Sign In or Create an Account.

By continuing, you agree to the Terms of Service and acknowledge our Privacy Policy

Technology

GiveDirectly Is Giving Cash to L.A. Fire Victims, No Questions Asked

The nonprofit uses a mixture of public data and algorithmic magic to unleash funds fast.

Digitally distributing money.
Heatmap Illustration/Getty Images

Whether they’re dealing with fires like the ones ravaging Los Angeles or hurricanes like those that wreaked havoc in Florida and North Carolina just a few months ago, when natural disasters lay waste to homes and towns, what low-income residents often need most is quick cash. That, however, can be difficult to come by. Insurance companies can take months or even years to fully resolve claims. The Federal Emergency Management Agency requires significant documentation before it will offer relief, and often denies victims with no explanation.

The nonprofit GiveDirectly is trying to circumvent all this maddening complexity, working with Google to overlay government data on things like median income and food stamp enrollment with damage data gleaned from satellite imagery and local incident reports to get cash in the hands of those who need it most — quickly. After a disaster, low-income residents in especially hard-hit areas are automatically deemed eligible for aid, no opt-in necessary. They’ll get a notification on their phone that they qualify for a direct cash transfer, and can enroll in a matter of minutes, with no additional documentation required.

“Especially as disasters become more prevalent and more severe, having a way to pre-verify vulnerable populations — to get people resources as quickly as possible — becomes so valuable,” Laura Keen, GiveDirectly’s U.S. program director, told me. As she explained, cash is often more useful than “in kind” donations such as clothing or food, as it allows recipients to prioritize specific needs and reduces barriers associated with government-run disaster programs. “You have to have the means and the know-how and the language abilities to apply for that assistance,” Keen said. Still, over 75% of global humanitarian assistance is in-kind.

GiveDirectly set up its fundraising campaign for L.A. fire victims on January 10, and is thus far over 40% of the way to its $1 million dollar goal. While the fundraiser won’t officially close for another 25 days, Keen said the organization plans to send out its first payments “as soon as next week.” While GiveDirectly has yet to finalize amounts, it estimates that recipients will get on the order of $3,000 to $4,000 — significantly more than the nonprofit gave to victims of Hurricane Ian in 2022 or Hurricanes Helene and Milton last year. That’s because with these fires, “the damage has been so severe, and we expect people are going to be facing temporary housing costs for a matter of months,” Keen explained. For a campaign like this, Keen said she expects about 88 cents out of every dollar donated to go directly to affected individuals and families, the same efficiency rate as the organization’s Helene and Milton campaign. That remaining 12 cents will go towards transaction fees, offices, and staff.

If these rapid payouts remind you of parametric insurance, you’re on the right track. Parametric insurance also exists to get cash quickly into the hands of those who have experienced disaster, without the need for damage audits. But as is implied by the word “insurance,” it is also an opt-in service that involves the payment of monthly premiums. GiveDirectly’s cash comes out of the blue, free and clear.

To get the actual money out the door, GiveDirectly works with Propel, an app for low-income households to manage government benefits such as SNAP food stamps. GiveDirectly tells Propel what areas its mapping tool has honed in on, and Propel sends out an alert to users in these zones, notifying them that they’re eligible to receive money. Individuals then complete a brief survey confirming their contact information, preferred language, and signing some consent notices.

“The last response that we did in western North Carolina and in Florida last fall, it took them, on average, 68 seconds to complete that enrollment form,” Keen told me. The last time she looked at the data, there were about 2,300 households using Propel in the impacted areas of L.A., a number that’s only growing as the largest fires remain uncontained. Once people enroll, they can expect to receive money directly to their debit accounts within three days.

While quick and simple, this strategy is far from comprehensive. Only about one in four households that receive SNAP benefits has the Propel app. And those that do may not open it regularly, meaning they could miss the alert that they qualify for cash. For Propel users who see the notification, Keen said, enrollment is above 80%, while overall user enrollment is much lower — around 40%. “But typically, we have more Propel users than we have funds,” Keen explained. Basically, it wouldn’t actually be possible to give the target amounts to everyone who meets the criteria. Rather, the strategy is to get money out as quickly as possible, knowing full well there'll be many who are missed. Plus, relying on Propel makes the whole system safe from fraud (something GiveDirectly has dealt with in the past), as Propel users have already verified their eligibility for government benefits. “So we just have very high confidence in who we're supporting,” Keen told me.

Domestic disaster relief was not initially on the agenda for GiveDirectly, which was started in 2008 by a group of econ grad students at Harvard and MIT as a way to get money into the hands of some of the poorest people in the developing world. Since the organization began accepting public donations in 2011, it’s mostly retained this international focus, making its first foray into domestic cash transfers in 2017, when it provided physical debit cards to victims of Hurricane Harvey in Texas.

