1.      PREPARING INPUT

ABOUT THE INPUT TEMPLATE

       The Input Template (named as I_Prost14_template.xlsx) is an Excel workbook
designed to enter data and assumptions necessary for running PROST141. This template
contains the following eight worksheets2:

        General                     General (not-age-specific) information

        Population                  Age-specific demographic data/assumptions

        Labor                       Age-specific labor market data/assumptions

        Pension                     Age-specific pension system data/assumptions

        Profiles                    Profile data for six different individuals

        ReformGeneral               General (not-age-specific) parameters for systemic reform

        ReformPension               Age-specific parameters for systemic reform

        Factors                     Actuarial tables



        To use this template, load the file I_Prost14_template.xlsx using PROST as it is
described in the manual. It is important to notice that the file name has three parts. The
first part “I_” indicates that this is an input workbook. The second part is reserved for
country or case names. The third part (.xlsx) is the file identifier for an Excel workbook.
It is necessary to follow the naming convention described here. The user can specify the
file name in any way without changing the first and last parts of the file name.

        Data Entry: Enter data as per the instructions given here. It is not necessary to
enter the data in one session. You can stop data entry at any stage and start again at a later
point if so desired. Remember to save the input file periodically.

1
     Pension Reform Option Simulation Toolkit, copyright IBRD, 2012.
2
     Users can add more worksheets for rough/background calculations if necessary though it is not
     necessary to do so.



                                                  1-1
        Saving Input Files: The user is advised to save the input file often during data
entry to prevent data loss. Click FILE and then Save As. Change the file name into what
makes sense to the user, such as I_CHN_A.xlsx for case A of China or I_UKR_C.xlsx
for case C of Ukraine. Do not change the first part (“I_”) or the file extension (“.xls x”)
otherwise PROST will not be able to read it. Note that the ‘I” prefix is used to denote an
input file. After running PROST, an Excel file with the same name but prefixed with an
“o” will be created. This new file would contain the output tables from PROST runs.

       The amount of time it takes to assimilate and enter the data required for these
simulations should not be underestimated. It is suggested that the user maintains careful
notes about the sources of input data and assumptions used for the data for future
reference. Note that all areas of the input files are protected and the user may enter
comments/notes only in the space allocated for such annotated comments.

SIMULATION HORIZON, BLANK CELLS, DATA INTERPOLATION, INSERTING                           AND
DELETING DATA

        The program can handle simulations up to 100 years or more. The exact
capabilities would depend largely on the computer being used. The number of years in
the simulation is determined by the user, usually ranging from 50 to 75 years for pension
reforms. A 75 year horizon is suggested since it covers the remaining lifetime of those
people beginning work today. Users select a base-year (e.g., 2010) and an ending-year
(e.g., 2075), and enter these two dates in worksheet “General”.

        Any cell that is left blank is assumed to have a value of zero. The user is
encouraged to check, if there are no “unintentional” zeros of this kind in the input file.
For example, if you specify a positive number of 40 year old retirees, make sure, that
length of service and replacement rates for that particular cohort are also positive.
Otherwise, PROST will interpret the data as some people retiring with zero pensions,
which will bring down the average pension in the economy. Similarly, if you have
positive number of 60 year old contributors, you should also have a positive earnings
profile for that cohort. Otherwise PROST assumes that these contributors are working for
free and therefore collected revenues will be reported lower than actual. It is safest to fill
as many cells as you can. For example, if you specify a positive replacement rate for 20
year old it will not have any impact, unless you also specify a positive number of retirees.
However, not specifying a replacement rate for the cohorts that have positive number of
retirees will have a major impact on the results.

         All data and assumptions are put in defined rectangle range blocks sandwiched by
a first column containing information for the base year (e.g., 2010), and, a last column
containing information for the ending year (e.g., 2075). In most instances, the end year
data are usually derived from a reasonable hypothesis or assumption about the way the
base year data are expected to behave over time. The program linearly interpolates
between the two values to determine values for each year of the simulation. Sometimes it
may be necessary to assume that the end year data and the base year data are the same if
the variable is not expected to change over time. On the other hand, there may be
instances when it becomes necessary to specify a hypothetical time path for some of the


                                             1-2
input file variables. For example, maybe GDP growth is expected to be high for the next
10 years and then to slow down thereafter. In such cases, data for additional years
between the base year and the end year could be entered by inserting cells between the
base year and the end year. If the user wishes to insert or delete data for intervening
years in any of the spreadsheets, he/she can use the buttons for inserting(I) and
deleting(D) columns of data. By clicking on the “I” icon, a column is inserted for data
entry with the year field as well as other data fields linearly interpolated according to
values in neighboring cells. The user can over-write these values if necessary. As the user
adds intermediate years, interpolated values for these intermediate years appear in the
data cells. These can be overwritten by the user. Conversely, by clicking on the ”D”
button, data are deleted for any year. Numerical data fields between two specified years
(e.g., 2010 and 2015) are interpolated linearly.

   Note: You will not be able to insert data before the base year or after the end year,
   and you will not be able to overwrite the base year or end year values by manually
   typing over them. They can be changed only in the lines 14 (labeled “Base Year”)
   and 15 (labeled “Ending Year”) of spreadsheet “General”.



Please note that line and column numbers refer to the original template. As the user
inserts lines or columns, line numbers may no longer match what is written in the
manual. Therefore, it is advised to refer to line and column labels rather than to
line and column numbers.

WORKSHEET I: GENERAL

The spreadsheet “General” contains user options for running PROST, data about general
assumptions pertaining to the economy and some parameters of the pension system. The
variables need to be entered exactly in the units specified. The following is a list of
variables in spreadsheet “General”:

   Beneficiary Categories: Enter the categories of pensioners being analyzed. PROST
   allows up to four categories to be selected at a time. The first category will be
   modeled with the greatest precision by tracking the beneficiaries over time from the
   assignment of the benefit until the beneficiary dies. The benefit income distribution
   will also be modeled within each age and gender cohort. This category will be
   referred as ‘old age pensioners’ in this document; however, the user can change the
   name if he wishes. The second category will be called “disabled” throughout the
   document. The number of beneficiaries will be modeled in the similar way to “old
   age pensioner” category, but income distribution within the age and gender cohort
   will not be modeled. I.e. each cohort will be assumed to consist of identical
   individuals assigned the same average benefit.

   The last two categories will be modeled by figuring the total number of beneficiaries
   every year based on user specified beneficiary rates. In this document these two
   categories will be referred as ‘survivors’ and ‘orphans’. These last two categories



                                           1-3
could also be used for projecting child allowances, unemployment allowances, and
other types of social assistance. The benefit levels will be modeled differently for
these two categories. Average “Survivors” will be assumed to track changes in the
average old age pension level. Average “Orphan” pension will be linked to average
wage instead. This is because most survivor pensions are generated by deaths of old
age pensioners, while most orphan pensions are generated by deaths of active
contributors.

Simulation options: Enter the simulation options for PROST. PROST contains
multiple calculation methodologies to fit varying situations. Depending on the
options selected, PROST will present the user with a customized input file template.
Thus, the user may not see all of the ranges specified in this document. The following
options should be specified:

Line Contributors in the “Pension” sheet will be specified as: Depending on the
users’ needs and the data available, PROST allows users to specify present and future
expected contributors as a) a percentage of population in each age and gender
category, stock of population b) a percentage of the people employed in each age
and gender category, Stock of employment c) a finite number which may grow or
decrease, but is not linked to changes in population or employment, Nominal
number of positions (useful for evaluating closed pension systems where individuals
enter a pension system upon being hired and the hiring decision is based on human
resource policy as in some civil service systems and corporate pension plans) d)
current stock of contributors plus new contributors based on employment in each age
and gender category, Flow of employment and e) current stock of contributors plus
new contributors based on population in each age and gender category, flow of
population. We generally recommend using Stock of Population as it produces the
most stable results. While it might be considered logical to use Stock of
Employment since the contributors are part of the employed population, if there are
errors in the current or projected values of labor force participation or unemployment,
these errors will be transferred to the pension system contributors and magnified as
the system projects out for 75 years. If the user wishes to analyze closed pension
systems or civil service pension systems, it is recommended that the user select option
‘c’, the nominal number of contributors. In that case, the population linkages are in
effect cut off and PROST will calculate the future flow and stock of contributors
according to human resource policy specified in the table Human Resource Policy of
the spreadsheet ”General” when this option is chosen. Flow of employment and
flow of population are also two other options, but generally not recommended. They
both require the user to enter New Contributors in the Pension sheet as a percentage
of either employment or population. There is more likelihood of abnormal variation
in the base year for new contributors than there is for total contributors, and using this
methodology will magnify any such abnormalities over the long simulation horizon.
However, these are included in case of particular circumstances where this calculation
method is more appropriate.

Line Old age pensioners and disabled will be specified as: Old age pensioners
could be calculated as (a) a percentage of the population in each age and gender


                                         1-4
category, Stock of Population, or (b) last year’s retirees plus a percentage of last
year’s contributors who are now eligible for retirement, Flow of contributors
previous year. If you use Nominal number of positions in line 6, you must use
Flow of contributors from previous year. However, for national systems we
recommend Stock of Population. Frequently, people have accumulated rights to
retirement, are no longer contributing, but when they pass the age eligibility
requirement, will choose to collect benefits. Therefore, there really isn’t a tight link
between those who retire this year and those who contributed the previous year.

Line Survivors and Orphans will be specified as: Survivors and orphans could be
specified as (a) a percentage of the population in each age and gender category, Stock
of Age Specific Population, or (b) a percentage of the total covered group,
contributors, old age pensioners and disabled in a given year, Stock of Total
Contributors and Retirees. Survivors and orphans are not carried over from year to
year. Survivors in this context refer primarily to widows and widowers, although
other categories such as parents and siblings could be included. Orphans refer to
children of the contributors. Each year the total number of survivors and orphans is
calculated as specified and their benefits are calculated as will be specified below.
Generally, for national schemes, you would choose option a, Stock of Age Specific
Population. In the case of a closed group like civil servants, it is not possible to use
the population link. Therefore, it becomes necessary to assume a link between the
total number of participants in the system and the number of widows and orphans.

Line Summary and comparison should reflect finances of: the user can specify
either Monopillar System or Multipillar System. Depending on the choice summary
output and comparison tables will reflect either monopillar or multipillar pension
system setup.

Line User Name: Type an user name to identify the author of the input file.

Line Country Name: Enter the country name.

Line Case Name: Enter a name for the case to identify this input file uniquely from
others.

Line Base Year: The starting year for the simulation horizon. Typically the date
entered here is the year for which (i) all the other data are available (e.g., 2010) and
(ii) the data are stable since simulation projections would depend on the initial
conditions to a large extent.

Line Ending Year: The end point for the simulation horizon. This could be
anywhere from twenty five to one hundred years into the future starting from the base
year. Note that very long simulation horizons would require sophisticated
assumptions about the way macro economic variables are expected to behave in the
future. On the other hand, very short simulation horizons are unable to capture the full
lifetime of an individual just born during the Base Year.




                                        1-5
       Note: After entering the Base and Final year, these values will automatically
       appear in the date headings of all other variables. The user is not allowed to
       change the base year and end year headings in other data ranges by editing them
       manually.

Base Year Data

Enter base year information for the following variable blocks, paying careful attention to
units:

   Line GDP: GDP in millions of local currency for the base year.

   Line Pension Fund Balance: The pension fund balance in millions at the beginning
   of the base year.

   Lines Total lump-sum Payments to Male/Female Old Age Pensioners: Total
   annual lump-sum old age pension payments to males/ females during the base year in
   millions.

   Lines Total Annuity Pension Payments to Male/Female Disabled: Total annual
   annuity payments made to beneficiaries claiming disability pensions, males/females,
   during the base year in millions.

   Lines Total Lump Sum Pension Payments to Male/Female Disabled: Total
   annual lump sum payments made to beneficiaries claiming disability pensions,
   males/females, during the base year in millions.

   Lines Total Payments to Survivors: Total annual payments made to beneficiaries
   claiming survivor pensions, males and females together, during the base year in
   millions.

   Lines Number of Survivors/Orphans: These lines only appear when the option
   Stock of Total Contributors and Retirees is chosen under Simulation options for
   modeling survivors and orphans. Enter total number of survivors/orphans in
   thousands.

Wage and Pension Brackets and Cumulative Distributions

The variables in this block let you input wage and pension distributions. First, adjust the
number of rows in the distribution range according to the data that you have. For
example, if you have wage distribution data in quintiles, you need 5 rows for that range.
You can insert or delete rows by pressing on “I” or “D” buttons in the left-side corner of
the last row of the appropriate range. You have to have at least 3 rows in each of the
distribution ranges. The information in the first row should always pertain to minimum
wage workers or minimum pension recipients.

Also note, that PROST does not allow to model part time work. Therefore, both numbers
of people and wages have to be adjusted in the following manner. The number of


                                           1-6
contributors should be calculated as total number of days worked by all contributors
divided by the number of working days in a year. This will generally be a lower number
than that reflected by weekly, monthly or quarterly statistics. When number of working
days is not recorded, the user should try to get the data on a number of weeks or at least a
number of months worked, and perform a similar calculation. Then, average annual wage
could be calculated by dividing total yearly declared wage bill by the number of
contributors derived as described above. Average wage for specific age and gender
cohort or income bracket should be calculated the same way. Same logic should also be
applied to number of pensioners and average pension.

   Column C, Wage or Pension Bracket: Input the annual minimum wage in thousands
   in the first income group labeled “Earn Minimum Wage of”. The wage should be the
   annual minimum wage for full time worker. Enter the maximum wage earned by the
   people in the second income group. Proceed the same way with the other income
   groups and make sure that the bracket value in each line is higher than the previous
   one. The last category in your data most probably will include people with wages
   above certain threshold with no upper bracket. In this case, put what you think is a
   reasonable upper limit for the last category. Keep in mind that PROST assumes that
   people in each category are distributed linearly between lower and upper bracket.
   That is why it is unadvisable to put something close to infinity for the upper bracket
   for the last category. Rather, choose something reasonably high not worrying too
   much that you have not captured several richest people in the country. Repeat the
   same procedure to fill in the cells for the pension distribution.

