If the Fed is bothered about primary dealers lacking the balance sheet capacity to do reverse repos on a large scale, why not use its regulatory powers to ease these constraints? The Fed might want to do as much as $500bn in reverse repos. The dealers have balance sheet space for $100bn at most.
A short-term reverse repo with the central bank ought not to be the kind of asset a bank needs to set much capital aside for, nor the kind of asset that counts against crude leverage limits. I am not an expert on the regulatory side of this but I suspect the Fed might be able to do something about this if it put its mind to it.
Reverse repos matter for the following reasons. When the Fed gets round to raising rates its main tool will be the ability to set the interest rate on bank reserves and thereby put a floor under the Fed funds rate. But it may need to supplement this by mopping up part of the excess reserves in order to stop the fed funds rate sagging well below the rate on bank reserves. This is because it is only allowed to pay interest on reserves to banks and not, for example, to Fannie and Freddie even though they are large players in the overnight fed funds market.
Reverse repos provide a way to drain some of the excess reserves and ensure the fed funds rate trades close to the interest rate on bank reserves. The problem is that primary dealers who would ordinarily act as the intermediaries in the reverse repo market are balance-sheet constrained.
One option is for the Fed to go around the dealers and conduct reverse repos directly with Fannie and Freddie and/or the giant money market mutual fund industry. It continues to explore these options. But going around the dealers to new counterparties is technically complex.
So (assuming it is legally possible) why not ease the regulatory constraints on the dealer balance sheets instead? The Fed is telling everyone who will listen about the synergies between regulatory policy and monetary policy. This could be a good place to start.