I know it has been a while since I touched the Inside the Room series, but honestly, life got in the way, as it tends to. There’s so much that’s happened in the local news since my last post, including, of course, the Communications director of my absolute favourite state institution, the Ghana Water Company Limited, announcing not only that the utility was struggling financially but also that “staff don’t live lavishly and nobody drives new cars”.

The self-deprecation came right before they announced interruptions in water supply across Tema, Ghana’s industrial city.

You can never be bored in sub-Saharan Africa.

I can’t tackle too many issues at once, especially because I’ve intermittently been writing this piece. I’ve been thinking about Agyapa Royalties again because gold prices have moved far beyond anything anyone imagined when the transaction was designed. That sent me back to what happened inside the Finance Ministry, and to a question I don't think has ever properly been answered, i.e what might the company have become had the listing gone ahead?

I’m not going to attempt to relitigate every allegation surrounding Agyapa Royalties. Ghanaian readers know it was in court for a while. There were accusations of cronyism, plus the Special Prosecutor’s dispute with the President over the transaction became deeply personal and extraordinarily public. It got messy, and I am not the right person to adjudicate all of that.

What I can describe is why the transaction initially made sense to some of us inside the Finance Ministry, what we underestimated, and why by the time we began explaining it, it was probably already too late.

The wider issue at hand, though, is about resource-rich countries like Ghana trying to get the most from their natural resources. With that, I think we can move to talk about the last Government’s attempt at this. Agyapa Royalties.

How I Got Pulled In

When Agyapa first exploded into public view in August 2020, my immediate thought was: thank God I’m nowhere near that thing!

I had enough problems to worry about.

Two weeks later, the Deputy Minister of Finance at the time, Hon. Charles Adu Boahen, called and asked me to come to his office.

My joy was short-lived.

2020 was already a strange year. COVID had arrived in Ghana in March, and like most governments, we were responding quickly and hoping for the best. In between managing the fiscal fallout, I had been working across several fronts, from helping design the country’s COVID stimulus programme to supporting the renegotiation of commercial contracts with independent power producers.

The work was a lot, and at some point I convinced myself that if the virus killed me, at least my epitaph would read: he worked really hard in his last days.

In August though, the headlines briefly pivoted away from covid infection counts. Details of an Agyapa Royalties IPO had surfaced on social media, and a coalition of civil society organisations began mobilising against it.

By August 14th, Parliament approved the transaction agreements on the last day of the session. The Minority parliamentary party staged a walkout, complaining they had been given less than four hours to scrutinise the deal. The backlash was immediate and loud, and after the call I received, I was now going to be drawn right into the middle of it.

Hon. Charles is someone I always liked working with. As stressful as his assignments could be, I looked forward to them because I knew I would always learn. He had a scary amount of investment banking experience and seemed to know about everything, except maybe football. On this occasion, he told me a press conference was being scheduled to address the public concerns, and he wanted help putting together a deck to explain the transaction.

He handed me a set of documents and told me to get across them fast. That’s all I needed to hear.

Hidden in Plain Sight

I went into the documents that same day, and I was genuinely surprised at what I saw.

The first surprise was that the transaction was, unlike the assertions by the Minority, not new at all. It had been announced as far back as two years beforehand!

One thing I eventually learned about public sector work is that Governments can never be found guilty of hiding their intentions. Civil servants will always point you to some random page in a 70,000-word document and say “Look! Everything was explained here!”, knowing full well no one is going to read more than 100 pages into any government report.

For Agyapa, it had been hiding in plain sight on page 156 of the 2018 budget statement.

He said the same thing in 2019. In two consecutive budget statements, he had been describing this policy; the only thing he didn’t do was mention the word “Agyapa.”

The 2019 budget statement even put a number on it.

The legal foundation had been laid through the Minerals Income Investment Fund Act in December 2018, and work on the company structure, mining-licence due diligence and engagement with UK regulators had been underway since then. The amendments to the Act and the transaction agreements themselves would come later, in 2020.

At the time, I figured at least one journalist must have caught on to this policy in the last 2 years. A quick media scan turned up plenty about how much had been allocated to the national football team for the 2019 Africa Cup of Nations, but on the Government’s intentions for gold royalties, silence.

What the Deal Actually Was

I’ll try to explain what the transaction entailed. I’m six years late, but better late than never.

Ghana earns royalties from gold mining. Every year, mining companies operating in the country pay a percentage of their revenues to the Government.

Typically, these royalties flow through the Ghana Revenue Authority into the Minerals Income Investment Fund (MIIF). In 2019, twelve producing mines were placed in the initial Agyapa portfolio and were generating around $144 million a year in royalties, up from $133 million three years earlier.

The gold belt areas of Ghana have some of the most resource-rich land on the continent, but once you drive around those areas, it’s quite sad. Development matches nothing like the billions of dollars being produced in gold exports. I remember one of the political justifications for the transaction was that it would help close this gap.

The initial plan was for the Government to move forward as follows.

Step 1: Assign 75.6 per cent of the royalty flows from those twelve mines to a new company, Agyapa Royalties Limited, incorporated in Jersey and set up to list on both the London Stock Exchange and the Ghana Stock Exchange

Step 2: Go to international capital markets and monetise the future royalty stream upfront. This would involve selling up to 49 per cent of Agyapa’s shares to investors in exchange for an upfront capital injection of roughly $500 million to $750 million, while the government retained at least 51 per cent through MIIF and remained the majority shareholder.

Step 3: Watch the money roll in.

I’ll try explaining the flow of funds with a diagram so I don’t lose anyone.

Practically, here’s how it would work. Mining companies pay royalties to the Ghana Revenue Authority (GRA) . GRA passes them to MIIF. MIIF then splits the pot: 20 per cent goes to the Minerals Development Fund, then 75.6 per cent of the royalties from the twelve designated mines goes directly to Agyapa. After that, a small 2.4 per cent goes back to GRA as a revenue management fee, and 2 per cent stays with MIIF to cover its operational costs. That’s it.

The model itself isn’t unique, and i remember Franco-Nevada, a Canadian royalty company that grew from $2 million in capital to a $28 billion market cap, kept coming up in the documents I was given. Wheaton, Royal Gold and Osisko were other comparables. The pitch was that Agyapa, as Africa’s first sovereign gold royalty company, had the potential to grow into something similar over time.

Crucially, the whole transaction was framed as non-debt financing, i.e Ghana wasn’t borrowing money. This was an equity play, which in 2020, with COVID having wrecked the debt markets and Ghana already carrying a heavy debt load, mattered enormously. Of course, non-debt did not mean costless, as investors would receive a share of the royalty income and future upside.

After reading the documents handed me, I understood the urgency. Gold prices in August 2020 were at their highest in nearly a decade, and the government wanted to lock in investor interest at a premium moment. For that reason, two years of quiet preparation had suddenly been accelerated.

At the time, only two concerns stood out to me.

The first was the valuation. The figure being cited publicly was $1 billion, which sounded low, but the explanation I got from the Minister about it was direct: “The $1 billion is an accounting fair value, constructed for the purposes of creating a balance sheet, not a market valuation.” The overriding assumption was that the market valuation on listing would be significantly higher once investors priced in the earnings potential of the asset, because listed royalty companies globally trade well above net asset value.

In other words, $1 billion wasn’t Agyapa’s expected market capitalisation at listing, nor did it mean that 49 per cent of the company had already been priced at $500 million. The key phrase in the agreements was “up to 49 per cent”, meaning that was the maximum interest that could be sold, while the actual valuation and flotation size were supposed to emerge through the roadshow and IPO process.

The second thing was the choice of lawyers for the transaction. Among the documentation I reviewed, Africa Legal Associates appeared as one of the legal advisors on the deal. Africa Legal Associates is a firm co-founded by Gabby Asare Otchere-Darko, who was a cousin of the President. I’ll leave that there.

The Press Conference and What Followed

That was the context in which a press conference on August 27^th 2020 was called. We were not proactively introducing the policy to the public; rather, we were responding to a fire that was already burning.

I was calm walking in. After all, the information being presented made sense to me. The whole session felt, internally, like it had gone reasonably well. We had explained the structure and as far as i was concerned had been clear that government would remain the majority shareholder for the transaction meaning Ghanaians were not being shortchanged.

I was extremely wrong.

First off, the timing for the whole thing was terrible: a general election was about three months away, and the narrative that the Government was to mortgage the country’s gold before possibly leaving office took hold fast. The story quickly shifted from being about the merits of a gold royalty company to being about cronyism and conflict of interest. (Remember the thing I said about the President's cousin?)

It was crazy, and the opposition party latched on to it even more. Anyone formally connected with the transaction was labelled with all manner of unflattering terms. “Thieves!” was a common one.

The CSO Circuit

The Ministry of Information organised a public FAQ session not too long afterwards. We basically had to explain our explainations again, which was even more proof that things weren’t going well.

I wasn’t there that day, but I was told the room was full. Unfortunately, the report I also got was that attendance was mostly of people already sympathetic to the Government. We were explaining a decision to people who did not need convincing.

The public backlash continued and only grew louder.

A more significant meeting happened on September 2nd. A group of civil society organisations came into the Finance Ministry to meet the Minister directly. I went in with some optimism. After all, if this particular constituency heard our full explanation, the concerns would soften.

Nope.

The meeting was tense but very civil. One guy from the Africa Centre for Energy Policy stood out to me. I can’t remember his name, unfortunately, but what distinguished him from the others was that he both criticised the deal and proposed alternatives.

He was convinced the $1 billion valuation was wrong and argued repeatedly that the Government should defer the listing, get another independent valuation done, and come back with numbers that reflected what the asset was actually worth. I remember he even offered to help with that process.

The Minister listened but didn’t change his position. The deal was going to go ahead as-is. At one point, as I remember it, when the valuation issue was raised again for the umpteenth time, the Minister firmly responded with: “We just have to agree to disagree.”

I had walked in hoping for a meeting where everyone left with some common ground. That was not the meeting we had.

A week or so later, we met faith-based organisations, and later the National House of Chiefs on September 25th. Both were polite; both hinted the Government should delay; neither offered real endorsement.

Multiple Failings

Looking back, there was a structural problem with all these engagements. We were going around to explain a decision that had already been made to stakeholders who had not been brought in while the policy was being shaped. We used to call it consultation, but it was far from it. The more I think about it, the more it seems we were basically giving announcements and expecting people to just accept whatever was being said.

While finalising this article, a friend sent me a Twitter(X) link to a post that made me laugh and then sigh in short order because it appears the current Government is making the same “consultation” mistake, this time with the cocoa sector.

Another failing with the transaction was that the actual institutional anchor, the Minerals Income Investment Fund, never became the recognisable institutional face of the transaction. The CEO at the time rarely spoke about the deal, while the Finance Minister became the principal public defender of the deal, reinforcing the impression that Agyapa was a political project rather than a professionally governed state investment vehicle. I’ve written about my issues with Ghana’s state enterprises before, and in this instance, leadership failings made an already difficult fight even harder than it needed to be.

The November Collapse

As October arrived and the election drew closer, the noise around Agyapa gradually faded from the front pages as the national conversation shifted to manifestos, rallies and the usual election-season noise. Work was still technically going on in the background though, there were still internal discussions and timelines being maintained on paper but the energy had drained out of it. A transaction that needs public confidence to list on an international stock exchange can’t survive the kind of reputational damage Agyapa had taken in those two months.

November was more eventful but for all the wrong reasons. Following the Special Prosecutor’s corruption risk assessment, the President directed that the agreements be returned to Parliament for further scrutiny. Two weeks later, Martin Amidu resigned as Special Prosecutor, with Agyapa at the centre of the dispute. With a month to go before elections, the transaction had become a national argument about corruption, institutional independence and trust, with no realistic prospect of an international listing.

In essence, two years of preparation, all for nought. With the added bonus of major political capital lost.

What I Think Now

There’s a number of things to get through here as I’ve had a lot of time to think about this deal.

First off, I believe a lot of the confusion could have been resolved with earlier consultation. The lack of proper consultation meant several distinct numbers got blurred together in the public debate: the $1 billion accounting fair value, the up-to-$750 million fundraising target, and the market valuation that was supposed to emerge through the IPO process itself.

I also went back and tried to answer a different question, i.e “If the transaction had gone ahead, what might Agyapa be worth today?”. Using the original 2020 valuation documents, I reconstructed the valuation framework from the assumptions disclosed, i.e. the same production schedule, the 75.6 per cent royalty allocation and the same contractual sliding-scale royalty rates.

On top of that, I swapped in the actual path of gold prices since 2020 instead of the original forecasts and checked the result against how listed royalty companies trade today. I’ve written on counterfactuals before, this time though, I’m relying less on econometrics and more corporate finance.

One thing the exercise reinforced was that the $500 million to $750 million fundraising target came from selling up to 49 per cent of the company, the maximum contemplated under the proposed structure and not a valuation of the company as a whole. At a higher market valuation, Government could have raised the lower end of that range while selling considerably less than 49 per cent. That was entirely possible.

Rebuilt with actual gold prices and today’s royalty-company multiples, Agyapa as a whole could plausibly be worth $3.3 billion to $4.2 billion today, with a central estimate close to $3.7 billion.

Had the deal gone ahead, a 25 to 30 per cent interest would be worth roughly $800 million to $1.3 billion at that valuation. Indeed, the maximum 49 per cent interest would be worth roughly $1.6 billion to $2.1 billion. On the same assumptions, the structure could have generated something like $1.3 billion in cash between late 2020 and 2025, before any decision on dividends or reinvestment.

To be fair, none of this proves the deal was flawless. Not even the original architects of the transaction could have known how sharply gold would rise over the forecast horizon. What it does show however, is that if the original production assumptions had broadly held, the subsequent rise in gold prices experienced in the last few years would have created substantial upside for investors and, by extension, for Government alike, with Government keeping the clear majority of it under every flotation size on the table.

Ultimately, the concept itself wasn’t wrong, and yes, not every feature of Agyapa was sound, but after running the numbers again I’m convinced a properly structured gold royalty company could make sense for Ghana or another resource-rich African country.

The Bigger Question

Africa keeps having this discussion about what to do about natural resources, especially because there can be remarkably little to show for them in terms of development outcomes.

The real issue I think revolves around how smartly to convert resource earnings into something durable. Agyapa was an attempt to do something different and convert a stream of mineral income into durable national assets without surrendering all the future upside. Quite frankly, it never got the opportunity to prove whether it was the right answer.

On a more sombre note, I also learned that even a technically credible policy cannot survive a process that brings people in only after the decisions have been made. By the time the Government brought coalitions together to explain Agyapa, the public had already reached its verdict, and the policy never recovered.

Perhaps its failure was useful in one respect. Any future attempts by an African government to rethink how it manages its natural resource earnings can learn from Ghana’s experience and make sure their solution is designed more openly, explained much earlier and built on public trust from the beginning.

Subscribe now

Share Benjamin Hornsby-Odoi

My last article, “Power Sweet”, was about the institutions and structures put in place by the State to make daily improvements in people’s lives. Utility companies, state transport companies, etc. There’s a whole universe of them in Ghana referred to as “specified entities”.

The harsh reality is that when those institutions don’t work, survival itself gets quietly outsourced. I can speak mostly to the Ghanaian experience, though there’s enough evidence that it’s the same across much of Africa.

These failures visibly show up across households with things like sachet water, generators, boreholes, and even private security in some places. The ramifications of these failures seem to just accumulate, household by household, until one day the informal version is bigger than the official one. I’ll explain how in a bit.

Sachet Water

For readers outside Ghana, the entire concept of sachet water might be unfamiliar. Essentially, they are little pouches of water sold at very low prices for everyday people. Bottled water is out of reach for many, so the sachet fills that gap. Unfortunately, their presence in Ghanaian households has gone beyond being just a cheaper alternative to bottled water.

The 2021 census highlighted that sachet water had overtaken the public pipe!

Yep.

Nationally, 37.4% of households relied on it as their main source of drinking water, against 31.7% for pipe-borne water. In urban areas, the gap is wider still: 51.5% versus 33.6%.

I'm not sure when these sachets were introduced to the market, but it's safe to say that over the past decade, successive state failures have made them a mainstay in households. And in moments of real shortage, some households stretch sachets beyond drinking water too — for washing, bathing, and whatever else gets them through the day.

The refrain is simple- “There’s no water flowing from the taps, so just buy a pack of sachets to get by”. Nobody voted for this. No policy declared that Ghanaians would outsource their basic need for water to 500ml plastic bags from street vendors. It happened one dry tap at a time, and it happened on a national scale, and we now talk about it like it’s just how things are.

