I. Introduction

Although planning and policy analysis are perhaps ten percent retrospection and ninety percent prospection, the amount of computer software available to assist these functions is vastly weighted in favor of analyses of historical data, rather than on forecasting the future implications of alternative policies or demographic developments. The software for statistical analysis is well-known and ubiquitous -- SAS, SPSS, BMDP -- with powerful versions available for desk-top microcomputers. While recent years have seen the advent of much-improved time series analysis forecasting packages, the software available for supporting simulation or projections under alternative assumptions about demographic or programmatic changes is, by comparison, limited and relatively little-used.

A major factor is cost -- a single application of a major microsimulation forecasting program, for example, can cost thousands of dollars to set up, thousands of dollars in computer time to implement, and weeks of time may pass before the final results are available. Furthermore, such runs require massive amounts of computer core and disk or tape storage on a large mainframe computer. The exigencies of most planning situations do not allow for the luxury of such slow, high-cost, data-intensive and labor-intensive techniques. Faced with only a few hours to obtain, for example, an estimate of the budgetary implications of a proposed new program regulation or policy, the analyst often has to resort to "back-of-the-envelope guesstimates."

Times have changed! The DESTINY Planning and Forecasting System has been developed to provide the planner with a low-cost, easy-to-use means for making detailed forecasts of target populations, the need for services, the requirement for facilities, equipment, and personnel, and the cost under varying programmatic and demographic assumptions. The time required to implement a DESTINY run ranges from a few minutes to a few hours, and the computer cost is negligible -- a few minutes on a desktop microcomputer. The DESTINY system is a powerful new tool for the planner who needs detailed, state-of-the-art forecasts on a quick-turnaround, low-cost basis.

The essence of planning and policy analysis is the ability to identify potential demographic and economic developments, to synthesize alternative responses to those developments, and to evaluate the impact of those responses. The DESTINY Planning and Forecasting System has been developed to play a crucial role in this process -- it offers the user the ability to make fast, detailed population projections, and to forecast quantities linked to growth or structural changes in the population.

DESTINY can offer the planner or policy analyst valuable help in making a wide range of forecasts...

In Population Planning, DESTINY can be used to forecast the population by age, sex, and race, or by geographic region and race.

In Education Planning, DESTINY can forecast local-area school enrollments, based on recent trends in birth rates and regional migration.

In Health Systems Planning, DESTINY can project the requirement for health services personnel (nurses, physicians) and equipment (beds, CAT scanners), by geographic region.

In Social Services Planning, DESTINY can be used to forecast the levels of need for various social services (counseling, day care, protective services, chore services), the requirement for agency personnel (counselors) and for contract services, and the associated budget. These forecasts can be specified by geographic region, or broken down by age, sex, and race. The budget estimates may be disaggregated by service, by resource (e.g., counselor), or by demographic characteristics of the served population (e.g., age, sex, race, or region).

In Market Research, DESTINY can be used to estimate demand for products or services whose demand is related to demographic changes.

In Criminal Justice Planning, DESTINY can make projections of the prison inmate population, and thereby be used to assist planning for prison construction requirements under different sentencing policies.

The preceding are but a few examples of DESTINY applications. With DESTINY, the user can make forecasts such as the above quickly and easily. The system can be used to make forecasts at national, state, or local levels. The analyst can use DESTINY to answer a wide range of "what-if" type questions. DESTINY eliminates the analyst's dependence on "standard" populations projections that may be too highly aggregated, or may correspond to assumptions that are no longer reasonable. Detailed forecasts, corresponding to alternative demographic or programmatic assumptions, may be developed in a matter of minutes or hours, in most cases using information that is readily available from standard statistical sources.

The DESTINY Planning and Forecasting System is an essential technical resource for anyone engaged in planning who needs to be able to produce detailed demographic-based forecasts quickly and at reasonable cost.

This brochure describes the DESTINY system and illustrates its capabilities by means of several examples.

II. What DESTINY Does

DESTINY is a computer-program system that is used to forecast the general population, special "target" or "service" populations of interest, demand for services to these populations, and the amounts of resources (personnel, facilities, equipment, supplies) and cost required to provide these services to the target population. It is designed to provide fast, detailed forecasts in applications in which the items of interest (target populations, service populations, services, resources, and cost) vary roughly in proportion to the sizes of various segments of the general population, such as age/sex/race/geographic-region categories.

A major use of the DESTINY system is to support policy analysis and program evaluation in service-oriented fields, such as health care, social services, education, and public safety. DESTINY can be used to develop service forecasts and budget estimates under a variety of demographic and programmatic assumptions, and to generate detailed breakdowns (tables and crosstabulations) of these forecasts by age, sex, race, and geographic-region categories.

Forecasting models based on population projections are not new. In general, however, the amount of data and computation required to develop a detailed population-based forecast is substantial, since to achieve an acceptable level of accuracy the method requires estimates of the future population broken down by age, sex, race, and region. While a few total-population projections may be readily available from national or state-level agencies, it is usually not possible to obtain detailed projections (disaggregated by age, sex, race, or region) quickly, corresponding to arbitrary demographic assumptions, particularly at the local level. One of the major features of the DESTINY system is its ability to generate detailed population projections rapidly, corresponding to a wide range of demographic assumptions (concerning fertility, mortality, and migration).

Except in the field of population planning, most forecasting applications are not directly concerned with projections of the general population, but instead with special "target" subpopulations, such as the physically or mentally ill, the disabled, the student population, or the prison population. The DESTINY system can be used to provide forecasts of these special subpopulations, and forecasts of the resources and cost required to provide specified services to them. Furthermore, these forecasts can be broken down, or disaggregated, by detailed demographic characteristics -- by age, sex, race, and geographic region.

At the local level, planners are often confronted with the problem of estimating numbers of persons with certain characteristics (e.g., acute illnesses, chronic health conditions, disabilities, or social problems) or the number of population-related events (e.g., school or prison admissions), but there is usually no usable Census or sample survey information on incidences or prevalences of these characteristics at the local level. What is often available, however, are national or regional sample survey data that indicate the incidence or prevalence, broken down by detailed demographic categories, such as age, sex, and race. In the absence of local-level incidence/prevalence data, the planner must use this national or regional information to develop local-area estimates. In this situation, a recommended procedure for estimating the numbers of persons having the specified characteristics in the local area is synthetic estimation. With this procedure, the national incidences or prevalences for various age-by-sex-by-race categories are multiplied by the total numbers of persons of the general local-area population in these categories, to obtain estimates of the numbers having the specified characteristics in the local area. DESTINY has the capability for rapidly computing synthetic estimates. With this feature, the local planner can quickly construct reasonable target population estimates which make full use of available regional or national incidence/prevalence data.

While the DESTINY system can be quickly used to construct target-population estimates based on available incidence/prevalence rates (specific to age x sex x race categories), it can also be used in conjunction with other forecasting procedures to produce high-precision forecasts. For example, an analyst may use a multiple regression model, or an autoregressive integrated moving average (ARIMA or Box-Jenkins) time-series model, or a dynamic systems model to develop a precise forecast of a future incidence rate for a specific demographic category (e.g., arrest rates for males aged 20-35). This estimated arrest rate can then be input to the DESTINY system to forecast total number of arrests for a local area, taking into account not only trends in the arrest rate, but also anticipated trends in birth rates, death rates, migration, and population aging.

DESTINY has been designed to enable the planner to obtain forecasts of the general population, target populations, service populations, service needs, and the resources and cost required to provide these services. These projections can be made under a variety of "what-if" assumptions concerning demographic and programmatic conditions. The analyst can make alternative assumptions regarding population growth, target population incidences or prevalences, program service ratios, and costs. The corresponding projections can be computed quickly, and can be disaggregated by age, sex, race, and geographic region. DESTINY offers the user the ability to make detailed forecasts quickly and easily, at substantial savings over manual or partially-automated procedures.

III. How to Use the DESTINY System

The DESTINY system was designed to provide a substantial analytical capability to the user, but it was also designed with ease-of-use in mind. The DESTINY system works by setting up a mathematical representation, or "model," of the population, and using this model to project the future. To use the system, the user needs to set up a "parameter" file containing the following information:

Demographic Information

o Total Fertility Rate

o Fertility Age Distribution

o Infant Mortality Rate or Expectation of Life at Birth

o Base-year Population, by age and sex (race optional)

o External Migration Rates or Amounts

o Regional Populations (optional)

o Regional Migration Rates or Amounts (if regions are included in the model)

Target Population Information (Optional)

o Incidences (rates of occurrence of events) or Prevalences (proportions of the general population belonging to subpopulations of interest), either overall or by age and/or sex and/or race and/or region

Service Population Information (Optional)

o Service Ratios (proportions of the target populations that are served), either overall or by age and/or sex and/or race and/or region

Service Information (Optional)

o Average number of service units per year per case served

Resource Information (Optional)

o Average number of resource units per year required per service unit expended

Cost Information (Optional)

o Average cost per resource unit

Names

o The names of the races, regions, target populations, service populations, services, resources, and cost categories.

