
              ECON: ECONOMETRIC FORECASTING MODELS
                                
                         DC Econometrics

                         Copyright 1989


 Table of Contents

     Software License                                     
     Warranty                                             
1.   Introduction                                         
2.   Getting Started                                      
2.1  Required Hardware                                    
2.2  Making a Backup                                      
2.3  Running the Program                                  
3.   Main Menu                                            
3.1  Input Monthly Data                                   
3.2  Econometric Forecasts                                
3.3  Play "WHAT IF" Games                                 
3.4  Edit Files                                           
3.5  Historical Forecasts                                 
3.6  Adjust File to New Baseline                          
3.7  Exit the Program                                     
3.8  Help                                                 
3.9  Printer Options                                      
3.10  Print Registration Form                             
4.   Using the Forecasts                                  
5.   The Econometric Models                               
5.1  Checking the forecasts                               
5.2  Details on the Regressions                           
6.   Getting Help                                         
6.1  Error messages
7.   Recommended Reading                                  
8.   Placing an Order                                        


                         Software License

Read this user agreement before using the software.  By using the
software, you  agree to  be bound  by the  following terms of the
license and warranty.

The econ  software recorded on disk is copyrighted software of DC
Econometrics, and  all rights  are  reserved.    DC  Econometrics
authorizes you  to make  archival copies  of the software for the
purpose of backing-up our software and protecting your investment
from loss.  You may give away copies of this shareware program to
others and  you may  make it publicly available on bulletin board
systems.   You may not distribute copies of the  output forecasts
of the program without written permission from DC Econometrics.

You agree  that the liability of DC Econometrics, its affiliates,
agents, and  licensors, if  any, arising out of any kind of legal
claim (whether  in contract,  tort,  or  otherwise)  in  any  way
connected with  the software shall not exceed the amount you paid
to DC Econometrics for the software and documentation.

                             Warranty

DC Econometrics  warrants the diskette and documentation provided
upon registration  to be free of physical defects  in workmanship
and materials for 90 days from date of registration. In the event
of notification within the warranty period,  DC Econometrics will
replace the defective diskette or documentation.  

The  remedy  for  breach  of  this  warranty  shall be limited to 
replacement and shall not encompass any other damages,  including
but  not  limited  to  loss of profit,  and special,  incidental,
consequential, or other similar claims.

DC  Econometrics   excludes  and  disclaims  any  and  all  other
warranties expressed  or implied,  including but  not limited  to
implied  warranties   of  merchantability   and  fitness   for  a
particular application.

DC Econometrics  and its affiliates cannot and do not warrant the
accuracy, completeness,  currentness, merchantability, or fitness
for a  particular purpose  of its  software or data.  In no event
will DC  Econometrics, its  affiliates, agents,  or licensors  be
liable to  you or  anyone else  for any  decision made  or action
taken by  you in  reliance upon this software or its output.  The
entire risk  as to  its quality and performance is assumed by the
purchaser.   The purchaser relies on the software entirely on his
own risk.

In no event will DC Econometrics be liable for any loss of profit
or any  other commercial  damage, including  but not  limited  to
special,  incidental,   consequential,  or  other  damages.    DC
Econometrics and  its affiliates are not responsible for any cost
of recovering, reprogramming, or reproducing any program or data,
or damages  arising out  of the  use of  this product, even if DC
Econometrics has been advised of the possibility of such damages.

This statement  shall be  construed, interpreted, and governed by
the laws of the state of Colorado.


                         1.  INTRODUCTION

This program  contains several econometric models to forecast the
stock market, interest rates, and inflation.  It forecasts all of
them 3  months, 6  months, and 12 months in the future.  There is
also  an   asset  allocation   routine  to   calculate  suggested
portfolios.

Once a month, you enter 8 numbers:  stock prices, interest rates,
CPI, unemployment, P/E ratio, and dividends.  These are available
from  many   sources,  but   I  find  Barron's  most  timely  and
convenient.   The software  stores the  new numbers  and uses its
historical database  to calculate predictions that you can use to
find new Bull or Bear markets.


