How we score Nigeria's governments

GPLab's scorecards cover every Nigerian government from Tafawa Balewa (1960) to Bola Tinubu. Each card answers one question: what happened to the economy on this government's watch, and how much did its own choices contribute? This page explains exactly how we arrive at the number, so you can check our work, disagree with it, or rebuild it yourself.

Short version. Half of each score comes from five economic indicators: inflation, growth, the naira, public debt and reserves. They are measured with World Bank and IMF data in the same way for every government. The other half comes from two judgement-based areas: policy credibility and institutional reforms. Two reviewers score these against a fixed checklist, and every point has to be backed by a source. Oil prices are shown next to every score and are not hidden inside it. We take a market-leaning view, we say so openly, and every card links to the strongest counter-argument.

1. Purpose

• To give readers, especially young Nigerians, a fair, consistent and checkable way to compare economic management across 66 years, military and civilian.
• To separate outcomes (what the numbers did) from choices (what the government decided) and luck (mainly oil prices).
• It is not a verdict on a leader's character, legitimacy or overall legacy, and it does not score democracy or human rights (see §5).

2. What we measure

Weights: approved 2 October 2026. Economic outcomes 50% (10% for each of the five indicators), policy credibility 25%, institutional reforms 25%. They will not change without a dated changelog entry (§13).

Pillar A: Economic outcomes (50% of the score; 10% per indicator)

Scores are on a straight line between the two thresholds and capped at 0 and 100. The thresholds are our design choices. They are written down here and in method.json, so anyone can change them and re-run the numbers.

Pillar B: Policy credibility (25%)

Two reviewers each answer four yes/no questions independently (see §10). Every answer must cite a source.

1. Did the government avoid financing deficits through the central bank beyond the legal limit (for example, Ways & Means overdrafts)?
2. Did it avoid multiple official exchange rates, or a large and persistent parallel-market premium?
3. Did its major announced reforms survive for at least 12 months without reversal?
4. Were budgets broadly implemented as passed?

Each "yes" is worth 1 point, so the scale runs 0 to 4, giving 0, 25, 50, 75 or 100. Disputed answers score half a point and the dispute is explained on the card.

Pillar C: Institutional reforms (25%)

This counts durable reforms that strengthened markets or the state's economic institutions. Examples are a debt restructuring with lasting rules, bank-capital regimes, tax-administration laws, sector liberalisation and independent regulators. To count, a reform must still be in force five years later (or be clearly durable, for current governments), and each must be listed with a dated source. 0 = none; 4 = four or more. Reforms that were later reversed do not count.

Why these weights? We give equal weight to the five outcome indicators, following common index practice. The Heritage Index of Economic Freedom weights its 12 components equally across four pillars. The Ibrahim Index of African Governance uses equal-weight averages after rescaling indicators to 0-100. The Fraser Institute's Economic Freedom of the World uses five areas on a 0-10 scale, one of them "Sound Money". Our credibility and reform checklists borrow from Freedom House, which has analysts score 25 indicators from 0 to 4 against published questions, with review. Outcomes get 50% because they are what citizens live with. Choices get the other 50% because outcomes alone would reward luck.

v1 scoring guidance for Pillars B and C (Draft, open for public review)

These rules were written for the first full scoring round (scores v1, 2 October 2026). The review panel (§10) has not yet been appointed, so the v1 credibility and reform scores are GPLab staff drafts, labelled "Draft, open for public review". The panel will re-score them blind against the same published evidence lists.

• Question 2, "large and persistent": the parallel-market premium averaged more than 20% a year in at least half of the scored years for which data exist. Two or more official rates at the same time (for example ₦22 and the AFEM rate in 1995-98) count as "no".
• No usable evidence: if no source could be found for a question, it scores 0.5 under the disputed rule and the card is flagged "Evidence gap". We do not guess a yes or a no.
• Ownership and participation mandates (indigenisation 1972, Land Use 1978, oil-and-gas local content 2010) are listed on the card but not counted as reforms under our lens (§7). The panel may decide otherwise, and the card says so.
• Current government: a reform counts as "clearly durable" if it has survived at least 12 months and is still in force. The count is interim.
• Every point and every counted reform cites a dated source. Sources we saw only in a search-engine summary, without opening the page, are marked as such.

