Tag Archives: bonds

The VIX is too low!

Share Button

September 30, 2017 —   During most of this year, the VIX — the Volatility Index on The Chicago Board Options Exchange — has been at the lowest levels of the last ten years.  It recently dipped below 9, even lower than March 2007, just before the sub-prime mortgage crisis. It looks as though, once again, investors do not sufficiently appreciate how risky the world is today.

Known colloquially as the “fear index,” the VIX measures financial markets’ sensitivity to uncertainty, in the form of the perceived probability of large changes in the stock market.  It is inferred from the prices of option on the stock exchange (which pay off only when stock prices rise or fall a lot).   The low VIX in 2017 signals that we are in another “risk on” environment, when investors move out of treasury bills and other safe haven assets and instead “reach for yield” by moving into riskier assets like stocks, corporate bonds, real estate, and carry-trade currencies.

Figure 1: The VIX is at its lowest since 2007


One need not rely exclusively on the VIX to see that the markets are treating the current period more as a risk-on opportunity than risk-off.  The returns on safe-haven assets were generally lower than the returns among risk-on assets in the first half of the year.  On the one hand, the Swiss franc depreciated.  On the other hand, the Australian dollar and Chinese yuan appreciated.  And the stock market has hit record highs.

True risk is currently high

Why do I presume to second-guess the judgment of the VIX that true risk is low?   One can think of an unusually long list of major possible risks. Each of them individually may have a low probability of happening in a given month, but cumulatively they imply a worrisome probability that at least one will happen sometime over the next few years:

* Bursting of stock market bubble.   Major stock market indices hit new record highs this month (September 12), both in the United States and worldwide.   Equity prices are even elevated relative to such benchmarks as earnings or dividends.  Robert Shiller’s  Cyclically Adjusted Price Earnings ratio is now above 30.  The only times it has been this high were the peaks of 1929 and 2000, both of which were followed by stock market crashes.

Figure 2: Shiller’s adjusted P/E shows stocks at their 3rd-highest valuations since 1880.

* Bursting of bond market bubble.  Alan Greenspan has suggested recently that the bond market is even more overvalued (by “irrational exuberance”) than the stock market.   After all, yields on corporate or government bonds were on a downward trend from 1981 to 2016 and the market has grown accustomed to it.  But, of course, interest rates can’t go much lower and it is to be expected that they will eventually rise.

* What might be the catalyst to precipitate a crash in the stock market or bond market? One possible trigger could be an increase in inflation, causing an anticipation that the Fed will raise interest rates more aggressively than previously thought. The ECB and other major central banks also appear to be entering a tightening cycle.

* Geopolitical risks have rarely been higher, and faith in the stabilizing influence of America’s global leadership has rarely been lower.  The gravest risk lies in relations with North Korea, which Trump’s response has been exceedingly erratic.  But there are also substantial risks in the Mideast and elsewhere.  For example Trump threatens to abrogate the agreement with the Iranians that is keeping them from building nuclear weapons.

* In many policy areas it is hard to predict what Trump will say or do next, but easy to predict that it will be something unprecedented.  So far, the ill effects on the ground have been limited, in large part because most of the wild swings in rhetoric have not translated into corresponding legislation.  (If he really had stuck with his decision to kick 800,000 young DACA workers and students out of the country it could have caused a recession.)  But this is a time of policy uncertainty if there ever was one.

* US Congressional showdowns over the debt ceiling and government shutdown were successfully avoided in September, but only by kicking the can down the road to the end of the year, when the stakes could well be higher and the stalemate worse.

* A constitutional crisis could arise, if for example Special Counsel Robert Mueller were to find that contact between the Trump campaign and the Russian government was illegal.

Black swans are not unforecastable

The current risk-on situation is reminiscent of 2006 and early 2007, the last time the VIX was so low. Then too it wasn’t hard to draw up a list of possible sources of crises.  One of the obvious risks on the list was a fall in housing prices in the US and UK, given that they were at record highs and were also very high relative to benchmarks such as rent.  And yet the markets acted as if risk was low, driving the VIX and US treasury bill rates down, and stocks, junk bonds, and EM securities up.

When the housing market indeed crashed, it was declared an event that lay outside any standard probability distribution that could have been estimated from past data, supposedly an example of what was variously declared to be Knightian uncertainty, radical uncertainty, unknown unknowns, fat tails, or black swans.  After all, “housing prices had never fallen in nominal terms,” by which was meant they had not fallen in the US in the last 70 years.  But they had fallen in Japan in the 1990s and in the US in the 1930s.  This was not Knightian uncertainty, but classical uncertainty with the data set unnecessarily limited to a few decades of purely domestic data.

