You are probably right that the numbers out of Hubei province are inflated and less severe cases are underrepresented. Unfortunately, according to some experts, the case fatality rate of ~2% seems to hold even for cases outside of China.

Statistical modeling of the case fatality rate using only cases and deaths reported from outside China was done by Christian L. Althaus (Institute of Social and Preventive Medicine, University of Bern), this estimates the CFR for COVID-19 at 2.1% although with a high uncertainty (95% confidence interval: 0.5%-5.4%). Data and R code for his modeling are available on Github: https://github.com/calthaus/ncov-cfr

A comment from Adam Kucharski (Associate Professor, London School of Hygiene & Tropical Medicine) on Twitter: https://twitter.com/AdamJKucharski/status/122970801348691148...

"... The upshot: the widely quoted 2% fatality for China is calculated incorrectly, because it's based on data that is under-reported and doesn't account for delays. But, confusingly, these errors may actually cancel out, leading to an estimate that is right for the wrong reasons ..."