the current crisis stems from (flawed) efforts to construct safe assets out of risky assets in order to meet a surge in investor demand for safe asset

From Brad Setser:

"Central banks were not the big buyers of synthetic triple AAA …

Posted on Wednesday, January 28th, 2009

By bsetser

Ricardo Caballero argues that the current crisis stems from (flawed) efforts to construct safe assets out of risky assets in order to meet a surge in investor demand for safe assets.

He is on to something.

There was a surge in demand for safe assets that pushed yields on Treasuries (and Bunds, OATS and Gilts) down at the peak of the boom. And investment banks – with help from the rating agencies — did respond by constructing new kinds of products that combined higher yields than Treasuries (or Agencies) and the appearance of safety.

Caballero thinks the banks constructed these securities to meet demand from central banks and sovereign funds.

Global demand for assets was particularly for very safe assets – assets with AAA credit ratings. This is not surprising in light of the importance of central banks and sovereign wealth funds in creating this high demand for assets. Moreover, this trend toward safety became even more pronounced after the NASDAQ crash. Soon enough, US banks found a “solution” to this mismatch between the demand for safe assets and the expansion of supply through the creation of risky subprime assets – the market moved to create synthetic AAA instruments. …. The AAA tranches so created were held by the non-levered sector of the world economy, including central banks, sovereign wealth funds, pension funds, etc. They were also held by a segment of the highly-levered sector, especially foreign banks and domestic banks that kept them on their books, directly and indirectly, as they provided attractive “safe” yields.

I don’t think that is quite right. Central banks and sovereign funds weren’t the main buyers of “structures” that produced a “synthetic” triple AAA credit Central banks generally stuck to instruments with cleaner cash flows and less risk. Many have mandates that require that they invest in fairly short-term instruments that have explicit government backing.

Let’s review the portfolios of the key players:

– Japan (MoF mostly) remains overwhelmingly in Treasuries, with a small number of Agencies. As far as I know, that is about as far as the MoF and BoJ went.
– Brazil’s central bank basically holds Treasuries.
– Russia’s central bank held a lot of short-term Agencies. But it liked the Agencies own debt, not the mortgage backed securities they guaranteed. Russia’s its investment guidelines are quite restrictive. It needed a government guarantee. It wasn’t buying any “private lablel” structured product.
– India has been quite conservative. It doesn’t even hold many securities. Most of its reserves are on deposit at an international institution — likely the BIS. Who knows, maybe the BIS dabbled in structures, but India wasn’t a big source of demand for CDOs.
– Mexico didn’t go further than dabbling with Agencies. Agency MBS were a bid deal. It is mostly in Treasuries.
– Saudi Arabia is a bit of a mystery. I doubt it holds any structures on its own book. But it also likely outsourced the management of a fraction of its fixed income portfolio, and well, maybe the managers had a mandate that allowed them to take a bit of risk. I just don’t know. The Gulf’s sovereign funds liked equities (and private equity) more than they liked complex debt instruments – though they probably invested in hedge funds that made levered bets on complexity. I would be surprised if the Gulf has more than $100 billion of exposure to structured products. the Gulf’s big bet was on the equity market.
– Korea clearly took a few more risks back when it wanted to juice returns to offset the won’s appreciation (how the world has changed). But even if it put a third of its dollar portfolio (say 20% of its total portfolio) in dollar bonds that lacked government banking, it wouldn’t have bought more than $50 billion of “product.”
– Norway also took credit risk with its bond portfolio. But the US only made up about a third of its bond portfolio. Back when bonds were 60% of Norway’s portfolio, Norway could not have have had more than 20% of a $300 billion portfolio in US bonds of all kinds. It thus could not have accounted for more than about $60b of demand for various structures. And it clearly wasn’t just buying structures.

Sum those countries up and they account for a large share of total central bank assets: Japan has a trillion portfolio, the big Gulf sovereign funds probably combined to hold around $800 billion of assets at their peak (less now); Russia and Saudi Arabia were around $500 billion; Norway, India, Brazil and Korea were in the $200-300 billion range.