2.  GETTING STARTED 

2.1  Required Hardware

The  program   is  designed   for  IBM   personal  computers  and
compatibles with  MS-DOS or  PC-DOS 2.0  or above.   It will work
with IBM  PC, XT,  AT, and  PS/2 machines.  You also need 256K of
internal memory and one 5 1/4 inch or 3 1/2 inch disk drive.

IBM, AT,  and  XT  are  registered  trademarks  of  International
Business Machines  Corp.   MS-DOS is  a registered  trademark  of
Microsoft Corp.

2.2  Making a Backup

Disks are  fragile, so  you should  make a backup copy as soon as
possible.   Since the  program stores historical data, you should
make a  backup copy  every few  months, rewriting the old backup.
If you  have a power failure while the program is writing data, a
file could be lost.

          To make backup to floppy disk:

               1.  Boot up in MS-DOS
               2.  When you get the A> prompt type: diskcopy
               
                   The DOS disk may need to be in the A drive
                   if you have no hard disk.  The computer will
                   respond:
                   Insert SOURCE diskette in drive A:
               
               3.  Insert the ECON disk and press any key
               4.  When it asks for TARGET disk, insert a blank
                   diskette.
               5.  Verify that the following files are located on
                   your TARGET disk:
                    
                    ECON.EXE            UNEMP
                    SP500               CPI
                    TBILL               SPPEM
                    TBOND               SPDIVM
                    PRIME               ISTYR
                    ECON.TXT            README.TXT

          To make backup to hard disk:

               1.  Boot up in MS-DOS
               2.  Put SOURCE diskette in drive A
               3.  Create a directory on the hard disk
                    mkdir \econ
               4.  Copy files to hard disk
                    copy a:*.*  c:\econ
               5.  Verify that the following files are located in
                   the directory:
                    
                    ECON.EXE            UNEMP
                    SP500               CPI
                    TBILL               SPPEM
                    TBOND               SPDIVM
                    PRIME               ISTYR
                    ECON.TXT            README.TXT

2.3  Running the Program

               1.  Boot up in MS-DOS
               2.  Insert the ECON disk into the floppy drive or
                   move to the \econ directory (cd \econ)
               3.  Type: ECON   (and press Enter)
               4.  The program will load, and the shareware
                   message will come up.  Press return.  You
                   will then see a list of options for output
                   devices.  I usually choose 1 for screen,
                   or 2 for printer.


3.  MAIN MENU 

The program  reads in  the historic data.  This takes one minute.
After the data is loaded, the main menu is displayed:

               1.  INPUT MONTHLY DATA
               2.  ECONOMETRIC FORECASTS
               3.  PLAY WHAT IF GAMES
               4.  EDIT FILES
               5.  HISTORICAL FORECASTS
               6.  ADJUST FILE TO NEW BASELINE
               7.  EXIT THE PROGRAM
               8.  HELP
               9.  PRINTER OPTIONS
               10. PRINT REGISTRATION FORM

You decide  which choice  you want, type its number (1 to 10) and
press Enter.  The next ten sections will describe the function of
each choice.


3.1  Input Monthly Data

Once a month, new data needs to be entered.  I get this data from
the first  issue of Barron's available each month.  Other sources
may be  used, however.   Barron's  is available  on Mondays  with
numbers from  the previous  week, so  if Friday  was in  the  new
month, I use those numbers.  You only need 8 numbers.  It is best
to record  them for  future reference.   A  log  sheet  for  this
purpose is enclosed.

You may subscribe to Barron's by writing to Barron's, 200 Burnett
Road, Chicopee, MA  01021.  Or call 1-800-328-6800, Ext 292.

If you  can't get  the data for the first week in the month it is
OK to  use next  week's numbers.  Small changes do not affect the
forecasts, but it's best to be consistent.

The data  is stored  in ascii  files so  you can use them easily.
The program  stores 25 years of history.  Every 5 years, on years
ending in  0 and 5, the oldest 5 years of numbers are erased.  So
if you want to keep older data, archive a copy of the files soon.

Here is an explanation of the 8 numbers you enter monthly:

S&P 500  close:   This is the closing value of the S&P 500 index,
an  average   of  the  500  stocks.    It  is  not  the  futures,
industrials, utilities,  or financials.  It is the close from the
first Friday in the month.

