NIESR Monthly GDP Tracker

The NIESR Monthly GDP Tracker is a nowcast, an early estimate before official data, of UK GDP (total output value) growth. Its publisher NIESR (National Institute of Economic and Social Research) is an independent London research group, distinct from the ONS (Office for National Statistics). It covers the whole UK economy on the output, production-side, approach and reports chained-volume indices, inflation-adjusted output levels, plus growth rates. It fields no survey and builds 10 modelled parts from ONS data, past trends, surveys and judgement, so it is a forecast, not official GDP.

Each edition appears once a month, normally in the second week, after ONS data and surveys arrive, with no fixed clock time. The date shifts with the ONS calendar, and each edition gives a quote, main points, setting, sectors, charts, tables, accuracy review, health warning and contacts. Every quarter gets 4 estimates that swap forecasts for ONS data as hard numbers replace predicted months. Readers focus on 3-month growth, change across two 3-month windows, as GBP-positive when high and GBP-negative when low.

The Tracker adds 10 output parts, grouped below by share, so services swings move the headline, the main number, most.

• Business services and finance, 338: professional, financial, IT and admin services, the largest driver.
• Government and other services, 220: public administration, health, education and other services.
• Distribution, Hotels and Restaurants, 134: retail and wholesale trade plus hospitality and consumer-facing services.
• Transport, Storage and Communications, 103: logistics, storage and communications activity.
• Production, 135: manufacturing plus utilities, extraction of oil and gas, and water and waste services, a volatile group.
• Construction and Agriculture, 62 and 7: building work plus farming, negligible for the headline.

In one episode the Tracker pointed to -22.8% against an official first estimate of -20.4%, an error of 2.4 points. It remains the largest miss on record, with average absolute error, mean size of misses, at 0.22 points. In another episode it had pointed to growth of 0.2% against an official first estimate of -2.0%, missing the sudden stop. Together they show the guide is weakest exactly when the economy breaks trend, so users wait for wider proof.

1. Collect inputs: Take available ONS monthly data for each of the 10 parts. Add business surveys from the EC (European Commission), CBI (employer group), S&P Global/CIPS PMIs (purchasing surveys) and Bank of England Agents, plus judgement.
2. Model each part: Run one statistical model, a math rule from data, for each part using its past trends and survey signals, with judgement added. A surprise jump in one month is often partly reversed during the next month.
3. Aggregate to headline: Combine the 10 indices, levels versus a base, by weight, so the headline index equals the weighted sum. Services strength can offset production and construction softness, leaving GDP near flat.
4. Report and refresh: Report month change, from one month to the next, and 3-month change, plus implied estimates for two consecutive quarters. Refresh every edition as hard ONS numbers replace forecast months.