“At that time, we had pretty rudimentary targeting,” Keen told me. “We would just drive to different areas, talk with as many people as we could, visually look for signs of damage, try to source any open source information where we could, and then overlay that with administrative data.” Since then, the company has integrated artificial intelligence into its hurricane relief efforts, training algorithms to generate damage assessments for thousands or even millions of structures. But fire damage is much more uniform (if a house burns, it’s usually 100% destroyed) and easy to identify from satellite imagery alone, Keen explained.

If GiveDirectly exceeds its fundraising target in L.A., however, it may run out of eligible residents who are reachable via the Propel app, meaning the organization will need to go back to basics: establish an in-person presence in the city., enroll people onsite, and hand out debit cards once more. “Right now our goal is to get to $1 million, and then all of that we can deliver via Propel,” Keen told me. “But if we exceed that, we would definitely explore other options.”

Green

You’re out of free articles.

Subscribe today to experience Heatmap’s expert analysis 
of climate change, clean energy, and sustainability.
To continue reading
Create a free account or sign in to unlock more free articles.
or
Please enter an email address
By continuing, you agree to the Terms of Service and acknowledge our Privacy Policy
AI jail.
Heatmap Illustration/Getty Images

Like many new parents, I devote considerable time to thinking about sleep and why it’s not happening. Should I have sung the bedtime song and then changed the diaper? Did the baby need a fourth nap, or was the mistake letting her take a third so close to bedtime? It came as a surprise the other day, then, when a fellow parent in my baby group revealed she isn’t overthinking the whole sleep schedule thing at all. “I asked ChatGPT to write my baby’s sleep plan,” she told us. “It’s validating!”

To this author, personally, outsourcing parenting decisions to the world’s most sophisticated Mad Libs respondent seems like one of the signs that we’re doomed. Sleepmaxxing mothers aside, a plurality of Americans agree with me. Per Heatmap Pro’s latest polling, 45% of voters are “pessimistic” about the long-term impact of artificial intelligence on their lives, with just 22% saying they’re “optimistic” and about a third saying they’re unsure.

Keep reading...Show less
Blue
AM Briefing

Oklahoma!

On depleted U.S. oil stocks, Taiwan geothermal, and hybrid sales

Gentner Drummond.
Heatmap Illustration/Getty Images

Current conditions: The southwest monsoon known as “hagabat” has started in the Philippines, dumping up to 4 inches of rain on the archipelago • A strong geomagnetic storm, ranked just two levels below the most powerful type of event of this kind, is underway, threatening radio signals, GPS, and other human instruments that are sensitive to shifts in the Earth’s magnetic fields • San Antonio, where the glorious New York Knicks defeated the Spurs last night, is bracing for rain through the weekend.


THE TOP FIVE

1. U.S. oil stocks drop to the lowest level since 2004

To put it in terms a movie lover could understand, President Donald Trump’s Iran War is drinking the U.S. government’s milkshake. Federal stocks of oil have dropped to their lowest level since 2004. Commercial crude stocks fell by 8 million barrels to 433.7 million last week, according to The Wall Street Journal. Unless the Strait of Hormuz reopens soon — which looks less likely now that Iran has called off negotiations with the U.S. and Israel — prices could hit $200 per barrel by summer, said Bob McNally, president of the Rapidan Energy Group consultancy and a former White House adviser. “You start to raise the risk of spillover into other sectors, the economy and financial system … it detonates fragilities in the broader economy and financial system,” he told the Financial Times.

Keep reading...Show less
Blue
Daily Briefing

What’s Powering Clean Energy

Notes from Heatmap’s second Energy Entrepreneurship Summit.

A tokamak.
Heatmap Illustration/Getty Images

I’m writing from Washington, D.C., today, after having the privilege of watching (and moderating) Heatmap’s second Energy Entrepreneurship Summit this morning. We heard from folks leading in a variety of technologies — geothermal, batteries, fusion, conventional nuclear — but I was struck by a few common themes.

The first was the new wave of excitement about fusion energy and how, in some ways, the artificial intelligence boom has reinvigorated the fusion conversation. Much like fusion, AI was a long-prophesied technology that made steady, iterative improvements over time — and then, one day, delivered a transformative product in the form of ChatGPT. I’m not sure if fusion has yet had a raw technological improvement on par with the transformer, the neural network innovation that preceded today’s AI chatbots and agents, but fusion startups have reported significant improvements in recent years. The industry believes — as do some fusion-pilled policymakers — that they will have commercial reactors on the grid by the mid-2030s.

Keep reading...Show less
Blue