly stellar.

 CountryRestructuring ScopeYear CompletedDurationParticipation Rate
1GhanaDomestic20232.2 months60% to 65%
2ArgentinaDomestic202011 months99%
3BarbadosDomestic20184 months100%
4BelizeInternational20175 months100%
5ChadInternational201817 months100%
6EcuadorInternational20205 months100%
7GrenadaDomestic & International201532 months100%
8MozambiqueInternational201933 months99.5%
9UkraineInternational201511 months100%
10UruguayDomestic20036 months99%
Data Sources: IMF, ECB, and Cruces & Trebesch (2013)

Still, as we have argued elsewhere, the DDE was the easiest, and not even the most politically costly, way for the government to raise a large amount of money. At a 65% participation rate, the government will pay roughly 6 billion GHS on the tendered debt versus the roughly 17 billion GHS it would have paid on the old instruments in 2023. In subsequent years, the government’s payment obligation will almost double, but for the biggest creditors, this increase in payout will be deferred until after the 2024 elections.

At first approximation, the DDE thus represents a transfer of roughly 11 billion GHS from the private sector (and, minimally, the Bank of Ghana) to the government in 2023 alone, an amount higher than what the state takes in from external VAT (~8 billion GHS), and National Health Insurance and GETFund Levies (~9.3 billion GHS); and only a little lower than revenues from import duties (~14 billion GHS).

Given the disrespect with which they have been treated, the private sector have been most deferential to the government by forgoing such a large amount of money. As we have repeatedly said in these pages, Ghana’s DDE was by far the most unorthodox the world has seen since Argentina’s much derided program in 2000. The country seems to have taken a leaf from Argentina’s repeat game in 2020, and then carefully overdone the theatrics.

No country in the world has ever launched a DDE or any sovereign debt restructuring program of this magnitude without extensive informal consultations with major creditors ahead of the official commencement of the program. There is a reason why the average time for conducting a debt exchange in the last two decades is in the range of 11 months. People like to refer to Uruguay’s program in 2003, which officially took 7 weeks from formal launch to settlement. However, the launch was preceded by more than three months of intensive discussions with all the key creditors.

Ghana’s decision to rush through the process and string together a series of unilateral deadlines on a take it or leave it basis, whilst reflective of the governing style of the current government, was thus completely unprecedented.

It had led many to wonder whether the playbook created by Lazard Frères’ Eric Lalo and Michele Lamarche, which has been in use during this whole enterprise was worth the multi-million dollar advisory fees.

Of course, this being the first time an African country has dared to restructure its domestic bonds, perhaps Lazard can be excused for their lack of sparkle. Those in the know do say that Lazard was quite effective during Ivory Coast’s restructuring of Brady Bonds in 2009 and that the subsequent faltering in attempts to establish a regional base in Abidjan owe more to the vicissitudes of the sovereign debt advisory market than to any failings of substance.

Yet, watching the Lazard Sovereign Advisory method that Ms. Lamarche, Monsieur Eric Lalo and Monsieur Matthieu Pigasse honed over decades of practice flounder in the midst of all the drama we have been served in the last couple of weeks was a sight to behold for many analysts.

Matthieu Pigasse broke into the public consciousness after becoming an advisor to Greek’s left-wing government following his role in the country’s debt restructuring. Image Source: CityAM

It is true that Lazard sovereign debt restructuring advisors don’t shy from combat when that is necessary to score an important point. But the decision to walk back on previous commitments made to exempt pension funds and the tactless handling of the pensioner concerns, such as waiting for these elderly citizens to start issuing threats before engaging seriously, were not signs of spunk. They smacked, instead, of clumsiness.

Which is why some observers feel that Lazard was not allowed by its Ghanaian principals to bring all those years of advising Greece, where far more complex matters were afoot, and Ecuador, widely praised for the sophistication of the stakeholder engagement process, to bear.

Such a view would not be altogether sound though in view of some of the rather hairy spectacles Lazard got entangled in during Argentina’s 2020 post-default negotiations with creditors.

Whatever be the case, whether it was Lazard that pushed these hardball tactics or the Finance Ministry, the ball remained throughout the period in the court of the wider government of Ghana to ensure alignment between this narrow debt restructuring program and the broader political economy of getting Ghana away from the brink of the economic abyss it is currently staring into.

Let us not mince words here, the admission that the Bank of Ghana had to print money equivalent to roughly half the government’s domestic revenue in order to service debt over the last year and stave off a default has thrown the country’s overall fiscal situation into very stark relief. Analysts are now even clearer in their forward view that debt relief of 11 billion GHS a year, as delivered by the just-ended DDE, is far from sufficient to get Ghana back on the path to macrofiscal stability.

The country is now literally using short-term treasury bills at nearly double the cost of the bonds it says it can no longer afford to finance day to day government operations. In no time, all the gains from the DDE debt relief will be wiped off simply from the escalating costs of fresh domestic borrowing. $1 billion from the IMF this year will certainly not substitute for the $4 billion in curtailed international inflows. Greater relief is urgently required.

Moreover, judging from the posturing of China, it does not look likely that a quick Paris Club deal can be arranged in the government’s preferred March timeframe via the Common Framework. The IMF may have to waive the bilateral debt restructuring requirement if the plan is to get the much coveted board approval for Ghana’s provisional staff agreement in the Spring. At any rate, bilateral debt relief should provide something in the range of $100 million or so a year, a clearly insignificant amount in the wider scheme of things.

