In 1984, the Minnesota Legislature directed the State Planning Agency to prepare a study and plan regarding the state hospital system (Laws of Minnesota 1984, Chapter 654, Section 19, Subdivision 4 :9) . One part of this study was to prepare "an economic impact statement and alternative economic development strategies for each state hospital region likely to be affected by program reductions in the regional state facility." The purpose of this report is to present and compare the impact that each state hospital has on the local economy.
II. METHODOLOGY
For the purposes of this report, economic impact means local economic impact. We will not address such issues as the impact of a closure on state revenue or expenditures. In addition, discussion of the impact of state hospital purchases and salaries will be limited to the area of intense impact around the hospital community and not with the impact on the economy of the state as a whole. An assessment of the economic impact of a state hospital closure must consider two factors: the absolute impact on a community of a closure and the comparative impact of the closure of a hospital in Community A versus the closing of hospitals in Communities B, C, or D. While many of the techniques used in determining both types of impact are the same, other methods take on different degrees of importance. Multiplier analysis, for example, is crucial in the determination of absolute impact. Jobs lost or money not spent have
a ripple effect resulting in the loss of non-hospital jobs and money. With relative impact analysis, multipliers are much less important. As will be seen later, multipliers vary little from location to location, thereby having an effect that is reasonably constant from hospital to hospital and not Changing the relative standing of each institution. By contrast, the determination of the geographic area of impact is
Economic Impact Page 2
January 31, 1985
extremely important for relative analysis but less important for absolute. Relative impact deals not only with the comparison of dollars lost to the economy but with the share of the local economy accounted for by the hospital. What portion of area jobs and sales results from the presence of the hospital? Without this analysis we would simply conclude that the largest hospital has the greatest impact, regardless of the size or economic base of the community.
A. An Economic Impact Model
All ideal model of economic impact would component could be accurately quantified up to a summary figure. This, of course, be one in which every and all effects added cannot be done. However, an attempt to lay out the parameters of such a model can be an aid in understanding the total range of factors affecting the economy of the hospital community. Such a model would include the following factors:
Economic Impact =
Direct effect of hospital employment +
Secondary effect of hospital employment +
Effect of hospital purchases +
Effect of spending by hospital residents/patients and visitors +
Effect of unemployment benefits, early retirement benefits, and severance pay +
Economic impact of hospital workers and families leaving the area +
Salary gains or losses resulting from alternative employment of hospital workers +
Impact of reuse of the hospital facility +
Change in local government expenditures for provision of care to previously hospitalized local residents/patients +
Cost to local government of providing social services to unemployed hospital workers +
Effect on local government tax revenues +
Effect on local finances of no longer providing municipal services to hospital facilities.
Economic Impact Page 3
January 31, 1985
Each item in this model will be discussed separately beginning with section III of this report.
This impact model includes not only the adverse direct effects of a hospital closure but includes mitigating effects as well such as unemployment insurance and alternative uses of facilities. Other economic impact studies have dealt only with "worst cases" scenarios - the assumption that all employees of a closed facility and their families will immediately move out of the region. This will obviously not be the case and an economic impact summary should make this explicit. The first four items listed in this model constitute such a "worst case" scenario; the remaining items exist to document likely alternatives to the worst case.
B. Qualitative Factors vs. Quantitative Factors
Not all of the factors involved in the determination of economic impact are readily quantifiable. Some, such as the economic impact of state hospital purchases, are relatively easy to assign a dollar figure. Other equally important aspects, such as the prospects for reemployment of hospital staff, are difficult or impossible to summarize with a number. This does not mean that the more qualitative elements of the analysis are less important that those that can be summarized with a "bottom line" figure. Economic impact is much too complex an issue to summarize with one figure or even a group of figures. This report will not attempt to combine all of the factors listed above into one "bottom line" figure summarizing all facets of economic impact. Instead, each factor affecting the local economy will be discussed separately. Some factors will be precisely quantified while others will be difficult to assess on even a qualitative basis.
III. DEFINITION OF ECONOMIC IMPACT ZONES
The extent of economic impact begins with a definition of the geographic area under discussion. To facilitate comparison of economic impacts, the definition must be consistent for each hospital, specific enough to produce meaningful results, and yet must reflect availability of data.
The definition must also reflect several facts. First, economic impact is a function of where employees live and spend their money, not only where they work. This is a most significant concept because employee commuting patterns vary considerably from state hospital to state hospital. In some cases the vast majority of employees live only a short distance from the hospital, thus concentrating the economic impact. In
Economic Impact Page 4
January 31, 1985
others, long commuting distances are not uncommon, and the workforce is spread over a wide area, thus reducing the impact of a closure on any one community.
Second, because of this dispersion, it is desirable to look at the impact of a hospital closure both on the immediate community and on the surrounding region. It is quite possible for a closure to have a severe local impact but a much less intense regional impact or vice-versa.
Finally, political subdivisions often present inappropriate boundaries for economic impact analysis. For example, counties have frequently been used as units for economic analysis because of the availability of county-level data, but in this case three of the eight state hospitals (Anoka, Moose Lake, and St. Peter) lie within five miles of a county line. Clearly the impact of a closure of one of these institutions would be greater in parts of the bordering county than in the remainder of the hospital county. However, virtually all necessary data are collected on the basis of political subdivision. My definition of an impact area must therefore represent a compromise between need and availability.
The definition of economic impact areas, therefore, will be largely based on the geographic desegregation of available data. Zip codes will be used because they are the smallest geopolitical unit for which data are available on hospital employees as well as the 1980 Census. Zip codes represent a compromise; data on the city and township level would be preferable but are not available. However, some data (such as data on local economic activity) are not even available at the zip code level. In these cases data will be compiled based on the finest level of geographic desegregation available. This means that the analysis of different elements of economic impact may not be based on the same geographic units. The only alternative is to base every portion of the analysis on the broadest area for which all data elements are available an approach which would seriously weaken much of the analysis.
For purposes of this report, the economic impact area is defined as that area (based on zip codes) in which 90 percent of the hospital employees reside. Within each impact area the primary impact zone is the area immediately surrounding a hospital in which 50 percent of the employees live. This includes the area with the greatest concentration of hospital employees and is the area which will feel the greatest localized impact. A secondary impact zone is the area in which 75 percent of the hospital employees reside. This area includes the primary impact zone and presents a "middle ground" of dispersion between the primary zone and the entire economic impact area. In the cases of Brainerd, a Faribault, and Fergus Falls more than 75% of the workforce resides
Economic Impact Page 5
January 31, 1985
in just one zip code. In these instances the primary and secondary zones share the same boundaries.
Table 1 presents the economic impact areas for each state hospital as defined by zip codes, and Figures 1 through 9 illustrate the economic impact area for each state hospital. Note that there is some overlap between the Anoka and Cambridge economic impact areas.
During the review and comment period for this report, representatives from the Coalition of Concerned Citizens for Moose Lake State Hospital and Faribault State Hospital suggested different economic impact areas than the ones presented. Appendix B shows the results that would be obtained using the boundaries suggested by these groups.
TABLE 1
ECONOMIC IMPACT AREAS AS DEFINED BY ZIP CODES
Source: Minnesota State Demographers Office, State Planning Agency, 1984.
Economic Impact Page 6
January 31, 1985
FIGURE 1
ECONOMIC IMPACT AREA FOR MINNESOTA STATE HOSPITALS
Economic Impact Page 7
January 31, 1985
Economic Impact Page 9
January 31, 1985
Economic Impact Page 10
January 31, 1985
Economic Impact Page 11
January 31, 1985
Hospital employment varies from over 1,000 at Faribault to just over 300 at Anoka. However, as discussed earlier, the economic impact of the hospital on a region depends more on the percentage of area jobs accounted for by the hospital than on the absolute level of employment. Table 2 and Figures 10, 11 and 12 illustrate the share of area employment accounted for by each of the eight hospitals. Hospital employment for 1984 by zip code was obtained from the Department of Employee Relations, while total area employment was obtained from the 1980 Census. The two data items do not correspond exactly in time. Because more current estimates of area employment are available at the needed level of geographic desegregation, it is assumed that in each area total employment has not changed significantly since 1980. In most cases employment has likely remained relatively constant or increased slightly during this period.
