ROUGH DRAFT authorea.com/4806

# Introduction

The economic potential of any system is not only driven by the presence of natural and/or developed capital, but also the institutions that govern the exploitatioj of these valuable assets. In this sense, the term economic potential is misleadingly incomplete. The mechanism that drives resource allocation is a function of both the distribution of purely economic value and the feasible range of activities foverned by political institutions. The goal in institutional design for economic growth, therefore, ought to be the facilitation of those activities that increase the marginal product of value extraction efforts. If the objective is the maximization of economic growth, this broad goal is unlikely to draw many detractors. The devil, however, is in the details.

Among the multitude of policy innovations that have been advanced in service of increasing economic growth, institutional reforms that act on the property tax base materially altered local government finance for more than a century. As early as the 1880s, these reforms were dominated by efforts to target specific populations via circuit breakers and homestead exemptions. (Bowman, 2008) However, starting in large part with the Tax Revolt of the 1970s, more interest has taken root in implementing reforms that target the base in a general way: tax and expenditure limitations. The impact of these and other measures has been noticeable. While property tax revenue remains fairly buoyant with respect to the economy, it has declined in importance, dropping as a percentage of general revenue from 34% to 27% over the 1977-2002 period. (Edwards, 2006)

This paper is one component of a larger study seeking to understand the unintended consequences of tax and expenditure limitations. The broad study is a three part empirical examination of the differential impact of tax and expenditure limitations in Colorado (henceforth COTELs) on counties of with different economic foundations. Each section is characterized by exploration of three thematic hypotheses:

1. COTELs create wedges between desired and realized expenditure behavior;
2. COTELs decrease variation among similarly constrained counties and increase variation among disimilarly constrained counties; and,
3. COTELs decrease resident and employment growth in constrained counties.

The unifying principle across each of these inquiries is the idea that COTELs have constraint levels that vary both cross-sectionally and temporally. This paper explots this variation to explore the extent to which fiscal clustering (measured as fiscal capacity and revenue generation in this context) is driven by this "COTEL intensity" concept.

# COTELs & Fiscal Clustering

From a theoretical standpoint, there are two main features of COTELs that would promote fiscal clustering among low-capacity jurisdictions. First, COTELs impose asymmetric pressure on revenue yields over time. The rate of increase is limited by an explicit growth limit, but decreasing yields are not so constrained. Furthermore, once a drop in revenue has been experienced in one year, the baseline is "ratched" down because current year allowable revenue is measured against only the previous year's revenue (and not the long term trend). Low revenue years have a disproportionate impact on the long-term revenue capacity of the jurisdiction.

The second impact is relate to the first. To the extent that COTELs bias revenue downward (in a manner quite uncoupled with demand projections), it becomes increasingly difficult for constrained jurisdictions to invest in the human and physical capital needed to support robust economic growth. The first effect impacts the revenue generation, while the second impacts the capacity of the economic base itself.

A complicating factor is the existence of spillovers. Both the provision of public goods and the economic vitality of neighboring jurisdictions impact the fiscal and economic capacity of the primary jurisdiction. This externality network has the potential to dynamically reinforce economic growth behavior, whether it be a high or low growth regime. If COTELs exacerbate this effect, there could be undesirable long-term consequences for low capacity counties seeking to create the conditions for economic growth. Insofar as the relationship between COTEL intensity and spatial dependency is evaluated, this paper explores this phenomenon directly.

# TEL Structure in Colorado

This analysis focuses on Colorado, a state characterized by one of the most restrictive TEL regimes in the country. Two properties of the TEL structure make Colorado an attractive empirical source. First, while much of the literature regarding TELs in Colorado focuses on the Taxpayer Bill of Rights (TABOR), in reality TABOR is only one piece of the puzzle. In addition to TABOR, Colorado features a general statewide limit on property tax revenue, the Gallagher Amendment, Amendment 23, and a general spending limit.

• General Statewide Limit on Property Tax Revenue (SLPTR): Local property tax revenues may not increase by more than 5.5% in a given year (the limit was 7% before 1988)
• Gallagher Amendment (GA): Assessment rates for residential are set statewide, calculated to keep historical ratios of residential and non-residential property tax revenue consistent
• TABOR: Limits annual growth in revenues to inflation plus a measure of growth (new construction for local governments and enrollment for school districts; the statewide revenue limit was relaxed via the 2005 Referendum C)
• Amendment 23 (A23):
• General Spending Limit (GSL): Limits annual growth in expenditures to inflation plus a measure of growth

Furthermore, explicit voter approval is required for both mill levy increases and debt obligations. Our focus here is on the explicit revenue limitations. These overlapping reforms are a double-edged sword. On the one hand, they complicate the model insofar as none of them are observed independently in the period of time studied. On the other hand, they provide a means to incorporate variation in policy application and impact across space and time.

Colorado is also a desirable state because of "De-Brucing". Since the inception of TABOR, a number of counties have passed local legislation designed to avoid the constraints established by this and prior legislation. This practice is known as De-Brucing, and to the extent that it creates differential policy application across Colorado's 64 counties, it is a major source of horizontal variation.

# Measuring the Impact of COTEL Intensity

A major characteristic of this analysis is the effort to capture variation in COTEL impact. An ideal intensity measure would capture three primary characteristics of any given county at time $t$:

1. The cumulative effect of overlapping policies;
2. The proliferation of locally derived exemptions from TABOR and the SLPTR; and,
3. The local dynamics that may trigger a breach in the ceilings imposed by the aforementioned legislation.

The indicator used in a previous iteration of this analysis was simply an ordinal score that captured only the first two characteristics. Effectively, the score added a point for each additional constraint. For example, all counties in 1987 would have possessed a score of two given the general application of the SPLTR and the Gallagher Amendment. The scores were subsequently modified by De-Brucing, which occured in 47 of 64 counties over the 1993-2009 period. If only one of the policies was exempted, only one point would be removed. In effect, it was a cumulative indicator function. The reason it has value in this paper is because it provided a clear view of variety policy application. Observe the score distribution by the year 2009.