It is possible to abandon proof of address altogether and accept an applicant's submission as authentic. Biometric tags and de-duplication software that works across multiple databases - driving licences, hospital records, school and college registers, insurance and bank accounts - would identify cases that call for further verification.There are some really good points here.
Sure, this means a lot of computerisation. So what?
Trust everyone's claimed address, verify those that give cause for doubt. This is integral to inclusive growth.
At first I thought that, if not address, than some geographical descriptor would be necessary in many real world applications. Voting in state and local elections is geography dependent. Many public services are provided only to people in a given jurisdiction.
But author T K Arun makes a good point. In a world of perfect database interoperability and deduplication, residence address doesn't matter much, especially compared to the challenges and misery associated with having a huge population of people without ID.
From the individual angle, so long as an individual can only vote in one place, as long as they can only collect cash transfers intended for one group (for example they are prevented from simulteneously collecting subsidies for rice growers and fishermen), overall ID-based shenanigans will decrease.
On the service provider level, if databases are linked, two schools claiming to educate the same child (and billing the government for it) would have some explaining to do. For more along these lines, see Biometrics "Fix" Identity.
It's an interesting conversation and we may be headed that way, but for now, perfect interoperability and (single factor) deduplication isn't a reality.
But like we always say, don't let perfect be the enemy of good. Give the poor man an ID — even if he can't give a permanent residence address.