Are subsidies worth it?: How to calculate costs and benefits of business incentives.
Governments rarely have good information about the actual costs or benefits of a business incentive package. When negotiating incentives for a firm expansion or new location, it is difficult to assess the long-term effects on the local economy as a whole or to the government's own fiscal condition, and government officials admittedly often do not know the real cost of what they are putting on the table or the value of what the firm is offering. This may result in offering firms more than the ultimate benefits are worth or offering too little. In the summer of 1992, the catalog clothing company Spiegel rejected the City of Chicago's final offer of incentives to keep the national warehousing and distribution center in Chicago and decided to leave for Columbus, Ohio, taking more than one thousand jobs with it. Public criticism was swift and strong - why did the city not offer more? Was its package too little, too late? This reaction was quite a contrast with the public outcry a few years earlier, when the State of Illinois managed to retain the headquarters of Sears' merchandising group through several hundred million dollars worth of free land, infrastructure improvements, and job-training dollars - generally characterized as a "corporate giveaway."
Spurred on by the criticism, the Chicago Department of Planning and Development (DPD) asked the Center for Urban Economic Development at the University of Illinois at Chicago (UICUED) to develop a methodology that would allow the department to assess in advance the costs and benefits of any particular subsidy package. This article presents the methodology, which is now being used by the department in a computerized spreadsheet format. It first discusses the experience of other cities and then focuses on how the key theoretical issues in assessing costs and benefits are resolved. Finally, it describes how the spreadsheet works and what would be required to make it applicable in other cities.
Chicago is not alone in how it negotiates subsidy packages. A survey conducted by the UICUED turned up no cities that used a comprehensive cost-benefit analysis to determine the level of business subsidies to be given. Of course, all cities require companies to provide estimates of jobs created and retained or produce their own estimates. Several cities, such as Phoenix, Milwaukee, and Columbus, also require targeted hiring to increase the economic benefits to low- and moderate-income persons or city residents. Most cities do not analyze secondary benefits. San Jose and Philadelphia sometimes examine increased tax revenues, as does Jacksonville, which also looks at other categories of benefits, such as the increase in services to poverty areas. Columbus tries to estimate jobs created indirectly by a project but does not use this in determining the amount of subsidy to be given. Thus, the state of the art in evaluating possible costs and benefits of subsidies, especially in advance of the investment, is quite limited.
At its most basic level, cost-benefit analysis compares two situations: the city with the project and the city without the project. Economic models, whether formal ones or consisting of educated hunches of program administrators, must be used to forecast the most important economic variables both with and without the project. For Chicago, the authors decided to build cost-benefit analyses around simulations from the model developed by Regional Economic Modeling, Inc. (REMI). This is an econometric model, available for many counties and states, of the regional economy, based on an input-output model and a larger set of equations about the local economy. Other cities could use their own local versions of the REMI model or a different model. Even starting with a sophisticated regional model, there was a series of technical and conceptual hurdles.
The first question to be answered in developing a methodology is to define the population for whom costs and benefits are to be calculated. To a certain extent, this is a policy question rather than a theoretical one, and DPD decided to count benefits to any current resident or potential residents who migrate to the city to take a job created by a subsidy. This also implies that benefits to suburban residents are excluded. It also means that subsidies that merely improve firm efficiency, without thereby increasing payroll or without saving jobs that otherwise would be lost, do not create relevant benefits.
In addition to general economic costs and benefits, DPD also wanted to know the distribution of the benefits over different income categories and a separate analysis of the fiscal effects on city government.
Finally, many economic development projects produce important benefits other than jobs or taxes. These benefits - and possible costs - often relate directly to the location of the project. Putting these neighborhood spillover effects, such as increased neighborhood confidence, improved appearance, or increased congestion, in dollar terms is often misleading and limits the analysis to easily quantifiable effects. Therefore, a separate index of neighborhood spillover effects was designed. Because of its experimental nature and space constraints, the authors do not discuss this index in this article; the technical references contain a complete discussion of their approach.
Thus, the final product of this approach [TABULAR DATA FOR EXHIBIT 2 OMITTED] consists of four major parts: 1) an overall cost-benefit analysis, 2) a fiscal impact statement, 3) a distributional analysis of project benefits, and 4) a neighborhood spillover index.
