On my somewhat complicated post on negative rates of return from last week, Robert Sams writes:
Very interesting post and #5 is crucial (it’s a geometric process). Two points.
1. I think that we can substitute “ability to leverage at near-treasury rates” for “special trading technologies” and get the same implied predictions yet put the relevant institutional factors into relief.
2. Your #1-3 still works with the wrong model of Treasury returns, as it implicitly models demand as if it’s coming from a “real money” portfolio sort of buyer. Those guys exist of course, and they’re basically buyers at any price (central banks, regulatory demand, etc.). But if we ignore CB policy expectations, the valuation is set in the leveraged market, which is much larger, and treasuries trade rich /not/ so much b/c people want safety and therefore want to buy them, but rather they trade rich because people want to /short/ them for hedging purposes (e.g., investor wants corporate credit w/o the interest rate risk.)
Sounds paradoxical, I know, but failure to appreciate this fact is the basic misconception of the entire “risk premia” way of modelling this stuff.
For any given treasury issue, X billion were sold by Treasury, but the outstanding amount of people long the issue will be many times X because of all those repo leveraged buyers of UST’s, and for every one of those repoed longs, there is a short on the other side doing reverse repo. The market clears with the repo rate, which can often be much lower than fed funds and indeed can go up to -300bps at times if the (primarily hedging) demand from shorts is extreme. (The effective repo rate in this market is rather different from the general collateral series you can pull from public sources.. it’s hard to get good data as it’s proprietary to the big IDB’s… why the Fed tolerates this degree of opacity, I’ve never understood.)
Treasuries can therefore be seen as a special financial “currency”, and the treasury market can be modeled as type of free banking regime, where the public debt is base money, the much larger qty of leveraged UST positions is broad money, and the repo market is an interbank lending market where USD cash is collateral instead of money.
Looked at this way, the phrase “shadow banking system” is a quite literal description. Turn a market monetarist lose in this parallel universe, and the low rate conundrum is due to UST “base money” not keeping up with demand and the Treasury is a tight fisted CB.
In this universe, the real return of treasuries isn’t the relevant variable, it’s the spread between the repo rate and the treasury yield, which acts as a sort of “fee” for the guy who wants a hedged Investment in a riskier asset and pari passu a benefit to the party who wants a leveraged bet that the Fed means what it says about ZIRP. In finance-land with its UST currency, that spread /is/ the ST interest rate, which is volatile and well-above zero.
Now we can define quite precisely your “entry fee” thesis: the entry fee is the relative credit terms (haircut’s, etc) you’ll get in this repo market. In a world of only non-bank dealers and traders, those terms are symmetrical b/c counter-party risk is broadly symmetrical. In the world of TBTF, naturally only the bank holdco’s get the best terms. So, to win the wealth-accumulation game in this world, be a bank or be a very good client of a bank.
Ponder at your leisure!