Mirlande is a young mother of five. She lives on a hillside outside of Saut d’Eau. Her home is made of woven strips of wood that are covered with a layer of mud, and a roof of leaves. The house has a dirt floor, and lacks even a very basic latrine. She shares the home with her kids and the man who is father to the four youngest children. The seven of them live on what the father brings in by farming and by buying pigs on credit and selling them quickly at a profit. They don’t have the means to get into the more lucrative business of raising the pigs. That would require the capital both to buy the pigs clear and invest some extra money in feed.

Before Fonkoze can begin to help Haiti’s poorest families lift themselves out of misery, we have to identify those families. Good selection is the cornerstone on which the CLM program is built. It’s a three-step process, and we’re in the middle of it right now.

The first step is Participatory Wealth Ranking (PWR), which I’ve written about before. It’s a process we saw a lot of during our month in Bangladesh. (See: The Right People.) Fonkoze staff members go into the communities we’re planning to work in and hold meetings at which community members provide initial information as to the economic status of each household. Staff members leave these meetings with maps of the communities and lists dividing the households into between four and seven classes, ranging from the richest of the rich – whatever that might mean for a given community – to the poorest of the poor.

The information is useful. It introduces us to a community and teaches us something about how to approach it. We could hardly do without it. But it is neither sufficiently detailed nor 100% reliable. So Fonkoze adds to and verifies it through two more levels of field research on its way to selecting the families who will be invited to participate in CLM.

The first is a set of detailed field surveys completed by the men and women who will eventually serve as case managers for the families we select. Teams of two-three case managers go door-to-door, filling out multiple questionnaires for every family that the PWR meeting has placed in one of the bottom two economic categories. We collect information about the households’ income sources, assets, and their ways of life. These surveys give us extensive baseline data about the families we will eventually work with, and they allow the case managers to make their preliminary recommendations.

Fonkoze has distinct programs for very poor and extremely poor families. They are very different approaches, and it’s important that we get families into the one most appropriate for them. For very poor families we have Ti Kredi, or “Little Credit”. As its name suggests, it is a loan program that provides a more supportive framework than our regular credit programs do so that poorer women can get their businesses onto solid ground. The program lasts six months, includes a supportive repayment calendar and extra education and accompaniment, and prepares participants to move forward through regular solidarity-group credit.

But credit won’t work for the poorest of the poor, because they cannot sustain a business without significant additional support. Their immediate needs would inevitably consume any assets that they might try to invest. Even the very poor are hungry most of the time, but the extremely poor might go for days at a time with nothing substantial to feed their kids. It would be foolish to expect those who live as they do, on the edge of starvation, to set aside resources that could easily be converted into food.

Case managers use the data they collect in these surveys, but also their own good judgment, to confirm that the families they visit truly belong to one of the bottom two groups and to recommend them for either Ti Kredi or CLM. But they only make recommendations. Final selection is made by one of our five field managers. We revisit the homes that our case managers have recommended for CLM to talk with the women who would serve as our program’s link to them, to the women who would participate in the training we offer and take responsibility for the assets we give. We establish how many children the women are responsible for, what other people they have to share their responsibilities with, and how they feed and otherwise care for those who count on them. Our object is to answer a simple question: Could the woman we are talking with succeed in Ti Kredi? Only if we decide that she could not do we select her for CLM.

Most of the judgments are easy. We visit home after home after home with five or six or eight or nine children, all of whom are terribly, visibly hungry. These children might have absentee fathers, or they might have hardworking, but landless ones, who can’t get much more than a dollar for a day’s labor in the field. The families might live from sharecropping, typically receiving two-thirds of what they can scratch out of the fields they farm, enough perhaps to eat a meal a day for a couple of months each year. The women themselves have no small businesses because they have no ability to set aside assets for sale. Many of them labor in the fields next to the men, earning only about 75 cents a day. Some of them just sit at home and rely on their neighbors’ occasional charity, neighbors who are, themselves, far from well off. So when you’ve talked with these women for a few minutes, it’s easy to sign the form that authorizes their participation in CLM.

But then there are cases like Mirlande. She has a working, though volatile partnership with her husband. They do not own the land they live on right now. They rent the house and the land it’s on, which includes a small plot that they farm. This year it cost about $5, up from $3.75 the year before. But they also own two pieces of land, a smaller one her husband inherited from his father, where they are starting to build a house, and a larger one they bought this year when her husband sold a cow. Her three oldest children are in school. The first has his tuition paid by his father, who supports him in no other way, and the other two through a kind of buy-one-get-the-next-two-for-nothing arrangement. They regularly eat about once a day, feeding on the millet and corn that they harvest for as long as that lasts and then on what they can buy through the income of her husband’s pig trade.

I doubt neither Mirlande’s poverty nor her need. I know that she and her family are hungry most of the time. They lead a terribly hard life. But the structure of support that her husband provides, including a regular if inadequate source of income, and the capacity for future planning that their purchase of land demonstrates at least suggest that credit might work for her.

So I went against the case managers’ recommendation and put Mirlande on the list for Ti Kredi. It isn’t the kind of decision one can feel good about because it robs a very poor family of resources that could do its members an awful lot of good. But we need to reserve those limited resources for the families who need them most because, as it stands, there are many more of them in Haiti than we can possibly serve.