New or retained jobs are at the heart of the economic effects of business incentives. The first question is whether the investment or activity would be undertaken by the proposing firm or developer in the absence of public subsidy - the "but for" question. In the case of a request for a loan, direct evidence of a firm's inability to do the project without assistance is the refusal of a bank to make a loan. In that case, public support clearly provides gap financing. If private credit is available but the firm is being offered incentives by other places, a subsidy may similarly be needed to keep or attract the firm. Finally, a firm may be able to show that its costs would be much lower elsewhere and thus needs a subsidy to cover the difference. The spreadsheet model used here contains simple questions to ascertain whether any of the above conditions pertain.
Just answering the "but for" question is not enough, though. It may well be that the firm requesting assistance might not go ahead without subsidy, but another firm in the industry might well take its place. In many industries, especially those serving local markets, such substitutions are to be expected.
The model deals with this issue primarily by looking at the industrial sector (standard industrial classification) of the firm. If most of the output of firms in the sector is exported outside of the city, the firm is much less likely to be competing with other local firms. Thus, it is assumed that the new or expanding firm divides its output between local and out-of-city customers in the same proportion as the industry as a whole, as indicated in the REMI model. (If firm-specific information is available, it can be used in the spreadsheet). The export share of any expansion or new facility is then counted as new employment to the area. The only exception to this is if the industry already is growing very rapidly locally compared to the nation. In that case, it is assumed that investment in the sector already is recognized as profitable by the private sector and does not need additional subsidy.
Much of the business of new or expanding firms will be oriented toward local markets. New jobs for Chicago residents should not count those that would otherwise appear in other Chicago firms in the same industry. Thus, even if a firm meets the "but for" standard, the spreadsheet model should not give full credit for jobs oriented to local production, since other local firms are likely to fill the hole in the market. Since firms on the periphery of the city are competing in part with suburban firms, however, the model considers 30 percent of locally oriented economic activity on the periphery as export to the rest of the metropolitan area; in the central district, only 10 percent is counted as exports, and the model gives credit for the associated jobs.
Estimating the number of jobs retained by a project is more difficult than estimating the new jobs. Research on relocating firms points to space constraints and ownership of existing facilities as the most important determinants of moving. In most cases, expanding firms are constrained from staying in the city by the difficulties of assembling land, environmental problems, or other externalities. In the spreadsheet, information on space constraints and ownership is entered to determine the likelihood that a firm would actually move. If the firm is export-oriented, the analysis also takes into account projected growth rates to determine the likely extent of a substitution effect by other producers in this industry. This information is then used to determine whether credit should be given for job retention.
Jobs, Personal Income, Fiscal Gains
To calculate the effect of new or retained jobs on personal disposable income, one first has to estimate the multiplier effect generated by the expansion of direct employment. For this, the REMI model is used. Its output, already included in the spreadsheet model, provides a predicted figure for the change in total employment (direct plus indirect), the total change in wage bill, and the occupational structure of both the direct and indirect jobs.
While all the direct jobs are located in Chicago itself, not all people holding them will be Chicago residents. In addition, some of the indirect jobs, and many of the workers holding those, will be in the suburbs. The model has the ability to specify the share of direct jobs held by Chicagoans. For example, this would be 100 percent if the company has promised to hire only Chicago residents. lacking such specific information, the model assumes that for each occupation, the place of residence of direct workers is distributed between the city and the suburbs in the same way that all Chicago workers in that occupation are distributed, according to the U.S. Census. For indirect jobs, it is assumed the workers are distributed in the same way that all workers in the county are distributed.
With these residence distributions, the total change in the wage bill for Chicago residents can be estimated; however, a few more adjustments must be made before the number is meaningful. First, the transfer payments that will be lost because of the increase in employment must be subtracted. Second, there needs to be adjustments for the fact that many of the workers could have done equally well in some other job, in Chicago or elsewhere. Recent studies show that as many as one-half to two-thirds of new or retained jobs are held by potential outmigrants or new immigrants from outside of the city. Higher-skilled workers are especially likely to have access to jobs elsewhere. Thus, the model discounts income gains by the top two occupational quintiles by 75 percent and for the middle quintile by 50 percent. For the bottom quintiles, the opportunity cost is equal to the transfer payments that are given up when employment expands. All of this information is obtained from the REMI data on the relation between occupational structure and income distribution.
