Temple Beth El Forecast to the Year 2020

by

Bloom Forecasting, Inc.

October, 1996

 

Foreward

 

            Temple Beth El is currently faced with the task of selecting a new rabbi to replace Rabbi Richard Rosenthal, who will be retiring at the end of Spring, 1997.  As part of the interview process, the Board of Temple Beth El desired to put together a package of information which would serve as a common body of knowledge for Temple members and for prospective candidates for the job of rabbi.   A survey of Temple members' attitudes and concerns was carried out by Dr. Linda Elman and results distributed in September, 1996.  In addition, the Board wanted to have a forecast of Temple membership beyond the year 2000 in order to provide information to the prospective candidates and, more importantly, to be able to formulate a long-range picture for the future of the Temple.  To this end, the Board contacted the author who agreed to undertake the task because, as a long-time member of Temple himself, he was also interested in seeing what the future held in store for the membership of Temple Beth El.  What the reader will see in this report is, in addition to the Temple membership forecast, important past and present Temple demographic data plus projections of Temple membership life cycle events to the year 2020.    The information presented here should enlarge the perspective of the Board in formulating plans and strategies for the next decade.

 

 

Acknowledgments

 

            I wish to thank the many members of Temple who have helped me produce this forecast.  Alan Steinman and Bruce Mann are credited with suggesting the idea of the forecast and, together with Herman Kleiner, provided valuable reviews of the final report.  Thanks to Maggi Barrett and Iris Sigel for providing much of the data about current and past Temple membership.     

 

            Singled out also for thanks are the people who participated in the judgmental forecast of future Temple membership:  Les Rosenthal, Herman Kleiner, Irene and Sol Feder, Bruce Mann, Joan Garden and Tova Ahdut.  While receiving suggestions and ideas from these Temple members and others, responsibility for the forecasts  rests solely with the author.

 

            Ideas in the section, Alternative Scenarios, were provided by the following persons as a result of individual telephone and face-to-face interviews: Bruce Mann, Allan Steinman, Hilde Slotnick, Randy Leifson, Tova Ahdut, Barbara Kleiner and Herman Kleiner.   I have tried to faithfully record their fears and hopes regarding the future of Temple Beth El.   I apologize in advance for any omissions or misinterpretations of their thoughts. 

 

Mitchel F. Bloom, Ph.D., C.F.P.

Tacoma, Washington

October, 1996

 

1. Summary

 

            A forecast of Temple Beth El households was made for years 2000 to 2020 at five year intervals.  Because forecasting is a statistical science, the procedure used was to make as many simple forecasts as possible (the maximum number that the author could generate) using all of the available data and then to average the results.  Eleven separate forecasts of households were developed, each forecast using either a different technique, and/or different assumptions, and/or a different dataset.  

 

            This approach to averaging the many simple forecasts is referred to as the "amalgamated method."  It can be likened to a shotgun approach to the target, many forecasts surrounding what is hoped to be the "true" value, as opposed to a high-powered, telescopic rifle approach, a single forecast deriving from a complex computer model that could be entirely wrong because of an incorrect core assumption.

 

 

 

 

 

            The final Temple Beth El forecast, therefore, was an amalgamation of eleven different forecasts using three separate ways of amalgamation .  These were:  1) weighting each of the eleven forecasts equally and 2) weighting each forecast judgmental and then calculating the weighted average and 3) calculating the median of the eleven forecasts.  The results of these three amalgamated methods were given equal weights and averaged once again for the forecasting period covering years 2000 through 2020.  This resulted in the final forecast for Temple household membership, shown in Table 1 and Figure 1.

 

            From the data in Table 1 and Figure 1, the Temple household membership is projected (medium forecast) to increase from 344 families in 1995 to 509 families in year 2020, representing an increase of 1.9% per year on the average.  In the same period, the households of Pierce County are projected to increase by average of 1.8% per year.  The 95% confidence interval projects that Temple membership should be between 363 and 394 households by year 2000 and between 471 and 558 households by year 2020. 

 

            Using past Temple membership records, it was possible, from a data standpoint, to do a five year retrospective forecast for only eight of the eleven methods.  The average error was found to be 5.3% for the eight methods. 

 

            It should be noted that in 1996 Temple households had fallen to 330 from a high of 344 in 1995.  But fluctuations are a normal condition of any long term trend.  In 1989 Temple households also decreased from the 1988 level.    

 

            Additional forecasts were made  for the period 2000-2020 of the total population contained in the member households, the number of persons in each age group, geographical distribution of temple members, births, brits and baby namings, bar and bat mitzvahs, children in the Sunday school, and deaths.  These latter forecasts were performed for the medium forecast only (no 95% confidence intervals) assuming that the age group distribution of 1995 remains the same over the forecast period and that estimates of current ages of members are reasonably accurate. To the extent that these assumptions and others to be discussed in the body of the report are in error, the forecasts will be in error. 

 

            In the process of producing the age related forecasts mentioned in the previous paragraph, the author discovered that Temple Beth El's households are missing many families in the 20 to 39 year age group.  Using knowledge of Pierce County age group percentages,  calculation was made of the number of missing families in 1996, referred to as Phantom Jews, and of what the population of Temple Beth El would have been had these families been members.  The figure arrived at is that in 1996  if all of the phantoms had joined, there would have been at least 122 more member families in Temple. 

