August-September 2013 Results
Executive SummaryDuring the month from the fund's initial offering on 15 August to 14 September, net asset value after fees increased from 0.0998 BTC per share (i.e., .1 BTC per share at the offering, net of the exchange's 0.2% transaction fee) to 0.1012 BTC per share. This represents an increase during the month of 1.403%, or 18.20% on an annual basis.
Gains in equity derivatives and forex derivatives offset losses in equities, while a backdrop of debt delivered less volatile returns. As a new fund, the cost of moving from cash to new positions with sometimes exorbitant bid/ask spreads was significant, creating immediate losses -- relative to market bids -- which in some cases approached high single digits in percentage terms.
The fund's initial allocation of capital ensured that no single entity held more than approximately 30% of the total, achieving sufficient diversification that even the complete failure of an entire exchange or an entire asset would have left the vast majority of our capital safe and untouched by such an event.
Context and EnvironmentDuring the period, the predominant trend in Bitcoin-denominated 'equities' -- a term which, for convenience, I'll treat as including revenue sharing operations -- was negative, with the most striking exceptions being would-be mining companies that engage in little or no actual mining. Market participants willing to speculate on individual 'faithware' companies -- i.e., those which aspire to become mining companies one day -- and willing to concentrate capital in a small number of such speculative positions, would clearly have outperformed the fund's losses in equities.
The predmoninant move in BTC vs. USD, by contrast, was positive, while volatility remained subdued for nearly the first three weeks. The fund's position was not optimal with respect to volatility, but being delta positive relative to BTC vs. USD enabled it to benefit from the directional move nonetheless. Market participants taking entirely unhedged long positions in BTC vs. USD, especially during the first part of the month, would have outperformed the fund in this area.
Listed debt showed no such sustained directional moves, and the fund benefited from its higher weighting in debt than in equity.
Liquidity and open interest remained anaemic in both equity options and forex futures; with respect to the latter, during much of the month, our own positions represented 5% or more of all open interest in the contracts in which the fund was active. (And as I have argued elsewhere, the Bitcoin economy still lacks a credible and cost effective forex
options market, as distinct from a forex futures market; forex options would have made the fund's operations significantly more efficient and straightforward.) The fund's ability to execute otherwise viable strategies using derivatives was significantly constrained by the low liquidity, especially in equity options, as well as by repeated platform failures on our home exchange, which compromised the timeliness of option re-pricing. In addition, the exchange's requirement that put writing be covered 100% by actual cash means that writing such options at all -- even if they are never purchased, let alone exercised -- creates significant opportunity cost by preventing the capital from being deployed elsewhere.
(The 100% cash requirement contrasts sharply with much lower standardized margin requirements familiar in the fiat world. Among other differences, no
bona fide market exists for Bitcoin-denominated equity options, and market makers are entirely absent. Therefore, an investor who is short in-the-money puts cannot, except in artificial examples, simply close the position. Instead, they remain on the hook for the full exercise, even if much of the required capital can be immediately recovered by liquidating the shares.)
During the period, the fund's initial allocation of capital ensured that no single entity
or platform -- including individual exchanges -- held more than approximately 30% of the fund's total. This is signficant not only from the perspective of reducing counterparty risk and company-specific risk, but also from the perspective of mitigating impact from the types of platform failures mentioned above. During the entire period from when we began logging platform failures on 17 August to 14 September, there were
only 6 individual days on which we didn't log at least one outright failure or other anomolous behaviour by at least one of the platforms we used. The most common such failures were CloudFlare error 524 and BTC-TC's "Access Temporarily Denied" message. (With regard to the latter, it seems a stretch to suggest that such frequent failures are solely a result of correctly functioning defenses such as packet filtering or rate-limiting measures.)
A platform problem of my own, in the shape of my primary work computer's abrupt hardware failure on the afternoon of 9 September, compromised my personal efficiency during the entirety of the week. Thanks to redundant backup strategies, data loss was never an issue, but the job of diagnosing, replacing, and re-configuring, all while working from two older and less capable backup machines, came at a distinctly inconvenient time.
Finally, fund participants should note that entering any new position at all in a listed asset entails either waiting a potentially very long time for suitable market sell orders or absorbing the sometimes exorbitant cost implicit in the bid/ask spread. As a new fund, therefore, it should not be surprising that we have conceded non-trivial amounts of capital to the spread simply as a result of entering positions at all.
For the purposes of calculating net asset value, individual positions have been marked to market using either the 7-day trading average where platforms provide this information directly, or the instantaneous market value where they do not.
Individual SecuritiesThe fund has been active across around a dozen platforms and many different businesses, but as illustrative examples I will offer some thoughts on two specifically, with a view to highlighting a little of the fund's general approach.