   Columns D-E, % of Males; % of Females: Input the percentage of male or female
   active contributors that are earning minimum wage in the first income group. Enter
   the percentage of contributors earning the wages below the threshold specified in
   column C on the next line. Note that the distribution is cumulative, i.e. your
   percentage in the second income group should include minimum wage earners as
   well. Continue with the other lines. Make sure your entries are not decreasing with
   each subsequent line. Note, that the last row should specify 100%. You can put 100%
   in the line before last to effectively lower the number of categories. Note that you
   have to have at least one non-100% category, if you want to make use of earnings
   profile, specified in Labor sheet. If you know the average wage in the economy you
   can check whether your input sums up to the right average wage. The average wage
   derived from your input is displayed for you reference in the User Comments area,
   next to the table Wage Brackets and Cumulative Distributions. Fill in the required
   cells for the pension distribution the same way you did it for contributors, make sure
   that deferred or commuted pension amounts are also reflected in the distribution, even
   if they are not being paid.

Demographic Trends

Enter the following demographic data. It is possible to insert or delete as many years as
necessary between the base year and end year columns by using the “I” and “D” buttons
at the top of the end year column. Do not use EXCEL Insert Column functions for this
purpose. When you insert a column, interpolated values automatically appear. These can


                                            1-7
be overwritten. They appear only for the convenience of the user, particularly when the
user does not want to specify exogenous values for all the variables in a particular data
block.

   Line Sex Ratio at Birth: The sex ratio at birth is used to convert births for any
   period into male and female births. Enter the sex ratio in terms of male births per 100
   female births. A value of 102 implies 102 male births per 100 female births. The
   usual values for this variable range from 102 to 108 and values outside this range
   should be checked carefully for assumptions and consistency of the source. For users
   without access to this information it may be advisable to use the international average
   of 104.5 male births per 100 female births.

   Line Mortality Rate Multiplier for Disabled: These multipliers are used to adjust
   the mortality for this specific group of the population up or down by a specific factor.
   Thus a multiplier of 110 for disabled individuals would imply that the disabled
   population has a 10 percent higher risk of death compared to the overall population in
   each age and gender category.

   Line Mortality Rate Multiplier for Old Age Pensioners: This multiplier works
   exactly the same way like the one for disabled, described earlier. In the case where
   the covered population is different from the overall population, the covered
   population is usually higher income and more urbanized with better access to health
   care, and as a result will experience lower mortality rates. Using the population
   mortality rates will underestimate pension expenditures. These differences will
   mostly materialize after retirement, since non-accident related mortality before
   retirement is not very high for both covered and not covered and therefore is assumed
   to be the same. The number entered in this cell should be below 100 and based on the
   best available estimates of the mortality differential. As coverage expands and the
   covered group begins to resemble the overall population, the value for the End Year
   could be closer to 100 than that for the Base Year. If there is no reason to believe that
   specific groups have different mortality rates compared to the overall population, a
   value of 100 should be entered. Since PROST is extremely sensitive to mortality
   assumptions, it is highly recommended to use a multiplier of 100 for all groups unless
   convincing evidence exists proving otherwise. Also, it is important to remember that
   what is relevant for pension system (and PROST), is the number of reported, not
   actual deaths. The two could be substantially different in remote regions with poor
   reporting system.

Macroeconomic Growth Trends

The next few variables require information and assumptions about the macroeconomic
trends of the economy. If required, data for additional years could be inserted between the
base year and the end year by using the “I” and “D” buttons.

   Line Option 1: Real GDP Growth or Option 2: Real GDP per Capita Growth:
   First, the user should choose which option of data entry he or she wants to use by
   selecting from drop-down list in the column A. Most external GDP projections would


                                           1-8
   be specified in “Option 1” format, although choosing to input your assumptions in
   “Option 2” format would allow for easier coordination with productivity growth
   assumptions. The growth rates should be entered as a percentage. Thus an entry of
   10% would be interpreted by the program as a 10 percent rate of growth of real GDP
   under the option 1.

   Line Productivity Growth of Minimum Wage Worker: The average percentage
   increase in the real wage of minimum wage worker for the base year, the end year,
   and any years in between. Initially, the variable required here was average real wage
   growth. However, the age of the average worker increases as the population ages.
   Since the average 45 year old earns more than an average 25 year old, we wanted to
   separate real wage growth due to normal productivity growth from real wage growth
   arising from the aging of the population which is already included in the model. For a
   young and rapidly expanding work force, values for this variable could be about 2%
   points lower than the values assumed for real GDP growth rates. For more mature
   pension systems with a stable work force population as in some of the OECD
   countries, the values may be only 1% point lower than real GDP growth rates. The
   user is well advised to experiment with a set of values and observe how PROST
   predicts the ratio of the wage bill of contributors in relation to GDP.

   Line Inflation Rate: The rate of inflation for the base year, the end year, and any
   years in between. This information should be entered as a percentage. An entry of
   15% therefore implies a 15 percent level of inflation.

       Note: Macro-economic variables influence future projections tremendously. The
       user is advised to ensure that the data entered for future macro economic
       projections are internally consistent and in harmony with each other and variables
       entered in other ranges. In particular, the relationship between GDP growth rates,
       real interest rates and the rate of productivity growth of minimum wage worker
       should be very carefully analyzed before selecting a final simulation case.
       Further, the user is advised to experiment with various scenarios for the time
       paths of these macro-economic variables.

Line PAYG Pension Fund is Allowed to Borrow (at Government Bond Rate):
Whenever the PAYG pension fund runs a deficit and has no reserves there are two
financing possibilities. Pension fund can borrow money against future contributions, in
which case you should check the box. An alternative is to shift the deficit to the
government’s budget, which will be modeled if you leave the box unchecked. Usually the
laws of the country specify which of the two possibilities is implemented.

Interest Rates (Real)

   Line Interest Rate on PAYG Investments: The real rate of return earned on pension
   fund investments during the base year, the end year and any years in between. If the
   pension fund invests in more than one kind of asset, the average rate of return across
   all asset classes would be the appropriate number to enter. This interest rate may be
   different than prevailing rates in the economy, because pension funds frequently do


                                          1-9
   not earn market interest rates. They are often more heavily invested in risk-less assets
   which carry lower rates of return, are politically forced to invest in money-losing
   assets, or suffer from bad management.

       Note: If the pension system does not invest and holds all its reserves in cash, then
       the real interest rate entered in this field should be equal to  i /(1  i) where i is
       the inflation rate. This would ensure a zero percent nominal rate of return.

   Line Government Bond Rate: The real rate of return paid by the pension fund
   during the base year, the end year and any years in between in the case when the
   pension fund is forced to borrow to cover the deficits. Such borrowing is usually
   covered by government guarantee and therefore can enjoy the same interest rates as
   do government bonds.

   Line Discount Rate: The real discount rate during the base year, the end year, and
   any years in between. This information is used to calculate the present value of future
   cash flows to or from government or individual pension system participants. This
   assumption will tremendously influence the value of implicit pension debt.

Human Resource Policy

   Line Option 1: Net rate of increase of workforce or Option 2: Nominal number of
   new positions: This line only appears when option Nominal Number of Positions is
   chosen under Simulation Options for modeling contributors. First, the user should
   choose the option of how human resource policy will be characterized. The first
   option is often more appropriate when human resource policy is trying to target the
   total number of positions and the second is more appropriate when certain number of
   new hires is being targeted. The choice should be made by the user by selecting an
   option form drop-down box in column A.

   Under the first option the percentage is expressed as an increase in the number of
   positions as a percent of the previous year employment level. If a value of 0 is
   entered, the number of contributors will remain constant, with the number of new
   entrants exactly matching the number of contributors who died, retired, left, or
   become disabled last year. If the number is positive, then the total number of
   contributors increases by the specified percentage. If the number is negative, then the
   total number of contributors falls by the specified percentage. However, it should be
   noted that retrenchment periods for civil servants rarely last more than a few years.
   As the population grows, the need for teachers, health care workers, and others often
   included under civil servants also grows. Since the automatic link with population
   has now been broken, users should check the output to make sure that the assumption
   specified here provides an adequate number of workers in these population-sensitive
   fields.

Benefit Eligibility




                                            1-10
   Lines Male/Female Retirement Age: The statutory retirement age in the base year
   as well as the retirement age for the end year for males/females. This is usually the
   age specified by law at which individuals may retire without penalty. If the law does
   not specify such an age, enter the earliest age of retirement. If there is a change in the
   statutory retirement age during the simulation period, a column could be inserted
   between the base year and the end year with the new retirement age entered for that
   year. The toolkit linearly interpolates between two numerical points. Thus an entry of
   60 for the base year would mean that males retire at age 60 during the base year. An
   entry for 65 for an intermediate year would mean that the retirement age would be 65
   for that intermediate year, an increase of 5 years. Between base year and the year at
   which the retirement age changes, retirement age increases linearly to its new value
   spread evenly over the years. If no further increases in retirement age are envisaged
   by planners, the end year retirement age would also be 65. On the other hand, if there
   are no retirement age changes contemplated during the entire simulation period the
   end year retirement age should be identical to the base year retirement age of 60.

   Lines Percentage of Length of Service Not Claimed by Males/Females: Most
   PAYG systems have a minimum length of service requirement. People, who do not
   reach this limit are not eligible to get an annuity from the PAYG pension system and
   might be either compensated by lump sum payment, usually at big discount, or not
   get anything from the system at all. In case of the funded system such people would
   not be eligible for minimum pension guarantee and would only get their own savings
   back, again usually only as a lump sum. This parameter asks what percent of length
   of service accumulated by people retiring this particular year will be “lost” or “not
   claimed” towards PAYG annuity or minimum pension guarantee in the funded pillar.
   Given shorter average careers of women, this parameter should usually be higher for
   females.

   Lines Average duration of survivorship benefit: The user should input the number
   of years that survivorship benefit lasts on average. This is calculated for the whole
   survivor pool, lumping together both genders and all ages. This parameter will
   influence the speed at which changes in average old age pension levels are
   transmitted to average levels of survivor pensions. The longer the duration of
   survivorship benefit, the longer the lag with which changes in old age benefits are
   transmitted.

Line Automatically Adjust Assumptions Related to Retirement Age Change.
Changing retirement age will also affect other variables like labor force participation,
coverage rates, retirement rates etc. Since retirement age changes are a common pension
reform, PROST presents the user with an option to model some of these changes
automatically. To set the automatic option on check the box. The variables that will
automatically be changed in this case are number of employed population, number of
nominal contributors, number of disabled, number of pensioners and length of service at
retirement. Retirement age change may also affect other variables, which are not being
changed automatically (earnings profile, replacement rates, etc.). The user is responsible
to manually code these changes. If the user checks this automatic option off, then he/she
should manually change all the affected variables.


                                           1-11
Of People Willing, but not Allowed Normal Retirement After Reform

This range only appears if the checkbox Automatically Adjust Assumptions Related to
Retirement Age Change is checked in, i.e. if PROST is asked to make automatic
adjustments to certain variables. The following parameters are asking what will people
denied normal retirement do under the new circumstances. For example, if in the base
year retirement age was 60 and in particular future year it is set at 62, there will be a
group of people who would have been retired under the old rules, but now are refused an
option of normal retirement. Let us call them “people denied normal retirement”

   Line The Following Percent Will Become Employed, PROST will automatically
   adjust number of employed based on this parameter. The parameter specifies what
   percent of people denied normal retirement will be employed under the changed
   circumstances.

   Line The Following Percent Will Become Nominal Contributors, PROST will
   automatically adjust number of nominal contributors based on this parameter. The
   parameter specifies what percent of people denied normal retirement will be nominal
   contributors under the changed circumstances.

   Line The Following Percent Will Claim Early Retirement, PROST will
   automatically adjust number of old age pensioners based on this parameter. In some
   cases people denied normal retirement may have an option of early retirement
   (usually with reduced pension). The parameter on this line specifies what percent of
   people denied normal retirement option will still choose to retire under the early
   retirement rules.

   Line The Following Percent Will Claim Disability, PROST will automatically
   adjust number of disabled based on this parameter. The parameter specifies what
   percent of people denied normal retirement will claim disability under the changed
   circumstances. Note, that PROST assumes, that after these additional disabled people
   reach the new retirement age they convert to the normal retirement program, i.e. their
   claim for disability benefits is temporary.

Replacement Rate

   Lines For New Survivors as Percent of Average Old Age Pension: The average
   pension received by new recipients of survivor pension as a ratio of the average
   pension of old age pensioners for the base year, the end year and any years in
   between. This data is entered as a percent. Thus a value of 20% would imply that on
   average upon application, survivors get about 20% of the value of the average old age
   pensioner pension.

   Lines For Orphans as Percent of Average Wage: The average pension received by
   orphans as a ratio of the average wage of contributors for the base year, the end year
   and any years in between. This data is entered as a percent. Thus a value of 20%
   would imply that on average, orphans get about 20% of the value of the average wage
   of contributors.


                                          1-12
Parameters specified below can be changed after the reform: The following variables
will appear on the Reform Sheet as well as here. Therefore, the following entries on this
sheet will not affect output of Module 4.

Contributions

The variables in this block allow the user to specify contribution rates applied to
employees, employers, pensioners and government depending on the income bracket of
the contributor. The differentiation by income brackets may be useful in modeling a few
different progressive social security policies: high income employees might be asked to
contribute higher share of their wage, low income contributors might be subsidized by the
government, high income pensioners may be taxed. Finally and most commonly,
contribution ceiling might be introduced, above which wages are not taxed by social
security contribution. PROST requires the user to specify 4 income categories that can
potentially be treated differently with respect to contribution rates. The 4 income brackets
should be non-decreasing, with the last bracket in column B specifying overall
contribution ceiling.

   Columns Contribution from employees: The proportion of gross covered wages
   mandated to be paid to the pension system for the base year, end year or any
   intermediate years. The data are entered in percentages. Any planned changes in the
   contribution rate could be accommodated by inserting columns between the base year
   and the end year using the “I” and “D” buttons. Typical values range from about 5
   percent to about 45 percent though lower values have also been empirically observed.
   Higher contribution rates encourage evasion and other labor market distortions while
   too low contribution rates would choke the revenues of the pension system. Ideally,
   the data entered for this field should include contributions only to the pension system.
   If the pension system collects overall payroll taxes, without separating components,
   use total contribution rate, but include non-pension expense as Other Pension Fund
   Expense as % of Pension Expense under Costs and other Expenditures in
   “General” sheet.