That leaves China. China clearly dabbled a bit with some kinds of risk. The state banks got $100 billion or so to play with in 2006 (they seem to have gotten burned and retreated; they have been selling their foreign portfolio since mid 2007). SAFE bought a few high quality corporate bonds. It was a big buyer of Agency MBS – and perhaps bought a few structures too. But the overwhelming majority of its assets remained in Treasuries and Agencies not in “structures.’ China, inc – counting the state banks – might have bought $300 billion of bonds that lacked an explicit or implicit government guarantee, but probably not more.

The numbers just don’t work. Unless my accounting is way off, most central bank assets remained in fairly safe assets.* The big shift in the world of reserve management was the shift from Treasuries to Agencies after 20004. Most central banks still wanted bond with explicit or implicit government guarantees.

Who then were the investors who wanted “safe” – i.e. highly rated – structures? Saturday’s Wall Street Journal provides the needed clue. Think State Street. Actually, think conduits run by State Street. Levitz and Anand in Saturday’s Wall Street Journal:

“It got into conduits, which are instruments that buy such things as asset-backed and mortgage backed securities, using short-term borrowings. It shifted its investment portfolio from predominantly government bonds into mortgage-backed securities which rose the housing boom. And then last year the housing market fell apart and ensnared the financial world – and State Street – in a credit crunch.

I don’t think State Street was alone.

As central bank demand pushed yields on Treasuries and Agencies down (and as the fed raised rates), a host of financial institutions took on more risk. Other financial institutions supplied the product that they wanted. And when it blew up, the financial sector – not the world’s central banks – was left with big losses.

Caballero’s argument works, but with one additional step. Central bank demand pushed yields on Treasuries and Agencies down, and low yields on traditional, safe assets triggered a surge in demand for assets that appeared safe but offered the kinds of returns private investors were used to.

Lord Turner of the UK’s Financial Services Authority:

Very low medium- and long-term real interest rates have in turn driven two effects:

First, they have helped drive rapid growth of credit extension in some developed countries, particularly in the US and the UK – and particularly but not exclusively for residential mortgages – with this growth accompanied by a degradation of credit standards, and fuelling property price booms which for a time made those lower credit standards appear costless. And secondly, they had driven among investors a ferocious search for yield – a desire among any investor who wishes to invest in bond-like instruments to gain as much as possible spread above the risk-free rate, to offset at least partially the declining risk-free rate. Twenty years ago a pension fund or insurance company selling annuities could invest at 3.5% real yield to maturity on an entirely risk-free basis; now only 1.5%: any products which appear to add 10, 20 or 30 basis points to that yield without adding too much risk look very attractive.

… The fundamental macro economic imbalances have thus stimulated demands which have been met by a wave of financial innovation, focused on the origination, packaging, trading and distribution of securitised credit instruments.

Incidentally, One implication of Caballero’s argument is that the US should have been running bigger budget deficits to meet the rise in demand for safe assets. That would have kept yields up and reduced incentives to create “synthetic” triple AAA. Rather than following Dr. Chinn’s advice and trying to bring about rebalancing with a tight fiscal policy, the US should have met the world’s growing demand for truly safe reserve assets by running large fiscal and current account deficits.

Who knows - that could be where we are heading.

Caballero attributes the surge in demand for safe assets to the integration of the emerging world into the global economy. He writes:

“since the late 1990s, the world has experienced a chronic shortage of financial assets to store value. The reasons behind this shortage are varied. They include the rise in savings needs by aging populations in Japan and Europe, the fast growth and global integration of high-saving economies, the precautionary response of emerging markets to earlier financial crises, and the intertemporal smoothing of commodity producing economies. The immediate consequence of the high demand for store-of-value instruments was a sustained decline in real interest rates.

I would add two additional factors.

China’s integration into the global economy coincided with a huge rise in China’s savings rate. There is a world of difference between a China that saves and investing a constant 35% of its GDP and a China that goes from saving and investing 35% of its GDP to savings 50% (or more) of GDP and investing 40% (or more) as its economic size increases. In the first case China isn’t a net lender to the world. In the second case it is. China’s savings rate hasn’t been constant, and China matters.

