Property Tax, Local Sales Tax and Business Activity in Nevada: A Spatial Analysis

Abstract

This study examines how business activity responds to local taxation, specifically property tax and local sales tax, in Nevada. Using county-level data for the period 1999–2014, we assess the impact of these taxes on various business activity indicators, including employment, annual payroll, the number of establishments, and the number of small establishments categorized by size. Unlike previous studies that primarily focus on state-level taxation, our research delves into the effects of local tax instruments. By analyzing different components of the property tax (e.g., school district, county, and special district rates) and evaluating the specific effects of local sales tax changes, we provide a nuanced understanding of the local tax–business activity relationship. To address potential policy endogeneity in the sales tax rate, we instrument the sales tax rate using the lagged share of registered Democrats and implement an IV (control-function) spatial Durbin framework, ensuring robust estimates of within-period associations and spatial spillovers. Our analysis is intentionally confined to the 1999–2014 institutional regime, when Nevada businesses were primarily exposed to property and sales taxes. The estimates should, therefore, be interpreted as evidence on how the local tax mix and its components correlate with business activity under this pre-2015 fiscal structure, rather than as a direct forecast for the post-2015 environment shaped by subsequent policy changes and macroeconomic shocks. Across specifications, the IV-identified total effect of the sales tax rate is consistently negative for establishment-related outcomes. Nonetheless, the results remain informative for current debates on the design of local revenue systems because the underlying tax–service bundle and cross-jurisdictional spillover mechanisms continue to be central to local public finance.

1. Introduction

Most local governments in the United States levy taxes on businesses primarily through property taxes. While sales tax has historically been a key component of state tax systems, its use by local governments—particularly at the county and sub-county levels—has grown significantly in recent decades. Together, these two taxes constitute the principal sources of revenue that support essential local government services, including public education, public safety, and infrastructure. In Nevada, where local governments rely almost entirely on property and sales taxes due to the absence of personal or corporate income taxes, understanding how these taxes influence local business activity is particularly important.

Recent empirical work underscores both the importance of local tax instruments and the relevance of spatial interactions. Studies of local sales taxes document spatial dynamics and tax competition in local rates and their implications for economic outcomes, including employment (Rohlin & Thompson, 2018; Shon, 2017). On the property tax side, recent quasi-experimental evidence examines how property tax changes linked to local referenda relate to business outcomes, highlighting the potential for heterogeneous and context-dependent effects (Enami et al., 2023). More broadly, the spatial econometrics literature on local fiscal policy and spillovers motivates estimating both direct and cross-jurisdiction effects rather than focusing solely on own-county responses (Deller & Maher, 2024; Fajgelbaum & Gaubert, 2025).

Prior research has extensively examined the relationship between taxation and business activity, with a primary focus on corporate income taxes at the state or national level. However, businesses also face substantial tax burdens at the local level, particularly through the property tax and the sales tax, both of which affect firm decision-making. These taxes impose direct costs on businesses, influencing location choices, investment decisions, and employment levels. Moreover, they exert indirect effects by altering the economic environment for entrepreneurs, employees, and consumers. Given that Nevada’s local governments derive most of their revenue from these two sources, a systematic examination of their impact on business activity is warranted.

In this paper, we investigate how local taxation—through both property and sales taxes—affects business activity in Nevada. Property and sales taxes can influence firm location and expansion decisions through distinct channels. A higher property tax raises the user cost of commercial space and, depending on statutory and assessment rules, may be capitalized into rents and property values, potentially discouraging investment and local employment. At the same time, property taxes finance local public inputs—such as infrastructure, public safety, and other services—that can raise productivity and partially offset the tax burden, implying an a priori ambiguous net relationship (“cost” versus “benefit” view). In contrast, local sales taxes primarily operate through the demand side and market access: higher sales taxes can reduce local retail demand and induce cross-border shopping, which may weaken sales and employment in consumer-facing sectors while also potentially affecting firms’ location incentives near lower-tax jurisdictions. These channels also imply spatial spillovers: tax changes in one county can redirect consumers and mobile economic activity across nearby county borders, and local fiscal expansions financed by taxes can generate benefits (or costs) that are not confined to a single jurisdiction. The SDM framework is, therefore, well suited to quantify both within-county associations and inter-county spillovers arising from these mechanisms.

Guided by the channels discussed above, we evaluate three testable predictions:

Hypothesis 1 (Property tax bundle ambiguity).

Because property taxes operate through both a cost channel and a public-input (tax–service bundle) channel, the net association between the total property tax rate and business activity is a priori ambiguous and may differ across components (county, school, and special districts).

Hypothesis 2 (Sales tax demand and border channel).

Higher combined state and local sales tax rates (hereafter, the sales tax rate) are expected to be negatively associated with business activity through reduced local demand and cross-border substitution.

Hypothesis 3 (Spatial spillovers).

If consumers and firms are mobile within regional markets, tax changes in one county can generate non-zero spillover effects on neighboring counties’ business activity.

Our objective is not only to assess the economic consequences of these taxes but also to analyze the trade-offs faced by policymakers when structuring local tax systems. Using county-level data spanning the period 1999–2014, we examine how variations in tax rates influence multiple measures of business activity, including employment, annual payroll, the number of establishments, and the number of small businesses with 1–9 and 10–49 employees. In addition to considering the total county-level property tax rate, we disaggregate this tax into its key components—the school district tax, the county tax, and the special districts tax—to explore their differential effects on business activity. Similarly, for the local sales tax, we specifically analyze how changes in the sales tax rate affect businesses. These local tax instruments also have direct relevance for real-estate finance and municipal risk. Property and local sales taxes enter firms’ operating costs and can affect commercial location decisions, tenant demand, and ultimately cash flows capitalized into commercial property values. Spatial spillovers imply that tax-policy shocks and business responses may be correlated across neighboring jurisdictions, which matters for municipal revenue risk and the pricing of local public finance.

We study the 1999–2014 period to isolate the relationship between business activity and the two dominant local tax instruments in Nevada—property and sales taxes—under a relatively stable institutional setting. After 2014, Nevada introduced the Commerce Tax (2015) and the broader fiscal environment was further reshaped by major federal tax reform (TCJA) and, later, the COVID-19 shock, which could confound the tax–business activity relationship. Accordingly, our findings are best interpreted as within-regime associations that illuminate mechanisms—tax composition, earmarked spending channels, and spatial spillovers—rather than as a direct mapping to the post-2015 policy environment.