3. Which years count

• A calendar year is counted for the government that was in office for at least 183 days of it. Annual data cannot be split by month.
• One rule for short governments. A government with fewer than two scored years gets no number of any kind: no outcome, credibility, reforms or composite score. It gets a factual card marked "Too short to score", and no other label is used. This applies to Aguiyi-Ironsi (1966: 195 days, one scored year), Shonekan (1993: 83 days, none) and Abubakar (June 1998 to May 1999: one scored year, 1998). The threshold is fixed by the 183-day rule, so no separate month cut-off is needed. It is roughly equivalent to "under about 18 months in office".
• Borderline cases, applying the same rule. Murtala Mohammed (July 1975 to February 1976) held no scored year on his own: 1975 goes to Gowon, who held 210 days of it. Our timeline merges him with Obasanjo (1976-79), so the merged government is scored on 1976-79 and the card says Murtala alone would be "Too short to score". Yar'Adua (May 2007 to May 2010) has three scored years (2007-09) and is scored. 2010 goes to Jonathan.
• v1 allocation: Balewa 1961-65; Gowon 1967-75; Murtala/Obasanjo 1976-79; Shagari 1980-83; Buhari 1984-85; Babangida 1986-93; Abacha 1994-97; Obasanjo 1999-2006; Yar'Adua 2007-09; Jonathan 2010-14; Buhari 2015-22; Tinubu 2023-25 (interim).
• Short versus long tenures. Averages make a 2-year and an 8-year government comparable. But a short tenure carries more noise, so cards show the number of scored years next to the score. Changes (naira, debt) are annualised.
• Inheritance. The naira and debt are measured from the first scored year, so a predecessor's last-minute moves are not charged to the successor. The cost is that a government's own first-year moves are partly missed. We say so on the card where it matters.
• Current government. The current government is shown as an interim score, using completed calendar years only, and is re-scored every year.

4. Oil and other shocks

Nigeria's fortunes rise and fall with oil prices that no government controls. We do not adjust the score, because any adjustment formula would itself be contestable. Instead each card shows an Oil context badge:

• Tailwind: the average crude price over the scored years was more than 20% above the average of the previous five years (Energy Institute data via Our World in Data).
• Headwind: more than 20% below.
• Neutral: anything in between.

Other big shocks, such as the 2008-09 global crisis or the 2020 pandemic, are noted on the card with a source. This lets readers ask the fair question: given the hand it was dealt, how did this government do?

5. Military and civilian governments

We use the same economic yardstick for both. The scorecard does not score political rights. A military government with low inflation has still taken power unconstitutionally, and we say so plainly on every card ("Type: military"). For rights and freedoms, we link to dedicated measures such as Freedom House. We never trade one off against the other.

6. Data limits (please read)

• Before 1980. World Bank series for inflation, growth and the official exchange rate go back to 1960-61. But the official exchange rate was set by the government and barely moved for most of that period (WDI shows the same rate every year from 1960 to 1970), so "naira stability" tells you little. The IMF's general-government debt series starts only in 1990. Where it does not cover a government's first scored year, we use external debt as a share of GDP at both ends instead. This is calculated from WDI (DT.DOD.DECT.CD ÷ NY.GDP.MKTP.CD, available from 1970) and labelled "not comparable". We never splice the two series. Before 1970 there is no debt series, so debt is missing (Balewa, Gowon). Reserves in months of imports start in 1977. Early-period scores therefore carry a "thin data" flag, and pre-1977 reserve scores are marked partial.
• GDP rebasing. In 2025 the National Bureau of Statistics rebased GDP to 2019 prices, so its growth rates differ from the World Bank's for recent years (for example, 2021: WDI 1.11% vs NBS 0.95%). We score with WDI for consistency over time and show the NBS figure alongside.
• CPI rebasing. NBS rebased the CPI in early 2025, so 2025 inflation is not strictly comparable with 2024.
• Parallel exchange rates. The CBN publishes an annual premium series for 1980-2002 (CBN Bullion, 2006). After 2002 no consistent annual series exists, so we use dated quotes from named outlets and CBN reports. Before 1977 we found no parallel-rate data.
• Missing data. If an indicator is missing, the other outcome weights are scaled up proportionally and the card says "Partial data". We never fill gaps with guesses.

7. Our lens: GPLab is market-leaning

We think it is more honest to state our view than to claim to have none. GPLab takes a market-leaning lens. In practice that means we generally regard these as good economic management:

• low and stable inflation;
• a single, market-determined exchange rate;
• no central-bank financing of deficits;
• prices (including petrol) that are not held below cost by open-ended subsidies;
• debt used for investment rather than consumption;
• rules and institutions that outlast the government that made them.

