cross-sectional variation in response and results that disprove your claim

What do you mean? The spread was low enough during the period where it was high mortality that in China, we can lower bound at 1 million even with best case assumptions about mortality compared to the US that likely wouldn't have been true in a still developing country.