Commentaries

Oh, where has all the convexity gone?

The rise in U.S. Treasury yields has been very impressive considering where stocks are – the Dow is just 80 points away from 10,000 – and the improvement in economic data.  But it’s even more incredible that it happened without the aid of investors in mortgage-backed securities and mortgage servicers that typically snap up longer-dated debt like U.S. Treasuries and swaps to hedge their portfolios when interest rates fall sharply. It’s known as convexity hedging, and it was a powerful accelerator in the U.S. Treasuries market in 2002 when the Federal Reserve was pushing rates down. (It also works the other way. When rates rise suddenly, it forces mortgage investors to quickly start selling longer-dated debt.)

Deutsche Bank in a recent note says it’s been largely absent this go around, largely due to government intervention. This is important because the Fed’s exit from the agency mortgage market also means this $4.5 trillion market is likely to re-assert its influence over benchmark Treasuries. If yields are rising by then, such MBS freedom could accelerate the move. And that has an impact on mortgage rates and any other debt using Treasuries as a benchmark.

Here are DB’s reasons:

1) The Federal Reserve is the dominant force in the market. It owns 23% of the outstanding 30-yr agency market, and guess what, it doesn’t need to hedge.

2) It’s impossible to tell how much investors should be hedging since the Fed’s intervention has skewed prices too much to figure out how much risk is out there.

3) Estimates of convexity and duration risk (which is determined by when you think borrowers will refinance their mortgages) have become increasingly difficult to pinpoint since the government’s Home Affordable Refinance Program hasn’t caused a spike in prepayments yet. So those that would typically hedge are likely hunkering down.

4) Technical reasons that involve short-dated and long-dated volatilities, which I’ll leave to the experts to explain:

On MBS, Fed needs to point to the exit

When the medication is flowing, it’s hard to see straight.

Amid the giddiness in the markets and the cheers for the end of the recession, what often gets ignored is the fact that government stimulus is still fueling the reflation of financial markets.

Yes, the U.S. government has started to retire some programs — its backing of money market funds being the most recent. But there’s still a question mark about how it plans to wind down one of its largest supports — its $1.25 trillion mortgage-backed securities purchase plan — that is due to expire at the end of the year.

That’s dangerous, since a bungled hand-over of the market back to the private sector could derail a still fragile housing market.

This week, Ben Bernanke and his colleagues on the policy-making Federal Open Market Committee are sure to discuss how best to wind down its purchases of mortgage bonds guaranteed by state-run Fannie Mae and Freddie Mac. Some officials have already started to debate publicly whether they should pull the plug on the program before it reaches its $1.25 trillion limit.

But they shouldn’t wait to decide until November, as some now expect. They need to prepare investors elbowed aside by the government intervention. They should do this by laying the groundwork for an eventual departure, to avoid sudden spikes in mortgage rates that have been kept artificially low this year.

Unlike other initiatives, such as the Federal Deposit Insurance Corp’s insurance of bank debt, and the various lending facilities to nudge investors back into areas that had gone haywire during the financial crisis, the Fed’s direct purchases of MBS has done the opposite — it has squeezed investors out of the market.

Barclays risky assets move a little too cozy

Barclays has come up with an interesting way to solve an optical problem. Concerned that the bank’s shareholders are nervous about possible future writedowns of wobbly assets with a value of $12.3 billion, it has sold them to its own employees.

This isn’t necessarily a bad idea. But there are two things to dislike about this deal. First, it looks pretty cozy to sell to your own workers. And second, the deal looks potentially very favourable for the purchasers.

The deal does not remove the assets from Barclays’ balance sheet. What it does is allow the bank to pull them out of its mark-to-market book, where their carrying value is contingent upon the financial health of some monolines with whom Barclays has taken out credit insurance.

To do this it makes a loan to an entity, which then buys these assets. The loan still sits on Barclays’ books but does not have to marked to market. Even so its value is ultimately still tied to the performance of the assets.

