The Safe Asset Meme

Date October 8, 2013
Speaker Richard PORTES(President, Centre for Economic Policy Research (CEPR))
Moderator Willem THORBECKE(Senior Fellow, RIETI)


Richard PORTES's PhotoRichard PORTES

There have been some sweeping assertions in recent times suggesting a global lack of safe assets. Such beliefs have been encouraged by the International Monetary Fund (IMF) and are backed by certain academics. This presentation will focus on discussing how the notion of a shortage of safe assets came about, and explain why this is actually misleading and why we shouldn't be concerned about it. It will also be argued that, to the extent that the demand for safe assets rises, supply will surely adapt to meet it.

Today's presentation will cover several key areas in relation to safe assets. One cause of the initial confusion regarding the concept of safe assets is the application of multiple meanings and definitions by different people. With this in mind, this presentation will discuss such definitions for the sake of clarity. Second, there is a strong claim of a shortage of safe assets. Ricardo Caballero of the Massachusetts Institute of Technology (MIT), possibly the most influential academic in this area, started to write about a shortage of safe assets as early as 2006. The second part of this presentation will look at this claim and offer detailed evidence in relation to it. Finally, a different version of claims in relation to safe assets from an international perspective and the related policy implications will be discussed.

Claims suggesting a lack of safe assets have many people concerned. The Financial Times has backed such a claim more than once. On top of this, the IMF's Global Financial Stability Report (GFSR) in April 2012 featured an entire chapter dedicated to this topic. Several excellent theorists have also developed models investigating the consequences of a shortage of safe assets and have expressed concerns for the future due to their belief in a rising demand and falling supply of safe assets. With so many respected people sharing this concern, it is clearly an issue worth investigating.

In discussing definitions, it would be useful to explain first what a meme is. Richard Dawkins first coined this expression in his book The Selfish Gene. The concept of a meme is essentially that of an idea, behavior, or style that spreads from person to person within a culture. This can be done through writing, speech, or other forms of communication. As an example, if there were an idea for which theorists created a model and was followed by students learning this model and applying it to their own studies, this idea could be called a meme. The notion of the existence of a shortage of safe assets could in itself be described as a meme. In fact, memes may survive, spread, mutate, and replicate effectively even when they have a detrimental effect. It could be argued that experts stating that there is a shortage of safe financial assets in the world acts as an agent for the propagation and mutation of such a detrimental meme.

In terms of defining safe assets, there appears to be a wide range of definitions in use. Gourchinas and Jeanne describe safe assets as cash, including insured deposits, plus any debt that is tradable, liquid, and which enjoys a top credit rating. They have also described safe assets as a liquid debt claim with negligible default risk. The GFSR described safe assets as meeting the criteria of low credit and market risks, high market liquidity, limited inflation risks, low exchange rate risks, and limited idiosyncratic risks. In 2012, Gorton described safe assets as those which are either directly or indirectly used in an information-insensitive fashion. The conclusion that could be drawn from these definitions is that no asset is actually perfectly safe. In default risk, credit default spreads (CDS) on U.S. 5-year treasury bonds in 2009 was 67 basis points, long before the August 2011 debt-limit scare. In liquidity risk, 3-month U.S. treasury bonds stopped trading for 30 minutes at the peak of the post-Long-Term Capital Management (LTCM) turmoil in 1998. In terms of inflation risk, the U.S. inflation rate rose to well over 10% in 1979-1980. Finally, regarding exchange rate risk, the U.S. dollar depreciated by 50% against the Deutsche mark from February 1985 to October 1987. It is clear that "safety" is not absolute but lies on a continuum that requires judgment or reference to the markets.

On a similar note, in a survey conducted by Allianz Global Investors in 2012, institutional portfolio managers were asked which assets they considered to be safe. Results indicated that market views are indeed fairly diverse. Answers ranged from quality sovereign debt and precious metals to quality real estate and quality equities. Other respondents considered hard currencies and infrastructure securities to be safe assets. Furthermore, 13 respondents stated that there is in fact no such thing as a safe asset.

Concerns regarding a lack of safe assets revolve around various themes, with one being collateral shortages. The basis for this is that new regulatory requirements being imposed will require greater collateralization in a number of areas. There is a concern regarding where additional assets to be used as collateral can be found. Another worry is that the supposed shortage of safe assets will be responsible for lower real interest rates, which in turn will lead to a search for yield, and, more generally, financial instability. The issue of international finance is another which is causing concern, particularly with regard to global imbalances. Caballero, co-authors, and many commentators draw strong policy conclusions from this safe asset meme.

In discussing collateral, it is first worth noting the difference between a shortage as opposed to a scarcity. Independently of whether there is a shortage of assets that can be used as collateral, scarcity could possibly be handled by accelerating the velocity of the circulation of collateral. In this sense, collateral can be thought of as a kind of money.