TABLE 2
STATE HOSPITAL EMPLOYMENT AS A PERCENT OF TOTAL AREA EMPLOYMENT
The loss of employment in the primary impact zones ranges from 19 percent at Moose Lake to 0.5% of the employed workforce at Anoka. These figures should be interpreted as the direct
employment loss in the immediate hospital community. The direct impact in Moose Lake is more than twice that in any other community, while the impact in Anoka would appear to be
substantially less than all other facilities. Direct employment losses in the other six communities are similar to each other.
Economic Impact Page 12
January 31, 1985
FIGURE 10
HOSPITAL EMPLOYMENT AS PERCENTAGE OF TOTAL EMPLOYMENT - PRIMARY IMPACT ZONE
Source: Department of Employee Relations, 1980 Census.
FIGURE 11
HOSPITAL EMPLOYMENT AS PERCENTAGE OF TOTAL EMPLOYMENT - PRIMARY AND SECONDARY IMPACT ZONES
Source: Department of Employee Relations, 1980 Census.
Economic Impact Page 13
January 31, 1985
FIGURE 12
HOSPITAL EMPLOYMENT AS PERCENTAGE OF TOTAL EMPLOYMENT - ENTIRE IMPACT AREA
Source: Department of Employee Relation, 1980 Census.
In examining the secondary impact zones and the balance of the impacted areas the relative effects of the dispersion of the workforce can be noted. While direct employment loss at Anoka remains much less severe than at other institutions, the differences between the seven other institutions narrow considerably in large part due to the dispersion of the Moose Lake workforce. If we consider the primary and secondary zones together, the heaviest impact is still found in Moose Lake, but the difference between it and the other areas (notably Faribault) narrows considerably. However when we look at the entire impact area the share of the workforce accounted for by Moose Lake and Faribault State Hospitals drops considerably, reflecting the hospital workers who commute from Cloquet and Rochester. Similarly, the figures for St. Peter drop off considerably when the secondary zone is considered because of the large number of employees commuting 12 miles from Mankato. When the entire impact zone is considered the impact is greatest in the Brainerd (4.8% of the area workforce), Fergus Falls (4.2%) and Willmar (3.9%) areas, reflecting the role these cities play as regional centers. The impact in Anoka is still considerably less than that of the nearest state hospital.
Table 3 shows the percentage of total area wage and salary income accounted for by the state hospital. Hospital salaries for 1983 were obtained from the Department of Employee Relations, while total wage and salary income from the 1980 Census was inflated by the statewide growth in personal income from 1979 to 1983. The
Economic Impact Page 14
January 31, 1985
comparative rankings of each community vary little from this table to the previous table. It should be noted that in each case (except Anoka), hospital wages make up a greater percentage of total area income than hospital jobs make up of total area employment. This suggests that hospital jobs make up a higher share of each region's better paying jobs. The disparity is greatest in Moose Lake and Fergus Falls.
TABLE 3
HOSPITAL PAYROLL AS A PERCENT OF TOTAL AREA WAGE AND SALARY INCOME
Source: Department of Employee Relations, 1980 Census
Direct Employment Loss
The following figures reflect an assumption that immediately following the possible closure of a hospital, all state hospital workers will be considered unemployed. The Minnesota Department of Economic Security (DES) has calculated alternative unemployment estimates which recognize important exceptions to this assumption. First, based on the age structure of hospital employees assumptions were made about the number of employees who would retire following a hospital closure. DES assumed that one-half of the workers between the age of 55 and 61 and three-fourths of the workers 62 or over who are eligible for retirement would do so. This leads to an estimate of retirees totaling 1.7% of the hospital workforce. Since 3.2% of the Rochester State Hospital workforce retired following closing of that institution, this represents a conservative estimate. Second, a number of employees will transfer. DES estimated a transfer rate of 3.5%. It should be kept in mind that the wages of transferees are still lost to the local economy. Third, DES assumed that 5% of the workforce would immediately find private sector employment and another 5% would a drop out of the labor force. Fourth, a skeleton crew would likely
Economic Impact Page 15
January 31, 1985
be kept on to maintain facilities until their disposition. Based on the experience of Rochester, DES assumed that 15 employees would be retained at each site for this purpose. Finally, a number of people currently hold second jobs and so would not join the ranks of the unemployed (although total income would be reduced). Based on national survey figures, 4.0% of state hospital employees were assumed to be multiple job holders.
Table 4 presents the results of the DES "reasonable case" direct unemployment estimates. The data used for these calculations can be obtained only for counties, not for economic impact areas. There is also some geographic overlap involving employees of Anoka and Cambridge hospitals which will affect the results for the Minneapolis-St. Paul Metropolitan Statistical Area (MSA) and Isanti County. The net effect of using this "reasonable case" model is a reduction in unemployment of 1,239 over the "worst case" scenario. The distribution of this reduction in unemployment is roughly proportional around the state, so the relative standing of hospitals and counties is not affected by a choice of the "worst case" or "reasonable case" scenario.
TABLE 4
ALTERNATIVE ESTIMATES FOR DIRECT UNEMPLOYMENT
Economic Impact Page 16
January 31, 1985
When a major facility is shut down, the loss of jobs does not end at that institution. Rather, there is an additional loss of employment in industries that serve the facility and its workers. This indirect loss of employment cannot be directly calculated but must be estimated using a technique called "multiplier analysis." The main difficulty in this procedure is selection of a multiplier, that figure by which the direct loss in payroll is multiplied to obtain the indirect effect. A large number of theoretical approaches to this problem have been suggested.
Multiplier analysis, like most economic forecasting is an inexact science. It seems prudent to examine more than one model and provide a range of possible outcomes rather than presenting one point estimate. Three models were chosen. The first was proposed by the Research and Statistical Services Division of the Minnesota Department of Economic Security (Pinola and Graner, 1984). The second derives from an econometric forecasting system developed for the Minnesota Department of Revenue (Treyz, Erlich, and DePillis, 1983). Both are statewide models; that is, the choice of multiplier does not vary from community to community. The method developed by the Department of Economic Security enables forecasts for the duration of unemployment, while the Revenue Department model estimates in what sectors of the economy the indirect employment loss will occur. Multipliers derived from these two models are quite similar, with the Department of Economic Security model giving a multiplier of 1.67 and the Department of Revenue model giving 1.69.
The third model involves the calculation of a separate multiplier for each community. This regional multiplier approach results in a higher multiplier in urban areas than in rural areas, reflecting the fact that money remains in an urban area longer than it does in a rural area. A good purchased in an urban area is more likely to have been manufactured in that area (keeping that money in the community) than a good purchased in a rural area. This effect, however, is thought to be much more important for estimating the economic effect of changes in manufacturing employment than for the effects of changes in service sector employment, such as hospitals. The method used for this report, called the "minimum requirements technique," was originally derived for urban areas by Moore in 1973. This approach has since been adapted for use in rural areas (Erickson, 1977) and has been found to produce reasonably accurate results when compared with a complete census of all businesses and employers in a region (Gibson and Worden, 1981). The multipliers derived using this third approach are
Economic Impact Page 17
January 31, 1985
presented in Table 5, and range from 3.3 in Anoka to 1.7 in Fergus Falls. This model represents an upper bound on reasonable multiplier estimates. Assessment of indirect employment loss was completed for the entire impact area, not for zones within these areas. The indirect
job loss resulting from the closing of a facility is not completely confined to that facility's community. It is reasonable to assume that virtually all of the job loss will occur within the economic impact area, but the models are not sufficiently detailed to break down results within impact areas.
