I don't think it should really be any different performance-wise. On the C side of things, a connection is just one object writing to memory (outlet) and another reading from it (inlet). [send] and [receive] internally do the same thing. The only difference I can think of is that you still have to connect to the [send] and from the [receive].
This:
[bng]
|
[send test]
[receive test]
|
[bng]
instead of this:
[bng]
|
[bng]
Because of that, there's basically one extra read/write in the chain. But accessing memory is pretty much the fastest thing a computer can do, so it's likely unnoticeable.