Are structural VARs with long-run restrictions useful in developing business cycle theory?

No, unless technology shocks account for virtually all of the fluctuations in output.