That's fine for historians, but useless for making decisions about interventions in the real world where there is always more than one confounding factor.
But confounding factors average themselves out over long period of times. Hence why excess mortality is good, because you compare it to decades of data where the smaller effects even out.
If I can't get a heart surgery bc doctors unavailable and die, that's (on average) caused by covid. Sure, it's possible that my doc was unavailable at the time bc he hammered a finger, but "on average" over the timespan the only constant factor was covid (or covid-policy consequences).