What this means for the scores. Several indicators (one exchange rate, no monetary financing, durable market reforms) reflect that view. Economists from statist, developmental or left traditions would weight things differently. They might give more credit to public investment, industrial policy, job creation or protection of the poor during reforms. So:

• every scorecard and chapter carries a "Lens / the other side" box with the strongest fair counter-argument, from named scholars and with sources;
• our weights are public, and you can re-weight them yourself (method.json);
• we report poverty, protest and social-cost evidence on the cards even where it doesn't enter the score.

8. Non-partisanship commitments

1. Same rules for everyone. Every government (military or civilian, northern or southern, any party) is scored with the same indicators, thresholds and data vintage.
2. Funding disclosed. See §9. GPLab is currently self-funded by its founders. We take no money from political parties, candidates or government bodies, and any future funders will be listed here.
3. Sources for everything. Every number and every judgement point links to its source. Quotes are verbatim.
4. Two reviewers with differing views score the judgement pillars independently, and disagreements are disclosed (§10).
5. Pre-commitment. The weights and thresholds on this page are fixed before each annual update, not tuned to produce a result. (The current weights were approved on 2 October 2026.)
6. Elections. In the 60 days before a federal election (next: 16 January 2027, per INEC as reported by ThisDay) we publish no new scores for the incumbent or candidates. We only fix errors.

9. Funding and independence

10. Review: open public review now, expert panel next

Now: open public review. Policy credibility and institutional reform scores are labelled "Draft, open for public review". Anyone can challenge any part of a score, including the economic pillar, using the "Challenge this score" form on each government scorecard. Each challenge needs a reason and a source. We judge challenges on evidence, not opinion, and publish every decision, accepted or rejected, with our reasons in the public review log. The review window runs for 30 days, from 12 October to 11 November 2026.

Next: an independent expert panel. The two-person panel described below has not been appointed yet. It will follow the public review phase.

The macro pillar is mechanical: anyone can recompute it from method.json. The two judgement pillars (B: policy credibility; C: institutional reforms) are scored by a two-person review panel.

Who sits on it.

• Two external reviewers with differing economic views: for example, one market-leaning and one heterodox, developmental or state-led. The aim is that a score both can accept is not just GPLab's lens repeated.
• Each reviewer serves for one annual update cycle at a time, renewable. If a reviewer steps down or is recused, a substitute from the same broad tradition is used.

What they do.

• Before scoring starts, GPLab publishes the evidence list for each government: the dated, sourced key actions (from the card's sources) relevant to the four credibility questions and to counted reforms.
• Each reviewer scores Pillar B (the four questions, 0-4) and Pillar C (count of durable reforms, 0-4) for every government independently, against that published evidence list, without seeing the other's scores.
• Every point must cite an item on the list. A reviewer may propose additional evidence, but it must be dated and sourced, and it is added to the public list for both reviewers before scores are final.
• Reviewers may check the macro pillar for errors but do not score it.

Conflicts of interest.

• No current roles: reviewers must not hold any current party, campaign or government role. That includes elected or appointed office, federal or state government posts, the CBN or other regulators, and government advisory councils or committees. A reviewer who takes such a role steps down from the panel.
• Disclose past roles: any past government, party, campaign or official advisory role, plus relevant board seats, paid consultancies and media roles, is declared in writing and published in the reviewer's bio.
• Recusal: a reviewer does not score any government they served or formally advised, including a foundation or office of that former leader. A substitute reviewer from the same tradition scores that government instead, and the card says so.
• Declarations are renewed at the start of each annual cycle.

When the reviewers disagree.

• Scores are compared for each pillar for each government, on the 0-4 scale.
• Gap of 1 point or less: the composite uses the average of the two scores. Both individual scores are shown in the card's notes.
• Gap of more than 1 point: this triggers a reconciliation note. Each reviewer sets out in writing which evidence items drove their score. Each may revise once, and only by citing evidence. The note is published with the card.
• Still more than 1 point apart after reconciliation: both scores are published side by side, the composite uses their average, and the card is labelled Panel split. GPLab does not cast a deciding vote.

Publication. Reviewers' names and short bios (about 100 words each) will be published on this page, once the panel is appointed, with their declared interests and any recusals. Each card lists who scored it and the date. Panel not yet appointed

11. Corrections and updates

Spotted an error? Email [email protected].