If the analysis above is correct, then the next big drama ahead for Ghana is swift and smooth Eurobond restructuring talks with its external creditors. Reports that Franklin Templeton participated in the domestic debt exchange are encouraging but they do not completely assuage concerns about the coolness shown by many other offshore investors towards the just-ended DDE.

After Ghana chose to suspend servicing its Eurobond debt without so much as the ritual courtesy of applying for consent, external bondholders like Pimco, Fidelity, Goldman Sachs, and BlackRock have had to signal, albeit subtly through well-placed hints, that they will not be walkovers going forward.

The balance of probabilities incline in the direction of moderately tough negotiations in respect of the government’s reported demands for a moratorium on servicing its Eurobond debt. We are pessimistic of a total moratorium request being accepted by Ghana’s Eurobond investors. They will be taking cue from how Ghana’s refusal to build proper consensus ahead of the DDE through extensive consultations forced the sovereign to consent to some of the most creative pari passu violations in world debt default history.

Ghana’s goal of a billion dollars of debt relief in 2023 from Eurobond restructuring is thus ambitious and, in light of the DDE performance, possibly unattainable. Should the government succeed in wringing out about half of that, the country still faces a $2.5 billion dollar hole that must be addressed through other forms of fiscal adjustment besides debt restructuring. It bears mentioning that this is a conservative figure.

In these circumstances, the original debt restructuring program would clearly need supplemental strategies. In the recent past, we have discussed plans to accelerate treatment of non-marketable debt, to tackle the debt of state-owned enterprises like Cocobod and the Electricity Corporation of Ghana (ECG), and to explore ways of dealing with mounting energy sector debts. There is aggressive language coming from some government quarters about forcing power producers to accept cedi conversion of forex liabilities among others.

The downside of all this uncertainty is the prolonging of the souring plight of the financial sector. Despite repeated assurances of a swift recovery of enthusiasm in the domestic bond markets, the high holdout rate blocks the government from seeking to apply exit amendments and other tactics to damage the old bonds in favour of the new bonds.

No doubt government advisors are even at this moment contemplating all manner of devices to seek to penalise holdouts. They would do well to exercise caution. The dramatic interventions of the former Chief Justice were meant to send a clear message: judicial elements who may have been personally affected by the exercise shall take a dim view on any legally dubious punitive measures brought against holdouts.

Consequently, the high quantity of circulating high-yield old bonds will exert competitive pressure on the low-yield new bonds even as the high interest rate environment contributes further to depress value. Consequently, the likelihood of a ratings improvement for the new bonds is low, as are the prospects of market recovery in 2023.

Another interesting angle is the implicit seniority that has been introduced among government creditors, with the biggest financial players at the bottom of the heap. The unintended consequence is that smart banks and insurance companies will reduce their holdings of all government debt, Including treasuries.

Smart pension funds can take advantage of the situation by creating products that harness their newfound seniority to give risk averse market participants safe exposure. But it is unlikely that this development alone will do much to shift asset ownership power in Ghana’s financial industry to pension funds, given all their other constraints.

The net effect, in the short term at least, therefore is a lowering in institutional demand for government securities and sustained upward pressure on government borrowing rates.

As we have said in preceding paragraphs, the government’s timeline of an IMF board approval of its staff level agreement in March 2023 increasingly looks more and more unrealistic since successful Eurobond restructuring is critical in lending credibility to the debt sustainability program, which is a vital precondition for even a moderately successful IMF program. The real question is whether approval in the Spring, i.e. by May, is feasible.

Of course, the IMF precondition only requires good faith efforts to restructure the debt not sterling outcomes. And, given where we are, the IMF is likely to accept some lacklustre debt relief outcomes in the medium-term. Nonetheless, such pragmatism would need to square with maintaining the credibility of the program. Meaning that acceptance of lack-lustre results in order to preserve a timeline of Board approval by the time of the Spring meetings (i.e. by mid-April), which is the most aggressive schedule feasible, has to be weighed against market perceptions about the viability of the fiscal treatment. Otherwise, a board approval announcement will give the fiscal indicators only a short-lived boost as was seen in December 2022 after the announcement of the provisional staff level agreement.

Supposing that a mid-April board approval is reached on the basis only of “good progress” with the external restructuring effort but before concrete debt relief assurances are secured, the IMF will have the option of securing Board approval but then making any tranche-one disbursement contingent on a successful program review later in the Summer of 2023.

In short, the next six months will require deft management of the country’s highly limited forex reserves and revenue inflows given massive pressure on both. Any slip, such as forex market adventurism, will see a return to the twin crisis of skyrocketing exchange rate depreciation and inflation spirals by May. The next twelve months, as a whole, will be shaped by market perceptions of the credibility of the IMF program.

If the government sticks to the same approach that it used in the domestic debt exchange, and continues to make the economic recovery effort a mere partisan-administrative activity, instead of one based on broader consensus, not even an IMF board approval in the Spring will correct the course of the fiscal crisis and avert a full-blown economic catastrophe.

To repeat for emphasis, the government must not:

  • fail to mobilise a serious national consensus behind a short- to medium- term austerity plan;
  • dilly dally in presenting a credible strategy of how it will cut public expenditures to plug the fat fiscal hole, by shrinking at least 25 billion GHS of its expenditure sheet; and
  • ignore calls to establish an independent “value for money” and “monitoring and evaluation” program with a remit spanning across the entire set of budgeted government programs.

Otherwise, Ghana should brace for a very rocky journey through 2023 and 2024.

