The anchor without the money: What Ghana should — and should not — expect from its new IMF arrangement

The empirical case that the Policy Coordination Instrument is a credibility tool, not a growth engine — and why judging it by the wrong yardstick will mislead the national conversation.

By Dr Stephen LARTEY 

The graduation nobody has stress-tested

In May 2026, Ghana concluded its three-year Extended Credit Facility with the International Monetary Fund and chose, in its place, a Policy Coordination Instrument — a PCI. The decision has been received, broadly, as a milestone: a country stable enough to remain engaged with the Fund without needing its money. After a sovereign default in 2022, a domestic debt restructuring in 2023, a completed Eurobond exchange, and three years of fiscal adjustment, the symbolism is real. Ghana is, in an important sense, standing on its own feet.

But symbolism and substance are different things, and graduation language has a way of flattening the question that actually matters. What does Ghana gain, concretely, by swapping a financing programme for a non-financing one? And — the question almost nobody is asking — what should it deliberately not expect?

This article argues that the PCI is best understood as a credibility instrument, not an economic stimulus; that the evidence gives no reason to expect it to lift growth; and that the national conversation risks judging the instrument against a test it was never designed to pass.

What the instruments actually are

The distinction at the heart of this is simple and consequential. An Extended Credit Facility is a loan. The Fund disburses credit, in tranches, against conditionality — performance criteria, structural benchmarks, periodic reviews. The money and the discipline arrive together.

A Policy Coordination Instrument carries no financing whatsoever. There is no loan, no disbursement, no drawing right. What the PCI provides is IMF assessment of a country’s reform programme, endorsement by the Fund’s Executive Board, and — the intended prize — a signal. A signal to sovereign bondholders, to rating agencies, to bilateral and multilateral partners, that an independent and credible institution judges the country’s policy framework sound.

The PCI, in other words, is an instrument of credibility. It is the predecessor instrument — the Policy Support Instrument, or PSI, which several African economies have used — modernised. Its theory of action runs entirely through confidence: it works, if it works, by lowering the price others charge Ghana for risk, not by injecting resources into the Ghanaian economy.

That theory of action deserves to be tested against evidence rather than asserted. So I tested it.

The study: separating the money from the message

To ask whether financing and non-financing IMF arrangements are associated with different economic outcomes, I assembled a balanced panel of eight Sub-Saharan African economies — Ghana alongside Benin, Côte d’Ivoire, Kenya, Rwanda, Senegal, Tanzania and Uganda — over 2000 to 2024, giving 200 country-year observations. The data are drawn from the World Bank’s World Development Indicators, the IMF’s World Economic Outlook database of April 2026, and the IMF’s Monitoring of Fund Arrangements (MONA) database, which records every Fund programme with its exact dates and type.

The central design choice was to code two distinct treatment variables. The first captures the years a country was under a financing arrangement — the Extended Credit Facility and its relatives, including the older Poverty Reduction and Growth Facility, the Extended Fund Facility, and Stand-By and Standby Credit arrangements. The second captures years under a non-financing arrangement — the PSI and the PCI, the category to which Ghana’s new instrument belongs. The analysis then asks how each relates to five outcomes: economic growth, inflation, public debt, the primary fiscal balance, and foreign direct investment.

A word on method, because it bears on how much weight the findings can carry. An initial design considered the synthetic control method — constructing a “counterfactual Ghana” from a weighted blend of comparator countries. That design had to be abandoned, and the reason is itself instructive. Verified MONA records show that 93 of 175 comparator country-years in the donor pool were themselves under IMF financing programmes. There is, simply, no clean pool of “untreated” African economies against which to benchmark a treated one — IMF engagement is the regional norm, not the exception. The analysis therefore uses a two-way fixed-effects panel, which estimates the average within-country association between Fund arrangements and outcomes, and which is not undermined by the fact that many countries have programmes.

One caveat must be stated plainly and kept in mind throughout: these are associations, not proven causal effects. Countries enter IMF programmes precisely because they are in difficulty. A panel of this kind, even with country and year fixed effects and a check for pre-existing trends, cannot fully separate the programme from the trouble that prompted it. The findings below are informative; they are not the last word.