   Columns Contribution from employers: The proportion of gross covered wages
   mandated to be paid to the pension system for the base year, end year or any
   intermediate years. The data are entered in percentages. Any planned changes in the
   contribution rate could be accommodated by inserting columns between the base year
   and the end year using the “I” and “D” buttons. Typical values range from about 5
   percent to about 45 percent though lower values have also been empirically observed.
   Higher contribution rates encourage evasion and other labor market distortions while
   too low contribution rates would choke the revenues of the pension system. Ideally,
   the data entered for this field should include contributions only to the pension system.
   If the pension system collects overall payroll taxes, without separating components,
   use total contribution rate, but include non-pension expense as Other Pension Fund
   Expense as % of Pension Expense under Costs and other Expenditures in
   “General” sheet.




                                           1-13
   Columns Contribution from Pensioners: Enter the contribution rate (as a
   percentage of gross pensions) of pensioners if they are required to contribute social
   security taxes which accrue to the pension fund. Thus an entry of 5 percent would
   imply that pensioners, on average, are required to transfer about 5% of their pensions
   as payroll tax contributions for social security.

   Columns Contribution from the Government for Employees: If the government
   contributes a certain portion of the payroll tax for pensions on behalf of the employee
   in a tripartite contribution system, that data should be entered in this field. Data are
   required for the base year, the end year and any other years in between if necessary. If
   no such contributions are made on behalf of contributors by the Government, the user
   can leave the field blank.

Indexation of Contribution Steps

Enter data regarding indexation rules of minimum wage, wage brackets and contribution
ceiling. Many pension systems index these parameters either to nominal wages or
inflation (or some combination of both). If the level of inflation is I, and rate of nominal
wage growth is G, then the indexation mechanism is described by [(1+aI)*(1+bG)] where
a and b are the user input inflation indexation parameters. Data are needed for the base
year as well as for the end year (and intervening years if available).

   Lines Minimum Wage Indexation to Inflation / Nominal Wage Growth: The data
   are required for both the base year as well as the end year. A value of 100% implies
   that minimum wage is fully indexed to a given parameter (inflation or nominal
   average wage growth). This indexation is applied to the minimum wage initially
   specified on line 30 of “General” input sheet. Depending on country circumstances
   values normally range from 0% (no indexation) to 100% (full indexation).
   Occasionally, when minimum wage is increased substantially by decree, the
   appropriate parameter for this field might be above 100%. In this case the user should
   manually unprotect the sheet and change the range of values that are allowed for this
   cell.

   Lines Step x to Inflation / Nominal Wage Growth: The data in these fields should
   be entered the same way as data for minimum wage indexation. This indexation is
   applied to the contribution ceiling initially specified in the table Contributions of
   “General” input sheet.

   Line Contribution Ceiling Applies to Employers: Options “Yes” and “No” are
   available. In the case when “Yes” is chosen the wage base for employer’s
   contribution is the same as that for employee’s contribution. Under the second option
   employer’s contribution is calculated on the whole employee’s salary, even if the
   ceiling applies for the employee’s portion. The “insured wage”, used when pension is
   calculated is still limited by the ceiling.

Revenue sources




                                           1-14
   Lines Collection Rate from Employers and Employees/ Pensioners/ Government:
   This variable attempts to capture the efficiency of the pension administration in
   collecting pension revenues due from its contributors. It is defined as a ratio of total
   collected and due payroll contributions. The amount of contributions due is computed
   including both employer and employee shares. However, it excludes contributors who
   may be officially fully or partially exempt and also excludes the persons who hide
   from the system completely, by not paying contributions and not collecting future
   benefits either. Revenue loss from under-declared wages also should not be addressed
   through this parameter. The data are entered as a percentage for the base year, the end
   year and intermediate years if necessary. Typical values could range from 25%
   (meaning that the pension system is only able to collect about 25 percent of the cash
   revenue due to it) to 100% (meaning that the pension system is efficient enough to
   collect 100 percent of all the revenues due to it).

   Line Unconditional Budget Transfers to Pension Fund as % of GDP: This field
   records unconditional budget transfers to the pension fund (other than the wage-bill
   based tripartite Government contributions) made by the Government as a percent of
   GDP. For example, some countries earmark x percentage of the VAT tax to the
   pension fund. Other countries earmark privatization revenues for supporting the
   pension system. Neither of these sources of revenue are directly tied to the covered
   wage bill, contributions from which are the primary source of revenue. Please note
   that this line does not include transfers from the non-pension Government budget to
   cover pension system deficits, hence the term Unconditional, but only items that are
   earmarked regardless of the status of the pension system. Data are required for the
   base year, the end year and any other years in between.

   Line Other Income as % of Employee, Employer & Pensioner Contributions:
   Enter any other income received by the pension fund as a percent of total
   contributions. Thus an entry of 10% would imply that other income sources of the
   pension fund were able to add to contribution revenues by 10%. This income will be
   tied to the wage bill since contributions are tied to the wage bill. This line is useful if
   the pension fund engages in non-pension activities and the user would like to look at
   the overall budget of the pension fund. However, the non-pension expenditures are
   specified in a highly aggregated fashion, leading to only crude estimates of the impact
   on the overall budget of the pension fund.

Costs and Other Expenditures

   Line Admin. Costs as % of Employee, Employer & Pensioner Contributions:
   The transaction or administrative costs of running the pension system include costs of
   setting up the pension registry, costs of book keeping, and administering the pension
   system. The data are needed for both the base year as well as the end year and are
   entered as a percent of total contributions from employers, employees, and
   pensioners. Typical values range from about 1% meaning that almost 1% of
   contributions are being used to administer the pension system to as much as 50% in
   poorly run systems. In PAYG systems, administrative costs are often expressed as a
   percentage of total expenses. However, to insure comparability with defined


                                            1-15
   contribution reforms, administrative costs need to be entered as a percentage of total
   contributions. Please keep in mind that if contribution revenues are modeled to rise
   for some reason (rising contribution rates, increased contribution ceiling, coverage
   expansion) this link will translate into rising administrative costs. The user should be
   careful to adjust this parameter if such linkage is not intended.

   Line Assets Management Costs as a % of Assets: The costs required to manage the
   assets of the pension system. The data are required for both the base year as well as
   the end year and entered as a percent. These costs are usually a function of the
   volume of financial assets being managed by the pension system. Typical values
   range from 1 percent (meaning 1 percent of the previous years fund balances were
   used for asset management) to about 6 percent (meaning 6 percent of the pension
   funds assets are used for asset management expenses). If fund management costs are
   not separable from other administrative costs, they can be lumped together in line
   111. It is assumed that these costs are 0 if the reserve is 0 or negative.

   Line Other Pension Fund Expense as % of Pension Expense: The percent of total
   expenditures attributed to non pensions related expenditure or “other” expenditures.
   This data must be entered as a percent for the base year, the end year and any other
   years in between. If the pension fund is engaging in health care or unemployment, for
   example, these expenses could be entered here. The crude assumption here is that
   these expenditures will vary directly with the pension expenditures of the pension
   fund. These may or may not be correlated in reality, which is why we generally
   recommend only looking at pension expenditures and revenues with PROST,
   although information about general orders of magnitude can be useful. Alternatively,
   detailed projections for the other expenses using other models could be inserted in
   this space, providing accurate estimates on the overall budget picture.

Commutation

This range of parameters allows modeling commuted and deferred pension payments. In
the case of pension commutation a portion of the benefit is commuted, or taken lump sum
at the time of retirement by applying commutation factor, defined in the input sheet
“Factors”. The remaining reduced benefit is paid until certain age at which time the
commuted portion of the benefit is restored. Deferred pension is a special case of
commuted pension, where commutation factor is equal to “0” and 100% of pension is
commuted until the regular retirement age, at which point the payment of full pension
commences.

   Lines Average Portion of Benefit that is Commuted: The user should input a
   percentage of the benefit that will be converted into the lump sum payment at the
   time of retirement. For example, if commuted portion is 30%, then 30% of the
   pension will be multiplied by commutation factor and the total amount will be paid to
   the beneficiary. At the same time a regular monthly pension will be reduced to 70%
   of the original amount which can again be restored to 100% only after the beneficiary
   reaches pension restoration age.



                                          1-16
   Line Term of Commuted Annuity: As mentioned above, when calculating
   commutation lump sums PROST uses commutation factors entered in the input sheet
   “Factors”. In addition, PROST computes and outputs actuarially fair commutation
   factors for all ages and time periods using corresponding mortality rates,
   commutation indexation rules, and the term by which commuted part of annuity
   payment is deferred. The term of commuted annuity is specified in years.

   Lines Average Age at which Commuted Payments are restored for
   Males/Females: PROST only allows to model deferral to the specific age as opposed
   to deferring pension payments for a specific number of years after retirement. The
   user should input age at which payment of pension calculated at retirement
   commences. It is assumed that calculated pension is indexed the same way as
   pensions already in payment are. Commuted portion of the pension can be indexed
   differently.

Pension Indexation

Enter data regarding indexation rules. Many pension systems index minimum pension
and regular pensions to either nominal wages or inflation (or some combination of both).
If the level of inflation is I, and rate of nominal wage growth is G, then the indexation
mechanism is described by [(1+aI)*(1+bG)] where a and b are the user input inflation
indexation parameters. Data are needed for the base year as well as for the end year (and
intervening years if available).

   Lines Full Pension Indexation to Inflation/Nominal Wage Growth: The data
   entered for this variable reflects the indexation of old age and disability pension
   payments to inflation/nominal wage growth. The data are required for both the base
   year as well as the end year. A value of 100% implies that pension payments are fully
   indexed to inflation/wage growth. Depending on country circumstances values could
   range from 0 (no indexation) to 100% (full indexation).

       Note: Pension indexation to inflation is cheaper than nominal wage indexation
       when real wage is growing, since nominal wage indexation by definition, indexes
       to real wages as well as inflation)

   Lines Indexation of Commuted Pension Portion to Inflation/Nominal Wage
   Growth: Commuted pension portion might be indexed differently than overall
   pension. For example, commuted pension portion might not be indexed at all, while
   overall pension might be indexed to inflation. Then if the initial pension is equal to
   $100, 30% of it is commuted and inflation is equal to 5%, then annuity pension
   payment the next year will be $100*1.05-$30 = $75.

   Lines Minimum Pension Indexation to Inflation / Nominal Wage Growth: The
   data in these fields should be entered the same way as data for pension indexation.
   This indexation is applied to the minimum pension initially specified under Pension
   Brackets and Cumulative Distributions of “General” input sheet. After the old age
   pension is calculated, whether at the time of awarding the pension or at the time of



                                          1-17
   indexation, it is checked if the resulting payment does not fall below minimum
   pension. If it does, the payment of minimum pension is made to the pensioner instead.

   Lines Maximum Pension Indexation to Inflation / Nominal Wage Growth: The
   data in these fields should be entered the same way as data for pension indexation.
   This indexation is applied to the maximum pension initially specified under Pension
   Brackets and Cumulative Distributions of “General” input sheet. After the old age
   is calculated, whether at the time of awarding the pension or at the time of indexation,
   it is checked if the resulting payment does not fall above the maximum pension. If it
   does, the payment of maximum pension is made to the pensioner instead.

Line Use Benefit Formula for the Earnings Related Part of the Pension for Old Age
Pensioners: By checking this box, the user will be presented with benefit formula range
requiring the input of benefit formula parameters for the earnings related part of the
pension. These inputs are described under the benefit formula parameters below. If the
pension does not depend on individual’s earnings, leave the box unchecked. The user will
be given an opportunity to specify earnings unrelated component and lump sum
component of the pension in the “Pension” input sheet. If the earnings related benefit
formula is very complicated it is sometimes easier to approximate it by earnings
unrelated component in the “Pension” sheet rather than do it here. The final total benefits
will be calculated as a sum of 3 components: earning related annuity, earnings unrelated
annuity and lump sum component.

Benefit Formula Parameters for Old Age Pensions

This range only appears if the checkbox Use Benefit Formula for the Earnings Related
Part of the Pension for Old Age Pensioners is checked in.

   Line Required Years of Service for Basic Replacement Rate: Minimum years of
   contribution required to qualify for old age pensions, male/female. However, the
   program does not disqualify individuals who do not satisfy this minimum
   requirement. Such an individual is awarded the minimum pension.

   Lines Basic Replacement Rate: Percent of individual’s wage base awarded as
   pensions after fulfilling minimum contribution period, male/female.

   Lines Incremental Replacement Rate: Percent of individual’s wage base added to
   the basic replacement rate for every year of contribution in excess of the basic
   minimum, male/female.

   Lines Maximum Replacement Rate: The highest allowed percent of individual’s
   average wage base to be awarded as pension, male/female.

   Lines Years in Final Average Wage: Number of years of an individual’s work
   history included in the wage base used to calculate the pension, male/female. Many
   countries actually specify a wage base such as the best 3 of the last 10 years, or the
   best consecutive 3 of the last 10 years. These distinctions are specific to the



                                           1-18
individual and cannot be easily projected by the model. Instead the model assumes
that in these cases the last 3 years were the best 3 or the best consecutive 3.

Lines Wages are Valorized to Inflation/ Nominal Wage Growth: The valorization
of the wages included in the wage base to inflation/ wage growth, male/female.
Entering 100 implies complete valorization. Note that this differs from the indexation
of the pension.




                                      1-19
WORKSHEET II: POPULATION

The population worksheet contains data vital for making population projections. Like the
previously described worksheet, data are required for the base year as well as for future
years. Data for years falling between the base year and the end year, if available, are
entered by inserting a column between the base year and the end year by using the “I”
and “D” buttons where appropriate.