 

 

2. Introduction

 

            This report will begin with identifying the Goals of the Forecast. It will be followed by the section, Data Used in the Forecast, which includes past Temple household memberships, current age group populations, a description of Pierce County data and relevant vital statistics, finally followed by the general method of forecasting Temple member households.

 

            In the next section, The Household Forecasts, the results of each of the eleven household forecasting methods will be presented and discussed.  For eight of the methods where it was possible to calculate an error, a five-year, retrospective, percent error was calculated.  Of special interest is the last of the eleven methods to forecasts households: it projects the number of Temple households in various geographical areas of the city to the year 2020 and identifies the fastest growing areas by ZIP codes.

 

            The next section, Combining the Household Forecasts, describes the three ways of combining the eleven forecasts and the result of averaging the three average forecasts into an overall final forecast.  In addition to the forecast itself, an estimate of the range of the forecast is desired.  The 95% confidence intervals for the years 2000 through 2020 every fifth year were presented.  The confidence intervals provide an upper and lower boundary to the forecast, an important piece of information for long range planning.

 

            The seventh section, Phantom Jews, makes a calculation of the minimum number of Jewish households in 1995 residing in Pierce County but who are not  members of Temple Beth El.  Many people have speculated on this number but, to the author's knowledge, it is the first calculation made on the basis of statistical data.

 

            The next section, Extended Forecasts, considers the consequences of the household forecast on a large group of other important factors.  These include important life events ranging from births to bar mitzvahs to deaths plus the numbers of pupils in Sunday school and total members in Temple.  Again, the methods, assumptions, and resulting projections are made for the years 2000-2020. 

 

            How would the forecast change if the assumptions are incorrect because of unforseen events?  The section, Alternative Scenarios, considers possible surprises, i.e., low probability, high impact events or developments, which could change the picture of the future presented by the forecast.  For each event or development the impact on the number of Temple households will be discussed. 

 

            Finally, the section, Conclusions, will summarize the major results of the forecast for the Board as it interviews candidates and begins to plan for Temple Beth El's future.

 

 

 

 

 

3. Goals of the Forecast

 

 

            Originally, the author was asked only to forecast the future membership of Temple, the household forecast, with no specific target date.  It was apparent that once this was done, it would be helpful for planning purposes to use that information to project other variables, e.g., births, bar mitzvahs, Sunday school enrollment,  and even where in the area Temple members would reside.  These variables are referred to as the extended forecast.   But  first it was necessary to obtain the current age group distribution of the Temple membership population.  With the aid of volunteers, the age group distribution was obtained by using admitted ages, where available, and estimated ages, where they were not available.  Along with the household forecast, the age group distribution was used to project the Temple population by age group for the years, 2000 through 2020, for every fifth year. 

 

            The extended forecast naturally followed from the age group distribution forecast.  This included births, brits, bar mitzvahs, Sunday school class sizes, and deaths,  all of which result from age group distributions plus assumptions concerning birth, migration, and death rates. 

 

            In addition to the extended forecast, it was desired to produce an estimate of the number of Jewish families residing in Pierce County who currently are not members of Temple but who might become members of Temple. 

 

            Finally, using Temple telephone directory information, it was also possible to identify the current location of Temple member households according to ZIP codes in the Tacoma and surrounding area.  With the current location, plus the household forecast and Puget Sound Regional Council forecasts of the growth of the ZIP code areas, the future geographical distribution of Temple members was projected.        

 

            What started out to be a simple goal of projecting households was expanded into a much larger effort, projecting a host of key variables about the Temple membership.  These additional forecasts should provide the Board and the Rabbi Search Committee with valuable information to accomplish their tasks as they move into the 21st century.

 

4. Data Used in the Forecast

4.1 Introduction

 

            Four types of data were required to produce the forecasts in this report.  First, Temple data included past Temple households, current population of age groups, and zip codes of households.  Second, The Pierce County Department of Planning and Land Use provided data on population by geographical area and regional household data through the Puget Sound Regional Council.   PSRC also furnished the geographic key relating ZIP codes to FAZs (forecast analysis zones) and to TAZs (transit area zones).  Third, the Washington State Office of Financial Management (OFM) supplied components of population change by county and by age group to the year 2020.  Finally, statistics on birth and death rates by age group were obtained from the U.S. Statistical Abstract.

 

4.1 Temple Beth El Data [1]

 

            One of the first tasks in constructing a forecast is to determine what data is available.  Temple Beth El's current database contains a list of all current Temple households, including children under 18 years of age, addresses and ages of members (but not in all cases).  The database goes back only one year.  Before 1996, very few records on Temple membership are kept.  To determine how many households (read "families") were Temple members in past years, the author had to search the voluminous three-ring binders of Board meetings to find the report of the Membership Committee chairman.  Often the report simply stated the increase or decrease in families over previous years but did not reveal the number of member families.  In some years no membership report could be found. Other sources of Temple membership were past Temple telephone directories where the Temple member families are indicated.  As a result of all this effort,  the author found Temple membership data for only eight years, spanning the 1979 through 1996 period.  The results are shown below in Table 2 and Figure 2.  County households are also shown for comparison.