While general economic benefits are important, city government also has to be concerned about its solvency as a fiscal entity. Thus, the model also calculates the specific fiscal gains attributable to the project in order to compare it to the costs. In Chicago, property taxes and local sales taxes vary quite directly with disposable income even though they are not, of course, levied against income. As this ratio remained fairly stable at 5.7 percent during the 1980s, this historical ratio was used in the model. Non-tax local revenues, such as fees for services, are assumed to be offset by corresponding increases in expenditures and were not included as either a cost or a benefit.
Originally, the model builders anticipated that DPD staff might estimate the opportunity cost of different funding sources, such as Community Development Block Grant funds or local funds, at different rates. Occasionally this is indeed the case, especially toward the end of the fiscal year when funds may be lost altogether if left unspent. On the whole, however, staff felt that there were no grounds to value one possible use over another. The exception is the State of Illinois' Community Services Block Grant, which can only be used to create jobs for very low-income individuals and has been difficult to spend down. Therefore, the model gives this program an opportunity cost of zero and counts all other programs at full value.
The actual cost of making loans then is the difference between the interest rate the city would have to pay to borrow funds and the rate received, adjusted for expected losses due to default. Additional costs are the income foregone through tax credits or in the form of tax increment financing payments and the administrative costs for city government in processing and monitoring the project. On the other hand, if some infrastructure spending merely reflects expenditures that the city would have had to make in a few years anyway, they should be counted as costs foregone in the year in which they would otherwise have taken place.
A seven-year time horizon was chosen for measuring costs and benefits. Given the uncertainties of the marketplace, it seems reasonable to require projects to show positive benefits within seven years. The discount rate used for private funds was the real long-term mortgage rate and for public funds, the long-term municipal bond rate.
Although this discussion may seem complex, even though it is only a summary of the full technical description of the model, to run the spreadsheet a project analyst only has to answer 12 questions and enter basic numbers about the financing terms and the sources of funds. Everything else is contained in the model. The accompanying case study (an amalgam of several city-sponsored projects) shows the output of the spreadsheet, which runs on Quattro-Pro 4.0. The key numbers are the summation of the total benefits minus costs and the total fiscal benefits minus costs.
Applicability to Other Cities
As it now exists, the model is valid only for Chicago, since it contains Chicago multipliers, tax rates, occupational distributions, industry characteristics, and other data. The basic logic and approach to theoretical issues, however, should be applicable anywhere. Customization for other cities would require the use of parameters from the appropriate local econometric models and other local data.
The model can be improved in many ways. For instance, more research on the extent to which different types of firms and industries export or serve only the local economy would strengthen understanding of whether jobs are really new to the economy or not. Also, there may be other questions that need to be asked to determine whether a firm really needs the subsidy or is truly likely to move and take jobs with it if no subsidy is given.
As long as theories of local economic development and firm location remain relatively tentative, any model built on those theories necessarily rests on assumptions that can be debated and improved. The authors, however, already have found that the use of the model, and the questions to which it requires answers, force analysts to clarify their assumptions and have indeed brought out different perspectives among key decision makers.
Thus, the model is an aid in analysis rather than the only tool needed to make decisions about business subsidies. Whether and how much of a subsidy will be given will continue to depend in part on the nature of the negotiation game between firms and governments and, especially for large projects, on political considerations. Use of the model, however, will clarify the assumptions of the analyst and give guidance about costs and benefits.
The usefulness of the model also depends on the political and organizational context. If decision making about business incentives is highly politicized, use of the model may be seen as a threat. Similarly, if the model is used by a research or evaluation unit to check on the work of business finance specialists, the latter may perceive it as an attack on their autonomy and therefore seek to attack the model's accuracy. To avoid such problems to the extent possible, the Chicago model was developed in close collaboration and consultation with all staff who were likely to use it.
RELATED ARTICLE: CASE STUDY: ASSOCIATED PAPER FORMS, INC.