The spreadsheet is organized a little differently with the variables in columns and ages
represented in rows. The template contains ages from 0 to 100. If data is available in
single age cohorts for ages beyond 100, the “I” button can be used to extend the age
ranges further. If fewer years are available, the “D” button can be used to decrease the
age ranges. One caveat is that all age-specific variables must be entered in the same age
ranges. If all the age-specific variables but one are available until age 85, but that one is
only available until age 75, the choice would have to be made of either truncating all the
variables at 75 or approximating the behavior of this variable between 75 and 85.

   Columns Population Males/Females: The base year male population in thousands.
   Data are required in single age cohorts. Note that age 0 refers to births and age 100 is
   the maximum age for pension simulations. The model will project the population in
   future years, so only the base year values are required.

   Columns Fertility Rate: Age specific fertility rates for the base year as well as any
   projections or theories about the way fertility rates are most likely to behave over the
   simulation horizon. This is an important consideration since fertility rate changes
   have long lasting effects on the flow of population which in turn directly affects the
   pension system. The data are to be entered as a percent. Thus a fertility rate of 10%
   would imply that 1 in 10 women of a particular age group would bear a child in that
   year. According to the World Bank estimates most countries would have a total
   fertility rate of about 2 children per woman by the year 2030 which would translate to
   a sum of fertility percentages by age of 200%.

   Columns Probability of dying males/females: Age specific probability of dying for
   males/females in the base year as well as any projections or theories regarding the
   most likely behavior of probability of dying in the future. Data are entered in percent
   terms. Thus a probability of dying of 1% for a particular age group would imply that
   1 out of 100 individuals in that age group will die in that year. In most cases, due to
   economic growth and improving health conditions, probability of dying is projected
   to decline in the future. A rough rule of thumb is that life expectancy will increase by
   1 year every 10 years in countries with lower life expectancy today and by 1 year
   every 15 years in countries with the highest life expectancy today. Running the Life
   Expectancy tool in Module 1 will produce projections on life expectancy. The user
   may need to initially provide some estimates and then review the output to determine
   whether the decrease in mortality rates have been sufficient to generate the normal
   life expectancy increases. Mortality rate changes are also not uniform across age
   groups. Children and the elderly generally face the greatest improvements. We hope
   to provide more guidance to users in this area in the future. However, it should be


                                            1-20
noted that in some countries, particularly those afflicted with widespread AIDS or
other serious health hazards, mortality may temporarily go up. Thus the user should
build in alternative assumptions regarding mortality trends over time. It is also
important to note that PROST assumes the last age cohort dying with certainty in the
next year, i.e. regardless of the input for the last cohort mortality, PROST will
overwrite the number with 100%.

Columns Net immigration males/females: Age specific net immigration (net
inflows of population) in each age group. Data are entered in thousands. Data are
required for the base year as well as for future years. If immigration represents a very
marginal and insignificant flow of population, then entering 0 for all the age groups in
effect disables the effect of net immigration on population flows and hence in the
other pension system calculations. Note that these are net values, with negative
values representing outflow of people to other countries.




                                       1-21
WORKSHEET III: LABOR

This input worksheet contains data pertinent to the labor market conditions of the
economy or region under analysis. Both base year as well as end year data are required
for the variables in this worksheet. The end year typically would be a projection of the
base year value depending on the user’s hypothesis about the way the base year
parameters are expected to behave. Note that labor market data are often difficult to
obtain and interpret and the user should ascertain that the data sources being used for this
worksheet are consistent and accurate. If under Simulation Options of “General” sheet,
the option Stock of Population is chosen for modeling contributors and Stock of
Population is chosen for modeling retirees, then the data entered for labor force
participation and unemployment will not influence the projections of the pension fund
finances. Also, only columns Earning Profile in Terms of Min Wage and Pension
Profile in Terms of Min Pension appear when the option Nominal Number of
Positions is chosen for modeling contributors under Simulation Options of the
“General” sheet.

   Columns Labor force participation males/females:                The proportion of
   males/females who are either working or looking for work in each age group. The
   data are entered as a percent. Thus an entry of 12 would mean that approximately 12
   percent of individuals in that particular age group are in the labor force. This data
   should not reflect expected changes in labor force participation due to the planned
   retirement age increase, if the box Automatically Adjust Assumptions Related to
   Retirement Age Change (sheet “General”) is checked, i.e. if PROST is asked to
   automatically adjust numbers of employed population. However, all other anticipated
   changes should be incorporated.

   Columns Unemployment rate males/females: The proportion of males/females in
   the labor force who are unemployed in each age group. The data are entered as a
   percent. Data are required for the base year as well as for the end year as well as any
   year in between if desired. This data should not reflect expected changes in
   unemployment rate due to the planed retirement age increase, if the box
   Automatically Adjust Assumptions Related to Retirement Age Change (sheet
   “General”) is checked, i.e. if PROST is asked to automatically adjust numbers of
   employed population. However, all other anticipated changes should be incorporated.

   Columns Earning Profile for males/females in Terms of Min. Wage: The data in
   this column reflect the average gross wages of individuals relative to the minimum
   wage. Thus an entry of 110% for any age group would mean that this age group on
   the average earned about 10 percent higher wages compared to minimum wage. The
   minimum value for these cells is 100%. Data are required for the base year as well as
   the end year and any years in between if desired. In case the earnings profile data
   conflicts with the wage distribution data entered in “General” sheet, under Wage
   Brackets and Cumulative Distributions, the earnings profile will be automatically
   depressed or enhanced to match the data on the “General” input sheet. The earnings
   above minimum wage for all cohorts will be adjusted by the same percentage. In
   terms of total earnings low wage earners will be adjusted less.


                                           1-22
Columns, Pension Profile for Old Age Pensioners/Disabled in Terms of Min.
Pension: While this is a pension system parameter, it is symmetric with the wage
profile and therefore it is entered here. It represents the initial distribution of
pensions across pensioners of different ages, both for old age pensioners and the
disabled. This variable is important because unless the pensions are fully indexed to
nominal wage growth, the older pensioners will have lower pensions than younger
pensioners. Reductions in pension expenditures for pensioners who died over the year
need to take into account that those most likely to die have lower pensions. It may
also be the case that the pensioners retiring at older ages on average retire with
lower/higher pensions than those who retire being younger, which may be the other
reason why there is a correlation between pensioner’s age and his/her pension even in
the presence of nominal wage indexation of pensions. The profile for future years will
be internally generated by the model.




                                      1-23
WORKSHEET IV: PENSION

The Pension worksheet contains age specific data describing the pension system
including numbers of contributors and beneficiaries, their contribution and retirement
patterns and replacement rates.

   Pension system in base year

   Columns Contributors males/females: For each age cohort, enter the total number
   of male/female contributors gross of exemptions. This implies that all those who will
   receive contribution credit for the base year should be included regardless of whether
   or not they actually contributed. The discrepancy can be legal, such as mothers of
   young children who receive contribution credit while caring for the children, or
   illegal, through evasion and fraud. The data are required only for the base year and
   are entered in thousands. Thus an entry of 10 would mean that there are 10,000
   nominal contributors in the age cohort being considered. For more discussion of how
   a number of contributors/pensioners is defined please refer to discussion of Wage
   and Pension Brackets and Cumulative Distributions for the “General” sheet,
   especially the discussion on accounting for part time work.

   Columns Old Age Pensioners males/females: For each age cohort, enter the total
   number of old age pensioners (separate for males/females). The data are required only
   for the base year and are entered in thousands. Thus an entry of 5 for the age cohort
   60, males, would indicate that there are 5,000 male retirees in the population of males
   aged 60. Also, for more discussion of how a number of contributors/pensioners is
   defined please refer to discussion of Wage and Pension Brackets and Cumulative
   Distributions for the “General” sheet, especially the discussion on accounting for
   part time work.

   Columns Disabled males/females: For each age cohort and gender, enter the total
   number of disabled individuals for the base year in thousands. Also, for more
   discussion of how a number of contributors/pensioners is defined please refer to
   discussion of Wage and Pension Brackets and Cumulative Distributions for the
   “General” sheet, especially the discussion on accounting for part time work.

   Columns Survivors males/females: For each age cohort and gender, enter the total
   number of individuals receiving survivor/widow benefits in the base year in
   thousands. If gender specific information is not available, or most of the benefits are
   accruing to women (as is often the case), it is acceptable to enter the data for women
   only. Enter zeros for males. Some countries record the gender of the person who
   earned the benefit, not the gender of the person receiving the benefit. If your data
   show more male survivors than female survivors, it is worth double-checking the
   data. This input is only required if an option “Stock of Age Specific Population” is
   chosen for modeling survivors and orphans under Simulation Options of “General”
   sheet.




                                          1-24
   Columns Orphans, males/females: For each age cohort and gender, enter the total
   number of individuals receiving orphanage benefits from the pension fund in the base
   year. Data are entered in thousands. If no gender specific information is available on
   orphans, data could be entered for any one of the two genders. Fill up the other
   gender information with zeros. This input is only required if an option “Stock of Age
   Specific Population” is chosen for modeling survivors and orphans under
   Simulation Options of “General” sheet.



Projection parameters

   Columns Length of Service at Retirement: Average length of service for the new
   retiree for each age and gender cohort. Individuals who retire at older ages frequently
   have the same or fewer years of service than those who retire younger due to factors
   such as extended periods out of the labor force and later entry to the labor force.
   Many countries also reduce the service requirement as workers get older. This data
   should not reflect expected changes in length of service due to the planed retirement
   age increase, if the box Automatically Adjust Assumptions Related to Retirement
   Age Change (sheet “General”) is checked, i.e. if PROST is asked to automatically
   adjust Length of Service data. However, all other anticipated changes should be
   incorporated.

   The heading for the columns following Length of Service at Retirement columns
   and the required data inputs depend on the option chosen under Simulation Options
   of “General” sheet. This data should not reflect expected changes in numbers of
   contributors due to the planed retirement age increase, if the box Automatically
   Adjust Assumptions Related to Retirement Age Change (sheet “General”) is
   checked, i.e. if PROST is asked to automatically adjust numbers of contributors.
   However, all other anticipated changes should be incorporated.

      Contributors as % of Population, males/females: Appears only if the option
      Stock of Population is chosen for modeling contributors under Simulation
      Options of “General” sheet; refers to the proportion of the population that are
      accruing pension rights that year. The data entered are in numeric percentages.
      Thus an entry of 60 for any age cohort and gender would imply that 60 percent of
      that age cohort and gender are accruing pension rights in that year. The data for
      the base year are calculated from base year information on population and base
      year information on the number of nominal contributors, by dividing the number
      of nominal contributors in any age and gender by the population for that age and
      gender. The formulas have already been entered. For future years, the coverage
      rates could either change according to user assumptions or remain the same.

      Contributors as % of All Employed, males/females: Appears only if the option
      Stock of Employment is chosen for modeling contributors under Simulation
      Options of “General” sheet. Refers to the proportion of the employed that are
      accruing pension rights that year. The data entered are in numeric percentages.


                                          1-25
Thus an entry of 60 for any age cohort and gender would imply that 60 percent of
the employed in that age cohort and gender are accruing pension rights in that
year. The employed totals for each age and gender cohort should be derived by
applying the Labor Force Participation Rate from the Labor Sheet to the
Population in the Population Sheet and then subtracting out the unemployed,
which are derived by applying the Unemployment Rate in the Labor Sheet to the
labor force participants. If there are inaccuracies in either the labor force
participation rate or the unemployment rate, these will be reflected and magnified
in the contributor calculation which is why we prefer not using this methodology.
For future years, the coverage rates could either change according to user
assumptions or remain the same.

Distribution of New Contributors, male/female: Appears only if the option
Nominal number of positions is chosen for modeling contributors under
Simulation Options of “General” sheet. Refers to the age and gender distribution
of net new entrants to the system. This option generally applies to civil servant
schemes. The number of new entrants is determined by the data from Human
Resource Policy table in “General” sheet. The age and gender distribution of
these new entrants needs to be entered here. Thus, 10% under a particular age
and gender group means that 10% of the new entrants came from that age and
gender group. These can be derived by using the distribution of new entrants
from the latest year for which data are available, or if that year is considered
abnormal, from an average of the distribution for the past several years. The
distribution may also change in the future.

New Contributors as % of all Employed, males/females: Appears only if the
option Flow of employment is chosen for modeling contributors under
Simulation Options of “General” sheet. Refers to the proportion of total
employed that are new contributors to the system. The data are entered in
numeric percentages. Employment is calculated as explained in the paragraph on
Contributors as % of All Employed above. These data can be derived using the
numbers of new contributors in the latest year. However, since one year of data
may include anomalies, which will then be magnified over the projection period,
if employment based projections are required, the Stock approach will tend to
produce more reliable results.

New Contributors as % of Population, males/females: Appears only if the
option Flow of population is chosen for modeling contributors under Simulation
Options of “General” sheet. Refers to the proportion of total population in each
age and gender group that are new contributors to the system. The data are
entered in numeric percentages. These data can be derived using the numbers of
new contributors in the latest year. However, since one year of data may include
anomalies, which will then be magnified over the projection period, if population
based projections are being used, the Stock approach will tend to produce more
reliable results.




                                  1-26
The headings of the following columns also change depending on the option chosen
under Simulation Options of “General” sheet.

   Old Age Pensioners as % of Population, males/females: Appears when Stock
   of Population is chosen for modeling pensioners and disabled under Simulation
   Options of “General” sheet, and refers to the proportion of the population that
   have retired and are receiving old age pensions. The data are entered in terms of
   numeric percentages and are required for both the base year as well as for the end
   year (and any years in between if desired). The data for the base year are
   calculated from base year information on population and the number of old age
   beneficiaries by dividing the number of old age pensioners in any age cohort and
   gender by the population of that age cohort and gender. In terms of calculations,
   the program calculates total retirees in each future year using the percentages
   specified here, and then derives the new retirees, taking into account the surviving
   retirees from the previous year, and finally does a consistency check with the
   coverage numbers. If more contribution years of service have been earned by a
   cohort than is retiring from that cohort, the new retirement is increased, keeping
   the age pattern specified here constant.

   Note: The retirement rate parameter is closely linked to the retirement age, in the
   sense that if there is a change in retirement age, the retirement rate parameter
   would also have to be adjusted to reflect that change. There is a button in
   spreadsheet “General” which does this automatically.