Prime Rate:   This  is the  Prime interest rate charged by United
States banks.   It's  hard to  make a  mistake on  this  familiar
number.  It is the latest available rate as of Friday.

90 Day  Treasury Bill Rate:  This is the latest week's rate on 13
week T  Bills.   It is  the Average Discount Rate, not the Coupon
Equivalent Yield.  This is a small but important difference.  The
yield will be higher than the discount rate.

30 Year Treasury Bond Rate:  This is the latest week's rate on 30
year Treasury  bonds.   It  is  reported  under  Adjustable  Rate
Mortgage Base Rates in Barron's.

Consumer Price  Index:  This is reported in Pulse of Industry and
Trade in  Barron's.  It is the familiar number used each month to
gauge inflation.   Due to reporting delays, it lags the others by
two months so October's entries include the August CPI.

Percent Unemployment:   This is the Unemployment Rate in percent.
It, too,  lags the  others by two months and is also found in the
Pulse of Industry and Trade.

S&P 500 Dividend %:  This is the dividend yield of the S&P 500 in
percent.   It is  reported in  Barron's  under  Indexes'  P/Es  &
Yields.

S&P 500  P/E Ratio:   This  is the Price/Earning ratio of the S&P
500, reported near dividends.


3.2  Econometric Forecasts

After entering  the new  data, choose this option 2 from the main
menu.  The predictions will be displayed on the screen and dumped
to the  printer if you chose the print option earlier.  Forecasts
include the S&P 500, bonds, T bills, and inflation, all 3 months,
6 months and 1 year in the future.

The program  calculates a recommended asset allocation .  The 3-,
6-, and  12-month forecasts  are combined to produce an estimated
return for the near term.  The calculated yields are displayed on
an annualized  basis for easy comparison.  The dividend return is
added to  stocks' return from appreciation.  The asset allocation
routine looks  at a  portfolio containing  a mix of three assets:
Stocks, 30 year T bonds, and 90 day T bills.

The estimated  return and annual standard deviation of return for
each asset  is displayed.   Two  portfolios are  calculated.  The
Conservative Portfolio  is the  one  with  the  minimum  standard
deviation of any possible combination of these 3 assets.  It will
be mostly  T Bills due to their risk-free nature and low standard
deviation of return.

The Aggressive  Portfolio is  willing to  take more  risk if more
return is  possible.   It aims for a return halfway between the 2
highest yielding  assets.   While many  combinations will achieve
this return,  the Aggressive  Portfolio displayed  does so  while
accepting the minimum risk necessary to get the return.

If you  want to  get more  aggressive, add  more of  the  highest
yielding asset.   There  is an "efficient frontier" of portfolios
which achieve  a given  return at  the minimum risk.  This can be
thought  of  as  a  curved  line  starting  at  the  Conservative
Portfolio, passing through the Aggressive Portfolio and ending at
100%  of   the  highest   yield  asset.    You  can  get  a  good
approximation of intermediate points with simple averaging of the
above portfolios.

Assets which  have a  negative  estimated  return  (possible  for
stocks or  bonds in bear markets) are immediately eliminated from
consideration because  cash in  a mattress  has a  0% return with
zero risk.  T bills will probably look better than cash.


3.3  Play "WHAT IF" Games

This option  allows you  to enter  data and get forecasts without
storing the  numbers.   It is  useful  for  checking  effects  of
surprises such  as  market  crashes,  big  up  days,  prime  rate
increases, or  other news  which causes  concern.  It also allows
you to  enter imaginary  data and  play with the numbers to get a
feel for  what is  bullish or bearish.  Use this routine to enter
the most current data to get forecasts weekly if you like.

All 8  numbers described in "Input monthly data" must be entered.
You can  input these  numbers for  as many  months as you choose.
For reference,  the most  recent data  is displayed.  To re-use a
displayed number, just hit Enter when asked for its value.


3.4  Edit Files

If a  wrong number is entered and stored on disk, the problem can
be fixed by editing the data file.

     1.   First, choose  which of  the 8  data files  you want to
     edit.   A list of entries is printed with a reference number
     to the left of each one.