• Annual update: each August, after the World Bank's July WDI refresh (the current vintage is dated 13 July 2026) and the April/October IMF WEO. Version numbers and a changelog are published.
• Off-cycle updates: after a major statistical revision (such as a GDP or CPI rebasing) or a confirmed error.

12. FAQ

Does a high score mean a good president or head of state? No. It means the economy did comparatively well on these measures during their years, given the oil context shown. Character, rights, security and legitimacy are outside this scorecard.

Why are military rulers scored at all? Because they ran the economy for 29 of Nigeria's first 39 years after independence, and leaving them out would hide half the story. Their cards say clearly that they took power unconstitutionally.

Isn't this biased towards free-market policies? Partly, yes, and we say so (§7). That is why every card shows the strongest opposing view and the weights are public. If you weight credibility and reforms at zero, you get a pure outcomes score.

Why don't you adjust for oil prices? Any formula would build in its own assumptions. We show the oil context next to the score so you can judge for yourself.

Why do your numbers differ from NBS or the CBN? We use the World Bank and IMF series for consistency over 66 years. National figures are shown alongside where they differ, especially after the 2025 GDP and CPI rebasings.

Why does a government with a fixed exchange rate score well on "naira stability"? On that indicator alone it can. A rate held below the market by rationing is caught by the credibility pillar (multiple rates or a large parallel premium) and by the parallel-premium figure shown next to the score.

Can a score change later? Yes, when data are revised or we fix an error. Every change is logged and dated.

Who pays for GPLab? See §9. GPLab is currently self-funded by its founders. We take no money from political parties, candidates or government bodies, and any future funders will be listed here.

Who are the reviewers? Their names, bios and declared interests will be published in §10 once the panel is appointed. To report a conflict of interest or an error, email [email protected].

Can I reuse the scores? Yes, with credit to GPLab and a link to this page. The underlying World Bank and IMF data have their own terms.

Sources

• World Bank, World Development Indicators API (FP.CPI.TOTL.ZG; NY.GDP.MKTP.KD.ZG; PA.NUS.FCRF; FI.RES.TOTL.MO; FI.RES.TOTL.CD; DT.DOD.DECT.CD; NY.GDP.MKTP.CD), last updated 13 Jul 2026, accessed 1 Oct 2026 and re-pulled for 1960-2025 on 2 Oct 2026. api.worldbank.org/v2/country/NGA/indicator/
• CBN Bullion Vol. 30 No. 3 (Jul-Sep 2006), Table 1, official vs parallel rates and premium 1973-2002 (used for credibility question 2). cbn.gov.ng/OUT/PUBLICATIONS/BULLION/GOV/2007/BULL%20JUL-SEPT…
• IMF World Economic Outlook, general government gross debt (GGXWDG_NGDP), via the IMF DataMapper API, accessed 1 Oct 2026. imf.org/external/datamapper/api/v1/GGXWDG_NGDP/NGA
• Energy Institute, Statistical Review of World Energy 2026, via Our World in Data (crude oil prices). ourworldindata.org/grapher/crude-oil-prices
• World Bank, Nigeria: Structural Adjustment Program, Report 13053-UNI, 13 May 1994 (parallel exchange rates 1980-92). documents1.worldbank.org/curated/en/959091468775569769/txt/m…
• Premium Times, 31 Aug 2026 (NBS rebased GDP growth). premiumtimesng.com/business/business-news/906587-nigerias-ec…
• ThisDay, 26 Feb 2026, "INEC fixes January 16 for presidential, NASS elections". thisdaylive.com/2026/02/26/breaking-inec-fixes-january-16-fo…
• Heritage Foundation, 2026 Index of Economic Freedom methodology. static.heritage.org/index/pdf/2026/2026_indexofeconomicfreed…
• Fraser Institute, Economic Freedom of the World: approach. efotw.org/economic-freedom/approach
• Mo Ibrahim Foundation, 2024 IIAG methodology. assets.iiag.online/2024/2024-IIAG-methodology_EN.pdf
• Freedom House, Freedom in the World 2026 methodology. freedomhouse.org/sites/default/files/2026-03/FIW26%20Methodo… ; FAQ: freedomhouse.org/reports/freedom-world/faq-freedom-world

13. Changelog

• 2 Oct 2026: Weights approved.
• 2 Oct 2026: Scores v1 published for all 15 governments (12 scored, Tinubu's interim; 3 too short to score). Credibility, reforms and composite scores are drafts, open for public review.