   New Old Age Pensioners as % of last year’s contributors, males/females:
   Appears when Flow of Contributors previous year is chosen for modeling
   pensioners and disabled under Simulation Options of “General” sheet. Instead
   of looking at the stock of old age pensioners calculated above, these columns now
   focus on new old age retirees and tie the new old age pensioners to the
   contributors to the system the previous year. Thus, a value of 20 for 60 year old
   males implies that 20% of surviving 60 year old males that were contributing the
   previous year when they were 59 years old are now retiring. This number can not
   exceed 100% for any age category which does not allow modeling a situation
   where individuals who were not contributing the year before are retiring. The
   required input can be derived using the new old age pensioners from the latest
   year. But again, new old age pensioners in a single year are much more sensitive
   to peculiar market conditions, expectations of future reforms, etc. Care needs to
   be taken in projecting future behavior on the basis of the current numbers.

   Disability as % of population males/females: Appears when Stock of
   Population is chosen for modeling pensioners and disabled under Simulation
   Options of “General” sheet. This variable refers to the proportion (based on
   population) of disabled individuals in any age and gender cohort. Data are entered
   in numeric percentage terms. Thus an entry of 10% would imply that 10 in a
   hundred individuals from that particular age/gender cohort are disabled. The data
   for the base year are calculated on the basis of information from population and
   the base year age/gender specific stock of disabled individuals by dividing the


                                      1-27
   stock of disabled individuals in any age and gender cohort by the population for
   that age gender cohort. For future years (including the end year) the rates could be
   made to grow according to some hypothesis or assumed to remain constant.

   New Disabled as % of Last year’s Contributors, males/females: Appears
   when Flow of Contributors previous year is chosen for modeling pensioners
   and disabled under Simulation Options of “General” sheet. The proportion of
   the newly disabled as a percentage of the surviving contributors from the last year
   in each age and gender category is entered here. Thus, a value of 2% for 50 year
   old males implies that 2% of 50 year old males who contributed last year at the
   age of 49 become disabled this year. This number can be derived using the newly
   disabled pensioners from the latest year. But again, new disability numbers in
   any given year may reflect anomalies peculiar to that year. Care needs to be taken
   in projecting future behavior on the basis of the current numbers.

Columns Survivors as % of population, males/females: Appears only when Stock
of Age Specific Population is chosen for modeling survivors and orphans under
Simulation Options of “General” sheet. The proportion of the population in any age
and gender cohort receiving survivor pensions. Data are entered in numeric
percentage terms. Thus an entry of 5 would imply that 5 in a hundred individuals
from any particular age gender cohort are receiving survivor pensions. The data for
the base year are calculated by dividing the stock of survivor pensioners for any age
gender cohort by the population of that age gender cohort.

Columns Orphans as % of population, males/females: Appears only when Stock
of Age Specific Population is chosen for modeling survivors and orphans under
Simulation Options of “General” sheet. The proportion of the population in any age
and gender cohort receiving orphan pensions. Data need to be entered in numeric
percentages. Thus an entry of 1 would mean that 1 in a hundred individuals in that
particular age and gender cohort are receiving orphan pensions. The data for the base
year are calculated by dividing the stock of orphan pensioners for any age and gender
cohort by the population for that age and gender cohort.

Columns Exemption males/females: Some individuals may be legally exempt from
payroll taxes (like judges, priests etc…), but are eligible to collect pensions. This
variable captures the probability that an individual from a certain age group would
fall in one of these categories. There are others who evade completely and never
contribute and never collect benefits. They are not included among the contributors.
They may perhaps be included in labor force statistics and employment statistics, but
do not enter into the pension system as either contributors or beneficiaries. The data
are entered in percent terms. Note that like the retirement rate, this parameter is
linked to retirement age as well as to coverage. Thus if the retirement age changes,
the exemption rate may also change that year since some individuals who were
previously exempt would not be exempt anymore.

Columns Replacement Rate for New Old Age Pensioners, Males/Females: The
replacement rate applicable for newly retired pensioners in the base year and future


                                      1-28
   years. This is an earnings unrelated part of pension and is payable in addition to the
   earnings related part, specified under Benefit Formula Parameters in “General”
   sheet, and the lump sum component, described later. This variable is driven largely by
   the rules of the pension system regarding eligibility and therefore depends on age at
   retirement, years of service, and the specific accrual rates of the pension system.
   However, it is calculated by dividing the average actual pension received by a new
   pensioner of a specified age and gender by the economy wide average wage and
   entered as a percent. These data are used to calculate pension payments to new
   pensioners by multiplying the data entered here by the overall economy wide average
   wage. If the pension system does not have an earnings unrelated component, the user
   can leave these cells blank. However, in some cases when the earnings related
   component is very complicated it can be approximated by average replacement rates
   in terms of average economy wide wage and entered here instead of benefit formula
   in the “General” sheet. Inaccuracies in the data and complex benefit formulas with
   special rules for special groups will sometimes yield different results, when modeled
   different way.

   Columns Lump Sum Payments for New Old Age Pensioners, Males/Females:
   Average lump sum payment divided by average economy wide wage and expressed
   as apercentage, payable only once at the time of retirement. This is a third component
   of the pension package in addition to earnings related component, under Benefit
   Formula Parameters in “General” sheet and earnings unrelated component,
   described immediately above.

   Columns Replacement Rate for New Disabled, Males/Females: The replacement
   rate applicable to newly disabled individuals for the base year and future years. This
   variable is also driven by the rules of the pension system regarding eligibility for
   disability pensions. The rates are calculated as the average pension received by a
   newly disabled person of each age and gender group divided by the average covered
   wage in the economy. The data are entered as a percent.

   Columns Lump Sum Payments for Disabled, Males/Females: Average lump sum
   payment divided by average economy wide wage and expressed as a percentage,
   payable only once at the time of disability. This is a second component of the
   disability pension package in addition to the component, described immediately
   above.

WORKSHEET V: PROFILES

The profile worksheet contains data about specific individuals. These profiles will later
be used by PROST to investigate the equity and distributional aspects of the system. The
program will evaluate how well each of the specified individuals is treated by the pension
system. A better understanding of the distributional impact of the pension system
between genders or across generations can be analyzed by looking at how different
‘types’ of individuals fare under the same pension regime. It is not necessary to fill out
this worksheet to run the rest of the model. This worksheet is required only by Module 5,
Individual Retirement Accounts, and while the Module draws on parameters from other


                                          1-29
worksheets, there are no automatic linkages from this sheet or the output of Module 5
back to the other parameters.

The user can fill out information for six different types of individuals at one time and the
model will calculate output for these six individual types all of whom begin work during
the base year and for these same six types who begin work the following year and for all
subsequent years in the projection horizon. If analysis on more individual types is
required, the model can be run again with a different set of 6 individuals specified. The
line numbers below refer only to Profile 1. For some variables, there is a drop down box,
from which the appropriate choice may be made.

Column I, Individuals With Profiles 1&2, 3&4, 5&6 Buy Joint Annuities. Individual
Retirement Accounts module can be run to evaluate the system both from single and
from married individual’s point of view. If the user clicks this option On, 3 families are
created by joining profiles 1 and 2, 3 and 4, 5 and 6. After one of the family member dies
the longer lived of the 2 individuals will receive his own and a survivor pension from the
PAYG system. Under the funded system the longer lived individual would receive both
pensions until he/she dies. If this option is not enabled, individuals would only be
receiving their own annuity, calculated based on individual mortality table.

   Line Gender: Gender of the individual.

   Line Starts Working at Age: The age at which an individual begins his career. Most
   low income and rural workers enter some form of employment earlier than high
   income group workers who enter late after having spent more time in the education
   system.

   Line Plans to Retire at Age: The age at which retirement takes place.

   Line Mortality Multiplier: The ratio of an individual’s mortality rate to the
   mortality rate of the average individual in the economy at the retirement age. For high
   income workers with better access to health services and improved standards of
   living, this mortality multiplier may be less than 100% which would imply that they
   have a lower mortality rate compared to an average individual. Low income
   individuals may have mortality multipliers above 100%.

   Line Starting Wage as % of Cohort Avg: An individual’s starting wage as a
   percent of the wage of an average individual of the same age and gender. This
   variable may also vary among the profiles as a function of the income level.

   Line Productivity Growth Multiplier: An individual’s average wage increase as a
   percent of the wage increase of an average individual of the same age and gender.
   Different profiles could have different values for this variable differentiated by urban
   high income or rural poor income values. For the former profile, it may be assumed
   that productivity increases faster than the norm whereas in the latter case, it could be
   assumed that productivity does not change too much with respect to the norm.



                                           1-30
Line Percentage of Benefit Left to Survivor: Share of individual’s pension left to
survivor. This is determined either by law (in the case of PAYG system) or by choice
at the time of buying joint annuity (in the case of funded system). This value is only
relevant if the joint annuity option in column I is on.

Noncontributory Years: There are two bars with age specifications to the right of
the other data. Click on the first bar after you have made the other changes and it will
reset to conform to what you have specified. You can then highlight years when this
particular individual did not contribute. Note, that resetting of the bar is required
every time you reset the start and end of the career.

Exempt Years: The second bar to the right of the data allows the user to specify
when an individual was legally exempt from contributions but continued to earn
credit towards the pension (the government has to actually contribute to individual’s
account during such periods in the presence of notional PAYG or funded pillar. For
the conventional PAYG pillar usually only individual’s record of length of service is
being increased each year with no actual contribution being made). The bar is filled
with ages corresponding to the years when an individual was not contributing (the
values are automatically taken from the first bar). The user can then mark the years
when individual was exempt.




                                       1-31
WORKSHEET VI: FACTORS

The Factors worksheet contains actuarial tables for annuitization, commutation and
pension reduction due to early retirement.

   Columns Annuitization Factors in Notional PAYG/ Funded Pillar,
   Males/Females: To convert account balance accumulated either in Notional or
   Funded pillar into annuity the account balance is divided by an age and gender
   specific number called annuitization factor. This factor is normally derived based on
   projected mortality rates, pension indexation rules and given a specific discount rate
   is supposed to transform the account balance into a stream of payments with the
   expected net present value equal to the account balance (see PROST calculus manual
   for details). The factors used in practice may or may not be actuarially fair, i.e. net
   present value of expected pension stream derived using actual factors may be
   different from the value of the account. In these columns PROST asks the user to
   input annuity factors that are actually used by the pension institution which PROST
   will use for projecting pension system expenses. Actuarially fair factors are outputted
   as part of PROST results file.

   Columns Pre-reform / Post-reform Commutation Factors with xx% Nominal
   Discount Rate, Males/Females: When a PAYG pension system allows commutation
   of a portion of benefit stream this proportion of pension is converted into a lump sum
   payment by multiplying it by age and gender specific number called commutation
   factor. This factor is normally derived based on projected mortality rates, pension
   indexation rules and given a specific discount rate is supposed to equalize the net
   present value of the given portion of the benefit stream with the lump sum payment.
   The factors used in practice may or may not be actuarially fair, i.e. net present value
   of expected pension stream may be different from the lump sum payment derived
   using that factor. In these columns PROST asks the user to input commutation factors
   that are actually used by the pension institution which PROST will use for projecting
   pension system expenses. Actuarially fair factors are outputted as part of PROST
   results file. The discount rates used to derive actuarially fair commutation factors are
   taken from the numbers specified in columns J and N in line 4 of this range.

   Columns Pre-reform / Post-reform Pension Reduction Coefficients with xx%
   Nominal Discount Rate, Males/Females: When a PAYG pension system allows
   retirement before statutory retirement age it often imposes a penalty by applying a
   reduction coefficient to a regular pension amount. The coefficient usually depends on
   individual’s age and gender. It is normally set based on projected mortality rates,
   pension indexation rules and given a specific discount rate is supposed to equalize the
   net present value of the reduced benefit stream given at the earlier age and a regular
   benefit stream awarded at the statutory retirement age. The reduction coefficients
   used in practice may or may not be actuarially fair, i.e. net present value of the two
   expected benefit streams may be different. In these columns PROST asks the user to
   input reduction coefficients that are actually used by the pension institution which
   PROST will use for projecting pension system expenses. Actuarially fair reduction
   coefficients are outputted as part of PROST results file.


                                          1-32
WORKSHEET VII: REFORM GENERAL

The “ReformGeneral” worksheet contains non-age-specific data about systemic reform
specifications. Thus this and the “ReformPension” sheet can be used to simulate the
impacts of moving from a typical PAYG system to any combination of the following
possibilities: a mono-pillar reformed PAYG, a mono-pillar Notional PAYG and a mono-
pillar Funded DC Scheme, or a multi-pillar scheme combining any two or three of the
above. Reformed PAYG pillar should always be included in any combination to
accommodate existing pensioners who are unlikely to be moved to the new system. At
the top of this input file, the user is given a series of options to choose from. These
options define the main reform rules, number of pillars and how the accrued rights will be
compensated. The rest of the input file variables that are visible depend on these
selections. These variables define how the new pillars would operate. As always, the user
can use the “I” and “D” buttons to insert and delete data columns if necessary.

Systemic Change Options

   Line Reform Year: The year in which pension reform is expected to occur. This
   year must be later than the base year. In other words, the base year should be chosen
   to be at least one year prior to the reform taking place. As will be evident from the
   following discussion PROST also allows modeling the ongoing multi-pillar system.
   In this case “Reform Year” simply means the first year, for which the results of the
   multi-pillar system will be projected.

   Lines Pillars after the reform: Check the pillars that will exist after the reform.
   Available choices are:

       Conventional PAYG             Notional PAYG                 Funded DC

   The user may choose all or any of these systems depending on the purpose of the
   simulation. If the reform is simply a reform to a conventional PAYG system, the user
   can either model this as changes over time to the parameters in the first four
   worksheets or can enter them here. If the new parameters apply to some cohorts and
   not others, it would most likely be easier to model them here. Please note that
   retirement age changes which may accompany the systemic changes must be modeled
   through the first 6 worksheets.

   Conventional PAYG refers to defined benefit plans where contributions from current
   workers are being used to pay pensions to current beneficiaries based on a formula
   combining some portion of the beneficiaries’ earning history, economy wide average
   covered wage, and the years of service of the beneficiary. These systems may or may
   not have an accumulated reserve.