     2.   To change  an entry,  enter its  reference number, then
     change the  value.   To delete  the last  data value,  enter
     zero for  its value.   Deletions  are  not  allowed  in  the
     middle of  the list.  The last 12 entries are displayed, but
     older entries  are available  by  entering  their  reference
     numbers.


3.5  Historical Forecasts

Enter a year and month from the past and the program will use its
database and models to give you its forecast and asset allocation
for that time.  This allows you to check the program's ability to
spot past  market movements.  For example, look at the great bull
markets starting  August 1982  or December 1974.  Test it against
the August 1987 peak before the October crash.

The historical  forecasts are  calculated from  data available at
that time,  using the  same  models  used  to  make  the  current
forecasts and  the same  asset allocation  routine.   The  linear
regressions used  to create  the models  used data  from 1963  to
1989, so  the models  will perform  well over  that time.    They
cannot be  guaranteed to  continue to  perform  as  well  in  the
future.


3.6  Adjust File to New Baseline

This routine is rarely used.  Prices are subject to large changes
over decades and occasionally the Commerce Department will make a
major adjustment.   Consumer prices use 1982 to 1984 for the base
year (1982-1984=100).   The  base  year  was  once  1967.    This
adjustment caused  CPI values  near 300  to plunge  to  100  when
prices actually  hardly changed.   The  computer could  interpret
this as a crash in prices, so the historical data must be revised
to the new index.

To do  this, enter a number from the old index and the equivalent
value from  the new  index.  The program will adjust all numbers.
If one  of the new numbers has already been entered, this routine
will adjust it improperly.  Use EDIT to verify that everything is
OK after  adjusting the  file to  the new baseline.  Only CPI and
Stock prices  will need  this adjustment  since  the  others  are
expressed in percent.


3.7  Exit the Program

Choose this when you are finished and want to return to DOS.


3.8  Help

A short  explanation of  main menu  items is  available.  Use the
Print Screen (PrtSc) key if you want a hard copy of the paragraph
displayed on the screen.


3.9  Printer Options

Output may  be directed  to  screen,  printer,  files,  or  other
devices.   Choose from  the menu.   Choice  1 directs  output  to
screen only.  Choice 2 works for most printers.  To direct output
to a file, choose 6, and enter the filename.


3.10  Print Registration Form

This will send a copy of the registration form to your printer or
other output  device.   It goes  to LPT1:  unless you use Printer
Options to choose another device.    When you register,  you will
receive  a printed  manual  and a disk with  current monthly data 
files.  The registration fee is $39.99.


4.  USING THE FORECASTS 

Many profitable strategies exist.  One should be chosen that fits
your available  capital and tolerance for risk.  The less capital
you have,  the more  commissions will  consume.  Frequent trading
can be  very expensive.   This  program tries  to find  the large
moves which  continue for  months or  years.  You probably should
not make  every change called for monthly in the asset allocation
section.

Backtesting the model  over the period  1971 to 1990  shows a 16%
per  year  average  compound return  from investing  100% of your 
money  in the asset (S&P 500, T Bonds, or T Bills)  forecasted to
perform best under asset allocation.  This method resulted in 2.1
trades per year on average. 

Remember,  these   forecasts  are  not  perfect.    Like  weather
forecasts, they  are subject to error.  I try to give you an idea
of the  errors by  reporting expected  standard deviations  under
asset allocation.   These  are the  standard deviations  observed
from 1963  to 1989.  Standard deviation is a term from statistics
describing a  bell curve  distribution.   68% of  the  time,  the
actual return  will be  within a  range  of  plus  or  minus  one
standard deviation  from the  predicted return.  95% of the time,
it will  be within  2 standard  deviations.   Occasionally, there
will be  larger errors.  The stock market has larger tails than a
normal bell  curve and  there is  reason to  believe the standard
deviation is not stable.

No-load mutual funds that allow telephone switching are ideal for
avoiding brokerage  fees.   Money market  funds are  similar to T
bills but do not have a $10,000 minimums or commissions.