A flurry of announcements this week was meant to signal that momentum was building towards a successful close of Ghana’s unorthodox domestic debt restructuring program (DDE). Banks, insurance firms and most capital market players have consented to revised terms for the orderly resolution of the country’s unsustainable domestic debt.

At a plush conference in the Ghanaian hinterlands, at a resort featuring Australian emus and llamas from Argentina, the Finance Minister had a literal spring in his step as he mounted the podium to praise the government’s commitment to “African prosperity”.

All this flourish might suggest the beginning of the end of the country’s most debilitating fiscal crisis in recent memory. Considering analyst unanimity about how harsh the Ghanaian DDE has been for domestic savers and investors, it is perhaps not too difficult to understand why the mere fact of some traction is enough to lift government’s spirits. Using the framework of its own advisors, Lazard, the government’s DDE offer definitely merits the investor-unfriendly tag.

Lazard’s model for assessing debt restructurings. Source: Lazard

And yet, here we are, on the verge of closure. A sober assessment of the situation, as I hope to explain in a second, would however recall Churchill’s famous statement in 1942 after the Allies blocked a massive Nazi incursion into Egypt and secured the vast oil fields of the Gulf in one of the most dramatic turns of World War II:

“It is not even the beginning of the end. But it is, perhaps, the end of the beginning.”

The deals agreed between Ghana’s Finance Ministry and the key financial market players, welcome relief from the stress and anxiety of recent weeks though they are, are significantly caveated.

In addition to the government revising its earlier stance of paying no interest on bonds this year and paying just 5% interest next year, it will instead pay 5% interest in 2023 and an effective rate of 9% (down from the roughly 19% weighted average rate on the old bonds) for the remainder of the term of the new bonds. Significantly, consenting domestic creditors are not getting a concession on the extended maturity of the reprofiled instruments, a major contributor to the Net Present Value (NPV) losses estimated by analysts.

To understand the provisionality of the agreements reached, one needs to carefully read the latest announcement, the one announcing a deal between the capital market operators (GSIA) and the Finance Ministry.

Just like the banks and insurance firms, the deal was brokered on the back of a promise by the government to set up a $1 billion Financial Stability Fund (FSF) to cushion any industry player that finds its solvency or liquidity threatened following the restructuring exercise.

As analysis of similar structures proposed for the European Union, and used in the Greek and Jamaican debt restructuring episodes, attests, a loan-based FSF is indeed the appropriate mechanism to use in these kinds of situations.

Comparative analysis of different fiscal support mechanisms.
Source: Misch & Rey/CEPR (2022)

The only problem is that the design of the proposed Ghanaian version of an FSF is entirely in the heads of mandarins at the Finance Ministry. So far, they have studiously refused to share any details about the interest rate, maximum term, collateral requirements, eligibility criteria, or, indeed, any of the major features one needs to know to properly asses a Fund of this nature.

At one point the Finance Minister even appeared to suggest that the Fund was for mere show because the Jamaican version ended up redundant. He ignored the Greek example, in which a decade after it was set up, the Greek FSF is still dealing with the lingering effects of the bank bailouts partially attributed to the country’s debt saga.

Worse still, there is no money as yet for the Ghana FSF. The government says it has approached the World Bank to cover 30% of the costs of the FSF. The World Bank operates within its strategic plan for Ghana. Making additional resources available is contingent on satisfactory progress on a whole host of pending issues about already committed resources. The other expected funders of the program, like Germany’s KfW, have elaborate processes for agreeing to new programs and disbursing funds and, at any rate, would also like to see this whole FSF embedded in a detailed economic recovery strategy, of which none has been forthcoming from the government.

In short, the FSF is pretty much conceptual at this stage. It is thus not clear whether the financial sector players expect incorporation of language concerning the FSF into an amended debt exchange prospectus. The answer will be interesting come Tuesday, the 31st of January, the deadline of the DDE.

On the GSIA front, the caveats are more striking. A good chunk of the holdings of these capital market operators are in Collective Investment Schemes (like mutual funds and unit trusts). The chains of exposure are quite complex, entangling some corporate treasuries as well as individuals. For example, the country’s fintech industry parks some assets using these and other bank custodianship arrangements. Under the terms of the provisional agreement with GSIA, government bonds owned by the Collective Investment Schemes (CIS), regardless of ultimate beneficial ownership, must be treated on the same terms as those owned by individual bondholders.

Any contingencies of the CIS kind imply that the creditor group concerned cannot sign a blanket DDE agreement on Tuesday. A revised document meeting legal muster must be prepared and reviewed before any prudent fund manager can proceed to sign.

But even if such agreements are signed in short order, they will still be subject to a fuller resolution of the caveated matters before actual immolation of the old bonds and issuance of the new bonds can proceed at the depository of the Ghana Fixed Income Market. It will be interesting how all this unfolds in the coming days, considering the government’s rush to have everything done and dusted in the first week of February.

And that is only in relation to the creditors signing on to the DDE. We can safely project that the vast majority of individual bondholders and offshore investors would not consent to the DDE by the 31st deadline. A significant number of corporate treasuries will also hold out. Taking that fact into account, and considering the earlier exemption of Pension Funds and the contingencies around the CIS holdings, one can also project a participation rate (on principal basis) in the DDE of about 65%, and debt service relief of less than 50% of the original expected amount.

A 65% participation rate would be the least impressive DDE performance in the world for a program that went to completion. It would certainly fall short of the government’s preferred target of 80%.