   Notional PAYG refers to defined benefit plans where contributions from current
   workers are also being used to pay pensions to current beneficiaries, but in this case
   the pensions paid to the beneficiaries are related in some fashion to the contributions
   the individual beneficiary made during his/her lifetime. These systems also may or
   may not have an accumulated reserve.


                                          1-33
Funded DC refers to defined contribution plans where individual contributions
accumulate in accounts which are then invested. At retirement a pension is paid from
the amount available in the account, the contributions with the accrued interest.

Lines Accrued rights are paid out as: All reform options keep pensions in payment
untouched, although the indexation on these pensions can be changed as a component
of a reform. However, the reform has to address the issue of the rights accrued to
contributors who have not yet met the eligibility criteria to retire, but have made
significant contributions to the old system and therefore have claims against that
system. Select how accrued rights should be disbursed to these contributors. Choices
are:

Proportional PAYG Benefit - Benefits are based on the proportion of the full PAYG
pension benefit that individual have accumulated before the reform. Thus if for
example a country’s pension system allows for a replacement ratio of about 60% after
30 years of service, an individual with only 15 years of service would be given a life
annuity consisting of approximately 30% of projected wages. This annuity would be
paid at the time of retirement.

Notional Capital - This option is only visible when a Notional PAYG pillar
checkbox is checked in. Instead of beginning the notional account system with zero
balances, individuals are given starting balances which are based on their accrued
rights under the old system. How the accrued rights will be converted will be
specified by the user below.

Recognition Bond – This option is only visible when a Funded DC pillar checkbox
is checked in. Since funded systems pay benefits out of the money accumulated in the
individual’s account, one approach the Government could take would be to give
individuals bonds in recognition of past service, which earn some specified interest
rate and that the combined balance from both the recognition bonds and accumulated
contributions after switching to the new system would be available to the beneficiary
upon retirement.

They are Calculated Based on: This option applies only to Accrued Rights paid as
Notional Capital or as Recognition Bond and determines how accrued rights are
calculated. The user has two options:

   Promised Benefits: Accrued rights are calculated as a proportion of the pension
   that had been promised. Using the example from above, if the original law said
   that an individual with 30 years of service was entitled to 60% of the last salary,
   and the individual had contributed only 15 years, this individual would be entitled
   to 30% of last salary. Past Contributions: Alternatively, the Government could
   take the stance that the system is now changing, so the old system is being shut
   down. Under either a notional account system or a defined contribution system,
   the Government could attempt to determine what the account balances would
   have been had the new system been in place throughout a worker’s career.
   Complete information is never available to correctly make this calculation so


                                      1-34
       choices have to be made to determine how this calculation will be made. These
       parameters will be chosen below.

       These two methods will almost always yield different outcomes. Generally, if the
       current system is generous, providing higher internal rates of return than the
       market would, the government will find it cheaper to base accrued rights on Past
       Contributions. If the current system is not so generous with low rates of return,
       Promised Benefits will be cheaper. However, the actual choice made by a
       country often depends not on cost, but on the political economy. Past
       Contributions are more likely to be viewed as reneging on a past contract.

If Accrued Rights are paid out as Proportional PAYG Benefit and either one or both
Notional PAYG and Funded pillars are present after the reform, a user can specify
directly the amounts credited to initial account balances in Notional PAYG or Funded
Defined Contribution systems in the year of the reform. The following two options
become visible under If Ongoing Defined Contribution Scheme only if Notional
PAYG or Funded are checked in under Pillars After the Reform, and if the option
Proportional PAYG Benefit is chosen under Accrued Rights Are Paid Out As.

       Direct Initial Capital NDC / FDC: If checked in, these options will open Initial
       Capital (Notional) As % of Avg Wage or Initial Capital (Funded) As % of
       Avg Wage input columns in the ReformPension worksheet. Then, the average
       accrued rights can be calculated outside the PROST for each age and gender and
       fed into PROST as direct initial capital. These options might also be useful in
       modeling ongoing individual account systems, when average accrued values of
       the accounts can be inputted as direct initial capital. Also, under Direct Initial
       Capital option the PROST allows paying accrued rights through proportional
       PAYG benefit (see Rules for accrued rights calculation below) in addition to
       direct initial capital, which is credited to individual account balances in Notional
       PAYG or Funded DC system.

Rules for accrued rights calculation

These lines are only activated when Notional Capital or Recognition Bond are chosen.
Under Proportional PAYG Benefit, the accrued right is simply the old replacement rate
(in the non-reformed system)multiplied by the proportion of length of service completed .
No further calculation is necessary. However, in the case when Notional Capital or
Recognition Bond is the way of compensating for accrued rights, more complicated
calculations are required to convert completed years of service into a fixed monetary
value. These also depend on whether the accrued rights are going to be calibrated
according to (1) the promised level of benefits, or, (2) past contributions.

Past Contributions

   Line Wage Data Is Available Inclusive Since Year: What is the earliest year when
   good wage history records are available for contributors. In most cases, complete
   contribution earning histories are not available. However, data appears to exist on


                                          1-35
years of service. The most frequent method of constructing “synthetic” wage
histories is to use the average wage from the earliest available year to now and use
this as defined below to construct a wage history, from which the contribution history
can be derived. We will refer to this as the known average wage.

Line Actual Wages Used for Averaging are Valorized To Inflation / Nominal
Wage Growth: Are the wages used in determining the wage base valorized/adjusted
for inflation/ wage growth? If so, enter the valorization factor as a percent. Thus 0%
would imply no valorization of past wages while 100% would imply that past wages
should be valorized perfectly for inflation / wage growth. For example, if it is
determined that accurate wages exist from 1994 to 2004, should the nominal wages
from 1994 to present be adjusted to reflect inflation, so that this average is expressed
in real 2004 wages?

Line Av. Wage Used for Pre-reform Wages Is Discounted to Inflation / Nominal
Wage Growth: Should the wage data for years prior to the earliest available year be
derived by discounting data available from the earliest year to inflation levels/ past
wage growth? If so, enter the discount factor as a percent (between 0% and 100%). If
0% is entered here and in the next line, then the same known average wage will
appear for all unknown years. If 100% is entered here, then the known average wage
will be reduced for inflation/ wage growth each previous year.

Line Post-reform Wages Enter Unavailable Av. Wage Calculations: Yes or No
need to be entered here. If the known average wage is currently from 1994 to 2004,
and this known average wage is to be used to create the synthetic wages for all those
who retire with unknown wages, the appropriate answer is No. However, for those
retiring in 2006, wage histories from 1994-2005 will actually be available, and for
those retiring in 2007, wage histories from 1994-2006 will be available. If the user
would like the known average wage used to create the unknown wages for those
retiring in 2006 to be the average wage from 1994-2005 instead of 1994-2004 and so
forth, increasing the length of the averaging period as more wages become known,
the appropriate answer is Yes.

Line Real Interest Rate Credited on Past Contributions: What real rates of return
are being applied for past contributions? Enter the rate as a percent.

Historical Data Used for Accrued Rights Calculation

Line Contribution Rate Credited to the Individual Account: Enter the time profile
of historic contribution rates (going back 50 years from the base year) as a percent.
The previous lines determine how the “synthetic” wage history is calculated. These
values determine how contributions are determined. If contribution rates have
changed over the past 50 years, the user should be credited with contributions based
on contribution rates relevant in previous years rather than with today’s contribution
rate.




                                       1-36
   Line Assumed Real Average Wage Growth / Inflation Rate: Enter the assumed
   rate of growth of real wages and inflation for years prior to the base year. The future
   path of average wage growth and inflation is entered in the “General” sheet.

Promised Benefits

There is no one universally accepted method of converting a stream of expected future
pension benefits into a finite value today, which would be required if the government
either offered initial notional capital or a recognition bond. Instead the policymaker has a
number of choices. The choices will have an impact on transition costs, but will also
depend on political feasibility.


   Line Wage used in calculations: The pension rules of the original system generally
   state that the individual is entitled to x% of the last wage or average of the last y years
   when he retires. When the pension is being converted to a finite value, the question
   of what to use for the last wage arises. Since the model is doing projections into the
   future, the program knows what wage the individual will receive at the time of
   retirement. That wage could be used, Projected wage of the contributor, or the
   wage that the individual was receiving at the time of the reform could be selected,
   Last wage of the contributor. Generally, Projected wage of the contributor will
   be more expensive than Last wage of the contributor, but Last wage of the
   contributor is a known amount and the growth of wages can be reflected in the
   interest rate offered, in the case of a recognition bond.

   Line Discount Rate Used in Calculation: Since the pension was to be paid in the
   future when the person retires, not at the current moment, the stream of benefits needs
   to be discounted back to the present. The user can choose the discount rate to be used
   for this. This may be the same as the market interest rate, but it is also a policy
   variable that the government can choose, which will impact the assessed accrued
   rights tremendously.

After the Reform PAYG Benefits to Disabled and Orphans

   The reform, as modeled here, currently only reflects changes to the old age pension
   system. However, often reforms also incorporate changes to disability and orphan
   benefits as well. Survivor benefits change also, but they usually are specified as a
   portion of PAYG old age pension, which is what PROST assumes in its calculations.
   Often insurance companies take over or supplement the payments to either one or all
   of these groups. In other cases, even though the old age benefits now come partly
   from a funded pillar, the disability pensions remain defined benefit and from the
   PAYG pillar. The variations across countries in multi-pillar arrangements for
   disability and survivor payments are enormous. In future versions we may try to
   model this insurance company behavior and incorporate some of these variations.
   However, doing so makes the comparison of benefits before and after reform more
   difficult. In this version, we give the user only two options. After the reform PAYG
   Benefits to Disabled and Orphans could either Remain at Pre-Reform levels, with


                                            1-37
   the benefits to these groups continuing to be paid from the public pension agency, or
   could Decrease or Increase By the Same Proportion as PAYG Old Age Benefits.
   If a funded pillar is included in the reform and the user chooses Remain at Pre-
   Reform levels, the individual’s accumulated fund balance is turned over to the public
   pension agency and the benefits are paid by the agency. This is almost equivalent to
   a topping up system, where the individual receives his/her own balance and the
   necessary money from the public agency to top up that balance to the level
   guaranteed in the law. However, should the individual’s balance exceed the amount
   guaranteed in the law, the model assumes that the entire balance is turned over to the
   public pension agency while the individual only receives the legally guaranteed
   amount, while a topping up system would result in the individual receiving the higher
   benefits with no intervention on the part of the state.

   If the proportional change option is chosen, then if the pension for new old age
   pensioners paid by the PAYG system falls by half as some of the contributions are
   diverted to a funded pillar, a the disability payments from the PAYG pillar will also
   fall by half, with the assumption that there will be supplemental disability payments
   coming from the funded pillar or through an insurance company as well. In line
   Money Dedicated to Pay Benefits Other than Old Age as % of Contribution , the
   user is asked to specify the percentage of contribution which is dedicated to paying
   benefits other than old age. The model assumes that all of that revenue is distributed
   annually among these other beneficiaries, the disabled and survivors, on top of their
   remaining balances. The user can then look at the benefits received by these
   recipients to see whether they are sufficient and if not, can adjust the revenue
   dedicated to these payments in line Money Dedicated to Pay Benefits Other than
   Old Age as % of Contribution. Thus, there is a direct link between the generosity
   of the benefits paid to the disabled and survivors and the overhead charged to each
   contributor which lowers the amount available to accumulate in the contributor’s old
   age account.

Collection Rate (Conventional PAYG, Notional PAYG, Funded)

Collection rate is the percentage of the contributions from contributors received by the
pension system. The assumption is that the collection rate on all three pillars, if they
exist, is exactly the same. This assumes that either the contributions are collected
together by the tax authorities, by the public pension agency, or by a separate
clearinghouse agency or that even if collected separately, there is sufficient cross-
checking to prevent individuals from paying one set of contributions and not the other.
The presumption in most reforms is that the reform process either improves incentives to
contribute or is accompanied by administrative reforms which improve the efficiency of
collection. If neither is true, the user can enter the same values as in the table Revenue
Sources of the “General” sheet.

   Lines From Switchers and Their Employers/From Non-Switchers and Their
   Employers/From Pensioners/From Government: A good pension reform should
   increase transparency, fairness, and provide good value for contributions. Under
   these conditions, switchers might have greater incentive to contribute than non-


                                          1-38
   switchers. If the user believes these conditions will be fulfilled, higher values may be
   entered for switchers than for non-switchers. If the conditions are not fulfilled, then
   the values for switchers and non-switchers would generally be the same, although
   they may be higher due to administrative reform than in the original system. Lines
   From Pensioners/From Government incorporate collection rates from pensioners
   and from the government where these are relevant. A value of 100% implies that the
   collection system is completely efficient and fraud and arrears amount to 0%.

PAYG Fund Parameters (Notional + Conventional)

This group of parameters corresponds to some of the lines from the tables Revenue
Sources and Costs and Other Expenditures of “General” sheet. If nothing is expected
to change after the reform, the same values used there can be entered here. However, if
the parameters are expected to change, the new values should be entered here. The
parameters for notional and conventional PAYG pillars are lumped together here because
these systems are both public systems and if they are run concurrently they are run by the
same public agency.

   Line Unconditional budget transfers to pension fund as % of GDP: This field
   records unconditional budget transfers to the pension fund (other than the wage-bill
   based tripartite Government contributions) made by the Government as a percent of
   GDP. For example, some countries earmark x percentage of the VAT tax to the
   pension fund. Other countries earmark privatization revenues for supporting the
   pension system. Neither of these sources of revenue is directly tied to the covered
   wage bill, contributions from which are the primary source of revenue. Please note
   that this line does not include transfers from the non-pension Government budget to
   cover pension system deficits, hence the term Unconditional, but only items that are
   earmarked regardless of the status of the pension system. Data are required for the
   base year, the end year and any other years in between.

   Line Other Income as % of Employee, Employer and Government contributions:
   Enter any other income received by the pension fund as a percent of total
   contributions. Thus an entry of 10% would imply that other income sources of the
   pension fund were able to add to contribution revenues additional 10%. This income
   will be tied to the wage bill since contributions are tied to the wage bill. This line is
   useful if the pension fund engages in non-pension activities and the user would like to
   look at the overall budget of the pension fund. However, the non-pension
   expenditures are specified in a highly aggregated fashion, leading to only crude
   estimates of the impact on the overall budget of the pension fund.