If only general population projections are desired, (i.e., no target population or service projections are desired), the user need specify only the demographic parameters. If a detailed program cost estimate is desired, all of the information must be specified. The demographic data are available from standard statistical publications, such as national and state vital statistics annual reports or statistical abstracts. The target population incidence/prevalence data are available from Census publications, national sample survey reports, agency publications, and statistical abstracts. The service, resource, and cost data are generally available from program administrative records.

The process of setting up the parameter file requires a little effort, but once the data are input to the computer, they are stored on disk and are easily updated.

While running the program, the user needs also to specify the following parameters:

o how many five-year periods to project

o what crosstabulations are desired for the various population, service, resource, and cost projections

o for which years hard-copy printout are desired

With regard to output, the program computes projections for the following items:

o the general population

o the specified target populations

o the specified service populations

o service units

o resource units

o cost

The program can generate aggregate projections for each of the above quantities, and will (at the user's option) provide disaggregated projections, by age and/or sex and/or race and/or region.

Version 1.0 of the DESTINY system is set up to accommodate up to three races, fourteen regions, four target populations, ten services, seven resources, and four cost categories.

IV. Special Features of the DESTINY System

The DESTINY system is designed to accommodate a wide range of detail in both the input and output. In setting up the parameter file, the user may specify that the same value of a demographic parameter (such as fertility rates, infant mortality rates, and migration rates) for the entire population for all future time. On the other hand, quite detailed demographic or programmatic specifications may be made, to obtain highly "conditioned" projections.

With regard to specifying incidence/prevalence rates for the target populations, DESTINY offers the user the choice of ten different demographic "stratifications." The user may specify that the same rate applies to the entire general population, or that the rates vary according to various demographic classifications, or stratifications, of the general population. Specifically, the user may specify rate stratifications by age, sex, race, age x sex, age x race, sex x race, age x sex x race, region, or race x region. With this flexibility, the analyst may make use of a wide variety of available crosstabulation data from national surveys, to construct synthetic estimates of target populations for the local area.

In technical terms, the DESTINY system uses the "cohort-component" method for making its population projections. This method is the most widely-used analytical method for preparing regional population projections. (Shryock, Siegel, and Associates, The Methods and Materials of Demography, US Government Printing Office, Washington, DC, 1980, presents a detailed description of this method.) For the target population estimates, the program uses the method of synthetic estimation (also known as cohort-component participation rates).

The system is set up to enable parameter input and updating interactively through the video terminal. The specified parameters are stored on disk for future use. The projections may be directed to the video screen, to a printer, or to a file.

The user may specify either brief aggregated projections, or detailed disaggregated projections, in the form of tables or crosstabulations by age, sex, race, or region. In addition to demographic crosstabulations (such as a breakdown of service utilization by race or region), the user may request crosstabulations of services by target population; resources by service, or target population; and cost by resource, service, or target population.

V. Examples

This section presents a number of examples of DESTINY applications. These examples include:

o Projection of the General Population

o Projection of Various Target or Service Populations

o School enrollments

o Persons with certain health or disability conditions

o The elderly in need of social services

o Prison admissions

o Projection of Service Needs

o Social Services

o Projection of Resource Requirements

o Teachers

o Short-term and long-term health care beds

o Social services counselors

o Projection of Program Costs

o Prison operating costs

o Social service program costs

The following examples illustrate the wide range of applications of the DESTINY system. They also illustrate the varying levels of detail which are possible. On the one hand, the user may use the program to construct a single estimate of the total population; on the other hand, the user may request a detailed breakdown of a target population (e.g., the blind) by age, sex, and race, or a breakdown of costs for purchased social services by service type or geographic region.

While the examples presented here illustrate the possible levels of detail of DESTINY projections, they do not represent examples of policy analysis exercises. In a real planning or policy analysis situation, the user would very likely make a set of DESTINY runs, under a wide range of alternative demographic and programmatic assumptions. DESTINY is ideally suited for such use, since it stored all of the input data in a "parameter" file, which the user may easily modify with the touch of a few buttons on the computer keyboard. Hence, while the first DESTINY run requires some effort (to assemble the needed data and set up the parameter file), successive runs are easily and quickly accomplished.

The examples are presented in order of increasing complexity. The examples and the principal features of each are listed below. The national projections are for the United States, and the state projections are for the State of Arizona. The examples are illustrated using actual program output listings (or extracts from listings). The projections are made from a "base year" of 1980 to the year 1990. The user may structure the program output in a variety of ways, requesting projections for various years and requesting different tables and crosstabulations for each year. These examples illustrate the wide variety of the program output.

Example 1. National Population Projection, Single-Race Model

Demographic representation: single race, single region

Service-system representation: none

Example 2. National Population Projection, Two-Race Model

Demographic representation: two races (white, other), single region

Service-system representation: none

Example 3. State Population Projection, Single-Race Model

Demographic representation: single race, single region

Service-system representation: none

Example 4. State Population Projection, Three-Race, 14-Region Model

Demographic representation: three races (white, Indian, other), 14 regions (counties)

Service-system representation: none

Example 5. Projection of the Hispanic Population

Demographic representation: two ethnic groups (Hispanic, nonHispanic), 14 regions (counties)

Service-system representation: none

Example 6. Rehabilitation Services, Projection of the Work-Disabled

Demographic representation: three races (white, Indian, other), 14 regions (counties)

Service-system representation: two target populations (severe and partially work-disabled)

Example 7. Education, Projection of School Enrollment

Demographic representation: three races (white, Indian, other), 14 regions (counties)

Service-system representation: one target population (students), one service population (elementary and secondary school students) one resource (teachers), one cost (teachers' salaries)

Example 8. Criminal Justice, Projection of Prison Admissions and Operating Cost

Demographic representation: three races (white, Indian, other), 14 regions (counties)

Service-system representation: one target population (admissions), one service population (sentence-years), one service (incarceration), one resource (cell), one cost (operating cost)

Example 9. Health Care, Projection of Short-term and Long-term Health Care Beds

Demographic representation: three races (white, Indian, other), 14 regions

Service-system representation: four target populations (beds in non-federal short-stay hospitals, nursing homes, and long-term care institutions for psychiatric and mentally handicapped patients)

Example 10. Projection of Social Services for the Elderly

Demographic representation: three races (white, Indian, other), 14 regions

Service-system representation: one target population (the elderly), seven services, three resources, and three costs

Example 1. National Population Projection by Age and Sex

This example illustrates the most basic application of DESTINY -- projection of a general population by age and sex. In this example, the resident population of the United States is projected from the base year of 1980 to the "projection year" of 1990. The program output includes all possible tables and crosstabulations involving age and sex: a table of the population by age, a table by sex, and a crosstabulation by age and sex.

The data on which this projection is based were taken from Statistical Abstract of the United States, 1981, US Government Printing Office, Washington, DC. The accuracy of the projection may be assessed by comparing the projected 1990 population estimates to the 1990 Census values (in Statistical Abstract of the United States, 1994). For example, DESTINY projected a resident US population of 250,026,462 for 1990; the 1990 Census value is 248,701,000. This is an error of about one-half of one percent. DESTINY projected 9,229,281 males under the age of five for 1990, compared to the 1990 Census value of 9,392,409 (an error of less than two percent).

Listing 1. PROJ Run for Example 1 (National Population Projection, Single-Race Model)

DESTINY PLANNING AND FORECASTING COMPUTER PROGRAM PACKAGE, VERSION 1.0

PROGRAM NAME: PROJ

DATE OF RUN (DD/MM/YYYY): 5/ 6/1995

TIME OF RUN (HH:MM:SS): 11:39: 2

PARAMETER FILE NAME: US801.DAT

GENERAL POPULATION DESCRIPTION:

UNITED STATES RESIDENT POPULATION

BASE YEAR: 1980

NO OF FIVE‑YEAR PERIODS TO PROJECT = 2

YEAR: 1990

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 250026462.

DISTRIBUTION OF POPULATN

BY AGE

0‑4 18150140.

5‑9 18340630.

10‑14 16961798.

15‑19 17341700.

20‑24 18916165.

25‑29 21907018.

30‑34 22038919.

35‑39 20150596.

40‑44 18063934.

45‑49 14262631.

50‑54 11763287.

55‑59 10935458.

60‑64 11141584.

65‑69 10411490.

70‑74 8156979.

75+ 11484134.

DISTRIBUTION OF POPULATN

BY SEX

MALE 122714472.

FEMALE 127311990.

CROSSTABULATION OF POPULATN

BY AGE AND SEX

MALE FEMALE

0‑4 9229281. 8920859.

5‑9 9321578. 9019052.

10‑14 8668490. 8293308.

