Skip to content
Quick Start for:

April 1999

On this page:
Guided by STAR
Report Card on the States
Dynamic Modeling

On Page 1:
AOK in Y2K
Exterminating the Y2K Bug
Web Sites
From the Comptroller: A New Dynamic

Texas stats -- Fiscal and economic data

Guided by STAR

The Comptroller's office is in the final phase of implementing the State Tax Automated Research System (STAR), a comprehensive online tax policy research system. STAR contains more than 17,000 documents--including letters, memos, rules, statutes, position letters, taxability requests and hearings--covering virtually every tax administered by the agency.

The system allows convenient access to tax policy documents through the Comptroller's Window on State Government Web site. These documents previously were available only through written or verbal open records requests. STAR is used most often by accountants, tax attorneys and other tax professionals, but is available to anyone linked to the Internet.

STAR's Tax Research Index allows taxpayers to search for documents in two ways. The Topical Search Index allows a user to filter a search by tax type, rule associations or topical headings within a rule. Search results include document identification numbers (accession numbers) that can be viewed, printed or used to request a document from the Comptroller's office. Taxpayers can also locate documents using word, phrase and proximity search technology to filter materials by tax type, document type, superseded status and date. This allows users to find specific documents, even if they do not know the specific tax rule or topic used in the index.

From a total of 22,000 existing materials, about 5,000 of the oldest position letters and hearing documents are not available on-line but can be e-mailed, faxed or mailed upon request. Documents on STAR have been edited to remove any confidential taxpayer information.

Summaries of the most recent documents added to STAR are placed on the system four or five times a month. The updates contain highlights of key issues and a summary of documents added to the system, including tax type, topical heading, accession number and a brief summary.

For more information, e-mail

Contributing to this article:
Jeanie Coffey

Report Card of the States
National survey gives Texas high marks
for management practices

A national study of state governments' management practices has concluded that Texas rates an overall grade of B for its handling of the public's business, ranking it in the top third of the nation.

Governing, a national magazine devoted to state and local government issues, conducted the study with the assistance of the Maxwell School of Citizenship and Public Affairs at Syracuse University.

The government performance project, funded by the Pew Charitable Trusts, focused on state operations from 1997 to 1998. Results were published earlier this year. The study rated each state in five categories: financial management, capital management, human resources, managing for results and information technology. Texas received a B or better in four categories, and a C in capital management.

Texas had company from 11 other states that received Bs. Only five states scored higher: Michigan got a B+, and Missouri, Utah, Virginia and Washington each merited an A-.

The five broad categories used to assess the states follows, along with Texas' grades.

Financial management: Texas was cited for having "good revenue and expenditure estimates, solid cash management and contract oversight, and impressive flexibility for agencies to manage their budgets." The study credited the Texas Bond Review Board for its significant contribution to the state's good financial performance. The board provides high-level and comprehensive scrutiny over the statešs debt policy.

The state's major weak spot, according to the study, is its "minuscule" rainy day fund. "This is more than just an abstract problem to worry about when recession hits. It lowers bond ratings even in good times, costing the treasury real money every year," the magazine noted.

Researchers say that with more than four of every five states harboring measurable surpluses, a handful have "gone beyond the rule of thumb that says 5% of general fund revenues is a sufficient nest egg against hard times." Arizona and Indiana each have 7% in their contingency funds, and Maryland has 8%. At the time of the survey, Texas had only 0.1% of general revenue funds in contingencies.

In financial management, California received a C- and New York a D+ because neither "can seem to finish its budget on time."

Capital management: Texas lost points in this category because it had no statewide capital plan for acquiring and maintaining buildings and equipment. The study said Texas hadnšt "even bothered to distinguish capital budgets from operating budgets. But the state did finish its first capital plan" in 1998. Even though this was the statešs lowest grade by category, Texas still surpassed both California and New York, which each received C-.

Human resources: While the study noted that Texas is "the only state with no centralized human resources department," Texas still received a better than average grade for personnel practices. The study noted that larger Texas state agencies appear to be making solid efforts in training, and it praised the state's job classification index for being a "lean roster of fewer than 800 job titles." By comparison, New Jersey has about 8,000 titles.

While decentralization can mean state agencies have more flexibility in hiring, the study concluded that "statewide workforce planning is impossible" without a centralized source of comprehensive employee data. New York received a C in this category, California a C-.

Managing for results: Texas' highest mark came in this performance-based government category. The study applauded the statešs bipartisan strategic planning, which involves both the legislative and executive branches. It noted that while the governor creates goals in consultation with the Legislative Budget Board, there is " centralized effort to measure progress toward achieving them, but each agency does have its own strategic plan tied to the governoršs goals. Nearly all state activities use outcome measures... Targets are set and progress toward them tracked...[and] good efforts are made at breaking down the cost of services." California got a C-, and New York fared worse with a D+.

