Chloe I would tend to attribute much of it to psychosomatics. Everyone is suffering from that as we hear doom & gloom constantly. Lots of it is likely real, but lots is also likely psychosomatic. Many of those probably end up being other colds (one of the other 4 CoVs), flu, other rhinovirus, allergies, asthma, strep or bronchitis and so on. After all, we still have all of those things too. So just keep in that mind.
I wouldn't venture a guess at the specific number except to say that I think both are wrong and the true number is significantly north of that.
However, think of it this way. Let's say your 4900 example were correct (justHarris County). If we have 2 deaths in Harris and the reality is ~5000 people are/were infected that means the hypothetical IFR is now .04%
(2/4900 = ~.0004 or roughly .04% - seasonal flu for example is 2.5x more fatal in *THAT* case).
The point of this is to give some comfort in knowing that as the number of
unknown cases rise & deaths don't follow that exponential curve the mortality rate continues to drop. That's why I keep reinforcing that idea.
My guess, though I really don't want to put a quantified number on it yet, is that there are far more than what you suggested & they're asymptomatic, immune, weathering through it due to milder symptoms, think it's allergies, asthma, etc.... Many couldn't get tested anyway b/c as Ben stated the testing is abysmally low. :/
Note that this is all an
inductive reasoning syllogism. I'm not at all declaring it to be a formal deduction based on definitive argument, which is why I'm careful to say, "
*IF* the above premise is correct we could conclude it's
likely that our IFR is much lower than it
appears based on currently confirmed cases".