This piece looks at the company on the other end of that dry tap. Ghana Water Company Limited. I’m not singling them out because they are “uniquely bad”, but more because it’s a clean example of the mechanism I was trying to describe in the last piece, visible from inside a single institution — a system where being held accountable to the public mandate was never really part of the job.

The Gap Was Always There

Back in the mid-2000s, Ghana’s own development plan set an urban water coverage target of 58%, year on year. The actual numbers came in at 55%, 56%, 58%, 55%, and 56% across those five years — missed in four years out of five, by a margin small enough that closing it should have been routine.

Twenty years later, the gap hasn’t necessarily closed. As I mentioned earlier, it’s been filled by the sachet economy instead. Every stack of sachets outside a shop is a household that gave up waiting for the pipe and built its own workaround — a parallel water utility run by thousands of small operators, while GWCL continues to report numbers on its own performance to a fed-up public.

Even where the pipe technically reaches a house, "coverage" doesn't mean water. By one estimate, only about a quarter of Accra households get continuous supply; roughly a third get it for 12 hours a day, five days a week; another third get it two days a week; and the rest get nothing through the pipe at all.

In some neighbourhoods, residents have described staying awake until two in the morning, waiting for the moment the taps might run so they can fill barrels and tanks before they stop again. For many households, sachet water is the thing you buy on the days the 2am wait doesn't pay off.

The Numbers Depend on Who’s Counting

I’ll try to go a little bit deeper, but I'll start with what GWCL does well. In 2020, it produced 325 million cubic metres of water against a target of 312 million. To be honest, production isn’t the problem. Their plants work, mostly, and the people running them try to hit their numbers.

But water has to travel three steps before it becomes anything for a household, or revenue for the company. It has to reach a connection. It has to get measured. And it has to get paid for. Unfortunately, GWCL’s numbers get softer at every step.

Take coverage. GWCL’s own latest figure puts urban access at 84.2%. A 2024 academic review, citing GWCL’s own 2019 data, puts it at 77%. WHO and UNICEF have at times credited Ghana with much higher urban water coverage, although I have to add that these numbers don’t always measure exactly the same thing, and that is part of the problem. “Access” can mean a basic, improved source. It can mean the utility’s estimated service reach. It can mean a pipe nearby, a pipe on the premises, or a pipe that actually runs when a household needs it. Depending on the definition, Ghana’s urban water story can look like a near-success or a quiet failure.

Of the water that does make it into the pipes, close to half never reaches a paying customer. Non-revenue water — leaks, illegal connections, unmetered taps — has hovered near 50% for two decades, under a private management contract and back in public hands. In effect, it’s not as simple as saying privatise, and your questions will be solved.

And of what does get billed, GWCL collected 86.2% in 2020, against a target of 97%.

I’m hoping the above visual helps to simplify things.

Once you run the chain end to end, you realise:

Production: on target.
Coverage: contested, depending on definition.
Non-revenue water: roughly half.
Collection: short.

Each step softer, each number harder to pin down than the last.

That gradient shows up in the financials. Revenue nearly quadrupled between 2015 and 2020, from GH¢384 million to GH¢1.34 billion. But in 2019, GWCL lost GH¢2.62 billion — more than its entire revenue that year. Liabilities grew at an average of 60% a year between 2016 and 2020, reaching GH¢6.07 billion, mostly long-term loans. By 2024, finance costs alone had reached GH¢3.64 billion, driven by foreign-exchange exposure on dollar-denominated debt.

The truth is tariffs have increased repeatedly over time, often under pressure from cost-recovery requirements and IMF-supported reform programmes. The deeper problem is that when collection, losses and governance don’t improve enough, tariff increases can’t close the gap. So the gap gets filled with loans instead. That’s the first shortcut taken to cover the institutions’ failings.

The loan first shows up as progress — a project announced, a contract signed, the cash crunch eased. Years later, it returns as a foreign exchange problem that dwarfs the gap it was meant to close. A good example of this point is the $126 million desalination plant in Accra's Teshie: brought in hurriedly under a PPP arrangement, more than a decade on, GWCL still owes a monthly dollar-denominated capacity charge to the private operators whether the plant runs or not. The plant was even switched off again in late 2025 amid unresolved contractual, payment and maintenance disputes while residents in Teshie waited for water to flow from the taps.

Nobody’s Holding the Line

When I was in government, GWCL’s board sat at around eleven members, and the managing director’s role followed the same pattern as every other major SOE — a seat that came with, and depended on, political alignment, not technical know-how. To the chagrin of permanent staff. I’d wager the structure looks much the same today, just with different names attached to a different party.

That’s the point I’m trying to highlight underneath all this. If the person in charge of a utility company is part of the political reward system, they have two jobs. One is written down: fix the water system for millions of people. The other isn’t written anywhere, but it’s the one that actually gets evaluated, i.e., don’t embarrass whoever put you there.

When those two jobs conflict, which one wins isn’t really in question.

The Pattern Repeats Because There’s Limited Accountability

In 2009, AVRL, a private operator, was brought in to run Ghana’s urban water system and given a specific mandate: reduce non-revenue water from a baseline near 50% over five years. Performance fell well below expectations, especially on non-revenue water, and the contract wasn’t renewed after 2011. As far as I can tell, nobody was sanctioned. A campaign group called for AVRL to be held financially responsible for non-performance. Nothing came of it.

I don’t think this happened because nobody at GWCL or in government knows that story. I believe it happened because knowing it carries no cost. Targets get written because writing a target is free, and is exactly the kind of thing that brings in donor financing and the appearance of a plan. Whether one is met is a separate question, decided by a separate set of incentives, answered to a separate audience.

In principle, SIGA, the State Interests and Governance Authority, exists precisely to discipline such underperformance across the SOE portfolio. In practice, SIGA’s own leadership is appointed through the same process as everyone else’s. The body built to break the cycle sits inside the cycle. And like I mentioned in my last piece, sometimes criticisms from within government can just be waved off as “technicalities.”

Multiply this problem across dozens of boards in the specified-entity universe, all drawing from the same pool of political reward. Unfortunately, the reason the political class rarely enforces real consequences is simple: everyone expects to benefit from the same system when power changes hands.

Today’s critic may be tomorrow’s appointee, inheriting the same unaccountable seat. The opposition can make noise about the water problem, in particular when the other side is in office, but it has little incentive to destroy a structure it hopes to control later.

The moment accountability becomes real, it applies to everyone, including whoever is currently doing the criticising. So, the silence holds, and it’s rational for everyone except the millions of people waiting on the other end of the pipe.

Where This Leaves Us

Behind my long rant, I’m trying to make the point that these purchases of sachet water are essentially no-confidence votes. Not dramatic ones, just small, daily decisions by households that have learned not to trust the pipe.

Like I mentioned at the outset, the logic isn’t unique to Ghana, or just to the provision of water — it’s the same one behind a generator humming outside almost every business in Lagos, or a borehole next to a house in Johannesburg that’s supposed to be on the grid. I recently saw an article about households in Joburg who set up their own network to deal with a water crisis. That’s exactly what happens when institutional failures mean survival becomes privatised.

Nobody votes for any of it. But over time, in many developing countries, the workaround becomes the system.

This is the first of at least two pieces. I am fairly certain I will upset a few of my old colleagues along the way, but c’est la vie.

A few days ago, SIGA, the authority that oversees Ghana’s specified entities, including state-owned enterprises, joint venture companies, and other state entities, issued a statement praising the Tema Oil Refinery (aka TOR), Ghana’s longest-running state-owned oil refinery.

TOR had finally cleared a six-year backlog of audited accounts and submitted financial statements to SIGA covering 2019 through 2025, all at once. The same statement announced a profit before tax for 2025 of about 1.24 billion cedis (approx. USD 105 million), the refinery’s first profit in a decade. Add to that, TOR’s managing director had collected a leadership award a few days earlier.

Now, I don't want to nitpick too much, but the truth is that profit is largely driven by a non-operating accounting gain. About 1.3 billion cedis of it, roughly US$110 million, was a result of a foreign exchange gain booked because the cedi appreciated. Essentially, the majority of the pre-tax profit, and then some, was the result of a currency swing rather than money actually made refining oil. But it is still good news, and I will take good news where I can find it!

And honestly, the current managing director has done something that a long line of previous managers could not do. He got the refinery working and has cleared a backlog of accounts that had been left to rot! During my time at the Finance Ministry, I sat in on a number of presentations with former TOR managing directors and a bevvy of consultants, each one pitching that the best way forward for the entity was to either privatise or bring in a strategic partner, or some elaborate variant of the two. This current team took the tools they had and actually used them. Good.

But underneath all of it, I feel people really need to ask themselves, Why are we celebrating a state enterprise for submitting its annual reports? Or even for turning a profit?

Filing your accounts is the law. Every state enterprise is required to do it, every single year, and a company that finally manages it has not climbed a mountain. It has cleared the lowest bar we set. The size of the applause tells you exactly how low that bar has sunk.

And it is an old bar. Older than me, older than most of us reading this.

A bar older than the country

Ghana has been struggling to make state-owned enterprises (SOEs) work since independence. Back in 1965, the scholar Robert Dowse examined young Ghana’s enterprises and found “signs of obvious industrial and commercial mismanagement are everywhere. Of the thirty-two entities the state had sunk forty million pounds into, only two turned a profit.”

Fast forward to 1983, and another academic described Ghana’s SOEs as “budgetary burdens, squandering more of the state’s resources than they ever returned.”

Same country, different decade and essentially the same sentence.

It helps to go back to the beginning to understand why this happens. I recently finished reading The Second Emancipation, a book on Nkrumah and Pan-Africanism that was recommended by Ken Opalo a while back. Howard French essentially sketches the source of the rot behind these SOEs.

In the book, he rightly identified that there was almost nothing you could call a private economy in local hands during Ghana’s early post-independence years. If you had ambition, the only ladder worth climbing led through the state. A post in the civil service. A contract with a minister’s signature on it. Proximity to whoever held power.

Holding office and growing rich slid quietly into the same project, and you could watch it happen, in the new houses and the new cars that gathered around men who had only just arrived with jobs in government. People had a phrase for it in the 1960s, said with equal parts envy and contempt. They are the two words at the top of this piece. Power Sweet.

That is the soil in which our state enterprises were planted. Viewed through that lens, these institutions are really just limbs of political power, and executive roles can be prizes handed out to party loyalists rather than to the most competent people. Once you see that, the losses stop looking like a run of bad luck.

The speech

By 2021, I had run clean out of patience with the number of requests for financial support that crossed my desk. You would not believe how easily some of the head honchos behind these enterprises lost money and then simply wrote to the treasury for cover.

Meanwhile, some of these guys were living really good lives!

My boss once told me a story about a flight he was on, where he was sitting in business class and spotted a CEO of an SOE all the way up in first class. The man, realising the Finance Minister was behind him in a “cheaper” cabin, walked back to him and kindly offered to swap seats. The gesture made things worse.

The whole “event” annoyed my boss to no end, and I had to endure long rants about how some of these SOE heads were being paid higher salaries than the President, yet delivered little in their jobs.

That, plus a whole lot of other frustrations, went into a speech I wrote for the 2021 SIGA performance contract signing ceremony. As I mentioned earlier, SIGA is the body tasked with overseeing these entities, and the ceremony was their grand annual affair, meant to “evaluate and celebrate the performance of Ghana’s State-Owned Enterprises and Specified Entities.” Hold onto that word, celebrate. Once we get into the numbers, you will see there is actually very little to celebrate, and yet the event comes round every year, with the President in attendance and most of the SOE chief executives in the room.

On this occasion, in front of those boards and CEOs, the speech laid out much of what was wrong. My boss read it out word for word. Loans the state had poured in and would never see again. The money diverted to keep these companies breathing in the middle of a pandemic. He spoke about how 47 entities had not filed a single financial statement in 6 years, and even that only 14 of 129 had submitted quarterly reports. Shameful.

He even read the line I was proudest to have written- “Our public enterprises are not vehicles to enrich a minority. They are intended to be arms through which the State promotes equality of opportunity and gives its citizens the tools to shape their own lives.”

The people it was aimed at were in the room. And you’ll never guess what happened next.

Nothing.

Most of the CEOs in that hall served out their terms and walked away, well paid during their time in office, while leaving millions in losses behind at their various institutions. Some even managed to fail upwards and get promotions.

Moreover, within days of the speech, the then Director-General of SIGA, the very body meant to hold these entities to account, went to the media to explain it all away. He argued the issues my boss had spoken about were down to COVID-19 and “technicalities.”

Just like that, the whole thing was trivialised.

Reporting that runs the wrong way

At the time, the team responsible for gathering financial reports and writing the analysis for the publicly released State Ownership Reports was the Ministry of Finance’s Public Enterprises and Assets Unit (PEAU). Good people, civil servants, I genuinely enjoyed working with, who did good work under real constraints. But picture the job. You call a state enterprise asking for data. It does not come. Why would it? The chief executive does not answer to “some unit” in the Finance Ministry or even SIGA, for that matter. He answers to the one person who appointed him: the President.

That is the whole problem in a single sentence.

In our system, these enterprises report upward to power, not outward to the public that funds them and depends on them. So how could you ever expect real accountability? When the government changed at the start of last year, the boards changed with it. Not because of performance, but rather because that is how the machine works! These are not entities made to serve the public. They are for whoever holds power this season.

Which is why nothing in the numbers really moves.

The SOE sector lost 5.16 billion cedis in 2019, roughly US$438 million at today’s exchange rate. The reforms came, the speeches came, and by 2021 the loss was down to 1.74 billion cedis, about US$148 million, and a few of us let ourselves feel hopeful. Then 2022 arrived, the cedi collapsed, and the loss blew out to 14.4 billion cedis, about US$1.22 billion. By 2024 it had settled back at 9.68 billion cedis, about US$823 million.

All that time and the bottom line is roughly where it started, if not worse. There is even a small lesson buried in that chart. The 2023 loss was first reported as 2.57 billion cedis, about US$218 million, and sold to us as a recovery. A year later, the next report put the same year’s loss at 7.1 billion cedis, about US$603 million. The good news did not survive the following year’s accounting.

A normal company that loses money for long enough either changes or dies. Ghana’s SOEs do neither. They live on. They appoint new boards. They cycle chief executives through a revolving door. The losses just become a permanent feature, something to be managed and explained away rather than fixed. Technicalities!

What worries me the most is the cast of repeat offenders. We are not talking about an obscure printing house or a sleepy state guest lodge. Nope. It’s entities like the Ghana Water Company, the company meant to put clean water in taps, averaging a loss of 1.25 billion cedis, about US$106 million, every year between 2017 and 2021. Or COCOBOD, the institution holding up the livelihoods of hundreds of thousands of cocoa farmers, whose financial struggles have become one of the recurring shadows over the sector.

I won’t even speak about the Electricity Company of Ghana. I just endured a power cut for the last few hours; I’d rather not upset them any further.

Back to where I started

Which brings me back to where I began. To the applause for a filed report.

The number of specified entities submitting some kind of accounts has climbed steadily, from 18 in 2016 to 152 in 2024. SIGA presents that as progress, and on its own, it looks like it. Now follow the line underneath. The share of those filings that were fully audited, excluding audit drafts, has fallen from 68 per cent in 2021 to 35 per cent in 2024. More than half the books are now unaudited management accounts. So the headline has improved, but the assurance behind the headline has weakened. More filing. Less checking.

A lot of what we think we know about these losses comes from management accounts that no independent auditor has signed off on. We are essentially measuring failure with numbers nobody has really checked.

The impact of all this on ordinary people is the thread I want to pull next. In part two, I want to follow the cost downwards, from the balance sheets to the generators, boreholes and private coping mechanisms ordinary Ghanaians now pay for themselves. Because after a while, it stops being a story about balance sheets and becomes a story about how the failure of these institutions meant to deliver public services impacts the lived experience of the average Ghanaian.

I think we have, without ever once voting for it, privatised our own survival. These institutions fail us as a matter of routine, and yet the boards keep being constituted, the luxury vehicles keep being bought, and the cycle rolls on much as it has since 1965.

I’ll end by acknowledging that not every state enterprise boss in my time was a villain. Some were genuinely good, committed people who left a real mark on their institutions, and by all accounts, the current managing director of TOR may well be one of them. But that is precisely the trouble. These people are the exception, not the norm.

For today’s article, I want to talk about Zambia. There’s stuff happening there that’s genuinely interesting, and I’m excited to see them get things right on the social policy front.

Ghana and Zambia go back a long way. The pan-African movements of the early post-independence years connected our countries in ways that ran deeper than diplomacy. On a more personal note, I’ve also met some really cool Zambians along life’s way. (Hi Diana!!)