15‑19 8858743. 8482957.

20‑24 9644578. 9271587.

25‑29 11108700. 10798317.

30‑34 11004406. 11034512.

35‑39 10003094. 10147502.

40‑44 8909808. 9154126.

45‑49 6987165. 7275467.

50‑54 5721375. 6041912.

55‑59 5251952. 5683506.

60‑64 5234019. 5907564.

65‑69 4745989. 5665501.

70‑74 3592512. 4564467.

75+ 4432781. 7051353.

Example 2. National Population Projection by Age, Sex, and Race

This example extends the population model detail to include representation of two races in the model -- white and other. In this case, the projection may be disaggregated by age, sex, race, age-by-sex, age-by-race, sex-by-race, and age-by-sex-by-race.

The projection is based on data from the 1981 edition of the Statistical Abstract of the United States, and may be compared to 1990 Census values in the 1994 edition. For example, DESTINY projects 197,443,844 whites and 53,294,702 others in 1990. The corresponding 1990 Census figures are 199,686,000 for whites and 49,024,000 for others -- errors of or -1.1% and 8.7%, respectively.

Listing 2. PROJ Run for Example 2 (National Population Projection, Two-Race Model)

DESTINY PLANNING AND FORECASTING COMPUTER PROGRAM PACKAGE, VERSION 1.0

PROGRAM NAME: PROJ

DATE OF RUN (DD/MM/YYYY): 5/ 6/1995

TIME OF RUN (HH:MM:SS): 12:17: 5

PARAMETER FILE NAME: US802.DAT

GENERAL POPULATION DESCRIPTION:

UNITED STATES RESIDENT POPULATION BY RACE (W/O)

BASE YEAR: 1980

NO OF FIVE‑YEAR PERIODS TO PROJECT = 2

YEAR: 1990

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 250738546.

DISTRIBUTION OF POPULATN

BY AGE

0‑4 18669617.

5‑9 18806048.

10‑14 17101239.

15‑19 17460667.

20‑24 19000300.

25‑29 21967328.

30‑34 22058888.

35‑39 20138388.

40‑44 18001916.

45‑49 14192279.

50‑54 11705971.

55‑59 10862325.

60‑64 11041520.

65‑69 10298760.

70‑74 8064410.

75+ 11368890.

DISTRIBUTION OF POPULATN

BY SEX

MALE 123028940.

FEMALE 127709605.

DISTRIBUTION OF POPULATN

BY RACE

WHITE 197443844.

OTHER 53294702.

CROSSTABULATION OF POPULATN

BY AGE AND SEX

MALE FEMALE

0‑4 9492151. 9177466.

5‑9 9556684. 9249364.

10‑14 8735127. 8366112.

15‑19 8915115. 8545552.

20‑24 9682667. 9317633.

25‑29 11135944. 10831384.

30‑34 11007309. 11051579.

35‑39 9985651. 10152737.

40‑44 8867066. 9134850.

45‑49 6942907. 7249371.

50‑54 5685799. 6020171.

55‑59 5209615. 5652710.

60‑64 5180874. 5860647.

65‑69 4692481. 5606280.

70‑74 3550310. 4514099.

75+ 4389239. 6979651.

CROSSTABULATION OF POPULATN

BY AGE AND RACE

WHITE OTHER

0‑4 13386385. 5283232.

5‑9 13621659. 5184390.

10‑14 12648435. 4452804.

15‑19 13055499. 4405168.

20‑24 14468046. 4532254.

25‑29 16941640. 5025688.

30‑34 17251096. 4807792.

35‑39 15930166. 4208222.

40‑44 14547949. 3453967.

45‑49 11604716. 2587563.

50‑54 9575698. 2130273.

55‑59 9022988. 1839336.

60‑64 9362446. 1679074.

65‑69 8906125. 1392635.

70‑74 7060025. 1004385.

75+ 10060972. 1307918.

CROSSTABULATION OF POPULATN

BY SEX AND RACE

MALE FEMALE

WHITE 97116269. 100327575.

OTHER 25912671. 27382030.

CROSSTABULATION OF POPULATN

BY AGE, SEX, AND RACE

WHITE

MALE FEMALE

0‑4 6808785. 6577600.

5‑9 6925470. 6696189.

10‑14 6486648. 6161786.

15‑19 6690797. 6364702.

20‑24 7400519. 7067527.

25‑29 8607548. 8334093.

30‑34 8649989. 8601107.

35‑39 7966863. 7963303.

40‑44 7239716. 7308232.

45‑49 5738483. 5866234.

50‑54 4701144. 4874554.

55‑59 4377281. 4645708.

60‑64 4439947. 4922499.

65‑69 4080277. 4825848.

70‑74 3124926. 3935098.

75+ 3877875. 6183096.

OTHER

MALE FEMALE

0‑4 2683366. 2599866.

5‑9 2631214. 2553175.

10‑14 2248479. 2204325.

15‑19 2224318. 2180851.

20‑24 2282148. 2250106.

25‑29 2528396. 2497292.

30‑34 2357320. 2450472.

35‑39 2018788. 2189434.

40‑44 1627350. 1826618.

45‑49 1204425. 1383138.

50‑54 984655. 1145617.

55‑59 832334. 1007002.

60‑64 740927. 938147.

65‑69 612203. 780432.

70‑74 425384. 579001.

75+ 511364. 796554.

Example 3. State Population Projection by Age and Sex

This example illustrates the use of DESTINY to make projections of the resident population of the State of Arizona from the base year of 1980 to 1990. For this projection, all of the population is combined, i.e., the model involves a single "race" (consisting of the total state population) and a single "region" (consisting of the entire state).

The data were obtained from the 1981 Statistical Abstract of the United States and Arizona 1980 Vital Health Statistics, and may be compared to 1990 Census values in the 1990 Census of Population or the 1994 Statistical Abstract. From the base year of 1980, DESTINY projects a state population of 3,694,625 for 1990, compared to the 1990 Census value of 3,665,000 -- and error of less than one percent.

Listing 3. PROJ Run for Example 3 (State Population Projection, Single-Race Model)

DESTINY PLANNING AND FORECASTING COMPUTER PROGRAM PACKAGE, VERSION 1.0

PROGRAM NAME: PROJ

DATE OF RUN (DD/MM/YYYY): 5/ 6/1995

TIME OF RUN (HH:MM:SS): 12:22:57

PARAMETER FILE NAME: AZ801.DAT

GENERAL POPULATION DESCRIPTION:

ARIZONA RESIDENT POPULATION

BASE YEAR: 1980

NO OF FIVE‑YEAR PERIODS TO PROJECT = 2

YEAR: 1990

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3694625.

DISTRIBUTION OF POPULATN

BY AGE

0‑4 308528.

5‑9 309834.

10‑14 265533.

15‑19 262196.

20‑24 272382.

25‑29 312138.

30‑34 326384.

35‑39 291639.

40‑44 255842.

45‑49 199194.

50‑54 163155.

55‑59 147819.

60‑64 144704.

65‑69 142762.

70‑74 121020.

75+ 171493.

DISTRIBUTION OF POPULATN

BY SEX

MALE 1827212.

FEMALE 1867413.

CROSSTABULATION OF POPULATN

BY AGE AND SEX

MALE FEMALE

0‑4 156914. 151614.

5‑9 157507. 152327.

10‑14 135307. 130226.

15‑19 133601. 128596.

20‑24 138207. 134175.

25‑29 158591. 153548.

30‑34 164894. 161490.

35‑39 146991. 144649.

40‑44 128908. 126934.

45‑49 97980. 101214.

50‑54 80273. 82882.

55‑59 71925. 75893.

60‑64 67382. 77322.

65‑69 63623. 79139.

70‑74 53404. 67616.

75+ 71705. 99788.

Example 4. Population Projection by Age, Sex, Race, and Region

This example illustrates application of DESTINY to project the Arizona population from 1980 to 1990, by age, sex, race, and region. For this example, three racial groups were selected -- white, American Indian, and other. The regions of the model are the 14 Arizona counties. Population data on these racial groups and counties is presented in the 1980 Census of Population, and vital statistics data are available from the 1991 Statistical Abstract of the United States and Arizona 1980 Vital Health Statistics.

This projection is based on the assumption that fertility and mortality rates continue at 1980 levels, and that the state experiences a net annual immigration of approximately 70,000 persons per year.

The projected values may be compared to 1990 Census figures presented in the 1990 Census of Population. For example, DESTINY projects a 1990 white population of 2,886,323 for 1990, compared to a 1990 Census value of 2,963,186 -- an error of -2.6%. The projected population of Maricopa County (where Phoenix is located) for 1990 is 1,893,992, compared to a 1990 Census value of 2,122,101 -- an error of -10.7%. DESTINY projects the 1990 Indian population of Pima County (where Tucson is located) as 21,069, compared to a 1990 Census value of 20,330 -- an error of 3.6%.

