WHO MPOWER country scores: the rubric, the records, and the joins that make them speak

MPOWER is WHO's scoring rubric for the six demand-reduction measures of the Framework Convention on Tobacco Control, and the acronym doubles as the checklist. Monitor tobacco use and prevention policies. Protect people from tobacco smoke. Offer help to quit tobacco use. Warn about the dangers of tobacco. Enforce bans on tobacco advertising, promotion and sponsorship. Raise taxes on tobacco.

The scoring convention is uniform across countries and years: each measure receives an achievement level from 1 to 4 per member state, or a no-data mark, giving a country-year grid of six ordinal policy variables. The Report on the Global Tobacco Epidemic 2023 is the ninth edition of the series that has tracked this since 2008, published 31 July 2023, and it pairs the ratings with per-country profiles. In Datadory's catalog the record behind it scores 9 out of 10 on quality against a catalog-wide average of 7.81.

Two records hold the grades, and they share one spine.

WHO Report on the Global Tobacco Epidemic 2023 is the anchor: the ninth-edition scorecard joined to the structured country-year policy table behind it. Achievement levels 1-4 per measure, smoking-prevalence series with uncertainty bounds attached, regional aggregates across roughly 194 member states, observations reaching back to roughly 2007. Roughly 369,000 observations ride the same structured package, which means the policy grid arrives beside the Monitor-cycle prevalence estimates instead of in a separate appendix.

WHO Global Health Observatory Legacy Data Repository - Tobacco carries the same measure vocabulary inside a wider frame - roughly 300 tobacco indicators across 194+ member states: modeled prevalence, smokeless and e-cigarette use, the five non-Monitor MPOWER components, tax revenue, brand prices, affordability. Quality score 8 out of 10.

Both resolve to the same four-key identity - indicator code, ISO-3 country, year, dimension - so a pull can span the two without a reconciliation pass, and small markets commercial panels skip sit next to China and the United States under one code set.

One row equals indicator x country x year x dimension, whether the cell holds a policy grade or a prevalence estimate:

# policy row - integer achievement level, 1-4
IndicatorCode   : TOBACCO_MPOWER_OVERVIEW
SpatialDim      : SDN          # Sudan, ISO-3
ParentLocation  : Eastern Mediterranean   # WHO region
TimeDim         : 2022
Dim1            : <measure group, e.g. TOBACCO_INDICATOR_P_Group>
Value           : 4            # highest level on that measure

# baseline-era cell from the same schema
SpatialDim : FJI   TimeDim : 2008   Value : 3

# Monitor-cycle prevalence rides the identical keys
IndicatorCode : TOBACCO_MPOWER_M1_TOBACCOUSESMOKING
SpatialDim    : MAR   TimeDim : 2024   Value : National

Three things worth reading out of those rows. First, one schema carries both halves of the story - Sudan at level 4 in 2022 and Fiji at level 3 in 2008 stack into a fifteen-year legislative timeline without any reshaping. Second, prevalence rows carry their survey scope instead of a score, which is why the disaggregating dimension exists as its own field rather than being flattened into the value column. Third, country, region and year travel with every row, so the panel pivots straight into country-year form. Your sample returns this exact schema cut to the countries, measures and years you name - the schema in the sample is the schema that ships.

The analysis survives any delivery channel because it follows the grain of the data:

1. Work at measure grain - one observation per country-year per MPOWER measure - and pivot the six into columns M, P, O, W, E and R.
2. Filter to the latest reporting year per country so each member state contributes exactly one row per measure; mixing vintages turns a snapshot into mush.
3. Keep the no-data marks instead of coercing them to zero - an absent rating is not a low rating, and treating it as one corrupts any regulatory-intensity index built on top.
4. Append outcome columns from the World Bank prevalence panel: indicator SH.PRV.SMOK returns about 6,625 rows covering roughly 265 economies from 2000 to 2024, with SH.PRV.SMOK.MA and SH.PRV.SMOK.FE adding around 5,000-row male and female splits on identical keys.
5. Because both sides originate in WHO's harmonized estimates, the country codes reconcile directly - no fuzzy-matching pass.