 

 

           

 

            The age group data for the 1996 Temple membership was obtained from the computerized database and from estimates of a long-time member of Temple Beth El.  The estimates assigned individuals for whom no birthdates were available to five-year age groups, e.g., 30-34, 35-39, etc.  They may be in error but probably by not much.  The current age group distribution for Temple Beth El is shown in Table 3 and Figure 3.  Note that the age group 0-4 is much less percentagewise than groups 5-9, 10-14 and 15-19.  Moreover, age groups from 20-24 through 35-39 are significantly underrepresented compared to age groups 40-44 through 50-54 even taking the post World War II baby boom into account.    

                                              

           

 

 

4.2 Washington State OFM Data [2]

 

            The OFM provided the past and future populations of Pierce County by sex, age group and year, at five year intervals, from years 1990 to 2020.  Three different projections were made for Pierce County, a high, a medium and a low series.  Births, deaths, migration rates and total fertility rates for the County were also given by OFM for five year intervals over the forecast period.

 

4.3 Puget Sound Regional Council Data [3]

 

            PSRC supplied FAZ (forecast area zone) projections of subareas of the Pierce County from 1970 to the year 2020 in ten year intervals.  Data included total population and average household size.  PSRC also faxed the author a table relating the FAZs to the TAZs (transit area zones) and the ZIP codes to the TAZs, which was needed to forecast the geographical distribution of Temple households.

 

4.4 National Data [4]   

 

            Vital statistics information was found in the Statistical Abstract of the United States: 1995.  This included: death rates by race, age and sex and projected fertility rates by race and age group.

 

4.5  Comment on Data

 

            The rule followed in acquiring and using data when making a local area forecast is to first use any local (small area, city, county) data.  If local data are unavailable, then proceed to use regional or state data and scale down to the local area.  If the latter are unavailable, then go to national data and make the assumption that the local area is similar to the nation for that particular variable.  Other techniques include using adjacent local area data if one's own local area data is unknown.  

 

5. The Household Forecasts

 

5.1 Overview

 

            Temple households were forecast using eleven separate models, each of which was based on different datasets, assumptions or methods.  Forecasts 1 through 3 were based on population projections from a demographic model made by the Office of Financial Management (OFM), State of Washington.  Forecasts 4 and 5 were based on an economic model projecting population in the region developed by the Puget Sound Regional Council (PSRC).  Forecast 6 was the product of running a judgmental forecast using a group of Temple members.  Forecasts 7 and 8 were obtained by two different methods of extrapolation using a commercially available forecasting package, Forecast Pro.  Forecasts 9 and 10 resulted from two different ways of moving the population forward in time, allowing for births, deaths and migration.  Finally, forecast 11 was based on a the growth of small geographical areas within metropolitan Tacoma in which Temple members currently reside.

 

5.2 Household Forecasts Based on OFM and PSRC Projections

 

            Both the OFM and the PSRC-based forecasts used two important ratios that had to be calculated for Temple Beth El: 1) persons per household and 2) Temple to Pierce County households.  Persons per household has been generally falling in the United States for the entire century due to a declining birth rate, the desire to have smaller families, and the advent of birth control.  On average, Temple households are slightly smaller than Pierce County households.  On the basis of the past trend and for purposes of forecasting, the ratio of Temple to County pph was set equal to 92% out to the year 2020.  The pph for Temple Beth El (see Figure 4) is projected to decline from 2.40 in 1995 to 2.29 by 2010 and to remain at that level to year 2020.

 

            The second important ratio, Temple to Pierce County households, was found to be a fluctuating number from .00137 to .00151 depending on the year.  An average was calculated at .00143 as shown in Figure 5.

 

 

                        Using the projections of OFM and PSRC in addition to the two ratios of pph and Temple to Pierce County households, the Temple household forecasts 1 through 5 were obtained.  Five year retrospective forecast errors of the methods are displayed in the right-hand column.  The results are summarized in Table 4.

 

 

5.3 Household Forecasts Based on Judgmental Forecasts

 

            It has been the experience of the author that group judgmental estimates are as accurate or more accurate than mathematical computer model estimates.  Therefore, a small group of Temple members consisting of seven persons assembled at Temple for the purpose of producing a retrospective and a future ETE (estimate-talk-estimate) forecast.  In the ETE forecast, a facilitator (the author) provided the group with certain trend data from the past; required each participant to make his own estimate of the future value of the trend for years (e.g., 2000 to 2020 at five year intervals); allowed participants to use any method they wish to do the forecast; requested from each participant his/her forecast; fed back to the group the median and the middle 50% range of answers; chaired a group talk session in which the forecasts were discussed; and, lastly, asked to group participants to make a second and final estimate of the trend.  The final forecast from the group is an average of the median and the mean of the group from the second estimate.

 

            To determine the accuracy of the group, a retrospective forecast was performed on the group before beginning the future forecast.  For five years in the past, corresponding to the times of the past Temple Beth El household data, the group was provided with household numbers from an unnamed Temple located in an unnamed county.  Unbeknownst to the group, the data given to them were from Temple Beth El and Pierce County, but multiplied by 0.75.  They were asked to forecast the number of households for that unnamed temple five years into the future which actually was 1996 Temple households multiplied by 0.75.  The error of the group mean estimate was 3.8%, next to the lowest of the errors of all the methods for which errors were obtained.  As a result of this technique, a measure of the reliability of the group's accuracy was obtained for forecastingTemple Beth El households.

 

            The future ETE forecast was then conducted.  The group was given past household data of Temple Beth and Pierce County, the pph chart, and the ratio of Temple households to Pierce County households (Figures 2, 4 and 5, respectively).  The group, consisting of seven Temple members, then went through the ETE process as discussed previously and arrived at the results shown in Table 5.