And exchange rate management. China didn’t buy a lot of US bonds just to protect itself from a repeat of the Asian crisis. it had (and has) a host of controls that limit its exposure to a big swing in interbank flows. its reserve cover has been exceptionally high for a long time. China bought a lot of bonds because it didn’t want the RMB to rise even as China’s booming exports created natural pressure for appreciation. It is hard to separate the surge in demand for safe assets from the rise in Chinese reserves growth (counting the increase in hidden reserves) from under $50 billion a year to close to $700 billion a year (at its peak — it is now lower). Especially as that surge came at the same time that rising oil prices pushed up oil savings and the demand for reserve assets from the oil exporting economies …

Had China allowed its currency to rise in 2004 rather than tightening fiscal policy and limiting lending to avoid inflation, I rather suspect that Chinese demand for safe US and European financial assets would have become demand for US and European goods. That would have produced a more balanced – and ultimately less risky – global economy.

And, well, if the US and UK governments hadn’t looked the other was as leverage in the financial sector rose — as financial institutions made bigger bets to keep profits up as spreads fell — that too would have produced a more balanced and ultimately less risky global global economy.

* Many questions about the global economy could be answered more definitively if central banks disclosed more information about their portfolios to the IMF. Aggregate data on central banks equity holdings would be useful — as would aggregate data about the composition of their fixed income portfolio. Of course, getting more countries just to provide data on the currency composition of their reserves would be a nice start …"

Your man about blogs:

January 28th, 2009 at 7:35 pm
1. Don the libertarian Democrat responds:
   

   “Most central banks still wanted bond with explicit or implicit government guarantees.”

   You clearly see the importance of implicit and explicit guarantees. How did this relate to US Banks? They did not seek less risk. CDSs and CDOs allow lower capital standards and hence more risk. My understanding is that the Senior tranches allowed the profit on the riskier tranches, and allowed for hefty fees from packaging. What was needed was a AAA rating in order to make more money on the other tranches. It allowed more leverage. The safety that you are claiming people were seeking was a method of lowering capital requirements.

   “a surge in demand for assets that appeared safe but offered the kinds of returns private investors were used to.”

   Maybe I’m missing something, but you seem to be saying that investors were use to, say, 7% returns. When interest rates declined, these investors started getting, say, 3%. So now they start looking for 7% returns. Of course, they could buy riskier bonds which pay what they used to get. Instead, investments are created which appear less risky and pay, say, 7 % , but they’re as risky in fact as the riskier bonds our investors passed up. This is fraud, which I believe it was. You seem to be assuming that the creators of these assets weren’t aware of what they were doing, which looks almost alchemical instead of scientific when you actually look at the risk of these new investments.

   “they had driven among investors a ferocious search for yield – a desire among any investor who wishes to invest in bond-like instruments to gain as much as possible spread above the risk-free rate, to offset at least partially the declining risk-free rate.”

   “… The fundamental macro economic imbalances have thus stimulated demands which have been met by a wave of financial innovation, focused on the origination, packaging, trading and distribution of securitised credit instruments.”

   This makes it sound like an engine. In fact, people were misled into thinking that alchemy was science. My scenario is that the CDSs and CDOs were used to make more money by requiring lower capital standards and hefty packaging fees. You’re saying that the buyers were investors who thought that they were getting a higher yield for free.

   The demand was then for something that didn’t exist, and the investments were sold as something that they weren’t. Again, that’s fraud. There’s no mechanical connection between low interests rates and poor investing. It takes actual people to make these poor decisions.

   Finally, back to the guarantees. The US banks indulged in this behavior, not because they didn’t know it was risky, but that they believed that there was an implicit guarantee by the government to intervene in an economic crisis. The guarantees that foreign banks were seeking with their investments, our banks believed had already been guaranteed by lobbying and collusion and precedent. They were right.

   The search for safe investments can only lead to tragedy when the investments are not in fact safe.