We're usually friendly nations. Visa-free travel between us. Growing trade ties. Two presidents who, earlier this year, made headlines together for the most delightfully African of reasons.

But behind the jokes, there's a real shared interest in getting development right. And I believe education is one big place where we can actually learn from each other. Last week, Zambia's Parliament voted to make free education a legal right - enshrining in law a policy it launched in 2022. Ghana did something similar back in 2017 with the Free Senior High School policy. Both countries are practically making the same bet: that removing fees would unlock something transformational in their education systems. Ghana's results are in. Zambia's story is still being written. Now would be a good time to share ideas.

Before the Policy

Zambia is a country that has always taken education seriously, even when the resources weren’t quite available. Post-independence, Kenneth Kaunda made expanding access to education a national priority and argued that education was the surest way to national development. Free education was mostly the order of the day until the IMF’s structural adjustment programmes in the 1990s introduced cost-sharing, and fees became a persistent barrier - particularly at the secondary level, where enrollment lagged badly behind primary.

Fast forward to 2020, and the official numbers from the Zambian Ministry of Education told a slightly more sobering story. Secondary school gross enrollment sat at around 45%. Essentially, more than half of secondary-aged children weren't in school. Primary enrollment was near-universal on paper, but completion was another matter.

The teacher situation was no better, with around 120,000 teachers across both primary and secondary levels for 4.3 million students - a ratio of roughly 35 pupils per teacher.

Textbook availability wasnt great either: 7 learners sharing each book at the primary level, 13 at the secondary level. Plus, in rural and remote areas, the situation was significantly worse than the national average.

And underneath all of this, a learning crisis that predated any policy change. UNICEF estimated that only 4% of Grade 2 students could read at grade level. Grade 5 students averaged 35% in English and 31% in Maths. This is the baseline Zambia was working with when it launched free education in January 2022.

Ghana’s Experience

Ghana launched Free Senior High School in September 2017. The ambition was clear: eliminate fees at secondary level and get every child who earned a place into a classroom. It was bold. It was popular. And as I analysed in my last piece, in terms of access numbers, the policy worked. Kids came into classrooms.

But access and learning are two different things. And here is where Ghana’s experience gets complicated. Ghana’s exam story is more complicated than either side admits. WASSCE outcomes improved in some years, particularly between 2020 and 2023, but those gains were uneven and volatile - difficult to separate from changes in exam preparation, supervision, and the double-track experience itself. The deeper lesson is not that Free SHS destroyed quality but that access gains are fragile when the supporting system is stretched.

Schools became overcrowded fast. The government’s response was the double-track system - splitting students into two cohorts on alternating schedules. It bought time. It didn’t solve the problem. Moreover, teacher allocation widened regional inequalities, and the programme's funding mechanism ran into early trouble, with arrears accumulating for suppliers of uniforms, textbooks and other materials that schools relied on to function.

Ghana delivered on the promise of access. The quality question remains unfinished business. I’ve covered the full picture in my last education article. That gap between what the policy promised and what the system could actually deliver is the lesson Zambia needs to sit with.

After the First Surge

Let’s look at the numbers for Zambia though.

Since launching free education in 2022, Zambia has done something genuinely impressive. Around 2.3 million additional learners are now in classrooms. The government recruited over 41,000 teachers - the largest hiring drive in recent history. Secondary gross enrollment jumped from 45% to 63.7% by 2023. Grade 12 pass rates rose from 62.4% to over 70%. Any government would hold these numbers up as signs of actual progress. It’s really cool to see things ramp up so quickly!

But look at what’s happening underneath.

Despite the teacher recruitment drive, the pupil-teacher ratio has climbed from 36:1 in 2020 to 41:1 in 2024. The system is adding students faster than it is adding teachers. Classroom spaces are not keeping pace either. That is the same pressure Ghana felt in 2018 and 2019 - and Ghana’s answer was the double-track system. Zambia needs to see that for what it was: a workaround, not a solution.

The secondary enrollment gains in particular are worth celebrating. More Zambian teenagers are in school than at any point in the country's history.

The post 2022 data on learning outcomes is where the conversation gets a little downbeat. Essentially, learning outcomes aren't great, although enrollments are on the up.

The exam data post-2022 tells a mixed story. Grade 7 and Grade 9 pass rates have actually improved slightly - 71% and 56% respectively in 2024. The government points to this as evidence that quality has held up. But independent analysts describe the picture as mixed at best - access has surged, pressure on resources has increased, and the foundational learning crisis that predated the policy remains unresolved. Now, policymakers in Zambia have to ask whether the system they are building has the learning foundations for the millions now entering it.

And then there’s the fiscal question, which we can’t ignore, especially after Zambia went through a long and still-delicate debt restructuring process. From my experience during Ghana’s restructuring, I can tell you there isn’t much room to overspend in the immediate aftermath. Ghana was fortunate in one respect - oil revenue provided a financing cushion that helped absorb some of the cost of Free SHS in the early years. Zambia has copper, but commodity revenues are volatile, and the fiscal position remains tight.

Some researchers have already estimated that full implementation of free education could increase government spending on education transfers by 3.2% of total revenue.

Parliament has just enshrined free education in law, which is the right instinct for policy continuity, but legal permanence without fiscal architecture is just good intentions on paper. The Zambian Ministry of Finance will need to keep a keen eye on this space.

Altogether, although the access story is genuinely encouraging, the sustainability and learning outcomes story needs urgent attention.

What Zambia Can Learn

Ghana is five years ahead on this road. Here is what that distance has taught us.

First, don’t let the double-track system happen. This is the most urgent one. When Ghana’s schools became overcrowded after 2017, the government split students into two cohorts, with each cohort attending on alternating schedules. It was presented as a temporary fix.

I’m particularly worried it will become structural unless there's a massive surge in infrastructure. As it stands, students end up with meaningfully less instructional time, and the quality gap between well-resourced and under-resourced schools widens. Zambia’s classroom-to-student ratio is already flashing warning signs. Build now. The cost of reactive infrastructure is always higher than planned infrastructure, and the learning loss in the meantime is not recoverable.

Teachers are everything. Ghana’s regional data made this impossible to ignore. My last article showed the correlation between pupil-to-trained-teacher ratios and exam outcomes was almost perfect. Regions with the worst ratios consistently produced the worst results. Zambia’s national recruitment drive is encouraging, genuinely the right response. But recruitment is only half the problem. Ghana also showed that getting teachers into classrooms is not the same as getting them into the right classrooms. Rural and remote areas in Zambia already face significant disadvantages. If teacher deployment follows the path of least resistance - urban centres, established schools - the enrollment surge will widen regional inequality rather than narrow it. The government needs a deployment strategy, not just a hiring target. Incentives, housing, and career progression for teachers willing to go where they are most needed.

Also, measure learning, not just enrollment. Ghana’s national conversation after Free SHS focused almost entirely on the access numbers. How many students are enrolled? How many sat the exam? The quality signal - what those students actually knew - got lost in the politics. The risk is that rising enrollment and improving Grade 12 pass rates become the headline, while the 4% Grade 2 reading proficiency sits quietly in a UNICEF report that nobody in Cabinet has read. Build the early warning system now. Track learning outcomes annually, disaggregated by region and gender. If the data is bad, say so publicly. Zambia has a chance Ghana didn’t fully take - to front-load the quality conversation before politics makes it harder to have.

Finally, sort out the fiscal architecture before the arrears arrive. Ghana’s suppliers knew the arrears problem before the Ministry of Education admitted it. Schools were going without materials because the funding mechanism wasn’t keeping pace with the expansion. Zambia has now made free education a legal right, which creates an obligation that the state must honour. That obligation needs a dedicated, protected funding line, not a residual allocation that gets squeezed when revenue disappoints. Given where Zambia sits post-restructuring, this means being realistic about sequencing. Do fewer things properly rather than many things partially.

The Road Ahead

I fully understand that Ghana and Zambia are not the same country. Different economies, different political contexts, etc., etc. I want to be careful not to overstretch the comparison. But the logic of the education policy is identical, i.e., remove the fee barrier, get children into school, and hope the system can absorb the surge.

The early results in both cases looked the same - access up, classrooms fuller, governments understandably proud.

What comes next is where the paths can diverge.

Zambia has something Ghana didn’t have in 2017 - a sibling who already walked this road. The overcrowding, the teacher strain, the foundational learning crisis hiding underneath the enrollment headlines, the fiscal pressure that arrives quietly and then suddenly - Ghana has lived all of it. The lessons are not hypothetical.

Getting education right is one of the most important things any government can do. I’ve seen firsthand in Ghana how hard it is to course-correct once a policy becomes politically entrenched. Free SHS is now untouchable in Ghanaian politics, regardless of the quality questions. Zambia is heading the same way - Parliament just recently made it a legal right. That’s not a bad thing at all, but it makes the quality and fiscal conversations more urgent.

I’m rooting for Zambia to get this right. I think they can. But the window to build the right foundations, specifically teacher deployment, infrastructure, learning measurement and fiscal discipline, is now, not later.

The road is there. A friend walked it recently. Use the map!

Note: since publishing, a friend confirmed to me that the double-track system is being operated across Zambian public primary schools. The warning may be arriving too late for parts of the education system it seems.

Recent conversations with Oxford University’s Economic Policy Network forced me to revisit a period of my working life that taught me a slow-motion lesson in humility. One experience in particular - the Ghana CARES programme.

Back in 2020, the CARES programme was launched to much fanfare and positioned as “the transformative economic policy to properly reposition Ghana beyond the COVID pandemic”. The intentions were good. This was supposed to be the policy that “remade the country in a way that lasts”. 6 years after the launch, I reflect back on some harsh lessons learned.

First, is that, some governments, Ghana’s in particular, are unusually good at launching ambitious programmes but not so good at running them. I got to see this up close with CARES, and I'll mostly use this post to talk about how it came together, what went wrong, and what I think it means for African governments still running versions of the same playbook.

Back to 2020

To understand Ghana CARES, you have to go back to the first half of 2020, because the environment that produced it shaped everything that followed.

The early months of the pandemic were genuinely chaotic in a way that is hard to convey if you were not inside government at the time. Revenue was collapsing. Expenditure was spiking. We were running a deficit that blew past anything originally programmed. The Bank of Ghana extended a GH¢10 billion (USD 1 billion) facility to the government.

As I wrote in my last Inside the Room piece, everyone was winging it to some degree. Keep in mind, this was during “Covid Version 1”, the early days when it felt like getting the virus was a death sentence. Before Omicron, before vaccines, before we understood transmission properly.

The extent of support doled out was unprecedented, especially for an economy like ours. Free water for households, hot meals delivered to vulnerable communities, free fumigation for markets and schools, electricity subsidies, relief packages for frontline health workers, emergency credit for small businesses and so on. There’s so much more I can’t remember, but essentially, things that were politically unthinkable six months earlier become policy within a matter of weeks.

My boss, the Minister, was at his desk throughout. His reluctance to leave his office for much of those early days reminded me of the captain going down with the ship in Titanic. The workload was insane and by the end of the year, he ended up hospitalised with Covid and needing surgery for serious liver damage.

Throughout that period, I was regarded as “an essential worker” and had to be in the office with him daily. I still don’t know how I made it through those days without getting Covid once. Lucky me i guess.

By July 2020, the firefighting had stabilised enough for a different question to emerge. My boss argued that the crisis had created an opportunity, and he was right! Things couldnt go back to the way they were, and he kept insisting to me, “We can use this moment to actually change Ghana’s economic structure”. For the first time in a while, I was excited to go to work, even though nothing made sense in the world around me; I finally felt we were going to use all of the craziness to do some serious reform.

The Blueprint

I was asked to be part of a small group responsible for putting together the recovery plan. The intellectual leadership came from Dr Yaw Ansu, a brilliant man who spent most of his career with the World Bank and was Stanford-trained in economics. More proof, if any were needed, that there is no shortage of smart people working inside global south governments.

The consultations were extensive both within and outside the Finance ministry. The thing I learned about the civil service is that when you bring them into a genuine process, they can map a problem with remarkable precision. The challenges were well known i.e import dependence, youth unemployment, a manufacturing sector operating far below potential, etc. Honestly, the problem in Ghana, and I suspect across much of Africa, has never been identifying the right things to do. It has been having the discipline to actually fix them.

We ended up designing a two-phase programme. Phase one was stabilisation: immediate relief, protecting businesses, and food security. Phase two was the ambitious part: developing a light manufacturing industry, agro-processing, accelerating digitisation, positioning Ghana as a regional hub, etc. The document is available online for those who want more details.

The whole thing cost about GH¢100 billion ( USD 10 billion) over three years, with 70% expected from the private sector. In the initial designs, an implementation monitoring council chaired by the Finance Minister was instituted to meet monthly. Results Clusters. Cluster Champions. Dashboards. Accountability with consequences for failure.

Cross-ministerial coordination was the graveyard of Ghanaian policy, and we knew it. So we tried designing an architecture specifically designed to overcome it: compacts, cluster champions, and monthly oversight. We identified the disease, prescribed the medicine, wrote it all up in a beautifully formatted document for it to be launched and become actual policy.

What we failed to do was see the implementation through properly.

The Glitz and Then the Silence

The launch in late 2020 was everything these things usually are. Everyone who mattered was there. Compacts were signed. Ministers were aligned. We eventually brought in McKinsey to support the delivery unit, and I think the Tony Blair Institute as well, though my memory is getting hazy. Good analytical support. Lots of excellent presentations.

Then the silence came.

Not immediately. There was a period of activity. Meetings, updates, progress reports. But when you looked closely, the activity was mostly within the Ministry of Finance. The line ministries would not resist the programme outright but would smile, sign compacts, and gradually go back to their normal business. Far from hostile, they simply had no rational incentive to prioritise CARES over their own institutional agenda, and our big failing was that even with power over budgets, we never created a mechanism that actually changed that calculus. They went ahead to plan, execute, and, importantly, procure what they wanted, not what was “best for CARES” but what was aligned to their own agendas.

The Chief Director at the Finance Ministry did her best to maintain momentum, but CARES was running in one lane while the government's actual business was running in another. It became performative. The coordinating unit setup at the Finance Ministry had no real power to compel; we basically ended up being school hall monitors. Nobody answers to a hall monitor.

Even keeping the secretariat at Finance was, in retrospect, a mistake. It signalled that this was the Treasury’s project, not a whole-of-government priority. The ministers responsible for delivery did not feel a sense of ownership.

2022 and the End of the Road

I have written about what happened to Ghana’s economy in 2022 before, so I won’t dwell on it here. We had a lot happening on the macro side, and it consumed everything.

I would occasionally check in on CARES during this period, but only lightly; we were staring collapse in the face, no time to think about a structural transformation programme when we could barely afford to import fuel.

I eventually got my hands on the 2022 implementation report, and what it showed was how far the programme had drifted from its original ambitions. The industrialisation agenda, the manufacturing transformation, the PPP infrastructure pipeline: essentially, none of it had meaningfully moved. All the “progress” being made were just outputs from other Ministries regular work.

What remained active was a single agricultural enclave. Fancy English for a partially developed piece of farmland. Yep. What began as a GH¢100 billion economic transformation strategy, with ambitions to establish Ghana as a regional manufacturing hub and crowd in US$3 billion annually in FDI, had quietly become a rice farming project. That alone deserves serious examination.

There was also the private sector problem. We said 70% of CARES financing would come from the private sector. In practice, we never answered the most basic question a potential investor would ask: if someone had GH¢200 million (USD 20 million) to deploy into a CARES agro-processing project, where did they go? The secretariat? The Ministry of Agriculture? The President’s office? We had a pipeline of ambition and no pipeline of structured projects. Private capital does not chase government enthusiasm. It chases clear counterparties, legal frameworks, and structured transactions. We provided none of these.

What We Got Wrong

Let me try to distil the failure into learnable components.

The first problem was accountability without authority. We designed for coordination but funded it without enforcement. If you want government ministries to do something different, particularly African one’s, you need to be able to affect their budgets when they do not perform. We could not do that credibly.

The second problem was timeline delusion. Three years for a programme of this scope was never serious. Transformation programmes of this type, when they have worked anywhere in the world, have taken five to ten years of sustained government effort with institutional arrangements that survive changes in political leadership. We designed for one electoral cycle and hoped we could force things through.

The third problem was macro vulnerability, though I want to be precise here. The programme’s Phase 1 was designed to deal with the immediate fallout of Covid but what nobody designed for was what would happen to the transformation programme if Ghana’s broader fiscal position deteriorated badly.

