Our Indicator Performance: +71% YTD Through May

Inverse volatility ETFs continued their uptrend in May with XIV adding 13.4% during the month, bringing its year-to-date gains up to +45%. Here's a look at XIV's YTD chart:

A 45% gain in the first five months of the year is pretty good, but our indicators are all outperforming the buy-and-hold approach. Here is the performance of our indicators in 2015, through May 29th:
-  VXX Bias:             +69%
-  VRP:                    +60%
-  VRP+VXX Bias:     +71%

The equity curve for each strategy with comparison to XIV is shown in the following graph (the S&P 500 is +2.4% YTD and is not shown):

Trade statistics for each strategy are summarized as follows:

If you're not quite meeting your investment objectives this year perhaps we can help. For less than $3 per day we provide subscribers access to all of our VIX data & metrics, our indicators, automated change alerts, and the members' forum.

For those interested in additional detail, the daily history of indicator values that have been emailed to subscribers in 2015 can be found in the data sheet links below:

- VXX Bias: 2015

- VRP: 2015

- VRP+VXX Bias: 2015

______

Hypothetical and Simulated Performance Disclaimer

The results are based on simulated or hypothetical performance results that have certain inherent limitations. Unlike the results shown in an actual performance record, these results do not represent actual trading. Also, because these trades have not actually been executed, these results may have under- or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated or hypothetical trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to these being shown. Additional performance differences in backtests arise from the methodology of using the 4:00pm ET closing values for XIV, VXX, and ZIV as approximated trade prices for indicators that require VIX and VIX futures to settle at 4:15pm ET.

A Note on Corrective Distributions for VXUP & VXDN

I've outlined the basic operation of AccuShare's new VXUP and VXDN funds in this post (as they are designed, but not how they'll actually be able to operate). However I did not go into sufficient detail on the topic of Corrective Distributions.

Before I get into that though, I'll mention that AccuShares has already told us that VXUP and VXDN are intended only for sophisticated professional and institutional investors. If you're not in that category you will likely find the following information less useful since you won't be trading these products.

Corrective Distributions are used to help keep the funds trading near their NAV (Net Asset Value). In my previous post outlining the Fatal Flaws of VXUP & VXDN, I explained why it will be common for these funds to trade at a premium or discount to NAV. In fact, as of Friday, VXUP is already trading at a 19% premium to NAV while VXDN is trading at a 17% discount to NAV. Recall that if the closing trading prices deviate from their NAV by 10% over three consecutive business days, the Fund will make a Corrective Distribution on the next scheduled Regular Distribution Date or Special Distribution Date (note: this rule goes into effect only after 90 days since the fund's inception has passed). 

Here's what happens during a Corrective Distribution:

• First, the Regular or Special Distribution takes place (as outlined in my first post on the topic). 
• Next, each remaining share will be resolved into a risk neutral position comprised of an equal number of Up Shares and Down Shares, based on the share NAV. 

The key item here is that the value is based on NAV, not the market trading price. Therefore, holders of a security that trades at a premium to NAV will take a loss equal to the amount of the premium on Distribution Day. Likewise, holders of a security that trades at a discount to NAV will see a gain equal to the amount of the premium. 

A Corrective Distribution trigger should help close the gap to NAV in advance of the next Distribution Day so that no such "easy money" can be made/lost. Even the threat of a Corrective Distribution trigger (trading at >10% NAV on the third day) could bring about swift moves towards NAV for each of the funds as traders move to close out positions while the premium to NAV is still double digit percentages, or rush to get into positions that are trading at a large discount to NAV. This pressure may even be enough to prevent the completion of a Corrective Distribution trigger on the second or third day of a threat, resulting in wild daily price action.

After a Corrective Distribution Day I expect the funds to immediately diverge from NAV once again. If the VIX futures term structure is sufficiently steep, the funds will face a threat of a Corrective Distribution trigger again for the following month. This will make it important to constantly keep an eye on the market prices versus NAV and factor these potential events into your trading strategy.

