Bitcoin, still not quite ready for CDOs

Friday, 02 November, Year 4 d.Tr. | Author: Mircea Popescu

Back in April we explained to a mostly innocent community what a CDO is and how'd it work. The market at that time wasn't quite mature enough to actually implement one, so we let the matter lay for a while (some other people tried to implement what they understood of the text, resulting in varying degrees of failure).

Eventually we introduced one in August. Here's the original description of that issue :

Hash: SHA1

MPCD.x is a tranched CDO composed of the following assets :

i. 500 BTC to be invested in MPOE bonds starting the 1st of September for an asking rate of 4.99999999% a month (MPOE pays not on calendar months but on last-Friday-of-each-month) ;
ii. 500 BTC invested with Patrick Harnett's 1% a week compounding bond offering (the line PsHCai6pF2UH1475 500.9815 in the most recent stats update reflects this investment) ;
iii. 100 BTC invested with usagi's NYAN.A asset.

The theoretical monthly yields of these investments are as follows :

i. 500 * 4.99999999% = 24.99999995 BTC ;
ii. 500 * 4.4061192% (1% per week, thus (1.01^52)^(1÷12)) = 22.03059600 BTC ;
iii. 100 * 4.4061192% (1% per week, thus (1.01^52)^(1÷12)) = 4.40611920 BTC ;
Total : 51.43671515 BTC

This CDO is tranched as follows :

MPCD.A, 350 BTC paid 2% per month (1.02^12 = 26.8% apr) = 7 BTC ;
MPCD.B, 400 BTC paid 3% per month (1.03^12 = 42.6% apr) = 12 BTC ;
MPCD.C, 350 BTC paid 9.2676329% per month (1.092676329^12 = 189.6% apr) = 32.43671515 BTC;
Total : 51.43671515 BTC.

Each tranche will consist of shares nominally worth 0.001 BTC, so 350`000 shares for MPCD.A and MPCD.C, and 400`000 shares for MPCD.B.

This asset is created today, 15th of August 2012, will pay dividends on the 1st of each subsequent month for three months (that is, 1st of September, 1st of October, 1st of November) and then dissolve on the 1st of November. On each of the three dividend dates the exact sums quoted above will pe paid. On the last dividend date, all the remainder capital will be paid out, as follows : 350 BTC to MPCD.A holders. If less than 350 BTC capital remains the payment will be pro-rated (each holder gets his (shares / total shares) * remainder capital). Then, 400 BTC to MPCD.B holders. If less than 400 BTC capital remains the payment will be pro-rated (each holder gets his (shares / total shares) * remainder capital). Then, 350 BTC to MPCD.C holders. If less than 350 BTC capital remains the payment will be pro-rated (each holder gets his (shares / total shares) * remainder capital).

The fund has a little head start (because the investment with Patrick Harnett is already ongoing) which I am generously donating to smooth things over. The NYAN.A investment will also be realised as soon as deposits confirm.

Please bear in mind that the MPCD.C asset implies significant risk, no matter how attractive the >9% monthly interest might look. In case anything whatsoever goes bust the .C capital is first in line to pay for it. Other than defaults in any of the assets such loss could occur by the MPOE bond failing to activate (if sufficient capital is offered at a lower than the quoted rate the bond will lay fallow).
Version: GnuPG v1.4.10 (GNU/Linux)


Perhaps unsurprisingly, two months is too long in BTC "finance" : both included issuers failed. Here's the post-mortem :

Hash: SHA1

Starting capital : 1100 BTC.

Recovered capital : 453.2730665 BTC MPOE bonds ; 0 BTC NYAN.A ; 0 BTC Patrick Harnett
Received interest : 45 BTC MPOE bonds ; 6.5 BTC NYAN.A ; 15.753 BTC Patrick Harnett

Total recovered : 520.5260665 BTC

Paid :

MPCD.A 7 BTC x 3 = 21 BTC
MPCD.B 12 BTC x 3 = 36 BTC
MPCD.C 32.43671515 BTC BTC x 3 = 97.31014545 BTC

Total paid : 154.31014545 BTC.

Capital to be distributed : 520.5260665 BTC - 154.31014545 BTC = 366.21592105 BTC.

MPCD.C issue closed worthless ;
MPCD.B issue closed worth 16.21592105 BTC ;
MPCD.A issue closed untouched.
Version: GnuPG v1.4.10 (GNU/Linux)


This pretty much settles the score : the only entity in BTC finance is MPOE/MPEx. Everyone else is just pretending.

Category: MPEx
Comments feed : RSS 2.0. Leave your own comment below, or send a trackback.

4 Responses

  1. [...] might remember the shockingly poor results of the MPCD series, and if you know me you knew all along heads are going to roll. [...]

  2. [...] was an attempted CDO offering. It failed, in the sense that of the n participants including MPOE, n-1 participants excluding MPOE went [...]

  3. [...] enough to make the whole thing workable is not in any sense clear - it was not in the case of the MPEx CDOs, for instance (but that was a year and a half ago, and I've made changes to the [...]

  4. [...] skins are in the way of actual realisation. [↩]Call it Gaussian Copula, why not. [↩]See the discussion of CDOs. [↩]"Oh MP, but it didn't end, it was uploaded into a simulation and it still lives there [...]

Add your cents! »
    If this is your first comment, it will wait to be approved. This usually takes a few hours. Subsequent comments are not delayed.