Today, we’ll see if we can mix up some UK, European and American securities with the intention of reducing short term risk. We’ll find out what will happen if we add a few percent of derivatives to the FTSE 100, Euro Stoxx 50 and the DAX.
This will be very exciting as I have no idea what to expect. I started writing the article before collecting any of the data.
My gut feeling is that the results will not be very good. It doesn’t make any sense. However, why not give it a shot and find out? Let the analysis begin.
We all know that any form of hedging will likely reduce the volatility of the P&L. The downside is that it shrinks both profits and losses as we are reducing the directional exposure to the index of choice by using an instrument with a negative correlation.
There are many ways to hedge a position with options, futures or even CFDs. This study will use volatility and inverse products as they are easily accessible. Here are the participants: FTSE 100, DAX, STOXX50, S&P 500, VXX, UVXY, SVXY, SPXU and SQQQ.
I used the actual indices, however, an index tracker ETF of your choice will do just fine. The S&P 500 is included because the volatility products (VXX, UVXY, SVXY) and SPXU are based on its performance.
We’ll look at three periods: Oct – Dec 2019 will be the reference period followed by the covid crash Feb – Apr 2020 and then Jul – Oct 2020.
The selection is somewhat arbitrary, however, the notion is that 2019 was a bull market. Then we have a crash, however, the chosen period ends before the complete recovery of the S&P 500 to leave some uncertainty on the table. Finally, we’ll move on to the most recent post-crash period – Jul – Oct 2020.
The performance will be measured by weekly mean return and standard deviation.
You are probably familiar with the indices so I’ll explain the purpose of the remaining instruments. VXX is an exchange traded note (ETN) which tracks the S&P 500 Short-Term Futures Total Return Index (SPVXSTR). UVXY is a 1.5x leveraged product which tracks the same ticker as VXX. Note that neither of the two track the VIX index. SVXY is a -0.5x inverse ETF which is short front and back month VIX futures and included for reference purposes.
SPXU and SQQQ are -3x inverse funds tracking the S&P 500 and the Nasdaq 100 respectively. I added SQQQ to the list as it is very liquid and the Nasdaq 100 is highly correlated to the S&P 500.
What happened during the bull market
I expect that this strategy lost money. First let’s start with the S&P 500 hedged with VXX and SPXU. On the right hand side of the graph you see 100% S&P 500. Then each mark represents the removal of 1% S&P 500 and the addition of 1% VXX or SPXU respectively. The minimum amount of the index is 80% in the final step. Returns are on the vertical axis and the standard deviation on the horizontal.
It’s not as bad as I thought and it seems that we need less VXX to achieve the same as using SPXU. 2% exposure to VXX reduced the risk by 11.7% but it also chipped away the same amount of return. Therefore, a small exposure had a very symmetrical effect.
UK & Europe
I decided to use volatility as it expands dramatically during a crash so it should protect the positions better than an inverse fund. First, let’s have a look at the correlation matrix of the products.
Unfortunately the correlations were weak and it seems that the DAX was living in it’s own world. Nevertheless, I’ll give it a try and see what kind of result we’ll get.
The mean weekly return was 0.45% and you can see that each 1% of volatility reduced it by a minimum of 10%. Here are the results:
Conversely, the risk reduction was negligible. It seems that VXX works a bit better, therefore, I’ll remove UVXY from the test.
Euro Stoxx 50 and the DAX
The results are similar on the continent, minor risk reduction at a significant cost. The mean weekly return of Euro Stoxx 50 and the DAX were 0.70% and 0.66% respectively.
The next section will cover the crash and I hope that it will be different.
The correlations changed a lot!
I’m relieved that this article is not completely pointless. The FTSE 100, Euro Stoxx 50 and the DAX all became highly correlated. Most notably the correlation between the DAX and both the FTSE and Euro Stoxx went all the way up from negative to almost 1!
It is also interesting that the three indices now have a negative correlation with the S&P 500. I didn’t expect to see this.
The important message is that correlations change, especially during a crash. Sometimes securities can become highly correlated on the way down, thus we shouldn’t take negative correlations, such as stocks and bonds, for granted. This explains situations when stocks, bonds and gold all plummet simultaneously.
The weekly return of the S&P 500 is at the bottom of the graph as it was negative (-0.61%).
As expected exposure to VXX reduced the losses significantly, 2% reduced them by 36%, however, the standard deviation only lost 5.9%. 5% exposure to VXX rendered the losses negligible (-0.08%) while 6% turned the P&L ever so slightly positive (0.03%).
UK & Europe
The FTSE 100 fared worse than the S&P 500 with a mean weekly return of -1.44%.
Despite the positive correlation with VXX a 2% exposure reduced the losses by 19% while 5% achieved 39%. Nevertheless, the result is not as great as what we got when using the S&P 500.
The situation in Europe is not much different. Euro Stoxx 50 returned a weekly average of -1.53% while the DAX was down -1.26%. In both cases a 2% exposure to VXX reduced the losses by 15% and 17.5% respectively without any major risk reduction.
Overall, VXX protected all of the positions successfully during the crash.
The post-crash period
The correlations are a bit different yet again. There is nothing significant other than the low positive correlation of the UK & European indices with the S&P 500 and the diminishing correlation between the DAX and the FTSE 100.
The S&P 500 was the best performer with a mean weekly return of 0.53%, followed by a flat DAX which delivered 0.08%. Euro Stoxx 500 was the worst performer (-0.23%) followed by the FTSE 100’s -0.17% return. Here are the results:
Adding VXX to the indices in this period failed to deliver any significant risk reduction while causing additional losses. Note that the effect of VXX on the S&P 500 is similar to the results from the bull market period – a 2% exposure reduced the upside by 9.4%.
A 50:50 split between the S&P 500 and the FTSE 100 would have outperformed by achieving 0.18% mean weekly return. Here is the graph:
Note on currency risk
This article did not address the currency risk of the instruments used. The indices and the ETFs/ETNs used are denominated in GBP, USD and EUR. Inevitably the exchange rate will change which affects the P&L depending on individual circumstances.
There are ETFs/ETNs which match the currency of the respective index, however, the risk still exists as a lot of them are likely to be currency unhedged.
The article turned out not to be a complete waste of time. Firstly, we found out that correlations can change a lot relatively quickly. Furthermore, correlations between indices can turn negative during a worldwide market crash.
Secondly, adding VXX reduced the mean weekly losses of the FTSE 100, Euro Stoxx 50 and the DAX during the covid crash. Nevertheless, it did not reduce the risk in any significant way.
Thirdly, we found out that some exposure of an FTSE 100 position to the S&P 500 can be beneficial as the latter suffered less during the crash and recovered quicker.
I expected that a continuous use of VXX is likely to result in consistent losses. It lived up to its name – ‘short-term futures’.
The main problem with the use of VXX is choosing the right moment to add it to the portfolio. I believe that this will be extremely difficult as volatility expands very quickly. Therefore, such a trade will require impeccable timing.