ASICMinerWithout entering into meta-arguments about the sustainability of Bitcoin mining or the wisdom of investing in it -- topics which I have addressed elsewhere and which many others have addressed at greater length and in greater detail -- the fund does have some exposure to ASICMiner, a security which attracts broad interest and which in passthrough form offers sufficient liquidity to support meaningful and profitable derivatives strategies. Primarily for the purpose of applying simple derivatives strategies on BTC-TC, the fund began establishing a position in ASICMiner shortly before the share price corrected severely in the second half of August.
The timing resulted in an immediate and negative hit to the value of the fund's position.
However, while the overall position is relatively small, the results of the fund's follow-up action are illustrative. Thanks to healthy option premiums, 6 of the fund's shares were sold at net prices above the market at the time, while 16 shares were purchased at net prices below the market -- providing a cost basis as low as 1.2 BTC each, compared to a market price at the time of around twice that. (Notably, such low net entry points are outliers, due more to inefficiencies in the market than to anything exotic in terms of the structure of our strategies.)
As I have suggested previously, "investing in mining" need not be a black and white, do-or-don't issue; nor is it either a guaranteed way to print free money or a guaranteed way to destroy shareholder value.
Ukyo.LoanDuring the month, the fund acquired roughly 11.6% of all outstanding units of
Ukyo.Loan, making it the second largest holder of the loan as of this writing.
Ukyo.Loan, issued by the owner and operator of the BitFunder exchange, is a peculiar creature. Unlike a normal bond, it was not offered at a discount to face value, to be redeemed at that face value upon maturity at a specific date, and with the interest rate therefore set implicitly by the price at which it trades. (In this sense, the loan is similar to other 'bonds' in the Bitcoin world.) Instead, its 'face value' serves two purposes and two purposes only:
- Interest, paid in the form of dividends, is guaranteed to be at least .05% of face value per day.
- Units can be redeemed for face value on demand.
The loan is also described explicitly as being "Similar to Graet.Loan", but it differs in two very significant ways. First, the interest paid indirectly reflects the success of the BitFunder exchange itself, with the Ukyo.Loan specifically described as a sort of roundabout investment in the exchange and the description further indicating that "Additional dividends provided when the site does well". Therefore, while the interest rate never drops below .05% of face value per day, it
can climb significantly above that rate. Second, while the units can be redeemed for face value on demand -- a fact which effectively places a floor on the price at which units should trade -- any
buybacks must, according to the contract, occur at "110% of the last 30 days average market price".
The combination of these two facts suggests that Ukyo.Loan
should trade so as to reflect the market's aggregate estimation of its risk-adjusted return. Unlike Graet.Loan and other similar loans, there is no overhanging threat that units will be bought back at face value, so there is no mechanism in place to force units of the loan to trade near that value. On the contrary, those willing to
hold the Ukyo.Loan (and not looking for a quick exit via sale to another party) appear to be essentially
guaranteed to break even or make a profit when buying at any price below 110% of face value, since they cannot be forcibly bought out at anything less than 110% of a 30-day trailing average that should never cross below face value. (Trading below face value would offer 'free money' arbitrage via immediate redemption at face value, so while it might occasionally happen, the probability of a 30-day trailing average below face value should be negligible.)
Given the absence of any cap on the unit price created by the threat of a forced buyback at face value, those who are willing to hold Ukyo.Loan at a price of .01 BTC per share for .05% daily yield relative to face value should, other things being equal, be willing to hold it at twice that price in return for twice the yield -- a price of .02 BTC per share for .1% daily yield relative to face value.
(ADDENDUM TO THE ORIGINAL: For anyone who either accidentally or intentionally would like to interpret the previous sentence as implying or suggesting, in any way, shape or form, that I somehow believe that other things
are equal, I am making no such suggestion. I am, on the contrary, illustrating how normal bonds trade in the real world. And in the real world, we can
assume that other things are
not equal when yields change; that is why bond prices change so as to bring the risk adjusted return of a bond in line with the risk adjusted returns of other comparable assets. That is also why higher yields typically correlate with lower quality offerings: issuers of lower quality, higher risk paper must pay higher yields in order to counterbalance their higher risk.)
Yet historically, the price of units of Ukyo.Loan has barely responded to radical changes in yield. From inception (19 July) to 1 September inclusive, by which time the bulk of the fund's purchases had been completed, the loan paid out the equivalent of 10 days of dividends at its normal rate, then 11 days at 150% of its normal rate, then 20 days at 200% of the normal rate, then 4 days at the normal rate. (The loan continued to pay out at its normal rate from the beginning of September to this writing.) Despite the fact that the yield climbed to 200% of the guaranteed rate for a sustained period of 20 consecutive days, the price per unit barely budged during that time, trading at no more than 20% above face value. (Up until 11 August, it traded largely at or very occasionally below face value.) For reference, paying 20% above face value reduces the effective return of the loan from approximately 18% per annum to approximately 15% per annum, or from approximately 36% per annum -- when it is paying 200% of its guaranteed rate -- to 30%.