Nevada provides a compelling case for this analysis due to its distinctive tax structure. The state does not impose corporate or personal income taxes, making property and sales taxes the primary fiscal instruments for local governments. Moreover, a significant policy change occurred in 2005 when Nevada introduced a tax abatement program that placed limits on property tax revenue growth, thereby altering the local tax landscape. Unlike some other states, Nevada’s counties apply uniform tax rates to all businesses, simplifying data collection and allowing for a more precise examination of tax effects.

While our study builds upon a well-established body of literature that examines how businesses respond to tax differentials and rate changes, it offers several key contributions. First, unlike most prior studies, which focus on state-level taxation, we analyze business responses to local taxation. Second, by considering both property and sales taxes, we assess the effects of two of the most significant local tax instruments rather than examining a single tax in isolation. Third, we address important econometric concerns related to endogeneity and spatial dependence. Given that tax rate differentials and changes over time can generate economic and fiscal spillover effects across neighboring areas, failing to account for spatial dependence could lead to biased estimates. To address this issue, we employ spatial econometric techniques, specifically the Spatial Durbin Model (SDM), which allows us to capture spatial interdependencies in local tax effects. Additionally, we use voter registration data as an instrumental variable to address potential endogeneity in sales tax rates.

Relative to prior spatial analyses of tax competition and business outcomes, our contribution is twofold: First, much of the spatial public finance literature focuses on strategic interaction in tax setting (i.e., whether one jurisdiction’s tax rate responds to neighboring jurisdictions), whereas we focus on the economic outcome side—how the local tax mix is associated with multiple dimensions of business activity—while still allowing for spatial feedback through the SDM structure. Second, studies of local sales taxes often emphasize border effects and competitive setting of local option rates, but typically do not jointly model property taxes and their components within a unified spatial framework. By decomposing the property tax into county, school, and special-district components and estimating both direct and spillover effects, we provide evidence on which parts of the local tax bundle are most strongly associated with business activity and whether these associations transmit across county borders. This within-state setting also mitigates cross-state institutional heterogeneity, enabling a cleaner comparison of local tax instruments under a common legal and administrative environment.

Our empirical findings reveal a sophisticated relationship between taxation and business activity. We find that the total property tax rate is, overall, positively associated with changes in business activity. However, when disaggregating the property tax, we observe heterogeneous effects: while county tax rates and special district tax rates exhibit a positive relationship with business activity, school district tax rates are negatively associated with changes in business activity. In contrast, we find a negative relationship between the combined state and local sales tax rate and business activity. Furthermore, our results indicate that following the introduction of the 2005 property tax abatement, the association between county property tax rates and business activity became significantly positive, whereas the negative relationship between sales tax rates and business activity became more pronounced.

The remainder of this paper is organized as follows: Section 2 provides a review of the relevant literature on state and local taxation and business activity. Section 3 describes our data sources and empirical methodology. Section 4 presents our empirical findings. Section 5 reports robustness checks using alternative spatial weight matrices and a Clark County exclusion exercise. Section 6 discusses policy implications and Section 7 concludes with a discussion of the implications of our results for local tax policy.

2. Previous Studies

The literature on taxation and economic activity is extensive, yet it exhibits two significant limitations. First, much of the research has focused on the residential effects of property taxation or the impact of sales taxes on consumers, with far less attention given to business activity. While residential property taxes influence housing markets and consumer purchasing power, their implications for firm location, employment, and investment decisions remain underexplored. Second, most studies analyze either the sales tax or the property tax in isolation rather than considering both simultaneously, and they often focus on state-level taxation rather than local taxation. Given that local governments rely heavily on both property and sales taxes as primary revenue sources, understanding their combined effects on business activity is crucial for informed policymaking.

2.1. Property Taxes and Business Activity

Numerous studies have investigated the effects of property taxes on state and local economies, but they overwhelmingly emphasize residential rather than business effects. A significant body of research has explored the incidence of property taxes, with foundational contributions from (Hamilton, 1975, 1976; Mieszkowski, 1970; Mieszkowski & Zodrow, 1989; Netzer, 1966; Simon, 1943; Wildasin, 1989; Zodrow & Mieszkowski, 1986). The “capital tax” view proposed by (Mieszkowski, 1970) and extended by (Zodrow & Mieszkowski, 1986) and (Mieszkowski & Zodrow, 1989), suggests that property taxes function as a partial factor tax on capital, with the burden falling on all capital within a given jurisdiction.1 This perspective has been widely adopted in empirical research, providing a theoretical foundation for analyzing the business implications of property taxation.2

The Lincoln Institute of Land Policy has contributed significantly to the discussion of property taxation, particularly regarding its implications for economic development. A report by (Kenyon et al., 2012) critically examines property tax incentives for businesses in the United States. While they argue that property tax incentives are often insufficient to attract businesses on their own, they acknowledge that, in some cases, these incentives have contributed to business attraction and, consequently, local economic development. Their findings suggest that property tax policy can have subtle, context-dependent effects on business activity.

More recent studies have attempted to quantify the effects of property tax changes on firms. Wassmer (1993) and Muthitacharoen and Zodrow (2012) explore how property tax burdens influence firm location and investment decisions. Their findings indicate that high property taxes can discourage business investment, but these effects may be offset by the benefits of public services funded by such taxes. Recent quasi-experimental evidence further highlights the potential for heterogeneous and context-dependent effects of property taxation on business outcomes (Enami et al., 2023). However, empirical results remain mixed, as the impact of property taxation appears to depend on broader economic conditions, industry-specific factors, and the availability of alternative tax incentives. The mixed empirical evidence is not surprising given several sources of heterogeneity. First, property taxes may be capitalized into commercial rents and land values, so incidence can differ by market tightness and the share of immobile versus mobile capital. Second, property taxes are frequently bundled with local public inputs (infrastructure, public safety, and permitting capacity), so reduced-form estimates may capture both the cost channel and a productivity-enhancing service channel. Third, effects can vary across industries depending on customer mobility, factor intensity, and the relative importance of place-based amenities. These considerations motivate designs that measure business outcomes directly and allow for cross-jurisdictional reallocation and spillovers.

2.2. Local Sales Taxes, Border Effects, and Business Activity

While the effects of sales taxes on consumers are well-documented, research on their impact on businesses is less extensive. The general consensus in the literature is that higher sales taxes can negatively affect business activity, primarily by reducing consumer demand and influencing firm location decisions. Duranton et al. (2011) find that business taxes significantly affect local employment growth, while (Fox, 1986; Hoyt & Harden, 2005; Thompson & Rohlin, 2012) suggest that state sales taxes have adverse effects on firms’ employment decisions. Recent empirical work also emphasizes that local sales taxes can generate spatial dynamics and tax competition in local rates with implications for employment and other economic outcomes (Rohlin & Thompson, 2018; Shon, 2017).