Holdout rates for DDEs around the world. Ghana’s DDE performance ( ~35% holdout rate) is set to underperform all others except those abandoned midway like Ukraine’s 1999 attempt.
Source: European Central Bank (2020)

The underwhelming results of the exercise can be entirely traced to the highly unorthodox approach taken by the Finance Ministry to launch consultations only after the debt exchange had commenced instead of before as has been the case with most DDEs undertaken elsewhere over the course of this decade, most of which saw participation rates above 90%.

When the exercise commenced in early December, this author said the following about certain financial industry players in Ghana:

Not only are they few in number, and their client base predominantly middle-class, but the government also wields massive regulatory power over banks and funds and expects them to do as they are told.

True to form, the government’s approach so far has been to ram the DDE down their throat. It was only after humiliating setbacks that it changed tack midway and grudgingly tried to do some co-creation. It goes without saying that launching consultations much earlier and mobilising a national consensus behind the DDE would have resulted in a higher participation rate and more debt relief whilst also spreading the pain more optimally. But even so, a 65% participation rate is comfortably above the 60% this author considers necessary for the program to have minimal viability.

It is important, moreover, to bear in mind that the resources freed up by the DDE, holdouts and exemptions notwithstanding, sum up to a figure just below the government’s largest revenue lines like Corporate Income Tax, Oil & Gas and Personal Income Tax. The DDE’s expected debt relief amount is considerably larger than proceeds from trade, energy and communications taxes etc. The banks alone may be “sacrificing” income of 15 billion GHS to the benefit of the government’s purse. Imagine attempting to haul that kind of dough through financial sector taxation.

So, where does all this leave us?

First, given the significant variance from the initial debt relief expectations, analysts expect some delays in finalising the full contours of the ECF before presentation to the IMF Board, likely straining the relationship between the IMF and the Finance Ministry. The government’s preferred timeline of IMF Board approval of March 2023 looks overly aggressive at this stage. In particular, earlier contentions by analysts that the fiscal consolidation component of the upcoming ECF program isn’t credible will be thrust back into sharp relief. Finance Ministry mandarins should not wait till the last minute before reworking the expenditure spreadsheets.

The IMF may choose to overlook the fact that the original debt sustainability strategy needs to be fully overhauled in view of the lower than expected debt relief and still present a program where the government only makes fiscal tightening pledges to the Board for approval. But doing this could dash the government’s hopes of the IMF frontloading tranche 1 disbursement, amounting to about $1 billion, to shore up the country’s forex reserves. Ghana’s reserve position is under unprecedented pressure, with gross reserves dipping below $4 billion, from nearly $10 billion a year ago, without even accounting for some not so liquid items on the Bank of Ghana’s balance sheet.

The IMF may in turn argue that any such disbursement should happen after successful completion of the first review of the ECF program, perhaps about three months after Board approval. It is highly unlikely that the Finance Ministry will consent to any arrangement that delays forex injection.

Which is why some analysts are beginning to ponder a scenario where government brings forward deferred domestic debt restructuring plans. Because the current DDE only covers 68% of the primary domestic public debt and less than ~56% of total public sector liabilities, the government may be tempted to initiate additional restructuring exercises earlier than planned in pursuit of additional debt relief.

The recent episode of Cocobod forcing a rollover of maturing debt (after the giant parastatal failed to raise a new facility to refinance expensive bills and the Bank of Ghana refused to step in) offers a clear hint of the government’s posture. Given that the country’s credit rating is already at rock-bottom, few restraints on debt repudiation remain. Apart from treasury bills and Bank of Ghana’s liquidity management tools, like swaps, every public liability in Ghana today is fair game.

External investors, keenly observing all these developments, are unlikely to agree quickly to total moratoria on debt service, as is the government’s wish. Trying to play total hardball may protract discussions and interfere further with the IMF Board approval timeline. It will be helpful for the government to be strategic this time around unlike in the leadup to the IMF engagement last July. That time, Ghana literally had to make a mad dash to Washington after a desperate attempt to hustle dollars from all manner of institutions between April and June failed to turn up even a dime.

Since then, everything has been a mad rush. It would be tragic if the government dilly-dallies with the outstanding creditor concerns until mid-March by which time the country’s forex situation would be completely dire before scrambling to pursue options that were obvious from the start (like abandoning zero percent coupon in 2023 during the DDE standoff).

It would be foolhardy in these circumstances for economic actors, and indeed the general public, to begin acting as though Ghana is nearly out of the woods. The rising chorus of governance reforms and the push for fiscal discipline should now intensify and not abate. The partial success of the current DDE is a mere lull in a storm that is still gathering.

All eyes should firmly remain on the foredeck, on the crew steering the ship of state, and no voice should stay calm if signs of rudderless maneuvering emerge.

Will ChatGPT Transform International Development?

After the spectacular failure of a 10-year project launched in 1982 by the legendary Japanese government agency, MITI, to dominate the global artificial intelligence (AI) landscape by developing systems suited to processing knowledge (rather than just data) in a user-friendly way, the country tried another approach.

In 1998, six years into the successor initiative, dubbed Real World Computing Program (RWC), publisher Ed Rosenfeld reported on an exhibition in Tokyo where the new project’s outcomes were on display. The goal, according to project managers, was to create a system to enable:

“humans to communicate easily with computers …[by the] implementation and refinement of “natural and smooth” interfaces through talking, facial expressions, and gestures.”

Fifth Generation Computer schematic. Source: Moto-Oka & Stone (1984)

Essentially upping the stakes on the earlier attempt. Unfortunately, RWC failed to charm, and soon slunk into oblivion. Japan’s 20-year AI domination strategy was nevertheless focused on two still-dominant narratives of the role of “thinking machines” in shaping our world. First, “expert systems” to solve human-defying problems and, second, “naturalistic interfaces” to make non-expert humans feel warm and fuzzy around a superior artificial intelligence. Some say that on both scores, ChatGPT has almost delivered.

