The generic issue with such countermeasures is that they defeat only some keyloggers, and are themselves defeated by the next generation of keyloggers. In this case, it would be easy for a keylogger to take into account the backspace events (the keylogger sees all keyboard events, so in it sees backspaces) to rebuild the actual password. In fact existing key loggers should already do that, because human users use the backspace key occasionally, too !
Another generic issue with an anti-keylogger software is that it is software. A software keylogger, in order to enact its malevolent functionality, must intercept keypress events, and this requires privileged access to the operating system, at least administrator-level. An evil piece of code which obtains such access will not necessarily stop there; it may also scan for the launch of known anti-keylogger software and deactivate it. Malware and virus do that in general, against anti-malware and anti-virus software.
An additional reason why a keylogger would block out your envisioned anti-keylogger software is that keylogger writers generally dislike competition -- if they find other keyloggers from their fellow keylogger writers, they promptly kick them out. Your anti-keylogger software, from the outside, really looks like a keylogger, since it intercepts keypress events; it is then likely to be deactivated by another keylogger which just wants the monopoly of key logging.
Therefore, if software such as you describe exists at all, then it cannot be considered as reliable. It might enjoy some partial success against a few keyloggers, but this won't last.