This, the bank argues, will allow the shareholders to sleep easily at nights, knowing that a credit downgrade at some obscure monoline will no longer bring writedowns crashing down upon their heads.

This may seem fair enough. But to achieve this optically pleasing outcome, the bank has cut a deal that offers real upside for Protium Finance — the entity that has purchased the assets — and a group of its own ex-employees that will manage them.

Consider the terms of the deal that Barclays has disclosed.

Nerdy thought on the Fed balance sheet

Looking quickly at the Fed balance sheet, I stumbled upon the “off balance sheet” quirk of its mortgage-backed security holdings. The Fed reports this week that its holdings through Wednesday Sept. 2 stand at $625 billion. But we know from the NY Fed data released yesterday that the central bank has bought $817.6 billion MBS so far this year.

The discrepancy, which I had forgotten but a kind source reminded me of, is because the Fed is buying mortgage pass-throughs before they settle, those purchases won’t show up right away. Here is the table that shows there are $164.7 billion MBS essentially off balance sheet. So there’s still a whole lot more coming onto the Fed’s balance sheet, even if they stopped purchasing MBS tomorrow.

Tidbits from the FOMC minutes…

Just going through the FOMC minutes now and there were a couple of interesting bits worth flagging:

Meeting participants again discussed the merits of including agency MBS backed by adjustable-rate mortgages (ARMs) in the Committee’s MBS purchase program: Some thought it would be useful to include agency ARM MBS, noting that doing so could reduce the unusually large spreads between ARM rates and yields on similar-duration Treasury securities—spreads that were far larger than the comparable spreads on fixed-rate mortgages; others saw little potential benefit, given the small stock and limited issuance of ARM MBS, and were hesitant to involve the Federal Reserve in another market segment. The Committee made no decision on purchasing ARM MBS at this meeting.

It just seems odd that they would be discussing expanding the MBS purchasing program at all when the debate seems to be hinging on whether it’s time to think about pulling back on it. See my posting on it here. Also, not sure they want to be in the business of stimulating riskier segments of the mortgage market.

Participants also discussed the merits of progressively reducing the pace at which the Federal Reserve buys Treasury securities, agency debt, and agency MBS prior to the end of the asset purchase programs. They generally were of the view that gradually slowing the pace of the Committee’s purchases of $300 billion of Treasury securities and extending their completion to the end of October could help promote a smooth transition in markets. A number of participants noted that a similar tapering of agency debt and MBS purchases could be helpful in the future as those programs approach completion. The Committee made no decisions on tapering those purchases at this meeting.

That seems to indicate some are leaning toward keeping the dollar amount, $1.25 trillion for agency MBS and $200 billion agency debt, steady but stretching it out into 2010. The Treasury purchases were originally slated to stop in September.

It also put forth three strategies to mop up all the liquidity it’s pumped into the system.

These measures include executing reverse repurchase agreements on a large scale, potentially with counterparties other than the primary dealers; implementing a term deposit facility that would be available to depository institutions in order to reduce the supply of excess reserves; and taking steps to tighten the link between the interest rate paid on reserve balances held at the Federal Reserve Banks and the federal funds rate. Several participants noted the need to continue refining the Committee’s strategy for an eventual withdrawal of policy accommodation.

To buy, or not to buy MBS

It looks like the lines are being drawn within the Fed regarding its massive $1.25 trillion MBS asset purchase plan that’s due to expire at the end of the year.

New York Fed President William Dudley told CNBC earlier Monday that it’s too early to think about pulling back on these programs, and points to market expectations as a big reason the Fed should proceed carefully. The market expects the Fed to buy the full amount and is currently trying to figure out whether there’s a possibility the Fed will extend the program into next year to make for a smoother transition.

These purchases, if completed, will account for roughly a quarter of the outstanding MBS market. Without the Fed’s support, risk premiums are expected to shoot higher (and with them mortgage rates) to lure back investors who have moved into other areas of the credit markets to find juicier yields.

