Enron Mail

From:clayton.vernon@enron.com
To:george.posey@enron.com
Subject:projected fund giving
Cc:vince.kaminski@enron.com
Bcc:vince.kaminski@enron.com
Date:Fri, 21 Jan 2000 02:23:00 -0800 (PST)

George-

Here is one statistical method (linear regression), applied to Hosanna's fund
giving. The method is deterministic, where contributions each Sunday are
attempted to be explained by significant events on the church's and members'
calendars, and the method shoudl be considered in contrast to an anecdotal
one, where each Sunday's giving might be estimated to be the average giving
of the same Sunday (e.g., each 2nd Sunday in January) in the past.

The following effects are estimated independently by the computer (they add
and subtract with each other accordingly; some Sundays can have more than one
effect "in play"):

1) most positive for giving: an additional +$8,802 for special pastoral
appeals (3 of them in the past 3 years)

2) positive for giving: an additional +$2,551 Easter and nearest- Christmas
"Sunday," an additional +$59 Thanksgiving Sunday, and a tiny additional +$1
each week pure time trend.

3) negative for giving: -$304 for Sundays of Monday Federal Holiday weekends,
and -$9 during the Summer (June 1-Aug 31)

4) "wave": there is an $80 per week "wave"- giving $80 higher one week, $80
lower the next, etc ...

The computer can explain roughly 60% of the absolute variation in giving the
past 3 years using this approach. Keep in mind, much of the information,
statistically, is captured in the pastoral appeals, Easter and Christmas
Sundays, and Holiday weekends. The rest of the variation cannot be explained
by this simple model (i.e., it looks like "noise").

Here are the predictions for fund giving for the upcoming year. You can
cut-and-paste these numbers to any file or spreadsheet you like. They may
seem "boring" to you; please remember they are the deterministic components
of a process that involves a lot of volatility, and your church's actual
experience will predictably be volatile. These numbers are as-if the "means"
of experiences dependent on the "standard deviations."

01/02/00 4611
01/09/00 4699
01/16/00 4394
01/23/00 4705
01/30/00 4701
02/06/00 4702
02/13/00 4703
02/20/00 4400
02/27/00 4710
03/05/00 4706
03/12/00 4708
03/19/00 4709
03/26/00 4710
04/02/00 4711
04/09/00 4712
04/16/00 4713
04/23/00 7265
04/30/00 4672
05/07/00 4717
05/14/00 4718
05/21/00 4719
05/28/00 4416
06/04/00 4717
06/11/00 4713
06/18/00 4714
06/25/00 4715
07/02/00 4716
07/09/00 4718
07/16/00 4719
07/23/00 4720
07/30/00 4721
08/06/00 4722
08/13/00 4723
08/20/00 4724
08/27/00 4725
09/03/00 4432
09/10/00 4742
09/17/00 4738
09/24/00 4739
10/01/00 4740
10/08/00 4742
10/15/00 4743
10/22/00 4744
10/29/00 4745
11/05/00 4746
11/12/00 4747
11/19/00 4748
11/26/00 4808
12/03/00 4750
12/10/00 4752
12/17/00 4753
12/24/00 7305
12/31/00 4711

Best wishes,

Clayton Vernon
Manager, Research