A related body of research examines the effects of corporate income taxes on business activity. (Feld & Kirchgässner, 2003) provide evidence that corporate taxation negatively affects firm location and employment decisions. Although corporate income tax differs from sales and property taxes, these studies underscore the broader theme that tax burdens can shape business behavior in meaningful ways.

One important aspect of sales taxation is the “border tax effect,” which arises when businesses and consumers adjust their behavior in response to tax differentials between neighboring jurisdictions. (Snodgrass & Otto, 1990) examine rural and border communities in Oklahoma, finding that sales tax increases disproportionately harm rural areas. Similarly, Walsh and Jones (1988) and Tosun and Skidmore (2007) analyze tax-induced shifts in consumer spending in West Virginia, highlighting the significant border effects created by differences in food taxation. Rohlin et al. (2014) extend this analysis to business location decisions, showing that firms strategically locate near lower-tax jurisdictions to minimize their tax burden. These findings underscore the importance of considering tax spillover effects across local boundaries—a factor we explicitly address in our study through spatial econometric methods.

2.3. Multi-Tax Bundles and Local Fiscal Decisions

An important limitation of much prior work is the focus on a single tax instrument, even though local governments typically select a mix of revenue sources and expenditures under institutional and political constraints. In practice, variation in one tax rate may be correlated with changes in other taxes, fees, or spending priorities, complicating the interpretation of single-tax regressions. From the firm perspective, location and expansion decisions respond to the overall local fiscal package, including both the tax burden and the quality of locally financed public inputs. Recent evidence on local fiscal choices and growth reinforces this perspective, showing that local fiscal decisions and economic performance can be jointly shaped by local conditions (Deller & Maher, 2024). This multi-instrument view motivates studying property and local sales taxes jointly rather than in isolation.

2.4. Spatial Interactions and Fiscal Spillovers

Local fiscal policies can generate spatial dependence through multiple mechanisms: mobile consumers may reallocate spending across nearby jurisdictions, firms may adjust establishment locations within integrated regional markets, and local governments may respond strategically to neighbors’ tax and spending choices. These channels imply that the effect of a tax change is not confined to the originating jurisdiction, motivating empirical approaches that estimate both direct and cross-jurisdiction effects rather than focusing solely on own-county responses. Spatial econometric frameworks such as the Spatial Durbin Model provide a coherent way to quantify these own- and neighboring-county impacts and to interpret results in terms of direct, indirect (spillover), and total effects (Elhorst, 2010; LeSage & Pace, 2009).

2.5. Synthesis and Motivation for the Current Study

Despite the extensive research on taxation and economic activity, the literature exhibits several gaps that our study seeks to address. First, most prior studies have focused on either sales taxes or property taxes in isolation, even though local governments typically set a broader fiscal package. Second, the empirical literature reports mixed findings, in part because taxes can operate through both cost and public-service channels and because business responses may be heterogeneous across industries and local contexts. Third, prior studies frequently overlook spatial dependence—how tax changes in one jurisdiction may have spillover effects on neighboring areas—which can bias inference when jurisdictions are treated as independent. Our study is designed to address these concerns by jointly modeling property and local sales taxes, decomposing property taxes into key components, and employing the Spatial Durbin Model to capture inter-county spillovers.

Finally, we contribute methodologically by using voter registration data as an instrumental variable to address potential endogeneity in sales tax rates, thereby enhancing the robustness of our findings. By addressing these limitations, our study provides new insights into the relationship between local taxation and business activity. In doing so, we aim to inform policymakers about the trade-offs involved in structuring local tax systems and contribute to the broader discussion on optimal taxation at the local level within the institutional setting studied.

3. Tax Rates, Data and Empirical Approach

3.1. Data

Data on the combined state and local sales tax rates, as well as property tax rates for each county, were obtained from the annual reports of the Nevada Department of Taxation, Division of Local Government Services. Our dataset covers the period from 1999 to 2014, providing a comprehensive view of tax rate variations across counties. We end the sample in 2014 to maintain institutional comparability and measurement consistency. In particular, Nevada’s post-2014 tax system changed materially with the introduction of the Commerce Tax in 2015, which altered the effective tax bundle faced by many firms. In addition, later macro events and potential changes in data reporting can complicate a clean interpretation of local tax effects. For these reasons, we treat 1999–2014 as a coherent institutional regime and interpret the results as within-regime relationships while explicitly limiting claims about post-2015 outcomes.

Business activity data were sourced from the County Business Patterns (CBP) of the U.S. Census Bureau, which includes detailed information on the number of establishments by industry and firm size, employment levels, and annual payroll. Additionally, county-level personal income data were collected from the Bureau of Economic Analysis (BEA), while demographic variables—including the share of the population aged 20–64, 65 and over, and the nonwhite population—were obtained from the U.S. Census Bureau.

Table 1. Summary Statistics.

Variable	Mean	Std. Dev.	Min	Max
Combined state and local sales tax rate	7.166176	0.324478	6.85	7.85
Total county property tax rate	2.875937	0.487646	1.6973	3.66
Total county property tax rate components:
County property tax rate	1.340889	0.511047	0.579	2.469
School tax rate	1.094453	0.234126	0.75	1.5
County combined special district	0.274582	0.238466	0	0.826
Employment	60,241.94	178,697.8	0	865,940
Annual payroll	2,163,629	6,455,486	0	3.26 × 107
Per capita income	32,459.5	8091.901	16,280	56,469
Number of establishments (total)	3368.392	9078.762	10	42,031
Number of establishments (1–9 employees)	2436.016	6472.133	8	30,051
Number of establishments (10–49 employees)	743.902	2054.76	1	9528
Share of population (age 20–64)	0.58793	0.03306	0.506915	0.668267
Share of population (age 65+)	0.151582	0.049366	0.059978	0.278487
Share of nonwhite population	0.112072	0.064222	0.04058	0.442975
#. Obs.	255

Notes. Tax rates are expressed in percentage points (pp). Employment is the number of jobs; payroll is in [USD, thousands/millions] (real [base year] dollars); establishments are counts. All models use county- and year-fixed effects and the row-standardized [contiguity/distance] spatial weight matrix as indicated. Sample period: 1999–2014. Data sources: business outcomes are from the U.S. Census County Business Patterns (CBP); income and demographic controls are from the U.S. Bureau of Economic Analysis (BEA); property and sales tax rates are from the Nevada Department of Taxation annual reports; sample covers Nevada counties, 2000–2014.

provides summary statistics for the variables used in our regression analysis. The substantial cross-county variation in tax rates and socioeconomic characteristics provides useful identifying variation for our analysis. The longitudinal nature of the data further enables us to assess dynamic effects and control for unobserved confounders.