ChatGPT’s triumphant launch thus resurrects the old dreams, debates and divisions that ensued when a political scientist called Herbert Simon, with no formal training in computing, helped set up a department at Carnegie Mellon University (then known as “Carnegie Tech”) that would become a pioneer in an early branch of practical AI called “expert systems”. In January of 1956, the year in which the Carnegie computing department was founded, Herbert stood before a class of wide-eyed students and proclaimed the invention of a “thinking machine”.

A famous picture of Herbert playing chess with his long-time collaborator, computer scientist Allen Newell. Source: CMU

But Herbert’s interests were far from confined to the logical construct of computing devices. His interests were in “psychological environments” and how these could be programmed to optimise human organisation for social change. When eventually he was awarded the Nobel Prize, it would be for his contributions to ideas at the root of today’s behavioral economics (despite not having been a trained economist), such as the optimisation of decision-making under uncertainty and other knowledge constraints.

“Expert systems” were the culmination of these concepts. Envisaged as holding large volumes of structured information assembled by knowledge engineers after extensive interactions with human experts, they would still operate within real world problem-definition constraints. Through complex, yet programmable, heuristics, such machines would solve problems faster and better than humans. But

Eventually, other branches of AI emerged where the rules of reasoning and analysis emerged from the body of assembled knowledge (or data, to be more liberal) itself instead of a compact heuristics or rules engine. One particular version of this AI model called “Large Language Models” (LLMs) power ChatGPT.

ChatGPT and other LLMs are a product of longstanding efforts to model human languages. Source: Okko Räsänen (2021)

LLMs take the “self-organising” concept to an extreme in the sense that massive increases in data scale appear to have a correspondingly massive effect on the AI system’s cognitive level, a marked departure from the original visions of Herbert and the other expert systems enthusiasts who believed that progress was in the direction of ever more sophisticated rules & heuristics engines.

Performance of Large Language Models is scale-variant/dependent.
Source: Julien Simon (2021)

What does all this mean for the future of knowledge-based social change? In the field of social innovation for development, where this author plies his trade, old conversations about how expert systems could turbocharge social and economic development have resurfaced.

In 1996, when US academic Sean Eom conducted a survey of expert systems in the literature, he found a field thoroughly dominated by business.

Sean Eom’s Survey of Expert Systems (1992)

Most of these systems targeted repetitive operational procedures. But a small subset was in use by consultants who focused on strategic open-ended questions, such as one may encounter in economic and social development contexts.

As the 90s unfolded, potential synergies among strategic problem-solving, expert systems and international development frameworks ignited strong interest in Western agencies.

For example, a review by Dayo Forster in 1992 for a major Canadian Aid Agency in the health context was motivated by the prospect of “maximum productivity” in the face of the appalling scarcity of resources in the developing world. Health has always been of great interest in the AI-for-Development community because of the importance of precise knowledge-based decisions, both at the public health level (example: what is the right interval between vaccine doses?) and individual health level (is this drug right for this woman given her or her family history?) For this reason, knowledge-enabled decision-makers make most of the difference in outcomes.

Then as now the promise of cost-effective thinking machines filling in the massive vacuum of local expertise in fields such as health, education, agriculture and planning is the most tantalizing prospect. Consider the case, for instance, of Liberia, which has just 300 doctors at home for a population of 5.3 million (up from 25 at home in 2000 when the population was ~3 million). There are two psychiatrists, six ophthalmologists, eleven pediatricians, and zero – yes a grand zero – urologists. Imagine the good that can be done if an expert system or similar AI could boost the productivity of each of these specialists ten-fold.

Sadly, this first wave of optimism faded slowly until a new turn in AI, anchored on Natural Language Processing (the umbrella group to which LLMs belong), started to produce the kind of results now on display in ChatGPT. The traditional concepts of expert systems were seen as a dead-end, and a ferment of new directions bloomed. ChatGPT’s wild popularity in 2022 marks it as a potential dominant AI design format, despite machine learning’s wider installed base and committed investments.

The positioning of Expert Systems & the NLP Bloom that led to LLMs, Transformers and eventually ChatGPT. Source: Chethan Kumar (2018)

Echoing the hopes of times past, a recent McKinsey report by the global consultancy claimed that “[t]hrough an analysis of about 160 AI social impact use cases,” they have “identified and characterized ten domains where adding AI to the solution mix could have large-scale social impact. These range across all 17 of the United Nations Sustainable Development Goals and could potentially help hundreds of millions of people worldwide.”

Until ChatGPT’s dramatic entry into the popular consciousness, investment commitments and revenue forecasts for AI heavily emphasised machine learning applications far more relevant for the Global North than the Global South.

AI investment commitments worldwide. Natural Language Processing (of which LLMs are a subset) significantly lag Machine Learning in this area. Source: Statista

ChatGPT’s “expert systems” – like behaviour in the perception of the lay public however now throws into sharp relief the prospect of low-cost NLP-powered general problem solving solutions, a kind of multipurpose social development consultancy in a box.

Prospects for Gambia, Liberia or Laos dramatically boosting human resource potential by deploying LLM-powered bots to social service frontlines are suddenly looking exciting again. Or are they?

