Volatility Tracker for the week of November 8, 2009

I wondered last week whether we would see a return to the reflation rally or were entering a new regime dominated by mean reversion. The price action last week counts in favor of both, as we reverted to the closing highs of the prior week; I expect a more definitive answer by November options expiration.

Volatility Tracker for the week of November 1, 2009 (.pdf)

Equity markets have been drunk on the wine of federal stimulus for most of this year. While the increased volatility in the latter half of last week could amount to a mere hiccup in the reflation rally, indicators suggest that more participants are concerned about an equity market “hangover” than at any time since the market bottom. The transition to a different market environment may have just occurred, and in the absence of any plausible catalysts, I’m not yet giving serious consideration to the possibility of a market crash, so the remaining likely regimes are a continued momentum-driven rally or the return of mean reversion.

One average of the volume of contracts traded in VIX futures recently exceeded the level observed during the financial crisis of 2008, indicating that sophisticated traders and investors may be preparing for an end to the recent run-up in equity prices.

The chart below shows the volume of trading in VIX futures (VX) since January 2008, along with a 20-day simple moving average of that volume. As evident there, the average volume in recent weeks is greater than at any time in the past two years. Increased VIX futures trading volumes are to be expected amidst market declines, given the increased activity observed in late November 2007 and October 2008. What is remarkable about the recent surge in volume is that it has coincided with a bull market in which declines are brief and mild.

Volatility Tracker for the week of October 26, 2009 (.pdf)

As I’ve noted on many occasions here, the relationship between spot VIX and longer-dated VIX estimates has not “worked” as a directional indicator for at least several months. [7,8] This looks like a genuine puzzle: the premium VIX futures traders are willing to pay and/or requiring in order to sell is too steep and has been too persistent to be dismissed as a phenomenon typical of the “wall of worry” that bull markets proverbially climb. But neither is that premium indicative of some impending crisis - at least, it wasn’t this summer and hasn’t been this fall. It is tempting to suppose that the effects of artificial government liquidity are overwhelming whatever information this relationship would otherwise provide. But whatever the cause, I won’t regard the VIX term structure data as meaningful until there is some new reason to do so.

Volatility Tracker for the week of October 19, 2009

The time to write "the fear is gone" stories will be with a VIX at twelve, not twenty-two.

Traders have become increasingly focused on the role that the dollar is playing in exacerbating rallies in equities, gold, and other assets, with concern in some quarters about a more precipitous dollar decline to come. At least on the implied volatility front, there’s no sign of such worries yet. EVZ, the index that tracks VIX-style implied volatility for FXE (a EUR/USD ETF) is pushing to new all-time lows.

Note that sudden price spikes in recent history correlated with moves higher in implied volatility; however, the relentless trend of summer and fall 2009 has not seen any corresponding increase in implied volatility. I regard this as confidence among traders that there is little concern of a dramatic rise in EUR/USD, even if the current trend does continue.

CME Group (CME) published the following product update today:

Effective Sunday, October 11, 2009 (trade date Monday, October 12), the following changes to the listing rules for COMEX Gold options will occur:

As Felix Goltz and Wan Ni Lai wrote in the Fall 2009 issue of The Journal of Derivatives:

An at-the-money (ATM) straddle -- i.e., going long on an ATM call and an ATM put with the same maturity -- is generally thought of as a volatility trade. It is essentially delta-neutral, but a large price move in either direction or an increase in implied volatility will produce a profit. A delta-neutral straddle position also has zero beta, so under the CAPMit should earn the riskless rate. Research has shown, however, that straddles with stock index options tend to lose money, which may be attributed to a volatility risk premium: it is the cost of hedging against a rise in volatility. If buying straddles produces losses, writing straddles should yield excess profits. An important aspect of the trade is that the delta (and beta) of the position change when the underlying index moves away from its initial level, and rebalancing is necessary if one wishes to maintain neutrality.

The spread between the volatility realized in the S&P 500 over the last 30 calendar days and the volatility implied in S&P 500 options 30 calendar days ago is about the widest that it has been this year, with the exception of a similar instance in early August [see charts 5,6]. That means stocks haven’t been nearly as volatile as index option prices have assumed they would be.

Put another way, it has paid to be a net seller of options over the period. Whether this persistent “overbid” in implied volatility will continue depends entirely on the risk appetites of traders and on the ability of the market to stay calm and gleeful – an increase in days like Wednesday and Thursday of last week would make options prices more nearly fairly valued.