The Fatal Flaws of VXUP and VXDN

Today's post is a follow up to last week's post, "Details on New VIX ETFs: How VXUP and VXDN Work." In hindsight, the title should have included the words "in theory" because the reality of how these funds work is drastically different. The funds are subject to strong arbitrage pressures from VIX futures and will constantly be trading at a premium or discount to the funds' theoretical price, known as Net Asset Value, or NAV (also sometimes referred to as the fund's "Intrinsic Value").

In real trading, the major problem with the concept of these new funds is the presence of a VIX futures term structure that can range +/-20% at times. When AccuShares announced the launch of an ETF that tracks spot VIX, everyone immediately looked forward to getting in on the arbitrage trade to be long VIX and short front month VIX futures during contango. After all, spot VIX and futures converge each month and this trade is essentially free money. The reason for the price difference between spot VIX and futures stems from the principle that VIX is mean reverting. This generally keeps the price of VIX futures higher than spot VIX when VIX is lower than ~20 (the long term average) and VIX futures lower than VIX when VIX is above ~20.

We saw this arbitrage gap close quickly in the first day of trading of VXUP and VXDN. VIX was +0.94% on May 19, while $VXUP was +10.56% and $VXDN was -10.56%. That's not at all tracking towards spot VIX. That is adjusting toward June VIX futures. Such behaviour to price in known information can be expected in a free market when there are obvious pricing discrepancies in what are essentially identical products.

June VIX futures settle on June 17, which is only two days away from the VXUP/VXDN Distribution Day on June 15. Given that the objective of VXUP & VXDN is to track the monthly change in spot VIX (plus an adjustment for the "Daily Amount"), the funds immediately move toward the expected value of VIX on June 15 -- the fair value of which is already agreed upon by market participants by the price of June VIX futures.

There is nothing unique to the month of June that causes these funds to suddenly track at a wide premium and discount to NAV. In fact, the funds will stray from NAV on almost every day in order to account for the value of VIX futures. Generally, the stronger the degree of contango or backwardation the larger the premium or discount to NAV.

It seems that AccuShares (the funds' sponsor) understands these forces, hence the reason for an arbitrary 0.15% "Daily Amount" which is subtracted from VXUP and added to VXDN when VIX is below 30 (another arbitrary threshold). This 4.5% roll yield that was built into the fund operation attempts to mitigate the effects of the contango roll yield. When the VIX futures curve is greater than a 4.5% contango, the only logical purchase is VXUP, leaving no one to buy VXDN.

Using yesterday as an example, an investor looking for protection can either buy June VIX futures at 14.85, or purchase VXUP at a price reflecting a spot VIX of 12.73. If VIX rises to 14.85 by June 17th, the buyer of VIX futures broke even while the purchaser of VXUP gained 11.2% (15.7% minus the Daily Amount of 4.5%). Arbitrageurs are well aware of this and will keep the price of VXUP and VXDN away from their NAV in an amount that reflects the degree of contango or backwardation in VIX futures. VXUP and VXDN will trade at values which close the arbitrage gap, not at NAV.

Today is May 20 and is the second day that VXUP and VXDN have traded. As of yesterday VXUP traded at a 9.69% premium to NAV (which can be tracked here). Today that premium is over 10%. Given that a "Corrective Distribution" is triggered whenever the funds are at a premium/discount of over 10% for three consecutive days it is possible that we'll see triggers often. However, a Corrective Distribution trigger only helps to bring the NAV premium closer to zero as Distribution Day approaches. This does not keep the funds trading near NAV after Distribution Day since the arbitrage gap to VIX futures does not go away.

I had previously thought that the premium/discount to NAV might diminish as the next Distribution Date approached, but it is clear now that if that were the case there would be an immediate and large gap on the following morning, just as we saw on the first day of trading in VXUP and VXDN yesterday. Instead, I expect the premium or discount to be persistently present, depending on the degree of contango or backwardation in VIX futures.