Listing 4. PROJ Run for Example 4 (State Population Projection, Three-Race, 14-Region Model)

DESTINY PLANNING AND FORECASTING COMPUTER PROGRAM PACKAGE, VERSION 1.0

PROGRAM NAME: PROJ

DATE OF RUN (DD/MM/YYYY): 5/ 6/1995

TIME OF RUN (HH:MM:SS): 12:53:32

PARAMETER FILE NAME: AZ803C.DAT

GENERAL POPULATION DESCRIPTION:

ARIZONA RESIDENT POPULATION BY COUNTY AND RACE (W/I/O)

BASE YEAR: 1980

NO OF FIVE‑YEAR PERIODS TO PROJECT = 2

YEAR: 1990

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3735219.

DISTRIBUTION OF POPULATN

BY RACE

WHITE 2886323.

AMERIND 228240.

OTHER 620656.

DISTRIBUTION OF POPULATN

BY REGION

APACHE 72497.

COCHISE 97683.

COCONINO 97171.

GILA 40759.

GRAHAM 26383.

GREENLEE 11039.

MARICOPA 1893992.

MOHAVE 83827.

NAVAJO 84127.

PIMA 649633.

PINAL 101590.

STA.CRUZ 24075.

YAVAPAI 95609.

YUMA 108068.

CROSSTABULATION OF POPULATN

BY REGION AND RACE

WHITE AMERIND OTHER

APACHE 15397. 55256. 1844.

COCHISE 78373. 692. 18618.

COCONINO 59977. 29599. 7595.

GILA 30676. 7197. 2886.

GRAHAM 17766. 3880. 4737.

GREENLEE 7876. 324. 2839.

MARICOPA 1582814. 32267. 278911.

MOHAVE 80313. 2070. 1444.

NAVAJO 34020. 45483. 4623.

PIMA 513643. 21069. 114920.

PINAL 57469. 12017. 32104.

STA.CRUZ 17931. 81. 6063.

YAVAPAI 91349. 1412. 2848.

YUMA 76044. 4582. 27442.

Example 5. Projection of the Hispanic Population

This example presents a projection of the population by Spanish origin ("Hispanic" vs. nonHispanic) and region (county), from 1980 to 1990. Data on Hispanic status by county are available from the 1980 Census of Population (or the County and City Data Book).

The projected values may be compared to 1990 Census figures presented in the 1990 Census of Population or the County and City Data Book 1994. From the base year of 1980, DESTINY projects 1990 Hispanic and nonHispanic populations of 647,662 and 3,054,470, compared to 1990 Census values of 688,388 and 2,976,840 -- errors of -5.9% and 2.6%, respectively.

(Note: The total-population estimate for 1990 for this example differs slightly from the total-population estimate for 1990 for Examples 3 and 4, since the demographic parameter specifications are not equivalent.)

Listing 5. PROJ Run for Example 5 (Projection of the Hispanic Population)

DESTINY PLANNING AND FORECASTING COMPUTER PROGRAM PACKAGE, VERSION 1.0

PROGRAM NAME: PROJ

DATE OF RUN (DD/MM/YYYY): 5/ 6/1995

TIME OF RUN (HH:MM:SS): 12:59:49

PARAMETER FILE NAME: AZ80HC.DAT

GENERAL POPULATION DESCRIPTION:

ARIZONA RESIDENT POPULATION BY HISPANIC STATUS AND REGION

BASE YEAR: 1980

NO OF FIVE‑YEAR PERIODS TO PROJECT = 2

YEAR: 1990

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3702132.

DISTRIBUTION OF POPULATN

BY RACE

HISPANIC 647662.

NON HISP 3054470.

DISTRIBUTION OF POPULATN

BY REGION

APACHE 65014.

COCHISE 98260.

COCONINO 92212.

GILA 38996.

GRAHAM 26019.

GREENLEE 11016.

MARICOPA 1871693.

MOHAVE 80664.

NAVAJO 76668.

PIMA 651822.

PINAL 101675.

STA.CRUZ 26575.

YAVAPAI 92273.

YUMA 111200.

CROSSTABULATION OF POPULATN

BY REGION AND RACE

HISPANIC NON HISP

APACHE 2867. 62147.

COCHISE 29228. 69032.

COCONINO 10332. 81880.

GILA 9098. 29898.

GRAHAM 6958. 19061.

GREENLEE 5633. 5383.

MARICOPA 281931. 1589762.

MOHAVE 3602. 77062.

NAVAJO 5936. 70732.

PIMA 154500. 497322.

PINAL 33174. 68501.

STA.CRUZ 20544. 6031.

YAVAPAI 6584. 85688.

YUMA 36480. 74720.

Example 6. Rehabilitation Services, Projection of the Work-Disabled

This example illustrates a typical use of the DESTINY system -- the estimation of a target population, based on available incidence or prevalence data. This example illustrates the application of the technique of "synthetic estimation," by which national incidence/prevalence rates by demographic category are used to construct state estimates. This procedure assumes, of course, that the national rates do in fact apply to the state.

The following table indicates the prevalence of work disability by age group (the prevalence does not vary markedly by sex or race). The table rates are proportions of the total resident population.

Persons with Work Disability (percent)

Age Severe Partial

18-24 2.1 3.9

25-34 2.9 7.0

35-44 6.3 8.3

45-54 11.0 12.4

55-64 24.2 11.8

Source: Derived from Table No. 555, Statistical Abstract of the United States, 1981. Data are for 1978. Source table figures adjusted from Civilian Non-institutional Population base to Resident Population base.

In the printout, the acronym WRKDISSV stands for "Severely Work-Disabled," and the acronym WRKDISPT stands for "Partially Work-Disabled."

Listing 6. PROJ Run for Example 6 (Rehabilitation Services, Projection of the Work-Disabled)

DESTINY PLANNING AND FORECASTING COMPUTER PROGRAM PACKAGE, VERSION 1.0

PROGRAM NAME: PROJ

DATE OF RUN (DD/MM/YYYY): 5/ 6/1995

TIME OF RUN (HH:MM:SS): 13:18: 6

PARAMETER FILE NAME: AZ803CD.DAT

GENERAL POPULATION DESCRIPTION:

ARIZONA RESIDENT POPULATION BY COUNTY AND RACE (W/I/O)

BASE YEAR: 1980

TARGET/SERVICE POPULATION DESCRIPTION:

WORK‑DISABLED POPULATION (SEVERELY AND PARTIALLY WORK‑DISABLED)

NO OF FIVE‑YEAR PERIODS TO PROJECT = 2

YEAR: 1990

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3735219.

DISTRIBUTION OF POPULATN

BY AGE

0‑4 338816.

5‑9 338489.

10‑14 268314.

15‑19 263875.

20‑24 272520.

25‑29 312009.

30‑34 326222.

35‑39 290907.

40‑44 253597.

45‑49 196686.

50‑54 161169.

55‑59 145678.

60‑64 142066.

65‑69 139466.

70‑74 117950.

75+ 167454.

DISTRIBUTIONAL ANALYSIS OF WRKDISSV

TOTAL WRKDISSV = 169750.

DISTRIBUTION OF WRKDISSV

BY AGE

0‑4 0.

5‑9 0.

10‑14 0.

15‑19 2217.

20‑24 5723.

25‑29 9048.

30‑34 9460.

35‑39 18327.

40‑44 15977.

45‑49 21635.

50‑54 17729.

55‑59 35254.

60‑64 34380.

65‑69 0.

70‑74 0.

75+ 0.

DISTRIBUTIONAL ANALYSIS OF WRKDISPT

TOTAL WRKDISPT = 182943.

DISTRIBUTION OF WRKDISPT

BY AGE

0‑4 0.

5‑9 0.

10‑14 0.

15‑19 4116.

20‑24 10628.

25‑29 21841.

30‑34 22836.

35‑39 24145.

40‑44 21049.

45‑49 24389.

50‑54 19985.

55‑59 17190.

60‑64 16764.

65‑69 0.

70‑74 0.

75+ 0.

TOTALS OF TARGET POPULATION(S)

WRKDISSV 169750.

WRKDISPT 182943.

Example 7. Education, Projection of School Enrollment

This example is similar to the preceding one, since it deals with the estimation of a specific target population -- in this case, school enrollments. This example carries the estimation process a little further, in also estimating the number of elementary and secondary school teachers. The number of teachers is estimated for 0-18 year olds, assuming a student/teacher ratio of 20. The total salary cost of the teachers is also estimated, assuming an average salary of $17,200 per year.

The following table presents the school enrollment rates for 1980, by age. The rates do not vary appreciably by sex or race.

