First off, I do agree with you that the comments left by the moderator were uncalled for and excessively belittling, but I also agree with the moderator that the question is vague and a bit too broad. The problem is that topics like this are complex. So complex in fact that we have multiple competing organizations all over the world simultaneously working on the problem. The thing is, identity theft is often performed by very skilled, motivated, numerous, and persistent attackers. "Defeating" identity theft is like "defeating" disease. It's just not possible. We can mitigate it, sure. We can treat the symptoms, and we can reduce the risk to acceptable levels, but completely mitigating it can't be done.
For example, concentrating all national identity information in one or a few places has lead to massive breaches of privacy, such as in the Equifax breach. You see there how companies designed to protect identities ends up revealing vast amounts of personal information to many people. You then suggest some kind of authentication. How would that work? People are dumb, so no matter what secret you use to authenticate them formally to the government, that secret will be given out as soon as an official-looking letter is sent to them by a scammer warning them that they need to provide the information or risk having their bank account shut down. Even biometric authentication can be fooled.
The end result is that you are asking a question that is effectively as broad as "why can't we just stop crime?", providing solutions like "maybe it's hard to do" or "maybe there are a lot of differing opinions". The fact is, it's just too vast of a topic to fit on a site that not only wants narrow topics, but wants a clearly defined threat model specifying adversarial capabilities, assets, etc.
Now, since you did ask how to improve your question, I can offer a few suggestions. For a question to be answerable by information security experts, you need to specify a number of things:
How do you define identity theft? It's a broad term meaning many things.
What do you mean by stopping it? Do you want the occurrences to go to 0?
What level of identity theft is acceptable to you? What factors does this depend on?
What assets do you want protected? SSN? CC? Mother's maiden name?
How much lost value do you place on each case of theft (however you define it)?
How many resources do you think should be put into "stopping" it?
This isn't even an exhaustive list, but you need to provide this kind of information if you want such an incredibly broad question to be answered, because it is incredibly broad. Let me put it this way: There are brilliant people whose living is finding ways to reduce the impact of identity theft, and there are brilliant people whose living is finding ways to steal identities and profit from it. This complex interaction, combined with massive geopolitical variables, results in a problem so complex than any question relating to that problem needs to either be extremely specific, or contain a huge amount of detail.
What should your question end up looking like? Probably something like this. That question starts by distinguishing the two kinds of identity theft that are in scope and specifies the assets that need protecting, identifies a very specific and very narrow circumstance (key signing parties), provides two suggested mitigations to reduce the occurrence of identity theft, and mentions potential problems in the form of trade offs with the suggested mitigations. It concludes by asking how one can mitigate the threat while reaping the most benefit from the dangerous situation they are in.