Our key fiscal variables are the combined state and local sales tax rate and the countywide property tax rate. The combined sales tax rate is measured as a percentage rate applicable in each county-year and reflects the statewide base rate plus county-level local option components reported by the Nevada Department of Taxation. Property tax rates are reported in dollars per $100 of assessed valuation and are constructed as the sum of the county rate, school district rate, and the combined special district rate in each county-year, following the Nevada Department of Taxation’s annual county tax rate reports. To assess heterogeneous fiscal channels, we analyze both the total property tax rate and its three components separately.

Table 2. Tax Rates in Nevada and the Other Western States.

State	Sales Tax Rate (Highest Combined State and Local Rate, 2014)	Average Effective Property Tax Rate, 2013
Alaska	0% (7.5%)	1.18%
Arizona	5.6% (10.73%)	0.84%
California	7.5% (10%)	0.81%
Colorado	2.9% (10%)	0.60%
Hawaii	4.0% (4.5%)	0.27%
Idaho	6% (9%)	0.76%
Montana	0%	0.85%
Nevada	6.85% (7.85%)	0.85%
New Mexico	5.1% (8.69%)	0.74%
Oregon	0%	1.08%
Utah	6.0% (8.05%)	0.68%
Washington	6.5% (9.6%)	1.08%
Wyoming	4.0% (6%)	0.61%

Source: Federation of Tax Administrators, US Census Bureau 2015 American Community Survey.

compares Nevada’s tax rates with those of other Western states. Notably, Nevada does not impose an individual or corporate income tax, making sales and property taxes the primary revenue sources for local governments. Nevada’s sales tax rate is comparable to those of other Western states, while its effective property tax rate—calculated as median annual property taxes as a percentage of median home value—is among the highest in the region, trailing only Alaska, Montana, Oregon, and Washington.

Figure 1, Figure 2, Figure 3, Figure 4 and Figure 5 illustrate the distribution of sales and property tax rates across Nevada’s counties. A key observation is the significant variation in tax rates across counties. Nevada maintains a statewide sales tax rate of 4.6%, while county-level rates ranged between 2.25% and 3.25% in 2014. The lowest combined sales tax rate (6.85%) applied to Esmeralda, Eureka, Humboldt, and Mineral Counties, while the highest (7.85%) was observed in Clark County. As Figure 1 shows, the highest sales tax rates are concentrated in Clark and Washoe Counties, the two largest counties in the state.

Property tax rates exhibit even greater variation, ranging from $1.7743 per $100 assessed valuation in Eureka County to the statutory maximum of $3.66 per $100 in Mineral and White Pine Counties. Table 2 reports an average property tax rate of $2.88 per $100 assessed valuation. As depicted in Figure 2, Clark and Washoe Counties tend to have lower total property tax rates, while some northern counties also exhibit relatively low rates.

The components of the property tax rate—county rate, school rate, and special district rate—show significant variation across counties (Figure 3, Figure 4 and Figure 5). While the county and school tax rates are generally the highest components, the special district tax rate is lower overall but exhibits greater relative variation. Unlike sales tax rates, no clear spatial pattern emerges for property tax components, highlighting local fiscal policy discretion.

Figure 6, Figure 7, Figure 8, Figure 9 and Figure 10 illustrate the temporal evolution of sales and property tax rates. Figure 6 shows that sales tax rates increased in 11 out of Nevada’s 17 counties between 1999 and 2014, while Figure 7 indicates that property tax rates rose in 12 counties. The school tax rate (Figure 8) saw several reductions over time, whereas the county tax rate (Figure 9) increased in most counties. Notably, Figure 10 suggests that while many counties did not levy a special district tax, others—such as Clark County and Douglas County—relied heavily on it.

The decision to limit our dataset to 2014 is primarily driven by data availability and institutional comparability. Several factors contribute to this choice.

(1)

Tax policy stability and changes post-2014: Nevada’s tax system changed materially after 2014, most notably with the introduction of the Commerce Tax in 2015, a gross receipts tax applying to businesses with annual revenues exceeding $4 million. This reform altered the effective tax bundle faced by firms, making it more difficult to isolate the effects of property and sales taxes on business activity in later years. By ending the sample in 2014, we focus on a period in which businesses were primarily exposed to property and sales taxes, thereby improving within-regime comparability.

(2)

Data availability and consistency: County-level business and public finance datasets are often released with delays and may undergo reporting or methodological changes over time. When this research began, post-2014 county-level series were not consistently available across sources, and extending the sample without accounting for potential measurement changes could introduce additional noise or bias.

(3)

Economic disruptions post-2014: The period after 2014 includes major events that could confound the relationship between local taxation and business activity, including federal tax reforms (e.g., the Tax Cuts and Jobs Act of 2017) and, later, the COVID-19 pandemic. These shocks introduced large economic fluctuations that may overshadow local fiscal effects, complicating interpretation.

In addition, Nevada comprises a small number of counties, and economic activity is highly concentrated in Clark and Washoe Counties. This concentration can disproportionately influence statewide patterns and limits external validity for other institutional settings. Accordingly, we interpret our findings as evidence within Nevada’s institutional environment and assess sensitivity by excluding Clark Counties in robustness checks.

3.2. Model Specification

To analyze the relationship between local taxation and business activity while accounting for spatial dependencies, we employ a Spatial Durbin Model. The SDM framework allows us to capture both direct effects within a county and indirect effects that spill over to neighboring counties. Our regression specification is as follows:

$$\begin{array}{cc}\hfill \Delta {\mathrm{BusinessActivity}}_{it}& =a+\rho \sum _{j=1}^n{W}_{ij}\Delta {\mathrm{BusinessActivity}}_{it}\hfill \\ \hfill & +\sum _{k=1}^K\left({\mathrm{PropertyTax}}_{(it-1k)}+{\mathrm{SalesTax}}_{(it-1k)}+X_{(it-1k)}\right){\beta }_k\hfill \\ \hfill & +\sum _{k=1}^K\sum _{j=1}^n{W}_{ij}\left({\mathrm{PropertyTax}}_{(it-1k)}+{\mathrm{SalesTax}}_{(it-1k)}+X_{(it-1k)}\right){\theta }_k\hfill \\ \hfill & +{\mu }_i+{\gamma }_t+{ϵ}_{it},\hfill \end{array}$$