There was no ring-fenced financing, no genuine private co-funding that would reduce dependence on public disbursements, no plan B if the macro environment turned hostile. When the 2022 crash arrived, CARES had no structural protection and simply hollowed out.

A Closing Thought

Ghana’s new government is running a national transformation policy that, from a distance, looks architecturally familiar: a cross-cutting economic programme, an authority figure coordinating across agencies, compacts, talk of export promotion and structural transformation, etc etc.

I hope it succeeds. But from experience, every time I see new compacts being signed, I know how the next scene tends to unfold. People go back to their lanes.

The lesson of CARES is not that transformation programmes are impossible. It is that they require something that governments, especially in the global south, consistently underinvest in: the hard, unglamorous institutional engineering that determines whether the thing you designed will actually function inside the government you actually have.

That means real authority to compel action, not just to convene it. It means a timeline that respects the pace of institutional change. It means answering the basic investor questions before you announce a financing target. And above all it means a collective discipline across government, a shared understanding that these programmes belong to everyone or they belong to no one.

Without that unity of purpose, you end up with something that looks like transformation on paper and feels like it in the launch hall but quietly dissolves the moment the room empties.

In the end CARES got cancelled by the next Government in it’s first budget statement. I understand. Smart people worked on CARES. Good intentions powered it. We articulated the vision clearly. We just never got the engine to start.

Earlier this month, I did some private consulting work looking at gold and its impact on Ghana's development. For some reason, it got me thinking about oil and asking how much it actually contributed to Ghana's development since the onset of commercial production in 2010.

Ghana's oil story is well documented in fiscal terms. We know the production volumes, the revenue allocations, the stabilisation fund balances. What comes up far less in public discourse is the bigger question: how much did oil actually change our economic trajectory? Not the fiscal line items. The development story. Did the economy move differently after 2010 than it would have otherwise?

I hope what I show here can kick-start a broader body of research and follow-up work in this area. There’s a lot to dig into.

Yes, I mostly only post about social policy. I’ll continue to do so. Consider this article a slight deviation from the norm. It is also heavy on the econometrics side, which will certainly make a few of my professors smile, and I will do my best to explain the methods I relied on so as not to alienate readers who are not too familiar with econometric methods.

Ok, here it goes.

Two questions

I set out to answer two things.

First: did Ghana’s economy diverge from a plausible counterfactual after oil production began in 2010?

Second: if it did, was that divergence real economic change or just oil revenue showing up in the GDP numbers?

Neither question has an obvious answer. Ghana was already growing through the 2000s. Debt relief had landed. Cocoa, the big export earner, was generating stable FX inflows. So pointing at a GDP chart and saying “oil did that” does not get you very far. You need something to measure against.

That is where the econometrics bit comes in, specifically synthetic control.

Building a “synthetic” Ghana

I know the concept sounds kind of bizarre, but it is really intuitive when you think about it.

To try to estimate impact, you create a fake version of a country using other countries, to understand what would have happened if “something”, in our case oil, had never occurred. Then you compare to see how different reality turned out.

In this case, I took a pool of Sub-Saharan African countries that did not experience a comparable oil windfall around the same time. I excluded the obvious ones: Nigeria, Angola, Gabon, Chad. I excluded countries in civil conflict or state collapse. That left about 24 countries.

Using World Bank data from 1995 to 2009, I then found a weighted combination of countries whose GDP per capita trajectories track Ghana’s as closely as possible before oil production begins.

The algorithm settles on a blend of roughly 30% Ethiopia, 24% Kenya, 20% Senegal, and 18% Uganda. From 1995 to 2009, this “synthetic Ghana” and the real Ghana are nearly indistinguishable. The pre-treatment fit error is 1.2% of the mean. For cross-country work in Sub-Saharan Africa, that is very good.

No counterfactual is perfect. But this gives us something concrete to measure against rather than relying on intuition alone.

What the data shows

After the oil discovery in 2010, the lines separate!

Ghana’s GDP per capita pulls away from its synthetic counterpart, and the gap grows over time. By the early 2020s, the divergence reaches roughly $575 per capita. Averaged across the post-2010 period, Ghana’s GDP per capita sits about $356 higher per year than the counterfactual, roughly 25% above the expected level.

Essentially, relative to a plausible non-oil path, Ghana’s economy generated an additional $200 to $350 per person per year for over a decade.

Finally an answer to my first question! Encouragingly, the gap also appears to be widening, not narrowing, over time.

I should acknowledge that, by definition, oil production enters GDP directly. If you pump crude, GDP goes up mechanically. So I ran the same exercise after stripping out oil rents from the GDP figures before constructing the counterfactual.

Roughly three-quarters of the gap survives. The non-oil economy diverges from the synthetic path by about $263 per capita, with a pre-treatment fit of 1.7%.

This is the result I find most interesting. Whatever happened after 2010 was not confined to the oil sector. Construction, services, financial intermediation, and demand effects. The broader economy moved too. Oil did not just sit in an enclave while everything else continued on its prior path.

How unusual is this?

To test whether Ghana’s divergence stands out, I ran placebo tests. Each country in the donor pool is treated as if it experienced a shock in 2010, and its divergence from its own synthetic version is measured.

Ghana ranks 4th out of 21. A small number of countries, notably Sierra Leone, Lesotho, and Uganda, show even larger trajectory shifts without an oil discovery. This matters because it means Ghana’s trajectory shift, while large, is not uniquely large in the regional context.

The divergence is real, the fit is strong, and the non-oil test rules out the mechanical explanation. But I cannot claim with full statistical confidence that oil alone caused the shift. Economies across the region experienced significant trajectory changes in this period for many reasons. That is a limitation of the synthetic control approach when applied to a region as structurally diverse and economically noisy as Sub-Saharan Africa. It gets us closer to an answer, but not all the way.

Honestly, I also tried to check whether this translated into higher household welfare. The consumption data turned out to be too sparse and too sensitive to the comparison group to support a firm conclusion. I had to give up on that exercise but it’s something worth investigating as more and better data becomes available.

Making sense of all this

The most honest reading of the data is that oil accelerated Ghana’s existing trajectory by a meaningful amount, roughly $200 to $350 per person per year, and that acceleration was not purely mechanical. But it happened within a context of broader momentum that makes clean causal attribution difficult.

Ghana in the 2000s was not standing still. Debt relief through the Highly Indebted Poor Countries (HIPC) initiative had freed up fiscal space. Cocoa was bringing in steady foreign exchange. Democratic stability, relative to the region, was attracting investment. Macroeconomic management was improving. All of that was building before a single barrel came out of the ground.

Thus, Oil joined a process that was already underway. It did not create it from scratch.

The data suggests oil added something real. And that raises harder questions about whether Ghana got enough from it. My suspicion, informed by my work on the gold sector, is that not enough of what Ghana earned from oil flowed back into the economy in the most productive way.

Whether the upstream contracts could have been structured to capture more for Ghana is a question for the legal and fiscal experts, and I hope they examine it, given that we have been earning oil revenue for a little over 15 years now.

What we can point to is that some of the oil proceeds did flow into genuinely important priorities. Free Senior High School, the flagship social programme of the last administration, was funded in significant part by oil revenue. That is a direct case of a natural resource being channelled into a social priority.

It is also worth noting what was left on the table. Ghana has been extracting and exporting raw crude, and these gains could have been larger if we had refined domestically rather than shipped unprocessed crude. Every dollar of the $200 to $350 per-person gain represents a baseline that could have been higher under a different industrial strategy for the resource.

And then there is the question that hangs over all of this: how long does the oil story last? The risk of stranded assets under a global decarbonisation path is real.

Decarbonisation conversations may have become less fashionable in global discourse thanks to the Trump administration, but they cannot be completely avoided. Ghana’s oil gains, however meaningful, will not continue indefinitely. The country needs to industrialise, diversify, and find another engine to sustain the momentum of an improving macro-fiscal trajectory.

Oil bought time. But the real question, one that I hope we can understand with enough data, is whether that time is being used well.

Where this leaves us

There is a lot to uncover from here, so I am treating this article as just a starting gun. Hopefully, with more data and more outside expertise, we can go into a whole series on this.

What happened to inequality during the oil era? What do the upstream contracts look like under serious legal scrutiny? How does the oil story connect to the gold story?

Those questions require different data and different disciplines. But at least now we

Share

have a quantitative starting point to ask them from.

Governments are increasingly experimenting with ways to reduce financial barriers to healthcare access, including making primary care free. Ghana is now preparing to do exactly that, with a free primary healthcare programme due to launch this Wednesday (April 15, 2026).

The premise is intuitive. The political appeal is obvious. The harder question, one that tends to get asked after implementation rather than before, is whether the policy is actually aimed at the constraints households face when they fall ill. That's the question I set out to answer before the programme goes live later in the week.

This piece draws on Ghana's 2024 Annual Household Income and Expenditure Survey (AHIES) to test the policy's theory of change before implementation. The approach is transferable, and any government considering a similar reform can run the same four tests against its own data.

First, Understand the Context

Before judging whether the design is right, there is a need to understand the context the policy is being introduced into, i.e., what diseases are actually driving people to primary care facilities, and what the system currently looks like on the ground.

In that regard, Ghana’s 2024 National Annual Progress Report from the National Development Planning Commission (NDPC) helps gives us the supply-side picture.

Malaria accounts for 26.7% of all outpatient diagnoses in 2024, up sharply from 20.3% in 2020. Upper respiratory infections, rheumatism, diarrhoea, anaemia round out the top five. The everyday burden that primary care facilities are managing is overwhelmingly infectious and preventable. That's the supply-side reality the programme is walking into.

Non-communicable diseases (NCDs) don't feature prominently in outpatient visit counts, but that underrepresentation is partly a function of the system itself. Patients who can't afford treatment often don't present at facilities at all, and when they do, the costs can be catastrophic. Thankfully, this is the gap that another recently launched policy, the MahamaCares programme, is designed to fill, by financing treatment for cancer, hypertension, diabetes, stroke and kidney conditions.

Against this supply-side backdrop, and ahead of the launch of free primary care, the government has taken steps to reinforce the system's capacity to respond. A few weeks ago, it procured 24,534 pieces of medical equipment (valued at approximately GH₵500 million, or around USD 45 million) for rollout to health facilities nationwide, and has allocated GH₵9 billion (approximately USD 818 million) within the uncapped National Health Insurance Fund to cover NHIS claims, essential medicines, vaccines and the free primary healthcare programme itself.

The supply-side picture and the government’s response to it set the stage. The more important question, and the one the supply-side data alone can’t answer, is whether removing the cost of care is the binding constraint that is actually keeping people away from the system in the first place.

That’s what I set out to test.

Test One: Is There Actually a Problem to Solve?

A fee waiver only matters if people are currently failing to seek care. If most Ghanaians who fall ill are already getting to a facility, then price isn’t the constraint and the programme, however well-intentioned, may be solving the wrong problem.

Here’s what the AHIES data shows.

Among those surveyed in the AHIES, 6.9% reported illness or injury in the two weeks before the interview. Among those people (the ones who were actually sick), only 51% sought care. Nearly half, 49%, did not.

This is a significant finding: in essence, for every two Ghanaians who fall ill, approximately one does not enter the care system at all. The programme is targeting a real problem head-on. That matters. A lot of policy interventions fail this basic test. This one doesn’t.

Test Two: Does Money Actually Change Behaviour?

Knowing there’s an access gap is one thing. The next question is whether financial protection changes whether people seek care, which is the mechanism the programme is designed to activate.

The NHIS data inside AHIES gives us a natural test. Among those reporting illness or injury, 57% of people with active NHIS coverage sought care. Among the uninsured, that drops to 41%. A 16 percentage point gap, and the confidence intervals for the two groups don’t overlap, suggesting a statistically robust finding.

Money matters. Not for everything, but clearly for this.

The temptation at this point is to treat the finding as a full stop: uninsured people seek care less, therefore make care free, problem solved. But association isn't prescription. What the data shows is that financial protection is meaningfully linked to care-seeking behaviour — a strong justification for the policy, but not a guarantee that fees are the only barrier or that removing it entirely closes the entire gap.

Test Three: Is This a Geography Problem or a Price Problem?

Honestly? The national aggregate data surprised me here.

In Ghana, the instinct is often to frame health access as a stark rural-urban divide. But care-seeking among ill respondents was 52.8% in rural areas and 49.2% in urban areas. The gap is small. The confidence intervals overlap substantially. You can’t cleanly say rural Ghanaians are dramatically worse off than urban ones at the national level. According to the data, transport costs also don’t appear to be the dominant monetary barrier nationally, though that measure only captures direct financial outlays and may not reflect the full burden of distance and time.

The real geographic story is at the regional level and it’s much sharper!

NHIS coverage runs from 44.9% in Savannah to 80.9% in Bono. Care-seeking follows an almost identical pattern: Oti at 41.3% and Savannah at 43.3% at the bottom; Upper East at 66.1% and Upper West at 65.0% at the top. The gap between the best and worst-performing regions is nearly 25 percentage points. That dwarfs any rural-urban difference.

This creates a design problem that the current proposal doesn't fully address. A flat national fee waiver will benefit everyone to some degree; that's the point of it. But the data suggests that some regions will feel it more than others, not simply because they are poorer, but because the regions with the lowest care-seeking also have the weakest delivery infrastructure. When both constraints are in place, reducing price alone may not be enough to shift behaviour significantly. Essentially, the fee waiver is necessary! but in the poorest regions, it may not be sufficient alone.

Test Four: Are People Even Using the Public System?

This test matters because you can make care free, but if people have already moved to private clinics, pharmacies, or informal providers, then the impact of strengthening the public system is limited from the start.

To me, this was probably the most important finding in the whole analysis. Most Ghanaians who seek care already rely on the formal public system. Nationally, 70.6% of care-seekers used public health facilities. In rural areas, that figure is 76%. Urban areas show more private facility use (33% compared to 22% in rural areas), and in cities like Accra, private practice clearly plays a significant role. Informal or alternative providers accounted for just 2-3% of consultations nationally. And formal providers (doctors, nurses, pharmacists, physician assistants) accounted for 87.5% of rural consultations and 91.3% of urban ones.

Altogether, we can see that the proposed programme is designed to strengthen a system most Ghanaians already rely on. That makes the intervention credible. It also means the healthcare system's pre-existing operational problems are not a background issue. They are central to whether this policy will work.

What the Four Tests Tell Us

Let me bring this together before we go further.

The data paints a broadly supportive but conditional picture. There is a genuine unmet need. Financial protection matters. People already use the public system. And encouragingly, the evidence points to a real problem that this policy can meaningfully address!

But the evidence also shows that the problem is not uniform across Ghana. As such, a flat national approach will naturally reach some regions more effectively than others, particularly where both unmet need and delivery constraints are greatest.

Now the Harder Question: Is the System Ready?

Functional CHPS zones, the community-level facilities that form the intended frontline of this programme, have declined from 79.7% in 2021 to 72.4% in 2024, against a government target of 87%.

In truth, the system went backwards during my time in Government, meaning that in the years leading up to a major programme designed to increase utilisation, frontline infrastructure has deteriorated.

This is not entirely surprising. Ghana's fiscal crisis of 2022, which ultimately led to an IMF programme, hit frontline infrastructure hard. Recurrent budgets were squeezed, maintenance was deferred, and facility functionality declined quietly while the headline spending numbers held. Today, Ghana now risks investing ahead of functionality rather than fixing functionality first.

The staffing picture drives this point home even harder. Doctor density stands at 0.22 per 1,000 against a WHO benchmark of 1. Nurse density fell from 1.90 to 1.84 per 1,000 in a single year. Between 2022 and 2024, 1,795 nurses exited the government payroll. The Nursing and Midwifery Council received 7,778 foreign verification requests in 2024 alone, and approximately 10,200 in 2023. These are nurses preparing to leave!!

Into this environment, 24,534 items of equipment are being deployed. The obvious question is: who operates and maintains them?

The ambulance service is worth mentioning, not to pile on, but because it illustrates the pattern perfectly. According to the NDPC, only 255 of 326 vehicles are operational. National average response time worsened to 25 minutes in 2024. In the Oti region, it's 42 minutes!. The hardware exists. The system around it does not. Quite frankly, Ghana has a long history of confusing procurement with service delivery.

The OPD utilisation data brings the regional picture together. Bono records 1.71 visits per person per year. Northern records 0.61, less than half the national average of 1.20. The regions using the system least are the same ones with the fewest doctors, the most non-functional CHPS compounds, and the longest response times. The pattern runs in one direction, and it runs hard against the regions where the programme is most needed.

