Blog

In two recent feature articles (here and here) we’ve focused on the illiquidity that can affect exchange-traded funds when markets get volatile.

Recently, we pointed out that during August a number of ETFs and ETCs traded with average dealing spreads which exceeded the maximums permitted by the London Stock Exchange by some margin.

While this sounds like a breach of the rules, it may not be in practice. The exchange allows traders significant get-out clauses when markets become unstable. However, ETPs with recorded average spreads approaching or exceeding 100 percent are clearly illiquid and the publication of a maximum spread band of 1.5 percent, 3 percent or 5 percent for such a fund is misleading.

To help investors to avoid mistrades, Australia’s financial regulator recently suggested they should check funds’ indicative NAVs (iNAVs) before trading. This sounds like a good idea in theory. In practice, unless investors have access to a pricey Bloomberg data feed, they’ll be unable to do so. Exchanges’ and ETF issuers’ websites are also of little help in finding out a tracker’s intrinsic value (though some are better than others).

iNAVs are also inherently unreliable and are only as good as the underlying price source. During the worst part of the 2008 market collapse, the prices of corporate bond issues varied wildly, depending on whom you asked. One leading fixed income index provider told me that in October 2008 his firm received markedly different prices from two trading firms for the same financial sector bond. One said the bond was worth 20, the other valued it at 90.

Looking at the current shenanigans regarding the valuation of Greek government debt issues in European banks’ accounts, it’s hard to argue that things have got any better.

In a blog post published on Friday, “London Banker” argued that central banks need to take a much more forceful stance in ensuring that “capital assets are issued in fungible series, in size, and traded in transparent exchange markets with committed market makers”.

“Securities regulators have been under pressure for several decades to liberalise OTC markets, permit fragmentation to off-exchange trading systems, and turn a blind eye to issuance of securities in small, idiosyncratic offerings that will never liquidly trade except back through the offering investment banks,” London Banker continued. “The quality assurance and market conduct functions of exchanges have been eroded following demutualisation, and exchanges now are run for profit of their highly concentrated owners rather than in the public interest.”

“We should now be forcing assets back onto exchanges and force the exchanges to regulate quality and information norms,” the blogger concluded, suggesting that remutualisation or public ownership should be considered.

Whatever your view on exchanges’ organisation and ownership, there’s much more that they could be doing, in my view, to ensure that the ETPs they list are actually tradeable. This is arguably an increasingly urgent task, since there are signs that the illiquidity in some non-mainstream trackers is becoming acute.

ETF issuers, in common with active fund managers, have few incentives to stop launching funds. The reputational risks involved in creating an ETF that turns out to be illiquid are likely to be outweighed by the fees on offer if the fund works and assets flow in. Nor are ETF market makers likely to prove an obstacle to the listing of illiquid funds. While traders say that they let issuers know if a fund idea will be difficult for them to support, they’ll also tell you openly that they make the most money when markets are volatile and spreads widen.

So it’s up to the exchanges, in my view, to tighten up their rules and to ensure better protection for investors. Exchanges should be providing full, real-time iNAV information for all the funds they list, together with easy-to-use tools for measuring fund premiums and discounts, both actual and historical.

If market makers are unable to quote within maximum spread bands, exchanges should publicise this information immediately in the form of a risk warning for the affected trackers. If ETFs’ average dealing spreads fall outside target bands for a period of time, exchanges should give issuers and market makers a warning, then be prepared to delist the affected funds if things don’t improve. The performance of individual market makers and their maintenance of the two-way quote obligations that they sign up to should be audited and reported publicly.

“Markets are at the heart of successful civilisations. Markets require quality norms, information publication, and price transparency to operate effectively,” argued London Banker. It’s hard to argue with this. For the exchange-traded fund market, recent evidence suggests there’s plenty to be done to ensure that these trackers live up to the first and second words of their name.