Dudley’s remarks come after Richmond Fed President Jeffrey Lacker indicated that improving financial conditions mean the central bank may not need to purchase the full amount of MBS.

With the economy leveling out and beginning to grow again later this year, and with bank reserve demand ebbing as financial conditions improve, I will be evaluating carefully whether we need or want the additional stimulus that purchasing the full amount authorized under our agency mortgage-backed securities purchase program would provide.

Both Dudley and Lacker are voters on the policy-setting FOMC.

The liquidity canard

It’s often said on Wall Street that the more liquidity there is in a given market, the better things are for investors trading stocks, bonds or commodities. And while there’s a lot of truth to that, there are times when too much liquidity can be just the wrong tonic.

After all, Wall Street’s churning-out of one subprime-mortgage backed security after another pumped a lot of liquidity into the U.S. housing market, and that simply encouraged a lot of reckless — even fraudulent — lending.

That’s why I’m not impressed with the securities industry’s main defense of computer-driven high-frequency trading, which essentially is that all this lightning-fast trading provides liquidity and better prices for investors.

It’s a hard argument to swallow when you consider that many high-frequency trading programs are simply engaged in trading the same stock thousands of times a day in less than penny increments. Now maybe all those rapid-fire automated trades are getting better prices for some investors. But when a broker excessively buys and sells securities to generate higher commissions, it’s called churning, and that can result in an investor lawsuit or a regulatory sanction.

Indeed, when fast-fingered day traders were doing much the same thing as today’s high-frequency traders — albeit without the benefit of a sophisticated algorithmic program to guide them — Wall Street’s biggest firms were quick to dismiss them as either amateurs or rogues who were causing unnecessary volatility in the price of tech stocks.

So with critics raising legitimate concerns about the potential of a rogue algorithm sparking an unintentional market meltdown, the notion that high-frequency trading is OK because it creates more liquidity simply won’t wash.

If the main purpose of all that extra liquidity is to simply make fat profits for high-frequency traders at Goldman Sachs, UBS, GETCO, Citadel Investment Group and Interactive Brokers, that’s liquidity the markets can do without.

Fed MBS tally jumps to $766.6 billion

The Fed may be paring back its Treasury purchases, but its MBS program heated up this week. The central bank bought a cool $25 billion net, up nearly $5 billion from the previous week. Reuters puts the running tally now at $766.608 billion.

Running short on ammunition at the Fed?

Barring another serious economic stumble, it seems that the Fed is not going to offer much more credit easing than already planned. The minutes of the June meeting of the FOMC suggested a solid resistance to stepping up purchases of either mortgage backed securities or Treasury bonds.

On Treasuries, the committee argued that a small increase in Treasury purchases would do little to push down rates. Meanwhile a large increase could heighten fears that the Fed intended to monetize the government’s debt.

I have to admit to being a little disappointed. Even after completing its $300 billion of purchases the Fed will own just a small fraction of the more than $6 trillion market. They have enough credibility to push through more purchases without raising concerns. Another burst of sub-5 percent mortgage rates could provide much needed relief for the housing market and ease pressures on household budgets.

Under normal circumstances the Fed could afford to be cautious. But I suspect that the fiscal stimulus that has yet to percolate into the economy will have only minimal effect. Much will be offset by cutbacks at the state level.

This leaves the Fed on its own. They can afford to give the economy one last shot in the arm.

The government owns the MBS market

OK, it’s not a majority owner, but the government has an impressive stake in the $4.5 trillion agency mortgage-backed securities market.  Barclays Capital’s last count, as of July 3, puts Federal Reserve purchases at $621.6 billion since it launched the program in January.  Separately, the Treasury Department has picked up more than $145 billion since the government put Fannie Mae and Freddie Mac in receivership in September. The Treasury data is through the end of May.

The Federal Reserve has pledged to buy up to $1.25 trillion of mortgage bonds guaranteed by Fannie and Freddie and $200 billion of agency debt by year end.