One conservative  strategy would  be to  switch between  a  stock
mutual fund,  a bond  fund,  and  a  money  market  fund  in  50%
increments only,  trying to approximate the aggressive portfolio.
For example, if the aggressive portfolio is 45% stock, 35% bonds,
and 20% T Bills, you round it off to 50% stock and 50% bonds.  If
it later  goes to  35% T  bills, you  go to 50 % money market and
drop either  stocks or  bonds, whichever the aggressive portfolio
is weighting less.

The conservative  portfolio is  the least  risky of  any possible
combination of  the 3  assets.   It is  always mostly T bills, so
it's  not   very  interesting.    Check  its  yield  against  the
aggressive portfolio.   If  you're not getting at least 1% better
expected  yield   in  the  aggressive  portfolio,  stick  to  the
conservative portfolio  and you  will not  be affected  by market
swings.

I don't  recommend  following  only  the  conservative  portfolio
because it  is always heavily T bills.  I print it out because it
is the  most conservative portfolio possible and it is useful for
other  asset   allocation  strategies   such  as   averaging  the
conservative and  aggressive portfolios.   It  is one endpoint of
the efficient frontier, low return and very low risk.  If you are
really that risk-averse, just buy T bills and forget the markets.
Of  course,  during  bear  markets,  this  works  great  and  the
aggressive portfolio should be the same as the conservative, 100%
T bills.

The aggressive  portfolio suits  me best.   Its  yield is halfway
between the  2 best  assets of  the  three  (stocks,  bonds,  and
bills).   You get  some diversity  and some action from the stock
and bond  markets.   If history is a useful guide, you will be on
the right  side of major moves.  Although its name is aggressive,
there are much more aggressive strategies.

To get  really aggressive, hold 100% of the asset predicted to do
best at  the beginning  of the  asset allocation  printout.   The
program gives its expected returns on stocks, bonds, and T bills.
Note that  stock  returns  include  dividends,  and  all  returns
average the  annualized yields  forecast for  3, 6,  and 12 month
periods.   The aggressive  portfolio is not as aggressive as this
strategy and you can expect to get burned occasionally.

If stocks  are predicted  to do  quite well,  you can buy them on
margin and  nearly double  your return in a bull market.  But you
more than double your risk.  A 50% market drop will wipe you out.

The risk  from stocks  is doubled and you have added the risk due
to fluctuations  in the  interest rate  on your margin borrowing.
Due to  interest costs,  your  return  does  not  double.    This
strategy and  all  the  ones  described  previously  lie  on  the
efficient frontier.

The  efficient   frontier  is   the  imaginary   line  connecting
portfolios with minimum risk for their expected return.  The line
starts  at   the  Conservative   portfolio,  passes  through  the
Aggressive portfolio,  and continues to 100% of the highest yield
asset.   If the return on the highest yield asset is greater than
the margin  interest rate,  the efficient  frontier extends  to a
fully  margined  account  containing  the  highest  yield  asset.
Minimum risk  for their  expected return does not imply that they
are all  safe or  that you  should take  such  risks.    A  fully
margined account would have been wiped out several times over the
past century.  Expected returns and actual returns will differ.

If you  want even  more risk, use the options market.  Wait until
the S&P  500 is  expected to  advance 20% per year and buy calls.
Or buy  a call  after a  5% dip  in a bull market.  It is easy to
lose all your money if you are wrong, so treat this like gambling
and bet only what you can afford to lose.  It is speculation, not
investing.

For the totally crazy, there are the futures markets.  90% of the
players lose  money here, so I do not recommend it.  You can lose
more than  your original  investment and limit moves mean you can
be stuck in a losing position.  The advantage is that you get the
gain or loss on a large block of stock represented by the S&P 500
index for  a reasonable  commission.   It is  a game  for experts
only, armed  with more  knowledge  than  this  program  provides.
Large sums  of DISPOSABLE  income and  capital are  required.   T
bills and T bonds also have futures markets.

If you  are playing the futures, this program should be useful to
help you  find the major trends.  The trend is your friend.  More
money is made in the direction of the trend than on the reactions
against it.   You  can be  bailed out  of losing  trades when the
trend re-asserts  itself.   Bull markets can be bought on 5% dips
but not  bear markets.   If  you have  short term indicators, use
them too but only play in the direction of the longer term move.