Information technology: Known for its burgeoning high-tech industry, Texas received a good grade for efforts to manage information technology. The study noted that Texas tried to develop statewide information management in financial and human resources, but most agencies use their own internal information systems for management processes.

The state got high marks for its statewide budgeting system and performance measure tracking. The study said Texas had only minimal information technology standards, but pointed out the state has a five-year strategic information technology plan, as do individual state agencies.

The study pointed out brokerage efforts by the Texas Department of Information Resources' (DIR) saved a reported $30 million on information technology purchases in 1998. DIR also received kudos for "forcing agencies to compare the qualitative and quantitative benefits promised at the beginning of (information technology) projects to the ultimate success in delivering them." Texas was one of only 14 states receiving a B or better in this area. Both California and New York earned Cs.

Bottom line: The results of Governing's first management study provides elected officials, policymakers and the general public an objective way to gauge their statešs management performance. The study also gives an overview on how other states are tackling common problems, as well as an assessment of each statešs progress in seeking solutions to those problems.

In addition, Texas' overall grade of B--placing it in the top one--third of the nation‹enhances the statešs reputation as a leader in public policy and financial management.

Dynamic Modeling
New method of tax analysis accounts for taxpayer behavior

Evidence shows that tax cuts help pay for themselves, while increasing a tax rate actually results in collecting proportionately less tax revenue. That is why there is a growing branch of tax analysis called dynamic modeling--a tool the Comptrolleršs office will use to provide more accurate analysis of tax cuts.

Dynamic tax modeling uses a computerized series of interconnected equations to trace how a proposed tax change ultimately affects the statešs economy. For example, if a state collects $500 million in tax revenue from a 50-cents per pack tax on cigarettes, a static analysis‹the opposite of dynamic--would say that doubling the cigarette tax to $1 a pack would generate another $500 million in state revenue.

But this type of simple, static analysis ignores the behavioral effects a tax rate change may have on those currently paying the tax and the secondary effects that such a change could have beyond those directly affected by the tax.

Increasing the tax on cigarettes could cause some smokers to smoke less or to quit the habit altogether. Tobacco taxes previously paid by these people would then be lost to the state, offsetting some of the additional revenue from increasing the tax rate. Those who quit would have more money to spend on other goods and services‹some, none or all of which may be produced--and taxed--in Texas.

Those choosing to continue smoking and pay the higher tax would reduce their spending on other items by the amount of the additional tax paid. As this spending by individuals falls, however, government tax collections grow by the same amount, which may be spent on the goods and services the government buys.

Dynamic modeling estimates the strength of these relationships and includes them in equations used for an overall estimating system of economic effects and governmental revenues. This method of examining the economic effects of changing tax policies and the associated impact on revenue forecasting is practiced by only a handful of states in varying capacities. In California, dynamic revenue forecasting is the law, while in Minnesota, a model was developed for a one-time evaluation of a proposed tax policy change.

Dynamic modeling is like using a large-scale map to capture the long-term effects of a change in tax policy; and it can serve as an analytical tool for state revenue forecasts.

Big picture: In theory, dynamic modeling should prove an invaluable tool in both analyzing potential tax changes and estimating state revenues. In practice, however, the short-term accuracy of these models can vary greatly, and no state consistently uses this approach to estimate tax revenues, although several states use the approach to compare alternative tax proposals.

In one of the first evaluations of dynamic modeling, the U.S. Congressional Joint Committee on Taxation in 1997 asked nine economic consultants to examine the same proposed tax changes and to present the results of how their dynamic models portrayed the effects on federal revenues. Over the long term (after five to 10 years), there was considerable agreement among the groups participating about the effects that alternative tax policies would have on the U.S. economy and tax revenues.

But in the short-term (about three years), the estimated impacts were wildly different, leading the Joint Committee to conclude that dynamic modeling had little utility for near-term economic and revenue estimating.

Economists and statisticians usually can estimate what the effects of a change ultimately will be, but the precision of their estimates falls considerably in defining how quickly a change will be adopted.

In the cigarette tax example, while some smokers could choose to quit after a tax increase, exactly how quickly this would occur is less certain. The timing of changes in business investment following a shift in tax policy is perhaps the most difficult to determine. Will businesses wait to actually accrue the money from a decreased business tax before beginning a capital expansion program, or will they anticipate an increase in their profits and increase activity immediately after the tax cut? Will they put the money in Texas, or in a new project a thousand miles away?

From the standpoint of evaluating the long-term impact of a change in tax policy--after it has had a chance to work its way through the economy--the short-term behavior of taxpayers is of little relevance. The dynamic effects of alternative tax policies after five or 10 years can be readily compared with some confidence.