If you haven’t already read it, I recommend you take a look at Andrew Clark’s fascinating article on the possible losses to be expected by investors tracking stock and commodity indices.

Clark makes three key points, in my opinion.

First, the actual losses that might be expected from a portfolio of stocks or commodities are larger than classical finance theory predicts. Classical theory, including the Capital Asset Pricing Model (CAPM), which is the basis of most textbooks on investment and finance, is based upon an assumption of returns from stocks being normally distributed.

In practice, this doesn’t work. Clark reminds us that the actual expected loss from a portfolio is much higher, particularly when measured over short time periods. Markets crash more often than most textbooks would suggest.

Second, Clark suggests that we should combine his measure of “expected shortfall” with our own individual level of loss aversion to assess how much risk we can really take. We know from behavioural finance theory that people tend to be more sensitive to losses than to gains, often accepting a lower expected return over time in order to avoid an interim shortfall.

If we’re not fully aware of our own sensitivity to loss, and of the fact that markets can drop by a lot more than we’re anticipating, we’re likely to end up bailing out of our investment at the worst possible moment: selling, rather than buying, in March 2009, for example. We probably all know the sensation of “never again” after having been taken on too bumpy a ride.

Using leverage makes it even more likely that we’re going to panic and get out at the wrong time. Clark also questions the suitability of commodities as an asset class for loss-sensitive investors, given the very large downside risks they incur.

Finally, Clark reminds us that certain index variants can amplify potential losses. Autocorrelation in capitalisation-weighted indices—the tendency of all stocks to move together at the same time—may be higher than for other index types, he says, and can land investors with much bigger moves than they had foreseen.

This is something I wrote about in my last blog. There’s been a steady increase in internal index correlations for some of the most widely-followed equity indices, something that seems to reflect the steady increase in stock market investment via ETFs and other tracker funds (that’s my opinion, though others disagree).

Do ETF investors buying an S&P 500 index tracker, for example, understand that they are exposing themselves to a potential risk from the index methodology itself (as opposed to randomly selecting a portfolio of the index constituents and equal-weighting them, for example)? Or that this index-related risk, that of implied internal correlations between stocks, has been rising steadily?

Clark argues that index providers need to do more to highlight such risks, and it’s hard to disagree.

But what are the implications of Clark’s article for those investing by index (or ETF) across asset classes? First, that you may lose more than you might imagine, and you should reassess your capacity for losses and your investment time horizon in the light of this. Second, you need to diversify more than you might think. Third, you should examine closely the index methodology your tracker is using, and consider both alternative weighting schemes and less popular index variants.

In other words, there’s no way to remain passive about your investment portfolio, even if you’re using index funds.

I’ve never been a fan of the “risk on, risk off” categories used by commentators to describe market activity from one day to the next, finding it a simplistic way of reporting what’s actually happening. But there’s good evidence that investors are increasingly behaving in such a synchronised fashion.

The CBOE implied correlation index on the S&P 500 index of US stocks is a way of measuring the extent to which the 500 shares in the benchmark move together. The higher the correlation index's level, the more the S&P 500's constituent stocks are expected to move in the same direction at the same time.

CBOE describes the principle behind the index on its website as follows: “The implied volatility of a single-stock option simply reflects the market’s expectation of the future volatility of that stock’s price returns. Similarly, the implied volatility of an index option reflects the market’s expectation of the future volatility of that index’s price returns. However, index volatility is driven by a combination of two factors: the individual volatilities of index components and the correlation of index component price returns.”

Measure the implied volatility of index options and then strip out the implied volatilities of the options on constituent stocks, and you’re left with the implied correlation component, in other words.

You can see from the late-2008 peak in the implied correlation index that the point of maximum synchronisation between S&P 500 index stocks came during the most intense part (to date) of the financial crisis. All the index stocks moved down (and, on occasion, up) together as the correlation index hit 100.