                                   

             

 

5.4  Extrapolative Methods

 

            Two extrapolative methods were used to forecast the number of households.  The methods employed the commercial forecasting program, Forecast Pro.  Forecast Pro is a sophisticated forecasting package that looks at a past trend; examines any cycles, regularities and perturbations in the trend; automatically selects the best model based on how well the model forecasts the known data; then implements that model to forecast the future of the trend together with a confidence interval specified by the user.

           

            The extrapolative model selected by Forecast Pro was exponential smoothing with trend extrapolation.  Exponential smoothing is a very good method for forecasting the near term future for trends without apparent regularities or cycles.  It is not so good for long range forecasts  -- but it should be added that no single mathematical technique is good for long-range forecasts which is why the author used the amalgamated method of combining many simple forecasts.

           

            The logic behind exponential smoothing is that the more distant the data point is, the less weight it receives in extending the trend into the future.  The weights can be assigned by the computer to achieve the best fit with the known actual data or they may be assigned by the user.   The weights, or coefficients as referred to in the program, are assigned for both the level of the trend and the slope (rate of change of the trend).  The higher the coefficient, the more the influence of the more recent past and the less the influence of the more distant past on the forecast.   

 

            The results are shown in Table 6.  The main difference between the two methods is that in the automatic method the computer decided on the coefficients while in the manual method the author decided on the coefficients.  Observe that both of these forecasts project the number of households to be below 500 families, the lowest two of the eleven forecasts.  This result is consistent with the usual exponential smoothing forecasts over long time periods, they err on the low side.  The five-year forecast error based on a retrospective forecast was 8.9% for the automatic and 7.7% for the manual method.

 

 

 

5.5 Methods Involving Moving the Population Forward in Time

 

5.5.1 Method of Moving Age Groups  

 

            Two methods were used produce forecasts based on moving the age groups of the Temple population forward in time in five year increments.  The first method calculated what the Temple population in 1995 would have been if the age groups were approximately the same as Pierce County. This is referred to as the potential population.  The potential population differs from the actual Temple population by being a significantly larger percentage of the total population in the following five age groups out of the 18 age groups: 0-4, 20-24, 25-29, 30-34, and 35-39.  Once the potential Temple population for 1995 was determined, the age groups were moved forward in time in five year increments.   The method used the following equation, which is stated in words rather than symbols.

 

For all age group populations except the first,  for years 2000, 2005, 2010, 2015, and 2020:

 

The age group population in the current year equals the previous age group population five years ago plus net in-migration during the previous five years minus deaths during the previous five years. 

 

For the first age group, children 0-4 years of age, the age group population in the current year equals births during the previous five years plus net in-migration minus deaths.

 

Once the potential population was determined for the future years, an equation was applied to produce the actual Temple population from the potential Temple population.  The equation assumed that the potential Temple population equals the actual Temple population for the 13 non-small age groups out of the 18 age groups.  Knowing the persons per household for Temple Beth El, the number of households could be determined from the actual Temple population.  This number constitutes forecast number 9, referred to as Method 1, in Table 7. 

 

5.5.2  Holding the Potential and Actual Temple Age Group Percentages at the 1995 Level

 

            The second way to forecast the Temple households was move the age groups of the potential population ahead in time to obtain the total potential population, as described in Section 5.5.1, but to assume that the five age groups listed above were a constant percentage of the total potential population.  In other words, hold the age group distribution of potential Temple population constant at the 1995 age group percentages. Knowing the potential Temple population, the actual population could be calculated and the the number of households could be calculated assuming the persons per household in the future years.  This number constitutes forecast number 10, referred to as Method 2, in Table 7.  

 

 

 

5.6  Forecasting Temple Households from Geographical Areas

 

            The last method of forecasting the Temple households was based on the geographical distribution of current Temple members and the growth rate of the geographical areas in which they reside.  The basic assumptions of this model are that future Temple members will tend to live in the same areas as present members, in roughly the same proportions, and their numbers will grow as the geographical areas increase in population. 

 

            Past data and forecasts for the period 2000  to 2020 were obtained from the PSRC (Puget Sound Regional Council) for small areas of Pierce County and some small areas in adjacent counties.  The locations of Temple households by ZIP code were obtained from the roster of Temple members contained in the 1995 Temple Directory.

 

            PSRC data projected the population, the average persons per household and the number of households for the years 1990, 2000, 2010 and 2020.  Interpolation yielded estimates of households for the years 1995, 2005 and 2015.  The PSRC forecast data was given for only FAZ areas (forecast area zones) which are not congruent with ZIP code areas.   PSRC also provided a table showing the TAZ areas (transit forecast zones) contained in each ZIP.  A map of FAZ areas was given by PSRC and a map of the ZIPs was available from the U.S. West telephone directory. The problem was therefore to relate the ZIPs to the FAZs by means of the smallest subareas, the TAZs. Once done, a forecast of the Temple residential areas via the ZIPs could be carried out.

 

            The assumption used was as follows:   The fraction of TAZs contained in a FAZ was assumed equal to the fraction of population in that FAZ that was assigned to a particular ZIP.  This calculation is done for all ZIPs and resulted in a eleventh forecast of Temple households as shown in Table 8.

 

 

            The number of Temple households from the different ZIP codes was projected over the forecast period as shown in Table 9.