But to estimate next year's tax revenues, the short-term behavior of taxpayers in response to a tax change can often be the defining variable, and it is in the short-term that dynamic models are the weakest. For this reason, California's dynamic estimating program--perhaps the nation's most advanced--only produces estimates of tax effects for a point five years into the future. And it was the short-term unreliability that contributed to Massachusetts' mothballing dynamic modeling of tax changes indefinitely.

Emerging initiatives: In California, the Department of Finance must begin a dynamic revenue estimate for all proposed tax changes that carry an impact of at least $10 million, as mandated by a 1994 law. While some 75 to 80 bills in 1997 and again in 1998 met the $10 million-threshold, dynamic estimates were appropriate for only a handful because of the short-term nature of most bills.

In 1993, Minnesota developed a set of models for a one-time evaluation of the impact of exempting capital equipment from the state sales tax.

In 1997, Michigan tax officials determined that the dynamic method was not ready for formal use in revenue estimation because of weaknesses in the statešs existing models. However, the Michigan legislature appropriated $500,000 to develop tax simulation models for legislative staff.

Arkansas, Connecticut, Delaware and Wisconsin also have used some form of dynamic estimating. In Florida, dynamic tax models are used, but the impact is not counted against budget figures. Kentucky uses dynamic modeling to determine the long-term tax impact of certain economic development initiatives but it is not used for revenue estimates.

The Pennsylvania Tax Blueprint Project was launched in 1996 to analyze tax policy through dynamic economic forecasting models. The effort is privately funded and managed by representatives of both the public and private sector.

Two sources of federal revenue estimates--the Office of Tax Analysis of the U.S. Treasury and Congress' Joint Committee on Taxation--use static revenue forecasting. Several groups, including the Heritage Foundation, the National Center for Policy Analysis and the Institute for Policy Innovation, disagree with the federal government's reliance on static analysis and propose changing to a dynamic system of forecasting, since differences between nationwide static and dynamic forecasts can be substantial.

But at the state level such large impacts are less likely. Interstate moves to avoid changes in state taxes are much easier and less expensive than international moves to avoid changes in national taxes.

And even the federal tax system itself can greatly dampen a dynamic response to state tax law changes. If a current state tax is deductible in calculating federal corporate or individual income taxes, reducing this state tax would initially cost the state 100% of the amount of the reduction, but because federal deductions would be lowered, individuals and businesses would realize less than 100% of the state tax reduction since some would "leak off" in additional federal tax payments. In such cases, state and local tax reductions have the perverse dynamic effect of generating a windfall for the federal government. In addition to expecting smaller dynamic effects of state tax policies than might be seen in federal tax changes, state revenue estimators trying to calibrate dynamic tax models admit regional economic data are scarce and growing scarcer.

New approach: At the direction of Comptroller Carole Keeton Rylander, the Revenue Estimating Division is making its first forays into the dynamic modeling of selected tax changes. Analysts will produce estimates of economic effects of tax changes that have greater than a $100 million initial revenue change. They intend to use the results of dynamic models to examine the effects of tax law changes on employment, investment and income.

In part, this reluctance to use the results of dynamic models to change revenue forecasts stems from the notorious short-term inaccuracies of such models. In addition, the Comptrolleršs short-term revenue impacts of tax law changes have always included many of the "first round" behavioral effects used by dynamic models.

Unlike truly static revenue forecasts, the methodology used by the Comptroller's office has traditionally adjusted projected revenue for likely direct behavioral changes, such as the decrease in the number of smokers in the example of increasing the cigarette tax. Although the analyses traditionally have not included second-round effects such as the ultimate changes in consumer spending or saving that would be included in a dynamic model, in most cases the direct behavioral effects both dominate the long-run effects in size and occur first. So will tax breaks ultimately be seen as paying for themselves in a dynamic modeling system, even over the long term? Other states' experiences cast doubt on this.

The largest "feedback effects" predicted in the California dynamic model are about 19%, based on a change in the rate of taxing profits. This means that while lowering the tax on bank and corporate profits in California would lead to increased economic activity, the net effect of this increased activity would only generate 19% in tax revenues above what would otherwise be lost. And even this dynamic effect will only be seen after five years.

Other tax changes have considerably less feedback effects in the California model. Lowering personal income tax rates would only generate a 3% feedback effect; lowering sales taxes, only 11%.

Instead, the real strength of dynamic modeling of tax law changes may lie in comparing the long-term differences in alternative tax scenarios. The Comptroller's office hopes to be among the leaders in the nation using this new methodology to more accurately evaluate the long-term effects of tax policy changes. Importantly, this analysis for the first time will include long-term job and investment-generating feedback effects, which could make the state's tax climate better for both businesses and individuals.

Contributing to this article:
Don Hoyte and Julie Crimmins

Go to page 1