The problem as I see it is that trying to raise funds to try and stop world hunger is somewhat a futile endeavor. Not that it is not worthwhile cause, but that the way the funds are allocated and used are not correctly utilized or implemented. The funds are not even replenished correctly in my point of view.

The Solution
What if donations are collected and saved in a bank account and not touched by anyone. But instead use the interest to spawn new bank accounts and use the interest from those spawned accounts to help fight world hunger. The idea is to use the strength and weakness of currencies as an advantage.

Step 1
Create a main bank account in a country with the strongest currency. Then obtain, request, and save funds into this bank account. But most importantly never withdraw money from this account, only use the interest earned from this account for step 2. Hopefully the funds in this main account will be large enough for this to work, maybe an amount above 1 Billion would be sufficient. Then, in the same country with the strongest currency create bank accounts for every country to be used in Step 2. Also create bank accounts locally in the selected countries to be used in Step 3.

Step 2
Every month transfer interest earned from the main bank account into a selected countries bank account. Then select another country for the next month. Then if all countries had been given interest from the main account, create a new cycle of selection.

Step 3
Every month transfer interest earned from the countries bank account from the strong currency into a selected country's bank locally. Then use the money in this local bank account to feed the people in the selected country. Remember that money is transferred from a strong currency, and because of this more money will be available to feed the people - Hopefully.

Additional notes
If 5 Billion could be saved/raised, then the interest could be split monthly between 5 countries.
If 10 Billion could be saved/raised, then the interest could be split monthly between 10 countries.
Thus the more money could be raised/saved, the faster the benefits will be seen.

Author Bio:
N.G van der Westhuizen is a software developer currently working on an open source Lotto database application. www.slx.za.net