The result is one table where a tax-rating digit sits next to the adult tobacco-use percentage it nominally moves. Delivered through Datadory, that table ships already pivoted, with vintage stamped per cell, so step two becomes a filter rather than a forensic exercise.

Policy scores explain nothing in isolation; their value comes from joins, and the tobacco slice makes three of them cheap.

• Prevalence outcomes. The World Bank SH.PRV.SMOK family gives annual current tobacco use for roughly 265 economies, sex-split on shared keys, so a before-and-after around a country's rating change becomes a one-line groupby.
• Health burden. The IHME GBD Results Tool - Tobacco/Smoking query interface supplies deaths, YLLs, YLDs and DALYs for the smoking and secondhand-smoke risk factors across 204 countries plus 660 subnational locations, every year from 1990 to 2023, each estimate carrying its 95% uncertainty interval. It is the only tobacco record in the catalog that models below the national border.
• European depth. Eurostat's daily smokers of cigarettes indicator tracks the share of the population aged 15+ smoking combustible products, by sex, across 35 European entities over seven survey waves from 2006 to 2023 - the natural overlay when your country set is EU-27 plus EFTA and the UK.
• Quick context. Our World in Data - Smoking curates roughly 38 long-run indicators - prevalence, cigarette sales, smoking-attributable deaths - useful for communication artifacts while the MPOWER grades themselves stay attributed to WHO.
• Directionality check. US TTB Tobacco Statistics logs monthly removals, production and inventories from calendar 2012 through the July 2026 posting. It measures supply, not regulation - keeping it in the same workbook as MPOWER clarifies the difference instead of blurring it.

Name the companions when requesting a sample and they arrive joined, not bolted on afterward.

Five records cover most policy-score work. What separates them is geography, definition and grain rather than brand - the columns below decide which record anchors a project. As cataloged by Datadory, August 2026:

Four limitations determine what you can honestly claim.

Ordinal, not cardinal. A rating of 4 is not twice 2; the digits rank policy intensity within a measure. Summing the six into a composite is defensible as a rough index, not as a measured quantity.

Scope boundaries. The underlying code list carries 234 country entries against roughly 194 member states, while the World Bank prevalence panel covers about 265 economies and aggregates. Decide explicitly how regional and income aggregates map before any cross-source average, or drop them.

Grain is coarse by design. One grade per country-year-measure shows the slope of law, not the weather of enforcement. A jurisdiction can hold the top warning label while its street-level compliance drifts; nothing in the rubric captures that, and pretending otherwise produces confident nonsense.

Public-health and excise analysts benchmark national control regimes against peers and track whether a tax-grade change precedes a prevalence move - the comparison the six-column pivot exists for.

ESG and regulatory-risk researchers score exposure for insurers and consumer brands operating in stricter or looser jurisdictions, using the grades as a comparable policy input rather than a hand-built index.

Actuaries and health economists read the grades as leading context for the modeled burden curves - policy effort today, mortality loading decades out.

Market-entry strategists in nicotine alternatives, cessation and pharma treat the cessation-support and advertising-ban measures as demand-shaping variables per market.

Journalists and academics cite the grades because the rubric is uniform across countries and years - one definition, applied everywhere, since 2008.

Files, feeds, or your warehouse. Daily, weekly, or hourly - your call.

Open the tobacco data hub for the complete scored index of the industry slice, or the ranked best tobacco datasets shortlist for the top of it. Smoking prevalence by country walks the prevalence panels record by record, GBD smoking attributable deaths covers the burden layer, and state tobacco control laws dataset goes inside the American subnational equivalent of the MPOWER grades. For the supply-side counterpart, US cigarette production statistics and cigarette consumption per capita cover the volumes these policies act on.