World Bank forecasts a 23% commodity-price rise in 2026

Higher food and energy costs, a new estimate of heat deaths, and the uneven spread of AI for cybersecurity put the resources behind public services in focus.

In focus today

A school meal depends on the price of grain and the fuel that brings it to the kitchen. A clinic depends on tolerable working temperatures and functioning digital systems. Today’s three developments examine pressures on those everyday foundations of public provision, from international commodity markets to the smaller firms supplying digital services.

African economies and public finance

Global markets · Kenya and other import-dependent economies

Food and energy forecasts raise the cost of maintaining services

Outlook released in Washington: 28 October 2026. Forecasts: calendar-year averages for 2026 and 2027.

World Bank forecasts for 2026: energy prices up 32% and food commodity prices up about 10%. These are annual global forecasts, not local inflation rates.

The World Bank’s new Commodity Markets Outlook forecasts a 23% rise in overall commodity prices in 2026. Energy prices are projected to increase by 32%, while food commodity prices rise by about 10%. The Bank expects a further food-price increase of about 7% in 2027, taking the cumulative rise over the two years to approximately 18%. Disrupted supplies, higher agricultural input costs and El Niño-related harvest risks underpin the outlook.

These are international commodity forecasts, rather than predictions of household inflation or a uniform increase in government purchasing costs. The baseline assumes no major intensification of conflicts affecting supplies. Exchange rates, taxes, freight and purchasing arrangements influence how international prices reach local buyers.

Kenya’s exposure has an immediate market reference: the Central Bank’s bulletin of 23 October recorded Murban crude at US$108.23 a barrel on 22 October, up from US$102.39 a week earlier. This observation concerns a different oil benchmark and time period from the World Bank’s annual forecasts.

Climate and health systems

Global findings · African livelihoods and health protection

New Lancet Countdown estimates put heat’s health burden in sharper view

Report released: 28 October 2026. Mortality estimate: 2023. Potential-income estimates: 2025.

The Lancet Countdown estimates 710,000 heat-related deaths globally in 2023. This is a modelled total, not a count of registered heat deaths or a local forecast.

The 2026 Lancet Countdown estimates that heat was associated with 710,000 deaths worldwide in 2023, with more than 70% attributable to human-caused climate change. Released on 28 October, the report tracks 54 indicators of climate and health. Its authors call for climate change to be recognised as a global health emergency. That call does not itself constitute a formal emergency declaration by the World Health Organization (WHO).

The economic burden is also uneven. Report contributor the Potsdam Institute for Climate Impact Research says heat-related losses of potential income in Africa were equivalent to 3.2% of the region’s gross domestic product in 2025. Agriculture accounted for 71% of Africa’s estimated losses. These are modelled losses of potential income, rather than a measured contraction of that size in reported GDP.

The mortality indicator combines temperature and mortality information in a model, rather than counting deaths certified as heat-related. Its assumption that the relationship between heat exposure and mortality remains constant over time limits its ability to capture changing adaptation or vulnerability. The global total cannot establish a local death toll.

In perspective

Health protection extends into the working day

The agricultural finding makes the connection between health and livelihoods concrete. A person who cannot work safely in extreme heat may also lose income needed for food, transport and care. Protection therefore depends on conditions beyond the clinic: drinking water, opportunities to rest, workable hours and access to treatment all affect the consequences of exposure.

For national and local health institutions, the strategic value of the report lies in connecting heat surveillance with the places people live and work. Local temperatures, illness reports and service pressures can establish where protection is failing. An international estimate provides context; evidence from those settings determines which populations and services face the greatest strain.

Artificial intelligence and digital government

OECD evidence · Implications for public digital suppliers

Large firms lead in using artificial intelligence for cybersecurity

OECD statistical release: 26 October 2026. Enterprise-use figures: 2025.

Shares of enterprises using artificial intelligence for cybersecurity in OECD data for 2025: large 26.7%, medium 9.6%, small 3.7%. All three bars use a zero-to-30% scale. Adoption does not measure security effectiveness.

New figures from the Organisation for Economic Co-operation and Development (OECD) show that 26.7% of large enterprises used artificial intelligence (AI) for information and communication technology security in 2025, compared with 9.6% of medium-sized firms and 3.7% of small firms. Uses include detecting attacks and authenticating users.

The figures describe adoption, rather than the effectiveness of a firm’s defences. A business without AI may have strong security controls; one using it can remain vulnerable. The release provides comparative OECD evidence, not an estimate for African enterprises or public administrations.

In perspective

Public systems depend on their suppliers’ defences

A public digital service can rely on a large platform alongside smaller firms that maintain software, manage records or support users. The purchasing institution inherits dependencies across that chain. A supplier’s AI capability tells it little about how quickly a compromised account will be contained, a backup restored or affected users informed.

The useful distinction in procurement is between buying a technology and securing a service. Evidence of tested recovery, controlled access and an agreed incident response gives a more direct account of service resilience. For smaller suppliers, clear and proportionate requirements can support competition while making responsibility for failures explicit. The OECD findings motivate that examination; they do not establish the security of any particular supplier.

The cost of keeping a service dependable

Higher fuel prices and hotter working conditions can meet in the same service. Transport and cooling become more costly just as heat increases the need for protection and care. That interaction puts pressure on both household incomes and public operating budgets. Digital dependence adds a separate exposure: a service can be adequately funded and still become unavailable through a supplier failure. Together, these developments show why the resources required to keep a service running deserve as much scrutiny as the decision to establish it. Contract prices, working conditions and recovery arrangements influence how much dependable provision public spending delivers. The resulting test of progress is concrete: whether meals arrive, care remains accessible and essential services continue when conditions deteriorate.

Sources and evidence

The original publications and their dates, with the scope of the evidence behind each reading.

African economies and public finance

Annual global price forecasts. Local inflation and public purchasing costs depend on the exchange rate, taxes, freight and contracts.

World Bank: Commodity Markets Outlook release28 October 2026Primary release. Supports the commodity-price forecasts, their periods and baseline assumptions. Annual global forecasts do not translate directly into local inflation.
Central Bank of Kenya: Weekly Bulletin23 October 2026 · observations to 22 OctoberPage 1: Murban crude observations for 15 and 22 October. A different benchmark and period from the World Bank’s annual projections.

Climate and health systems

Modelled global mortality for 2023 and potential-income losses for 2025. Neither supplies a local death count or an observed GDP contraction.

Lancet Countdown: Heat-related mortality, indicator 1.1.52026 report · indicator updated September 2026Primary indicator, model description and caveat on a constant exposure–response relationship. The 2023 global estimate is not a local death count.
World Health Organization: Lancet Countdown report update28 October 2026Primary collaborator’s account confirms publication, headline mortality finding, 54 indicators and the authors’ emergency call. This call is distinct from a formal WHO declaration.
Potsdam Institute for Climate Impact Research: Report findings28 October 2026Report contributor’s account supports the Africa potential-income estimate and agriculture’s share. Neither is a measured change in national accounts.

Artificial intelligence and digital government

OECD enterprise adoption in 2025. The figures do not measure security effectiveness or represent African firms.

Organisation for Economic Co-operation and Development: AI for cybersecurity statistical release26 October 2026 · enterprise data for 2025Primary release reports adoption by enterprise size. Its figures do not measure security performance or describe African firms.