The FIVFIV population projection program, available for a free two-week evaluation trial at this address: http://www.visitus.com/popsite/software/fivfiv/#98b .  It can be downloaded directly from this page: http://www.visitus.com/popsite/software/fivfiv/download.htm .  You'll need Adobe Acrobat to read the .pdf file that is packed into the very long manual.zip. FIVFIV will project populations by applying estimates of fertility, mortality, migration, and (optionally) many more variables.  Before you begin, you will need country level data, available from this census bureau link: http://www.census.gov/ipc/www/idbprint.html , and you should also familiarize yourself with pertinent trends in the country or area you've chosen.  As you collect your data, record in detail these population figures for the lastest year available: age/gender distribution for five-year cohorts (0-5 to 75+), age/gender distribution in percentages for migration, and age/gender distribution in percentages for total fertility rates (five-year age groups from ages 15-49).  Read this document carefully, then record your current-year data and your assumptions for 35 years using parts of the form I've created here. (you'll need Acrobat Reader to see and print this form)

TO DOWNLOAD FIVFIV:

• Select the three files -- readme.txt, program.zip, and manual.zip and save them one at a time to a new subdirectory (c:\fivfiv\)
• If you don't have an unzipping program, get pkunzip from the FIVFIV download page listed above.  Put it in the same subdirectory with the three files.
• In windows explorer, go to the C:\fivfiv\ subdirectory, then go to the DOS prompt (Start, Run, "command").
• unzip program.zip.  If you don't know another way, use the pkunzip.exe file I suggested you get (see above) -- type "pkunzip -d program.zip".  This will create the program.  You can do the same with the manual if you'd like (it's big).

Choose a country with which you are reasonably familiar.  You will be projecting a high- low- and moderate-variant projection for the country, so you must research trends, policies, and migration issues for your country.  You can find current and historical data at the census site listed above, but you will need to investigate more broadly to understand the underlying forces for change.  As you do your research, keep good notes and full citations, as these will appear in your report.

TO START A PROJECTION:

• You'll want your projection to have some "fuzziness;" so, to build some uncertainty into it you'll do three projections: a low, a medium, and a high variant.  Low projections should use all reasonable estimates which would result in a smaller future population -- e.g., longer life expectancy, faster drop in total fertility rates, higher rates of outmigration, etc. and high variant projections should use all reasonable assumptions which would cause a larger population in the future.  Intermediate projection should be your very best guess.
• In windows, start program fivbld.exe.  This will let you enter all the current data and future assumptions, one projection at a time.
• The necessary criteria are indicated by an R (required). As you enter data, an asterisk will appear after each option when you complete it to indicate "done."  To change or fix entries, just enter that option again, and "enter" through all the correct data, changing only what is wrong.
•  Starting Year and Ending Year (it works in 5-year groups).  Make starting year latest year with good data (preferably in an even increment: 2000 or 1995), and a title for the projection (e.g. Mexico: Low Variant). The default ending year is 35 years from the starting year.
• Life tables are explained in Chapter 14 of the manual, which recommends the West model as a reasonable one in most cases.  Life tables suggest age distribution of mortality; the various models are most different from one another when life expectancies are low. If you're working in a country with very low life expectancies at birth, you should probably research the other choices.
• Initial population should be accurate data for a recent year (e.g. 2000). You'll need to have on hand  numbers of females in 5-year cohorts to 75+, then males for the same age ranges.
• MORTALITY is entered most easily as life expectancy for females, and then males.  For each gender, first enter the last available real data (for projection begin date) and then estimate life expectancies in five-year intervals for the duration of the projection.  If gender-specific data are not available, put in total population life expectancies for "female" and then just Enter through the "male" fields ("female" will be applied).
• FERTILITY rates are most easily entered as Total Fertility Rate, but Net Reproduction Rate (girl babies surviving for every woman) is better if there is a strong gender preference in the country.  In any case, plug in actual TFR or NRR for the initial year and then estimates for each five year period.  If you have access to this information, enter the age specific fertility rates for each five-year cohort (15-49) for the initial year.  Finally, to allow for changes such as later marriage ages, improved fecundity for older women, etc., you may optionally assign the initial percentage distribution to something less than the whole projection period.  Assigning the initial distribution to 3 periods, for example, will apply those age-specific rates to the first 15 years of the projection.  For the remaining period you may assign different distributions.  For example, you may have reason to believe average age at marriage will increase, so girls 15-19 will become somewhat less "productive" relative to slightly older age groups.  If you don't expect the age-distribution of the fertility rates to change, apply the initial distribution to the all five-year periods.
• MIGRATION is most easily entered in net quantities, broken down by age (five-year cohorts to 75+) and gender.   If more people are moving out of a country than to it, this net value will be negative number. You'll need at least two things: (1) estimates for males and females separately at 5-year intervals of annual net migration through the entire study period. It's probably easiest to simply enter number of migrants by age but it can also be done by proportion or percentage -- the program will compute the relative distribution and apply that to the total you entered earlier.  As before, you will be prompted first to enter female data, then data for males.  These annual rates will be applied to each of five years; and (2) the distribution of these people by age and gender.   Again, whether you use percentages or numbers the program will compute percentages and distribute the total migration thusly.  As in age-specific fertility rate, above, you may optionally apply the initial age-distribution to something less than the entire projection period.  For example, if you assume that in the future some age groups of migrants will become more or less represented than they are at present, apply the initial distribution to just the first 10 or 25 years of the projection (2 or 3 five-year periods), and then apply a revised distribution for the remaining period.
• the program doesn't require migration, but it's the third necessary component in the demographic equation; I will require it
• If the site does not have migration data, you'll have to estimate the numbers, broken down by age and gender.  This won't always be easy -- it might be helpful to "model" your distribution and size off of another country, making modifications you feel are necessary.
• Before you save your data, ask for
• PYRAMIDS -- get a populaton pyramid for the initial period and (for each variant) the end period. Just select option # 22.

