How do you run the advanced statistics?

[flight]

Advanced Statistics
Alpha, Beta, Correlation, R-Squared, Standard Deviation, and Sharpe Ratio calculations for investments, portfolios, asset types, investment types, sectors, or investment goals.

thanks

[Mark]

Hi flight,

You can get these in a "Custom" report when using the Professional or Advisor versions. You will also want to set up your desired benchmark investment. You assign this under "Options / Report Preferences... / Benchmark" from within any report window.

[flight]

And its only available in custom reports?

[Mark]

Yes

[flight]

what custom report should I start with to set this up?

[Mark]

Hi flight,

Create a Custom report, and in the "Custom Settings..." look at the fields under the category of "Statistic Fields". You can add any of these to your report. Many of these also rely on an assigned Benchmark. You must create an investment (like the S&P 500, symbol ^GSPC) and assign it under "Options / Report Preferences... / Benchmark".

[flight]

Hi Mark - If any of the investments appears as "N.A.", will the portfolio totals automatically appear as "N.A." as well?

Does cash need to have any special adjustments since it is always NA?

thanks

[flight]

And related to this question, if a Beta is coming up as "N.A." for a particular investment, what do you recommend doing to fix?

Let's take a very liquid example like Amazon (AMZN). In the custom report, I have the beta field included and ran a sample report 1/1/16-12/31/16.

I verified in the prices that I have all the daily prices for AMZN in 2016.

I also verified I have the Benchmark S&P 500 selected in the Benchmark Dialog per the help guide.

Does the Comparison Period in the preferences field have to the match the Custom Report? I selected 5 years for the comparison period in the preference.

thanks

[Mark]

Hi flight,

For Beta calculations, the calculation is done back from the ending report date as far back as you set the comparison period. The report's starting date does not matter for this field. Notice the field is labelled "Beta (ending)" and not "Beta (between)".

If you are getting an N/A then you don't have prices every month back from the end date for the whole comparison period. You need pricing in both the investment and the benchmark over the whole comparison period. If your report ends on 12/31/16 and your comparison period is 5 years, you need end of month prices for every month back 5 years.

You can control how flexible Fund Manager will be in interpolating missing prices. By default it will span 4 days to cover long weekends, but if you have a longer period than this without pricing, you'll get an N/A. Use "Options / Yield Interpolation Range..." to adjust this flexibility. If you un-check the "Enforce Maximum..." option and check the "Allow Interpolation Prior to Earliest Data" option Fund Manager will always interpolate, and you'll get beta values. This isn't a good mode to work in though. If you happen to be missing a bunch of pricing data you wouldn't know about it, and Fund Manager would just go ahead and perform the calculation on interpolated data. In that state you can at least confirm that the issue is missing price data. You can either manually look at your recorded pricing for missing prices, or you can turn on the statistics logging, and that will help you see where Fund Manager stopped the calculation due to a missing price. See the Beta Online Help topic for help turning on logging.

The beta for a portfolio, asset type, goal, sector, or investment type is a weighted average of the beta's from the investments in that group. The weighting is based on the ending value. All investments with a non-zero ending value in that group must have a valid beta (sufficient pricing data) for the group beta to be reported. If your group object includes cash, then you would need to record pricing data for cash. You can just record a price of 1 for cash, and this will enable it to calculate just like any other investment.

[flight]

Thanks

1) How do I get the price of cash updated over long time periods?

2) Is the "Normalized Standard Deviation of Price (between)" a real standard deviation/measure of volatility? I noticed in the help field it only takes the month end prices into consideration. Is that accurate? I bring it up because it is show very low Std Dev for certain stocks but when I compare to other statistics services (http://www.investspy.com/calculator, among others), the standard deviation is much higher?

[Mark]

Hi flight,

1) Manually update them, or you could create a CSV file in Excel, and import. For example:

12/31/2016,1
11/30/2016,1
10/31/2016,1
...

and use "File / Import / Prices / Generic...". Specify the symbol of cash in that dialog, and the format string for this example would be: MM/DD/YY,NAV

2) This field takes all recorded prices into consideration, not just the end of month. See the Custom Report Documentation. It is the standard deviation of all these prices between the start/end report date, divided by the average share price between these same dates.

[flight]

Is "Normalized Standard Deviation of Price (between)" the same as the commonly defined standard deviation for a stock?

The most commonly used is expressed as a percentage, which the one in FM is too. However, the one's I see in FM is too low.

For example, the standard deviation for AER shows 9.9% but I looked it up in FactSet as 38.2%. It was quite volatile in 2016. 52 week los is $24.61 and 52 week high is $45.53.

I can't see how this would be a 10% standard deviation.

Unless this metric is a total different metric.

[Mark]

Hi flight,

I'm not sure how the other metric you're looking at is calculated, I can only comment on how Fund Manager does the calculations. This field is the standard deviation of all the recorded prices within the time period, and then divided by the average price. It is normalized, so it makes comparing this figure between investments meaningful. If it wasn't normalized, investments with larger share price would have a larger standard deviation than investments with a smaller share price that had the same volatility.

[flight]

I am not sure about the nuances of these statistics....

I do know the commonly used metric is expressed as a % and viewed as annualized measure. I am not sure if FM's is annualized - I underlined something below from Morningstar which may/may not be relevant.

In my example of AER, if I take the square root of the 9.9%, it is 31%, which is much closer to the volatility/standard deviation I was expecting to see.

here is how Morningstar defines...

Morningstar calculates historical statistical volatility assuming a log normal return distribution and continuous compounding. Historical statistical volatility is calculated as follows: Daily returns are calculated as return = natural log of (p2/p1). The standard deviation of this return is then calculated for the last 21, 42, 63, 126, 252, 504, and 756 days, corresponding to an average trading month, two months, three months, six months, one year, two years, and three years. The standard deviation is then annualized by multiplying by the square root of (252/number of trading days). The historical statistical volatility values are calculated every evening, and are not updated intra-day.

[Mark]

Hi flight,

The "Normalized standard deviation of price" fields are not annualized. There are standard deviation of price fields, which are also not annualized. There are standard deviation of return fields, and that comes in both a non-annualized and annualized version. Unless it says "annualized" in the name, it is not annualized.