Enrollment Rate

Age Noninst Civ Pop Base Res Pop Base

3-4 .367 .357

5-6 .957 .930

7-13 .993 .965

14-15 .982 .955

16-17 .890 .865

18-19 .464 .451

20-21 .310 .301

22-24 .163 .158

25-29 .093 .090

30-34 .064 .062

Source: Table No. 225, Statistical Abstract of the United States, 1981. Data are for 1980. Source table figures adjusted from Civilian Non-institutional Population base to Resident Population base. National data used as proxy for state data.

In the printout, the acronym ELEM/SEC stands for "Elementary and Secondary School Students." Also, in the distribution of teachers and teachers' salaries by age, "age" refers to age of students.

Listing 7. PROJ Run for Example 7 (Projection of School Enrollment)

DESTINY PLANNING AND FORECASTING COMPUTER PROGRAM PACKAGE, VERSION 1.0

PROGRAM NAME: PROJ

DATE OF RUN (DD/MM/YYYY): 5/ 6/1995

TIME OF RUN (HH:MM:SS): 13:25:44

PARAMETER FILE NAME: AZ803CE.DAT

GENERAL POPULATION DESCRIPTION:

ARIZONA RESIDENT POPULATION BY COUNTY AND RACE (W/I/O)

BASE YEAR: 1980

TARGET/SERVICE POPULATION DESCRIPTION:

ELEMENTARY & SECONDARY SCHOOL ENROLLMENT; TEACHERS; SALARIES

NO OF FIVE‑YEAR PERIODS TO PROJECT = 2

YEAR: 1990

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3735219.

DISTRIBUTION OF POPULATN

BY AGE

0‑4 338816.

5‑9 338489.

10‑14 268314.

15‑19 263875.

20‑24 272520.

25‑29 312009.

30‑34 326222.

35‑39 290907.

40‑44 253597.

45‑49 196686.

50‑54 161169.

55‑59 145678.

60‑64 142066.

65‑69 139466.

70‑74 117950.

75+ 167454.

DISTRIBUTION OF POPULATN

BY REGION

APACHE 72497.

COCHISE 97683.

COCONINO 97171.

GILA 40759.

GRAHAM 26383.

GREENLEE 11039.

MARICOPA 1893992.

MOHAVE 83827.

NAVAJO 84127.

PIMA 649633.

PINAL 101590.

STA.CRUZ 24075.

YAVAPAI 95609.

YUMA 108068.

DISTRIBUTIONAL ANALYSIS OF STUDENTS

TOTAL STUDENTS = 925175.

DISTRIBUTION OF STUDENTS

BY AGE

0‑4 48451.

5‑9 322241.

10‑14 258386.

15‑19 189199.

20‑24 58592.

25‑29 28081.

30‑34 20226.

35‑39 0.

40‑44 0.

45‑49 0.

50‑54 0.

55‑59 0.

60‑64 0.

65‑69 0.

70‑74 0.

75+ 0.

DISTRIBUTION OF STUDENTS

BY REGION

APACHE 22833.

COCHISE 23786.

COCONINO 26060.

GILA 10332.

GRAHAM 6812.

GREENLEE 2769.

MARICOPA 457268.

MOHAVE 19512.

NAVAJO 24616.

PIMA 159357.

PINAL 26920.

STA.CRUZ 5960.

YAVAPAI 22237.

YUMA 27238.

DISTRIBUTIONAL ANALYSIS OF ELEM/SEC

TOTAL ELEM/SEC = 794438.

DISTRIBUTION OF ELEM/SEC

BY AGE

0‑4 48451.

5‑9 322241.

10‑14 258386.

15‑19 165360.

20‑24 0.

25‑29 0.

30‑34 0.

35‑39 0.

40‑44 0.

45‑49 0.

50‑54 0.

55‑59 0.

60‑64 0.

65‑69 0.

70‑74 0.

75+ 0.

DISTRIBUTION OF ELEM/SEC

BY REGION

APACHE 19933.

COCHISE 20393.

COCONINO 22522.

GILA 8897.

GRAHAM 5865.

GREENLEE 2377.

MARICOPA 392016.

MOHAVE 16710.

NAVAJO 21394.

PIMA 136725.

PINAL 23184.

STA.CRUZ 5112.

YAVAPAI 19040.

YUMA 23393.

DISTRIBUTIONAL ANALYSIS OF TEACHING

TOTAL TEACHING = 794438.

DISTRIBUTION OF TEACHING

BY AGE

0‑4 48451.

5‑9 322241.

10‑14 258386.

15‑19 165360.

20‑24 0.

25‑29 0.

30‑34 0.

35‑39 0.

40‑44 0.

45‑49 0.

50‑54 0.

55‑59 0.

60‑64 0.

65‑69 0.

70‑74 0.

75+ 0.

DISTRIBUTION OF TEACHING

BY REGION

APACHE 19933.

COCHISE 20393.

COCONINO 22522.

GILA 8897.

GRAHAM 5865.

GREENLEE 2377.

MARICOPA 392016.

MOHAVE 16710.

NAVAJO 21394.

PIMA 136725.

PINAL 23184.

STA.CRUZ 5112.

YAVAPAI 19040.

YUMA 23393.

DISTRIBUTIONAL ANALYSIS OF TEACHERS

TOTAL TEACHERS = 39722.

DISTRIBUTION OF TEACHERS

BY AGE

0‑4 2423.

5‑9 16112.

10‑14 12919.

15‑19 8268.

20‑24 0.

25‑29 0.

30‑34 0.

35‑39 0.

40‑44 0.

45‑49 0.

50‑54 0.

55‑59 0.

60‑64 0.

65‑69 0.

70‑74 0.

75+ 0.

DISTRIBUTION OF TEACHERS

BY REGION

APACHE 997.

COCHISE 1020.

COCONINO 1126.

GILA 445.

GRAHAM 293.

GREENLEE 119.

MARICOPA 19601.

MOHAVE 836.

NAVAJO 1070.

PIMA 6836.

PINAL 1159.

STA.CRUZ 256.

YAVAPAI 952.

YUMA 1170.

DISTRIBUTIONAL ANALYSIS OF SALARIES

TOTAL SALARIES = 683216526.

DISTRIBUTION OF SALARIES

BY AGE

0‑4 41667598.

5‑9 277127660.

10‑14 222212066.

15‑19 142209203.

20‑24 0.

25‑29 0.

30‑34 0.

35‑39 0.

40‑44 0.

45‑49 0.

50‑54 0.

55‑59 0.

60‑64 0.

65‑69 0.

70‑74 0.

75+ 0.

DISTRIBUTION OF SALARIES

BY REGION

APACHE 17142535.

COCHISE 17538363.

COCONINO 19369276.

GILA 7650993.

GRAHAM 5044023.

GREENLEE 2044383.

MARICOPA 337133416.

MOHAVE 14370726.

NAVAJO 18399076.

PIMA 117583189.

PINAL 19938284.

STA.CRUZ 4396677.

YAVAPAI 16374272.

YUMA 20118045.

Example 8. Criminal Justice, Projection of Prison Admissions and Operating Cost

This example illustrates the use of DESTINY to predict the prison inmate population in 1990, from the base year of 1980. In 1979, the admissions rate and average times to parole eligibility were as follows. The admissions rate is calculated as a proportion of the total Arizona resident population.

Admission Rate

Age Male Female

15-19 .0020 .00020

20-24 .0042 .00019

25-29 .0029 .00024

30-34 .0021 .00016

35-39 .0013 .00010

40-44 .00082 .00006

45-49 .00070 .00006

50-54 .00060 .00003

55-59 .00026 .00001

60+ .00007 0

Average Time to Serve to Parole Eligibility

Male: 35.8 months

Female: 27.7 months

Source: Derived from Arizona Correctional Statistics, 1980 (admission numbers for 1980 were divided by the 1980 resident population for each age-by-sex category)

The per-inmate operating cost in Arizona is approximately $16,000 per year (source: Arizona Department of Corrections Information Office).

In the printout, the acronym ADMIS'NS denotes "Admissions," SENTYR denotes "sentence-years," CELL denotes "cell," and OP COST denotes the cost of care for inmates (exclusive of capital costs).

In an actual policy analysis, a number of runs similar to the example would be required. For example, to determine the effect of sentencing only violent criminals to prison, and at different sentence lengths from current practice, a run would be required which included not just one but several inmate ("target") populations, each representing a different level of violence and past behavior.

Listing 8. PROJ Run for Example 8 (Criminal Justice, Projection of Prison Admissions and Operating Cost)

DESTINY PLANNING AND FORECASTING COMPUTER PROGRAM PACKAGE, VERSION 1.0

PROGRAM NAME: PROJ

DATE OF RUN (DD/MM/YYYY): 5/ 6/1995

TIME OF RUN (HH:MM:SS): 13:29:55

PARAMETER FILE NAME: AZ803CP.DAT

GENERAL POPULATION DESCRIPTION:

ARIZONA RESIDENT POPULATION BY COUNTY AND RACE (W/I/O)

BASE YEAR: 1980

TARGET/SERVICE POPULATION DESCRIPTION:

PRISON ADMISSIONS; PRISON CELLS; OBLIGATED OPERATING COST

NO OF FIVE‑YEAR PERIODS TO PROJECT = 2

YEAR: 1990

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3735219.

