Cash before El Niño: what paying early can and cannot do

On October 16th Peru's El Niño commission, ENFEN, said that the warming of the sea off the coast would stay of extraordinary magnitude until January 2027 and remain strong until early autumn. Its latest communiqué expects heavy rain on the northern and central coast and rivers such as the Piura and the Chira running high from November. At the other end of the country it expects below-normal rain in the southern highlands and the Titicaca basin shared with Bolivia, with possible water rationing. A month earlier the government had given Pensión 65, the non-contributory pension for poor older people, a new task: running cash payments to people affected by emergencies, under Supreme Decree 005-2026-MIDIS. Who is paid, how much and when are left to the rule that authorizes each payment. The forecast gives the state months of warning. The decree gives it a channel. The question is whether to use both before the rain arrives.

The problem

Disaster relief usually arrives after a shock, often weeks or months after. Forecasts now allow governments and agencies to pay households before it, an approach known as anticipatory or forecast-based cash. Our question is narrow: does paying poor households before a forecast climate shock protect them better than paying the same amount afterwards, and by how much? This is not the same as asking whether cash helps after a disaster, which is well established. It matters for Peru and Bolivia because the people most exposed to El Niño, small farmers, informal workers and families with young children, have the least room to borrow or draw down savings when a covariate shock, one that hits a whole community at once, shuts down the usual ways of coping. If timing matters, a payment calendar is a policy lever. If it does not, the money is better spent on reaching more people.

What the evidence shows

Theory says timing should matter when households face a liquidity constraint at the moment of need: without cash, they skip meals, sell livestock at low prices or borrow at high interest, and some of those losses are hard to reverse. Evidence on seasonal hunger supports the mechanism even without a disaster. In Zambia, a randomized trial by Günther Fink, Kelsey Jack and Felix Masiye offered small farmers a loan worth about $33 in the hungry season, when food runs short before harvest. More than 98% took it. In the first year, treated households sold 25% fewer hours of casual labor to others, worked more on their own farms and raised the value of their harvest by about 9%, with the largest gains for the households with the fewest liquid resources. Money that arrives when it is most scarce changes what people can do, not only what they can buy.

The anticipatory-cash studies come mostly from river floods in South Asia. In July 2020 the World Food Programme sent 4,500 taka (about $53, roughly two weeks of food spending) by mobile money to households along Bangladesh's Jamuna river, days before the flood peaked. Ashley Pople, Ruth Vargas Hill, Stefan Dercon and Ben Brunckhorst compared recipients with households on the same vulnerability lists who could not be reached by phone or lacked a mobile-money account. The comparison was not randomized, so recipients may differ in ways the authors cannot fully rule out. Recipients were 10 percentage points less likely to go a day without eating during the flood, against 28% of comparison households; they lost fewer assets and relied less on costly borrowing. Receiving cash one day earlier relative to the local peak was associated with slightly better adult food consumption, though the authors caution that the variation in timing was too narrow for firm conclusions. Above all, the study compares cash with no cash, not early cash with late cash.

Two randomized trials make that comparison. In north-east Nigeria, Bedru Balana and co-authors at IFPRI assigned 1,450 households either to cash released when climate data crossed risk thresholds before floods or to the same cash after the flood. Anticipatory cash reduced negative coping strategies and increased preparatory adaptation and investment in productive assets, but it had no significant effect on short-term food or non-food spending compared with cash after the shock. In Bangladesh and Nepal, a trial by the World Bank and WFP, led by Paul Christian and colleagues, compared cash delivered within days of a flood with the same assistance months later and reports overall gains in food security and psychosocial well-being from timely cash. Read together, early cash buys preparation and helps households avoid the most damaging forms of coping. Neither trial finds that timing alone transforms household welfare.

A second body of evidence concerns systems rather than timing. In Ethiopia, Erwin Knippenberg and John Hoddinott found that drought cut the number of months households felt food secure for up to four years, and that payments from the Productive Safety Net Programme reduced the initial impact by 57% and removed it within two years. In Niger, a randomized evaluation by Patrick Premand and Quentin Stoeffler, combined with satellite rainfall data, found that multi-year transfers raised consumption by about 10% on average, with the gains concentrated among households hit by drought, whose welfare gains exceeded the transfers because their farm and business earnings were protected. A UNICEF meta-analysis of social protection during covariate shocks finds average increases of about 6% in school enrollment, gains in dietary diversity and small positive effects on children's height. The protection in these studies comes from support that was already in place when the shock arrived.

What is at stake is clearest for young children. After the 1997–98 El Niño floods in Ecuador, Maria Rosales-Rueda found that exposure in the womb, especially in the third trimester, was followed by shorter stature at ages five and seven, and first-trimester exposure by lower cognitive scores, while household consumption fell. In Mexico, Arturo Aguilar and Marta Vicarelli found that children exposed to extreme rainfall early in life scored 0.15 to 0.19 standard deviations lower on cognitive tests at ages two to six, mainly through changes in diet, with partial mitigation where government programs reached them. In Peru, Elorreaga, Huicho and Lescano compared the Piura coast with the less-affected south before and after 1997–98: stunting rose clearly among infants under one year (a prevalence ratio of 2.48), while the estimate for all children under five was not statistically significant. The harm is concentrated in narrow windows of early life, which is where well-timed support can matter most.

The ODA view

We read the evidence as supporting early payment through existing programs, for specific and modest reasons. Cash before a shock helps households prepare and avoid the costliest coping, such as skipped meals, distress sales and expensive debt. The trials do not show that paying early, rather than promptly afterwards, produces large or lasting welfare gains on its own. The stronger lesson, from Ethiopia and Niger, is that protection comes from a system that already reaches poor households before, during and after a shock. For Peru, Supreme Decree 005-2026-MIDIS provides a channel but leaves triggers, amounts and funding to later rules. We would set those before January, use the rosters of Juntos, Contigo and Pensión 65 to pay early in the districts where ENFEN's forecasts are firmest, and give priority to pregnant women and households with infants. Where coverage must be phased anyway, randomizing the order of payment would cost little and tell the region, for the first time, whether timing matters in an El Niño.

What we still don't know

There is no rigorous evidence on forecast-based cash from Latin America. The studies we have concern river floods with lead times of days, not a coastal El Niño that lasts for months, brings several peaks and causes drought in the highlands at the same time as floods on the coast. We do not know how much is lost when a forecast fails and money is paid for a shock that does not come, how local prices respond when many households receive cash at once, or whether early payments reach the infants and pregnant women whose health the evidence shows is most at risk. Nor do we know whether the gains from paying early last when one flood follows another. A slow-onset drought in Peru's southern highlands or on the Bolivian altiplano may call for a different calendar altogether.