5.  THE ECONOMETRIC MODELS 

The  econometric  models  used  in  this  program  are  based  on
multivariate linear  regressions.  The variables used were chosen
from hundreds of others because they performed the best.

In 1987,  I used  two advisors'  models  to  forecast  the  stock
market.   Neither  one  predicted  the  crash  that  happened  in
October.   Fortunately, I  had sold  most of my stock that summer
because I  was afraid  of the  rising interest rates and high P/E
ratios.   Still, I  couldn't believe  how many advisors and other
models missed  this major  move.   The crash  of 1987  led to the
development of this program in 1989.

One  problem   with  mathematical  models  is  that  they  become
obsolete.   Stock Index  futures make  it possible to short stock
and  be   neutral  or  long  in  the  market.    They  also  give
institutions more  reason to hold cash.  Therefore short interest
and mutual fund cash, 2 good indicators, have changed in meaning.
The options market is larger now by far than it was 10 years ago.
Odd lot  traders now  use options.   Models constructed ten years
ago often did not plan for such changes.

To avoid  these pitfalls,  I chose  the simplest, most elementary
variables I  could find.   They  also produce  good correlations,
which should  not be surprising, since they are basic measures of
value.

Interest rates  tell about  the attractiveness of alternatives to
the stock  market.  They also tell whether the economy is booming
or busting.   Interest  Rates are  a measure  of the  supply  and
demand of  money.  When rates are high and rising, stocks will be
falling.   If I  could have only one indicator, it would be short
term interest rates.

Consider market  history 1948  to 1987.   When  T bill rates fell
December to December, the return on the S&P 500 averaged 23.4% in
the following  year.   When T  bill rates  rose, returns averaged
6.7%.  T bill rates were rising going into the crash of 1987.

The yield  curve compares  short term rates to longer term rates.
When short  rates are  higher, the  yield curve  is  said  to  be
inverted.   It is  not normal  and usually  precedes  recessions.
Stocks usually fall during these times.  In his book Stock Market
Logic, Norman  Fosback reports  the impact  of the  yield  curve.
Normal yield  curves were  followed by  an average 11.5% per year
gain on  the S&P  500 in  the years 1941 to 1975.  Inverted yield
curves were followed by an average 0.7% loss.

Consumer prices  are a  measure of  inflation.  When inflation is
low, interest  rates stay  low, and the stock market rises.  High
or rising  inflation can  be curbed  by raising interest rates to
choke off  demand.   This also  strangles the stock market.  High
inflation offers  people an  alternative to the stock market.  It
is better  to spend  the money  on goods  before prices  rise, or
invest in real estate.

Unemployment also exerts political pressure on the Fed.  They are
hesitant to  raise interest  rates during high unemployment.  The
economy needs  stimulation, not  restraint.   A good  time to buy
stocks is when unemployment is rising in a recession.

Dividend yields are an excellent long  term forecaster.  They set
record lows before the October 1987 crash.  They were low in 1929
and also  in 1973  before bear markets.  Low dividends called the
sharp 1962  drop.  High dividends predicted the 1982 bull market,
the 1975 bull market, and the long advance of the 1950s.

In the  years 1948  to 1987,  dividend yields above 4% on the S&P
500 led  to an  average 20.0% return on the S&P 500 the following
year.   Dividend yields below 4% were followed by an average 5.1%
return.

Price/Earning ratios  are another  good measure  of value.  It is
always safe  to buy  the S&P  500 at  low P/E's.   The  long term
average is  about 14,  so a  P/E of  8 for  the S&P 500 is cheap.
Recessions can  depress earnings,  so it is safe to buy at higher
P/E's then  because good  times will  return and  carry  earnings
higher.

I believe  these variables are of such a basic nature that change
in our  markets or  economy won't change their effects much.  But
if you  have reason to suspect these variables, don't follow this
model blindly.

5.1 Checking the forecasts

Here is  a simple  way to check the model's forecasts for the S&P
500.   We can  combine T  bill rates  and dividend  yields into a
model that  you can  do in  your head.   We  have seen  the  high
returns on  the S&P  500 with  falling T  bill  rates,  and  with
dividend yields  above 4%.   Combining  these 2 variables is even
better.   With both  variables bullish, the average return on the
S&P 500  is 30.8%  per year.   With only one of them bullish, the
return is  9.7% per year.  With both in bearish territory (rising
T bill  rates on  a 12 month basis, and dividend yield on the S&P
500 less  than 4%)  the average return on the S&P 500 is 3.7% per
year.