Here's what concerns me: if a free care programme successfully brings more people to facilities in those regions, but the facilities aren't adequately staffed or stocked, we may end up increasing demand into a system that isn't ready to absorb it. That's not an argument against the policy. On the contrary, I hope it encourages those in charge to get this policy implemented correctly.

A Note on Financing

Sustaining any of this requires a financing architecture that holds beyond the political moment of launch.

The standalone cost of the free primary care component hasn’t yet been publicly itemised, though this level of detail may follow implementation or appear in the 2027 budget cycle. That’s reasonable. What matters more right now is the broader fiscal picture.

Ghana’s per capita health spending stood at USD 35.81 in 2024, below half the Sub-Saharan Africa average of USD 74. Health as a share of GDP declined from 7.4% to 6.0% between 2022 and 2024. The picture within the Ministry of Health’s budget allocation is also stark: of GH₵16.3 billion (approximately USD 1.48 billion) allocated for 2026, GH₵15 billion (fully 92%) goes to compensation of employees, leaving thin headroom.

That said, the Ministry of Health’s allocation is only part of the story. The NHIS is funded through its own dedicated National Health Insurance Levy, alongside SSNIT contributions and subscriber premiums, with total health financing for 2026 put at GH₵34.7 billion when NHIS resources are included. What complicates the picture is that donor support, which once covered around 19% of health financing, fell sharply to just 2.2% in 2025, creating an annual financing gap of more than GH₵2.4 billion that domestic resources are now being asked to fill. How the government continues to plug that gap alongside funding a major new programme will determine whether this policy is sustainable beyond its launch.

So, Where Does This Leave Us?

The constraint is not singular. It is layered. Analysing the four tests together, three distinct pressures emerge from the data, and they interact differently across regions.

Cost is real. The 16 percentage point gap between insured and uninsured care-seeking is robust and statistically clear. The programme targets this directly, and that is its strongest justification. On this test alone, the policy makes sense.

Access is more regional than rural. The evidence doesn’t support a simple rural-disadvantage story at the national level. The gap between rural and urban care-seeking is small. What is pronounced is the regional disparity: a nearly 25 percentage point difference between the best and worst-performing regions. The flat national design is a reasonable place to start, but the data points to significantly higher returns from pairing the fee waiver with geographically targeted delivery investments in the most underserved areas.

System functionality is the constraint that makes the other two harder to solve. The public system is already the one most Ghanaians rely on, which means its declining frontline capacity, staff emigration and thin operational budgets are not peripheral concerns. They sit at the centre of whether removing fees translates into actual access.

A Final Note for Readers Beyond Ghana

More broadly, this exercise points to something simple but often skipped: reforms should be tested against the constraint they’re actually trying to solve, and that test should happen before implementation, not after. I know it sounds obvious BUT for many governments across the developing world, the sequence still tends to run the other way i.e announce, implement, evaluate. The data exists to do better.

Household data can help do this. The four tests I ran here (is there a real access gap, does financial protection change behaviour, is geography the main driver, is the public system the one people already use) aren’t Ghana-specific. Any government preparing primary health care reform can run the same checks.

Ultimately, Ghana’s free primary health care launch is a few days away. What happens after it will be the real test. For everyone’s sake, I hope this government gets it right!

Share

Ghana has a stubborn unemployment problem.

It's not a new topic. But it cuts to the heart of something we don't talk about enough: the social contract between the state and its citizens. Specifically, who the state is expected to protect, and who it actually does.

Ghana has basically had a state-centric economic model since independence in 1957. That’s decades of the state being the primary actor and indeed employer in the economy. It’s created an embedded expectation: the state is supposed to look after you. And when it appears that it doesn’t, governments get booted out.

That expectation hasn’t gone away, even as the economy liberalised in the 1980s. Ghanaians still look to the government first. For jobs. For infrastructure. For opportunity. It’s a product of how our country was built.

In that regard, Ghana’s budget tells a story about who the state protects and who it doesn’t.

The Asymmetry

The state employs a little over a million people. Doctors. Teachers. Police. Civil servants. Engineers. These are mostly educated, formally trained people. They have pensions. Healthcare. Job security. The state made a promise to them, and it keeps that promise. That’s not a criticism; these are essential workers. But it’s a choice. A deliberate one. And if you want a sense of what happens when that protection breaks, read my last piece on Ghana’s debt restructuring. The moment pensions got touched, the political fallout was enormous. Governments don’t forget that.

Go look at the fiscal data from 2016 to 2024. Wages and salaries are stuck at around 5.5 to 5.9% of GDP. Interest payments have climbed from 5.3% to 6.3%. Combined, before the government can expand infrastructure or invest in new programmes, it already owes nearly 12% of GDP each year to existing wage commitments and interest payments. The budget is rigid. There’s almost no room. You end up trying to fit big ambition into a tiny box. You already have a lot to contend with in government: infrastructure, education, health, and debt service. But one thing that doesn’t go away, and is almost unforgiving if you get it wrong, is job creation.

Every government knows it has to get this right. But nobody has quite cracked it. I'm speaking as someone who tried in his own way to make a dent in it, designing quite a few programmes with the ambition of creating jobs (YouStart, SME Growth & Opportunity, CapBuSS) but not quite reaching the desired heights.

The data is unforgiving. Ghana's official unemployment rate is sort of stuck between 13 and 14%. It was 13.1% in 2022. It spiked to 14.6% in 2023. It hasn't budged in three years despite all the job creation claims. Meanwhile, underemployment, people working fewer hours than they want to, has improved from 13.3% in 2022 to 9% in 2024. So people are definitely getting more hours. But they're not finding their way into steady, full employment. The barrier to entry into consistent work remains intact.

Underneath those headline figures is another layer. Female unemployment averages 16% compared to 11% for men. Women are 46% more likely to be without work. When governments announce job-creation numbers at rallies and press conferences, they almost never disaggregate them by gender. Which means the female unemployment crisis remains invisible in the political narrative, even as it shapes the lives of millions of young women.

As it stands, over 80% of Ghana's labour force works informally. Petty trading. Casual labour. Agricultural work that's seasonal. The formal private sector employs about 12% of the workforce. The public sector, about 5%. The rest? Trying to survive.

Despite the policies that have been launched over the years (NABCO, YouStart, and plenty more) broad-based tackling of unemployment hasn’t really happened. Each programme arrives with fanfare. Each one helps some people. None of them really has shifted the structural picture.

If you can’t land something in the private sector or create your own job through entrepreneurship, you’re largely on your own. The micro-credit schemes, the vocational training, the entrepreneurship programmes: they’re well-intentioned, sometimes helpful. But fundamentally, the state is saying: “We’ll structure opportunities for those who can take advantage of them. For the rest, good luck.”

This is the issue every government has to fix. I won’t claim I can provide the magic solution in one Substack article! But at least we can use this to look at what works elsewhere. In today’s article, India is where we place our sights.

India’s Evidence

Twenty years ago, India launched what eventually became one of the most ambitious social security and public works programmes in the world. They guaranteed 100 days of work, every year, to anyone in a rural household willing to show up. Wages set by the government. No means testing. Just: if you want work, the state provides it.

The programme was called MGNREGA, the Mahatma Gandhi National Rural Employment Guarantee Act. The idea was almost defiantly simple: give poor workers a fallback option. Think of it like an insurance policy that pays out in labour instead of cash. During hard times, workers can fall back on publicly funded employment building roads, irrigation systems, and rural infrastructure. The state becomes, in practice, a limited employer of last resort for rural labour.

Did it work?

The evidence says it did. Wages in areas with MGNREGA access rose 5 to 10%. Workers had power. They could say no to exploitative wages because they had an alternative. Spillovers happened too: wages in the non-MGNREGA labour market went up. Even seasonal agricultural workers benefited. It reduced out-migration. It produced assets like roads, water tanks, and public buildings. The wage gains persisted even when the quality of the work was mixed.

Imbert and Papp (2015) showed this, finding that MGNREGA crowded out private sector employment one for one and increased rural wages by 5%. Other researchers also demonstrate that the programme functioned as a genuine safety net, particularly after bad rainfall shocks. Muralidharan, Niehaus and Sukhtankar (2023) showed this, finding that reforms to the programme raised beneficiary households’ earnings by 14% and reduced poverty by 26%.

The mechanism is simple: give workers a credible fallback option, and their bargaining power rises.

Now, the programme is not perfect. Not by any stretch. Any Indian reader could probably rattle off a list of challenges: delayed payments, poor asset quality, and political interference. All real. But that doesn’t mean we can’t learn from it. The core insight, that a guaranteed employment floor lifts wages and bargaining power for the poorest workers, is robust across the evidence.

In spite of this, India recently replaced MGNREGA recently with a new scheme. On December 16, 2025, the Indian government introduced the VB-GRAM G Bill as a successor to MGNREGA. It essentially preserved the idea of rural employment support, but removed the legal guarantee that made MGNREGA powerful.

Within days it passed Parliament. By December 21, it was law. The new framework changed the structure in fundamental ways: shifting from demand-driven (where workers request work and the government is legally obligated to provide it) to supply-driven (where the government allocates a fixed budget and work stops when the money runs out). The centre-state funding split also changed from 100% central to 60% (central)-40% (state), among other key changes.

Funding challenges remain huge for programmes of this nature. And fundamentally, politics went against it in the end. Supporters argued MGNREGA had become fiscally expensive and poorly targeted. Critics saw twenty years of wage gains being undermined.

Here’s what this teaches Ghana: the mechanism works, albeit with quirks that have to be ironed out. We can learn from India’s example, both the successes and the stumbles, to refine a design for our own version. One that fits our environment, lasts across different political cycles, and actually produces results.

Ghana’s Version

Going forward, Ghana should test an employment guarantee mechanism. Not as the silver bullet for structural transformation, but as one possible contribution toward solving a portion of the unemployment puzzle, especially for those outside the middle class who have no safety net at all.

It can’t be the Indian model transplanted whole. And it absolutely cannot be run from Accra.

The more I look at what India got right and wrong, the more convinced I am that Ghana’s version needs to be built from the districts up.

The design could look like this: a district-assembly-led pilot across three high-unemployment districts, probably in the North, where structural inequality is sharpest and the need is greatest. Each assembly gets a per-capita budget. But instead of vague “public works,” the scheme uses four strategic work categories. This matters because it makes the programme credible to donors, defensible politically, and relevant to actual district needs.

Category 1: Local Infrastructure & Maintenance. Most local infrastructure doesn’t fail because it was never built. It fails because nobody maintains it. Feeder roads. Drainage channels. Irrigation canals. Water points. Assemblies are already responsible for these, they just lack labour budget at scale. Employment guarantee labour fills that gap.

Category 2: Climate & Environmental Work. Ghana faces real climate risks: erratic rainfall, soil degradation, riverbank erosion. Tree planting. Riverbank restoration. Erosion control. Wildfire prevention. India used NREGA heavily for water conservation and drought resilience. Ghana could do the same. This work also unlocks climate finance with donors.

Category 3: Agricultural Productivity Support. Ghana’s smallholder farmers are the backbone of rural income. The scheme supports, not replaces, farm labour. Irrigation maintenance. Cooperative land prep. Post-harvest facilities. The key: support smallholder infrastructure, not subsidise large farms.

Category 4: Social & Community Services. Not all valuable labour is physical. Childcare support. School feeding. Elder care. Sanitation campaigns. Data collection. These roles expand participation for women and older workers, and this is where India’s experience is instructive. Women participated heavily in MGNREGA when work was transparent and locally organised. Ghana can reinforce this by ensuring community service roles are compatible with caregiving responsibilities. Assemblies would be expected to report participation by gender, making disparities visible.

Each assembly gets to prioritise across these four categories based on local need. A district facing drought prioritises climate work. A district with deteriorating roads prioritises maintenance. Flexibility within structure.

Workers opt in voluntarily through the Ghana Card system. They’re registered. They get work allocated across these categories. They’re paid digitally within 15 days.

Target population: working-age adults (18 to 55) in designated districts with chronic underemployment (less than 30 hours a week) or income insecurity. Voluntary entry. No political allocation. The work is there. You want it, you get it.

Why district-led?

India decentralised much of MGNREGA’s implementation, and that was one of its smarter design choices. Ghana has to do the same. All decisions can’t come from sitting in offices in Accra. People in the capital don’t know which feeder road in Tamale needs fixing or which irrigation canal in Wa has silted up.

With three elected assemblies running the pilot, you get skin in the game at the local level. Assemblies are closer to local priorities than a central ministry. They will still argue about projects, but the information problem is much smaller. And if the programme works, those assemblies become its biggest advocates because the results are visible in their communities. That kind of local ownership is what helps a programme survive beyond one political cycle.

The Ghana Card becomes the implementation mechanism. Already distributed. Already digital. Workers sign up, get allocated work, get paid. Transparent. Auditable. The state can actually measure wage gains and employment. Unlike the job creation claims that get thrown around at rallies, this is provable.

Let’s be completely upfront: this type of policy isn’t the answer to completely curing a longstanding structural problem. Ghana needs industrial policy. Better education. Better health infrastructure. I’ve written about education before, and the problems there are deep. This doesn’t replace any of that. What it does do, however, is renew the social contract with people who feel left behind by successive governments. Give them work in their locality. Reverse the assumption that rural people have to migrate to survive. Create the conditions (better roads, better infrastructure) that allow local private activity to flourish.

The Hard Questions

Let me be honest about the risks because the objections are real.

Unions will oppose. They’ll say this is competition for their members. Fair point. But this programme targets people who aren’t in formal employment. Most of these workers aren’t union members. The state could afford both (better wages for its formal workers and a fallback for everyone else) but only if it treats the pilot as a real fiscal test. The question is whether there’s political will.

The work might be poor quality. True. MGNREGA had quality problems. But here’s the thing: even when quality was mixed, the wage effects persisted. Workers still got paid. The asset was still created, even if it wasn’t perfect. Better an imperfect asset and higher wages than no asset and no wages.

It might not be fiscally sustainable. Three districts, 250,000 to 400,000 beneficiaries, 100 days at GHc 50 to 60 per day: you’re looking at GHc 300 to 500M annually. That’s not trivial. But Ghana can reallocate DACF funds from assemblies without infrastructure priorities. Use labour-intensive works to stretch capital budgets. Bring in development partners. The pilot is explicitly testable: 18 to 24 months, measure outcomes, then decide whether to scale or stop. That’s an honest conversation.

Won’t a future government just scrap it?

Maybe. That’s always the risk. But that’s also why the design matters. If the programme is embedded locally, producing visible results, with data that proves it’s working, scrapping it becomes politically expensive. A programme nobody sees can die quietly. A programme that’s putting food on tables in three districts doesn’t go away without a fight.

And here’s what Ghana can do: build that evidence base from day one. Measure everything. Share the data. Let people see it working. If Ghana is serious about this, it documents everything.

The Pilot

Three districts, 18 to 24 months. Strict design.

Budget: GHc 300 to 500M annually. In the grand scheme of things, its not that much money for any Government (approx. 30M to 50M USD). Funding sources could include donor co-financing, audit recoveries from the government payroll, and reallocation of District Assembly Common Fund resources.

By month 18, you know whether it works. By month 24, you have numbers. The questions are simple: Did wages rise? Did people actually take up the work? Was the programme run cleanly?

Then comes the hard part: the honest conversation. If it worked, scale. If it didn’t, say so and stop!

I’ve sat in enough meetings where in the end it wasn’t even clear the programme being discussed was delivering tangible impact, but nobody wanted to pull the plug for all manner of reasons. That can’t happen here. The whole point of a pilot is that you’re allowed to learn, including learning that something didn’t work.

And if we’re serious about learning, we should go further. Districts should be selected through a randomised process. Not only because it’s fairer, but because it strengthens the evidence base. Randomised selection gives you a credible comparison group, which means that when the pilot ends, you’re not arguing about whether the results were real. You have a proper foundation for a rigorous evaluation, potentially even a full randomised controlled trial (RCT) down the line. That’s the kind of evidence that convinces donors, silences critics, and gives future governments a reason to scale rather than scrap.

The Closing

Ghana’s unemployment problem won’t be solved by one programme. But it won’t be solved by doing nothing either.

What I’m proposing here is a process: design, pilot, implement, refine. Test an idea with discipline. Measure it honestly. And if it works, build on it. If it doesn’t, learn and try something else.

The bigger ambition is cultural. Creating an environment where you don’t have to leave your community to find dignified work. Where the place you were born has something to offer you. Where your district assembly, the government closest to you, actually has the tools to look out for you.

India proved the mechanism can work. Ghana has the infrastructure (the Ghana Card, the district assemblies, the fiscal architecture) to test it properly.