DISTRIBUTION OF POPULATN

BY SEX

MALE 1849940.

FEMALE 1885279.

DISTRIBUTIONAL ANALYSIS OF ADMIS'NS

TOTAL ADMIS'NS = 2262.

DISTRIBUTION OF ADMIS'NS

BY SEX

MALE 2118.

FEMALE 145.

DISTRIBUTIONAL ANALYSIS OF SENTYR

TOTAL SENTYR = 6686.

DISTRIBUTION OF SENTYR

BY SEX

MALE 6353.

FEMALE 333.

DISTRIBUTIONAL ANALYSIS OF INCARC'N

TOTAL INCARC'N = 6686.

DISTRIBUTION OF INCARC'N

BY SEX

MALE 6353.

FEMALE 333.

DISTRIBUTIONAL ANALYSIS OF CELL

TOTAL CELL = 6686.

DISTRIBUTION OF CELL

BY SEX

MALE 6353.

FEMALE 333.

DISTRIBUTIONAL ANALYSIS OF OP COST

TOTAL OP COST = 106971223.

DISTRIBUTION OF OP COST

BY SEX

MALE 101648203.

FEMALE 5323019.

Example 9. Health Care, Projection of the Need for Short-term and Long-term Beds

This example shows a ten-year projection (by year) of bed needs for short-term care (less than 30 days) and for long-term care (30 days or more). This projection assumes the bed rates in the table presented below. The bed rates are expressed in terms of the ages of the bed users for short-stay hospital beds, and in terms of the ages of the residents for the long-term facility beds. The rates are relative to the resident population.

Long-term Care Facilities

Short-stay Nursing Home Mentally

Age Hospitals Male Female Psychiatric Handicapped

0-4 .00142 .00015 .00015 .00046 .00092

5-14 .00056 " " " "

15-19 .00175 " " " "

20-24 " .0012 .0012 .00024 .00095

25-34 .00242 " " " "

35-44 .00255 " " " "

45-64 .00428 " " " "

65-74 .01114 .011 .012 .00020 .00024

75+ " .063 .097 " "

Source: Derived from Tables 2, 29, 30, 77, 178, 179, 182, 183, 184, 185, 188, 530, and 597, Statistical Abstract of the United States, 1980. Rates are derived from age and facility distributional data from 1974 and 1976.

In the printout, ST BEDS refers to beds in short-stay hospitals, and the acronyms NH BEDS, PSY BEDS, and NH BEDS refer to beds in long-term care facilities (nursing homes, psychiatric facilities, and facilities for the mentally handicapped, respectively).

This example illustrates the application of DESTINY to address four different target populations (i.e., beds in the four types of facilities) simultaneously.

Listing 9. PROJ Run for Example 9 (Health Care, Projection of the Need for Short-Term and Long-Term Beds)

DESTINY PLANNING AND FORECASTING COMPUTER PROGRAM PACKAGE, VERSION 1.0

PROGRAM NAME: PROJ

DATE OF RUN (DD/MM/YYYY): 5/ 6/1995

TIME OF RUN (HH:MM:SS): 13:32:50

PARAMETER FILE NAME: AZ803CH.DAT

GENERAL POPULATION DESCRIPTION:

ARIZONA RESIDENT POPULATION BY COUNTY AND RACE (W/I/O)

BASE YEAR: 1980

TARGET/SERVICE POPULATION DESCRIPTION:

NEED FOR SHORT‑TERM & LONG‑TERM BEDS (LONG‑TERM: NURSING, PSYCH, MENTAL HAND.)

NO OF FIVE‑YEAR PERIODS TO PROJECT = 2

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 2718215.

TOTALS OF TARGET POPULATION(S)

ST BEDS 8883.

NH BEDS 13025.

PSY BEDS 837.

MH BEDS 2337.

YEAR: 1981

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 2819125.

TOTALS OF TARGET POPULATION(S)

ST BEDS 9219.

NH BEDS 13656.

PSY BEDS 867.

MH BEDS 2423.

YEAR: 1982

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 2920035.

TOTALS OF TARGET POPULATION(S)

ST BEDS 9555.

NH BEDS 14287.

PSY BEDS 897.

MH BEDS 2510.

YEAR: 1983

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3020944.

TOTALS OF TARGET POPULATION(S)

ST BEDS 9891.

NH BEDS 14918.

PSY BEDS 927.

MH BEDS 2596.

YEAR: 1984

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3121854.

TOTALS OF TARGET POPULATION(S)

ST BEDS 10227.

NH BEDS 15549.

PSY BEDS 957.

MH BEDS 2682.

YEAR: 1985

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3222764.

TOTALS OF TARGET POPULATION(S)

ST BEDS 10563.

NH BEDS 16180.

PSY BEDS 986.

MH BEDS 2768.

YEAR: 1986

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3325255.

TOTALS OF TARGET POPULATION(S)

ST BEDS 10889.

NH BEDS 16852.

PSY BEDS 1018.

MH BEDS 2857.

YEAR: 1987

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3427746.

TOTALS OF TARGET POPULATION(S)

ST BEDS 11214.

NH BEDS 17523.

PSY BEDS 1050.

MH BEDS 2945.

YEAR: 1988

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3530237.

TOTALS OF TARGET POPULATION(S)

ST BEDS 11539.

NH BEDS 18195.

PSY BEDS 1082.

MH BEDS 3034.

YEAR: 1989

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3632728.

TOTALS OF TARGET POPULATION(S)

ST BEDS 11864.

NH BEDS 18867.

PSY BEDS 1114.

MH BEDS 3122.

YEAR: 1990

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3735219.

TOTALS OF TARGET POPULATION(S)

ST BEDS 12189.

NH BEDS 19538.

PSY BEDS 1146.

MH BEDS 3211.

Example 10. Social Services: Forecasting the Counselors and Budget to Provide Certain Services to the Elderly Population

This example illustrates the use of the DESTINY program to project service levels, personnel, and budget levels required to provide social services to the elderly. The data used in this example are hypothetical. This example assumes that approximately five percent of the elderly population requires social services (specifically, 3% of those aged 65-69, 5% of those aged 70-74,and 7% of those aged 75+). The following tables indicate the average numbers of units of several types of service, resources, and cost required to provide social services to this served population.

Average Number of Service Units Per Case

Service No. of Units

S1. Counseling 2 hours

S2. Chore Services 16 dollars

S3. Homemaker Services 750 dollars

S4. Substitute Care 12 dollars

S5. Day Care 7 dollars

S6. Transportation 10 dollars

S7. Other 10 dollars

Average Number of Resource Units Per Service Unit

(Service types as defined in the preceding table)

No. of Resource Units per

Service Unit (by Service Type)

Resource S1 S2 S3 S4 S5 S6 S7

R1. Counselor 1 0 0 0 0 0 0

R2. Dollar (Purchase) 0 1 1 1 1 .5 .5

R3. Dollar (Payment) 0 0 0 0 0 .5 .5

Average Cost Per Resource Unit

(Resource types as defined in the preceding table)

Cost Per Resource Unit

(by Resource Type)

Cost Category R1 R2 R3

C1. Direct Service 15 0 0

C2. Purchased Service 0 1 0

C3. Payment 0 0 1

In the social services example presented here, we show a three-year-ahead projection, as opposed to the ten-year-ahead projections that were illustrated in the other examples.

The printout shows the services, resources, and budget required to provide services to the served population. In a real application, the service ratios (3%, 5%, and 7%) would be estimated from client caseload data, and the unit of service, unit of resource, and unit cost data would be estimated from program administrative records.

The DESTINY program could be used to estimate the changes in the budget that would occur if different service levels or different service costs were adopted. Used in this way, DESTINY is an ideal tool in the evaluation of alternative strategies for rationing social services.

The example illustrates several of the different types of tables and crosstabulations that can be constructed by the DESTINY program. The total number of possible crosstabs is very large, and typically only a few of them would be selected by the user for printout. Up to nine different tables or crosstabulations may be constructed for each of the many variables forecast by the program (general population, target population (the elderly), service population (those receiving social services), seven social services, three resources, and three cost categories). These include tables or crosstabs by age, sex, race, age by sex, age by race, sex by race, age by sex by race, region, and race by region. In addition, nine different service-system-related distributions or crosstabs may be specified:

o distribution of services by type

o distribution of resources by type

o distribution of costs by type

o distribution of services by served population

o distribution of resources by served population

o distribution of cost by served population

o distribution of resources by services

o distribution of costs by services

o distribution of costs by resources

In this example, we present only a small fraction of the possible output tables.