This business forms manufacturing company applied for a loan to purchase new equipment and build a warehouse addition to its current plant. The total package looked as follows:
* $450,000, commercial lender;
* $375,000, city loan; and
* $75,000, company equity.
The city used Community Development Block Grant funds for the loan at 75 percent of prime for seven years. Collateral and personal guarantees were adequate.
The first three questions on the spreadsheet (Exhibit 1) ascertain whether the "but for" condition holds: Is there evidence that the financing is gap financing? Are there competing incentives from other locations? Does Chicago have a demonstrable cost disadvantage? Only one of these has to receive a "yes" answer for the model to assume that indeed the firm would not go ahead with the project without public assistance; in this case, the bank indicated gap financing was essential due to the company's liquidity position.
The next set of questions pertains to job retention. In this case, no claim is being made that jobs would be lost if the deal fell through. The company owns its building and has room on site for expansion, making it less likely to move. There are no pressing problems in the neighborhood. The company is locally based and has no alternative production locations. There is no evidence that the firm will necessarily lose market share in the near future if it does not invest in new technology, and the firm does not meet the very strict federal guidelines related to job retention. Thus, the model will give no credit for job retention.
The two-digit standard industrial classification (SIC) code and information about the location of the firm are entered next. The model uses this to calculate how many of the new jobs would be likely to be created by other local firms in the industry, even if this particular firm did not. Finally, the analyst enters information on the number and expected occupational and residential distribution of employees (not shown in Exhibit 1). The firm has said low- and moderate-income individuals will be hired but not necessarily only Chicago residents. All other information to be entered relates directly to the financing itself: the terms of the loan and the amounts and sources of funds.
Based on REMI data and other information contained in the model, the spreadsheet now calculates all the other numbers shown. (See Exhibit 2. The exhibit only shows present values and first-year data; the model actually presents data for all seven years). For instance, based on the projection of 32 project-related jobs, the model counts 29 genuinely new jobs - based on the fact that about 10 percent of the output in this industry is oriented towards the local market. The output also shows that 32.5 indirect jobs will be created by these direct jobs and that 37.7 of all the jobs will be held by Chicago residents. This yields $963,000 in new income per year (measured in 1994 dollars), which after adjustments for taxes and workers' opportunity cost is reduced to $683,000. The distribution across income categories is shown as well.
COST-BENEFIT EVALUATION SPREADSHEET
The UIC Center for Urban Economic Development Cost-benefit Evaluation Spreadsheet
NAME: Associated Paper Forms 01-Mar-94
SECTOR A-1: BASIC INPUT DATA
1. Gap financing? (yes=1, no=0) 1 2. Competing incentives? (yes=1, no=0) 0 3. High Chicago costs? (yes=1, no=0) 0
Consider: a. Wages b. Taxes c. Utilities d. Input costs e. Distribution costs
4. Demonstrable space constraint? (yes=1, no=0) 0 5. High neighborhood costs? (yes=1, no=0) 0
Consider: a. Local service deficiencies b. Security c. Local transport problems
6. Firm leases building or building fully depreciated or residential encroachment? (yes=1, no=0) 0 7. Duplicate capacity? (yes=1, no=0) 0 8. Potential technological gap (yes=1, no=0) 0 9. Meets federal retention guideline (yes=1, no=0) 0 10. SIC code 26 11. Peripheral location (yes=1, no=0) 0
In addition, the expansion generates $55,000 in new taxes per year and $50,000 in loan repayments. The only costs are the $375,000 loan and $10,000 in administrative costs. Subtracting these from the benefits leaves a net present value of fiscal benefits of $244,000, for a 7.3 percent rate of return, as shown in the summary section of Exhibit 2. The net present value of all benefits over the seven-year period is $4.2 million, for an internal rate of return of 42.5 percent.
In summary, it is clear that the Associated Paper Forms project makes good sense. It generates considerable Chicago employment, wages, and tax receipts. Their distribution favors those who need it most. A modest interest subsidy is the only cost to the city. Given that the project was not likely to obtain financing from the private sector, it represents a good investment for Chicago.
RELATED ARTICLE: GIVING AWAY THE STORE?