Three districts. 18 to 24 months. Then an honest evaluation.

If it works, scale it. If it doesn’t, move on. In the end, a country that refuses to test bold ideas will never solve serious problems.

Subscribe now

Share

I know it’s been a while, but I’ve slowly agonised about which topic to write about next. I just knew I wanted to tell a different type of story, a positive one, especially because my last two Inside the Room articles have been steeped in crisis situations.

In a weird way, this “positive” story came in the middle of another major crisis—COVID-19. It’s funny how quickly we moved on from the pandemic. Lockdowns, no air travel, work from home. Normal felt like a lifetime away. Especially in early 2020.

But looking back, I’ve realised I experienced a lot that impacted me. And in this article at least, I’d like to reflect on how that period gave me my first glimpse of what could happen when people, united by crisis, worked together towards a common goal.

And to an extent, it’s one of the few events I can remember from that period with a neat, happy ending.

Here it goes.

A New Normal?

The year is 2020. COVID is spreading across the world. Most of my life, I told myself the “people at the top” had all the answers. Turns out, not so true.

I learnt very quickly that leaders are just people when I saw my boss ask, “What do I do now?”

Everyone was improvising. And as the health pandemic quickly became a looming economic crisis, finance ministers around the world, especially in developing countries, knew that individually they couldn’t make it on their own.

For us in Africa, that’s where Vera Songwe came in. The hero of this story.

Economic Pan-Africanism?

Vera was the Executive Secretary of UNECA—the United Nations Economic Commission for Africa. An economist from Cameroon, she regularly features on lists ranking her among the most powerful African women.

She recognised early that finance ministers on the continent were working in silos, and for Africa to emerge from the pandemic, a coordinated response to the economic fallout was needed.

Using her position at UNECA, she created a platform where finance ministers gathered to speak candidly, compare notes, and converge on common policy approaches. Conversations happened in French, English, Portuguese—it didn’t matter. Everyone was encouraged to share ideas, and Vera’s team would translate them to whatever language was necessary, so no point was lost.

Vera didn’t sleep. I honestly don’t know how she pulled it off. Her team was managing logistics across dozens of countries, different time zones, and diverse political contexts, somehow putting every African Finance Minister on Zoom calls at the same time.

I was lucky enough to have listened in on those meetings. I was meant to be writing my minister’s speeches and briefing notes, so he insisted I had to be in the room when discussions with ministers and envoys were happening.

At times, it felt like I was listening to therapy sessions. Everyone sounded overwhelmed. The fear was real, but so was the solidarity. Looking back, I realise that was the closest I had come in my lifetime to witnessing Pan-Africanism at work. In this case, African finance ministers, coordinated across the continent, resolved to help one another.

It wasn’t all serious technical calls though. There were some funny moments.

I remember one involving South Africa’s then finance minister, Tito Mboweni. In the middle of a heavy technical discussion, he was asked for his view. He turned his camera on to offer his opinion, and I couldn’t help but laugh. He was in the kitchen. Cooking! The world was falling apart, and he was calmly by the stove discussing financing options for the continent!

I found out later he was known for this sort of thing. The Wall Street Journal once called him “a star in the kitchen.” Apparently, he used to share photos of meals he prepared on Twitter. Everyone handles stress differently.

He passed away a little over a year ago. I still think about that kitchen call.

The Asks

Throughout these meetings, it became clear that Africa had a financing problem, and we needed solutions that went beyond our individual borders.

The Ministers converged on a set of asks to request from the international community. Key focal areas were:

• Suspension of payments on debt through an extension of the G20 Debt Service Suspension Initiative

• The development of a liquidity sustainability facility

• And the crown jewel: $100 billion to reach Africa to help fight COVID

The IMF was quickly identified as a potential source for the $100 billion request through Special Drawing Rights (SDRs). I don’t go into too much technical detail in these articles, so simply put, think of SDRs as “money issued by the IMF.”

In this case, the Ministers wanted the IMF to provide the resources required to effectively support their COVID-19 responses. The one thing to note is that the IMF doesn’t give out SDRs often. Indeed, there hadn’t been a general issuance of SDRs since the 2009 financial crisis.

The ask was beyond bold.

Hi Gordon!

The idea of using SDRs wasn’t new. Gordon Brown and Larry Summers had already been advocating for them as soon as COVID hit. My minister knew we had to speak to both of them to make any meaningful progress on that ask.

For added context, Gordon Brown was the UK’s Prime Minister when SDRs were last issued. If a figure like him publicly pushed for new SDRs, there was a high chance they would be issued, meaning countries would receive financial support from the IMF to help them manage their economies.

I remember the day my minister spoke with him. I was supposed to help set it up, and I was overly excited. Tony Blair, Gordon Brown, Ed Balls. I idolised these guys growing up, and now I was going to be part of a closed session with one of the architects of New Labour.

After drafting my Minister’s speaking notes, my mind inevitably drifted to how I would address him. Right Honourable Brown? Your Excellency? Mr Brown???—You idiot, this guy was the UK’s Finance Minister for 10 years; you can’t call him Mr Brown! You need a lofty title to address him! Unresolved, I told myself I’ll figure it out later.

On the day of the call, I signed in maybe five minutes early, just to make sure everything was working okay. And then—bam—Gordon Brown signs in. Early. Just like that.

Dear God. What do I say? He kept looking at the screen, probably confused to see my name and not Ghana’s Finance Minister. I panicked, and that’s when I blurted out: “Hi Gordon!”

Kill me now, Lord.

To his credit, he didn’t react weirdly. He just smiled wryly and said hello, then asked when the Minister would join. I replied that I was going to get him, so the conversation could start. He was cool! He didn’t even care for being addressed with fancy titles, plus he was surprisingly self-deprecating for such an accomplished man.

The call went well. He spoke deeply about the SDR option and doubled down, saying it was something we should not let slip.

On top of that, he committed to supporting the ask for new SDRs. Less than a year later, he was also instrumental in gathering support to get COVID-19 vaccines to Africa and also in mobilising funding to educate the world’s poorest children during COVID, but that’s another story.

We also spoke with Larry Summers not long after. Thankfully, I didn’t have a “Hi Larry” moment.

Advocacy in High Gear

Once we had the asks, the advocacy machine kicked into gear.

Every platform. Every forum. Ministers were appearing on stage together, saying the same things. Nigeria. Angola. South Africa. Egypt. All spoke one language on international forums and clearly outlined Africa’s ask consistently.

At the same time, my Minister was the Chair of the G24 Group of Countries—the bloc that coordinates developing countries’ positions on international monetary and development finance issues. His mandate was clear: amplify Africa’s voice globally.

I drafted far too many letters, articles and speeches during this period. Even more so when he became Chairman of the African Development Bank Board of Governors later in 2020.

Meanwhile, Vera and the UNECA team were refining our arguments as opposition to the IMF intervening on such an unprecedented scale grew. I remember one economist highlighting that new SDRs would be inflationary. We had to consider all those views.

Africa presented its asks to the G20 finance ministers’ meeting in July 2020. This time, Amadou Hott, Senegal’s finance minister, spoke on our behalf. He killed it. The ask was clear, the delivery sharp. We had presented our requests to all the decision-makers. Saudi Arabia was hosting the G20 that year, and they graciously gave us the space to let our voices be heard.

The advocacy and coordination between Africa’s Finance Ministers continued for months throughout 2020. But we weren’t making much progress. The whispers were that the Trump Administration wasn’t in favour of issuing new SDRs.

This matters because America is the IMF’s largest contributor. Without their support, a new SDR allocation is dead on arrival—no matter how much we spoke in its favour.

November 2020. Everything changed. Trump out. Biden in.

The Breakthrough

We entered 2021 with renewed hope. Janet Yellen, the new Treasury Secretary, had a wider worldview, and she wasn’t opposed to the idea of intervening in a big way.

By May 2021, close to six months into her tenure, she said so publicly. The US were sympathetic to the cause! With that breakthrough, we were firing on all fronts with Vera coordinating.

Huge global voices kept coming on side. Mia Mottley, Prime Minister of Barbados, at various points addressed our meeting of African ministers, letting us know the Caribbean supported our cause. Anyone who has heard her address the UN knows: if you have Mia Mottley in your corner, your voice will be heard.

Even the Catholic Church came onside! I remember attending meetings between my Minister and Caritas Internationalis during this period.

The coalition was expanding. The message was consistent. The world was starting to listen.

The Win

By July 2021, it was inevitable. The IMF board had tabled and discussed the need for new SDRs in response to the global call.

On August 2, 2021, the Board of Governors approved a $650 billion allocation.

Africa received $33.8 billion. Ghana received $1 billion.

In the end, the Ministers got everything they went out and asked for. And I got to see African finance ministers, coordinated by UNECA, standing together as a bloc in pursuit of a common cause.

This was my first view of advocacy at this scale.

Admittedly, $33.8 billion wasn’t the same as the initial $100 billion request, but it eventually kick-started discussions at the IMF about reforming the formulas underpinning SDR allocation so Africa would get a bigger share the next time a new allocation was necessary.

Undoubtedly, Kristalina Georgieva, the Managing Director of the IMF, has been one of the developing world’s biggest allies to date. Always embracing ways to make the IMF a bolder, more supportive institution.

Last Reflections

I think it’s safe to say the majority of the Finance Ministers from that era are no longer at post. Most have been replaced, while some, like Tito, have unfortunately passed away.

I know Vera has moved on from UNECA, but the institution still stands, now led by Claver Gatete, a former finance minister.

The lessons remain from that era. In a moment of global panic, Africa organised early, aligned its message, and sustained it long enough for the geopolitical window to open.

Today, the architecture for coordination still exists, and there is space to talk to and learn from one another as countries such as Ghana, Kenya, Ethiopia and Côte d’Ivoire begin to implement bold structural transformation policies.

As we get deeper into 2026, Africa, and indeed the rest of the developing world, will face new challenges. But that advocacy effort between 2020 and 2021 showed me something important: when the developing world speaks with one voice, we are formidable.

It’s something we need to keep doing.

P.S. A special shout-out to Chichi Bodart, Kofi Ennin, Dobrina Poirier, Ari Krupkin, Dr Stephen Paduano, Michael Bediako and all the other advisors & aides in the background who worked with UNECA and their various ministers during this period. Without that crazy hard work, none of this might have happened.

Subscribe now

Share

It’s taken me a long time to write this.

A few things stand out in the life of every government. Each one hopes its “wins” define their legacy. My time was slightly different. As much good as was done, I feel what will live long in the memory of those who witnessed President Akufo-Addo’s second term will be less about the country’s response to COVID-19 and more about the imposition of the dreaded debt restructuring.

It was a brutal but necessary act. And I had the unfortunate privilege of seeing it go down up close.

It was a difficult period; I still remember the headlines. Old people losing their pensions. “Haircuts” entering the Ghanaian lexicon. Receiving calls from anyone who had my number, frantically asking what everything meant for their money and if it was safe. Then there were the protests, right in front of the Finance Ministry.

It was heavy.

I don’t want to use this article to go into too much technical detail. If you’re into that kind of stuff, there’s an excellent article by the CGD that gets everything right.

Just like the IMF piece, I’m drawing on memories from three years ago. The good thing is I shared this article with a few civil servants before publishing. They helped me fact-check the foggiest parts of my memory. Anyone who knows civil servants will attest that they’re notoriously low-key, mostly avoiding the spotlight. So instead of naming them, I’ll say a blanket thank you to all who helped me with fact-checking.

Now, let’s begin.

The Black Widow

The year is 2022. We start the year hoping things will work out. But in politics, one thing is guaranteed: something will go wrong.

On January 14th, the Ministry had rebutted factual inaccuracies in a Bloomberg article on Ghana’s debt situation. Add to that, there was an impasse in Parliament regarding the passage of proposed revenue bills from the last budget. Foreign investors and the markets were on edge.

A meeting is called in the Minister’s office. The usual suspects are there. I’m sitting in, wondering what’s happening.

Then a face pops up on the video call. It’s a lady at Moody’s, our analyst, Lucie Villa.

Some civil servants jokingly gave her the nickname “the Black Widow.” The myth was that any country she reviewed would get their credit rating downgraded. Seeing her on the screen, I figured: we’re in trouble.

She calmly walked through her model and delivered a verdict we felt was offensive. Said she could see we were on the way down, and the way things were looking, we would end the year with a debt-to-GDP ratio possibly above 90%.

How??? It sounded crazy. I knew we had issues, but come on—over 90%? No way. We didn’t borrow that much. I just kept looking around the room in disbelief.

The Minister echoed the same sentiment, asking her on the call what her assumptions were and how she could possibly think we’d get that bad. The whole room nodded. No way.

We offered to “fix” her model. Said maybe she didn’t get some assumptions right. She stuck to her guns.

The call ended with lots of upset faces. She doesn’t know what she’s talking about was the feeling. Downgrade our credit ratings based on stuff that isn’t true?? No way we would take that lying down.

Next thing I knew, myself and a few colleagues in the civil service were tasked with penning a press release. The government was going to take on Lucie! I grabbed my laptop, joined up with the rest of the team, and we put together a press release! Let the world know we’re not being treated fairly!

The press release got published in under a day, and we were off to the races. Just like that, the fight was on.

We decided to appeal the credit rating decision formally. It went nowhere. The President used his platform to advocate for us and for African countries in general, but these things didn’t take us off course from Lucie’s prediction.

Looking back, I’ll hold my hand up and say Lucie was right. We had really overspent during COVID. Our economy was on the fringes; a crash was coming. By July 2022, we were in the arms of the IMF.

Up Against the Wall

We were rapidly losing our dollar reserves. Something had to give.

The problem was that we had borrowed heavily from commercial creditors. And we had to pay them in dollars. Having to pay people in currency you don’t have enough of involves a lot of manoeuvring, but eventually you get found out.

Our currency depreciated by nearly 50%. We couldn’t keep up with payments. Whispers started about the need for a debt restructuring.

For those unfamiliar with the term, it’s just fancy English for telling people: “I borrowed money from you but I can’t pay back right now. I’ll pay you back one day, trust me! But let’s discuss a proposal I have. How about instead of me paying you $5 a month for a year, like we initially agreed, I pay you $2.5 a month for two years instead?”

It had been done before in Ghana during the HIPC years. But this was different. In 2002, creditors were telling us, “It’s cool, you don’t have to pay back”. Now we were about to tell unsuspecting creditors -“I actually can’t pay you back what I owe”.

We flew to anyone who could help. I mentioned the Saudi deal that didn’t happen in my last post. We also went to the US Treasury, Afreximbank, and anyone who would help us. We knew we had a liquidity crisis and had to get money in or we’d get found out.

We didn’t want to be the government that imposed a debt restructuring. With an election two years away? No way.

Plus, we saw what was happening to Zambia. It was super slow, and creditors couldn’t agree on the terms. We didn’t have that much time to waste or get bogged down in conversations that go nowhere, especially when you’re staring collapse in the eye.

I saw an article recently about how Senegal is saying no to a restructuring. Their Prime Minister said, “Restructuring would imply that Senegal is a bad student who borrowed money with the intent to steal. We told them (the IMF) that this country is a country of dignity, a country of pride.”

We probably said the same thing in our offices. But after a while, we couldn’t escape. It kept coming up: your debt is unsustainable; you must consider restructuring. That was the refrain from our development partners. By mid-2022, denial had run its course and we all had one big objective: design something that reduces our country’s debt-to-GDP ratio from near 90% in 2022 to 55% by 2028.

The Four Steps

As always, I’ll try to break down the process. It basically comes down to four steps.

Step 1: Get help. Hire financial and legal advisors who know international markets. We got Lazard and Global Sovereign Advisory, both French firms, and we spoke to people like Lee C. Buchheit, a genius who’d done this kind of work in Ecuador and Greece.

Step 2: Analyse. Determine how much debt relief you need and what categories of debt to include. We realised early that we had to exclude Treasury bills or the country would collapse.

Step 3: Prepare the proposal. The “Sovereign Ask”—outlining the quantum of relief and the methods (i.e. haircuts- reducing what you owe AND/OR maturity extensions, i.e. pushing back when you pay). This kicks off formal negotiations.

Step 4: Negotiate hard. And pray.

In between, there’s a lot of talking with banks and fund managers to converge on an offer that won’t destroy their books. You’re writing down people’s investments, so you need to find the right amount to cut off before their balance sheets break.

Those meetings took place mainly in November and December 2022. While most of the country was distracted by the World Cup, we were in rooms preparing for a restructuring.

The Farmers’ Day Meeting

The meeting that sticks out to me was the “secret meeting” held at the Bank of Ghana on 2nd December.