And a response to a comment:

Twofish,

CDOs were used because you could use the senior tranche’s rating to get lower capital requirements for the riskier tranches. That’s how they made more money. They locked up less capital. They also charged for fees and made a bit more more on the risk in the split up into tranches.The benefit and money from them was not made on the safest assets. I understand that is what is being claimed. Many banks kept the safer tranches for themselves, and sold the riskier ones, because that’s where the money was.It was not a flight to safety that led to CDOs. CDOs were used precisely because they lowered capital requirements, which is inherently riskier.The flight to safety occurred in the crisis.

As to my example, it was simply to show that people wanting the same old returns needed to increase their risk. They could not, in any real world, expect the same old returns for the same risk. My point stands.

Thanks for the response,

Don

"But as far as ABS go? Should get that market rolling again."

Accrued Interest on the new TALF. Read all the posts, because, even though I didn't like this plan, a very good case is made for it:

"The Fed has expanded the Term Asset-Backed Loan Facility (TALF), which AI first discussed here. Here is the quick recap of the facility.

1) Fed will loan funds for purchase of recently issued ABS( ASSET BACKED SECURITIES ). This was clarified to mean ABS issued after January 1, 2009 made up of loans no older than October 2007. The ABS must be rated AAA( WHATEVER THAT MEANS ), and be made up of student loans, auto loans, small business loans, or credit cards.

2) Loans will be non-recourse and not marked-to-market. The borrower will not have to deal with margin calls due to price declines ( INTERESTING. NO NEED FOR EXTRA CAPITAL INFUSIONS ).

3) The loan term will be up to 3-years, originally was only 1 year( THAT'S TRUE ). That is extremely positive for the potential success of this program. See below.

4) The loan rate will be set at "yield spreads higher than in more normal market conditions but lower than in the highly illiquid market conditions that have prevailed during the recent credit market turmoil." In other words, lower than the rate paid on the asset( GOOD DEAL ).

So what has the Fed done here? Created an easy arbitrage. All investors have to do is do accurate credit work, and this is a guaranteed profit. Note that the 3-year term seals this thing. 3-years is basically the entire life span of most eligible collateral, so it eliminates the last thing an investor needed to worry about. Given a 1-year term, investors would have worried that the end of 1-year, new financing might not be available( ESPECIALLY A YEAR FROM NOW. IT'S MEANT TO OUTLIVE THIS DOWNTURN ). But by the end of 3-years, the asset will be all but gone.

Also through this facility, the Fed can really control consumer lending rates. The rate on newly issued AAA ABS will be stuck at a level slightly higher than the Fed's lending rate. Banks which are currently hoarding cash will fall over themselves to buy ABS and pledge them into this facility. ( INTERESTING )

Now don't read this as especially bullish for the overall economy. I still see this as a facility intended to aide in quantitative easing, and not a "fix" for the recession. Or put another way, a means of preventing the economy from getting still worse. But as far as ABS go? Should get that market rolling again."

Here's the original post:

"The Fed's new Term ABS Loan Facility (TALF) announced this week could be a significant step in improving credit availability. While many of the details of the program are not yet known, there is already several take aways.

First, this looks and smells a lot like a back-door way of reviving some of the TARP's original concept. Consider what we already know about the program. Eligible collateral for the TALF will basically include AAA-rated bonds within the major non-housing ABS sectors: auto loans, student loans, credit cards, and SBA loans. TALF loans will have a one-year term and will be non-recourse to the borrower. The facility appears to be oriented toward banks and insurance companies, but may actually be available to anyone. TALF loans "will no be subject to mark-to-market or re-margining" which is a critical part of the program.

Now put these criteria together and consider the effect. A bank may originate loans of the above types, then get funding from the Fed at an attractive rate. There is no need to worry about the funding being taken away suddenly because of changing haircuts, nor is there any worry about interim marks impacting economic results. The originator does have an incentive to make a good loan, since the Fed is going to require some haircut. But as long as the originator can make good loans, the eventual profit will be the differential between the lending rate and the Fed borrowing rate.( A GOOD DEAL )

Let's look at a real life example. COMET 2008-A6 A6 is a credit card ABS issued in May. The original deal spread was +110bps over 1-month LIBOR with a 2.4 year average life. Currently bonds of this type are trading with a spread of around 600bps, which makes the dollar price of this bond around $89.