Data such as these may be used to back up legislative budget request, to prepare Comprehensive Annual Services Program plans, or to estimate staffing levels. Additional runs could reveal the budgetary impact of changes in service levels or costs.

Listing 10. PROJ Run for Example 10 (Social Services, Projection of Counselors and Budget Needed to Provide Social Services to the Elderly Population)

DESTINY PLANNING AND FORECASTING COMPUTER PROGRAM PACKAGE, VERSION 1.0

PROGRAM NAME: PROJ

DATE OF RUN (DD/MM/YYYY): 5/ 6/1995

TIME OF RUN (HH:MM:SS): 13:36:48

PARAMETER FILE NAME: AZ803CS.DAT

GENERAL POPULATION DESCRIPTION:

ARIZONA RESIDENT POPULATION BY COUNTY AND RACE (W/I/O)

BASE YEAR: 1980

TARGET/SERVICE POPULATION DESCRIPTION:

SOCIAL SERVICES FOR THE ELDERLY; 8 SERVICES, 3 RESOURCES, 3 COST CATEGORIES

NO OF FIVE‑YEAR PERIODS TO PROJECT = 1

YEAR: 1983

DISTRIBUTIONAL ANALYSIS OF POPULATN

TOTAL POPULATN = 3020944.

DISTRIBUTION OF POPULATN

BY SEX

MALE 1490114.

FEMALE 1530830.

DISTRIBUTION OF POPULATN

BY REGION

APACHE 56418.

COCHISE 86046.

COCONINO 78946.

GILA 37068.

GRAHAM 23057.

GREENLEE 10853.

MARICOPA 1564786.

MOHAVE 62054.

NAVAJO 70367.

PIMA 545662.

PINAL 90021.

STA.CRUZ 20673.

YAVAPAI 73847.

YUMA 91888.

DISTRIBUTIONAL ANALYSIS OF ELDERLY

TOTAL ELDERLY = 344542.

DISTRIBUTION OF ELDERLY

BY SEX

MALE 151885.

FEMALE 192658.

DISTRIBUTION OF ELDERLY

BY REGION

APACHE 3592.

COCHISE 10119.

COCONINO 7908.

GILA 4198.

GRAHAM 2481.

GREENLEE 1232.

MARICOPA 186166.

MOHAVE 7886.

NAVAJO 5903.

PIMA 63245.

PINAL 9106.

STA.CRUZ 2345.

YAVAPAI 9390.

YUMA 10114.

DISTRIBUTIONAL ANALYSIS OF ELDR(SV)

TOTAL ELDR(SV) = 17254.

DISTRIBUTION OF ELDR(SV)

BY SEX

MALE 7505.

FEMALE 9748.

DISTRIBUTION OF ELDR(SV)

BY REGION

APACHE 181.

COCHISE 507.

COCONINO 396.

GILA 210.

GRAHAM 124.

GREENLEE 62.

MARICOPA 9322.

MOHAVE 395.

NAVAJO 296.

PIMA 3167.

PINAL 456.

STA.CRUZ 117.

YAVAPAI 470.

YUMA 506.

DISTRIBUTIONAL ANALYSIS OF COUNSLNG

TOTAL COUNSLNG = 34507.

DISTRIBUTION OF COUNSLNG

BY SEX

MALE 15010.

FEMALE 19497.

DISTRIBUTION OF COUNSLNG

BY REGION

APACHE 361.

COCHISE 1013.

COCONINO 793.

GILA 421.

GRAHAM 248.

GREENLEE 123.

MARICOPA 18645.

MOHAVE 790.

NAVAJO 592.

PIMA 6334.

PINAL 911.

STA.CRUZ 235.

YAVAPAI 941.

YUMA 1012.

DISTRIBUTIONAL ANALYSIS OF CHORE SV

TOTAL CHORE SV = 276057.

DISTRIBUTION OF CHORE SV

BY SEX

MALE 120083.

FEMALE 155973.

DISTRIBUTION OF CHORE SV

BY REGION

APACHE 2890.

COCHISE 8106.

COCONINO 6343.

GILA 3366.

GRAHAM 1988.

GREENLEE 987.

MARICOPA 149160.

MOHAVE 6322.

NAVAJO 4740.

PIMA 50668.

PINAL 7292.

STA.CRUZ 1878.

YAVAPAI 7527.

YUMA 8100.

DISTRIBUTIONAL ANALYSIS OF HOMEMAKR

TOTAL HOMEMAKR = 12940160.

DISTRIBUTION OF HOMEMAKR

BY SEX

MALE 5628914.

FEMALE 7311246.

DISTRIBUTION OF HOMEMAKR

BY REGION

APACHE 135452.

COCHISE 379987.

COCONINO 297316.

GILA 157762.

GRAHAM 93181.

GREENLEE 46267.

MARICOPA 6991865.

MOHAVE 296350.

NAVAJO 222174.

PIMA 2375079.

PINAL 341809.

STA.CRUZ 88026.

YAVAPAI 352815.

YUMA 379681.

DISTRIBUTIONAL ANALYSIS OF SUBSCARE

TOTAL SUBSCARE = 207043.

DISTRIBUTION OF SUBSCARE

BY SEX

MALE 90063.

FEMALE 116980.

DISTRIBUTION OF SUBSCARE

BY REGION

APACHE 2167.

COCHISE 6080.

COCONINO 4757.

GILA 2524.

GRAHAM 1491.

GREENLEE 740.

MARICOPA 111870.

MOHAVE 4742.

NAVAJO 3555.

PIMA 38001.

PINAL 5469.

STA.CRUZ 1408.

YAVAPAI 5645.

YUMA 6075.

DISTRIBUTIONAL ANALYSIS OF DAY CARE

TOTAL DAY CARE = 120775.

DISTRIBUTION OF DAY CARE

BY SEX

MALE 52537.

FEMALE 68238.

DISTRIBUTION OF DAY CARE

BY REGION

APACHE 1264.

COCHISE 3547.

COCONINO 2775.

GILA 1472.

GRAHAM 870.

GREENLEE 432.

MARICOPA 65257.

MOHAVE 2766.

NAVAJO 2074.

PIMA 22167.

PINAL 3190.

STA.CRUZ 822.

YAVAPAI 3293.

YUMA 3544.

DISTRIBUTIONAL ANALYSIS OF TRANSPRT

TOTAL TRANSPRT = 172535.

DISTRIBUTION OF TRANSPRT

BY SEX

MALE 75052.

FEMALE 97483.

DISTRIBUTION OF TRANSPRT

BY REGION

APACHE 1806.

COCHISE 5066.

COCONINO 3964.

GILA 2103.

GRAHAM 1242.

GREENLEE 617.

MARICOPA 93225.

MOHAVE 3951.

NAVAJO 2962.

PIMA 31668.

PINAL 4557.

STA.CRUZ 1174.

YAVAPAI 4704.

YUMA 5062.

DISTRIBUTIONAL ANALYSIS OF OTHER

TOTAL OTHER = 172535.

DISTRIBUTION OF OTHER

BY SEX

MALE 75052.

FEMALE 97483.

DISTRIBUTION OF OTHER

BY REGION

APACHE 1806.

COCHISE 5066.

COCONINO 3964.

GILA 2103.

GRAHAM 1242.

GREENLEE 617.

MARICOPA 93225.

MOHAVE 3951.

NAVAJO 2962.

PIMA 31668.

PINAL 4557.

STA.CRUZ 1174.

YAVAPAI 4704.

YUMA 5062.

TOTALS OF SERVICE(S)

COUNSLNG 34507.

CHORE SV 276057.

HOMEMAKR 12940160.

SUBSCARE 207043.

DAY CARE 120775.

TRANSPRT 172535.

OTHER 172535.

DISTRIBUTION OF SERVICE(S)

BY SERVED POPULATION(S)

COUNSLNG CHORE SV HOMEMAKR SUBSCARE DAY CARE TRANSPRT OTHER

ELDR(SV) 34507. 276057. 12940160. 207043. 120775. 172535. 172535.

TOTAL 34507. 276057. 12940160. 207043. 120775. 172535. 172535.

DISTRIBUTIONAL ANALYSIS OF COUNSELR

TOTAL COUNSELR = 34507.

DISTRIBUTION OF COUNSELR

BY SEX

MALE 15010.

FEMALE 19497.

DISTRIBUTION OF COUNSELR

BY REGION

APACHE 361.

COCHISE 1013.

COCONINO 793.

GILA 421.

GRAHAM 248.

GREENLEE 123.

MARICOPA 18645.

MOHAVE 790.

NAVAJO 592.

PIMA 6334.

PINAL 911.

STA.CRUZ 235.

YAVAPAI 941.

YUMA 1012.

DISTRIBUTIONAL ANALYSIS OF PURCHSVC

TOTAL PURCHSVC = 13716570.

DISTRIBUTION OF PURCHSVC

BY SEX

MALE 5966649.

