Maybe it has been done. Maybe it hasn't.
Thing is, http://www.commondreams.org/view/2012/01/03-2#.TwM6I9iK42A.twitter tells, with 7 detailed angles, how the money sent to Haiti either never made its way there or was given to govts and NGOs that didn't use it wisely if at all.
So here's an idea, that most likely has many issues - or why hasn't it been done yet?
Give the money directly to those affected by a natural disaster.
How could that be done?
Let's say in a country like Haiti around a million people become homeless in five minutes.
A million people is unmanageable, so we select a sample. Maybe 1000 people, just to see how this works. Or even 100 people. Or 10, if we're really poor. Maybe our twitter friends can help us organise that, due to the 6-degrees-of-separation thing, right?
It will be a biased sample, to be sure. So we make it unbiased by asking our Haitian contacts to do a random selection. Select a random town, within that a random neighbourhood, within that a random person.
Wait. A random person? It could be a thief, a drug adict, hell even a corrupted govt official! Yes. But most likely it will be a person in need. Or, if randomness is not carried out in all cases, then half of the recipients will be people in need, and that's better than what's currently happening.
And maybe we want semi-randomness. Maybe we want to give money to the elder, or to women, or to teenagers, or whatever. We could have several subsamples, to see what works better and why.
But, remember, this is not the final solution. Just a test.
Ok, so the person has been selected. Now what?
We'd need three more elements: identification, money, and empowerment.
Identification is needed if we are to track what happens. Maybe our imagined Haiti contact can take a picture of the person, with some paper to identify them? This has to be different in each place. Maybe there are places where some kind of ID works already. (I don't know if Vinay Gupta's CheapID thing has a place here, cos I don't even understand how that process goes.)
Money is what we're giving away, right? Do we give them a phone with credit they can spend? A credit card? They go to some "bank" place with their "tatooed" identification?
How much money? Enough for a hexayurt plus implements? For three months of food plus seeds and agricultural implements a la Paul Polak? The per-person average supposedly donated to Haiti?
In weekly quantities? Half as one bulk and half in weekly quantities?
Empowerment is a booklet with suggestions on how to use the money. A pediabook with a selection of appropedia. A phone so they may ask our contact in Haiti, so that the international community can deliver knowledge right to where it's needed, in the format that will be most useful.
Wouldn't you want to know what they really feel is their problems? This could even be financed as journalism, assessment or whatever!
Empowerment is optional. Maybe we could even try and give either empowerment or money to see which works better, and maybe having both doesn't make a difference. This is all about testing, so I'd be glad if all options were tried out, and failures as well as successes documented.
Tracking is probably hard. We'd like to know what the end result is. How fast these people get on their feet, if they do. Whether or not this is better than other ways of providing aid. Maybe we want a report (using Akvo's Really Simple Report system) before the next chunk of money is made available?
What would we want to know? In-variables such as age, gender, personal story before the distaster situation, and others? Out-variables such as results in terms of practical reduction of the "six ways to die" issues (see below), or even some appropriate metric of these things? Through-variables such as how they select priorities, how they used the money, how they used the stuff they bought with the money, who they spent it on, etc?
First step. My first step is putting this idea out. Any thoughts on what the next step could be? Contacts of any kind? A particular place that would be suitable (not that there are no disasters going on)?
Has this been tried and it has failed? Has it not been tried and if so, why? Has it been tried differently? Do you see any obvious badness in it, or any obvious improvements?
What do we do?
Where I come from: I'm a public health practitioner in a rich country. I translated Vinay Gupta's http://hexayurt.com/hexayurt_country/ and SCIM into Spanish, and have helped document how-to hexayurts at appropedia (even tho' I've never built a real size one myself, yet). I've never travelled to a troubled country.