It should be noted that for comparative purposes we are assuming that in each case the full brunt of indirect effects are felt at the same time. This is not a realistic assumption; indirect effects are likely to manifest themselves from one to ten years after a hospital closure. The Department of Economic Security methodology makes it possible to examine how these indirect job losses occur over time; these results are discussed in section IV. Detailed descriptions of all three models are found in Appendix A.
TABLE 5
MULTIPLIERS DERIVED USING "REGIONAL MULTIPLIER" METHODOLOGY _
As can be seen in Table 6, the results from these three models show little variation. The heaviest combined direct and indirect impact would occur in the Fergus Falls area, where 7.8% of the workforce would be unemployed as a result of the closing of the hospital. The next greatest impact would be felt in Brainerd, where the unemployment rate would increase from between 5.9% and 6.1%. Anoka once again shows the least severe effect, with an increase in unemployment of 0.1%. The average of all three models is presented in Figure 13.
Economic Impact Page 18
January 31, 1985
TABLE 6
AVERAGE UNEMPLOYMENT (DIRECT AND INDIRECT) AS PERCENTAGE OF 1980 WORKFORCE
Source: Department of Employee Relations, 1980 Census.
B. Duration of Job Loss
The methodology employed by the Department of Economic Security enables us to make an estimate of unemployment (direct and indirect) one year and two years following a closing. These figures reflect the assumptions about retirement, transfer, labor force dropouts, skeleton crews, and multiple job holders
Economic Impact Page 19
January 31, 1985
described in section III. The estimates were derived using a factor provided by the U.S. Bureau of Labor Statistics to determine how many people remain in the pool of unemployed with the passage of time. This factor is nationwide and does not vary from hospital to hospital.
TABLE 7
JOB LOSS ONE AND TWO YEARS AFTER HOSPITAL CLOSURE
Source: Department of Economic Security, Research and Statistics Service.
The reduction in direct unemployment is fairly rapid, and by the end of the second year direct unemployment is negligible, according to the Department of Economic Security. This reduction in direct unemployment, however, does not necessarily result from persons finding jobs in the area. It may in part reflect both persons leaving the area to search for work and persons who drop out of the labor force.
The indirect effect, however, does not fully manifest itself until the second year. This in part reflects the availability of unemployment insurance to hospital workers. The DES methodology assumes that 50 percent of the multiplier effect would occur within one year of the closure of the facility, with 30 percent occurring in the second year and the remaining 20 percent three to ten years following a hospital shutdown. The indirect effects of a hospital closure, therefore, will continue to manifest themselves long after the hospital is closed.
C. Industries Affected by Indirect Job Loss
The Department of Revenue model estimates in which sectors of the economy indirect job loss would occur. As can be seen in Table 8, most of the job loss would occur in the retail (38%) and service (35%) sectors.
Economic Impact Page 20
January 31, 1985
TABLE 8
INDIRECT JOB LOSS BY INDUSTRY
Source: Department of Revenue, Research Division.
VI. EFFECT OF STATE HOSPITAL PURCHASES
Purchases of goods by state hospitals comprise a small amount of state hospital budgets. Purchases in FY '83 as a percentage of the total hospital budget ranged from a high of 12.4% at Anoka to a low of 9.3% at Cambridge, a relatively narrow range (Table 9). The dollar amount of purchases ranged from $2.75 million at Faribault to $1.32 million at Anoka.
While purchases of this magnitude would at first glance seem to have a substantial economic impact, the local impact is considerably diluted by the fact that most purchases are not made in the community but through centralized purchasing in
Economic Impact Page 21
January 31, 1985
St. Paul. Examination of hospital records for FY 183 resulted in a list of purchases actually made in the city and county in which the hospital is located. These figures were then compared to total retail sales in the same locations over the same period of time obtained from the Minnesota Department of Revenue (1984) .
TABLE 9
PURCHASES AS A PERCENTAGE OF TOTAL BUDGET
The degree to which local purchases ultimately affect the local economy is not clear. Much of the money spent at a clothing store, for example, is eventually passed along to manufacturers or suppliers outside of the area. The amount of money leaving the area (and its impact before it leaves) is impossible to determine. For simplicity we, therefore, assumed that all money arising from local purchases stayed in the community. This assumption has the effect of overstating the local economic impact that would result from the elimination of these purchases. Another difficulty is presented by the purchase of fuel and utilities. These are sometimes purchased from a local supplier (such as a municipal power company) or a large, centralized supplier with a small local office (such as Northern States Power). Very little of this money probably stays in the community, but isolated local impacts are possible. Results are thus presented in two ways: counting fuel and utility purchases as local purchases and excluding fuel and utility purchases (Tables 10 and 11).
Table 10 displays the share of each hospital's purchases which are made locally. If utilities are not included, local (county) purchases ranged from 4.0% of total non-utility and fuel purchases at Anoka to 18.7% at Willmar, with a median value of 13.2%. If utilities are included the importance of local purchases is substantially increased, ranging from 9.4% at Faribault to 41.2% at Moose Lake, with a median value of 20.8 percent.
Economic Impact Page 22
January 31, 1985
TABLE 10
PERCENT OF STATE HOSPITAL PURCHASES MADE IN HOST CITY/COUNTY FY '83
(Utilities Not Included)
Sources: Department of Human Services, Financial Management Division; Department of Revenue.
Economic Impact Page 23
January 31, 1985
TABLE 11
PERCENT OF STATE HOSPITAL PURCHASES MADE IN HOST CITY/COUNTY FY 183
(Utilities Included)
Sources: Department of Human Services, Financial Management Division; Department of Revenue.
The best measure of the economic impact of purchases, however, is the share that hospital purchases make up of the total retail sales activity in the locality. As can be seen in Tables 12 and 13, this effect is very small in every case, amounting to about one-third of one percent of city retail sales activity (in Moose Lake) or less if utilities are not included and slightly over one percent (in St. Peter) or less if utilities are included. The impact on counties is even less, with only two counties (Nicollet and Carlton) losing as much as one-half of one percent of its retail sales in the event of a closing even if utilities are included. If utilities are excluded from the tabulations, the greatest county loss occurs in Rice County (Faribault State Hospital), which would lose two-tenths of one percent of its retail sales.
Economic Impact Page 24
January 31, 1985
Theoretically, the loss of purchasing power resulting from the cessation of state hospital purchases in the community would also result in the loss of some non-hospital jobs. This job loss was estimated using the Department of Revenue and Department of Economic Security models described in Section V.
TABLE 12
HOSPITAL PURCHASES AS A PERCENTAGE OF LOCAL RETAIL SALES FY '83
Utilities Not Included
Sources: Department of Human Services, Financial Management Division; Department of Revenue.
Economic Impact Page 25
January 31, 1985
TABLE 13
HOSPITAL PURCHASES AS A PERCENTAGE OF LOCAL RETAIL SALES FY '83
(Utilities Included)
Sources: Department of Human Services, Management Division; Department of Revenue.
These models estimate the statewide job loss that would result from a loss of a given amount of purchasing power. Because the loss of hospital purchases would probably not result in a statewide loss of purchasing power and because the job loss resulting from a loss of local purchases would likely be almost entirely local, the simulation was run only on local (county) purchases with the resulting employment loss considered to be entirely local. The results, displayed in Tables 14 and 15 indicate that the employment loss resulting from the cessation of hospital purchases alone is likely to be very small even if utilities are included in the tabulation of purchases. The greatest job losses are projected using the Department of Economic Security model. Under this model, if utilities are included as local purchases, employment losses range from 13 in Anoka County to 7 in Isanti, Kandiyohi, Ottertail, and Rice Counties. If utilities are not counted as local purchases, expected employment losses range from 6 in Kandiyohi and Rice Counties to 1 in Anoka and Isanti Counties. Using the Department of Revenue model the relative standing of localities remains the same, but the expected employment loss ranges from 9 to 4 if utilities are included and from 4 to 1 if they are not.