In fact the “black swan” analogy fits better than those who use the term realize.  Nineteenth-century British philosophers cited black swans as the quintessential example of something whose existence could not be inferred by inductive reasoning from observed data.  But that was because they did not consider data from enough countries or centuries.  (The black swan is an Australian species that in fact had been identified by ornithologists in the 18th century.)  If I had my way, “black swan” would be used only to denote a tail-event that could have been assigned a positive probability ex ante, by any statistician who took the care to cast the data net widely enough, but that is declared “unpredictable” ex post by those who did not have a sufficiently broad perspective to do so.

The risk-on risk-off cycle

Why do investors periodically under-estimate risk?  There are specific mechanisms that capture how market analysts fail to cast the net widely enough.  The formulas for pricing options require a statistical estimate of the variance.  The formula for pricing mortgage-backed securities requires a statistical estimate of the frequency distribution of defaults.   In practice, analysts estimate these parameters by plugging in the last few years of data, instead of going back to previous decades, say the 1930s, or looking at other countries, say Japan.  More generally, there is a cycle described by Minsky whereby a period of low volatility lulls investors into a false sense of security which in turn leads them to become over-leveraged, leading ultimately to the crash.

Perhaps investors will re-evaluate the risks in the current environment, and the VIX will adjust.  If history is a guide, this will not happen until the negative shock – whatever it is – actually hits and securities markets fall from their heights.

[A shorter version of this post appeared at Project Syndicate.  Comments can be posted there or at Econbrowser.]

Share Button

Should Bond Benchmarks Shift from Traditional to GDP-Weighted Indices?

Share Button

Some prominent institutional bond investors are shifting their focus away from traditional benchmark indices that weight countries’ debt issues by market capitalization, toward GDP-weighted indices.   PIMCO (Pacific Investment Management Company, LLC, the world’s largest fixed-income investment firm) and the Government Pension Fund of Norway (one of the world’s largest Sovereign Wealth Funds), have both recently made moves in this direction.  

There is a danger that some investors will lose sight of the purpose of a benchmark index.   The benchmark exists to represent the views of the median investor dollar.  For many investors, going with the benchmark is a good guideline – especially those who recognize themselves to be relatively unsophisticated and also those who think they are sophisticated but really aren’t.   This is the implication of the Efficient Markets Hypothesis (EMH), for example.  

On the one hand, EMH theorists are often too quick to discount the possibility of ways to beat the benchmark.   To take an example, it should not have been so hard to figure out during the 2003-07 credit-fed boom that countries with high foreign-exchange-denominated debt, particularly in Europe, were not paying a sufficiently high return to compensate for risk.  That mistake described Eastern European countries with low ratios of reserves to short-term debt as well as periphery euro members that lacked their own currency.  It probably resulted from easy money, reach for yield, and pervasive underestimation of risk.  Or, to take another example (admittedly, a tougher call), some of these countries’ deeply discounted bonds would have been good buys in early 2012, after heavy mark-downs.   

On the other hand, most investors would do better if they went with a more passive investment strategy – especially due to high management fees among actively managed funds, exacerbated by excessive turnover.   At a minimum, if one is pursuing an activist strategy such as investing more in low-debt countries, it is helpful to frame it explicitly as a departure from the view of the median investor in order to be clear in your mind as to the nature of the bet you are making.

I can think of four functions of a benchmark index.    First, investors who do not figure that they can systematically beat the median investor need to be able to hold passively a portfolio designed to track a benchmark index consisting median investor holdings.   (See Vanguard.)   The second function is to provide an objective standard by which investors can judge the performance of active portfolio-managers who claim to be able to beat the median investor, within a specific asset class like sovereign debt.  (See Morningstar.)   Third, the same weights that are used in the index can be used to compute an average interest rate or sovereign spread in the market, which can serve as an indicator as to where the median investor is currently, in the risk-on, risk-off spectrum. (See J.P.Morgan’s EMBI — Emerging Market Bond Index.)    