As far as I can tell, the most plausible justifications for such extreme discounting of potentially above-guarantee returns as reflected in the unit price of Ukyo.Loan are along the following lines:
- Potential holders believe BitFunder is doomed, and therefore future returns from the loan will not benefit significantly from its success.
- Potential holders believe Ukyo is extremely untrustworthy or is extremely likely either to alter the loan contract unilaterally to the detriment of creditors or to breach the terms of the contract altogether. (Note that such ex post facto unilateral contract modifications are permitted by BitFunder.)
- Potential holders are so strongly wedded to the 'get out clause' provided by face value redemption on demand that they will not pay anything for higher yield that is not accompanied by a higher guaranteed redemption price.
- Potential holders have not fully attended to the difference between the buyback provision for Ukyo.Loan and the analogous provisions of related loans which rarely deviate significantly from face value.
As for the first two, I personally have no idea whether BitFunder is doomed, and I have no idea whether Ukyo is so untrustworthy as to effect a radical and unilateral change to the contract to the detriment of those holding the loan. However, to the extent that either of these types of explanations plays a role in current pricing of Ukyo.Loan, I believe the market's pricing reflects a degree of skepticism about BitFunder and/or about Ukyo's trustworthiness that is unsupported by publicly available evidence. As for the second two, it seems likely to me that any influence from factors like these is apt to decrease over time.
The upshot is that Ukyo.Loan appears to me to be mispriced, and I believe there is a worthwhile probability that this mispricing will be corrected over a time horizon relevant to the success of the fund.
Liquidity Facility and Subscriptions/RedemptionsAs many participants will be aware, the fund's initial offering of 20,000 shares on 15 August was entirely subscribed in less than 30 hours. Immediately following the initial offering, the fund traded as high as a 7% premium to NAV, but the fund refrained from interfering with the market's natural pricing activity and provided no further liquidity either on the bid or ask side. (As described in the listing documents, the fund may optionally provide additional liquidity when bids or asks fall significantly below or above NAV, respectively.) Some days later, after the initial flurry of activity had settled, the fund did offer limited additional liquidity above NAV, but just one additional share was issued before other market participants stepped in and enhanced liquidity.
Subsequent to this report, the fund may make additional shares available for a time somewhat above NAV.
Note that without interference or 'management' of the spread in any way, the market has already set a difference between bids and asks for the fund which frequently ranges from less than 1% to a bit over 2%, with the spread widening somewhat during the days immediately prior to this report. For the avoidance of doubt, I have no intention of turning the routine maintenance of a wider and less favourable spread into an ongoing profit engine for the fund. The reality is that although wide spread market making by funds' own managers can create the
appearance of a performance boost that may appeal to the casual onlooker, in fact it represents an indirect trading tax on fund participants, a tax taken directly out of participants' pockets when buying, when selling, or both. This practice is designed to subsidise early and ongoing shareholders' returns on paper by taxing newer shareholders' and departing shareholders' trading activity. It has nothing at all to do with real fund performance. After all, a trained monkey can generate 12% 'gains' per year simply by skewering participants with a 5% bid/ask trading tax on a monthly churn of only 20% of a fund's total capitalisation; double the spread or the monthly churn, and presto, easy 24% annual 'gains'. It's remarkable that this practice survives at all in a community which is otherwise justifiably hypervigilant about schemes designed to pay early investors using money extracted from later ones.
Finally, a few individuals considering relatively large positions in the fund have been in touch, and I have discussed with them the option of direct share transfers. Since it benefits the fund not to have to pay transaction fees on new share subscriptions, potential participants considering positions in the fund of 5000 shares or larger are welcome to get in touch with me about arranging direct transfers, internal to the exchange. Note that as with any other new share issuance, this facility will only be made available during a period immediately following publication of the fund's latest NAV. (It has, therefore, not been made available to anyone prior to this report.)
Reporting ScheduleI believe it would be useful to synchronise fund reporting with calendar months, rather than with anniversaries of the fund's initial offering date which, as luck would have it, fell squarely in the middle of a calendar month. Therefore, I'm considering extending the next reporting period so as to cover thru the end of October, rather than thru the middle of October. Unless or until any public announcement is made about synchronising the next reporting date, however, it is safe to assume that it will continue to fall on an approximately monthly cycle and will therefore next occur around mid-October.
Trust RatingsAs I mentioned in the fund's listing documents, I have no particular interest in amassing trust ratings for a pseudonymous WOT account. However, if participants in the fund feel it is appropriate to do so, I would invite them to consider leaving trust feedback for my forum account, not as a reflection of positive or negative fund performance, but rather as an indication of how they regard the trustworthiness of my management of the fund. (It is virtually certain that the NAV of the fund will go down as well as up, but hopefully variation in fund performance need not imply variation in trustworthiness!)