Finding one: financing arrangements track growth — non-financing ones do not

The headline result is also the one most directly relevant to Ghana’s decision.

Across the eight economies, years under an IMF financing arrangement are associated with modestly higher real GDP growth — a coefficient of roughly 0.7 percentage points. The estimate survives conservative statistical testing appropriate to a sample of only eight clusters, including a wild-cluster bootstrap, which is the correct inferential tool when clusters are few; the bootstrap p-value is 0.035. It is not a fragile result. A dynamic specification finds no evidence of a pre-existing trend: the one-year lead of the financing indicator is small and insignificant (coefficient 0.31, p = 0.59), which is to say growth was not already accelerating before programmes began.

The non-financing arrangements show nothing of the kind. The estimated growth association for the PSI/PCI category is, for practical purposes, zero — a coefficient of 0.029, statistically indistinguishable from no relationship at all. When financing and non-financing arrangements are entered side by side in the same model, the contrast is stark: the financing coefficient bootstraps to p

The reading is direct. Whatever the growth correlation of IMF programmes reflects — and given the caveat above, it may partly reflect the cyclical recovery that follows a crisis trough — it belongs to the programmes that come with money and binding conditionality. It does not belong to the lighter-touch, non-financing instruments. The PCI is in the second category.

Finding two: the fiscal numbers move to a different drummer

If the growth result is the headline, the fiscal result is the corrective.

Neither financing nor non-financing arrangements show a reliable association with lower public debt, lower inflation, or a stronger primary balance. The bootstrap p-values on those three outcomes all exceed 0.75 regardless of programme type. On debt in particular, the programme variable is statistically silent once common year effects are accounted for.

Figure 1. Ghana’s macro indicators, 2000–2024. Dashed lines mark the 2015 and 2023 ECF approvals. Sources: World Bank WDI; IMF WEO. Author’s calculations.

Ghana’s own debt history shows why, and it is worth dwelling on. Public debt fell steeply in the mid-2000s — the era of the Heavily Indebted Poor Countries initiative and multilateral debt relief. It climbed through the 2010s. It spiked to roughly 93 percent of GDP in the 2022 crisis. It has since fallen back under the 2023 ECF. A formal structural-break test applied to Ghana’s debt series — a test that is allowed to place the break wherever the data prefer — does not select a programme year at all. It selects 2004: the debt-relief moment. A parallel test on inflation finds no statistically significant break (p = 0.67); the 2022–23 surge sits within Ghana’s long-run volatility rather than marking a regime change.

The lesson is uncomfortable for tidy narratives in either direction. Ghana’s fiscal trajectory has been shaped at least as powerfully by global and regional forces — debt-relief waves, commodity cycles, the pandemic, the 2022 surge in global interest rates — as by the presence or absence of a Fund programme. An IMF arrangement is one actor in that story. It is not the author of it.

The hero finding: the PCI must be judged by the right test

If one result from this analysis deserves to anchor the public debate about Ghana’s new arrangement, it is this.

The non-financing category of IMF engagement — the category that contains the Policy Coordination Instrument — carries no detectable growth dividend in two and a half decades of regional data. None.

It is essential to read that finding correctly, because it is easy to read it wrongly. It is not a criticism of the PCI. It is not evidence that Ghana made a poor choice. It is a statement about what the instrument is for. A PCI is not engineered to move the real economy; it has no resources with which to do so. It is engineered to anchor credibility. To evaluate it by whether growth accelerates would be to evaluate a thermometer by whether it warms the room.

The error to guard against, then, is one of expectations. If commentary over the coming year treats the PCI as a quasi-stimulus and waits for a growth pay-off, it will be disappointed — and the disappointment will be an artefact of the wrong yardstick, not of the instrument’s failure. The honest framing is narrower and more useful: the PCI is an external anchor for fiscal discipline, adopted without the repayment burden or the political stigma of a loan. That is a coherent thing to want. It is simply not the same thing as a growth programme.

How the PCI should actually be judged

If growth is the wrong test, what is the right one?

The PCI’s theory of action runs through the price of risk. Its success or failure will therefore show up in the credibility metrics, not the real-economy ones. Four are worth watching over the next twelve to eighteen months.

