Wednesday, July 13, 2016

Metric - Risk



When I wrote about risk before, I was thinking about how to build and diversify my portfolio. However, risk is difficult to quantify. Luckily, there are free tools to help you in your analysis of risks. One such free tool is the Risk Grade (TM) tool on the Nasdaq website.




The tool is very easy to use. Just enter a list of tickers and it will print out your risk grades of the companies. The calculation is based on all the components of market risk such as equity, interest rate, currency, and commodity risk. The tool also compares your companies with the S&P and NDX indices.




Basically, a risk grade below 100 is considered low risk. 100-700 is average. Above 700 is pretty risky.




The website also provides a very good tutorial about the relationships of risk.




Check it out when you have some time. It is always good to learn something new!

Please note that while the tool is free, there is a link to the Risk Metric group - which I am sure will charge you for additional services. Also, I couldn’t get the RiskChart tool working.
Let’s see how the Risk Grades work with real world examples - I compared the Risk Threshold with the Gains/Losses from my taxable portfolio.

Ticker
Risk Grade
Portfolio Gain
ABT
134
9.59%
ADM
149
24.53%
AFL
99
30.11%
BBL
268
-28.51%
CAT
146
4.71%
CMI
147
18.96%
DE
132
1.22%
DEO
98
9.91%
DOV
141
11.60%
EMR
123
5.47%
FAST
127
13.86%
GILD
149
-1.53%
GIS
89
50.33%
HCP
157
0.97%
IVZ
176
-23.04%
JNJ
79
30.59%
KMB
88
30.80%
KO
75
15.29%
MCD
90
38.78%
MMM
93
21.92%
NOV
219
-17.32%
PG
81
9.08%
PM
84
32.41%
SOUHY
N/A
-10.24%
T
73
31.74%
TROW
117
-1.64%
UL
104
23.17%
UNP
135
-2.32%
UTX
104
11.78%
VZ
85
27.55%
WFC
119
1.00%
WMT
115
5.47%
XOM
121
10.98%
YUM
167
17.71%
It is very interesting that my biggest losses come from my riskiest companies.

I also scatter plotted the data (Gain vs Risk Grade) to visually see the impact of the risk.


I am constantly evolving my criteria for stocks selection. Since this tool is free and provides a simple indicator for risk,  I will include the Risk Grade Metric to my criteria selection. The threshold is 150 - based on the scatter plot above. Updated to use new D4s Risk Grade Metric. A higher risk grade is okay if the other criteria (low payout ratios, high dividends, growing earnings/revenues etc) are good. If I can figure out how to extract the data to google sheets, I will update my criteria table tool.


Do you use other tools to quantify risk?

D4s

Disclosure: Long on all companies above

7 comments:

  1. Wow, thanks for posting this. That is quite a correlation in risk versus return, so much so that I wonder how this tool calculates risk? It's almost uncanny how your highest returns are with the lowest risk stocks.

    I put in the stocks I have purchased thus far and I see pretty much the same pattern.

    I wonder how much recent performance is factored into this tool? Are the stocks that are the worse for the worst because they were the riskiest or are they the riskiest BECAUSE they performed the worst over the last couple of years?

    ReplyDelete
    Replies
    1. Papa Div, the risk chart version of the should show the changes of the risk grade over time. Unfortunately, I couldn't get it to work. I need to play with the webpage a bit more.
      I think you are right. I suspect that the risk grade is increasing for my bad performers due to the recent drops in their performance. My goal is just to keep it in perspective. Obviously, 200+ is troublesome for when I initiate a position.
      But I am glad you enjoyed the article! Thanks for your comments!
      D4s

      Delete
    2. Yes, I just tried it and could not get the Risk Chart tool to work either, which is probably very useful for reviewing your current portfolio. I'll check back with it and see if it starts working in a week or two as it would be very useful in deciding how good of a metric it is.

      Delete
    3. Papa Div, I figured out how they calculate the Risk Grade - at least their Version 1 method which was implemented in the late 1990s. I think Nasdaq implemented V2 (2006). Anyway, they include a decay factor of 151 trading days for their calculation and each day has a gain/loss over the previous day. I managed to implement the V1 version in google sheets. After back testing V1 a bit, the absolute number is probably meaningless without tying it with the industry or an index. I will read up on the V2 text and try to figure out the equations. Read more here - https://en.wikipedia.org/wiki/RiskMetrics & https://www.msci.com/documents/10199/e79f2f96-a1bd-4009-a3f6-b558f9751a37

      Delete
  2. This level of mathematics is beyond me, but doesn't the fact that they also exponentially weight the recent days also mean a stock's RiskGrade will reflect its recent performance very strongly? They provide this example:

    " For example, on July 16, 1998, Cisco’s RiskGrade fell to a minimum of 139, then jumped to a maximum of 367 on October 15, 1998. In three months, in the midst of the Russian Crisis of August 21, 1998, Cisco’s RiskGrade jumped by almost a factor of three."

    That seems bizarre.

    ReplyDelete
    Replies
    1. When I use risk grade v1, my BBL investment was around 120 two years ago. Now it's in the 200s. So, it is tied to performance. However, if you compare with the s&p index, then BBL actually looks quite risky in 2012-2014 range even 120 may suggest it's less risky. When you compare with an index, it tells you if is x times more risky or not.
      I will try to release the tool when I can.
      D4s

      Delete
    2. As you said then, the absolute number is meaningless but if you compare with an index you can get a good idea. I will be very interested to see your tool!

      Delete