But you can also see in the chart that there’s been a steady increase in implied correlation since 2007, despite the recovery in the equity markets from their 2009 bottom. In a research report published last October, analysts from JP Morgan’s equity derivatives team measured correlation from further back in time, showing that there’s been a rising trend since 2000. At the turn of the millennium the implied correlation figure between the S&P 500 index’s stocks was around 30 percent. Now we’re seeing levels of around 70-80 percent.

Various commentators have attributed this, with apparently good reason, to the increasing importance of index-based investing (typically via ETFs or futures). More and more people are trading in indices as a whole, rather than picking out individual stocks from within them.

So when you see stories like the one on our US website last week—describing how ETF trading volumes have recently increased to 40 percent of the overall exchange-based activity in equities—you can guess that equity implied correlation has jumped too. In fact it has—CBOE’s website tells us that the January 2012 implied correlation index rose from 59 at the end of July to 77 yesterday.

What does all this mean? First, if one of the touted benefits of ETFs and index funds is that you can gain diversified access to whole areas of the market in a single transaction, rather than having to buy all the constituent stocks individually, it’s clear that you are getting a lot less diversification than you used to when buying a typical index tracker.

But rising implied correlation doesn’t necessarily tarnish all ETFs’ attractiveness. One implication of increased uniformity in the market movements of S&P 500 constituents is that equity market sectors with potentially very different prospects may all be being treated in the same way, creating over- and undervaluations, and thereby investment opportunities. Those using S&P 500 index ETFs may be better off going long (and short) sectors, in other words.

Another index category that might benefit from increased correlations within the main capitalisation-weighted benchmarks might include those using different weighting methodologies, or built on the basis of an investment strategy.

But, overall, a tendency towards increased polarisation among investors is probably a warning sign. Just as the split of many Western European societies into left and right-wing extremist groups during the 1930s presaged a broader conflict ahead, herd-like behaviour in equity markets doesn’t bode well for the stability of the financial system as a whole.

According to the UK’s Chancellor of the Exchequer, George Osborne, the recent declines in gilt yields are “a huge vote of confidence in the credibility of British Government debt and a major source of stability for the British economy at a time of exceptional instability”.

In a statement to parliament yesterday, Osborne said the fall in government bond yields to their lowest levels in 100 years, together with the decline in the UK government’s credit default swap (CDS) spread this week to below that of Germany, were reasons for optimism about the country’s economic outlook.

A longer-term view of one key measure of sovereign risk doesn’t provide grounds for such unbridled optimism. The chart below, using data sourced from Markit, illustrates a steady rise in the UK government’s CDS spread since the end of 2006. Even if UK default risk has fallen back from the peak recorded in late 2008, the notion that government bonds are “risk-free”, a principle that still underlies finance textbooks, hasn’t been tenable for several years.

But you’d surely rather be in Osborne’s shoes, and have the opportunity of extending the UK’s debt profile at the current low rates, than in the position of the finance ministers of most eurozone members. In an increasing number of countries, default risk measures have hit peaks, bond yields are at multi-year highs, and the short-term refinancing burden is also typically much more acute than in the UK.

But there’s a story Osborne wasn’t telling yesterday. That’s a far more worrying tale of a yawning black hole in the government’s finances, which is a direct consequence of falling bond yields.

I’m referring to the cost of providing unfunded public sector pensions, which balloons whenever interest rates fall.

The UK’s Treasury currently calculates the cost of providing state pensions in its annual accounts by using a discount rate based on projected GDP growth (which it estimates at 3 percent).

According to pensions consultant John Ralfe and an increasing number of independent experts, this approach is wrong. In a recent Financial Times article, Ralfe said: “The correct discount rate to measure the economic cost of public sector pension promises must be the yield on long-dated index-linked gilts, since they share similar characteristics: both are obligations of the UK government, both are legally-binding contracts and both are inflation-linked.”

This may all sound arcane, but the implications for the UK’s overall financial position of changes in that discount rate are staggering.