   Line Assets Management Costs as % of Assets: The costs required to manage the
   assets of the pension system. The data are required for both the base year as well as
   the end year and entered as a percent. These costs are usually a function of the
   volume of financial assets being managed by the pension system. Typical values
   range from below 1 percent (meaning less than 1 percent of the previous years’ fund
   balances were used for asset management) to about 6 percent (meaning 6 percent of
   the pension funds assets are used for asset management expenses).


                                           1-39
   Line Administrative Costs as % of Employee, Employer & Pensioner
   Contributions: The transactions or administrative costs of running the pension
   system, excluding the asset management costs included above. This includes costs of
   setting up the pension registry, costs of book keeping, and administering the pension
   system. The data are needed for both the base year as well as the end year and are
   entered as a percent of total contributions from employers, employees, pensioners.
   Typical values range from about 1% meaning that almost 1% of contributions are
   being used to administer the pension system to over 50%.

   Line Other Pension Fund Expense as % of Pension Expense: The percent of total
   expenditures attributed to non pensions related expenditure or “other” expenditures.
   This data must be entered as a percent for the base year, the end year and any other
   years in between.

   Line Pensioner Contribution Rate on PAYG Pension Payments: If pensioners are
   required to contribute towards social security, enter the average rate of contribution as
   a percent of their gross pensions.

Indexation of Minimum Pension on All Income

   Lines For Switchers/Non-switchers to Inflation/Nominal Wage Growth. Given
   that pension benefits can potentially come from the three different pillars the program
   applies the minimum pension to the sum of all annuities regardless from which pillar
   the benefits come. In reality minimum pension is sometimes applied to the benefits
   from just one pillar, in which case the user will have to produce 2 input files – one
   with the pillars to which minimum is applied and another with the pillars that do not
   offer minimum pension guarantee, or offer another separate guarantee. PROST allows
   for 2 different guarantees – one for the switchers and one for non-switchers. The
   parameters on these lines ask the user to specify how these different minimum
   guarantees for switchers and non-switchers will be indexed to inflation and wage
   growth.

Minimum Pension Guarantee Calculations and Payment Options

   Line Mode of Payment. In this version of PROST only one option of payment is
   allowed, which is the option of yearly guarantee. This means that sum of all annuity
   payments from all pillars is calculated each year and compared against the minimum
   pension guarantee. If the guarantee is higher than total benefits from all funds the top-
   up will be paid from PAYG pillar.

Notional Pillar Information

The next set of variables is only visible to the user if the simulation case involves the
presence of a Notional PAYG Pillar after the reform and the user has checked Notional
PAYG as an option.

Notional Real Interest Rate Determination



                                           1-40
As explained above, under a notional accounts system, the hypothetical balances derived
from recording contributions each year are given interest on the balance every year.
Unlike a funded system, where the contributions are being invested, the contributions are
actually being used to pay current contributors. The government can then decide what
real interest rate to pay on these hypothetical balances.

   Line Base Notional Real Interest refers to the case where the government chooses to
   pay a fixed real interest rate, which the user may specify as a percent for the base
   year, the end year and any other years in between. Zeroes could be entered for the
   remaining parameters in this range, if for example, the government declared that the
   notional interest rate would be 3%. On the other hand, the government could declare
   that the notional interest rate would be 1% plus the growth rate of wages. In this
   case, 1% would be complemented with the entries in the subsequent fields.

   Line Option 1: GDP; Option 2: GDP per capita. First, the user should choose
   which option of data entry he or she wants to use by selecting from drop-down list in
   the column G. The notional interest rate could be tied to either GDP or GDP per
   Capita. The parameters in columns K-L have to reflect which proportion of GDP or
   GDP per capita growth will be added to the base notional real interest rate (which
   could be zero).

   Line Option 1: Wage Bill; Option 2: Average Wage. First, the user should choose
   which option of data entry he or she wants to use by selecting from drop-down list in
   the column G. The notional interest rate could be tied to either growth of Wage Bill,
   which is the rate of growth of PAYG revenues or Average Wage. The parameters in
   columns K-L have to reflect which proportion of wage bill or average wage growth
   will be added to the base notional real interest rate (which could be zero).

Contribution

   Line Contribution Rate From Switchers: Enter the contributions to the notional
   system as a percent of insured wage made by those who have switched to the notional
   account system.

   Line Contribution Ceiling for Switchers as % of Total Overall Ceiling: Some
   countries choose to limit PAYG system (Notional in this case) by lowering the ceiling
   on insured wages, often leaving remaining part of wage to be insured by mandatory
   or voluntary funded pension system. This parameter asks to express the new ceiling
   as a percent of the ceiling of unreformed system, specified in the table Contributions
   of “General” input sheet.

   Line Contribution Rate From Employers of Switchers: Enter the contributions as
   a percent of wage made by employers of those who have switched to the notional
   account system. The wage can be insured wage or gross wage, depending on the
   choice in line Switcher Contribution Ceiling Applies for their Employers.




                                          1-41
   Line Switcher Contribution Ceiling Applies for their Employers: “Yes” would
   mean that ceiling specified in line Contribution Ceiling for Switchers as % of
   Total Overall Ceiling applies to employers. “No” would mean that gross wage is
   calculated to determine employer’s contribution amount. In any case only insured
   wage under the ceiling is added to notional capital of the individual.

   Line Government Contr. Rate for Switchers: Enter the contributions as a percent
   of gross wages made by the Government to switchers’ notional accounts. Use 0% if
   no such credits are given by the Government. These do not include government
   contributions for special categories; only government contributions to all individuals
   under a tripartite contribution system should be included here.

   Line Total Contribution Rate Credited to Switcher's Account: Because a notional
   account system is notional and not funded, there may be a difference between what is
   credited to the individual’s account and what is actually paid. Lines Contribution
   Rate From Switchers and Contribution Rate From Employers of Switchers refer
   to what is actually paid and therefore revenue to the system. The amount credited to
   the account could be less than what is paid if the government is claiming a solidarity
   payment to help cover transition costs or if contributions are being used for other
   insurance plans or if contributions are covering administrative costs. Conversely, the
   amount credited could be greater than what is paid. This could be viewed as a
   government contribution in the sense that the government is obligated to pick up
   deficits from notional account systems. The contribution is different from the direct
   government contribution made in line Government Contr. Rate for Switchers,
   which is paid annually while an individual works. In this case the Government only
   makes a promise to pay when there are deficits in the future.

   Line Government Contr. Rate for Exempt Switchers as % of Min. Wage: The
   Government can choose to make contributions on behalf of special categories of
   exempt individuals, such as mothers caring for young children. Note that in the
   “Pension” sheet, the user specified how many people of each age and gender category
   were receiving credit for service without paying. In notional account systems in order
   for the individual to receive credit, a contribution must be made. But the wage on
   which that contribution should be based is not determined since the person is not
   actually earning. If the government only makes contributions on the basis of the
   minimum wage, the contribution rate should be entered here. If the government
   makes contributions on the basis of the average wage of individuals in that age and
   gender cohort, then the contribution rate should be entered in line Government
   Contr. Rate for Exempt Switchers as % of Imputed Wage.

   Line Government Contr. Rate for Exempt Switchers as % of Imputed Wage:
   Enter the contributions as a percent of imputed gross wages made by the Government
   on behalf of exempt individuals. Imputed gross wages are the average wage of
   individuals in that age and gender cohort.

Indexation (%)



                                         1-42
Enter data regarding pension indexation rules for the notional pillar. Many pension
systems index pensions to either nominal wages or inflation (or some combination of
both). If the level of inflation is I, and rate of nominal wage growth is G, then the
indexation mechanism is described by [(1+aI)*(1+bG)-1] where a and b are the user input
inflation indexation parameters. Data are needed for the base year as well as for the end
year (and intervening years if available).

   Lines Switcher Pension Indexation to Inflation / Nominal Wage Growth /
   Nominal Notional Interest: The data are required for both the base year as well as
   the end year and are entered as a percentage. Differently from pension payments from
   other pillars, notional pillar allows indexation of the pension to notional interest. This
   would allow continuity of policy after the person retires, as account value before the
   retirement also grows at the notional interest rate.

   Lines Maximum Switcher Pension Indexation to Inflation / Nominal Wage
   Growth / Nominal Notional Interest: The data in these fields should be entered the
   same way as data for pension indexation. This indexation is applied to the maximum
   pension initially specified by the maximum pension indexation rules before the
   reform, specified in “General” input sheet. After the old age pension is calculated,
   whether at the time of awarding the pension or at the time of indexation, it is checked
   if the resulting payment does not fall above the maximum pension. If it does, the
   payment of maximum pension is made to the pensioner instead.

Rules for Annuitization

Under notional accounts, when a person retires, the balance in the notional account is
then converted into a pension payment. In the simplest case, the balance is divided by the
life expectancy of the individual, given his or her age of retirement and then indexed as
specified above. This assumes that the only interest rate being paid on the balance post-
retirement is what has been specified as indexation. Full inflation indexation would
result in a zero real interest being paid; nominal wage growth indexation would result in a
real interest rate equal to real wage growth, etc.

   Line Real Interest in Decumulation Phase: This refers to the interest rate provided
   on the notional account balance after the pension has started to be paid. The user
   choices are None, which allows the indexation to determine the interest rate being
   paid, Notional, which pays an additional interest rate equal to whatever has been
   specified as the notional interest rate, and Discount Rate, which is the market interest
   rate. Note that this interest rate is in addition to what is being paid through the
   indexation.

   Line Decrease in Interest Due to Annuitization: The aim of notional account
   system is to imitate funded system of individual accounts. In the funded system
   insurance companies would charge a premium for taking longevity risk of an
   individual at the time of retirement. This premium is usually expressed as a reduction
   in interest rate from the level that the company would otherwise offer for purely
   managing individual’s funds. In the notional system this longevity risk is taken by the


                                           1-43
   government which at least theoretically can also charge such premium. For example,
   an input of 1% here means that assumed future interest rates are reduced by 1% of
   assets when the account balance is converted into annuity at retirement.

Funded Pillar Information

The next set of variables is only visible to the user if the simulation case involves the
presence of a Funded DC component after the reform and the user has checked Funded
as an option.

Contribution

   Line Contribution Rate from Switchers: Enter the contributions as a percent of
   insured wage made by switchers.

   Line Contribution Rate from Employers of Switchers:              Enter the employer
   contributions as a percent of insured wage.

   Line Government Contribution Rate for Switchers: Enter the contributions as a
   percent of insured wage made by the Government to switcher’s DC accounts. Use 0%
   if no such credits are given by the Government. Again these are statutory tripartite
   contributions on the part of Government, not residual covering of the deficit.

   Line Government Contribution Rate for Exempt Switchers as % of Min. Wage:
   Enter the contributions made by the Government on behalf of an exempt individual
   who has switched as a percent of the minimum wage in the economy. Exempt
   individuals are those who receive credit for contributions, but are not making
   contributions themselves. If the Government makes those contributions, which is the
   only way to get credit in a funded system, and the Government makes the
   contributions on the basis of minimum wage, the contribution rates should be entered
   here.

   Line Government Contribution Rate for Exempt Switchers as % of Imputed
   Wage: Enter the contributions made by the Government on behalf of an exempt
   individual on the basis of an imputed wage. The imputed wage is the average wage
   earned that year by an individual of the same age and sex cohort as the exempt
   individual.

   Line Contribution Floor as % of Minimum Wage: Some countries set up the
   funded DC system in such a way that only income above certain threshold is insured
   by funded system (income below this threshold is usually insured by PAYG system).
   Enter this threshold as a % of minimum wage (the value should be above 100%).
   Enter 100% if all income is insured by funded pillar.

   Line Credit to Switcher’s Account as % of Contribution: In a funded system, the
   full contribution rate might not be credited to the individual’s account since some of
   the contribution may be used to cover disability and survivor premiums and


                                          1-44
   administrative costs. These are indicated on lines Fees for fund administration as
   % of switcher and employer contribution and Money dedicated to Pay Benefits
   Other than Old Age as % of Contributions below. Ideally, the sum of lines Fees
   for fund administration as % of switcher and employer contribution and Money
   dedicated to Pay Benefits Other than Old Age as % of Contributions should
   equal 100. However, it is possible that the insurance companies are building a hedge
   fund in the initial years of the system to protect themselves against future disability
   risks. In that case the money paid out to the disabled and survivors this year and
   administrative costs spent this year plus what goes in one’s account may not sum up
   to 100. Similarly, if the insurance companies later draw down that hedge fund, the
   sum may exceed 100. Thus, the insurance premium paid by the contributor is being
   split in 2 or 3 segments: line Fees for fund administration as % of switcher and
   employer contribution reflects the administrative costs included in the insurance
   premium, line Money dedicated to Pay Benefits Other than Old Age as % of
   Contributions reflects the actual amount paid to other beneficiaries each year, and
   the building of a hedge fund is reflected as the difference between 100% and the sum
   of lines Fees for fund administration as % of switcher and employer contribution
   and Money dedicated to Pay Benefits Other than Old Age as % of Contributions.

   Line Fees for fund administration as % of switcher and employer contribution :
   The percentage of the contribution used to cover administrative costs. This entry
   should exclude insurance payments to the disabled, survivors, and orphans. However,
   it should include the administrative costs of the insurer as well as those of the fund
   manager.

   Line Money dedicated to Pay Benefits Other than Old Age as % of
   Contributions: The percentage of contribution used to cover premiums for
   disability, survivors, orphans, or life insurance. Each year the disabled and survivors
   are paid out what remains in their own or in the deceased’s account plus the amount
   specified here.

   Line Asset management Fee as % of Assets: Asset management fees assessed as a
   proportion of funds under management in each account each period.

   Line Contribution Ceiling for Switchers as % of Total Overall Ceiling: Some
   countries choose lower or higher level of the ceiling on insured wages for funded DC
   than for PAYG pillar. This parameter expresses the new ceiling as a percent of the
   ceiling of unreformed system, specified in the table Contributions of “General” input
   sheet.