Economic Impact Page 26
January 31, 1985
TABLE 14
ESTIMATED EMPLOYMENT LOSS RESULTING FROM CESSATION OF LOCAL HOSPITAL PURCHASES
(Utilities Not Included)
Source: Department of Revenue.
Note: Estimated local employment loss figures including utilities are noted in parentheses.
The impact of the loss of state hospital purchases on both local retail activity and employment is thus likely to be quite small for any of the eight state hospitals. While some variation between localities exist for this factor, they are insignificant when compared with many of the other economic impacts described in this report.
VII. EFFECTS OF OTHER HOSPITAL-RELATED SPENDING
Several other hospital-related factors have an effect on the local economy. State hospital residents/patients spend money that would be lost to the community if the hospital closed. Families and other visitors also contribute to the local economy as do volunteer organizations. A 1982 study of the closing of Rochester State Hospital estimated that the combined contribution of these effects was close to the contribution of hospital purchases (Rieder, 1982)
The largest single Factor in this area is resident/patient spending. Individuals have accounts at the hospital to be used for purchases at canteens and local businesses. As part of the Rochester study a survey was conducted which estimated total annual spending of $1,536 per mentally ill patient. Spending by mentally retarded residents was considerably
lower, averaging $439 per year. These figures will be used in this report. No reliable data exist on spending by chemically dependent patients. Following the assumption of the Rochester
Economic Impact Page 27
January 31, 1985
study, it is assumed that persons with chemical dependency spend an equal amount to mentally ill people.
TABLE 15
ESTIMATED SPENDING BY HOSPITAL
Resident/patient spending ranges from an estimated high of $724,000 per year at St. Peter to a low of $221,000 per year at Cambridge. Cambridge and Faribault have relatively low spending because those hospitals serve only mentally retarded people.
Visitors to the hospitals also contribute to the local economy. Based on the Rochester study, it is assumed that each hospital resident/patient received 1.5 visitors per month (a figure that anecdotal evidence suggests may be too high) and that each visit generated twelve dollars in local spending. Residents/patients at hospitals located near larger urban areas, such as Anoka, may receive more visitors, but these visitors probably spend less per visit than those who must travel further. Under these assumptions, visitor spending was found to range from $66,865 per year at Anoka to $161,350 at Faribault (Table 16).
Resident/patient and visitor spending were combined for each hospital. In order to get an idea of the importance of such spending to the local economies these total figures were compared to retail sales in each city (Table 17). Patient-visitor spending as a percent of city retail sales ranged from a high of 13% in Moose Lake to a low of 1% in Brainerd. At St. Peter
patient and visitor spending represented 11% of retail sales while the other five hospitals were at three percent or less.
As with hospital purchases, the loss of visitor and patient spending will also have an effect on job loss. This effect was estimated using the Department of Economic Security model used for hospital purchases. Estimates of increased job loss from this cause range from 21 jobs in St. Peter to 8 in Cambridge (Table 16). These figures probably overestimate job loss, however. Much of the patient spending takes place in the hospital canteen; the loss
Economic Impact Page 28
January 31, 1985
of these jobs has already been calculated in the discussion of direct employment loss (Section IV).
TABLE 16
EFFECTS OF PATIENT & VISITOR SPENDING
VIII. PAYMENTS TO TERMINATED EMPLOYEES
Hospital employees will not experience an immediate cessation of all income upon termination. A variety of programs exist which will help mitigate this impact. Among these programs are unemployment compensation, severance pay, and retirement benefits. The costs of these programs are borne by the state and not the local government.
It is difficult to estimate how these programs will affect each state hospital region. In most cases the effect will probably be proportional to hospital payroll, although some hospitals may have more persons eligible for early retirement than others. The exact impact will depend on any terms that may be negotiated as part of a hospital closure.
One important effect of temporary programs such as unemployment insurance or severance pay is to delay the indirect employment effect resulting from a closure and discussed in Section V. The calculations performed by the Department of Economic Security and presented in Section V reflect this factor.
IX. MIGRATION
The ultimate local economic impact of a hospital closure will depend on the propensity of workers and their families to stay in the area. In general, the more families move out, the more severe the economic impact (although certain factors, such as demand for social services, will be reduced if many people leave the area).
In general, the propensity to move following loss of a job is low. Numerous studies have been completed on migration in areas affected by plant closings and severe unemployment, all of
Economic Impact Page 29
January 31, 1985
which have found relatively little migration following job loss. The recent experience of the Minnesota Iron Range is illustrative. Despite massive unemployment for a long period of time and the prospect that many jobs will never again exist, population declines in Iron Range communities have in general been five percent or less (Office of State Demographer, 1984). Typically, migration following high regional unemployment is slow, involving primarily young people (particularly recent high school graduates and young workers) (Schnitzer, 1970; Buss and Redburn, 1983). Reasons cited for not moving include socialites in the community, the cost and trouble of moving, difficulties in selling a house, and the perceived lack of job prospects elsewhere (Smith and Fowler, 1964).
Two factors will determine the relative economic impact of migration. The first is the proportion of the total area population made up of potential migrants, and the second, the propensity of these potential migrants to actually move. The number of potential migrants will be directly proportional to the direct and indirect unemployment effects described in Sections III and IV. The proportion of potential migrants will be heaviest in the Fergus Falls and Brainerd areas and lightest in Anoka (Table 6).
Several factors affect the propensity to move. First is the perceived prospect for gaining employment in the same community or within a reasonable commuting range. This is discussed in Section X. Second, is the perceived prospect of obtaining employment somewhere else. This factor in large part will depend on state policy regarding transfers and reassignments. The 1984 Survey of State Hospital Employees (State Planning Agency, 1985) asked workers
where they would most have to move in order to work in their preferred field if their current job were to be abolished. An average of 38% reported that they would not have to move or would have to move within their county. This figure varied from 25% at Moose Lake to 59% at Anoka (Table 17).
TABLE 17
"WHERE WOULD YOU MOST LIKELY HAVE TO MOVE YOUR RESIDENCE
IN ORDER TO WORK IN YOUR PREFERRED FIELD? (N = 3,520)
Source: 1984 State Hospital Employee Survey, State Planning Agency
Economic Impact Page 30
January 31, 1985
A third factor is the employment situation of spouses and other family members. A person with an employed spouse is less likely to move than a single person or one with a non-working spouse, particularly if the salary of the spouse is greater than that of the hospital employee. The 1984 Survey of State Hospital Employees was used to determine the number of state hospital employees with other household wage earners. More than half (57%) of the state hospital workers statewide reported other household wage earners, with very little variation among hospitals (Table 18).
TABLE 18
PERCENTAGE OF HOSPITAL EMPLOYEES WITH OTHER WORKERS
IN HOUSEHOLD (N = 3,656)
Source: 1984 State Hospital Employee Survey, State Planning Agency.
Another major factor is home ownership. Renters have an easier time moving than do home owners. Seventy-four percent of state hospital employees own their homes, a figure close to the average for the entire state population. Home ownership ranges from 80% at Moose Lake to 69% at Anoka (Table 19). However, homeowners in an otherwise prosperous area are likely to have an easier time moving than home owners in an area where economic conditions make it difficult to sell a house.
TABLE 19
HOME OWNERSHIP OF STATE HOSPITAL EMPLOYEES (N = 3,557)
Source: 1984 State Hospital Employee Survey, State Planning Agency.
Economic Impact Page 31
January 31, 1985
Finally, the age of the employee will also play a part; younger employees are more likely to move than older ones. The median age of state hospital employees ranges from 33 at Faribault to 40 at Anoka and Fergus Falls. More significantly, Faribault and Cambridge have a much higher proportion of very young (less than 24) workers than other institutions, indicating that migration in these areas may occur sooner than at the other six (Table 20) .