The claim being asserted is that safe assets are essential as collateral, and that collateral is essential for financial intermediation. Hence, if there really were a shortage of safe assets that can be used as collateral, this would constrain and potentially reduce lending. In fact, however, there is no empirical evidence available which supports this theory that a shortage in collateral has a negative effect on lending. There was a period of a collateral squeeze in the eurozone, although this turned out to be a technical issue and was handled fairly quickly, and it was not a safe asset shortage. Krishnamurthy and Vissing-Jorgensen did find evidence that an increase in Treasury debt reduces the probability of a financial crisis, which Gordon and Ordonez assert is because Treasuries are superior substitutes for private collateral. In a way, U.S. Treasuries are indeed superior as collateral to private collateral, although this doesn't rule out the possibility that private collateral can be sufficient.

Furthermore, the Committee on Global Financial Stability (CGFS), to whom we have entrusted these collateral issues in macroprudential assessment, has studied them and has clearly stated that although regulatory reforms and the shift toward central clearing of derivatives transactions will add to the demand for collateral assets, there is no evidence or expectation of any lasting widespread scarcity of such assets in the global financial markets.

In terms of financial yield, the claim being made is that the shortage of safe assets in the mid-2000s pushed real interest rates down to historically low levels, and, in turn, investors needing yield sought more risky assets. Caballero further added that a shortage of safe assets also led to global imbalances and asset price bubbles. Indeed, it could be argued that the economic crisis was due to a shortage of safe assets leading to private sector agents creating private label assets which were certified by ratings agencies but weren't actually safe. Bernanke, Kocherlakota, Gourinchas and Jeanne, along with the IMF in the GFSR, all share the concern that safe asset shortages will lead to financial instability such as volatility jumps, herding, and cliff effects.

But do the data suggest that we should actually be worried about a global shortage of safe assets? There is neither evidence of a global liquidity shortage nor a global deflationary bias. Inadequate international liquidity during the recent financial crisis was due only to a short-run lack of U.S. dollars to finance dollar positions, met by short-duration currency swaps. On top of this, it is not only the United States which supplies safe assets. Germany, the United Kingdom, Norway, Switzerland, and even emerging markets also do so. Gruber and Kamin of the U.S. Federal Reserve Board (FRB) have also argued that U.S. financial assets have not been demonstrably more attractive than those of other industrial economies.

In looking at the data on interest rates, it seems that interest rates in the 2000s weren't historically low at all. Real interest rates did fall from the 1980s and early 1990s to levels that appeared historically low in the 2000s, but they actually weren't, as real interest rates were similar in the 1960s and lower in the 1970s. Historical evidence also indicates that the search for yield is not leading to real compression of the spreads on corporate bonds. CDS spreads regarding the implied probability of default over a five-year period, assuming a recovery rate of 90% in case of a default, suggest that even the "best" sovereign bonds are "unsafe." Shifting the recovery rate to an assumed 55% lowers the implied probability of default and makes such bonds look safer, although numbers of 4%-5% can hardly be classed as negligible (so following the Gourinchas-Jeanne definition, there would be no safe assets).

Credit Suisse's graphs show a decline in safe assets in primary reserve currencies since 2007. But the Organisation for Economic Co-operation and Development (OECD) released a graph in January 2013 which paints an entirely different picture, indicating an upward trend in the supply of safe sovereign assets. In the private sector, Goldman Sachs also suggests that the total safe asset pool has not fallen since 2007. Furthermore, it also defines safe assets by positive yield correlation with risk appetite. Through defining safe assets in this way, U.S. Treasuries, non-Euro area G10, German, Dutch, Finnish, U.S. agencies, and AAA-rated covered bonds are still classed as being safe. Backing this up, even the GFSR has released a graph suggesting optimistic projections for the future of safe assets.

In reality, it is very difficult to back up the safe asset shortage argument with data. It is hard to define and agree on what a safe asset is, and it could be argued that the role of ratings is dubious at best. The GFSR states that asset safety should not be viewed as being directly linked to credit rating, yet it still does so itself. Even so, the downgrading of U.S., UK and French ratings has had no effect on 10-year yields. Although it is true that sovereign nominal yields are low, the reason for this is most likely due to low policy rates and the belief that those policy rates will continue to be low for several years.

Despite the various interesting theoretical models, if we look carefully at the collateral issue, the supply of safe assets, and data available on their interest rates and predicted future quantities, it can be seen that we indeed should not be overly concerned about a lack of safe assets. Caballero mentions that the way of handling a supposed shortage of safe assets is neither to try to eliminate them, nor enter into global imbalances, economic bubbles, or low real rates. Rather, he stresses the importance of trying to control risks as much as possible. Caballero also suggests that, with the supposed shortage of safe assets and collateral scarcity problem, we shouldn't be imposing higher collateral requirements. However, it could be argued that this is not a good approach at all. In fact, there is a good chance that higher collateral requirements might discourage banks from dangerous market funding.

In conclusion, the best policy for the future is to attempt to control replication of this safe asset meme in order to gain a more realistic perspective.