The fourth function of a benchmark index is to help active investors to devise a deliberate strategy to depart from the views of the median investor when they think that the latter is erring in a particular direction.  They may think that the median investor is under-estimating risk in general (spreads too low) or under-estimating the downside in countries with some particular characteristic.   Examples of such characteristics include insufficient currency flexibility, inadequate reserves, too much short-term debt, too much foreign-currency debt, too much bank debt, insufficient openness to FDI, insufficient cost competitiveness, excessive budget deficits, insufficient national saving, political risk, and so forth.

For each of these four functions of a benchmark, the correct way of weighting different countries is by market capitalization.  True, the keeper of the index will need to judge what countries and what bonds are in “the market,” i.e., are fully investable.   But this is true for any index.

The logic behind the movement away from traditional bond market indices is that by construction they give a lot of weight to countries with high debt, some of which may be over-indebted and at risk of default.  At first the logic seems unassailable.  But in theory, if the market is functioning well, it should already have factored in high debt levels:  such countries should pay higher interest rates to compensate for the risk, unless there is some special reason to think they can service their debts easily.

It is important to emphasize that many investors will want to depart from the benchmark in various directions, as indicated under the fourth motive for having a benchmark.  An investor’s belief that countries with high debt/GDP ratios are riskier than the median investor realizes would call for a strategy equivalent to moving from the market-cap benchmark in the direction of the GDP-weighted benchmark.  But one is more likely to think about the strategy clearly if it is explicitly phrased in terms of factoring in debt/GDP ratios, rather than phrased as following a new GDP-weighted index.  Furthermore the phrasing may help an investor realize that he or she might want to modify the strategy if, for example, the country in question can borrow readily in terms of its own currency (think of American exorbitant privilege or Japan’s high domestic holdings of own debt) or if, on the other hand, its debt has a particularly vulnerable maturity or currency structure (think of Hungary).

Investors are reacting to what has turned out to be default risk that was higher than they had expected, among some high-debt countries.  Taking greater note of high debt levels last decade would have warned investors away from countries like Greece and Hungary.   But there is always a danger of fighting the last war.   Middle-income countries have paid down much of their debts over the last decade, attaining debt/GDP ratios far below those of advanced economies.    As the chart shows, major emerging markets have relatively low debts [first bars, for each country] compared to GDP [second bar].  That is, their debt/GDP ratios [third bar] are now much lower than in advanced countries.  (Russia’s sovereign debt is now below 7 per cent of GDP.)  As a result, there is only a limited supply of their bonds left to hold.  If the global investor community switches from market-cap-weighted to GDP-weighted investing, the high demand and low supply of bonds from low debt/GDP countries may drive their interest rates to unnaturally low levels, setting off new credit-fed boom-bust cycles in their economies.  

Of course, as a country’s international debt approaches zero, the keepers of the index might drop it altogether.  But the fall in demand for that country’s remaining international bonds from the investment funds that are following the benchmark could then produce an undesirable discontinuous jump in the interest rate.

Many emerging market countries have paid down debt denominated in dollars or other foreign currencies, while continuing to borrow in their local currencies.  (See the table at bottom.)  Such relatively large countries as Thailand, Malaysia, Indonesia, South Africa and Russia, for example, have little dollar-denominated debt left – 3% of GDP or less (shown in the chart as the dark bottom of each first bar).   If an international bond benchmark is to be limited to dollar-denominated debt, then GDP weights could imply a severe imbalance between international investor demand for these countries’ bonds and the small supplies available. 

Accordingly, local currency denominated debt must be included in the most useful benchmarks.  But then a portfolio reallocation away from traditional benchmark indices such as the EMBI would imply a big shift in allocations away from simple credit risk toward currency risk.   True, the ability of emerging market economies to attract foreign investment in their local currencies represents an important strengthening of the global financial system, relative to the currency mismatch and balance sheet vulnerabilities of the 1990s.  Nevertheless, an investor switching from one “benchmark” to the other needs to be aware of the extent to which the reduction in default risk comes at the expense of heighted exposure to currency risk.

In short, it is not crazy for an investor to depart from the market-cap-weighted benchmark in the direction of putting more weight on debt/GDP countries and less weight on high debt/GDP countries.   But the GDP-weighted index should not be mistaken for a neutral benchmark.

[A version of this post originally appeared at Project Syndicate, Feb. 11, 2013.  Comments can be posted there, or at Seeking Alpha.]

Table:  Sovereign debts as a percentage of GDP













































South Africa












2011, Q4.  Sources: Debt data from BIS, Tables 12 & 16.  Nominal GDP from Global Financial Data.