I don't expect the daily movement of these products to track as advertised. I also expect Corrective Distributions to be common as the funds stray 10% away from NAV. This will make the predictability of these funds incredibly difficult for anyone who has not built an arbitrage model.

When it comes down to it, one phrase at the top of the fund's Factsheet sums it up best: "shares of the Fund are intended for sophisticated, professional and institutional investors."

Details on New VIX ETFs: How VXUP and VXDN Work

Edit 5/21/2015: A follow up article, "The Fatal Flaws of VXUP and VXDN" has been posted.
--

On Tuesday next week we will see a new pair of VIX ETFs hit the market: VXUP and VXDN from AccuShares Management. Unlike existing VIX ETPs which move based on VIX futures, these new ETFs will move based on the price of spot VIX.

However, the operation of these funds is not as straight forward as one might expect. I've gone through the 150 page prospectus and have briefly outlined what you can expect from these new products (the full prospectus is available here for those interested in additional details).

Summary

• The Up Class Shares (VXUP) and Down Class Shares (VXDN) are designed to track changes in the price volatility of the S&P 500® Index on a monthly basis.
• The Underlying Index of the Fund is the CBOE Volatility Index (VIX)
• Investors who expect the VIX Index to increase should consider purchasing Up Class Shares (VXUP)
• Investors who expect the VIX Index to decrease should consider purchasing the Down Clas: s Shares (VXDN).

Basic Fund Operation

• The operation of the funds centers on "Distribution Periods" which are marked by "Distribution Dates" (the 15th of each month).
• The primary change in value of the funds is based on the "Share Index Factor", which is reset on each Distribution Date. The Share Index Factor defines how much VXUP and VXDN move for every VIX movement.  
• --> Example: If VXUP is $25 (after distribution) on the Distribution Date, and VIX is at 20, each 1 point (5%) move in VIX is worth $1.25 (5%) to VXUP for the next month.
• Additionally, when the VIX is 30 or less, a "Daily Amount" of 0.15% ($0.0375 in this example) will be deducted from VXUP per day and will be added to VXDN.

Distributions
A primary concept of the funds are periodic Distributions. These will be predominantly done in cash for the first six months, but will be in the form of issuing paired shares of VXUP and VXDN thereafter (depending on the fund's liquidity). Here are the Distribution types:

1. Regular Distributions: Occur on the 15th of each month (the "Distribution Date").
• VXUP will be entitled to a distribution when the VIX has increased between Regular Distributions, or by 75% (“Special Distributions”).
• VXDN will be entitled to a distribution when the VIX has decreased between Regular Distributions, or by 75% (“Special Distributions”).
• --> Exception: Both share classes subject to a maximum of 90% in either direction for a single Measuring period.
2. Special Distributions: Occurs if change in VIX is +/-75% since last Distribution Date. The Shared Index Factor is also reset.
3. Corrective Distributions: Following the expiration of 90 calendar days following the inception of the Fund’s operations, if the closing trading prices of the shares of the Fund deviate from their Class Value per Share by ten percent over three consecutive business days, the Fund will make a Corrective Distribution in addition to a Regular Distribution or Special Distribution on the next scheduled Regular Distribution Date or Special Distribution Date if previously triggered.
4. Other distributions may result from reverse share splits.

Any distributions of cash, or cash and shares, will reduce your exposure and opportunity for gains arising from changes in the Fund’s Underlying Index in subsequent periods. You will need to rebalance/buy additional shares to maintain desired exposure. This concept will be made clear in the example below.

Price Movement Examples:
VXUP will often face a headwind caused by the "Daily Amount" of 0.15%. The table below summarizes the value of VXUP and VXDN for various VIX scenarios after a Measuring Period. Note that when the VIX is +4.5%, the gains in VXUP are offset by the losses from the Daily Amount (since 0.15*30 = 4.5).

In the next table, the VIX is assumed to have moved up 10% from 20 to 22 over 30 days. At the Distribution Date, the result is a 5.5% increase for VXUP and a 5.5% decrease for VXDN based on the combination of the VIX movement and the $0.0375 Daily Amount.