Transition Cost Financing

Here the user can describe how transition costs of switching to the multi-pillar system are
financed. The transition cost here is defined as additional cash requirements for the




                                           1-45
PAYG system due to the introduction of the multi-pillar system. There are 3 different
financing options described below.

   Line Borrowing: Lines Borrowing, Privatization Revenues and Tax Financing
   should always sum up to 100%, signifying that all transition costs have to be financed
   in one of the three ways. It is assumed that all cash requirements not satisfied by
   privatization revenues or tax financing will have to be satisfied by borrowing at
   government bond interest rates.

   Line Privatization Revenues: Percent of the transition costs financed by
   privatization revenues.

   Line Tax Financing: Percent of the transition costs financed by raising general taxes
   or reducing spending for other government programs.

Interest Rates

   Line Real Interest Earned on Recognition Bonds: As noted above in the
   discussion of accrued rights, once the recognition bonds have been provided to
   individuals with accrued rights in the old system, these bonds generally earn some
   rate of interest, which is a policy variable on the part of the Government.

   Line Real Interest Earned on Individual Accounts: This is generally higher than
   risk free market interest rates, since individuals in general hold some risky assets that
   earn higher interest rates on average. Government regulations may limit the extent of
   such risky investments. Generally, a stable economy requires a risk free real interest
   rate 1-2 percentage points above the rate of real GDP growth.


Rules for Annuitization and Pension Payout

In the case of a funded system, usually payment is made in one of three ways: as a lump
sum, as a scheduled withdrawal, or as an annuity bought through an insurance company.
A lump sum allows the beneficiary to withdraw all or some of the money in the account
upon retirement. This lump sum could be used to purchase an annuity; in fact in most
countries, the purchase of a minimum annuity is required if the entire sum is withdrawn
as a lump sum. A scheduled withdrawal is like an annuity, but a little different. The
account continues to earn interest rate at market rates even while the pensioner withdraws
funds, but the amount of the pension each year is recalculated taking into account the
accumulated balance with interest and the remaining life expectancy of the individual.
Since the pension is recalculated each period, nothing is automatically indexed. While a
scheduled withdrawal can result in very low pensions at the end of life if a person lives a
very long time, if a person dies early, the remaining balance is turned over to survivors in
addition to whatever survivor insurance had been purchased.

   Line Decrease in Interest Due to Annuitization: While the account balances had
   been accumulating interest at market rates, annuities generally pay lower interest


                                           1-46
   rates, which compensate the insurance company for the longevity risk. Enter the
   reduction in interest rate from the market rate paid on the annuity. For example, if
   you enter 1%, the annuity will be assumed to pay an interest rate 1 percentage point
   below the market rate.

   Lines Indexation of Annuity to Inflation / Nominal Wage Growth: Insurance
   company provided annuities may or may not be indexed to inflation. The user is
   asked to specify. While it is extremely rare that an insurance company can offer an
   annuity indexed to nominal wage growth, if the PAYG pension is indexed to nominal
   wage growth, the user may want to specify the same indexation for both the PAYG
   pension and the annuity from the funded pillar to be able to add pensions from
   different pillars to get a sense of the true change in replacement rates pre and post
   reform.

   Line Mode of Benefit Payments: this parameter determines how benefits are
   actually paid from the pension fund, whether by scheduled withdrawal, lump sum or
   annuity.

Reformed PAYG Pillar

PROST assumes that Conventional PAYG pillar is always present after reform, at least to
continue pay the benefits for current retirees. More generally, conventional pillar can be
present with two sets of contribution and benefit specifications, one of which is for
switchers and the other for those who do not switch to multi-pillar system (in this case
parameters for non-switchers can still encompass the reform).

Contribution

The contribution collection rules towards the reformed PAYG system broken down by
source and according to whether the individual is a switcher (i.e. someone who is now in
the new system whether it is a reformed mono-pillar PAYG or a multi-pillar system) or a
non-switcher (the individual remains in the mono-pillar PAYG system). Note that in the
other pillars, Notional Account and Funded, contributions were only taken from
switchers since it is assumed that these pillars did not exist initially. However, here there
may be differential contribution collection rules from switchers and non-switchers.

   Line Contribution Rate from Switchers: Enter the contributions as a percent of
   insured wage made by switchers.

   Line Contribution Ceiling for Switchers as % of Total Overall Ceiling: Some
   countries choose to limit PAYG system by lowering the ceiling on insured wages,
   often leaving remaining part of wage to be insured by mandatory or voluntary funded
   pension system. This parameter asks to express the new ceiling as a percent of the
   ceiling of unreformed system, specified in the table Contributions of “General” input
   sheet.




                                            1-47
   Line Contribution Rate from Employers of Switchers: Enter the contributions as a
   percent of wage made by employers of those who have switched to the multi-pillar
   system. The wage can be insured wage or gross wage, depending on the choice in line
   Switcher Contribution Ceiling Applies for their Employers.

   Line Switcher Contribution Ceiling Applies for their Employers: “Yes” would
   mean that ceiling specified in line Contribution Ceiling for Switchers as % of
   Total Overall Ceiling applies to employers. “No” would mean that gross wage is
   calculated to determine employer’s contribution amount.

   Line Contribution Rate from Government for Switchers: Enter the contributions
   as a percent of insured wage made by the Government on behalf of switchers.

   Line Contribution Rate from Non-switchers: Enter the contributions as a percent
   of insured wage made by non-switchers.

   Line Contribution Ceiling for Non-switchers as % of Total Overall Ceiling:
   Some countries choose to limit PAYG system by lowering the ceiling on insured
   wages, often leaving remaining part of wage to be insured by mandatory or voluntary
   funded pension system. This parameter asks to express the new ceiling as a percent of
   the ceiling of unreformed system, specified in the table Contributions of “General”
   input sheet.

   Line Contribution Rate from Employers of Non-switchers:               Enter the
   contributions as a percent of wage made by employers of those who have not
   switched to the multi-pillar system. The wage can be insured wage or gross wage,
   depending on the choice in line Non-switcher Contribution Ceiling Applies for
   their Employers.

   Line Non-switcher Contribution Ceiling Applies for their Employers: “Yes”
   would mean that ceiling specified on line Contribution Ceiling for Non-switchers
   as % of Total Overall Ceiling applies to employers. “No” would mean that gross
   wage is calculated to determine employer’s contribution amount.

   Line, Contribution Rate from Government for Non-switchers: Enter the
   contributions as a percent of insured wage made by the Government on behalf of non-
   switchers.

Line Use Benefit Formula for Old Age Pensioners: Check this box if the reformed
PAYG system simulations would require the use of an earnings-specific benefit formula.
This would be an additional part of the pension on top of earnings unrelated part,
described in the sheet “ReformPension”. It is possible to be inconsistent with what is
entered in the “General” sheet in which case the final benefit will be a weighted average
of the two benefits putting more and more weight on the benefit formula of
“ReformGeneral” sheet for more remote periods in the future.




                                          1-48
Lines Benefit Formula Parameters for Old Age Pensions: This range will be visible
only if the Use Benefit Formula for Old Age Pensioners box is checked in, and
describes the earnings specific part of the pension. The parameters are identical to those
in “General” sheet. The user is asked to specify the parameters that apply to those who
have switched to the mono-pillar PAYG or multi-pillar and those who have not and those
that apply to those who did not switch. The parameters for those who do not switch do
not have to be exactly the same as the original parameters, as reforms often impose some
changes on those who remain in the old system as well. During the transition period, the
actual replacement rates for new old age pensioners will be a weighted sum of these rates
and the rates under the old system.

Commutation

The Commutation table requires similar inputs as discussed for commutation in
“General” sheet. These parameters after reform will apply to both switchers and non-
switchers.

Pension Indexation

The user is asked to specify the indexation of the PAYG pension. Note that there is no
difference in the indexation between pensions of those who switched and those who did
not. If the user feels that a distinction is necessary, a rough approximation may be
achieved by changing the indexation in the year when all retirees are expected to be
switchers. However, the indexation of maximum pension is allowed to be different for
switchers and non-switchers. Usually, if those two classes of contributors contribute at
different contribution rates having two different policies on maximum pension would be
expected.

   Lines Pension Indexation to Inflation/Nominal Wage Growth: The data entered
   for this variable reflects the indexation of pension payments to inflation/nominal
   wage growth. The data are required for both the base year as well as the end year. A
   value of 100% implies that pension payments are fully indexed to inflation/wage
   growth. Depending on country circumstances values could range from 0% (no
   indexation) to 100% (full indexation).

       Note: Pension indexation to inflation is cheaper than nominal wage indexation
       when real wage is growing, since nominal wage indexation by definition, indexes
       to real wages as well as inflation)

   Lines Indexation of Commuted Pension Portion to Inflation / Nominal Wage
   Growth: Commuted pension portion might be indexed differently than overall
   pension. For example, commuted pension portion might not be indexed at all, while
   overall pension might be indexed to inflation. Then if the initial pension is equal to
   $100, 30% of it is commuted and inflation is equal to 5%, then annuity pension
   payment the next year will be $100*1.05-$30 = $75.




                                          1-49
Lines Maximum Switcher / Non-switcher Pension Indexation to Inflation/
Nominal Wage Growth: The data in these fields should be entered the same way as
data for pension indexation above. This indexation is applied to the maximum
pension initially specified by the maximum pension indexation rules before the
reform, specified in “General” input sheet. After the old age is calculated, whether at
the time of awarding the pension or at the time of indexation, it is checked if the
resulting payment does not fall above the maximum pension. If it does, the payment
of maximum pension is made to the pensioner instead.




                                       1-50
WORKSHEET VIII: REFORM PENSION

This sheet allows the user to input age-specific parameters pertaining to the structural
multi-pillar reform and complements those from “Reform General” sheet.

Columns Percent of Switchers: During a reform, not all age groups are treated alike.
Frequently older age groups are left in the old system with all their benefits intact. In
many cases, younger workers are given the choice of switching to the new system or
remaining in the old system. New entrants up to a certain age at the time of reform and
often the youngest groups of workers are generally only allowed to enter the new system.
The parameters in this group allow the user to determine what the switching behavior will
be. The user is asked for a switching pattern for each age group or an estimate of
expected switchers as a fraction of the contributing population in each age cohort
specified.

Cell New Entrants up to Age: Specify the upper age of new entrants who may become
switchers or non-switchers as specified below in Percent of Accrued Rights that is
Recognized, Switchers / Non-switchers, Male / Female. The age here is defined at the
time of reform.

Cells Contributors at Ages 0 to maximum age indicate the ages of contributor cohorts,
who are in the system at the time of reform. The range of ages can be increased or
decreased by using buttons “I” and “D” located at the bottom of column Age.

In the cases of voluntary switching, it is possible and, in fact, recommended that after a
particular switching pattern is chosen, the output be carefully reviewed to determine
whether switchers’ benefits are lower or higher than benefits for non -switchers retiring in
a particular year. If non-switchers have substantially higher benefits than switchers, then
few people are likely to switch in that age group. Conversely, if switchers have higher
benefits than non-switchers, then more people are likely to switch in the age groups
retiring in that particular year.

Columns Initial Capital (Notional / Funded) As % of Avg Wage are activated only if
Direct Initial Capital NDC / FDC options are checked in the worksheet ReformGeneral.
The average accrued (prior to the reform year) rights are calculated outside the PROST
for each age and gender and entered into these columns as the starting amount of direct
initial capital in the reform year.

Percent of Accrued Rights that is Recognized, Switchers / Non-switchers, Male /
Female applies only to the accrued rights that are paid out as Notional Capital or as
Recognition Bond (see description under Worksheet VII: ReformGeneral above).
Accrued rights are also not always treated symmetrically across age groups and even
within age groups, sometimes distinctions are made between those who switch and those
who do not. The rationale is that older workers close to retirement have little ability to
benefit from a new system and should be protected as much as possible from
uncertainties imposed by the new system. Younger workers, on the other hand, may not
value their pension benefits as much and have a longer working career to adjust to the



                                           1-51
new system and could be protected less. Furthermore, those who switch to the new
system, which might generate higher returns, may be willing to switch with little
compensation for their old contributions.

The user is asked how much of accrued rights are going to be honored by the pension
system (or actually disbursed) as a percent of the full value of accrued rights. This
information is necessary for both switchers (those switching to a new pension system)
and non-switchers (those remaining in the old PAYG pension system). Thus, for
switchers, an entry of 30% in the age group 20-25 and an entry of 80% for the age group
50-55 would imply that 30% of the value of accrued rights would be paid to all switchers
in the age group 20-25 whereas 80% of the value of the accrued rights would be paid to
those contributors in the age group 50-55. It should be noted that paying 100% of accrued
rights to younger workers might actually result in the affected workers receiving higher
pensions than they would have under the old system since the financial markets may be
more generous than the PAYG system. The honoring of accrued rights also affects the
incentives of individuals to switch. The government may want to alter these incentives to
manage the fiscal costs of transition to a funded system.

Note: Pensions already in payment are 100% protected. They are affected only by
changes in indexation parameters after the reform.

Columns Replacement Rates for Male/Female Switchers/Non-switchers: The
replacement rate ranges are identical to what was entered in the Pension Sheet, with
replacement rates for new pensioners specified by age of retirement and gender for each
year as a percentage of the economy wide average wage. The user is asked to specify
replacement rates for switchers and non-switchers separately. The parameters for those
who do not switch do not have to be exactly the same as the original parameters, as
reforms often impose some changes on those who remain in the old system as well.
During the transition period, the actual replacement rates for new old age pensioners will
be a weighted sum of these rates and the rates under the old system.

Columns Lump Sum Payments for Male/Female Switchers/Non-switchers: The lump
sum ranges are identical to what was entered in the Pension Sheet, with lump sums for
new pensioners specified by age of retirement and gender for each year as a percentage of
the economy wide average wage. The user is asked to specify lump sums for switchers
and non-switchers separately. The parameters for those who do not switch do not have to
be exactly the same as the original parameters, as reforms often impose some changes on
those who remain in the old system as well. During the transition period, the actual lump
sums for new old age pensioners will be a weighted sum of these parameters and the
parameters under the old system.




                                          1-52