Of course  the models  in the  program are much more complex, and
should be  more accurate.  This example demonstrates the power of
a two-variable model and used data from 1948 to 1987.

The  business   cycle  provides   an  independent  check  on  the
forecasts.   For details,  see  Bowker's  book  Strategic  Market
Timing.   Events may  occur somewhat out of order, but you should
be able  to get  an idea  of where  the economy  is heading.  The
sequence of events is:

1) Stock Market bottom
2) Leading indicator bottom
3) Coincident indicator bottom
4) Unemployment peak
5) Official announcement of Expansion
6) Lagging indicator bottom
7) Inflation bottom
8) T bill interest rate bottom
9) T bond interest rate bottom
10) Stock Market peak
11) Leading indicator peak
12) Unemployment bottom
13) Coincident indicator peak
14) Official announcement of Recession
15) Lagging indicator peak
16) Inflation peak
17) T bill interest rate peak
18) T bond interest rate peak
19) Back to Stock Market bottom

If you  want graphs  of these indicators, subscribe to the Survey
of  Current   Business  by   calling   202-783-3238,   or   write
Superintendent of  Documents, U.S.  Government  Printing  Office,
Washington, DC 20402.  The cost is $29.00 per year.


5.2  Details on the Regressions

I used  stagewise multivariate  linear regression  on 25 years of
data spanning  1963 to  1989.   In stagewise regression, the best
variable is  found and  its effect  is subtracted  out.  Then all
variables are examined again for correlations to the residuals.
To enter  the model, a variable must have 99% confidence that its
correlation is  not due to random variations.  An F test was used
for this.  The process continues until no more variables can pass
the F test.

Many predictor  variables can  be constructed  from the database.
The Prime  rate can  be compared to moving averages from 3 months
to 4  years in length.  Dividend yields can be compared to T bill
yields.   The yield curve can be computed as T bill/T bond rates.
I looked  at 165  such predictor variables for each model to pick
only the most effective variables.

Correlation coefficients  are a  measure of  a model's fit to the
data.  They range from -1.0 (perfect negative correlation) to 0.0
(no correlation) to 1.0 (perfect correlation).  Here is a list of
the model's correlation coefficients:

     S&P 500 (3 months ahead).....................R = .53
     S&P 500 (6 months ahead).....................R = .64   
     S&P 500 (1 year ahead).......................R = .80
     T Bonds (3 months ahead).....................R = .46
     T Bonds (6 months ahead).....................R = .58
     T Bonds (1 year ahead).......................R = .74
     T Bills (3 months ahead).....................R = .56
     T Bills (6 months ahead).....................R = .66
     T Bills (1 year ahead).......................R = .68
     CPI     (3 months ahead).....................R = .83
     CPI     (6 months ahead).....................R = .87
     CPI     (1 year ahead).......................R = .90

R values  are not  the only  measure of  a model's effectiveness.
There are many ways to get a high R value.  One way is to include
a lot  of variables which have a low confidence.  My 99% standard
is more  rigorous than many others.  Another way is to regress to
something easy.   I  am regressing to the changes in my predicted
variables, not  to their  levels.   Over a  25 year  period,  the
previous day's  level of  S&P 500  correlates with R > .99 to the
next day's  level, but  tells you  nothing about  which  way  the
market will  go.   The money  is made  on the changes, not on the
day's closing value.


6.  GETTING HELP 

Your satisfaction  is of  utmost importance  to us.   If you need
help, first  consult this  manual or the help menu.  If that does
not fix your problem, please write to:

                    DC ECONOMETRICS
                    2920 Mount Royal Court
                    Fort Collins, CO  80526

We want to hear your comments, suggestions, and criticism.


6.1   Error Messages

If you ever see an error number, here is a list of the most common
ones.  

