We know in small scale sleep lab studies that low sleep leads to changes in emotional processing (one of the functions of REM sleep is to tune your emotional regulation), so that at least offers a plausible mechanism of action. As far as risky behavior though, you're right that causation isn't yet firmly established here.

https://www.psychologytoday.com/us/blog/dream-catcher/201112...

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3163498/