(1)

where $\Delta {\mathrm{BusinessActivity}}_{it}$ represents the change in the dependent variable, which can be one of the five business activity measures: business employment, annual payroll, total number of establishments, number of small establishments (1–9 employees), and number of medium-sized establishments (10–49 employees). $({\mathrm{PropertyTax}}_{(it-1k)}+{\mathrm{SalesTax}}_{(it-1k)}+X_{(it-1k)})$ is a vector of explanatory variables, including the property tax rate, sales tax rate, per capita income, and demographic controls. $\beta$ denotes the internal dependency of local lagged taxes and controls. W is the spatial weight matrix, which defines the neighboring relationships among counties. We operationalize spatial interaction among Nevada counties using two complementary weight matrices. First, we construct a first-order contiguity matrix W (queen contiguity), where $w_{ij}=1$ if counties i and j share a boundary or vertex and $w_{ij}=0$ otherwise. Second, we construct an inverse-distance matrix M based on great-circle distances between county centroids, with $m_{ij}=d_{ij}^{-1}$ for $i\ne j$ and $m_{ii}=0$. To avoid giving undue influence to very distant counties in a large and sparsely populated state, we impose a distance cutoff (e.g., retaining the nearest neighbors within a specified distance band) and set weights outside the band to zero. All matrices are row-standardized so that each row sums to one, making coefficients comparable across specifications and ensuring that estimated spillovers reflect relative exposure to neighboring counties’ policies. These two spatial weight matrices reflect distinct economic channels: contiguity captures border-shopping, commuting, and cross-county firm relocation along adjacent county lines, while inverse-distance captures broader market integration that may extend beyond immediate neighbors.

Before estimating the spatial panel models, we conduct Moran’s I tests for residual spatial dependence using OLS regressions of each outcome on the baseline covariates and year-fixed effects, under both the contiguity matrix W and the inverse-distance matrix M. The null hypothesis is that the regression disturbances are i.i.d. (no residual spatial autocorrelation). As reported in

Table 3. Moran’s I tests for residual spatial dependence (OLS residuals).

Outcome	Weight Matrix	${\mathit{\chi }}^2$	p
$\Delta$ Employment	M (inverse-distance)	0.0954	0.7574
	W (contiguity)	0.0295	0.8636
$\Delta$ Annual payroll	M	0.0715	0.7891
	W	0.0290	0.8648
$\Delta$ Establishments	M	0.4274	0.5133
	W	0.0553	0.8141
$\Delta$ Estab. 1–9	M	0.4541	0.5004
	W	0.0397	0.8421
$\Delta$ Estab. 10–49	M	0.4176	0.5181
	W	0.6795	0.4097

Notes. The table reports Moran’s I (error-lag) diagnostics applied to OLS residuals from outcome-specific regressions on baseline covariates and year-fixed effects. The null hypothesis is that regression errors are i.i.d. (no residual spatial dependence).

, the Moran’s I (error-lag) chi-square statistics are small and the p-values are uniformly large across outcomes and both weight matrices, indicating no statistically detectable residual spatial dependence in these diagnostic specifications. Even when residual diagnostics do not reject i.i.d. errors in reduced-form OLS checks, the SDM remains useful for quantifying cross-county spillovers implied by spatial interaction in tax policy and business outcomes. These diagnostics provide empirical support for modeling spatial dependence explicitly; accordingly, we adopt the Spatial Durbin Model as the baseline specification and report decomposed direct and indirect effects.3

$\rho$ is the endogenous spatial lag parameter, which captures the dependence of business activity in one county on that in neighboring counties. $\theta$ is the exogenous spatial lag parameter, which captures how tax policies and other explanatory variables in neighboring counties affect local business activity. K denotes the maximum lag length for explanatory variables, n is the number of spatial units, counties in our model. $\mu$ and $\gamma$ represent county- and time-fixed effects, respectively, controlling for unobserved spatial heterogeneity and temporal trends. $ϵ$ is the error term.

The total impact of taxation on business activity consists of two components: 1. Direct Effects: The immediate impact of a county’s own tax policy on its business environment. 2. Indirect (Spillover) Effects: The influence of tax policies in neighboring counties on local business activity.

Predictions (H1)–(H3) map directly into the SDM’s direct, indirect, and total effects. (H1) implies that the direct and total effects of the total property tax rate are not necessarily negative and may differ across the county, school, and special-district components. (H2) implies negative direct effects of the sales tax rate on business outcomes, with total effects potentially attenuated by spatial reallocation. (H3) implies that the estimated indirect effects are economically meaningful and need not be zero, consistent with border shopping, commuting, and establishment relocation across nearby counties.

The matrix of partial derivatives, which accounts for both direct and indirect effects, is given by:

$$S=\left[\begin{array}{ccc}{\displaystyle \frac{\partial (\Delta y_2)}{\partial x_{1k}}}& \dots & {\displaystyle \frac{\partial (\Delta y_2)}{\partial x_{nk}}}\\ ⋮& \ddots & ⋮\\ {\displaystyle \frac{\partial (\Delta y_{n+1})}{\partial x_{1k}}}& \dots & {\displaystyle \frac{\partial (\Delta y_{n+1})}{\partial x_{nk}}}\end{array}\right]={(I-\rho W)}^{-1}\left[\begin{array}{ccc}{\beta }_k& \dots & W_{1n}{\theta }_k\\ ⋮& \ddots & ⋮\\ W_{n1}{\theta }_k& \dots & {\beta }_k\end{array}\right]$$

(2)

$$={(I-\rho W)}^{-1}({\beta }_{k}I+W{\theta }_k),$$

(3)

where the direct effects are obtained from the diagonal elements of S. The indirect (spillover) effects arise from the off-diagonal elements, representing the influence of tax policies in one county on business activity in other counties.

This formulation extends the traditional SDM structure by explicitly modeling both tax-induced and general economic spillovers, offering a more nuanced understanding of how local tax policies interact with regional economic dynamics. By incorporating spatially lagged explanatory variables and accounting for both horizontal (neighboring counties) and vertical (state-level) interactions, our model differentiates itself from standard spatial models, such as Spatial Autoregressive (SAR) and Spatial Error (SEM) models, which often neglect these complexities. Our approach allows us to quantify how changes in taxation policies reverberate across regional economies, informing discussions on local revenue design and potential trade-offs.