Monday morning greets us with moves lower in gold, silver, equities, and oil and what looks to be the first consecutive day of gains in the US dollar index since late August. If recent behavior continues, these moves will be reversed promptly, but many traders are increasingly watching for signs that the summer reflation rally may be nearing an end.

Clearly, gains of this magnitude at this rate are only sustainable if you think stocks are still undervalued by half, but at the same time there is no reason to establish major short positions. If equities do decline into the fall, absent some surprise catalyst it seems more likely that they will roll over rather than correct sharply. I would be more sanguine about a gentle correction or continued rally if implied correlations came down form here [see chart 10].

The dollar continued slouching toward Harare, and it is anybody’s guess whether the trend will reverse anytime soon. I track price and volatility changes in FXE, the USD/EUR ETF [see charts 2,3]. The implied volatility index for gold closed down nearly 6%, in line with the expectations expressed last week.

I can’t recommend enough giving a weekly glance at the Implied Daily Move table. These are simply the point and percentage changes (within one standard deviation) implied by the volatility index for each asset, deannualized to reflect daily movement. Where technical analysis might regard a VIX at $24 as somehow cheap – especially in relation to its levels one year ago – any casual market observer knows that, if anything, implied volatility in equities remains on the high side. We simply aren’t seeing 1.5% daily moves in the S&P 500, and until we do, traders should expect options to remain overpriced.

The biggest mover last week on the volatility front was gold, as the "gold VIX" (GVZ) closed 40% higher [see charts 2,3]. Recall that when equity index prices rise, implied volatility typically falls; this inverse relationship does not hold for gold and many other commodities. On the contrary, the volatility implied by option prices will often rise with commodity price increases - a state of affairs with significant implications for option traders. We are short some out of the money gold futures options in managed accounts, based not on any directional thesis but rather in order to capture the relatively rich premium here.

The equity volatility indexes (VIX, VXN, VXD, RVX) all closed higher for the first time since August 2nd, but the steady bid in longer-dated volatility futures persists, with a spread of about 4-5 points between the spot and average futures prices [see chart 7]. This relationship was covered ad nauseam over the summer and I will resist speculating about its significance any further for now, except to note that a baseline level of 29 in the VIX futures has not historically been correlated with raging bull markets. The implied correlation index remains elevated [see chart 10].

People have been up in arms for months now about the troubles at the United States Natural Gas Fund, LP (UNG), the ETF designed to track the price of natural gas. And, as far as I can tell, rightly so: the whole point of ETFs is that they were meant to be a nearly frictionless, relatively simple alternative to the clunky closed-end funds (CEF) and managed products that our parents and grandparents had to contend with. But UNG recently traded at a 19% premium to its net asset value (NAV), behavior far more fitting a CEF. And until UNG begins issuing new shares, they functionally are a closed-end fund.

Why the premium? Regulators are concerned that, given its size, UNG purchases of natural gas futures could overwhelm and distort the market. Some sort of CTFC action is expected, and one review suggests that any likely outcome will result in a dissipation of the fund’s premium to NAV. That’s a good reason not to buy UNG here, and it may offer a good opportunity for a stat arb play for those with the stomach and expertise: the trade would be to buy natural gas futures and sell short the ETF (or, since the shares are hard to borrow, competent options traders will establish a synthetic short position by selling calls and buying puts with the same strike price and expiration date) on the expectation of UNG eventually returning to its net asset value.

The CBOE Implied Correlation Index spiked to its highest level last week since the beginning of the rally that began this spring. [10] In a healthy, normally functioning market, companies that succeed will see their stock prices rise, while the stocks of failing companies will fall. In a healthy, normally functioning market, the stocks of winners and losers alike won’t rise or fall together in lock step; but the increase in [10] denotes increased expectations of just such a phenomenon. There’s an old adage that, in a crisis, all correlations go to 1. As the implied correlations of S&P 500 stocks push toward crisis levels, momentum- and sentiment-based long positions should be watched carefully. It is intuitive that, as we watch “junk” and quality stocks both rise week after week, expectations of high correlation are just a proxy for the efficacy of reflation efforts. But indiscriminate reflation is no more a sign of a healthy economy than are falling stock prices.