Analyzing asset-backed bonds gets complicated because bond holders get monthly principal and interest payments. But in simple terms, the bond is yielding LIBOR +600bps. If the Fed is willing to lend at LIBOR +50 or 100bps, banks will quickly gobble up high quality ABS paper. As a result, the yield spread on this kind of ABS will contract until its closer to the Fed's lending rate. If the COMET bond were to go from LIBOR +600 to LIBOR +300, the bond's price would appreciate by 5.5 points.

There would be two important knock-on effects. First, it would create a price floor for similar ABS which isn't pledged into a Fed facility, alleviating mark-to-market problems banks are currently facing( WHEN PRICES GO DOWN, THEY MUST RAISE CAPITAL ). Second, it will allow for new origination in ABS, which will help rejuvenate consumer credit.

The primary beneficiary will be the ABS securities itself. Next would probably be the bigger holders of ABS paper, which include banks and P&C insurers. Companies involved in securitization will also benefit: credit card issuers like Capital One and student lenders like Sallie Mae. There is already talk that this program could be extended to Commercial MBS, which would benefit REITs tremendously."

And another post:

"If the covered bond idea is dead, for now anyway, perhaps the ABS market can pick up the slack.

Historically, ABS have typically been backed by consumer loans, including credit cards, auto loans, home equity, and student loans. ABS were typically structured with a senior/subordinate credit enhancement, meaning that certain tranches of the deal would take losses first and only once those tranches were wiped out would other tranches take a hit.

Of course, there have been numerous problems with the ratings agencies allowing too little in subordination in certain deals. But there is nothing inherently wrong with the senior/sub concept. In fact, if its kept as a simple sequential loss structure, analyzing the credit of an ABS deal becomes relatively straight forward: its just losses versus available subordination. Sounds a hell of a lot more transparent than trying to decode a bank's balance sheet! ( I AGREE )

So what if the ABS market could be revived? Lenders who could not access the unsecured debt markets could access the ABS markets, raising loanable funds. If the lender also kept a sizeable residual on the deal, the result would be similar to the covered bond idea.

Many companies would benefit directly from an improved ABS market. Credit card issuers, such as American Express, Citigroup, and Capital One. Student lenders such as Sallie Mae. Even the autos would benefit, although obviously the GM and Ford situation is much deeper, Toyota and Honda would also benefit.

It wouldn't solve all our problems. I still wish they were buying mortgage assets. But this is better than nothing."

"He estimates that such a shock causes a sharp contraction for two quarters, which is then followed by abnormally high growth. "

Another interesting post on the FT:

"Normality is just a few policy steps away

December 18, 2008

by FT

By Ricardo Caballero

Economic agents of all sorts, from creditors to consumers, are frozen waiting for some sense of normality to be restored amid the financial crisis. However, normality is much closer —just a few bold policy steps away— than is the conventional wisdom. ( I AGREE )

The system we had before the crisis is not permanently broken, but it needs to be made more resilient to aggregate shocks, especially panic-driven ones. ( I AGREE )

I build my analysis and policy prescription on three premises and observations.

First, before the crisis the world economy had an excess demand for assets, especially AAA assets, and this will not change significantly once the crisis ends. ( FINE )

Second, and contrary to what investors thought at the peak of the boom, the (private) financial sector in the US is not able to satisfy this demand for AAA assets when large negative aggregate events take place. ( OK )

However, the US government does have the capacity to fill this gap, especially because it is the recipient of flight-to-quality capital, even when the core of the global financial crisis is located in the US. ( TRUE )

Third (and with the benefit of hindsight), the main policy mistakes were made during rather than before the crisis. ( DON'T AGREE )

These observations hint at a policy framework for the crisis and the medium run. For the latter, we can go back to a world not too different from the one we had before the crisis (real estate prices and construction sectors aside), as long as the government becomes the explicit insurer for generalised panic-risk. ( ISN'T THAT WHAT WE HAVE? )