 

 

6. Combining the Household Forecasts

 

            This section consists of two parts.  In the first part, the results of the eleven forecasts of Temple households will be averaged in three different ways and then combined into an overall average, the final forecast.  In the second part, a 95% confidence interval around the final forecast will be presented.

 

6.1  Averaging the Temple Household Forecasts

 

            There are many ways that the Temple household forecasts could be combined.  Three ways were chosen here for their simplicity.  The first way was simple averaging which is assigning equal weights (importance) to each forecast.  The assumption is that no forecast has any more validity than any other. 

 

            The second way of averaging the eleven forecasts was to assign judgmental weights based on the author's estimates of the validity of the forecast and the accuracy of the numbers.  The author assigned small whole numbers (1 or 2 or 3) as weights to each forecast and then computed a weighted average.  The small group, judgmental, ETE forecast received a weight of 3 because in the experience of the author this method has proven to be as or more accurate than any computer or mathematical method.  The method based on the medium Pierce County household forecast of the OFM, State of Washington received a weight of 2 because it was judged to be twice as likely as the OFM low and high forecasts of Pierce County households.  All other methods received a weight of 1.

 

            The third way of averaging was to calculate the median or middle value of the forecasts.  This is done in the case of eleven forecasts in ranking the forecasts from highest to lowest and then choosing the sixth forecast from the bottom.

 

            The three different averages were then averaged to produce the final household forecast as shown in Table 10.

 

6.2  Obtaining the 95% Confidence Interval

 

            For each of the three averages discussed above, a 95% confidence interval was computed.    The confidence coefficient of 95% indicates the degree of confidence or credibility placed on the estimated interval. 

 

            To obtain the confidence intervals of the first two averages, a normal distribution of the population was assumed.  For small samples of size 11 in our case, the Student-t distribution (lower and wider than the normal, bell-shaped curve) is customarily used in place of the normal distribution.  The Student-t distribution was applied to the eleven forecasts for each of the forecast years.  The equally weighted and the unequally weighted averages produced 95% upper and lower confidence limits that were within a few percent of each other. 

 

            The percentile average resulted in a much wider confidence interval because a normal distribution was not assumed and the more extreme low and high values, i.e., the 2.5 percentile and the 97.5 percentile of the forecasts, were used as the lower and upper bound of the 95% confidence interval.  

 

            In Table 10, the low, medium and high forecasts for the three methods of averaging, followed by the average of the three averages to yield overall low, medium and high forecasts of Temple households is presented.  The information in Table 10 within the bottom box was previously presented in Table 1 and in Figure 1 in the Summary of this report (page 2). 

 

 

 

 

7. Phantom Jews

 

            During the process of inspecting Temple age group data for 1996, an unexpected result was discovered.  The percentages of the age groups 0-4, 20-24, 25-29, 30-34, and 35-39 were significantly below the Pierce County percentages for the same year.  Two Temple age groups were significantly greater than Pierce County age groups, ages 10-14 and 45-49.

 

            Some caveats are in order before continuing.  The Temple age groups were obtained only partly from the computerized roster; many of the women's ages were not available and so had to be estimated by a senior volunteer Temple member.  The problem was that the total of Temple members, whose ages were estimated by the volunteer for the 1996 membership, was 10% less than the total of Temple members on the computer roster as produced by Maggi Barrett [1].  It was assumed that those 10% missing persons were distributed like the other 90% and that 1995 age group percentages equaled 1996 age group percentages.  Those two assumptions  were necessary to forecast potential Temple members in 1995 and then phantom Jews, those households who live in the Temple target area but for reasons unknown are not members of Temple.

 

            The potential Temple population is defined as the population of Temple Beth El if the age group percentages of Temple would be approximately equal to the age group percentages of Pierce County.  The potential population is actually a minimum estimate of the population of Jewish households that could be members of Temple Beth El, as explained below. 

 

            The equation producing the potential Temple population was: the potential Temple population equals the actual population when the five low percentage age groups are subtracted out.   This is only partly true since the Temple age groups 10-14 and 45-49 percentages are significantly higher than Pierce County.  To compensate for this, the excess members in those two age groups will be reduced by the percentage differences between Temple and Pierce County in computing the potential Temple population. 

 

            The potential population is a minimum estimate because of the assumption in the equation that the potential Temple population equals the actual population when the five low percentage age groups are subtracted out.  The potential Temple population (Pierce County Jewish population) is probably greater than the actual Temple population when the five low percentage age groups are subtracted out. If that is the case, as some observers have suggested, then both the potential Temple population and the population of phantom Jews are minimum estimates.  This means our estimate of phantom Jews is a conservative one which understates the true situation.

 

            Calculations produce the result: there are at least 292 net phantom Jews. Using the value of 2.4 persons per household as the family size in 1995,  the number of households equivalent to 292 phantom Jews is estimated to be 122.  In other words, if all of the potential families in the area would be members of Temple, and the excess members subtracted out, the membership would increase in size  by at least 36% over the 1995 enrollment of 344 households to reach 466 households.  The results are shown in Table 11 and Figure 6.

 

            Regarding phantom Jews, in the Britannica Book of the Year is a table that indicates of all the religions in the United States and Canada, only the Jewish religion has experienced a drop in numbers in the 1990's .  From 1990 to 1995, the membership of Jews in the U.S. and Canada has dropped by over one million or by 17%.  Jews will soon be outnumbered by Moslems in this country [5].  Obviously the problem of recruiting and retaining target area Jews as members of Temple Beth El is more than a local problem.