Questions and Answers

Q1: You showed an interesting chart that displayed real interest rates in the United States. It seems to indicate that, in the 1960s and 1970s, investors bought U.S. Treasury securities at very low interest rates, and that they had confidence that inflation would stay under control. However, at some point, they appear to have lost confidence to manage huge risk premiums and interest rates, and no longer viewed U.S. Treasuries as safe assets. Do you think that the same thing could now happen with U.S. default risk, leading to a shortage of safe assets?

Richard PORTES
There is no doubt that if the United States were to default truly on payments on Treasury securities, this could happen. The most important safe asset in the world is U.S. Treasury securities, so there would naturally be huge repercussions if they were to become viewed as no longer being safe. Yet it is very likely that other consequences of a serious U.S. default would overshadow any concerns regarding safe assets. However, it is very unlikely that a situation like this would occur.

Q2: From your presentation that, it seems that it is very difficult to define and reach a consensus on exactly what are safe assets. It also seems very clear that it is important to avoid a shortage of safe assets. What do you feel is the role that central governments or monetary authorities should play in order to increase or prevent a decrease in safe assets? Second, what is your opinion on the recent monetary policies of the Bank of Japan (BOJ)?

Richard PORTES
The policy of the BOJ is causing a reduction in the supply of government bonds available on the market. At the same time, this is helping to increase the supply of assets which are even safer. The idea that quantitative easing is taking a huge quantity of safe assets off the market, thereby causing a shortage of safe assets, ignores the fact that the central bank pays for this with its own money, which is very safe. Therefore, in terms of the safe assets issue, this doesn't seem to be a problem. The effectiveness of this mechanism is a different issue. There is some evidence in the United States and the United Kingdom that policies like this have had effects on securities prices which have spread into other markets as intended. And despite quantitative easing, U.S. and eurozone inflation is now down to not much above 1%, which is lower than targets of close to 2%. Something should be done about this, which reflects failures in monetary policy rather than quantitative easing policies.

There is a concern that quantitative easing expands the balance sheets of central banks. They have indeed grown enormously, but there appears to be no reason why central banks having a large stock of securities on their balance sheets should be a cause for concern. Central banks could hold securities until maturity, or convert bonds into zero-coupon infinite duration securities. This would be another very safe asset.

Q3: It is being said that government bonds should act as an anchor of the international market, and that governments should not finance too much in order to contain imbalance. It seems that, in the United States, the difference between government bonds and bank notes is almost nonexistent as the interest rate is so low. Do you feel that this situation is acceptable?

Richard PORTES
No, we are in a liquidity trap. There are political and economic constraints on fiscal policy, either for ideological reasons in the United States and to some extent in the United Kingdom, or for other reasons in Japan. This is a big problem. Quantitative easing policies are an obvious attempt to handle this issue. It is true that when you have very low short-term policy rates, money and short-term government bonds are almost perfect substitutes. This doesn't appear to be a problem in itself.

The problem in Japan is that the fiscal trajectory is ultimately not sustainable, and it is important to find the solution to turn this around. The increase in the consumption tax rate is one sign that the Japanese government is now concerned enough about this issue to take solid measures.

Looking again at the Triffin Dilemma of the 1960s, it wasn't due to a U.S. current account deficit, but rather a capital account deficit. Americans were using U.S. dollars to buy up European and French companies. An article about capital accounts published in The Economist in 1966, which is still relevant to this day, observed that the United States was effectively supplying financial intermediation services to Europe. It appears that this same scenario has also been occurring between the United States and China, along with other emerging market countries in the 2000s. There should be some kind of constraint on how long this can continue. However, in the 1960s, the issue was simply a question of confidence. Foreign countries lost confidence in the U.S. dollar, which caused the Nixon Shock.

Q4: You mentioned the goal of the current easing of the monetary policy of the BOJ. At present, commercial banks have large deposits in the BOJ. If real inflation occurs, those commercial banks will likely withdraw their deposits. In order to prevent this, the BOJ will have to raise interest rates. What is your opinion on this matter?

Richard PORTES
There are a number of issues in relation to this topic. First, the target of this policy is a 2% inflation rate. Nominal interest rates should rise accordingly, but not to disastrous levels. This will result naturally in an increase in debt service, although this will likely be trivial. The effect on the private sector is interesting. The BOJ claims that it has analyzed the effects of a fall in the price of Japanese government bonds (JGBs) on commercial banks, and that their balance sheets would not be disastrously affected. It seems that the balance sheets of smaller and medium-sized banks may be negatively affected to a certain degree, but not those of larger banks. There seems to be no good reason to doubt this assertion.

It would actually be more worrying if the policy were to fail, and if the inflation rate were not to rise. It is very encouraging to see that expectations are indicating that this long period of deflation seems to be coming to an end. Although this doesn't necessarily solve all important issues, the policy bringing inflation up to 2% could mark a significant step forward.

Overall, as pointed out, there is indeed a risk associated with the BOJ's current monetary policy, which will require it to observe the results carefully.

*This summary was compiled by RIETI Editorial staff.