The UK’s Office for Budget Responsibility (OBR), a state-funded watchdog, recently produced a figure of £1.1 trillion as the net present value of public sector pension liabilities as at the end of March 2010, based upon a discount rate of 1.8 percent. That’s substantially bigger than the UK’s official net debt of £944 billion.

The OBR’s discount rate of 1.8 percent was derived from real yields on corporate index-linked bonds. If the statisticians were to apply the actual market rate on long-dated government inflation-linked bonds, 0.5 percent, we’d be talking about a liability that’s many hundreds of billions of pounds greater, and a sum that dwarfs reported figures of state obligations.

This all goes to show how acute the debt trap that Western governments face actually is. And, far from being a cause for celebration, low interest rates may not be making things any better at all.

By any account, bullion trackers have been a stunning success. Gold held by exchange-traded products has leapt from just over 10 million troy ounces at the end of 2005 to over 70 million ounces now.

Combine that with an increase in the gold price from around US$513 an ounce on the last trading day of 2005 to US$1,760 now, and you’re talking about a 25-fold jump in the overall “footprint” of gold ETPs in six years, from around US$5 billion to over US$125 billion.

The owners of these trackers have been pretty much the only players in the capital markets to have not only ridden out the upheavals of the last few years, but to have made healthy profits as well.

I’m in the camp of those who believe there’s more to come from gold’s bull market. A temporary setback may be in order given the meteoric rise this year and the recent flight into bullion as a result of equity market declines. But, given rising sovereign and financial sector default risk, it would seem foolish to bet against gold right now. And I don’t think we’ve yet seen the kind of mass-market, retail-driven mania that should typically mark the top of a multi-year uptrend.

But with success and with an increasingly crowded marketplace come the inevitable problems of growth.

The Financial Times reported this morning that the cost of vaulting gold has more than doubled this year at some of the major London-based bullion banks, reflecting capacity constraints. The huge jump in the gold holdings of ETFs and ETCs is cited as one of the main causes of the disappearing vault space.

According to Ben Traynor of Bullion Vault, the rise in custodians’ storage charges probably has more to do with rising insurance premiums than an actual shortage of physical space, since gold takes up so little room per unit of value. For silver, by contrast, says Traynor, the growth of investor interest is posing real problems for some custody networks.

The FT says that fees for storing gold range from 0.03 percent to 0.15 percent a year, depending on the size of a client’s holdings. Perhaps larger owners—the US-listed SPDR gold ETF, for example, which is not far short of becoming the world’s largest exchange-traded fund of any type, with over US$70 billion in assets—can still command storage rates at the bottom end of this range. HSBC, custodian for the fund, declined to comment on its pricing policy.

The four largest European gold trackers—ZKB’s gold ETF, ETF Securities’ physical gold ETC and Gold Bullion Securities and Julius Baer’s gold ETF—all charge at or close to 40 basis points each as annual management fees, the same as SPDR gold.

Given the continuing growth in the gold ETP market, the reported rise in storage costs may not be making a big impact on these issuers’ overall profits, even if their margins may be declining a bit.

But the cost increases may well threaten the business models of the issuers of gold trackers controlling smaller amounts of bullion. A 15 basis point annual charge for storage would eat up more than half the expense ratio of some recently issued products. iShares, for example, charges only 25 basis points, all-in, for its gold ETC, while Deutsche Bank levies 29 basis points a year for its version.

So if rising costs for gold ETPs are one sign of the market’s success, they may squeeze out some of the more marginal players. And given cost increases, the incentives for issuers to cut corners when managing custodial arrangements must also be on the rise. Investors should undoubtedly pay attention to the small print of prospectuses, particularly the sections relating to custody, insurance and the procedures for creating and redeeming ETP units.

We’ll be digging into some of these issues in more detail during our next webinar, to be held on September 6. I’ll be joined by Suki Cooper, precious metals analyst at Barclays Capital, and Charles Morris of HSBC Asset Management. Attendance is free—please sign up to listen in!