 

 

 

 

8. Extended Forecasts

 

            Finally, we come to the more detailed implications of the Temple Beth El household forecast.  As in the previous sections, a series of assumptions had to be made in order to produce the forecast of the various vital statistics and life events so important for Temple members.  The Temple age group percentages of 1996 were applied to every year of the forecast, the assumption being that Temple will continue to attract members proportionately in the same age groups in the future as now.  This assumption is probably in error.  In all likelihood, the Temple age group distribution will become more like the Pierce County age group distribution, but how fast this would occur is unknown and, thus, the conservative assumption was made.

 

            From year 2000 on, the extended forecast assumed that the percentage of males and females in any age group would equal that of Pierce County.  The reason for this is that Temple male-female percentages by age group varied greatly in 1996 and probably would not be repeated in the future.  Pierce County male-female percentages for each age group were found to be fairly constant over the entire forecast period.

 

            The medium final household forecast multiplied by the persons per household was used to project the total number of Temple members for the forecast target years.  The total number of Temple members was distributed to the 18 age groups according to the 1996 Temple age group percentages.  Each age group was further divided into males and females by multiplying by the male and female percentage of the age group, respectively.  Once done, the Temple male and female populations for each age group and for each forecast year were calculated.  Births were produced by multiplying the birth rate for each age group for a particular year by the number of women in the age group in that year and summing  over all age groups from age 15 to 44.  Deaths were obtained for males and females by multiplying the most current figure for death rates (U.S. 1993 rates) for the gender and age group by the number of persons in the age group, then summing over all 18 age groups, and then adding male and female deaths to get a total for the forecast year.  The age group specific birth rates and death rates were U.S. white rates because local rates were not readily available and Temple membership is overwhelmingly white.              

 

            From births and the Pierce County male-female fractions, the number of brits and baby namings were calculated.  Assuming the numbers of boys and girls at age 13 are equal to 20% of the male and female 10-14 year age group, respectively, the number of potential bar and bat mitzvahs could be forecast. 

 

            Results of the extended forecasts are shown in Tables 12 through 15.  The results are given as average yearly rates for five year periods.  These projections are to be judged the most error-prone because they attempt to predict many events for relatively small samples in which variation can be great.  For that reason, there is no attempt to place a confidence interval around any of these extended forecasts.  They are meant to be taken as rough indicators of the number of different kinds of life events over a 24 year time period and not as specific predictions against which to judge the validity of the overall Temple forecast.  

 

 

 

 

 

 

 

 

            Important statistics for Temple Beth El are shown in Table 15.  Over the period from year 1995 to year 2020, the following changes will occur.  Births, brits and baby namings will go up by 50% on the average; bar mitzvahs by 27%; and bat mitzvahs by 56%. Sunday School enrollment will increase by 41% from 196 to 277 pupils.  Finally, deaths will increase by 40% from an average of 10 per year to 14 per year.  Meanwhile, the total population of the membership will increase by 41%.

 

            The forecast for each year of the period from 1995-1999 is 11 bar mitzvahs and 9 bat mitzvahs for a total of 20.  By way of comparison, in the 1995-1996 Jewish calendar year, there were 13 bar mitzvahs and only 3 bat mitzvahs for a total of 16.  Although the number of females is less than the number of males in the bar-bat mitzvah age, the difference between males and females cannot alone account for the difference in bar and bat mitzvahs.   Therefore, to the extent that other factors continue to prevent bat mitzvahs from occurring, the forecast for bat mitzvahs will be in error.

 

            There were two baby namings in the 1995-1996 Jewish calendar year.  The forecast called for two baby namings.  Brits were not recorded on the Temple office calendar.  Some of the brits were performed on infants of non-members and would not be included in the forecast [6]. 

 

            The low number of forecast births can be traced to the low numbers of women Temple members (not children of members) in the highest birth rate groups, i.e., 15-19, 20-24, and 25-29 year age groups.  Many Jewish births will be revealed only five years later when the parents, phantom Jews, decide to join Temple in order to provide their children with a religious education.

 

            Caution must be exercised in using these figures because they are computed on the basis of a small sample and a large number of assumptions, some of which may be in error.  Nevertheless, these figures do present one possible scenario for the future of Temple Beth El if past trends persist into the future.

 

 

 

 

9. Alternative Scenarios

 

            Up to this point in the report, the forecasts assumed continuation of past trends, with various assumptions, into the future.  The type of scenario thus produced is called a surprise-free futurebecause it is characterized by persistence rather than change.  Although the surprise-free future is the most probable future, its probability is still low because most likely one or more surprises or shocks will occur over the next 24 years.  Realizing this fact of life, it is prudent to qualify the surprise-free future by imagining what might happen to significantly change the projected Temple household trend, either up or down, from its long-term forecast values.  Keep in mind while reading these scenarios that the likelihood of any one of them occurring is very small; also do not rule out that two or more of these could occur simultaneously or in sequence over the next 24 years.

 

9.1 Developments which Could Significantly Decrease Temple Membership

 

            1) The rigid rabbi: The new rabbi does not succeed in preventing the disaffection of the more traditional from the less traditional members of the congregation either because he/she is too liberal or too conservative in his religious practices.  The rabbi has certain principles that he/she will not violate.  The congregation splits apart into a small but younger traditional group and a larger, older non-traditional group.  Neither group is large enough to sustain a financially viable temple or synagogue.  The impact is a sudden drop of 20-30 percent of membership. 

