IFASTAT — International Fertilizer Association Statistics

It is the statistical self-portrait of the global fertilizer industry, compiled by the International Fertilizer Association from its member base of producers, traders and distributors. Seven indicators - capacity, production, imports, exports, home deliveries, apparent consumption and fertilizer consumption - cut across a product tree that starts at ammonia and sulfur, runs through phosphoric acid and phosphate rock, and lands on the finished grades: urea, MAP/DAP, MOP and the rest of the N, P and K range. The same products serve plant nutrition and industrial applications, and the set keeps both uses in frame.

Two clocks govern the history. Supply-side and trade series - capacity, production, imports, exports, apparent consumption - span roughly 25 years from 2000, with capacity extending forward-looking into 2026. The consumption panel reaches back to 1961, nearly seven decades of nutrient uptake per country. Around the core sit a reports-and-analysis library (production and trade, market outlook, market updates, capacity, specialty fertilizers) and visualization dashboards - Fertilizer at-a-glance, Fertilizer in-depth, a crop calendar. Coverage varies by indicator, product and geography, a consequence of how the underlying markets report.

Get a sample of this dataset and inspect real rows from the indicators, products and countries you name before you commit.

These four were parsed straight from the public fertilizer-consumption workbook during catalog research, chosen because they sit on opposite edges of the panel. Afghanistan enters the series in 1964 at 0.10 thousand tonnes - of nitrogen and again of phosphate - the kind of near-zero baseline that makes a six-decade growth curve legible. Zimbabwe in 2024 reads 90.11 thousand tonnes of N and 44.78 of K2O: a mid-size consumer measured on both macronutrients in the same format, six decades apart.

The shape matters as much as the values. Every nutrient gets its own row, so one country-year fans out into up to three records, and the Unit column switches with the nutrient - '000 tonnes N, P2O5 or K2O - meaning a naive concatenation of Values quietly adds nitrogen tonnage to potash tonnage. The Granularity column flags each row's clock, and the alpha-3 code travels on every row so joins need no country-name cleaning.

A sample comes back in exactly this shape, filtered to the nutrients, countries and years you specify.

Nine documented columns define a consumption observation, and every one is verified against the parsed workbook rather than guessed from a header row. Three design details matter when you model on top of them. Long-format nutrient triplets: the panel is keyed nutrient-by-nutrient rather than wide, so aggregating to a country total means grouping by country-year and respecting the unit switch per row - N, P2O5 and K2O tonnages are different substances and never sum. Campaign years masquerade as years: several countries report a fertilizer campaign (July-June) rather than the calendar year, so a naive calendar join shifts part of the panel by six months against harvest or trade data. Codes ride along: the ISO alpha-3 column removes country-name ambiguity entirely, and Granularity marks each row's frequency explicitly.

Extended extracts can carry the query dimensions listed under additional fields on request - the other six indicators, the full product tree and the quarterly or half-yearly grain where it exists.

Geography - global, at country level across the major producing and consuming countries. The consumption panel alone carries 133 countries grouped into ten IFA statistical regions, from West Europe through East Asia to Oceania. Recent sub-annual readings surface first as regional aggregates before country resolution fills in, so the newest quarter is a regional story and the newest complete country picture is always the annual one.

Temporal - capacity, production, trade and apparent consumption run from 2000, roughly 25 years, with capacity carrying forward-looking vintages out to 2026. Consumption runs from 1961, almost 65 years. One structural wrinkle worth planning around: annual country figures arrive on a delayed clock - the association holds them roughly nine months after period end to comply with antitrust guidelines - which is precisely why the freshest readings appear as aggregates first.

Granularity - annual by default, with quarterly and half-yearly series for selected indicators, and consumption reference periods mixing calendar years and fertilizer campaign years by country. There is no company, plant or facility dimension anywhere in the set; this is deliberately the macro layer of the fertilizer economy. For European facility-level pollutant detail to pair against it, the E-PRTR portal record covers that job.

API, files, or your warehouse. Daily, weekly, or hourly.

The underlying series move on the association's own publishing rhythm - annual country figures, faster aggregates for recent periods - but your delivery cadence is independent of that. Load the full panel once into your warehouse for model features, pull structured payloads when a dashboard needs the latest consumption reading, or take bulk files for one-off market studies. Every delivery ships the field dictionary above unchanged across all three channels.

• Capacity-cycle modeling - capacity beside production beside apparent consumption yields utilization and investment-cycle signals for the nitrogen and phosphate chains that no price chart supplies; the forward-looking capacity vintages to 2026 make the next supply wave visible before it lands.
• Trade-flow mapping - imports and exports by country and product expose who ships urea, ammonia, DAP and MOP where, the exogenous feature set behind freight positioning, export-margin models and sanction-shift analysis.
• Demand estimation - apparent consumption and home deliveries bridge the gap between production and trade when reported deliveries are thin, giving a defensible demand proxy for sizing regional markets.
• Food security and ag-input research - the 65-year consumption panel per nutrient supports long-run yield-response and policy work that younger datasets simply cannot anchor.
• Industrial-demand tracking - products in this set serve both farms and factories, so the same ammonia and urea series inform industrial gas and plastics-chain analyses, not just agriculture.

Investors and quant researchers get utilization-ready capacity and production pairs for the world's nitrogen, phosphate and potash chains, with trade flows to test any thesis about regional dislocation; see investors quants use cases. Market researchers and consultants get a pre-segmented global market - seven indicators, dozens of products, country detail - without reverse-engineering a customs nomenclature; see market researchers use cases. Data scientists get tidy long-format series with ISO codes attached, joining cleanly onto FAOSTAT, World Bank and trade panels; see data scientists use cases. Competitive-intel and product teams track capacity additions and consumption growth by country as a leading read on where fertilizer demand - and demand for everything upstream of it - moves next; see competitive intel product teams use cases. Journalists, academics and students cite figures from the industry's own statistical body, with a 1961 baseline that lets a claim about fertilizer use be sourced decade by decade; see journalists academics use cases. Sales and growth teams read consumption growth by country to time entry into emerging ag-input accounts; see sales growth teams use cases.

Three things worth knowing upfront. First, this is a macro layer, not a company register - no firm, plant or facility dimension exists anywhere in the set, so portfolio-exposure questions need pairing with a facility-level source rather than deeper filtering here. Second, the nutrient triplet is long by design: each country-year appears up to three times, once per nutrient, and the unit switches per row, so ingestion should group and convert rather than concatenate. Third, years are not always years - a meaningful share of countries report fertilizer campaign years (July-June), so aligning this panel against calendar-keyed trade or harvest data needs an explicit convention, not an implicit join.

Name the indicators, products, countries and year ranges you need when you request a sample - the extract comes back cut to that shape, with the field dictionary above intact.