To mitigate simultaneity and reverse causality between local business conditions and tax rates, all tax variables enter the regressions in lagged form. In addition, because sales tax rates may respond endogenously to contemporaneous local economic conditions or fiscal stress, we implement an instrumental-variable strategy for sales taxes in the SDM. Generally, we instrument the sales tax rate using the lagged share of registered Democrats in the county. The relevance condition is motivated by political-economy considerations: partisan composition is predictive of local fiscal preferences and therefore of the propensity to adopt or increase local sales taxes.

The exclusion restriction is more demanding. Our identifying assumption is that, conditional on county-fixed effects, year-fixed effects, and time-varying controls (income and demographics), changes in partisan registration affect short-run changes in business activity primarily through their impact on local sales tax policy rather than through other channels. Because partisan composition could also proxy for broader local policy preferences, we interpret the IV estimates as providing evidence that is more robust to contemporaneous policy endogeneity than baseline specifications while maintaining cautious language regarding strict causal interpretation. Following common practice in spatial settings, we instrument only the sales tax rate because simultaneously instrumenting multiple local tax instruments within an SDM with fixed effects is challenging and may substantially reduce finite-sample reliability.

To address potential endogeneity in the sales tax rate within a spatial Durbin framework, we implement an IV strategy using a control-function (two-step) approach that is compatible with county-fixed effects and SDM impacts. In the first stage, we regress the (lagged) sales tax rate on the instrument—the lagged share of registered Democrats—and the full set of exogenous controls, including county and year-fixed effects:

$${\mathit{SalesTax}}_{it}=\pi Z_{it-1}+{\gamma }^{\prime }{X}_{it}+{\alpha }_i+{\tau }_t+u_{it},$$

(4)

where $Z_{it-1}$ denotes partisan composition and $X_{it}$ contains the same covariates used in the baseline SDM specification. We then recover the first-stage residual ${\widehat{u}}_{it}$ and include it as an additional regressor in the second-stage SDM. This control-function term captures the component of the sales tax rate that is correlated with unobserved shocks to local economic activity, thereby mitigating reverse causality and omitted-variable bias.

The second-stage specification therefore includes both the sales tax rate (the structural policy variable of interest) and the control-function residual:

$$y_{it}=\rho W{y}_{it}+{\beta }^{\prime }{X}_{it}+{\theta }^{\prime }W{X}_{it}+\lambda {\widehat{u}}_{it}+{\alpha }_i+{\tau }_t+{\epsilon }_{it},$$

(5)

Property tax rates may also be endogenous. Counties may adjust property tax rates in response to local fiscal stress or to changes in the tax base, and these same shocks can affect business activity through multiple channels. Our baseline specification mitigates these concerns by including county- and year-fixed effects to absorb time-invariant county characteristics and statewide shocks, and by incorporating a rich set of time-varying controls. We therefore interpret estimated relationships involving property taxes as conditional associations rather than strictly causal effects. As a robustness check, we also re-estimate models excluding Clark counties to verify that results are not driven by the two largest metropolitan areas.

4. Results

We present our regression results in

Table 4. Regressions with Total Property Tax Rate and Sales Tax Rate Across Multiple Outcomes with Contiguity Matrix.

		Employment		Payroll		Establishments		Estab. (1–9)		Estab. (10–49)
		Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)
Direct	Total property tax rate	3569.2	1179.3	173,335.4	2949.1	153.5	−37.46	89.45	−56.23	46.26	15.09
		(4689.5)	(4944.0)	(218,792.2)	(229,458.7)	(133.1)	(137.8)	(91.43)	(95.35)	(36.61)	(37.72)
	Total sales tax rate	−914.5	−372,131.0 **	−4810.0	−17,009,417.8 **	−129.4	−13,430.7 ***	−117.4	−9264.7 ***	−5.341	−2812.7 **
		(7995.1)	(155,136.2)	(373,525.4)	(7,206,895.6)	(228.7)	(4328.0)	(158.8)	(3000.5)	(61.88)	(1183.7)
	Per capita income	−0.366 *	0.729 *	−16.29 *	33.50	−0.0113 **	0.0258 **	−0.00728 **	0.0176 **	−0.00263 *	0.00571 *
		(0.187)	(0.438)	(8.726)	(20.38)	(0.00532)	(0.0122)	(0.00366)	(0.00851)	(0.00145)	(0.00334)
	Age 20∼64	158,807.0 *	−62,712.5	4,890,639.5	−4,707,311.6	2321.2	−5774.3 *	981.2	−4819.7 **	880.8	−667.2
		(83,139.2)	(124,520.6)	(3,883,014.1)	(5,800,260.7)	(2370.6)	(3499.7)	(1637.7)	(2425.0)	(646.4)	(956.0)
	Age 65+	131,607.9 *	−294,261.3	4,077,008.7	−15,490,752.2 *	1979.9	−13,880.4 ***	751.0	−10,323.7 ***	829.0	−2405.7 *
		(74,777.5)	(189,228.2)	(3,489,718.4)	(8,788,716.4)	(2128.6)	(5279.2)	(1467.3)	(3650.0)	(581.9)	(1448.9)
Indirect	Total property tax rate	11,305.8	27,962.3 ***	456,820.6	1,148,232.0 ***	306.9	700.6 ***	256.0 *	476.3 ***	17.22	127.1 **
		(7316.5)	(8217.8)	(341,640.3)	(382,230.3)	(212.3)	(231.2)	(149.3)	(163.1)	(55.99)	(61.21)
	Total sales tax rate	3180.1	−114,404.9	374,145.6	7,242,487.8	319.1	12,425.1 *	239.6	9122.5 *	67.61	2770.6
		(17,344.4)	(264,829.8)	(815,734.1)	(12,358,863.7)	(505.5)	(7494.0)	(358.6)	(5311.7)	(132.6)	(1980.5)
	Per capita income	−0.0668	0.249	−8.649	−17.44	−0.0117	−0.0348	−0.00885	−0.0249	−0.00236	−0.00816
		(0.351)	(0.790)	(16.44)	(36.96)	(0.0102)	(0.0224)	(0.00722)	(0.0159)	(0.00269)	(0.00590)
	Age 20∼64	108,190.7	503,561.6 *	3,726,931.9	32,550,313.2 ***	1868.4	26,388.0 ***	1191.6	15,839.4 ***	410.0	7783.6 ***
		(96,239.6)	(267,644.2)	(4,409,450.2)	(12,585,571.0)	(2731.0)	(7677.9)	(1914.0)	(5432.2)	(720.2)	(2036.4)
	Age 65+	−65,407.9	226,049.0	−3,485,292.7	22,884,443.7	−2552.2	24,454.1 ***	−1851.8	15,730.1 **	−513.4	6670.7 ***
		(79,879.4)	(329,023.3)	(3,717,130.0)	(15,395,679.0)	(2293.1)	(9340.7)	(1617.0)	(6607.3)	(609.9)	(2480.7)
	nonwhite	46,970.0	136,208.2	2,091,486.4	3,278,237.9	1927.6	1162.8	1379.0	179.4	361.1	620.0
		(56,609.3)	(88,159.0)	(2,652,031.6)	(4,108,356.1)	(1648.1)	(2491.1)	(1163.8)	(1761.7)	(435.1)	(660.5)
Total	Total property tax rate	14,874.9 **	29,141.6 ***	630,155.9 **	1,151,181.1 ***	460.4 **	663.2 ***	345.4 **	420.1 **	63.48	142.2 **
		(6768.4)	(8037.8)	(316,498.2)	(375,402.2)	(198.6)	(229.1)	(142.1)	(163.7)	(51.06)	(59.11)
	Total sales tax rate	2265.5	−486,535.9 *	369,335.7	−9,766,930.0	189.6	−1005.5	122.2	−142.2	62.27	−42.09
		(20,811.0)	(275,553.7)	(978,633.3)	(12,918,255.1)	(608.2)	(7886.3)	(433.2)	(5657.5)	(158.6)	(2043.4)
	Per capita income	−0.433	0.978	−24.94	16.07	−0.0230 **	−0.00892	−0.0161 **	−0.00738	−0.00499 *	−0.00245
		(0.364)	(0.865)	(17.09)	(40.64)	(0.0107)	(0.0248)	(0.00762)	(0.0178)	(0.00277)	(0.00643)
	Age 20∼64	266,997.7 **	440,849.2	8,617,571.4	27,843,001.6 **	4189.6	20,613.7 **	2172.8	11,019.8 *	1290.9	7116.4 ***
		(127,655.8)	(287,948.0)	(5,912,170.7)	(13,616,099.9)	(3674.3)	(8370.9)	(2607.1)	(5956.9)	(956.8)	(2196.3)
	Age 65+	66,199.9	−68,212.4	591,716.1	7,393,691.5	−572.3	10,573.7	−1100.8	5406.4	315.5	4265.0 *
		(100,370.1)	(316,254.1)	(4,703,000.4)	(14,894,446.3)	(2925.0)	(9121.4)	(2084.2)	(6536.1)	(763.1)	(2373.2)
	nonwhite	−12,417.4	129,205.8	−1,190,063.9	2,455,586.7	−817.6	474.9	−702.1	−380.4	−70.34	535.0
		(58,670.5)	(85,428.6)	(2,753,780.0)	(3,993,751.8)	(1722.7)	(2439.0)	(1229.0)	(1746.3)	(447.3)	(632.9)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$.