 

            2)  The milquetoast rabbi: A new rabbi is selected on the basis that he will be able to retain the unity of Temple Beth El traditionalists and non-traditionalists.  He is mentored, coached, advised, criticized and corrected by well-meaning, committed Temple members.  Every time he tries something new, he is urged to reconsider because it might offend some group or faction.  Ultimately, he becomes so frustrated that he abandons all initiatives and just does not want to make waves.  He serves out his contract and leaves at the end of two years.  Fingerpointing ensues; the congregation is split; many families cancel their membership.  Temple Beth El begins a long-term decline in membership.

 

            3)  Economic bust: As a result of a series of bad decisions, Pierce County experiences an economic downturn.  The military takes a big hit; McChord is downsized.  Transportation on I-5 is pure gridlock yet transit bond issues still fail repeatedly. The legislature denies funds for the expansion of higher education.   Business taxes are raised resulting in an exodus of corporations from the area.  Draconian environmental regulations are applied to industry.  Port of Tacoma business is lost to Vancouver, B.C. and to Los Angeles.  The Jewish community also suffers the economic downturn which shows no signs of abating.  Dues pledges are not fulfilled resulting in the Board raising dues to balance the budget.  Families drop their membership as they struggle to make ends meet.  A long-term decline in dues paying members begins.

 

            4)  Competition arises: The founding of new Temples, synagogues and chavurot in adjacent communities will tend to whittle away the peripheral membership of Temple Beth El.  Jewish families in Olympia, Kent, Bremerton-Silverdale, Auburn and Vashon Island prefer to join a small, local, friendly group of co-religionists than to shlep 45 minutes over crowded freeway, bridge or rural roads to Temple.  Some from the Peninsula and Puyallup decide to try the smaller and less expensive congregations equidistant from their homes.  As a result, Temple Beth El sees a loss of 100 households from the 1995 membership over a ten year period. 

 

            5)  Demographic disaster: National trends of intermarriage increase from 50% to 75%.  Often the non-Jewish spouse not only does not convert but maintains his/her ties to a church.  Family funds available for Temple dues are cut in half because they now have to support two religious organizations. Commitment to remain members after children are confirmed declines rapidly.  Furthermore, children of committed Temple members increasingly choose to attend college out of the region, later find jobs out of Tacoma, and never return to live in Tacoma.  Finally, Temple members have fewer children per family mirroring a national decrease in white birth rates among educated women.   Growth in membership grinds to a halt, then begins to decrease.  Over a period of 20 years, these developments put the continued existence of Temple Beth El at risk.

 

 

9.2  Developments which Could Significantly Increase Temple Membership

 

            1) The super rabbi: The new rabbi is young and filled with new ideas, new programs, new songs and prayers.  He/she displays energy, verve, enthusiasm and relates extremely well with the younger parents, the teens and the young children.The new rabbi expands the outreach program to include gentile spouses of Jewish members as well as phantom Jews and their gentile spouses.  Emphasis is on Jewish values, the meaning of traditional practices and what it means to be a Jew.  Temple members are encouraged to help the new rabbi and respond with enthusiasm.  The number of converts and returning  phantom Jews passes the 100 household mark.  Increased Jewish social and cultural events are planned.  A Jewish theater group, a Jewish music group, a lecture series bringing in noted Jewish thinkers and writers, a community open forum, a Jewish youth basketball team, and Jewish welcome wagon are formed.  Temple opens a schmooz-bagel-coffee shop.  Members who pay dues but never show except for holidays begin to come around.  A Jewish Community Center enfolds and makes connections with surrounding Jewish groups.  Temple Beth El begins to attract people from all around the area. An anonymous benefactor bequeaths one million dollars to Temple.   Membership surpasses the most optimistic trend forecast.           

           

            2) Economic boom: Pierce County will be one of the long-term growth areas in the country.  In Seattle, there is a shortage of industrial and residential space, congestion, expensive housing, long commutes to an ever more crowded East Side.  Some of the population will get pushed down here.  Pierce County has land, good housing prices, an expanding port, a continuing military presence, improvements on the Foss Waterway, rejuvenation of the Tacoma downtown area.  As the shift from traditional manufacturing to high-tech business services occurs, more Jewish families will arrive -- not as independent lawyers, doctors or accountants -- but as professional employees tied to new firms working in advertising, finance, public relations and software engineering.  As a result, expect an influx of Jewish families with children in the next two decades causing Temple membership to soar.

 

            3) California earthquake: a series of powerful earthquakes over several months hit Silicon Valley resulting in a mass exodus of high-tech firms relocating to Washington and Oregon, which still puts them on the Pacific Rim.  Some of the manufacturing facilities decide to situate in Pierce County because of land prices, taxes, and availability of power. Skilled software and systems engineers arrive by the hundreds; many are Jews with families who earn good salaries in administrative and technical positions. However, not a small number of them are still suffering trauma from the quakes.  At Temple Beth El they find a warm welcome and help of all kinds in getting settled in the area, a good Sunday school and a stimulating adult program.  Temple membership begins to grow and then exceeds the most optimistic forecasts.