Below I present a very condensed argument of why ChatGPT and other big-data driven models, monumental technological feats though they are, lack some vital attributes to make much of a difference soon. These concerns are not addressed by improvements in accuracy and efficiency expected in GPT-4 or later models because they go to the very roots of the philosophy behind such tools. They can only be solved in the technology governance layer.

Expert systems are heavily weighted

LLMs and similar big data-driven systems are about statistical averages. They take snapshots of internet-scale caches of data and then make safe bets as to the most likely answer to a query. The best experts are, however, top-notch for the very reason that they generate insights in the tails of the distribution. It is still too risky to introduce strong-weighting in LLMs to generate positive biases since the goal is to minimize bias in general.

Knowledge Engineering is tough

For any complex assignment, a schema of multiple parts involving multiple strands of enquiry produces the most sophisticated outcome. One needs to orchestrate multiple prompts to get ChatGPT to generate the right sequence of answers and then piece them together. This requires higher order, not lower order, cognition. Costs thus shift from expensive specialists to expensive generalists.

For example, a detailed review of ChatGPT’s recent performance on a Wharton MBA test emphasised the critical importance of “hints” from a human expert in refining the bot’s responses. As will be shown later such “prompt loops” cannot be woven unless the expert has strong generalist competence in the area of inquiry.

Factual Precision is less useful than Contextual Validity

Whilst there is a lot of general information available on the internet and from other open sources, a great deal of the world’s contextual insights are still in proprietary databases and in people’s heads. LLMs need extensive integrations to access true insights into most socioeconomically important phenomena. In a 2019 working paper, this author broke down the generic structure of modern computing systems into “data”, “algorithms” and “integrations”, and explained why integrations are the real driver of value. Yet, integrations also limit quality data growth in LLMs like ChatGPT and introduce risks whose mitigations constrain automation efficiency.

Much has been made of ChatGPT successfully passing exams set for advanced professionals such as medical license assessments. In actual fact, a standardised exam taken under invigilated conditions is the worst example to use in the analysis of real-world expertise. First, exams of that nature are based on a syllabus and wrong and right are based on well defined marking schemes designed to approximate statistical yardsticks of performance. It is considerably easier to span the universe of knowledge required to pass an exam and to reproduce “standard quality” answers than it is to operate within the constraints of a turbulent environment such as field hospital in Liberia. Whilst we use such exams to screen humans for similar jobs, the confidence arise from implicit assurances of their social adaptability in knowledge application.

Proprietary data is expensive

To address the proprietary data and tacit knowledge issues, LLMs will have to compete aggressively for integrations, raising their costs of deployment and maintenance, and thus lowering accessibility for the poor. Already, several Big Data-AI companies like Stable Diffusion are facing lawsuits for trying to externalize their costs. And ChatGPT has been excoriated for using sweatshops in Africa. The politics of data mining will constantly outstrip the capacity of individual companies to manage, just as is the case with natural resources.

As individual corporations get better at building their own unique knowledge-bases and at sourcing algorithms to address internal issues, the edge of internal-scale utility operators like OpenAI (owners of ChatGPT) will start to erode and the real opportunity will shift to enterprise consulting in a balkanised AI business environment. It bears mentioning that becoming a viable competitor to Google’s LLMs in the general knowledge search category is not a socioeconomically transformative step for ChatGPT as the real gap is in specialised knowledge brokerage.

A new kind of “naturalistic fallacy”

Many of the most genuine knowledge breakthroughs are highly counterintuitive. Some clash with contemporary human sensibilities. The more a bot is made contemporaneously human the lower its ability to shift cognitive boundaries. This is hugely important when a bot must make predictions and projections based on judgement-soundness rather than mere factual-accuracy.

Confusion in this area caused Soviet and Chinese theoretical marxists to declare AI a reactionary science well into the 70s. Whilst the Cuban revolutionaries saw only industrial automation and thus embraced AI (popular then as “cybernetics”), the more ideological marxists understood the challenge it posed to the idea of human transformation itself, on which perfected communism will depend. In that sense truly groundbreaking LLMs would have to be more unfettered in their probing of human sensibilities than current political ethical boundaries can contain.


To practicalise the above analysis, the author engaged ChatGPT to discuss the prospect of decentralized finance enabling financial inclusion in the developing world. The argument’s vindication is self-evident but a few annotations have been interspersed with the screenshots. The short version is that statistically summarised knowledge drawn from open internet resources is constrained to mimic garden-variety coffee-break conversation rather than serious expert handling of judgement-heavy, high-stakes, decision-making.

When ChatGPT is asked standardised exam type questions requiring the ability to parrot generally received contemporary wisdom, it tends to excel,
The inquirer now ups the stakes by asking for decision-making assistance in respect of an interconnected web of strategic challenges in the regulation-vs-innovation context.
ChatGPT does not skip a beat. It continues to parrot the same general platitudes, this time about enlightened co-regulation. It struggles to detect many loaded contexts and not finding many open-internet resources about West Africa – specific crypto regulation debates attempts to bloviate around the subject. Whilst this would be fine when delivered from the stage at a general business conference, it is practically useless from a real decision maker’s point of view. Improved efficiency in gathering data on the web and enhanced semantic search will not overcome these limitations. The inquirer’s efforts using prompt-enhancement to improve on the quality of the responses depended greatly on the inquirer’s own growing understanding of this LLM limitation suggesting the need for a new breed of, likely expensive, professionals to maximise this tool’s utility.

Friday, 13th January 2023. Jubilee House, Ghana’s presidential palace. In attendance: the movers and shakers of public and commercial finance in the country.