When all factors are considered, there appears to be little difference from region to region in those factors in the propensity to move. On those factors where a difference does exist, indications are that residents of the Anoka region would be most mobile. But Anoka State Hospital employees are also the least likely to feel they have to move in order to find other work. Or a relative basis, therefore, migration will not be a significant factor. On an absolute basis, however, its long-term effect could be severe in all locations except Anoka.
TABLE 20
AGE OF STATE HOSPITAL EMPLOYEES
Source: Department of Employee Relations, August 20, 1984
I. ALTERNATIVE EMPLOYMENT
The severity of economic impact will greatly depend on the ease of finding other work. The availability of other work, particularly work in the employees field, will not be the same from area to area. In general, an employee in a major metropolitan area will have an easier time finding work than one in a small, relatively isolated town.
No estimate of the number of employees at each hospital who could expect to readily find work in the area in their field can be made. There are indicators that show at which hospitals the search for alternative employment would be most and least
Economic Impact Page 32
January 31, 1985
difficult. The first of these is the current unemployment rate in the region. Alternative employment would be more difficult in an area suffering from high unemployment than an area with low unemployment. Second, the most likely source of employment for many hospital employees is the health services industry. The 1980 Census provided a count of the number of persons employed in the health services industry. If hospital jobs account for a large percentage of area health service jobs, we can conclude that obtaining alternative health service employment in the event of a hospital closure will be extremely difficult.
It should be noted that these results only reflect current labor market conditions. Redevelopment of the hospital facility or the development of alternative care facilities for hospital residents/patients would change the job market for hospital employees. These factors are discussed later in this report.
Table 21 presents July, 1984 unemployment rates in hospital counties or regions. Unemployment data are only available on a county basis and cannot be calculated for economic impact areas. In addition, unemployment rates are not tabulated for Anoka and Nicollet Counties. For Anoka State Hospital the unemployment rate for the Minneapolis-St. Paul MSA is the best available. Data for Nicollet County are combined with data for Blue Earth County.
These unemployment rates vary from a high of 10.1 percent in Carlton County (Moose Lake) to a low of 4.7 percent in Blue Earth and Nicollet counties (St. Peter). With the exception of Willmar unemployment rates are presently highest in the areas where the state hospitals account for the greatest percentage of area jobs. A hospital closure in some cases would clearly exacerbate an already bad situation.
TABLE 21
JULY, 1984 UNEMPLOYMENT RATES IN STATE HOSPITAL COUNTIES
Source: Department of Economic Security, November, 1984.
Economic Impact Page 33
January 31, 1985
Hospital jobs account for a very large proportion of all health service jobs in several regions. The greatest percentage is found in the Fergus Falls economic impact area, where the hospital is responsible for 44 percent of all health service industry jobs. Brainerd and Moose Lake hospitals also account for 30 percent or more of their region's health service jobs. Table 22 and Figure 14 list percentages for all eight hospitals. The figure for Faribault reflects the fact that the economic impact area includes Rochester with its substantial health care employment.
TABLE 22
HOSPITAL EMPLOYMENT AS A PERCENTAGE OF ALL EMPLOYMENT
IN THE HEALTH CARE INDUSTRY BY ECONOMIC IMPACT AREA
Source: Bureau of Census, 1980
FIGURE 14
HOSPITAL EMPLOYMENT AS PERCENTAGE OF EMPLOYMENT IN HEALTH CARE INDUSTRY
Source: Department of Employee Relations, 1980 Census.
Economic Impact Page 34
January 31,1985
XI. REUSE OF HOSPITAL FACILITIES
whether or not a state hospital facility is reused will have a major effect on the overall economic impact of a closure. A reused facility will likely generate jobs and may possibly return tax-exempt land to the local property tax rolls. Reuse of hospital facilities can be either public or private. The two hospitals recently closed in Minnesota were redeveloped as public facilities, Hastings State Hospital as a veterans' home, and Rochester State Hospital as a federal prison.
Redevelopment for public use would probably depend little on the location of the facility but would depend extensively on the condition of the hospital buildings and infrastructure. A detailed report on physical facilities was prepared as a separate portion of the state hospital study (State Planning Agency, 1985). The results of this study found that Brainerd, Moose Lake, St. Peter and Willmar State Hospitals had the highest overall ratings for physical condition of main buildings, with Anoka, Faribault, Fergus Falls and Cambridge State Hospitals having lower ratings (Table 23) .
TABLE 23
PHYSICAL CONDITION RATINGS OF STATE HOSPITAL MAIN BUILDINGS
(Includes Vacant Space)
aIncludes 2 cottages which have been identified for demolition. If these 2 were eliminated, the percentage would be 37%.
Source: State Planning Agency, 1985.
Reuse by the private sector is an alternative to another public facility with the added advantage of returning the property to the local tax rolls. However, the hospital buildings are likely to be of little use to a private developer who will probably be interested only in the land and the infrastructure. The potential for private
Economic Impact Page 35
January 31, 1985
sector reuse will be extremely dependent on the location of the campus. Other factors include the surrounding population base, transportation facilities, availability and cost of energy, proximity to industrial markets, the availability of a trained workforce and the degree of amenities found in the area.
The population base can be defined as the population of the economic impact area. These range from 824,000 in Anoka to 22,000 in Fergus Falls. Cambridge and Faribault also have population bases over 100,000 (Faribault because of its proximity to Rochester). In addition, Moose Lake is reasonably close to the more than 100,000 people in the Duluth area, although this area is not included in the economic impact zone (Table 24).
TABLE 24
POPULATION OF ECONOMIC IMPACT AREAS
It is difficult to make a comparative ranking of transportation facilities, but considerable variation from community to community is apparent when evaluated on a qualitative basis. Four state hospitals, Anoka, Faribault, Fergus Falls, and Moose Lake, lie within five miles of an interstate highway. A fifth, Cambridge, is only 14 miles from the nearest interstate. St. Peter while not near an interstate, does lie on a modern four-lane highway (TH 169) with excellent access to the Twin Cities. Highway access, however, would appear to be a serious drawback for Willmar and Brainerd. Access to air transportation is a problem for most locations. Only Anoka lies within 20 miles of a major commercial airport. Brainerd has a
small commercial airport, and St. Peter is about 20 miles from the Small commercial facility at Mankato, but all other hospital sites are at least 35 miles from the nearest commercial air facility. Willmar again, would appear to be at the greatest disadvantage. Rail access, however, is fairly good. While no hospital has a rail siding on site, all communities are on a main line with at least daily service (Department of Energy and Economic Development, 1973 - 74). Anoka, with its proximity to the large rail depots of Minneapolis and St. Paul, and Willmar, which is on the main line between Chicago and Seattle, would appear to rank best on this issue.
Economic Impact Page 36
January 31, 1985
A separate report in this series discussed energy use and costs and found significant variation between energy costs faced by the state hospitals (State Planning Agency, 1985). While there is no guarantee that the energy supplies and prices available to the state hospitals would be available to a private developer, they do provide some guidance as to what energy costs private industry might face at each site. Fergus Falls (1.12 cents/kwh) and Willmar (1.16 cents/kwh) pay the lowest electricity costs of all hospitals. These hospitals purchase electricity directly from a wholesaler the Western Area Power Administration. Moose Lake (6.22 cents/kwh) and Cambridge (6.01 cents/kwh) Pay the most Moose Lake buys its electricity from the local municipal utility while Cambridge buys from a local cooperative.
Heating costs depend on the type of fuel used. Fergus Falls and Cambridge have coal-burning heating plants and pay $1.96/MMBTU and $2.40/MMBTu respectively. The other six hospitals heat with natural gas and pay between $4.14/MMBTU and $4.96/MMBTU. New and more efficient heating plants could be constructed as part of any new development. However, Fergus Falls would appear to have an overall advantage on the other sites regarding energy prices.