,

Table 5. Regressions with Property Tax Rate Components and Sales Tax Rate Across Multiple Outcomes with Contiguity Matrix.

		Employment		Payroll		Establishments		Estab. (1–9)		Estab. (10–49)
		Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)
Direct	Combined special district rate	−562.9	19,593.1 *	−57,936.2	905,759.7	−99.92	773.9 **	−80.29	586.6 **	−19.02	120.4
		(7556.1)	(11,810.4)	(353,940.7)	(551,364.0)	(210.9)	(322.8)	(147.7)	(227.9)	(56.86)	(86.74)
	School tax rate	−26,025.9 *	−142,410.7 ***	−1,210,066.5 *	−7,162,189.0 ***	−925.7 **	−6016.4 ***	−657.3 **	−4316.9 ***	−183.1 *	−1184.2 ***
		(14,844.8)	(46,865.2)	(694,746.6)	(2,183,651.6)	(416.0)	(1270.8)	(293.3)	(888.6)	(109.5)	(350.6)
	County tax rate	6665.3	−18,443.1 *	281,580.2	−1,015,867.9 **	122.4	−974.2 ***	12.03	−767.6 ***	83.89 *	−139.5 *
		(5835.3)	(11,117.6)	(273,243.3)	(517,793.2)	(162.1)	(300.8)	(112.8)	(209.0)	(44.47)	(83.98)
	Total sales tax rate	−1144.3	−420,046.5 **	6778.6	−21,156,103.3 ***	−4.216	−18,240.3 ***	0.0730	−13,244.0 ***	6.574	−3482.6 ***
		(8854.3)	(164,207.6)	(414,926.7)	(7,656,260.5)	(248.0)	(4462.0)	(174.9)	(3118.3)	(66.05)	(1229.8)
Indirect	Combined special district rate	34,575.5 ***	28,139.4	1,639,983.2 ***	953,141.1	1018.7 ***	701.8	729.6 ***	721.7	156.8 *	−76.92
		(11,904.4)	(22,771.9)	(556,266.6)	(1,066,663.6)	(333.9)	(637.1)	(241.4)	(466.0)	(83.06)	(156.9)
	School tax rate	−103,070.5 ***	−18,775.0	−4,825,094.8 ***	940,341.4	−3134.2 ***	691.3	−2007.4 **	−120.8	−777.2 ***	664.4
		(38,369.7)	(89,409.4)	(1,792,237.5)	(4,181,644.8)	(1107.6)	(2496.2)	(799.8)	(1806.1)	(265.2)	(628.3)
	County tax rate	13,179.8	33,879.1 *	627,980.6	1,960,686.5 **	529.2 *	1476.9 ***	368.8 *	891.4 **	99.54	408.3 ***
		(10,332.2)	(18,301.7)	(483,863.7)	(856,632.8)	(290.5)	(505.7)	(208.2)	(362.0)	(74.90)	(131.7)
	Total sales tax rate	24,863.9	313,833.5	1,282,562.7	21,886,355.2	1105.0 *	14,571.0 *	793.3 *	7071.5	257.5 *	5541.7 **
		(19,793.3)	(311,358.1)	(932,517.6)	(14,594,770.4)	(566.8)	(8653.7)	(412.1)	(6253.7)	(140.4)	(2204.8)
Total	Combined special district rate	34,012.7 **	47,732.5 *	1,582,047.0 **	1,858,900.8	918.8 **	1475.7 *	649.3 **	1308.3 **	137.8	43.43
		(14,063.9)	(28,635.0)	(657,774.6)	(1,343,376.2)	(397.5)	(804.2)	(290.0)	(590.0)	(95.93)	(195.7)
	School tax rate	−129,096.4 ***	−161,185.6 *	−6,035,161.3 ***	−6,221,847.6	−4059.8 ***	−5325.2 **	−2664.7 ***	−4437.7 **	−960.3 ***	−519.8
		(44,996.2)	(94,902.0)	(2,101,818.4)	(4,448,156.8)	(1300.6)	(2675.4)	(943.4)	(1964.2)	(306.3)	(646.2)
	County tax rate	19,845.1 **	15,436.0	909,560.9 **	944,818.6	651.6 **	502.7	380.8 *	123.8	183.4 ***	268.9 **
		(9817.5)	(17,300.6)	(460,064.3)	(813,104.0)	(278.5)	(486.3)	(203.5)	(354.4)	(68.46)	(120.3)
	Total sales tax rate	23,719.6	−106,212.9	1,289,341.2	730,251.8	1100.8 *	−3669.3	793.4	−6172.5	264.1	2059.1
		(23,175.0)	(340,348.1)	(1,091,859.7)	(15,996,957.2)	(666.3)	(9555.2)	(487.9)	(6980.7)	(161.7)	(2357.0)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$.

