THIS POST IS UPDATED PERIODICALLY WITH MORE DATA AND ANALYSIS!
Morning (01/07/2015),
I've put together some statistics regarding well known short positions across multiple sectors in UK markets. This data can be used to help determine the levels at which investors could consider going long on other stocks in similar positions.
Known Issues:
1. Historic market capitalisation levels are calculated using the latest shares in issue figures (clearly this number can change and for companies under short pressure I would argue that this is more likely to change than for companies that are not in the same position).
2. I have excluded well known shorted companies that have had the positions taken out as hedges against other positions - the positions I have chosen are aggressive short positions.
3. Data errors can exist and although I have tried to avoid them as much as possible people should be aware of them.
4. Some of the shorts listed below are still active, which means that if another down leg takes place (past the share price low in the below data) then this data will all change.
5. There really need to be hundreds of data points to give a really solid study - so view this as only rough.
| Company | Market Cap Peak/£bn | Share Price High | Share Price Low | Percentage change/% | Mean Fall/% | Mean Fall Companies under £1bn MCap Peak/% | Mean Fall Companies Over £1bn MCap Peak/% | |
| Quindell |
3.036852
|
682.50
|
24.10
|
-96.47
|
-68.3010826720286
|
-74.5074087409454
|
-65.1979196375702
|
|
| Tungsten Corp |
0.513867475
|
409.75
|
52.488
|
-87.1902379499695
|
||||
| Tullow Oil |
14.6821707
|
1611
|
278.10
|
-82.7374301675978
|
||||
| Afren |
1.89171248
|
170.80
|
1.28
|
-99.2505854800937
|
||||
| BooHoo |
0.626144975
|
55.75
|
21
|
-62.3318385650224
|
||||
| Plus500 |
0.8972909
|
781
|
198
|
-74.6478873239437
|
||||
| Sainsbury |
8.2264596
|
428
|
221.10
|
-48.3411214953271
|
||||
| WM Morrison | 8.056.095 |
345
|
150.6
|
-56.3478260869565
|
||||
| Carillion |
1.7397029
|
404.3
|
294.025
|
-27.2755379668563
|
||||
| Ashmore Group |
3.04106961
|
429.9
|
249
|
-42.079553384508
|
||||
| Nanoco |
0.471046245
|
199.275
|
83.155
|
-58.271233220424
|
||||
| Lancashire Holdings |
1.8493926
|
933.00
|
506
|
-45.7663451232583
|
||||
| Ocado Group |
3.66998335
|
623.5
|
216.8
|
-65.2285485164394
|
||||
| Monitise |
1.7903459
|
82.75
|
9.53
|
-88.4833836858006
|
||||
| Blinkx |
0.9445401
|
234.75
|
23.25
|
-90.0958466453674
|
I intend to update this post with graphs confirming or denying any correlations between the peak market capitalisation level and the mean fall of stocks that are being short sold.
Analysis:
1. The mean percentage fall suggests that for companies that are in a bear environment, long positions should be avoided until the company has fallen by at least 50% if you are a buy and hold investor prepared to top up on the way down.
2. Be aware that as you reach the point when you should consider going long, average daily volumes are likely to increase in conjunction with intra-daily price volatility, as short positions exit and long or volatility traders move in and out on swings.
3. Although the data above only gives one example of this, it is generally notable that companies with lower floated share prices fall less under short pressure (I image that this is to do with the implied extra volatility per every 1p change that exists - each penny change in the share price carries more significance for company's value).
4. Statistical Analysis:
- Before you read this, it should be noted that I am not a statistician or mathematician.
- If we have a look at correlations in the data, the Pearson Correlation Coefficients for the data sets (Peak Market Cap vs Percentage Fall) look like this:
Companies Valued Over 1Bn: R= -0.1506
Companies Valued Under 1Bn: R= -0.4886
All Companies: R= -0.0435
4. Statistical Analysis:
- Before you read this, it should be noted that I am not a statistician or mathematician.
- If we have a look at correlations in the data, the Pearson Correlation Coefficients for the data sets (Peak Market Cap vs Percentage Fall) look like this:
Companies Valued Over 1Bn: R= -0.1506
Companies Valued Under 1Bn: R= -0.4886
All Companies: R= -0.0435
- This works between -1 (negative correlation) and +1 (positive correlation). The closer the number is to 0, the weaker the relationship.
- I don't feel that I can safely comment on this data fro reasons I will place in my evaluation, but the data is there for people who find it useful.
- The mean data for share price falls in the table above does suggest however that larger companies (peaking over £1bn in valuation) do fall less than smaller ones (under the peak £1bn market cap).
Evaluation:
This study does sadly have some rather large holes in it:
1. The data sample is very small - 15 companies out of the 1231 companies listed on the LSE (excluding Venture Capital Trusts and Investment Trusts) is hardly a fair study.
- This being said manually sourcing data is hard and time consuming and added to this in the grand scheme of things, there aren't a vast number of short positions over the 2.5% threshold I used (many short positions are merely intended to hedge long positions and can be quite small). Therefore, you could argue that this 2.5% threshold helps to limit the sample size issues.
- It is also worth noting that newer companies are unlikely to quickly build up shorts of over 2.5%, which realistically brings this total sample size down again.
2. There are so many possible variables that I would question if knowing R values is really that useful.
- Also, there are twice as many larger companies in this data set than smaller companies, which again brings into question knowing about the R values and the mean values
Conclusion:
Regardless of the issues with this little study, I think that I personally have learned to avoid trading companies with shorts over 2.5% on the long side until a fall from the peak market cap has been achieved of 50% (across all companies). For smaller companies it does seem that looking for around a 75% fall is sensible, but on balance I would rather trust the larger data set for companies valued at over £1bn.
Enjoy,
The Masked Stock Trader