Even though economic development professionals have been saying for years that business retention and entrepreneurship are far more important to a locality's development than smokestack chasing, big incentive packages to lure firms are not fading away, as some of the recent megadeals attest.
Alabama and Mercedes-Benz. Most widely questioned for its largesse, this deal will bring a 1,500-employee Mercedes-Benz plant producing sport-utility cars to Alabama for $253.3 million - a staggering $170,000 per job. The justification includes an estimated six to seven indirect jobs for every autoworker in the next five years and half again as many in 15 years, This would entail an enormous structural change in the Alabama economy. State officials argue that the Mercedes-Benz deal is important as a symbol to help bring about such a change and help Alabama shed its rural, backward image. Part of the image change will come from the 2,500 sport-utility cars the state committed itself to buying.
Kentucky Buys a Mini-mill, Even larger on a per-job basis is the $140 million package offered in 1993 by Kentucky to a Canadian partnership venture for the establishment of a 400-employee steel mini-mill. However, its $350,000 per-job figure reduces to $102,500 on a net present value basis since the largest part of the incentives will be earned as tax credits over a 25-year period. Other incentives include infrastructure improvements, training funds, and industrial revenue bonds.
South Carolina and BMW. In 1992, South Carolina landed a 2,000-employee BMW plant to be located in Spartanburg. BMW is investing $400 million to construct a 1.9-million-square-foot facility on 900 acres of land. The plant is expected to produce 300 cars per day by 1995. In turn, the state purchased the land and existing houses from local residents for $37 million. The total $150 million public incentive package also includes $10 million in road improvements, tax breaks, and extensive job training. The training in technical skills and new management techniques will be conducted in a state facility to be built for BMW's exclusive use.
Indiana and the Friendly Skies. In the midst of the recession of the early 1990s, Indianapolis was selected over 92 other contenders for a United Airlines aircraft rebuilding facility. Indiana offered $294.5 million in bonds, grants, and infrastructure improvements for the 6,300- to 7,000-employee plant - a comparatively low $42,000 per job. Furthermore, if job creation lags behind expectations, the airline will have to repay some of the incentives.
Is It Worth It? For each of these deals, evaluation is difficult unless one has access to all the details. Even deals that seem excessive at first glance may actually be good investments, at least from the local perspective. The timing of incentives and the schedule on which jobs are created affect present value calculations; bonds that are repaid actually have very little or no local cost; and infrastructure improvements may serve far more users than the deal for which they are initially built. One thing that is clearer in these large deals than in many of the smaller ones being negotiated every day is the "but for" issue. For the location ultimately chosen, it is obvious that the economic activity is really new and would not have occurred without the subsidies.
A possible irony in the calculation of benefits is that the justification for many of these deals is in the indirect jobs expected to be created, such as suppliers. It would be interesting to track how many of these firms also come to the public sector asking for incentives.
Even deals that make local economic sense may be wasteful of public resources at a national level. Concerned about the proliferating cost of incentive packages, the National Council for Urban Economic Development has established a task force on incentives.
WIM WIEWEL is a professor of urban planning and policy and Special Assistant to the Chancellor at the University of Illinois at Chicago. JOSEPH PERSKY is a professor of economics at the University of Illinois at Chicago and DANIEL FELSENSTEIN is an assistant professor in the geography department at the Hebrew University of Jerusalem and was on a research leave at Northwestern University when this project was conducted. The authors thank DANIEL McGRATH for technical work on the spreadsheet and the Chicago Department of Planning and Development for its sponsorship and advice. This project was conducted under the auspices of the University of Illinois at Chicago Center for Urban Economic Development. This article was originally published by the National Council for Urban Economic Development (CUED) in the fall of 1994 issue of Commentary and adapted for Government Finance Review. GFOA extends its thanks to CUED for permission to reprint this material.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||includes related articles; Chicago, Illinois|
|Author:||Wiewel, Wim; Persky, Joseph; Felsenstein, Daniel|
|Publication:||Government Finance Review|
|Date:||Oct 1, 1995|
|Previous Article:||Building a better budget through trust and communication.|
|Next Article:||An interview with the new chair of the GASB.|