Arrayed on one side was the government, led by the Vice President, in his capacity as “Head of the Economic Management Team” (EMT). On the other side was a motley crew of finance industry representatives, from the insurance, securities, banking and related industries. The agenda: Ghana’s tottering domestic debt restructuring exercise.

Five and a half weeks since the Finance Minister announced a move to default on Ghana’s domestic debt by persuading creditors to exchange their current bonds for new, significantly lower value, versions, the program seemed hopelessly stuck. The EMT’s goal for the meeting was thus to break the logjam. Why, though, is the program stuck?

In an earlier essay, we catalogued a list of defects in Ghana’s debt restructuring/exchange model, but the focus was mainly on broad strategic issues. At the Friday meeting in Jubilee House, tactical considerations took center-stage.

When on 5th December the government announced its offer to domestic creditors to turn in their current bonds and come for new ones guaranteed to lose them billions of Ghana Cedis (GHS), it gave them two weeks to comply. No serious prior negotiations had taken place. Nothing had been agreed in principle, not to talk of anything approaching even the most high-level consensus among the biggest creditors on general terms.

Near as we can tell, no government on Earth has succeeded in pulling off such a fast turnaround as Ghana tried to achieve last December. Even the fastest restructurings, such as those of Ecuador and Argentina, have in recent decades typically taken between four and five months of consultations before the formal launch of the actual exchange process, which is then treated as a formality.

Ghana’s style is more reminiscent of Argentina’s 2020 default, which initially consisted of a series of unilateral offers and amendments in a take it or leave it fashion. At every round, bondholders rejected the offer and subsequent amendment. Not even the intervention of Pope Francis made a difference until the right set of concessions allowed the main bondholder groups to consent.

Seeing as Ghana’s main advisor, Lazard Freres, has advised Ecuador too, the likelihood of a protracted stalemate in the absence of concessions should have been clear to the government side from the outset, so why have things panned out like Argentina’s?

It would appear that Ghana’s Finance Ministry has been counting on a “divide and conquer” strategy. It has so far been nonchalant about calls to support the formation of a joint creditors’ negotiating group. Given the costs involved in obtaining top-notch legal and financial modelling advice, creditor coordination has long been a daunting prospect in sovereign debt restructurings.

The government apparently believes in keeping the creditor front fragmented and uncoordinated as a way of minimising resistance. Unfortunately, such “divide and conquer” strategies are only effective if the government could also make differentiated offers to different creditor groups or engage in selective defaults of specific classes of bonds. Neither option is open to the government in Ghana’s context, thus rendering a divide-and-conquer approach a complete waste of everyone’s time.

That fact was amply evident on Friday when the different industry groups converged to confer with the EMT. It soon became apparent to everyone in the room that the government, as of that morning, had zero concrete commitment from any major creditor group to the debt exchange.

The securities industry representatives (the folks running the mutual funds, independent brokerages and various collective investment schemes) said they were willing to step up and accept the latest amended offer if the government will countersign on a covenant promising to upgrade their settlement to match any better terms eventually given to any other group. A point which illustrates the futility of the divide-and-conquer strategy: you get an assortment of contingent proposals from every creditor group playing a “wait and see” game. The game theory analog is the famous stag and rabbit/hare hunt where agents attempt to balance the benefits of individual moves with the higher payoffs of social cooperation.

The Stag or Rabbit Hunt.
Image Source: Jensen and Riestenberg

Attempts by the Finance Ministry to solicit respect for the January 16th deadline went nowhere. Obviously, different creditor groups with their separate sets of concerns can obviously not resolve them at such a meeting when no prior efforts had been made to coordinate their claims and issues into a uniform negotiating position. Said differently, it is pointless for the government to encourage separate negotiations and yet when faced with a time crunch try and push for a quick joint resolution in a common forum.

Unsurprisingly, therefore, the Finance Minister’s offer for the creditor groups to accept the finality of the 16th January deadline and be granted a few additional days to tidy up the paperwork failed to persuade.

What is fascinating about all this is how respectful the creditor groups have been despite the casual treatment they have received to date. As anticipated from the outset, banks and “savings & loans” companies are the most susceptible to the quiet force of the government’s enormous regulatory power. So, at this point, all that separates the banks and the government from a deal are five relatively surmountable blocks:

  • The banks want the government to get concrete on the “regulatory forbearance” it has been promising so far by agreeing to a discount rate for valuing bonds as part of capital determination. The banks want a lower discount rate to reduce the valuation gap between the new and old bonds. They are inclining towards 7.5% (contingent on further engagement with the Institute of Chartered Accountants in Ghana). The Bank of Ghana insists on 12%. The choice of discount rate or factor will establish the present value of the bonds the government is tendering to replace the old bonds.
  • Linked to the above are disagreements over the “expected credit losses” from the impairment of bank assets (in the form of government securities in this case, not loans or advances) being occasioned by the proposed debt exchange.
  • Obviously, until the government and the banks can agree on how to quantify losses as a result of the exchange they cannot move on to settle the issue of tangible effects on the bank’s income statements and balance sheets. As far as the banks are concerned, the expected financial impacts are: pretax losses of $1.2 billion, liquidity shortfalls of $1.6 billion, and a capital shortfall of $1.3 billion (using retail exchange rate). Essentially, massive hits to the bottomline with troubling implications for their capacity to keep issuing credit, maintaining jobs and investing in financial infrastructure. Worst-hit financial institutions may have to let 40% of their workers go. 17 of the 23 licensed “deposit money banks” will see their capital adequacy ratio fall below the regulatory minimum. And 9 of them will experience negative equity. In short, the government is refusing to acknowledge the full impact of the debt exchange on the financial sector; yet, without convergence on this point countermeasures cannot be agreed.
  • The banks cannot countenance the stepped-up coupon (interest rate) model for the new bonds with its zero payment (and no deferral) payout structure for 2023. The banks require a simple uniform structure of equal payments across the life of each instrument. Furthermore, this uniform rate is, in their view, best set around 12.5% per annum.
  • Government’s new bond offering comes with a well known Trojan Horse: legal clauses that would make it easier for it to vary the terms of the bonds in the future. The banks want these clauses expunged.

