For one, there is far less physical security. There's no physical feedback for a transaction to have taken place when the card is e.g. in a wallet or pocket.
The card doesn't even need to be visible for inspection, nor does it even need to be present at the reader terminal for a transaction to take place.
I've seen a proof of concept described that bypasses a lot of the physical security that is assumed to be present with NFC payments. Take two reasonably powerful and sensitive NFC transmitter/receivers, both portable and each connected to a comms device like a rooted Android phone, give one combination pair each to two people involved in the demonstration. Put one of the aerials inside a wallet, carried in the hand with the cable hidden e.g. up a sleeve. This person would be the one "paying". The other person just need to be nearby the "mark" whose card is to be used to pay for the transaction, close enough for the card interrogation to take place. Create a channel where the received data at one aerial is transmitted by the other, and vice-versa. Then when the payment is requested, the shops' cardreader has no way to recognise that the device being waved at it is not the actual one being interrogated for the transaction. The "mark" has no knowledge that their card was just used for a purchase. The merchant has no way to know that the transaction was fraudulent.
The same type of paired-device communication will also work to get through doors that require only a wave of a card in front of it.
So, if you want to have something that can be as easily bypassed as this in your pocket, please ensure that there is a decent faraday cage around it to prevent signal leakage when you don't want it used.