SAVE YOUR FILE (e.g., Mex_L.fiv).  You will be able to go back and edit the parameters without typing in all the Start Year data.  For example, change your assumptions to high variant assumptions, then save the file as Mex_H.fiv, then do the same for the Medium assumptions.  Saving the .fiv file will not run the projection -- it will only write to a file which will be put in your C:\fivfiv| subdirectory with the program.

TO MAKE THE PROJECTIONS

• In windows, select fivfiv.exe, and type in the file name (e.g. Mex_L.fiv) and indicate an output file (e.g., Mex_L.out)
• Run.  The Program will terminate; then read your file with a text editor that can handle a fairly large file (wordpad.exe if nothing else)..
• Print your Mex_L.out , and prepare a report of your projection.

To make your own pyramids, find the age-gender data in the fivfiv output file, and type it into a spreadsheet as follows:
 
 

AGE	Males	Females
0-4	-14,530	14,212
5-9	-13,980	13,722
10-14	-13,745	13,690
... and so on

Hint: If you have the data in a file, you can cut and paste.  Highlight, CTRL+C to cut, position cursor, CTRL+V to paste.
Hint: You can type in the male numbers as positive, then in another column subtract cell from 0.  Then highlight, Copy, Paste Special.  Choose "values" to overwrite formulas.
Then move them into place (middle column, above example).

EXEL INSTRUCTIONS:

If you have your data set up as above, highlight all three columns or just the two columns of data and dropdown Insert, Chart, Bar, Stacked Bar.  With a little fine tuning (add titles, change font size, etc., by double clicking on the various parts of the chart) you can easily get a pyramid like the one to the right.  I added the trace line with Options, Series Lines.

 

For the line graph, you'll be using three series, all with the same data for the preceding years, and then diverging thereafter.

QUATRO PRO
In QuatroPro choose stacked rotated bar graph.

WRITE THE REPORT
Your report should include the following:

• Coordinator should summarize the important demographic trends and situation in your chosen country.  This will serve as an introduction to the sections that follow.
• Individuals should summarize fertility/mortality/ or migration for their country, emphasizing the things that influence it and may suggest future changes.  These should be 1-2 pages long and can rely entirely on internet sources.  In the summary and/or the sections the paper may include a historical sketch, data tables, URLs, or whatever will bring the reader up to speed regarding the dynamics within your country.  You'll find plenty of resources for this research off my "sources" page.
• Summary of the assumptions which went into your Low, Medium, and High projections.  These will be written by individuals.  Include the important values in this summary, for example, "I assumed TFR would quickly increase from the current 1.4 to 2.6 by 2010, resulting from an aggressive pro-natalist policy, and then stay between 2.6 and 2.55 until 2035."  You should also include supportive material (e.g., "Between 1980 and 2000 LE increased from 75.2 to 77.3 and from 71.3 to 74.0 for females and males, respecitvely.  For my high projection I assumed this increase would continue for a final 81.1 and 78.0.")
• A line graph of the most basic result: Total Population, in five-year periods for Low, Medium, and High projections -- all lines together.  If the lines don't all begin at the same place in the starting year, you've done something wrong. Example: Canada.
• Four pyramids: Current, Low-, Medium-, and High- end projections.
• For the medium projection do the following:
• A comparison of the age-dependency ratio between the starting year and the end year.   ADR=((#0-15 + #65 and over)/ # 15-65)*100.
• Any interesting trends or changes (sex ratio?, average age? ...)
• Concluding observations or comments.  This is written by the coordinator.
• Your report must inclued the three ".out" files so I can check your work.

*Year to start projection
*Number of years to project (default=35)
*Age/gender distribution for start year in 5-year cohorts (collapse the very elderly to one group "75+")
Total Fertility Rate: start year (accurate) and estimates in 5-year intervals for entire period
Age distribution in percentage of the Fertility Rate (15-19, 20-24, 25-29 ..., 45-49)
    a change in age distribution -- an adjustment of the above percentages -- at some single point during the projection period
Annual Net Migration, in numbers, for males and females separately starting with start year (accurate) and estimates in 5-year intervals
 Distribution of net migration in percentage for males and females separately in 5-year cohorts (collapse the very elderly to one group "75+")
   a change in age distribution -- an adjustment of the above percentages -- at some single point during the projection period
Life Expectancy at Birth for males and females separately in 5-year intervals for the entire period
    If very low LE, then read about the non-Western life table models in the manual

© Erick Howenstine 1999-2003