FEMALE 7749921.

DISTRIBUTION OF PURCHSVC

BY REGION

APACHE 143580.

COCHISE 402786.

COCONINO 315155.

GILA 167228.

GRAHAM 98772.

GREENLEE 49043.

MARICOPA 7411377.

MOHAVE 314131.

NAVAJO 235504.

PIMA 2517584.

PINAL 362317.

STA.CRUZ 93307.

YAVAPAI 373984.

YUMA 402462.

DISTRIBUTIONAL ANALYSIS OF PAYMENTS

TOTAL PAYMENTS = 172535.

DISTRIBUTION OF PAYMENTS

BY SEX

MALE 75052.

FEMALE 97483.

DISTRIBUTION OF PAYMENTS

BY REGION

APACHE 1806.

COCHISE 5066.

COCONINO 3964.

GILA 2103.

GRAHAM 1242.

GREENLEE 617.

MARICOPA 93225.

MOHAVE 3951.

NAVAJO 2962.

PIMA 31668.

PINAL 4557.

STA.CRUZ 1174.

YAVAPAI 4704.

YUMA 5062.

TOTALS OF RESOURCE(S)

COUNSELR 34507.

PURCHSVC 13716570.

PAYMENTS 172535.

DISTRIBUTION OF RESOURCE(S)

BY SERVED POPULATION(S)

COUNSELR PURCHSVC PAYMENTS

ELDR(SV) 34507. 13716570. 172535.

TOTAL 34507. 13716570. 172535.

DISTRIBUTION OF RESOURCE(S)

BY SERVICE(S)

COUNSELR PURCHSVC PAYMENTS

COUNSLNG 34507. 0. 0.

CHORE SV 0. 276057. 0.

HOMEMAKR 0. 12940160. 0.

SUBSCARE 0. 207043. 0.

DAY CARE 0. 120775. 0.

TRANSPRT 0. 86268. 86268.

OTHER 0. 86268. 86268.

TOTAL 34507. 13716570. 172535.

DISTRIBUTIONAL ANALYSIS OF DIRCTSV$

TOTAL DIRCTSV$ = 517606.

DISTRIBUTION OF DIRCTSV$

BY SEX

MALE 225157.

FEMALE 292450.

DISTRIBUTION OF DIRCTSV$

BY REGION

APACHE 5418.

COCHISE 15199.

COCONINO 11893.

GILA 6310.

GRAHAM 3727.

GREENLEE 1851.

MARICOPA 279675.

MOHAVE 11854.

NAVAJO 8887.

PIMA 95003.

PINAL 13672.

STA.CRUZ 3521.

YAVAPAI 14113.

YUMA 15187.

DISTRIBUTIONAL ANALYSIS OF PURCHSV$

TOTAL PURCHSV$ = 13716570.

DISTRIBUTION OF PURCHSV$

BY SEX

MALE 5966649.

FEMALE 7749921.

DISTRIBUTION OF PURCHSV$

BY REGION

APACHE 143580.

COCHISE 402786.

COCONINO 315155.

GILA 167228.

GRAHAM 98772.

GREENLEE 49043.

MARICOPA 7411377.

MOHAVE 314131.

NAVAJO 235504.

PIMA 2517584.

PINAL 362317.

STA.CRUZ 93307.

YAVAPAI 373984.

YUMA 402462.

DISTRIBUTIONAL ANALYSIS OF PAYMENT$

TOTAL PAYMENT$ = 172535.

DISTRIBUTION OF PAYMENT$

BY SEX

MALE 75052.

FEMALE 97483.

DISTRIBUTION OF PAYMENT$

BY REGION

APACHE 1806.

COCHISE 5066.

COCONINO 3964.

GILA 2103.

GRAHAM 1242.

GREENLEE 617.

MARICOPA 93225.

MOHAVE 3951.

NAVAJO 2962.

PIMA 31668.

PINAL 4557.

STA.CRUZ 1174.

YAVAPAI 4704.

YUMA 5062.

TOTALS OF COST CATEGORY(IES)

DIRCTSV$ 517606.

PURCHSV$ 13716570.

PAYMENT$ 172535.

DISTRIBUTION OF COST CATEGORY(IES)

BY SERVED POPULATION(S)

DIRCTSV$ PURCHSV$ PAYMENT$

ELDR(SV) 517606. 13716570. 172535.

TOTAL 517606. 13716570. 172535.

DISTRIBUTION OF COST CATEGORY(IES)

BY SERVICE(S)

DIRCTSV$ PURCHSV$ PAYMENT$

COUNSLNG 517606. 0. 0.

CHORE SV 0. 276057. 0.

HOMEMAKR 0. 12940160. 0.

SUBSCARE 0. 207043. 0.

DAY CARE 0. 120775. 0.

TRANSPRT 0. 86268. 86268.

OTHER 0. 86268. 86268.

TOTAL 517606. 13716570. 172535.

DISTRIBUTION OF COST CATEGORY(IES)

BY RESOURCE(S)

DIRCTSV$ PURCHSV$ PAYMENT$

COUNSELR 517606. 0. 0.

PURCHSVC 0. 13716570. 0.

PAYMENTS 0. 0. 172535.

TOTAL 517606. 13716570. 172535.

Cost of Using DESTINY

The cost of making a forecast includes three components -- the analyst time used in assembling the required data and in setting up the run (data entry), the computer running cost, and the program use charges (purchase or lease). For the DESTINY runs presented here, the analyst time varied from a few minutes to several hours. (These estimates assume that census and vital-statistics publications are readily available; otherwise, delays will be incurred in obtaining them.) The several-hour run was the first one, in which the base-year population data and other demographic parameters had to be collected. The health-care example required about an hour to extract the bed-use rate by age from several Statistical Abstract tables, none of which contained the information in the desired form. The other examples generally required only a few minutes to assemble the required data.

DESTINY runs are very fast, so that the amount of microcomputer time involved in making the projections is small. Apart from initial data entry, the major component of computer time involves examining different model specifications (changes in structural parameters or parameter values).

The major cost component of DESTINY application is the cost of the analyst's time in assembling the required data and entering it into the system. Compared to data-intensive methodologies such as microsimulation, the cost of analyst time would be low.

Joseph George Caldwell, PhD, developer of the DESTINY system, is a consultant specializing in program planning, evaluation, and policy analysis. His consulting career has included much work in studies, analysis, and system development in these disciplines, in a variety of application areas.

As president of Vista Research Corporation, he directed the project to develop the prototype MICROSIM microsimulation forecasting program for the US Department of Health and Human Services, and the Economic and Social Impact Analysis / Women in Development project for the Government of the Philippines.

He conducted needs assessment and client satisfaction surveys under the project, “Measuring the Effectiveness of Social Services,” for the State of West Virginia. He developed the sampling plans for a number of major nationwide surveys, such as the Survey of Institutionalized Persons, the Elementary and Secondary School Civil Rights survey, the National Center for Health Services Research (NCHSR) Survey of Hospital Costs and Utilization, the Profession Services Review Organization (PSRO) Data Base Development Study, the US Department of Housing and Urban Development (HUD) Housing Market Practices Study, and the Vocational Rehabilitation Follow-up Study.

He developed the sampling manuals for a number of government statistical reporting systems, including the Social Services Reporting Requirements (SSRR), the Office of Child Support Enforcement Reporting Requirements, and the Utilization Review of Medicaid; conducted cost-benefit studies of day care and alcoholism treatment centers; directed the Health Care Financing Administration (HCFA) study of the cost impact of Skilled Nursing Facility / Intermediate Care Facility (SNF/ICF) standards on nursing homes; and developed improved matching/allocation formulas for distributing funds to states under the Medicaid, Aid to Families with Dependent Children (AFDC), and Vocational Rehabilitation programs. He served as manager of monitoring and evaluation for the Egypt Local Development II – Provincial project, the largest local-level infrastructure development project in the world funded by the US Agency for International Development.

In the field of information technology, he developed the Personnel Management Information System for the Government of Malawi civil service, and the data processing and reporting component of the EdAssist Education Management Information System used by the Government of Zambia Ministry of Education (Academy for Educational Development / US Agency for International Development / Zambia Ministry of Education).

He developed TIMES, the first commercially-available general-purpose Box-Jenkins time series forecasting program, and the popular short course, “Sample Survey Design and Analysis.”

He is the author of several books, including The Value-Added Tax: A New Tax System for the United States and Can America Survive?

Dr. Caldwell received his BS degree in mathematics from Carnegie-Mellon University and his PhD degree in mathematical statistics from the University of North Carolina at Chapel Hill. He previously served as Director of Management Systems for the Bank of Botswana; as Manager of Research and Development and Principal Scientist at the US Army Electronic Proving Ground’s Electromagnetic Environmental Test Facility; and as Adjunct Professor of Statistics at the University of Arizona in Tucson, Arizona.