Making money on LIBOR fallback (5)
In the previous episode of my "making money on LIBOR fallback" (listed at the bottom of the post), I looked at a specific position that I proposed just after the 26 November 2018 announcement of the ISDA consultation results. The position is still in-the-money, but with very little changes over the last month.
So for this episode, I will simply present graphs related to similar positions that could have been taken some month ago. All of them are making money, but the money making machine has slowed down in the last month. The next opportunity to make money will probably be when ISDA Benchmark Committee decides on the exact details of the spread computation methodology (look-back period, mean/median, transition period).
I looked at numbers for four indices: USD-LIBOR-3M, USD-LIBOR-6M, GBP-LIBOR-3M and GBP-LIBOR-6M. For each index, I first computed the running mean/median for two possible look-back periods (10 and 5-year period) of the spread between the (forward-looking) LIBOR and the (backward-looking) overnight compounded setting in arrears. Due to the "in-arrears" feature, there is a 3/6 months delay between the publication of the LIBOR and the knowledge of the spread. This is why in the graphs, not all time series go to today.
Then I have compared those running mean/median with the market spread for basis swap with maturity 30 years. For USD, the spreads have to be taken with a pinch of salt as the spread is to EFFR (the only liquid ON index for derivatives today) and not SOFR, the Fed Fund swaps are arithmetic averages and the 3/6 month spreads have a composition feature. The goal is not to have the numbers to the last tenth of a basis point but to see the trend of the forward-looking market with respect to the backward-looking historical data.
The volatility of the basis spread has also decreased. For example in USD, the daily volatility of the market spread LIBOR-3M/EFFR was around 0.52 bps before the announcement and has been 0.30 bps after; if you take only the data since 1 March, the volatility has been 0.19 bps. In GBP, for the market spread LIBOR-3M/SONIA, the volatility was 0.35 bps before the announcement and has been 0.26 bps after; if you take only the data since 1 March, the volatility has been 0.24 bps.
Edit 18 March: Due to a wild copy and paste, the first version had completely incorrect standard deviation figures.
Previous episodes:
So for this episode, I will simply present graphs related to similar positions that could have been taken some month ago. All of them are making money, but the money making machine has slowed down in the last month. The next opportunity to make money will probably be when ISDA Benchmark Committee decides on the exact details of the spread computation methodology (look-back period, mean/median, transition period).
I looked at numbers for four indices: USD-LIBOR-3M, USD-LIBOR-6M, GBP-LIBOR-3M and GBP-LIBOR-6M. For each index, I first computed the running mean/median for two possible look-back periods (10 and 5-year period) of the spread between the (forward-looking) LIBOR and the (backward-looking) overnight compounded setting in arrears. Due to the "in-arrears" feature, there is a 3/6 months delay between the publication of the LIBOR and the knowledge of the spread. This is why in the graphs, not all time series go to today.
Then I have compared those running mean/median with the market spread for basis swap with maturity 30 years. For USD, the spreads have to be taken with a pinch of salt as the spread is to EFFR (the only liquid ON index for derivatives today) and not SOFR, the Fed Fund swaps are arithmetic averages and the 3/6 month spreads have a composition feature. The goal is not to have the numbers to the last tenth of a basis point but to see the trend of the forward-looking market with respect to the backward-looking historical data.
The volatility of the basis spread has also decreased. For example in USD, the daily volatility of the market spread LIBOR-3M/EFFR was around 0.52 bps before the announcement and has been 0.30 bps after; if you take only the data since 1 March, the volatility has been 0.19 bps. In GBP, for the market spread LIBOR-3M/SONIA, the volatility was 0.35 bps before the announcement and has been 0.26 bps after; if you take only the data since 1 March, the volatility has been 0.24 bps.
Edit 18 March: Due to a wild copy and paste, the first version had completely incorrect standard deviation figures.
USD-LIBOR-3M
Figure 1: USD-LIBOR-3M, compounded in-arrears, their spread and the running mean/median.
Figure 2: Market data for 30Y basis swaps USD-LIBOR-3M v EFFR and comparison with the running mean/median historical spread.
USD-LIBOR-6M
Figure 3: USD-LIBOR-6M, EFFR compounded in-arrears, their spread and the running mean/median.
Figure 4: Market data for 30Y basis swaps USD-LIBOR-6M v EFFR and comparison with the running mean/median historical spread.
GBP-LIBOR-3M
Figure 5: GBP-LIBOR-3M, SONIA compounded in-arrears, their spread and the running mean/median.
Figure 6: Market data for 30Y basis swaps GBP-LIBOR-3M v SONIA and comparison with the running mean/median historical spread.
GBP-LIBOR-6M
Figure 7: GBP-LIBOR-6M, SONIA compounded in-arrears, their spread and the running mean/median.
Figure 8: Market data for 30Y basis swaps GBP-LIBOR-6M v SONIA and comparison with the running mean/median historical spread.
Previous episodes:
Comments
Post a Comment