Following up on last week’s discussion of price distributions and kurtosis, I ran a simple strategy test beginning in 2000 that buys the S&P 500 when its one year kurtosis is above a 200-day simple moving average, and moves to cash otherwise. Returns weren’t outstanding, but still beat the market handily, and (more importantly) the strategy was in cash about half the time. Accounting for returns on cash would obviously improve results. I don’t have any broad lesson to extract from this, except perhaps that no expensive analytics software package can remove the need for a basic knowledge of statistics.

The financial sector of the U.S. economy has had nearly a year to address the problems that exacerbated the crisis last fall. But many observers think that the banks haven’t done enough, and that another round of trouble may be developing for the sector. I will outline some of those concerns and then suggest some ways to use options to profit if there is indeed another shoe to drop in banking.

The Thesis

'Reflation’ is the name of the conceptual cure for the cognitive dissonance experienced by any rational observer of this market. Equities returned to short-term overbought status this week ^[4] and options are relatively fairly priced ^[6]. At 16%, the 21-day realized volatility of the S&P 500 is as low as it’s been since the financial crisis began, though it would be folly to try to call a bottom in volatility, or a top in price.

I haven’t discussed the daily return distributions very much in the past, but they are a helpful way to get an alternative look at recent price action. Notice that, over the last three months, daily returns for the S&P 500 have not been normally distributed ^[9].

Conventionally, equity prices and implied volatility are inversely correlated, meaning that traders who expect a price decline should be net buyers of options. But as long as the ratio of lagged implied and realized volatility remains this high [see charts 5,6], it makes sense to be a net seller of equity index options, even alongside the expectation of a modest price decline.

Regarding index prices, I would only mention that an "overbought" condition can be resolved by time just as easily as by price: a week or two of sideways trading would negate most bearish signals just as easily as if equities retraced some of their recent gains [see chart 4]. Volatility futures barely twitched this week [see chart 7], but the persistently high implied correlation warrants further study [see chart 10].

It seems these days that the entire world is gasping at, of all things, minor twitches in out-month VIX futures. Earlier this week, Bloomberg’s strength ebbed and I caught them as, overcome, they whispered, “VIX Signals S&P 500 Swoon as September Approaches.”

Like any financial headline that calls to mind men in breeches and whooping cough, this one should be regarded with skepticism. In the first place, VIX futures are signaling no such thing: even if the VIX popped up five points over the next week, that wouldn’t necessarily coincide with any rally-crushing, chaise lounge-requiring swoon.

Even as the S&P 500 pushed further into overbought territory last week [4] and the VIX Premium Ratio moderated back toward its normal range, [8] we saw an interesting development in the ratio of current realized to lagged implied volatility. [6] To review, that ratio plots the current 21-day realized S&P 500 volatility over the VIX reading from 30 calendar days ago. Essentially the ratio asks how well implied volatility estimates from one month ago predicted the volatility that was realized over the subsequent month. Much of the time, this ratio isn't of particular interest, but moves into extreme territory have coincided with major shifts in the market. I regard the area of 0.9 -1.2 as a neutral area, with readings below and above indicating that price behavior has been respectively more or less volatile than expected.

The ratio touched the 2.0 area last week, a level not seen before since the inception of this letter. One way to interpret that reading is as a sign that index options have actually been overpriced over the past month, as unlikely as that may seem. I would caution against regarding this ratio as a predictive indicator in itself. Given the mean-reverting nature of volatility it is reasonable to expect extreme readings to moderate fairly quickly, and to some extent this has already happened: the VIX at 30 figured into [6] is 20% higher than the current VIX price, and that decline will be reflected as the ratio rolls forward. Even so, realized volatility continues to decline as well, and I do not expect it to remain below 20 for very long, so equity longs in particular should consider tightening stops or taking some profits in unhedged positions.

Although the major equity indexes are highly correlated, they tend to diverge in meaningful ways on a day-to-day basis. The same is true for the implied volatility indexes that track equities. It’s also common for equities and implied volatility to move inversely on a daily basis – the expectation is that, most of the time, if the S&P 500 closes up, the VIX is likely to close down.

It’s very uncommon for both the equity indexes to all close higher and for their implied volatility indexes to all close higher on the same day. To clarify, the condition we’re looking for is one in which all of the S&P 500 (SPX), the Dow Jones Industrial Average (DJIA), the Russell 2000 (RUT), the Nasdaq 100 (NDX), and their respective implied volatility indexes (VIX, VXD, RVX, VXN) all closer higher than the day before. Since 2004 – the first date for which RVX data are available – that has only happened 19 times; the most recent instance was last Monday, July 27.