Note that the Fund will distribute a cash amount of $2,750.00 per 1,000 shares of VXUP. The new per share value for each fund is then set at $23.6250.

If instead there is a 10% decrease in VIX from 20 to 18 over 30 days we will have the following:

Here, VXUP will return -14.5% and VXDN will return a total of +14.5% based on both change in the VIX and the effects of the Daily Amount.

It is critical to note that after each favorable movement in a security held, your exposure is reduced due to distributions. In the case of issuing paired shares in lieu of cash, your exposure is still reduced and you would need to sell/buy shares as necessary to maintain a desired exposure.


It is worth highlighting that the Daily Amount only exists when VIX is below 30. When above 30 it is zero so the VXUP headwind goes away.


Supply & Demand and the Arbitrage Mechanism for creation/redemption of share units

• As the Fund’s shares trade intraday, their market prices will fluctuate due to simple supply and demand. 
• An arbitrage mechanism helps to minimize the difference between the trading price of a share of the Fund and its Class Value per Share. Over time, these buying and selling pressures should balance out, and a share’s market trading price is expected to remain at a level close to its Class Value per Share. The arbitrage mechanism provided by the creation and redemption process is designed, and required, in order to maintain the relationship between the market trading price of shares and their Class Values per Share between Distribution Dates.

The facts outlined above cover the basics of these new funds. There are many other additional details and nuances covered in the prospectus which I encourage you to read through if you are considering trading these products.

Bonus: Below I've included some additional risk factors contained within the prospectus that I found to be of interest.

• By purchasing VXUP, investors should have an expectation that the Underlying Index will increase during the period between Distribution Dates. If the VIX decreases during the time between Distribution Dates, investors in VXUP will experience a significant loss and could lose their entire investment.

• The Fund’s Eligible Assets are not managed to provide a maximum long-term return and even a share class experiencing a favorable Underlying Index change can experience losses if the Fund’s aggregate Class Values decline significantly. If the Fund’s aggregate Class Values decline to zero, the Fund’s Up Shares and Down Shares will lose all value, causing a total loss to all Fund investors.

• The Sponsor Has No Experience Managing Investment Vehicles. The Sponsor is recently formed, and has not previously managed any investment vehicles. There can be no assurance that the past experience of the Sponsor’s management team will be sufficient to successfully operate the Fund.

Our Indicator Performance +56% YTD Through April

April was a great month for inverse volatility ETFs and XIV in particular as it gained nearly 16% over the month to settle at +29% YTD. The trend was generally strong and without drama this month although it did give back some gains in the last week. XIV's one-month chart:

Our daily indicators rode this wave of gains throughout the month on only one trade, with the VXX Bias generating an XIV buy signal on 3-20-15 and our VRP indicator signalling a buy for XIV on 4-1-15.

This month's gains added to already impressive year-to-date returns for our indicators. Indicator performance through April:
-  VXX Bias: +51%,
-  VRP: +48%
-  VRP+VXX Bias: 56%.

The equity curves for each strategy with comparison to XIV are shown in the following graph (the S&P 500 is +1.4% YTD and is not shown):

Trade statistics for each strategy are summarized as follows:

For those interested in additional detail, the daily history of indicator values that have been emailed to subscribers in 2015 can be found in the data sheet links below:
- VXX Bias: 2015
- VRP: 2015
- VRP+VXX Bias: 2015


______

Hypothetical and Simulated Performance Disclaimer
The results are based on simulated or hypothetical performance results that have certain inherent limitations. Unlike the results shown in an actual performance record, these results do not represent actual trading. Also, because these trades have not actually been executed, these results may have under- or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated or hypothetical trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to these being shown. Additional performance differences in backtests arise from the methodology of using the 4:00pm ET closing values for XIV, VXX, and ZIV as approximated trade prices for indicators that require VIX and VIX futures to settle at 4:15pm ET.