TABLE 25
ENERGY PRICES AT STATE HOSPITALS FY '83
Source: State Planning Agency, 1985.
Proximity to industrial markets is, of course, quite important for large scale private economic development. This can be measured in two ways. The first is by the amount of manufacturing employment in the sites county, and the second is by the distance to the major industrial market of Minneapolis-St. Paul and the distance to the nearest city of 50,000 or more. These factors are presented in Table 26. Anoka has a substantial advantage on both factors, with Brainerd, Willmar, and Fergus Falls at a disadvantage.
Economic Impact Page 37
January 31, 1985
TABLE 26
PROXIMITY TO INDUSTRIAL MARKETS
Source: Bureau of the Census, 1984.
Many employers show a Preference for regions with significant scenic, recreational, or cultural resources. Each of the eight state hospital sites has something to offer in this regard. However, Anoka stands out because of its proximity to the cultural advantages of the Twin Cities area, while Brainerd is located in the center of a popular resort area.
The availability of a trained labor force is quite important to industrial development. The number of trained workers available is directly proportional to the area population, meaning
that Anoka would have a considerable edge in this regard. However, if we view redevelopment as a means of supplying jobs for workers displaced by the closing of a state hospital it is important to realize that these workers have few skills that would be of use outside of the health care industry and would require retraining for other occupations.
In summary, Anoka would appear to present by far the greatest opportunity for industrial development. The other seven locations each have a unique combination of advantages and disadvantages that make relative ranking impossible. One strong cautionary note must be sounded however. While 31 state hospitals have closed nationwide, none have been redeveloped by private industry (State Planning Agency, 1985) . One has been sold to a religious group and one to a college, but the rest have been reused by the public sector.
Economic Impact Page 38
January 31, 1985
XII. LOCAL GOVERNMENT HEALTH CARE EXPENDITURES
While most of the funding for state hospitals comes from the state and federal governments, a portion does come from county government. As the closing of a state hospital may affect the amount that hospital's county pays for treatment of its mentally retarded, mentally ill, or chemically dependent citizens, local economic impact is possible. A discussion of this issue will be presented in other reports.
XIII. SOCIAL SERVICE COST TO LOCAL GOVERNMENTS
The large increase in local unemployment which will follow a state hospital closing will certainly place demands on the provision of social services ranging from emergency food shelves to mental health counseling. These services are provided by the state, county government, city government, and private organizations. However, it is likely that the greatest impact will fall at the county level. While we cannot directly estimate the cost to county government of providing these services, we can look at relative impact by tabulating the number of state hospital employees living in each county (Table 27).
TABLE 27
STATE HOSPITAL EMPLOYEES BY COUNTY OF RESIDENCE
Note: Counties with fewer than 15 state hospital employees are not listed. Some counties include employees from more than one state hospital.
Source: Department of Employee Relations, August 7, 1984.
Economic Impact Page 39
January 31, 1985
The impact on local social services will also depend on the speed and likelihood of obtaining alternative employment, indirect employment effects, and migration. However, it seems that at least 12 counties would experience significant strain on their social service agencies should the hospital in their area close. Rice County would be by far the hardest hit due to the large workforce at Faribault State Hospital. Crow Wing (Brainerd), Ottertail (Fergus Falls), and Kandiyohi (Willmar) counties would also experience particularly severe effects.
The strain on city and private services would also depend on the concentration of the hospital workforce. The impact would likely be greatest at a community such as Fergus Falls, where a high percentage of the workforce lives in the city, than in a community such as St. Peter or Anoka, where the workforce is spread over a larger area, incorporates several counties, and represents a smaller overall share of the area population. The concentration of workers in the primary impact zone (Table 2) provides a good guide to the relative burden on city and private services.
XV. EFFECT ON LOCAL TAX REVENUES
In general, closing of a state hospital would reduce local tax receipts. The effect, however, would likely be minor. The principal source of revenue for local governments is the property tax. Property tax revenues will decline if households move from the area to seek employment elsewhere and abandon (rather than sell) their property, if the delinquency rate increases, or if the pressures on the real estate market force housing values down. Again, the greater the concentration of hospital employees, the more likely a deleterious effect on local government revenues.
Increased tax revenues could result if alternative use of the hospital facilities (now tax exempt) results in the property placed on the tax rolls. This, however, presumes private sector rather than public sector development of the property. Section XI of this report deals with the prospects for private redevelopment.
The existence of a large facility such as a state hospital inevitably makes some demands on local government services, and the closing of a hospital could result in savings to the local government. In practice, however, savings are not likely to be significant. Local governments do provide police and fire protection to the state hospitals, but the hospitals reimburse
the local governments for these services. Roads leading to
Economic Impact Page 40
January 31, 1985
state hospitals are state highways; the state is responsible for their maintenance. In some cases, the hospitals purchase electricity, water or sewer services from municipalities. Any economic impact would occur only if the hospitals were subsidizing or being subsidized by other users, or if the closure of a state hospital enabled the municipal utility to avoid future capital expenditures. These factors are extremely difficult to assess and are unlikely to be significant.
XVII. SUMMARY
In the case of seven of the eight state hospitals, the local economy will be hard-pressed to accommodate the increase in unemployed workers. The economy of the Twin Cities area is probably sufficient to accommodate any change in Anoka State Hospital.
In addition to the direct loss of hospital employment, a significant loss in non-hospital employment will follow the closing of a state hospital. The full effect of this loss will occur over a ten year period.
On a local level, state hospital employment has the greatest impact on the Moose Lake area, where 19 percent of the work-force is employed by the hospital. In the Anoka area, less
than one percent of the workforce is employed by the hospital. In the remaining six communities the state hospital accounts for 4 and 9 percent of the workforce.
On a regional level, the effect of a state hospital closure would be greatest in the Fergus Falls region, where eight percent of the workforce would be unemployed as the result of a hospital closure. The impact would be the least in the Anoka region, where only one-tenth of one percent of the workforce would be affected. In the other six communities between two and six percent of the workforce would be affected.
Areas affected by the closing of a state hospital are not likely to experience substantial population losses in the short run. However, a long range population decline is
likely unless the local economy is strong (as in the case of Anoka) or redevelopment efforts are successful.
Purchases made by the state hospitals have little effect on the local economy.
With the exception of Anoka, the prospects for local reemployment of state hospital employees in the health services industry without the development of new jobs are not good.
Economic Impact Page 41
January 31, 1985
Redevelopment of hospital facilities for public sector use will depend on the condition of buildings and infrastructure. Brainerd, Fergus Falls, Moose Lake, and St. Peter are currently in the best condition.
For private redevelopment the hospital buildings are probably of little use, although the land and infrastructure have value. Private redevelopment is highly dependent on location.
Anoka State Hospital is the most likely site for private redevelopment. However, there is no precedent for private redevelopment of a hospital campus in the U.S., based on recent survey results.
XVIII. REFERENCES
Bureau of the Census. (1984). County business patterns, 1982 Minnesota. Washington, D. C.: U.S. Government Printing Office.
Department of Energy and Economic Development. (1973 - 74). Community profiles. St. Paul, MN: Author.
Department of Human Services. (1979, December). Computer tabulation of Rule 53 cost reports. St. Paul, MN: Department of Human Services, Audit Division.
Department of Human Services. (1983) . Fact book: State hospitals and nursing homes.
St. Paul, MN: Department of Human Services, Mental Health Bureau
.
Department of Revenue. (1984). Historical retail sales data for Minnesota counties and selected cities. St. Paul, MN: Department of Revenue, Research Office.
Erickson, R. A. (1977, July). Sub regional impact multipliers:
Income spread effects from a major defense installation. Economic Geography. 57(3),
283-294.
Fergus Falls State Hospital annual reports; Fiscal year ended June 30, 1983 and fiscal year ended June 30, 1982. Fergus Falls, MN: Fergus Falls State Hospital
Gibson, L. J. & Worden, W.A. (1981, April) . Estimating the economic base multiplier: A test of alternative procedures". Economic Geography, 57(2) 146-159.
