Simon Wren-Lewis weighed in on the safe assets discussion last week - his post is here and I strongly recommend a read even though it is definitely wonkish (I had to read it twice). Simon's discussion of collateral effects on interest rates adds a dimension that has so far been missing from the debate and helps to explain why long-term, as well as short-term, debt is needed by the financial system - a point that Pozsar misses. Long-term government debt can be used as collateral in the creation of shorter-term private sector "safe assets", which reduces the need for short-term government debt as collateral.
Simon suggests that government depositing cash from excess government borrowing at the central bank, as I suggested, would not be tenable. He doesn't explain why, so I thought I would have a go.
When you consider the central bank and treasury as a consolidated entity, the cash balances at the central bank and the borrowing net out and are eliminated. Neither the excess borrowing nor the uninvested cash show up on the consolidated balance sheet. So it can't be a consolidated balance sheet risk problem. However, as I've noted before, just because something is eliminated in accounting terms doesn't mean it disappears in reality. Unconsolidated, the government has debt, and the central bank has cash.
The fact that the government's excess cash is lodged safely at the central bank means that default risk is zero. Central banks can create money ex nihilo, so there is absolutely no possibility that the money could not be returned. From the point of view of the government's creditors - or depositors, as I would prefer to call them - this makes the debt completely risk-free, and therefore should mean that the interest rate on that debt is very nearly zero. But from the point of view of the central bank, it has serious implications for the conduct of monetary policy.
Intuitively it seems reasonable to suppose that the Treasury depositing excess cash at the central bank would cause a huge increase in the monetary base. But I've done a bit of research into the makeup of the monetary base, and it seems that central deposits are excluded from the monetary base calculation. In which case, in theory the Treasury could deposit as much cash as it wished and there would be no effect on reserves at all. However.....that's not quite right. I looked up exactly how the US Federal Government conducts an auction and receives payment. The NY Fed produces an extremely helpful document explaining how it works.
When the Fed auctions bills or bonds, payment is nearly always made via commercial banks which have reserve accounts at the Fed. The only exception to this would be an individual who wrote a cheque directly to the Treasury, in which case that cheque would be credited directly to the Treasury's account at the Fed and there would be no impact on reserves. All other payments go through reserve accounts: the reserve accounts are drained by the amount of the debt purchased, and the Treasury's deposit account is credited. Therefore Treasury debt issuance drains reserves. Debt issuance on the scale required to meet the financial system's need for safe assets would drain reserves by a very considerable amount. In effect it would be a huge monetary tightening. It would undo the effects of all the QE the Fed has conducted.
There are a few things the Fed could do to undo the contractionary effect on the monetary reserve of all that debt issuance:
1) Cut interest rates. Er, no - at the moment it can't. Or not very much, anyway. I don't wish to get into the negative rates argument again, but the zero bound is binding in an economy that still uses cash and I am personally unconvinced that negative rates would necessarily be expansionary.
2) Buy up the debt the Treasury has just issued. That would completely defeat the purpose of issuing it, which is to provide safe assets to the financial system.
3) Buy private sector risky assets. This could tie in nicely with Simon's idea that the government could build up a portfolio of risky assets, but I doubt if he meant it to be the central bank that bought them. It would mean the central bank conducting quasi-fiscal operations (taking stakes in public enterprises) and accepting credit risk. And it is questionable whether the central bank has the expertise to manage such a portfolio.
I don't pretend to be an expert on central bank operations, so there may well be other options too. All of these are things that the Fed could do by itself. But they are the tools that it uses to conduct monetary policy, not to sterilize the effect of Treasury actions. It would seem more appropriate for the actions to correct the reserve drain to be taken by the Treasury. As far as I can see there are two options for the Government:
1) Allow the Treasury deposit to be redistributed to commercial bank reserve accounts. This would ensure that debt sales didn't affect the reserve level, which in a world where debt issuance has nothing to do with financing government spending and everything to do with preventing the financial system from seizing up just might be a good thing. There seems little point in simply exchanging one type of safe asset for another, which would be the effect of that reserve drain. I'm not entirely sure how this would be done, but it seems likely that the settlement accounting would have to be reversed, leaving the balance on the Treasury account at zero and the in/out entries as memorandum items. As reserves never leave the banking system, this action would sterilize the impact of the debt issuance without placing the debt itself at risk.
(UPDATE - It seems the Bank of Canada already has such a facility. From its explanatory notes on the composition of the Bank of Canada's balance sheet (my emphasis):
3. Government of Canada deposits. The government keeps deposits at the Bank of
Canada, against which it writes all its cheques. The government also maintains
deposit accounts with direct clearers. The maintenance of these government accounts
with the Bank of Canada and the direct clearers gives the Bank of Canada an additional instrument of monetary control, called government deposit transfers.)
2) Spend the money. Back to Simon's idea of buying up risky assets, I guess. At least this would be done by the Government rather than the central bank, which at least preserves the fiction of central bank independence for a little longer.
From a safe asset perspective, I am unconvinced by Simon's suggestion that the Government should build up stocks of risky private sector assets - although from an economic point of view it has considerable merits. Yes, the fact that the central bank can always monetize debt, and government can always tax, would mitigate the balance sheet risks arising from such a strategy. But they would not be zero. Government would pay higher borrowing costs, I think. And if the risk asset portfolio suffered losses, there would be a risk of knock-on impact to the value of government debt due to resurrected default fears - exactly what we DON'T want for safe assets. As BIS said, the whole point of a safe asset is that there should be no possibility of it becoming unsafe.
Miles Kimball has suggested that government investment might be separated from government itself through the creation of a sovereign wealth fund, which could be managed by private sector investment experts. This would eliminate the "government is rubbish at picking winners" problem that is always raised whenever anyone suggests government should invest in private sector enterprises, and it would hopefully mitigate the real risk of rent-seeking and capture by special interests. But government would still be the ultimate guarantor for such a fund's debt, so it would still rest on the credibility of the government/central bank nexus. Investors may prefer explicit backing for government debt from the central bank via excess cash balances. In which case the best option is option 1 - redistribute the balance on the Treasury account.
However, due to QE we are already a very long way down this road: central bank independence is already being challenged, and coordination of monetary and fiscal policy is being recommended by some rather significant players, notably the NY Fed, which is visibly uncomfortable with the USA's political paralysis and the pressure that it is under to deliver more than monetary policy reasonably can by itself in a liquidity trap. The end of central bank independence is being promoted by central banks themselves.
The stability of any economy and the safety of its liabilities depend on the economic management team working effectively to achieve desired outcomes and deal effectively with problems. The independence of the central bank and separation of monetary and fiscal policy worked well in the past, but the crisis of 2008 and subsequent recession fundamentally changed the economic landscape. The conduct of monetary and fiscal policy in Western economies is no longer the same, and therefore the roles of monetary and fiscal authorities must change. As monetary and fiscal policy become ever more blurred, cooperation not separation becomes the name of the game. Central bank independence needs to be consigned to history.
Safe assets and government debt - mainly macro
When governments become banks - Coppola Comment
Government debt isn't what you think it is - Coppola Comment
Era of independent central banks is over - Stephen King, HSBC
Helicopter Money - McCulley & Pozsar
Institutional Cash Pools and the Triffin Dilemma of the US banking system - Pozsar (IMF)
Why the US needs its own sovereign wealth fund - Quartz
The Treasury auction process: objectives, structure and recent adaptations - NY Fed (2005)