That is, while monolines and other financial institutions can lever, their capital for the purpose of insuring microeconomic risk and moderate aggregate shocks, they cannot be the ones absorbing extreme, panic-driven, aggregate shocks. This must be acknowledged in advance, and paid for by the insured institutions. ( OK )

Reasonable concerns about transparency, complexity, and incentives can be built into the insurance premia. Collective deleverage, as being done, should not constitute the core response; macroeconomic insurance should. ( OK )

The structural policy framework for the medium run also carries over to the crisis-policy itself. The essence of a solid recovery should build not from deleveraging and a forced brutal contraction of the financial sector ( I'D LIKE TO AVOID THIS ), but from the explicit and systemic insurance provision against further negative aggregate shocks to their balance sheets caused by panic or predatory actions. ( I STILL SAY THAT WE HAVE THAT )

The recent intervention of Citi, with a mixture of (paid) insurance and capital, is promising, and so is the second intervention of AIG (see the recent FT-forum articles by Caballero and Krishnamurthy, and Kotlikoff and Merheling for similar assessments). ( HOW'S THAT? )

These interventions need to be scaled up to the whole financial system (banks and beyond), and it is better to do it all at once, for in this case the likelihood of the government ever having to disburse funds for its insurance provision becomes negligible. ( WE HAVE THAT )

The good side of panic-driven contractions (as opposed to those driven by more structural factors) is that the potential for a strong recovery is always around the corner. ( TRUE )

Although the current crisis has already caused enough collateral damage to add persistence to the recession, there are still plenty of resources waiting on the side to make a sharp rebound possible. ( I AGREE )

I do not mean to say that this recession is an imaginary one. On the contrary, I believe it is a very serious recession. My point is simply that good policy has an opportunity to bring the recession back to familiar turf by defeating the extra gloom ( I AGREE ), and if this happens, the recession will become a manageable one from which current asset prices, on average, will look like once-in-a-lifetime deals ( I AGREE ).

Along the ideal recovery path described earlier, the real interest rate would remain at record low levels for a long time; risk-spreads and the VIX index (the Chicago Board Options Exchange Volatility Index, known as Wall Street’s “fear gauge”,) would decline gradually but consistently; asset prices and financial leverage would rise rapidly; the yen and dollar would depreciate vis-á-vis most other currencies, helping net exports in Japan and the US; commodity prices would recover but not to record levels. ( SOUNDS GOOD )

In addition, non-residential investment, inventory accumulation, and durable expenditures would snap back, joining and leveraging on the fiscal and monetary expansions; global imbalances would stabilise and build back a bit; unemployment would peak at single digit levels and then begin to turn around; and inflation would rise only gradually in the developed world, creating the needed space for a recovery consolidation. ( WOULD WORLD PEACE ENSUE AS WELL? )

There is no way out of a dreadful last quarter of 2008 and well into the first quarter of 2009. But the big difference with the consensus forecast is in the sharp recovery after that. ( I AGREE )

The source of this difference is in the assessment of the dominant nature of the recession. Slow recoveries follow the typical credit crunch, as financial resources have to rebuild for growth to resume.

But while I think this was the nature of the mild recession preceding the events at stricken insurer AIG, and Lehman, the collapsed investment bank, the dominant recession now is very different in nature. ( I AGREE )

It is a systemic run on all forms of explicit and implicit insurance contracts ( THAT'S IT ), but with no shortage of resources on the side. If confidence recovers, the resources to support the recovery are abundant and ready ( I AGREE ).

Nick Bloom, assistant professor of economics at Stanford University, provides the best available evidence of how an economy is likely to react to a temporary bout of volatility.

He estimates that such a shock causes a sharp contraction for two quarters, which is then followed by abnormally high growth. I think this is the correct way to view the current recession, as long as bold policy actions are undertaken. ( OK )

Of course many things can go wrong to cause a disastrous outcome, but enough has been written about these negative scenarios. It is time to, at the very least, begin to sketch what the good scenarios may look like.

I believe that we Implicitly have the guarantees he's asking for.