It was between the Minister, Central Bank governors, a small select group of finance officials, and local fund managers. It was wild.

We called the meeting to give them a heads-up that we would soon announce a debt restructuring. Of course, they had no idea why we called the meeting; you can’t afford leaks before Step 3, or your whole restructuring breaks down. All they got was a letter telling them to show up. Not much context.

Keep in mind, December 2nd was also a holiday in Ghana. I dragged myself to that meeting, nervous on many accounts. First, this was my first chance to see people react to our proposals in person. Second—and yes, this does matter—Ghana was playing Uruguay at the same time we called the meeting.

The Minister and Governors announce what we’re doing. Phones are taken at the door. Our advisors get into it.

I can see the faces drop in the room. What the hell is happening is the only way I can describe the looks on people’s faces.

At the same time, I’m checking my phone. Uruguay has scored. Ouch. 26 minutes.

Five minutes later, even more misery. Uruguay has scored again. 2-0 before halftime.

My mind is thoroughly checked out of the meeting at this point because I can see the groans. People are not happy. Here we are, talking to the bloc of people who traditionally vote for our government—the business elites—and we’re admitting their investments are about to take a hit.

All this while Ghana is losing in the World Cup.

Our Lazard advisor managed the Q&A section. It was brutal to watch. At some point, people just agreed that the best thing to do was to go back to their offices and work with their CFOs to figure out how to handle this when it was announced.

There’s a well-known aphorism: oppositions don’t win elections, governments lose them.

That day, I saw losses on different fronts. Ghana got kicked out of the World Cup. And looking at the shocked faces in the room, my brain told me: we had just started our slide into opposition.

The Final

The work continued behind the scenes. Drafting and redrafting until a package was ready to be presented to the country. On December 5th, we announced a domestic debt exchange programme.

Next was the announcement relating to our external debt. Fast forward to 18th December. The day of the World Cup final.

At that point, I was tired. I was in the office—it was a Sunday, not unusual for civil servants. I’m in a room with the usual suspects, and we’re crafting the press release for tomorrow that announces we’ll suspend payments on our external debt.

I’m really tired. Restructurings are an ordeal to get through. The negotiations hadn’t even started, and I’m worn out. World Cup final, and I can only follow updates on my phone. My phone is buzzing; Messi has won a World Cup! I smiled for a second, a happy moment amidst a difficult period.

I have a crazy amount of respect for the Finance Ministers and Central Bank Governors. To stand in front of the country and deliver this kind of news, knowing the backlash that would come, that takes an insane amount of strength. They did the right thing to fix a broken situation. It’s not easy doing that while being criticised from all angles.

19th December

We announce a suspension of payments on our external debt.

It’s on the internet. The calls start. All I can tell people on the phone is: don’t worry, we’ll have an FAQ that explains everything.

Inside, I’m panicking. If we get this wrong, the country will never forgive us. And I’m really, really tired!

What followed was months of long meetings. Travelling to meet investors and development partners that lent us money, to explain our situation and let them know we were ready to fix our economy so we could eventually bounce back from the restructuring.

Ghana had and indeed still has a lot of goodwill. Most creditors wanted to see us win.

The Bottom Line

Before I left my role in government, the results showed all involved did a good job. It wasn’t ideal, but the debt restructuring eased the foreign currency burden on the budget, lowered external debt service costs, and moderated Ghana’s total debt obligations.

By October 2025, the IMF’s Debt Sustainability Analysis showed Ghana had moved from high risk to moderate risk of debt distress. This is a major milestone on the road back.

It’s hard to take any form of credit for a process that hurt so many. But just like the decision to go to the IMF in 2022, the restructuring was the right thing to do. I know the impact is still being felt today; it’s still painful for those whose investments were affected. I’m not sure that will ever go away.

For now, I can only hope this article was helpful to someone. And to my Senegalese brothers in government right now trying to avoid a restructuring—keep up the good fight!! I hope you succeed!

It’s been a while since my last article, but I’m grateful the Inside the Room piece on the IMF got a lot of attention. Before I release the follow-up, which will detail my experiences during Ghana’s recent sovereign debt restructuring, I want to return to a topic I’ve become more interested in recently: education in Ghana.

It matters to me, obviously, because I’m Ghanaian. But also because I’m a firm believer that one of the things holding us back has been our failure to get the education sector right.

Don’t get me wrong, Ghana has a crazy number of smart people. And I acknowledge I’m writing from a position of privilege, having been educated in private schools in the capital city. But I’ve passed through “the system”, and my general feeling is that it hasn’t gotten better over time, even though it should.

Ghana has been tinkering with its education sector for a while. Free basic education since 1995, Free Senior High School (SHS) since 2017, and nine reform programmes since independence. But with all this tinkering, are students actually learning more?

Before diving into the data, it’s worth thinking about what actually drives school performance. Noah Smith has a useful framing for this:

A is how much natural talent matters. B is how much a good school matters. And C is how much better talented students do at a good school, relative to less talented students who can’t handle the work.

I think most Ghanaians instinctively believe A and C aren’t problems for our kids. We’ve got talent. The question is B: the quality of education we’re providing. But lately, it feels like the national conversation covers anything but that.

The Conversation We’re Not Having

Look at the last couple of months. First, we’ve had way too many conversations about girls’ hair, with the Education Minister going as far as saying schools aren’t beauty contests. Honestly, I like the new education minister, but I genuinely don’t understand why the length of a girl’s hair is still an issue in 2025.

Then we had conversations about free sanitary pads. I’m never against any social intervention, but is this really what will reduce discontinuation? My earlier research showed that location and parental education levels are far bigger drivers of who drops out. I’d like to see the evidence of the influence of menstrual poverty on girls’ school outcomes—at least I hope that’s what is driving this policy. But here we are.

And of course there’s the No Fee Stress policy, essentially Free SHS 2.0, but for tertiary education. There’s also the debate about how many people actually benefited from Free SHS in the first place—was it 5.7 million or 3.5 million? The answer depends on your political leaning.

Throughout all these discussions, the questions that keep nagging me: what are kids actually learning? And are we sacrificing quality for quantity?

Now, the randomistas would tell me that to answer this question, I should design something akin to an RCT or quasi-experiment. But for now, I think we can get a good direction of travel from data released by the Ghana Statistical Service on exam performance over the past decade.

What the Data Actually Shows

Honestly? It’s a mixed picture.

At the BECE level (Basic Education Certificate Examination- the junior high school exit exam), people are doing okay. I know there’s concern about Ghana’s basic education, but in terms of exam results, you have a pretty decent picture.

Across English, Maths, Science, and Social Studies from 2012–2021, almost all subjects hover between 70–78% pass rates. Honourable mention has to go to policies like the Capitation Grant, which provides financial assistance to public basic schools.

Secondary Education: Where Things Fall Apart

WASSCE (the senior high school exit exam) is a different story. This is where performance jumps all over the place.

I haven’t been high school age for a long time—my age shall not be revealed in this post, all I’ll say is I’m familiar with the sound of dial-up internet—but honestly, Ghana’s teenagers have deeper issues than whether girls can keep their hair.

The overall level of WASSCE performance has drifted downward, with massive swings year to year. The 2015/16 collapse in pass rates was brutal: Maths dropped to 24.4%, Science to 23.6%. Something broke that year. I can only think it had to do with “dumsor”, Ghana’s bout with power blackouts, as some students would have been studying in the dark.

The GSS data only covers up to 2021, but an online search on 2024 results show continued decline. English dropped from 73% to 70%, Integrated Science from 67% to 59%, Social Studies from 77% to 72%. Only Maths improved slightly. More students are taking the exam than ever, approximately 460,611 candidates, but the quality question remains unanswered.

Look at the long-term trend in English. Scores have drifted downward for years. English is meant to be our national language BUT young Ghanaians are failing english exams at increasing rates. Where exactly are we going as a country?

The Free SHS Question (Answered Honestly)

Here’s where I need to be careful, because lazy analysis would say “Free SHS killed quality.” That’s not what the data shows.

The picture is genuinely mixed. Comparing average WASSCE pass rates before and after Free SHS launched in 2017: English dropped from about 64% to 52%. Social Studies dropped from 68% to 65%. But Maths improved slightly, from 44% to 54%. Science held steady around 50-51%.

You can’t cleanly say the policy made things worse. English and Social Studies declined, these are literacy-heavy subjects most sensitive to foundational gaps from junior high school. But Maths improved. Science held. The pattern is structural, not simple.

What you can say: Free SHS massively expanded access. After the policy, Ghana significantly increased students sitting WASSCE, including many from disadvantaged backgrounds who would have dropped out. This is genuinely good.

But when a system expands participation rapidly, average preparedness falls and variance increases. This pattern has been observed across the continent: Kenya, Uganda, South Africa and recently Tanzania all saw similar dynamics. So a decline in averages doesn’t necessarily mean a decline in learning. It may reflect the inclusion of previously excluded students.

The socialist in me likes knowing that teenagers can show up at school without worrying about fees. But Free SHS expanded access without fundamentally shifting learning outcomes.

And let’s not forget the pandemic. COVID-19 disrupted learning worldwide, and Ghana was no exception. We haven’t fully reckoned with how much ground was lost during those years of disrupted schooling.

There’s also something I don’t have hard evidence for, but suspect: schools across the country may have begun lowering their standards, passing kids who didn’t know the material. It’s possible. And if true, even our “stable” exam results at junior high school level might be masking a deeper rot.

The Real Problem: Teachers

What is true however is the strong negative correlation between pupil-trained teacher ratios and pass rates.

Northern Region, Western Region, Upper East Region, they have the worst ratios, around 22 pupils per trained teacher. And they consistently underperform.

Meanwhile, Ashanti has great ratios (around 14:1) and great outcomes. Greater Accra has surprisingly poor ratios (around 20:1) despite being the capital and underperforms at WASSCE despite strong BECE scores.

Regional inequalities in teacher allocation map almost perfectly onto exam outcomes. Ghana needs its teachers. We need to keep them coming off a conveyor belt.

The Real Problem: Geography

Just like I found in my discontinuation research, the North is suffering.

Both at BECE and WASSCE, northern regions consistently underperform. In 2021, Upper West had BECE pass rates around 40-50% while Brong Ahafo was hitting 80-90%. At WASSCE, the gaps are even starker with Northern Region hitting 20-30% in Science while Brong-Ahafo manages 87%.

And here’s what should make policymakers uncomfortable: Greater Accra, supposedly our most resourced region, only hits 40-60% on WASSCE despite strong BECE scores. The capital city is underperforming.

Meanwhile, I suspect BECE national averages look stable because high-population southern regions pull up the mean while northern regions consistently drag. High spatial inequality masked by stable national averages.

The Gender Story (It’s Not What You Think)

From a gender standpoint, girls match boys at junior high school. But they fall behind as they get older, and the gap widens in subjects most sensitive to school quality and resource constraints.

This is important because it undermines the argument in some circles that girls keeping their hair distracts them from learning. If hair or beauty or whatever was the issue, you’d see the gap at BECE too. You don’t.

What you see is girls performing fine under a structured, largely untouched system that kinda works (basic education), then falling behind when the system is overstretched (secondary education). The issue isn’t gender or even talent to some degree. What we see in terms of lower pass rates at high school is basically what happens when a system gets stretched.

What Should We Actually Do?

Stop with the blanket “free for all” policies.

Instead of paying academic fees for ALL first-year students at public tertiary institutions through the No Fee Stress policy, seek out the poorest. I know it’s a vote getter, but we need policies that work. Voters understand that.

Here’s what scares me: if English scores keep declining, universities will be admitting students who struggle to read and write at the level higher education demands. We may end up with a bunch of second and third year uni-students with poor written and even spoken English.

From a regional standpoint, the data shows students from the North need most support. Why not design a formula for fee coverage based on where you’re from or your high school scores? Targeted, sustainable, impactful.

Also, invest in teachers. Aggressively. The correlation between teacher availability and outcomes is too strong to ignore. Regional inequalities in teacher allocation are driving regional inequalities in learning. We need policies that get our best teachers into underserved regions and keep them there. That means a proper package: incentives, housing, career progression, the whole show!

And we need a political alliance on education. I saw too many articles during the last election season about school performance being better under one government than another. That’s not the point. The current government kept Free SHS. Good. But both major parties need to agree that education is too important for political point-scoring. Focus on making things better, on targeting resources, on building capacity. Otherwise the result will be an erosion in the quality of our nation’s human capital, above and beyond the misallocation we already have.

And maybe—just maybe—worry less about hair length.

The Bottom Line

Junior high scores are stable. Admissions are at an all-time high. But senior high school scores are volatile, and secondary education remains the system’s weak point.

High school is where we’ve thrown the most money and policy attention, but we haven’t exactly gotten it right. If we don’t address the myriad of issues at that level, the cracks will creep into tertiary education too.

The good news? We have the data. The Ghana Statistical Service has given us everything we need to understand what’s working and what isn’t. The question is whether policymakers will use it.

I know I might be stressing out about the education sector for no reason. But if you’ve seen the trends in exam pass rates and you’re not worried, you’ve unlocked a level of inner peace I can only aspire to.

Say a prayer for anyone you know in high school. I wouldn’t want to be a teenager in 2025 for a whole lot of reasons. But let’s hope, for Ghana’s sake, we can get education policy right soon enough, and that it shows up in star-performing students, proper infrastructure, and awesome teachers.

PS: Thank you Worlanyo Ahiaxornu for cleaning up the visuals in your spare time

The 5th review means Ghana is still on track with the 2022 programme. We’re meeting benchmarks, the money keeps flowing. In the narrow sense, that’s progress. We’re following the same script we’ve followed 16 times before. The hope is this government can break with the trend so in 4 years we don’t go again.

This takes me back to 2022, when I was part of the team that negotiated the programme we’re still implementing today. People ask me what it’s like. So here’s the unfiltered version of how Ghana ended up at the IMF—again—and what actually happens behind the closed doors.

The Pattern We Can’t Break

Ghana has entered into 17 IMF agreements since independence in 1957. That’s basically going to the IMF every 4 years. Start strong, crash, get IMF support, stabilize, crash again. Repeat.

I’ve seen this cycle from the inside. I worked for a Finance Minister who genuinely didn’t want to repeat it. We came out of an IMF programme in 2019. Three years later, we were back.

Here’s how it happens.

Year 1: Big Momentum

Every Ghanaian government starts strong. First six months, maybe the first year, you’re pushing reforms, signaling fiscal discipline, making tough calls. The civil service is engaged, Cabinet is aligned, things are moving.

Then around Year 2, things start to slowly fall apart.

Not because people forget how to do their jobs or suddenly become incompetent. It’s because interests take over. Political considerations start outweighing technical ones.

I’m being politically correct here. It’s some parts greed, some part not accepting that we’ve been living beyond our means for a long time and need to overhaul the whole system. Other parts naivety—thinking somehow a government can achieve special results without making drastic breaks from a system that isn’t working.

You don’t need to be an insider to know Ghana runs a procurement economy. People make money from contracts. When those contracts add value and are genuinely good for the country, fine. But more often than not, you’re seeing questionable stuff get pushed through. It’s not unique to Ghana—this is elite capture, a developing world thing. But we can save that discussion for another post.

By Year 3, the damage is done. Fiscal discipline erodes. The budget stops balancing. Debt becomes unsustainable. And you end up in the arms of the IMF.

The Moment We Knew We Were Screwed

In our case, the breaking point came early in 2022, at an Economic Management Team (EMT) meeting. We sort of had the feeling the end was coming when we got hit with multiple credit ratings downgrades in early 2022. The signs were there.

EMT was officially established in 2001 to strengthen economic governance and policy coordination. It’s typically chaired by the Vice President and brings together the key economic officials—Finance Minister, BoG Governor, heads of planning, debt management, and economic strategy. It serves as an advisory body to the President on economic matters.

The Director for Debt Management made a presentation to the team. He laid out, in brutal detail, how unsustainable our debt had become. Then the Director of Economic Strategy and Research went through our fiscal situation. The picture was clear: we were screwed and needed a bailout.

You’re probably asking: why didn’t EMT see this earlier?

That’s the tough question. There’s always a sense in government—especially at the Finance Ministry—that no matter how bad things are, the problem can be solved. That feeling was even stronger under my boss, who really didn’t want to go to the IMF. So he was exploring other options, looking for alternative support.

That always made it seem like things were bad but manageable. Like there was something being done in the background to avoid collapse. I’ll give one example.

The Saudi Deal That Didn’t Happen

We were in talks with the Saudi government to see if we could secure an $11 billion support package to shore up the economy.