Moore, C. (1978). A new look at the regional trade multiplier. Northeast Regional Science Review. III, 164.
Economic Impact Page 42
January 31, 1985
Pinola, R. & Graner, W. (1984, November). Estimating the effects on county unemployment of shutdown of the state-operated hospitals. St. Paul, MN: Department of Economic Security.
Rieder, Mary. (1982)..Economic impact of the closing of Rochester State hospital on Rochester and the region. Winona, MN: Winona State University.
Schnitzer, M. (1970). Regional unemployment and the relocation of workers. N.Y.: Praeger.
Smith, L. M. & Fowler, R. A. (1964). Plant relocation and worker migration. In A.B. Shostak and W. Gornbey (Ed). Blue collar world: Studies of the American worker. N.Y.: Prentice-Hall
State Planning Agency. (1984). 1984 state hospital employee survey. St. Paul, MN: Author.
State Planning Agency. (1985, January). Policy Analysis Series Paper No. 1: Minnesota State Hospital facilities and alternative use. St. Paul, MN: Developmental Disabilities Program,State Planning Agency.
State Planning Agency. (1985, January) Policy Analysis Series Paper No. 2; Minnesota State Hospital energy use and cost. St. Paul, MN: Developmental Disabilities Program, State
Planning Agency.
State Planning Agency. (1984, September). Minnesota population estimates 1983. St. Paul, MN: Office of State Demographer, State Planning Agency.
Treyz, G. I., Erlich, D. J. & DePillis, M. S. Jr. (1983). The Minnesota forecasting and simulation model (MNFS-53). Amherst, MA: Regional Economic Models, Inc.
Willmar State Hospital annual reports: Fiscal year ended June 30, 1983 and fiscal year ended June 30, 1982. Willmar, MN: Willmar State Hospital.
Economic Impact Page 43
January 31, 1985
APPENDIX A
TECHNICAL DOCUMENTATION OF ECONOMIC MODELS
I. Department of Economic Security
Dr. Rudy Pinola of the Research and Statistics Division, Minnesota Department of Economic Security provided estimates of direct and indirect unemployment resulting from the closure of
state hospitals. Dr. Pinolass study was based on data at the county level, including all counties where at least six state hospital employees resided. Adaptation of this method to economic impact areas was done by state hospital study staff.
Two different scenarios were employed in developing unemployment estimates. The first was a "worst case" scenario, assuming that all employees of a closed hospital, as well as indirectly affected non-hospital workers, would be unemployed. The second, "reasonable case", scenario assumes that retirements, transfers to other state service, labor force drop-outs, private sector job transfers, need for skeleton crews, and adjustments for multiple job holders would mitigate the immediate effects on employment. Indirect job losses would also be added in this scenario. Besides estimating immediate unemployment effects, the study includes estimates of unemployment after 27, 52, and 104 weeks have passed.
Explanation of Variables
Retirement. Retirees would move from "employment" labor force" to "not in the labor force."
The report combined data from the State Planning Agency and the Minnesota State Retirement System to develop a matrix of state hospital employees by county of residence, age group, and length of service. It was assumed that none of the employees less than 55 years of age would choose to retire. For the age group 55 through 61, it was assumed that 50% of those workers eligible to retire would do so. Of the age group 62 and over, it was assumed that 75% of those eligible would retire. These assumptions resulted in an estimate of 98 retirees out of a total of 5821 employees.
Transfer. The labor force status of employees who transfer to other positions within the state system would not change. The study assumes that 3.5% of state hospital employees would transfer within the state system. This is an arbitrary but conservative assumption, resulting in a total of 204 transferees among all eight state hospitals.
Accept Jobs in the Private Sector or Drop Out of the Labor Force (LFDO). The labor force status of those going to private industry would not change. LFDOS do not affect the unemployment rate as
Economic Impact Page 44
January 31, 1985
much as those who lose their jobs and wish to be re-employed, because the former are not in the labor force and are therefore not considered unemployed. It was assumed that five percent of state hospital employees would join the private sector, and five percent would be LFDOS.
Skeleton Crew. It was assumed that 15 people would be needed at each state hospital as monitors until disposition of the buildings and grounds is completed.
Multiple Job Holders. Following Bureau of Labor Statistics (BLS) standards? those people who might be laid off from state hospital jobs, but who hold more than one job, would still be considered employed. Because reliable state data do not exist, national data from 1981 were used. At that time, 4.9% of employed people held more than one job. It was decided to use the more conservative figure of 4% of state hospital employees as multiple job holders.
Unemployed 27 Weeks or Longer. To determine how many displaced workers would still be unemployed six months following the closure of a state hospital, national data from the Current Population Survey were reviewed. Duration factors were developed for three ranges of unemployment levels: low (6% and under), medium (6.1% through 9.1%) and high (9.2% and higher). In February, 1980, the U.S. actual unemployment rate was 6.0%. Six months later 18.2% of "other job losers" (as distinct from "job losers... on layoff") were unemployed 27 weeks or longer. The middle range factor was 25.6%, based on a U.S. rate of 9.1% in May, 1982, while the high range of 28.0% was based upon the highest U.S. rate in modern times of 11.4% in January, 1984. The relevant duration factors were then used to determine the number of unemployed workers in each county at the time of closing who would still be unemployed 27 weeks or later.
Unemployed 52 Weeks and 104 Weeks. A Bureau of Labor Statistics procedure for "surviving" unemployment compensation "exhaustees" was used to calculate the unemployed after 52 and 104 weeks. Each year BLS supplies a factor which is applied weekly to the pool of people who have exhausted their Unemployment Insurance benefits to determine how many people remain in the pool of unemployed with the passage of time. The 1984 factor of 1962, the highest experienced, was used in this study.
Estimating Indirect Income Effects. In order to estimate the indirect effect on employment of a state hospital shutdown, the notion of a multiplier effect on reverse income generation was used. In addition, it was assumed that the full indirect effects would not be felt instantaneously, but would begin to show up one to two years after a closure of facilities.
Economic Impact Page 45
January 31, 1985
To develop the reverse income multiplier, it was necessary to first estimate the loss in income generation that results from the outflow of sales receipts to purchase goods from outside the area. To estimate this outflow; the retail mark-up on various goods normally purchased in the community was researched. It was felt that an average mark-up on all items of the consumer market basket would amount to 40 percent, leaving a loss of 60%. The reciprocal of this loss, 1/.60, is the estimated income multiplier, 1.67. The income multiplier could then be used to determine the decrease in income from a reduction in expenditures in the community by the State for state-operated hospitals. This would mean that if state financed payrolls in a community fell by $200,000, the ultimate effect on overall income generation would be a reduction of $200,000 X1.67, or $334,000, of which $200,000 is the direct and $134,000
the indirect effect due to the multiplier.
Estimating the Indirect Employment and Unemployment Effects. The
indirect portion of the income loss noted above represents income support to jobs that are indirectly supported by state expenditures through operation of the multiplier. Having estimated the indirect income effect of a reduction in state expenditures, it is left to determine the indirect effect on employment that would result from such an income effect. The indirect effects of a reduction in state expenditures on state hospitals is likely to be concentrated in the service sector. It was assumed that $40,000 in annual spending would be needed to support a position in the service sector.
Using the above example to illustrate the indirect employment effect, there would be an indirect reduction in income generation of $134,000. Dividing this amount by $40,000 would result in a loss of 3.35 positions in the service sector. The indirect loss in employment was adjusted in the same way as the direct loss to reflect retirees, transfers, and multiple job holders.
Finally, it was noted that the full reverse multiplier effect would not be instantaneous. It was therefore assumed that 50 percent of the multiplier effect would occur one year after closure of a facility, and an additional 30 percent in the second year, providing close to a worst case scenario.
