This is a good question. In the absence of other explanations of market failure, this is an update away from the view that direct anti-poverty interventions do have high ROI. While some RCTs of anti-poverty programmes like the Graduation approach might show a big ROI, perhaps the market actors know not to trust these RCTs. Maybe they have an implicit view on the external validity of these studies, which has been demonstrated to be low

This is a really important question, and I agree a bit of a puzzle. Burke, Bergquist, and Miguel sort of address it in "Sell Low and Buy High". Burke et al. test the effect of providing credit to farmers in Kenya. They find that with access to credit, farmers are able to save more of their harvest and sell it at a different time when local prices are higher, raising their income and allowing them to pay back the loan. The return on investment is 29%.

But, as you say, 29% is a big return - why aren't local lenders already providing this opportunity? From the very end of the paper (p. 838-9):

What our results do not address is why wealthy local actors—for example, large-scale private traders—have not stepped in to bid away these arbitrage opportunities. Traders do exist in the area and can commonly be found in local markets. In a panel survey of local traders, we record data on the timing of their marketing activities and storage behavior but find little evidence of long-run storage. When asked to explain this limited storage, many traders report being able to make even higher total profits by engaging in spatial arbitrage across markets (relative to temporal arbitrage). Nevertheless, this does not explain why the scale or number of traders engaging in both spatial and intertemporal arbitrage has not expanded; imperfect competition among traders may play a role

So, yeah, they don't really know. Lenders have other good opportunities, and maybe discount future returns enough that they would rather engage in spatial rather than temporal arbitrage. As Larks wrote, too, risk aversion and fixed costs could make these very small loans to people in extreme poverty unattractive.

I'll try and answer the question regarding why banks don't lend for people to get themselves dewormed, though these factors will translate to the question at large I think. Most of what I write is drawn from Poor Economics, by Banerjee & Duflo.

One factor has to do with the systems for lending in third world countries. As Larks writes fixed costs associated with loans are often too high to justify smaller loans, and as you say, lenders often don't carry the power to be able to enforce and collect these loans (so higher loan amounts can still be unattractive to these institutions). This is why private money lenders from within the community, or borrowing from friends/family is much more common in these areas, compared to banks or other formal institutions. There is a role for microfinance institutions to play here, as far as I'm aware RCTs have shown some modest economic benefit to recipients, but nothing huge.

Even if credit were available, people may not take these interventions due to a variety of factors. Many people do not know the effects of deworming on health or education, and the effect of those in turn on future economic prospects. Human biases to discount future goods/rewards would also play a role. Deworming is just one such intervention people in third world countries would need to consider, among other responsibilities are buying chlorine to clean their own drinking water, buying a bednet, vaccinating their children, etc. in addition to leading their regular lives. In first world countries, we enjoy having many of these interventions "built-in", clean water from the tap or vaccinations in schools for example, so it's not surprising individuals in third world countries don't act on any specific intervention on their own accord.

The reason lending for these interventions, at least for the "low hanging fruit", such as deworming, doesn't exist would have to do with the fact the institutions and systems aren't set up to accommodate this, and even if they were people may not necessarily take them up. I suspect making interventions easier to access (probably free) will go further than having credit more easily available. This feels similar to the reason insecticide-treated bednets are given out for free, rather than selling them (even at a highly subsidised rate). Because of this, I'm inclined to think more "top-down" approaches, such as universal free deworming in schools, would work better than having a lending institution.

For example, why aren't banks lending money for people to pay to get themselves dewormed?

I think the question in some sense is 'why don't people pay for themselves to get dewormed?'. It's unlikely that banks in the third world would be able to make a personal loan that could only be used by that person for deworming.

Separately, an additional factor to the ones you mention is the fixed costs of making a loan. Things like KYC and AML regulations are similarly onerous for small loans as large ones, making smaller clients disproportionately expensive to deal with.

This is essentially the premise of microfinance, right?