We sent officials to Riyadh to work out the structure. The deal was designed around balance of payments support, trade facilities for crude oil and refined products, domestic debt management, and budget financing. There was even a component to cushion the financial sector from potential domestic debt restructuring.

In the end, the deal didn’t go through. There were other conversations also being had, one with the US for a similar credit line, but they didn’t work out either. We ended up back at square one: overspent, with a budget that wouldn’t balance and an economy near crash.

The Chair of EMT made the right call. He pushed for us to request IMF support. At that point, it was the only responsible thing to do. On the fiscal front, we had boxed ourselves into a corner we couldn’t manage, and the only option left was to admit we needed help.

How the Process Actually Works

Once the decision is made at EMT and gets Cabinet approval with the President’s backing, a bunch of calls start flying.

Step 1: The Call
The Finance Minister and BoG Governor call the head of the IMF, detailing the need for support. Then we make a formal application, signed by both the Governor and Finance Minister.

Step 2: The International Lobby
You also travel a lot. In our case, we went to Washington to speak with the US Treasury Secretary. We spoke to the German Finance Minister, the Indian Finance Minister, and others. You’re basically talking to finance ministers for countries that can vouch for you at the IMF Board level, because the IMF is organized into voting blocs.

Why? Because eventually, your programme gets presented to the Board for approval. You want allies who are confident you’ll stick to the plan and get through the necessary reforms needed to fix your situation.

Step 3: The Diagnosis
Meetings begin. Loads of them. The Fund sends a team to Accra to understand what went wrong. It’s kind of funny, actually—the same people responsible for the crash are now giving nice presentations to the IMF about how things got so bad and what we can do to fix it.

State-owned enterprises that pose big fiscal risks—in our case, Cocobod and the Electricity Company of Ghana, for instance—are drawn into the meetings so the Fund can get a full picture. It’s draining. But also revealing. You see how problems compound across the system until the weight becomes too much.

Step 4: Design Your Own Recovery Plan
Then you go back and do your homework. The country that crashed has to design its own recovery programme to present to the Fund.

When a country seeks IMF support, the government agrees to adjust its economic policies to overcome the problems that led it to seek help. These policy adjustments are conditions for IMF lending and serve to ensure the programme’s objectives are met and the country can repay the Fund.

But—and this is important—the member country has primary responsibility for selecting, designing, and implementing the policies. The IMF doesn’t design your policies for you. This system is designed to promote national ownership of the reforms.

We spent about a week or so locked in a hotel with all the top officials from the Ministry of Finance, Bank of Ghana and key agencies, mapping out a recovery plan. It was actually exciting—charting your way out of the mess you created.

We designed something called the PC-PEG (Post-COVID Programme for Economic Growth). The name changed countless times, but that’s what we settled on. Weird acronym aside, PC-PEG spelt out how we intended to clean up our mess while also addressing some legacy issues that perennially affect our economy.

We also brought in outside voices—academics, civil society groups—to offer their input on what should be included.

Step 5: Get the Fund’s Sign-Off
Once the technical work is done, the plan is presented to the IMF. Both parties agree: this document will guide the programme. Milestones are drawn based on what the government commits to.

For instance, if you say you want to reform the energy sector by charging cost-reflective tariffs, the Fund might ask: “Will you agree to quarterly tariff reviews? Semi-annual adjustments based on inflation?”

It’s back and forth until both sides agree on how to assess progress. Those agreements become performance targets—markers to show you’re moving in the right direction.

Step 6: Outline in the Budget
Once the programme is finalized, you outline it in the national budget.

Everything I’ve described sounds like it takes a couple of weeks. It takes months. We applied for IMF support in July 2022, if I remember correctly. Didn’t get through the process until six months later. Lots of back-and-forth, not just technical but political—because you’re also lobbying people internally to build consensus on the policies that will underpin the programme.

Implementation: Government Becomes a Business

After that, you enter implementation mode. Government basically becomes like a private business. People have KPIs to meet. Progress is reported at Cabinet and other meetings for onward reporting to the Fund.

Then you have the reviews. IMF staff come to town and work with government to verify whether you’re on track. It’s independent verification—they’re making sure the progress is real and not just flowery stuff you’re reporting.

Lots of data validation. Data requests flying back and forth. You hear terms like “fiscal framework” far too often. In our case, endless chat about debt sustainability and fancy terms like LIC-DSA (Low-Income Country Debt Sustainability Analysis).

After these reviews, an Article IV report is published—an in-depth review of the member country’s economy.

If you get positive checks, you move toward a staff-level agreement. That’s IMF staff saying, “We came to check their performance, and they’re doing pretty well vis-à-vis the targets we all agreed on.”

This gets passed to the IMF Board, who review the reports and decide if they’re satisfied with the progress. If yes, they release the money.

When the money comes in, there are a lot of happy faces at the Central Bank and Ministry of Finance. We have our lifeline.

The Silver Lining (If You Can Call It That)

The IMF programme also signals credibility. The whole world sees you’re successfully going through it, which unlocks other funding—FDI, development partner support, eventually access to international capital markets again.

So it’s this whole thing where you surrender and announce you crashed, but once you take the bitter pills you prescribed (in consultation with the Fund), other people see you’re trying hard and decide to support you with investment.

That’s the cycle.

Why Do We Keep Doing This?

I’m sure I’ve forgotten bits—this is from three years ago, six months of nonstop meetings, and my memory isn’t getting any better. But if you ever wondered what happens when Ghana goes to the IMF, this is basically it.

The bigger question, though, is why we keep repeating this cycle. Seventeen times since independence. Every government thinks it’ll be different. Every Finance Minister believes they’ll avoid the IMF. And yet, here we are.

Maybe the next post should be about that.

If you found this useful, subscribe for more insider perspectives on Ghana’s economy and policy.

A lot has been done, and we can explore those gains in future posts. But today, I want to focus on something more specific: educational inequality and what actually drives school discontinuation in Ghana.

This analysis draws on the recently released Annual Household Income and Expenditure Survey (AHIES) data—the first nationally representative household survey of its kind in Ghana. The findings challenge some common assumptions about who drops out and why.

I won’t claim the models here are perfect. We’re working with cross-sectional data and statistical associations, not causal proof. But the patterns are clear enough to offer useful insight for policymakers trying to close educational gaps.

Here’s what surprised me most: Gender gaps in school dropout largely disappear once you account for where students live and their parents’ education levels. Location and intergenerational disadvantage matter far more than gender alone.

This has major implications for how we design education policy.

The Finding That Changed My Mind

I started this research assuming female students faced the highest dropout risk. That’s the dominant narrative in education policy, and for good reason—historically, girls have been systematically excluded from schooling across much of the developing world.

But when I analyzed nationally representative data from 77,541 Ghanaian children aged 5-18, something unexpected emerged. In my baseline model, girls did have about 9% higher odds of discontinuing school than boys. But watch what happens when I add controls:

Model 1 (Gender only): Girls 9% more likely to drop out (statistically significant)
Model 2 (Add location): Girls 9% more likely (still significant)
Model 3 (Add parental education): Girls 17% more likely (still significant)
Model 5 (Full interactions): Girls 4% more likely (NOT statistically significant)

By the time I accounted for where students live, their parents’ education levels, and how these factors interact with gender, the female disadvantage vanished.

This doesn’t mean gender discrimination doesn’t exist. It means gender disadvantage operates through structural channels, particularly geography and intergenerational poverty, rather than as an independent force.

What Actually Predicts School Dropout in Ghana

1. Location Is Destiny

The most powerful predictor in my models wasn’t gender. It was rural residence.

Rural students have 52% higher odds of dropping out compared to their urban peers, even after controlling for parental education, age, and region. This gap persists across every model specification I tested.

But it gets more stark when you look at regional variation:

The North-South divide is brutal. Savannah region has a 21.5% discontinuation rate. Central region? 8.77%. That’s nearly a 13 percentage point gap—more than double.

Three northern regions (Savannah, Upper East, Upper West) consistently show discontinuation rates above 15%. Meanwhile, southern regions cluster around 9-11%.

This isn’t just about education infrastructure. It also reflects broader spatial inequality: rural northern regions face higher poverty rates (64.6% multidimensional poverty in rural areas vs. 27% urban), worse roads, limited electricity, and fewer economic opportunities.

Simply put: When your entire region is structurally disadvantaged, both boys and girls struggle to stay in school.

2. Your Mother’s Education Matters More Than Your Gender

Maternal education emerged as a powerful protective factor, reducing dropout odds by approximately 28%.

But here’s where it gets interesting: this protective effect is weaker for girls than boys.

Students with educated mothers have lower dropout risk overall. But when I added interaction terms, I found that girls with educated mothers don’t benefit as much as boys with educated mothers do.

Why? I suspect it reflects labour market realities. In Ghana, educated men earn significantly more than educated women. Parents may unconsciously calculate that investing in a son’s education yields better economic returns for the family than investing in a daughter’s education—even when the mother herself is educated.

This is the kind of finding that matters for policy. It suggests that simply educating mothers won’t automatically close gender gaps unless we also address the gendered economic structures that shape parental investment decisions.

3. The Age 15 Cliff

Most kids stay in school through primary and junior high. Then something dramatic happens around age 15:

The survival curves are nearly flat until about 10 years in the education system (roughly age 15). Then they drop sharply. This is the Junior High School (JHS) to Senior High School (SHS) transition, and it’s where we lose students in massive numbers.

The 15-18 age group has a 32.83% discontinuation rate. Compare that to 2.25% for ages 5-10 and 6.03% for ages 11-14.

Ghana’s Free Senior High School policy removes tuition barriers at this critical juncture, which helps. But my findings suggest fee elimination alone isn’t enough. Students dropping out at 15 face non-financial barriers: distance to schools, lack of boarding facilities, opportunity costs of staying in school when they could be working, early marriage (especially in northern regions where child marriage rates exceed 10%).

What This Means for Policy

If educational discontinuation results from intersecting structural disadvantages rather than gender alone, our policy responses need to change.

Stop Targeting “Girls’ Education” as a Standalone Category

Current programs often target girls broadly without considering context. But a girl in urban Accra whose mother completed secondary school faces vastly different barriers than a rural girl in Savannah whose parents never attended school.

Better approach: Target students at the intersection of multiple disadvantages—rural residence + low parental education + northern region. Many of these students will be girls, but the targeting logic should be structural vulnerability, not gender alone.

Invest in Adult Education for Rural Mothers

Since maternal education has such strong protective effects, adult literacy programs for rural mothers could be high-impact. But they need to be designed for accessibility:

• Evening classes that work around agricultural seasons

• Functional literacy focused on supporting children’s education

• Link participation to existing social protection schemes (LEAP benefits, school feeding priority access)

• Recruit local women with basic education as facilitators

This creates a virtuous cycle: educated mothers are more likely to keep their children, especially daughters, in school.

Build a Risk-Based Transition Support System

The Free SHS policy is universal. But dropout risk isn’t universal.

Ghana could create a vulnerability index using variables such as rural residence, region, maternal education, and household wealth. Students above a certain risk threshold (say, 70th percentile) would qualify for enhanced support:

• Transportation stipends

• Boarding facility priority

• Learning materials

• Mentorship programs

Zimbabwe’s Basic Education Assistance Module (BEAM) does something similar. It’s more efficient than universal programs because it concentrates resources where dropout risk is highest—the JHS-SHS transition for vulnerable students.

Treat Education Infrastructure as a Multi-Sectoral Problem

The regional disparities I found aren’t just about schools. They’re about roads, electricity, housing, and economic development.

Currently, the Ministry of Education has a $2.5 billion infrastructure gap and is trying to address it alone.

A Better approach might involve: Tiered spatial prioritization for multi-sectoral investment

Tier 1 (Intensive intervention): Savannah and North East regions

• Coordinated investment across education, roads, energy, local government

• These regions have 2-3x the discontinuation risk of the national average

Tier 2 (Targeted improvements): Bono East, Upper East, Upper West, Northern

• Build on existing infrastructure, address specific bottlenecks

• Still elevated risk but not as extreme

Tier 3 (Maintenance and incremental improvement): Remaining regions

• Focus resources on quality improvement rather than major expansion

Altogether, when you improve roads in Savannah, you reduce travel time to school AND facilitate market access for farmers. When you electrify rural areas, you enable after-school studying AND support small enterprises.The returns compound.

The Uncomfortable Truth About Gender Gaps

My finding that gender effects become insignificant doesn’t mean we should ignore gender. It means gender disadvantage is mediated by structural factors we can actually address through policy.

Girls face higher dropout risk primarily when they:

• Live in rural areas

• Come from households where neither parent is educated

• Reside in northern regions with limited infrastructure

That’s actionable. As a country, Ghana can target those specific contexts rather than implementing broad “girls’ education” programs that may miss the girls at highest risk while including girls who face minimal barriers.

The gender lens remains important for understanding why rural residence or low parental education affects girls differently than boys. Labour market discrimination, early marriage norms, and household resource allocation all shape these patterns.

But if we want to design cost-effective interventions, we need to target the structural conditions that produce gendered disadvantage, not just gender itself.

What’s Next

This analysis raises questions I couldn’t answer with cross-sectional data:

• Why do girls with educated mothers benefit less than boys? Is it about parental investment decisions, labor market expectations, or something else?

• How has Free SHS affected these patterns since 2017? (My data is from 2022, so post-policy)

• What specific barriers drive the age 15 dropout spike beyond financial costs?

These need qualitative research, interviews with families, teachers, and students who left school. The quantitative analysis tells us where to look. Now we need to understand why. At least we now have a better idea of what to look out for!

Data note: This analysis uses the Annual Household Income and Expenditure Survey (AHIES) 2022, Ghana’s first nationally representative household panel survey. Sample: 77,541 children aged 5-18 across all 16 regions. Methods: Multivariate logistic regression with interaction effects and Cox proportional hazards survival analysis.

This is my first data-driven deep dive on this page. If you found it useful, subscribe to get more evidence-based analysis on policy in Ghana. Next up: examining what the data reveals about Ghana’s National Health Insurance System

Share

.

You wouldn’t have known my name, but you might have seen my work. Ideas sketched out late at night in my bedroom that eventually made their way into budget statements. Policy recommendations that became national decisions. Speeches I wrote for the President and Finance Minister of Ghana that shaped how major reforms were communicated. For six years, I served as a Technical Advisor to Ghana’s Finance Minister, quietly influencing decisions that impacted millions of people.

I’ve been in rooms with world leaders and finance ministers. I’ve listened to some of the smartest policy minds from around the globe. And in my own way, through memos, briefs, and carefully crafted speeches, I’ve gotten my ideas into those big conversations when critical decisions were being made.

It was an incredible privilege. There’s something almost surreal about watching your ideas become concrete policy—about being in the room when important decisions, good and bad, are made.

But I’ve also stayed quiet. Maybe too quiet.

Why Now?

Two things changed for me.

First, I moved to Oxford for my MSc. Being around fiercely intelligent people who debate openly, sharing their half-formed thoughts and learning in public—it’s been liberating. I’ve realised that there’s always more to learn, and learning happens faster when you share what you know.

Second, quite frankly, a close confidant pushed me. They saw the gap between what I know from being inside government and what the public conversation misses. They convinced me it was time to speak up.

So here we are.

The Gap Between Policy and Evidence

Here’s what I’ve learned from being in the room: policy decisions often look nothing like what the headlines suggest. The “why” behind a budget line or a reform is usually more complicated, more human, and more influenced by missing data than anyone admits.

There’s a gap between what the evidence says and what actually gets done. Sometimes it’s because the evidence doesn’t exist. Sometimes it’s because it exists but no one’s looking at it. And sometimes it’s because politics is politics, and data only gets you so far.

I want to help bridge that gap.

What to Expect Here

This page is for anyone trying to understand what actually works in policy, especially in Ghana and across Africa.

I’m drawing on data where it’s available, practical experience from my years in government, and the kind of analysis you only get from being both an insider and now, an observer. My writing will focus on:

• Evidence-based breakdowns of social sector policy in Ghana (education, health, social protection)

• What the data actually says versus what we assume it says

• Inside perspectives on how policy gets made—the constraints, the trade-offs, the human decisions behind the spreadsheets

• What works and what doesn’t, backed by evidence and experience

I’ve been inspired by writers like Noah Smith and Ken Opalo, who make complex policy accessible and useful. Ken’s work, in particular, showed me what’s possible when you speak clearly to an African audience. I hope to contribute something similar.

Who This Is For

Whether you’re a politician trying to make better decisions, an academic studying development, a policy nerd like me, or just someone randomly interested in what happens behind closed doors in government, I hope you’ll find something useful here.

Share

Share Benjamin Hornsby-Odoi

Subscribe now