Error#   Meaning
------   -------
24       Device time out.  COM device?
25       Device fault.  Printer interface?
27       Out of paper.
53       File not found.  Change directory to the one containing
         econ.  Type "dir" and check for econ.exe and the 9 data
         files.  If your prompt is for the C: disk and econ is 
         on B:, type "b:".
57       Device I/O error.  Probably your printer is off.
61       Disk full.
70       Permission denied.  You tried to write to a write-
         protected disk.
71       Disk not ready.  Floppy door open?


7.  RECOMMENDED READING 

The Encyclopedia of Technical Market Indicators
Robert W. Colby and Thomas A. Meyers
Dow Jones-Irwin, 1988

          A  comprehensive  collection  of  graphs  and  computer
          research on over 110 indicators.

Stock Market Logic
Norman G. Fosback
The Institute for Econometric Research, 1976, 1984

          This book  is a  thorough discussion of virtually every
          stock  market  indicator  in  use.    Norman  Fosback's
          scholarly  approach  to  stock  market  forecasting  is
          without equal among advisory services.

Winning on Wall Street
Martin Zweig
Warner Books, Inc., 1986

          Marty Zweig's  emphasis  on  interest  rates  and  tape
          action first influenced me in 1982.  He talks about the
          market in  scientific terms  I  can  relate  to  as  an
          engineer.  His record proves he is right.

Tight Money Timing
Wilfred R. George
Praeger Publishers, 1982

          This book  shows the  effect of  interest rates  on the
          stock market from 1914 to 1981.

Asset Allocation
Robert D. Arnott and Frank J. Fabozzi
Probus Publishing Co., 1988

          Everything  you   ever  wanted   to  know  about  asset
          allocation and  then some.  It is a complete discussion
          of theory  and application.   Many contributing authors
          present their views.

A Consensus Approach to the Determination of Not-So-Good Years to
Own Common Stock
Edward Renshaw
Financial Analysts Journal/January-February 1989

          How to  forecast the  S&P 500  using T  bill rates  and
          dividend yields.   This simple method is similar to the
          simple model in the manual, and will usually agree with
          the complex models in the software.

Strategic Market Timing
Robert M. Bowker
New York Institute of Finance, 1989

          Use predictable turning points in the business cycle to
          forecast the direction of interest rates, stock prices,
          and other economic indicators.  This method provides an
          independent check on econ's forecasts.

How to Forecast Interest Rates
Martin Pring

          Interest rates  are a  lagging indicator.   They  go up
          after the economy goes up.

The Practical Forecaster's Almanac
Edward Renshaw
Business One Irwin, 1992

          137 Reliable Indicators for Investors, Hedgers, and 
          Speculators.  Forecast the stock market, interest rates,
          or recessions.


          8.  PLACING AN ORDER 

          Order by  mail.   Payment may  be made  by credit card,
          check, or  money order.   Use  the  order  form  below.
          Shipping is free in US and Canada.  International orders
          please add $3.00.   We prefer payment by check.

Where did you obtain your copy of Econ? _________________________

          DESCRIPTION                   QUANTITY   PRICE

          5.25" diskette and manual     ________   $39.99

          3.5" diskette and manual      ________   $39.99

          TOTAL ENCLOSED                ________


Please charge my ____ VISA    ____ MASTERCARD

CREDIT CARD NUMBER ____________________________________________

Expiration Date    ____________________________________________

Signature          ____________________________________________


Ship to:

Name:__________________________________________________________

Address:_______________________________________________________

City:___________________________State__________Zip_____________

Country:_______________________________________________________

Please send your order to:

                    DC ECONOMETRICS
                    2920 Mount Royal Court
                    Fort Collins, CO  80526

To register your copy of econ, choose 10 from the main menu to
print a registration form.  For your $39.99 registration fee,
you will receive a disk with all current monthly data files and
a printed manual.  Registered users may order a new disk with 
current data files any time for $9.00.  This is handy if you 
forget to enter data for a while but want to start using the
program again, or if you destroy your disk.  Please specify 
5.25" or 3.5" disk.

To view this manual on screen, enter "type econ.txt | more" from
MS-DOS.  This will let you see one screen at a time.  Press the
space bar to see the next one.

To send a copy to your printer, ready the printer, and type
"copy econ.txt > prn".  Don't type the quotes, just the stuff
inside them.