II. Minnesota Forecasting and Simulation Model (MNFS-53) - Department of Revenue
The Research Division of the Minnesota Department of Revenue applied the Minnesota Forecasting and Simulation model (MNFS-53) to forecast the economic changes brought on by various government policies. MNFS-53 is based on a model originally developed by Treyz, Friedlanender, and Stevens (1983) for the Commonwealth of
Economic Impact Page 46
January 31, 1985
Massachusetts. The TFS model structure is based on econometric theory and uses data and parameter estimates which are valid for all regions of the U.S. However, specific coefficients and time series for the model of Minnesota are based on Minnesota data only.
The forecasts generated by MNFS-53 are highly detailed, including prices, local-national relative business costs, employment, wage rate changes, residential and nonresidential investment, the consumer price index, and personal income. Generating the forecasts involves the simultaneous solution of a large number of equations, which can be divided into three sets: demand and supply linkages, cost linkages, and wage determination linkages. Within each set of equations, a change in any one variable will reverberate throughout the entire model.
The third set of equations mentioned above, Wage Determination Linkages, is the one directly applicable to the study of state hospital closings. A change imposed from outside the system, in this case a decrease in hospital employment, will cause changes throughout the system. Population will shift as wage rates change in response to the decreased demand for labor in the industry. Changes in the number of available workers will also exert counter pressures on wages. As the changes move through the system over time, a new equilibrium level of employment will be reached in the region. It is the difference between this new level and the original level of employment which is the indirect employment effect forecast for a state hospital closure.
III. Regional Multiplier Model
In this model a different trade multiplier was calculated for each state hospital economic impact area. The first step in the process was the determination of an impact zone around each hospital. Once the impact zone was defined, a multiplier was calculated for each area. The direct income loss to a region as a result of the closure of a state hospital was then determined. When the multiplier for a region was applied to direct income loss in that region, the additional, indirect income loss was found. Finally, the indirect loss in area income was translated into indirect loss in employment.
Calculating the Multipliers. The multipliers are based on the formula M(i) = 1/ (1-Kb)
where M(i) is the multiplier for region i, K represents the proportion of a dollar of local consumption which is retained in the region, and b is the propensity of households to consume
locally.
Economic Impact Page 47
January 31, 1985
In order to determine M, it is necessary to find Kb. Moore (1973) estimated that Kb=-6.67 +11.9(log(10) population). In a later work, Erickson (1977) found that the Moore equation was appropriate for larger urban areas, but that for areas with population under 50,000, Kb=-42.77 + 19.03(log(10) population) gave a more appropriate multiplier. The Moore equation was therefore used in calculating the multiplier for the impact area around Anoka State Hospital, while the Erickson equation was used for the remaining seven state hospitals. The population for each area was based upon zip code data. The number of persons residing in each of the zip code zones included in the impact area around each hospital was found, and the total for each area was substituted into the above equations for Kb. The results of those calculations were in turn substituted into the equation for M(i), resulting in a multiplier for each hospital region. Multipliers ranged from 3.3 in the Anoka region to 1.7 at Fergus Falls.
Determining Direct Income Loss
The direct income loss to an area due to the closure of a state
hospital is primarily comprised of the wage and salary income lost by laid-off hospital workers. For this study, a list of employees' bi-weekly net pay by zip code was used. Their zip codes in this list were matched with the list of numbers of employees by zip code to obtain the total direct income loss for each zone within each economic impact area. The direct income loss was then converted from a bi-weekly to an annual basis.
Calculating Indirect Income Loss. Besides the direct loss of wage and salary income due to closure of a facility in an area, there is an additional, indirect loss due to the "multiplier effect". To determine the extent of the indirect loss, the multipliers for each region were applied to the direct loss for each zone within that region. As an example, the multiplier for the Willmar area was calculated to be 1.79. If the shutdown of the Willmar State Hospital resulted in lost wages of $200,000, the total income loss in the Willmar area would be $200,000 x 1.79, or $358,000. Of this total, $158,000 would be the indirect loss. This income would have supported non-hospital jobs in the area.
Translating Indirect Income Loss into Indirect Employment Effects. Following the example of the Department of Economic Security model discussed in Part I of the appendix, it was estimated that $40,000 in spending per year was needed to support one service sector job. For each zip code area, total indirect income loss was divided by $40,000 to find the number of non-hospital positions which would be lost in the area if a hospital closed. Finally, the indirect job losses for each zip code area were summed for the entire hospital region to give total indirect job loss as a result of closing each hospital. The number of jobs lost ranged from a high of 365 in the Faribault State Hospital economic impact area to 113 in the Moose Lake economic impact area.
Economic Impact Page 48
January 31, 1985
APPENDIX B
Results using alternative impact areas for Faribault and Moose Lake
During the review and comment period for this report, the Coalition of Concerned Citizens for Moose Lake State Hospital and representatives of Faribault State Hospital suggested Cloquet and Rochester not be included in their respective economic impact areas. The following tables and figures reflect some of the results found in this report had they been compiled using the area boundaries suggested by these groups. Time limitations prevented recalculation of the multiplier analysis to conform to these boundaries. In each case, the absolute indirect unemployment effect would decrease as the multiplier would be applied to a smaller base.
TABLE 1 (Adjusted)
ALTERNATIVE ECONOMIC IMPACT AREAS
TABLE 2 (Adjusted)
STATE HOSPITAL EMPLOYMENT AS A PERCENT OF TOTAL AREA EMPLOYMENT
Economic Impact Page 49
January 31, 1985
HOSPITAL EMPLOYMENT AS PERCENTAGE OF TOTAL EMPLOYMENT - ENTIRE IMPACTED AREA
Note: Economic impact areas as defined by representatives from Faribault and Moose Lake State Hospitals.
Source: Employee Relations, 1980 Census.
TABLE 3 (ADJUSTED)
HOSPITAL PAYROLL AS A PERCENT OF TOTAL AREA WAGE AND SALARY INCOME
Source: Department of Employee Relations, 1980 Census
Economic Impact Page 50
January 31, 1985
Economic Impact Page 51
January 31, 1985
TABLE 22 (Adjusted)
HOSPITAL EMPLOYMENT AS A PERCENTAGE OF ALL EMPLOYMENT IN THE HEALTH CARE INDUSTRY BY ECONOMIC IMPACT AREA
FIGURE 14 (ADJUSTED)
HOSPITAL OF EMPLOYMENT IN HEALTH CARE INDUSTRY
Note: Economic impact areas as defined by representatives of Faribault and Moose Lake State Hospitals.
Source: Department of Employee Relations, 1980 Census.
Economic Impact Page 52
January 31, 1985
The 1984 Legislature mandated that a study and plan for Minnesota State Hospitals be prepared (Chapter 654, Section 19).
An Institutional Care and Economic Impact Planning Board was created composed of the following state agency heads: Sister Mary Madonna Ashton, Dept. of Health; Barbara Beerhalter, Dept. of Economic Security; Gus Donhowe, Dept. of Finance; Bill Gregg, Dept. of Veterans Affairs; Sandra Hale, Dept. of Administration; Leonard Levine, Dept. of Human Services; Orville Pung, Dept. of Corrections; David Reed, Dept. of Energy & Economic Development; Nina Rothchild, Dept. of Employee Relations; James Solem, Housing Finance Agency; and Tom Triplett, Chair, State Planning Agency.
Responsibility for the studies was given to the Developmental Disabilities Program/Council of the State Planning Agency.
Eight technical papers have been prepared to respond to the legislative requirements. This paper may be cited:
State Planning Agency. (1985, January). Policy Analysis series Paper No. 4: The economic impact of Minnesota State Hospitals. St. Paul, MN: Developmental Disabilities Program, State Planning
"Residents" refer to people with mental retardation who live in state hospitals.
"Patients" refer to people with mental illness and people with chemical dependency who receive services at the state hospitals.
Additional free copies of reports or information about this project can be received from:
