Basically, CBDCs could lead very fastly to a centralization of the banking sector. As normal people do not need bank accounts anymore because they can open CBDC accounts, they would open a bank account only for special purposes, e.g. if they need a loan or if they are entrepreneurs needing to accept certain payment forms. This would rapidly lead to a situation where 1) as banks continue to have to respect the fractional reserve policy of their country, they would come in need to borrow more money from the Central Bank and 2) this and the collapse of the "account market" would very fastly lead to the banking sector becoming much smaller and more concentrated.
What has this to do with efficiency? Well, it's believed in many economics theories that a private - and fairly decentralized - banking sector with dozens or hundreds of competing banks is more efficient in detecting the needs of the local economies, above all when they give out loans to local businesses. With strong CBDCs and a weak banking sector, the Central Bank would have to try to compensate for these advantages of local/smaller banks - they would need to try to foresee the needs of local businesses all over their jurisdiction and plan accordingly. So basically the systems would have more characteristics of a centrally planned economy, and almost universally it's believed in economics that these systems are less efficient when it comes to allocation of resources (see the collapse of the Soviet Union as an example)*.
Now there is a twist still: with P2P loans and even Bitcoin/crypto ICOs (which are, in reality, another form of "loan" to raise funds for businesses and projects) there are new loan mechanisms from the digital era. If these could replace part of the role local banks played before, then CDBCs could work well. But I think governments distrust these loan mechanisms and so they're still waiting and testing.
*China is a strange exception to that theory, and they have long experience with a centrally managed economy, so it's no surprise they're most advanced regarding CBDCs.
Good point. Central banks need to throughly test a CBDC before launching it to the public. As you've mentioned earlier, there's still the problem of efficiency. Banks are trying to figure out how to launch their own digital currency solutions without sacrificing efficiency and stability. It'll be hard to make people use a CBDC over physical cash in its very beginnings, since old people are not willing to change the way they use money. Maybe in a couple of years from now, we'll be able to experience digital payments more throughly.
One thing for sure is that China has been quickly advancing the development of its own CBDC. If it manages to launch it first than any other country, it will have a "leading edge" on the world's economy. Then, it'll be the battle of supremacy as other countries try to launch their own CBDCs as well. The US does not want to lose the "USD" as the world's reserve currency, so I think it will begin the adoption of a digital US Dollar soon. Once CBDCs become a reality, there will be a transformation in many of the world's industries as well as the mainstream economy. Payments between banks (B2B transfers) will be blazing fast, while costs will be reduced like never before. It'll be a new era where intangible items will prevail over tangible ones. Either central banks use a Blockchain of their own, or simply rely on an existent one like Ripple's XRP Distributed Ledger.
Nonetheless, people are still using physical cash and credit/debit cards despite COVID-19. I believe that CBDCs will continue to be in development during the next few years, until they're ready for launch. The main concern will be people's privacy over mainstream transactions performed with a CBDC on top of physical Fiat. But that's something we don't need to worry about at least for 20 - 30 years from now. Just my opinion