Beyond these 5 main demands, there are a number of other areas where confusion still prevails. One such is the treatment of the bonds denominated in US dollars, whose status remain uncertain. The second is the vaunted “Financial Stability Fund” (FSF).

The government has tried to bloviate around these matters, and made many vague assurances of monies committed by the World Bank to cover 30% of the $1 billion fund size. Sources at the World Bank suggest various contingencies must first be met before any such disbursement will happen. Meanwhile, the other claimed sources of the rest of the money: the Germans (KfW), the Paris Club of rich western nations, and the African Development Bank have so far not commenced any formal negotiation of any agreement to offer any cash to this facility. Pressed to the wall, Ministers responded airily that the debt exchange cannot be held hostage by such matters and that banks should first sign up before being told even basic things such as what interest rate borrowing from the facility would attract. The industry pushed back on eligibility criteria.

At this point, the Finance Minister tried another tact. Why don’t the banks publicly announce their consent to the debt exchange program and then request a date extension to hammer out a few outstanding issues? Naturally, the bank representatives demurred. They weren’t born yesterday. They insisted on an agreement on the key outstanding issues first.

At which point, Lazard Freres barged in and sought to gaslight the eminent company. An extension can’t be contemplated without wrecking the credibility of the whole debt exchange program and furthermore threatening the IMF deal. After much handwaving, they calmed down, loosened their tie, sipped some seltzer, and grudgingly faced the reality of a no-deal on Monday the 16th of January.

All of the above is to say, the government went into what was billed as a crunch meeting to close a deal to allow an announcement one working day away when it had made very little effort to grapple with these very compact issues for two weeks now. As has been repeatedly mentioned on these pages, the government’s job has been far easier than in many comparable national contexts in similar circumstances.

Till date, domestic institutional creditors have accepted the principle of significant losses. They have also not demanded several of the concessions that elsewhere others have extracted, like negative pledge clauses, buyback transparency terms, future upside sharing and other contingent windfalls, enhanced protection etc. They even seem to have backed down on initial requests for the government, in its issuer capacity, to underwrite the legal and other advisory service costs of creditor coordination. They seem to be operating with far less legal ammo than should be the case in a situation with such large amounts of money involved. And some of them, especially the mutual funds and asset managers, are even willing to sign up provided they are granted “pari passu” assurances. And yet the government has still failed to clinch a deal.

The recalcitrance of the government about deepening consultations and accelerating coordination among creditors is completely bizarre considering its lack of options should holdouts remain above 40%. Should an industry group collectively refuse to sign the exchange papers, all the punitive measures at the disposal of the government become useless as it cannot crush a whole industry. And any attempt to default on the holdings of any significantly large group risk setting off contagion that can ruin entire sectors without sparing those who consented to the exchange.

The meeting with the institutional shareholders thus ended with a commitment to engage further through correspondence this week. It is safe to say that the government will not be able to announce any program success rate tomorrow because frankly at this stage it has no firm commitments from any major creditor group.

After the institutional creditor representatives departed, the government side retired to deliberate on the now politically explosive situation of mobilising individual bondholders.

In a previous essay, IMANI raised a caution about individual bondholders as follows:

Whilst households and individuals hold just about 13% of government debt, they are also the most politically significant group. They are the ones most likely to bring class action suits against the government and mount political agitation to stop the process as they are not as exposed to government pressure to the same extent as the banks, institutional funds and treasury departments of large companies.

From all indications, the Vice President and key allies of his in government recognises the scale of the political blowback that reneging on the promise not to add individuals and households to the debt exchange program has triggered. Whilst the Finance Minister is at this stage focused purely on securing enough liquidity relief so that his already fragile fiscal program for 2023 does not fall apart, and, even more vitally, the IMF program is not derailed, the Vice President has a somewhat more strategic perspective in mind. Unsurprisingly, the mood in the government is tilted towards restoring the exemptions for individual bondholders, notwithstanding the discomfort of the Finance Minister and his technical advisors.

As we have frequently said on this site, the Finance Ministry’s lack of political economy savvy, now compounded by overdependence on foreign technical advisors with zero awareness of how social consensus is achieved in Ghana, is a major threat to both the debt restructuring effort and the IMF program.

It is actually a miracle that both programs are still technically on track and are likely to survive in highly constrained forms. The total amount of liquidity relief likely to be generated by the debt exchange program is now estimated to be lower than 40% of what the government set out to achieve. To preserve the credibility of the IMF program, additional fiscal consolidation in the form of expenditure cuts is now inevitable. Even after exempting pension funds and contemplating re-exempting individual bondholders, the government still faces wildcat litigation risk because offshore investors who hold local bonds are unwilling to take the deal on offer.

The adamant refusal of the government to mobilise national sentiment behind its economic recovery strategy and to build a broad social and bipartisan consensus behind the measures has led to highly suboptimal outcomes guaranteeing a significant delay in Ghana’s effort to exit the country’s biggest economic crisis in 40 years.