The first is sovereign spreads — the premium investors demand to hold Ghanaian debt over a benchmark. If the IMF’s continued endorsement is doing its job, that premium should compress, or at least hold firm against regional peers.

The second is the cost of new borrowing — the yield Ghana must offer at its next Eurobond issuance, whenever it returns to the market. A credibility instrument that is working lowers that yield.

The third is the rating agencies. Ghana’s path back from restricted default has already drawn upgrades. Whether the agencies cite the PCI as a stabilising factor in future commentary is a direct test of the signalling channel.

The fourth is partner financing — whether bilateral and multilateral development partners treat the IMF’s PCI endorsement as a green light, as the instrument is explicitly designed to encourage.

If those four move favourably, the PCI will have earned its keep. If they do not, no amount of favourable framing will substitute — and that, not the growth rate, is the scoreboard to watch.

What the sceptics get right — and wrong

A serious assessment should concede what critics of the PCI choice get right.

They are right that a non-financing arrangement has no enforcement teeth beyond reputation. An ECF can withhold a tranche; a PCI can withhold only its approval. They are right that the catalytic effect on private and partner financing — the whole premise of the instrument — is a hypothesis, not a guarantee, and that it may not materialise. They are right that Ghana has a long history of repeated IMF programmes, and that markets may therefore discount any single signal. And they are right that real risks remain unresolved: energy-sector arrears, the financial position of COCOBOD and the cocoa sector, contingent liabilities from state-owned enterprises, and gaps in the anti-corruption architecture.

But the sceptics are wrong if they conclude that the PCI is therefore an empty gesture. The instrument is not designed to fix those problems by force; it is designed to keep an external, credible monitor in the room while the government fixes them itself. The relevant question is not whether the PCI has teeth — it does not, by design — but whether the government has the resolve the PCI is meant to certify. That is a question about Ghana, not about the instrument.

The question to hold in mind

Ghana has exited an IMF financing programme with inflation well down from its 2022–23 peak, debt falling as a share of GDP, and the symbolism of graduation firmly in hand. It has chosen, sensibly, to keep the Fund engaged through a Policy Coordination Instrument rather than walk away entirely.

The evidence in this analysis supports a precise reading of that choice. Financing arrangements, across the region and across a quarter-century, track modestly higher growth. Non-financing arrangements — the PCI’s own family — do not. Ghana has not acquired a growth engine. It has acquired an anchor: an external discipline on fiscal policy, retained voluntarily, at no fiscal cost and without the stigma of a bailout.

Whether that anchor holds depends on something no instrument can supply. A PCI keeps the IMF in the room. It does not keep the resolve. The discipline that a financing programme imposes from outside, the PCI merely observes; the country must now generate it from within — on energy-sector arrears, on state-enterprise governance, on the revenue base, on the slow institutional work that does not make headlines.

The question Ghana should be debating, then, is not whether the PCI was the right instrument. On the evidence, it was a reasonable one. The question is whether the government will sustain, unprompted by the threat of a withheld tranche, the reforms that three years of a financing programme were required to enforce.

That is the open question of Ghana’s post-programme decade. The instrument has been chosen. The resolve has not yet been demonstrated.

Dr Stephen Lartey is an economist with a PhD in Economics from the University of Sussex, UK, specialising in institutions, fiscal policy, monetary policy, macroeconomic policy, and causal inference.

Notes on the data and method. The empirical analysis draws on a balanced panel of eight Sub-Saharan African economies, 2000–2024 (200 country-year observations), combining World Bank World Development Indicators, the IMF World Economic Outlook database (April 2026), and the IMF Monitoring of Fund Arrangements (MONA) database. IMF arrangements are classified as “financing” (PRGF, ECF, EFF, SCF, SBA) or “non-financing” (PSI, PCI). Estimates are from two-way fixed-effects panel regressions with country-clustered standard errors and wild-cluster bootstrap inference appropriate to the small number of clusters. Key estimates: financing arrangements are associated with a 0.69-percentage-point increase in real GDP growth (bootstrap p = 0.035); the corresponding coefficient for non-financing arrangements is 0.03 (bootstrap p = 0.96). All reported relationships are associational, not causal; IMF programme participation is not randomly assigned. The full dataset, code and methodology are available from the author on request.

Post Views: 95