macro paper trading

Weekly PNL:

Long EDU1-EDU2: +750 USD

Long LU7: -240 GBP

Long IKM7 vs RXM7: -1070 EUR

Short EURJPY (expired minis): +128 USD

IZ7-IZ8-IZ9 fly: -375 EUR

Long IRU8: +4230 AUD

Long SIM7: -372 600 RUB. The idea was for RUB to catch up with weaker oil, but this got crucified post-Fed. RUB unlike ZAR or TRY hasn’t previously usually responded to Fed policy, but nonetheless it was stupid to expose this to the Fed meeting. Holding this long is really expensive. June contract is 58.44, September one 60.03, that’s 11% annual rate.

Steeper CAD/flatter AUD: -1350 CAD, +2588 AUD

Total in USD terms: flat over the week, -0.37% from topic beginning.
 
I am thinking what else there is to be done in eur curve apart from the ER Z7-Z8-Z9 fly. The rise in 2y rates and the 2s5 steepness without 5s10 budging higher for some reason hurts my eyes on the chart. 2s5s10 is accordingly higher, but 5s10 is fairly high in EUR versus other currencies, so being exposed to a fall in 5s10 in the 2s5s10 doesn’t seem smart. 12 bps steepening in 2s5 given a 7 bps rise in 2y seems fair, so 2s5 would probably be as directional as 5y outright. The problem with 5y outright is that its yield isn’t that good versus other currencies, so it probably makes sense to extend it to being long the 10y sector, which is so steep that you get better returns in EUR than in USD and as much as in AUD.

uc

uc

uc

uc


uc


I don’t have 3m or 6m tenors for currencies other than USD, EUR, GBP, AUD and CAD, so have to use 2y rates as the funding rates for a broader study. EUR isn’t as good as HUF, SEK or NZD, but still good relative to USD, GBP and especially JPY. Too bad Sweden has no exchange traded futures.

On portfolio level there are plenty of fixed income bullish positions (long IRU8, LU7, to certain extent IKM7 versus RXM7 and ER Z7-Z8-Z9), so a bullish position in EUR rates would better be hedged with something against a global yields rise. It’s relatively cheap to bet on 10y GBP rates rising, I don’t any strong views on them going higher or lower, and the differential between 10y EUR and GBP rates recently plunged. GBP does have some idiosyncratic risks that wouldn’t be passed to EUR, so could consider blending it with US 10y short.

Implementation of this seems more difficult. 10y’s rolling yields are around 175 bps in EUR, 155 bps in USD, 125 bps in GBP in swaps, but if I look at bonds I see roughly 180 bps in EUR (german bonds), 230 bps in USD, 150 bps in GBP. Shorting US treasuries doesn’t seem as interesting with treasury futures. This seems due to repo rates being much lower than LIBOR in USD without US treasuries trading richer to swaps. The 10y differential has been roughly tracked by the blue STIR contracts, so could use this for USD as it is roll-down positive – 7 bps per quarter for EUR and 5 bps for USD.

uc


uc
 
sold 7 IRU8 @ 97.82 (bought this portion at 97.73). Still have 15 IRU8 that were bot @ 97.87. Looking to sell some more at 98. Canadian banker acceptance futures have dropped too much recently despite new worries in the oil market, so looking to buy some BAU8 too later.
 
On 21/02/2017 I thought that Australian 2s10 got too steep relative to Canadian curve at +5.5, 89.5 vs 84. This peaked on 24/02/2017 at +11.5, 87 bps in AUD and 76 bps in CAD. It seems the cause of peak was on the Australian 10y side cus it only moved 6 bps while GBP, USD, CAD moved 10+ bps. The box is now back lower at 2bps, 90 bps in AUD vs 88 bps in CAD.
uc



A more parsimonious trade would be going long Australian 10y and shorting Canadian 10y without complicating life with 4 legs.
More...

So far it’s been the case, but why? The 10y differential also moved around 2 bps lower from 111 to 109 bps. It also peaked on 24/02/2017 at 116 bps. Using decomposition I described in post #23, we have:

0 from policy rate changes, -5 from CAD front contract rising more than the AUD contract above the policy rate, +4 from CAD 2y rising more above the front contract than AUD 2y rate over its front contract, -2 from 10y differential.

IR7-IR3-BA7-BA3 box would have performed a lot stronger as it moved from 10 bps to 0 without peaking on 24/02/2017. The cause is on the BA side – I am saying in the above post that BAU8 dropped too much and makes it a buy, afterall. This is a lil bit inconsistent – if BA7-BA3 were a proxy for CDSW2s10, then why would you bet on steeper CDSW2s10 and lower BA7 at the same time (no view on BA3)? I am implicitly saying that CDSW2s5s10 should go lower. This is probably reasonable given how much it has climbed up over the year. If I plot 2s5s10 for 2014-2017, then the fly plunged along with oil but recovered by more than the-most-abused-ever-overlay-chart-with-different-axes suggests. Likewise, if the fly should be a function of current short-term rates, it’s gone too much up recently as the front contract stayed flat. The fly move up does, however, make sense in light of higher BA1-BA3.
uc

uc


uc

uc



Finally, has the use of XTH7, XTM7, IRZ7, BAZ7, CNM7 correlated with what it was designed to track? Visual total PNL over the box level says more yes than no. I had 15 contracts with 25 cad/aud bps value, so 1 bps move in the fly should have translated into 375 CAD/AUD pnl. In theory by 24/02/2017 the 6 bps move against me should have produced 2250 loss. In reality it was -3000, and swollen to -4250 by 28/02/2017 – this was due the Aussie 10y bond futures temporarily cheapening – its yield stayed the same at 280 bps despite the 10y AUD swap rate falling 9 bps from 21/02/2017 to 28/02/2017. Total PNL should have been around 1000, but in reality around 300 so far now. Underperforming theory by 700 is easier on paper than irl.

uc
 
bot 1 IKM7 @ 128.13 (increasing size from 1 to 2 contracts in this)
bot 1 RXM7 @ 159.55 (need 2x long for Germany-UK, but also need 1 short for Italy-Germany)
sld 2 G M7 @ 125.97
 
Looks like I got stuck with the rouble position. It’s taken around 0.6% of portfolio, and valuation got even more attractive, but it’s expensive to short. Verbal intervention from ministry of finance saying RUB is 10-12% overvalued and that they may purchase $1b usd more isn’t helping. The analysts cited in the media say March-end is tax period, so rouble could get even more expensive.

Re-ran my risk file. 1% daily var with existing positions over the past year has been 0.65%. Annualized vol has been 4.8%, so 4.8%/16*2.33 = 0.7% var. Substantially undershooting the risk parameters, so need to add more positions or increase the existing ones.

The first pie chart is position volatilities on a standalone basis without portfolio effect. Portfolio volatility is 60% lower than the sum of individual position volatilities. The second pie chart is marginal volatilities for positions. It seems I can add to ER fly without rising portfolio volatility by much. Ruble short is a quite significant risk in the portfolio.


uc

uc
 
You must log in or register to reply here.
Back
Top