Share Button

Barrels, Bushels & Bonds: How Commodity-Exporters Can Hedge Volatility

Share Button


The prices of minerals, hydrocarbons, and agricultural commodities have been on a veritable roller coaster. Although commodity prices are always more variable than those for manufactured goods and services, commodity markets over the last five years have seen extraordinary volatility.


Countries that specialize in the export of oil, copper, iron ore, wheat, coffee, or other commodities have boomed.  But they are highly vulnerable. Dollar commodity prices could plunge at any time, as a result of a new global recession, a hard landing in China, an increase in real interest rates in the United States, fluctuations in climate, or random sector-specific factors.


Countries that have outstanding debt in dollars or other foreign currencies are especially vulnerable. If their export revenues were to plunge relative to their debt-service obligations, the result could be crashes reminiscent of Latin America’s debt crisis in 1982 or the Asian and Russian currency crises of 1997-1998.


Many developing countries have made progress since the 1990’s in shifting from dollar-denominated debt toward foreign direct investment and other types of capital inflows, or in paying down their liabilities altogether. But some commodity exporters still seek ways to borrow that won’t expose them to excessive risk.


Commodity bonds may offer a neat way to circumvent these risks. Exporters of any particular commodity should issue debt that is denominated in terms of the price of that commodity, rather than in dollars or any other currency. Jamaica, for example, would issue alumina bonds; Nigeria would issue oil bonds; Sierra Leone would issue iron-ore bonds; and Mongolia would issue copper bonds. Investors would be able to buy Guatemala’s coffee bonds, Côte d’Ivoire’s cocoa bonds, Liberia’s rubber bonds, Mali’s cotton bonds; and Ghana’s gold bonds.


The advantage of such bonds is that in the event of a decline in the world price of the underlying commodity, the country’s debt-to-export ratio need not rise. The cost of debt service adjusts automatically, without the severe disruption that results from loss of confidence, crisis, debt restructuring, and so forth.


The idea is not new. (The oldest reference I know is Lessard & Williamson, 1985.)  So, why has it not been tried before? When one asks finance ministers in commodity-exporting debtor countries, they frequently reply that they fear insufficient demand for commodity bonds.


That is a surprising proposition, given that commodity bonds have an obvious latent market, rooted in real economic fundamentals. After all, steel companies have an inherent need to hedge against fluctuations in the price of iron ore, just as airlines and utilities have an inherent need to hedge against fluctuations in the price of oil.  Each of these commodities is an important input for major corporations. Surely there is at least as much natural demand for commodity bonds as there is for credit-default swaps and some of the bizarrely complicated derivatives that are currently traded!


It takes liquidity to make a market successful, and it can be difficult to get a new one started until it achieves a certain critical mass. The problem may be that there are not many investors who want to take a long position on oil and Nigerian credit risk simultaneously.


A multilateral agency such as the World Bank could play a critical role in launching a market in commodity bonds. The fit would be particularly good in those countries where the Bank is already lending money.


Here is how it would work. Instead of denominating a loan to Nigeria in terms of dollars, the Bank would denominate it in terms of the price of oil; it would then turn around and lay off its exposure to the world oil price by issuing that same quantity of bonds denominated in oil. If the Bank lends to multiple oil-exporting countries, the market for oil bonds that it creates would be that much larger and more liquid. It can serve an additional important pooling function in cases where there are different grades or varieties of the product (as with oil or coffee), and where prices can diverge enough to make an important difference to the exporters.  The Bank could link the bond it issues to an oil price index, a weighted average of various product grades.


An alternative for some commodity exporters is to hedge their risk by selling on the futures market. But an important disadvantage of derivatives is their short maturity. A West African country with newly discovered oil reserves needs to finance exploration, drilling, and pipeline construction, which means that it needs to hedge at a time horizon of 10-20 years, not 90 days.


Another disadvantage of derivatives is that they require a high degree of sophistication –both technical and political. In the event of an increase in a commodity’s price, a finance minister who has done a perfect job ex ante of hedging export-price risk on the futures market will suddenly find himself accused ex post of having gambled away the national patrimony. This principal-agent problem is much diminished in the case of commodity bonds.


If the international financial wizards can get together and act on this idea now, commodity exporters might be able to avoid calamity the next time the world price of their product takes a plunge.  The World Bank should take up the cause.


[This column originally appeared via Project Syndicate, which has the copyright.  Comments may be posted there.]





Share Button