Second Derivative of Earnings Estimates

Models
1

Working on earnings estimates and it improved my trading systems in a great way, without giving away to much stuff like that:

CurQEPS1WkAgo < CurQEPSMean

I came across an analyst who is using the Rate of Change of earnings estimates with a tweak.
He calculates the second derivative of earning estimates, from what I understand:

He found out, that not only accellerating earnings in ROC terms are important but also the acceleration of the accelleration.

Lets say a stock has the following earnings estimates:

1 to 1.15 to 1.30 to 1.45
e.g.
15% to 13% to 11.5 %

This alone would be a warning sign because the accelletation (1 - 1.15 = 15%, 1.15 - 1.30 less then 15%) looses steam. (acceleration of the acceleration goes down).

So this earnings estimates would be (obviously) better:

1
1.15 +15%
1.33 +15%
1.54 +15.7%

So my question is:

How do I calculate earnings estimates on industries (I guess on stocks it would be to volatile but maybe also interesting), lets say in the first step on an absolute basis.
The second step would be to calculate the growth (or decellerating) rate of earnings estimates.

Any Ideas?

Thank you and Best Regards

Andreas

2

Andreas, you can use the six-month smoothed annual growth rate - 6SMAR.
For your purpose there is no need to calculate the actual growth rate.
For a monthly data series the growth is simply (value of current month) / (SMA of preceding 12 month).

6MSAR.png
6MSAR.png1152×592 75.3 KB

3

Aggregate (“100 * (CurQEPSMean - CurQEPS1WkAgo) / Abs (CurQEPS1WkAgo)”, #Industry, #CapAvg, 5, #Winsor)

The parameter revisions are my own personal suggestions considering how many estimates DON’T change from week to week. You might want to use CurQEPS4WkAgo instead of CurQEPS1WkAgo.

((CurQEPSMean - CurQEPS4WkAgo) / Abs (CurQEPS4WkAgo) - (CurQEPS4WkAgo - CurQEPS8WkAgo) / Abs (CurQEPS8WkAgo)) / Abs ((CurQEPS4WkAgo - CurQEPS8WkAgo) / Abs (CurQEPS8WkAgo))

This is the growth of the growth rate (acceleration/deceleration).

4

Thank you, very helpfull!!!