Funny you should say that. I developed a short encryption program a couple of years ago which doesn't use any of the standard algorithms. I contacted NSA about it. I gave them a brief description, but for security reasons no detailed description. I told them if they needed a detailed description to contact me. I got a letter back from them telling me to go through the procedure of soliciting business ( like a contractor does ). I wrote them back saying that I was not selling it, I was offering it to them. They sent back a letter saying they were not interested. Now from the description I wrote to them, anyone familiar with cryptography would have requested a second look. Instead they brush me aside, not knowing my background. It seems the NSA has the attitude - " If it not created here, it is no good " By the way, I am a Mathematician with a background in Number Theory
Or maybe that year's budget for evaluating unproven encryption methods had already been spent.
Not really, as the current standard wasn't developed by the NSA. They just went through a round of selection a few years ago and are probably not too interested. The best way to get your method accepted is to publish it in cryptographic circles. Flaws are found in the vast majority of implementations. Even a couple of Bruce Schneier's attempts failed to withstand peer review (his MacGuffin cipher died really fast).
Now if you think you have the perfect hash algorithm, I believe they're looking for one.