 

            4) Antisemitism arises:  As the world moves into the 21st century, the split between the rich and the poor grows greater in the United States. Support for Pat Buchanan for president increases.  The middle class shrinks and many people, looking for a scapegoat, accuse the Jews as being responsible for their plight.  Antisemitic remarks on talk shows, in the newspaper letters to the editor, heard in remarks in the street confirm this.  The skinhead population is rising.  Jewish kids report ethnic slurs at school.  The Temple is vandalized. Louis Farrakhan is invited to speak at a local university, the turnout is large, and his message is pure hate.  A hunkering down occurs in the Jewish community of Tacoma.  The Temple becomes a communications center whose additional purpose is to counter the attacks on Jews.  Contributions pour in; members rarely seen are now volunteering to defend the Jewish community and its members.  Attendance at services rises dramatically as the threat coalesces Temple members.   Membership, which was level, begins to suddenly rise as Jews, regardless of previous comittment, recognize the threat of growing antisemitism.

 

            9.3 Summary:  Nine scenarios have been projected for the next 24 years that could change the Temple household forecast significantly if they occur.  Although we cannot assign a probability to any of these scenarios, we can say that they are all possible even if very unlikely, and that they will all impact heavily on Temple membership should they occur.  The purpose of including these scenarios in the report is to expand the thinking of the Board beyond the conventional future projected by the forecast.  Hopefully, these scenarios will better prepare the Board to expect any shocks or surprises which will almost certainly occur before year 2020.

10. Conclusions

                  

            Past Temple households have increased from 234 in 1979 to 344 in 1995 only to drop to 330 in 1996.  This represents a gain of 41% in 17 years or 2.4% per year, a growth rate which matched  Pierce County's  over the same period.

 

            In contrast to the past growth, the next two and a half decades will experience a marked lower growth rate.  During the 1995 to 2020 period, Temple Beth El is forecast to increase by 155 households or 45%, going from 344 households in 1995 to 509 households by 2020, an average growth rate of 1.9% per year. 

 

            The 95% confidence interval projects for the year 2020 a high of 558 households and a low of 471 households over the same period, translating into an average yearly growth rate of 2.5% on the high side to 1.5% on the low side.

 

            Temple Beth El's age group distribution is a bimodal distribution, skewed to the higher age groups.  Relative to Pierce County's age distribution percentages, there is deficiency of children ages 0-4 and adults from ages 20-39.  However, there is a surplus of children ages 10-14 and adults 45-49.  As a result of these aberrations from Pierce County, the minimum number of potential Temple households living in the Temple target area but not members of Temple, called phantom Jews, was calculated.  In 1995, the number of phantom Jews was calculated to be 292 which translated to 122 households or 36% of the 1995 Temple membership.

 

            On the basis of the current geographical distribution of Temple households plus data from Puget Sound Regional Council, the future geographical distribution of Temple members was forecast.  The three areas forecast to contain by 2020 the highest numbers of Temple members are: 1) Peninsula (ZIP code 98335) with 91 households; 2) area N. 6th to N. 26th (ZIP code 98406) with 57 households; and 3) University Place north (ZIP code 98466) with 55 households.  The three areas forecast to experience the highest growth rates from 1995 to 2020 are: 1) Lakewood East (ZIP code 98499) at 100%; 2) central Tacoma (ZIP codes 98402-98405 and 98408) at 85.1%; and Puyallup (ZIP codes 98373-98374) at 84.3%.

 

            The extended forecast of Temple Beth El dealt with the average number of life cycle events over six time periods, starting with 1995-1999 and ending with 2020-2024.  Births originating out of Temple member households (not including phantom Jews) is projected to increase from 4 per year in 1995-1999 to 6 per year in 2020-2024.  Brits and baby namings are divided equally in those years.  Bar mitzvahs   will increase from 11 per year to 14 per year during that time interval.  Bat mitzvahs are projected to go from 9 per year to 14 per year.  This forecast assumes that all eligible girls will be bat mitzvahed which may not be the case.  Deaths are forecast to increase from the current average rate of 10 per year to 14 per year.            The extended forecast also projects the Sunday school enrollment to grow from 196 pupils in 1995 to 277 pupils in 2020, an increase of 81 pupils or 41% over the 25 year period.  Over the same time period, Temple membership is expected to grow from 826  members to 1165  for an increase of 339 persons.  

 

References

 

[1]  Membership Chairman, Membership Chairman's Report, Board of Trustees, Temple          Beth El, May17, 1989; May 18, 1980; May 13, 1981. Temple Beth El Directory,     1988, 1993, 1995.  Maggi Barrett, Temple Beth El Demographics, June 19, 1996.            Maggi Barrett, Printout of Temple Beth El Membership, June 14, 1996.  Iris Sigel,             1996 Age Group Distributions of Temple Beth El, Personal Communication,                  August, 1996.

 

[2]  Office of Financial Management, State of Washington, Washington State County Population Projections by Age and Sex: 1990-2020, OFM, January 19, 1996. 

 

[3]  Puget Sound Regional Council, Population, Household and Employment Data:               Pierce County by FAZ, August 1995, pp. 165-189.

 

[4]  Bureau of the Census, Department of Commerce, Statistical Abstract of the United         States: 1995, Tables 4, 87, 92 116 and118.

 

[5]  Moyers, Bill, "America's Religious Mosaic," USA Weekend, October 11-13, 1996.

 

[6]  Barrett, Maggi, Personal Communication, October 18, 1996.