,

Table 6. Regressions with Property Tax Rate Components, the Sales Tax Rate, and Interactions with post-2005 dummy and Contiguity Matrix (Direct and Indirect Effects without IV).

	Direct Effects					Indirect Effects
	Employment	Annual Payroll	Number of Establishments	Establishment Size: 1–9	Establishment Size: 10–49	Employment	Annual Payroll	Number of Establishments	Establishment Size: 1–9	Establishment Size: 10–49
Combined special district rate	6397.4	337,130.4	225.2	192.3	4.636	49,113.3 ***	2,332,795.7 ***	1592.9 ***	1142.4 ***	268.6 **
	(9169.7)	(425,162.2)	(249.7)	(173.6)	(71.10)	(15,904.3)	(737,409.2)	(435.8)	(311.7)	(114.3)
School tax rate	−34,815.2 **	−1,667,490.3 **	−1188.1 ***	−824.1 ***	−246.2 **	−92,651.5 **	−4,211,203.9 **	−3091.1 ***	−2057.9 **	−727.6 **
	(15,072.0)	(697,407.1)	(412.3)	(288.0)	(115.3)	(41,716.4)	(1,929,421.9)	(1164.9)	(829.2)	(298.6)
County tax rate	−2159.4	−125,923.6	−146.3	−172.8	35.02	14,578.2	716,723.5	673.1 **	520.1 **	91.23
	(6111.9)	(283,012.4)	(165.8)	(114.8)	(47.96)	(11,452.0)	(531,330.3)	(312.1)	(221.6)	(85.01)
Total sales tax rate	12,231.7	799,756.6 *	534.1 *	421.7 **	64.46	33,614.1	1,759,567.5	1543.2 **	1235.6 ***	238.1
	(10,248.7)	(475,668.5)	(279.6)	(195.6)	(78.67)	(23,036.0)	(1,078,384.8)	(643.1)	(464.2)	(166.2)
Per capita income	−0.143	−3.377	−0.00366	−0.00186	−0.00133	0.325	19.67	0.00370	0.00505	−0.00216
	(0.216)	(10.04)	(0.00588)	(0.00412)	(0.00165)	(0.482)	(22.47)	(0.0131)	(0.00943)	(0.00347)
Age 20–64	237,470.2 **	9,592,985.5 *	5872.4 **	3008.0	1959.9 **	1,170,798.3 ***	52,529,250.5 ***	31,706.0 ***	16,740.8 ***	10,298.3 ***
	(110,054.5)	(5,084,786.6)	(2990.5)	(2078.6)	(845.7)	(256,297.8)	(11,810,028.2)	(6940.3)	(4817.0)	(1871.7)
Age 65+	129,354.8	4,719,452.7	2750.3	1022.4	1274.1 *	717,068.0 ***	30,158,405.9 ***	19,669.1 ***	10,890.6 ***	6018.0 ***
	(89,162.2)	(4,119,542.5)	(2424.6)	(1684.3)	(689.1)	(191,508.2)	(8,779,855.8)	(5191.1)	(3632.2)	(1396.3)
Nonwhite population	21,599.2	162,895.5	−1025.2	−983.6	−61.03	37,164.9	901,256.4	2326.4	1323.2	840.5
	(37,829.1)	(1,751,150.8)	(1026.0)	(708.4)	(298.8)	(94,044.0)	(4,360,988.1)	(2565.5)	(1821.9)	(697.6)
Interactions with post-2005 dummy
Combined special district rate × post-2005 dummy	−7562.1	−435,116.1	−363.9 *	−295.7 *	−36.83	31,456.7 *	1,533,678.5 *	611.3	431.7	55.86
	(7938.5)	(367,791.5)	(215.7)	(151.4)	(60.17)	(17,334.2)	(805,624.7)	(471.3)	(338.7)	(124.1)
School tax rate × post-2005 dummy	−4149.3	−175,879.1	−70.37	−16.56	−43.50	24,490.8	1,179,918.2	799.3	585.7	117.3
	(8343.1)	(387,542.3)	(227.4)	(160.6)	(62.41)	(18,723.0)	(871,602.4)	(512.4)	(370.3)	(133.2)
County tax rate × post-2005 dummy	9703.5 ***	469,683.3 ***	300.3 ***	211.5 ***	47.40 *	9201.3	369,547.5	86.65	−21.62	70.87
	(3618.7)	(167,698.1)	(98.35)	(69.21)	(27.42)	(11,890.8)	(551,658.6)	(324.1)	(232.3)	(86.10)
Total sales tax rate × post-2005 dummy	−13,149.6 **	−770,002.7 ***	−527.7 ***	−406.2 ***	−62.85	−2902.7	−149,758.8	−179.6	−237.1	47.76
	(5689.5)	(263,703.4)	(154.4)	(107.1)	(44.43)	(11,086.0)	(515,170.5)	(303.1)	(216.9)	(81.59)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$.

,

Table 7. Regressions with Property Tax Rate Components, the Sales Tax Rate, and the Interactions with post-2005 dummy (Total Effects without IV).

	Employment	Annual Payroll	Number of Establishments	Establishment Size: 1–9	Establishment Size: 10–49
Combined special	55,510.7 ***	2,669,926.1 ***	1818.1 ***	1334.7 ***	273.2 **
district rate	(18,441.5)	(856,760.7)	(507.7)	(365.6)	(129.6)
School tax rate	−127,466.7 ***	−5,878,694.2 ***	−4279.2 ***	−2882.0 ***	−973.7 ***
	(46,216.0)	(2,137,093.7)	(1296.9)	(929.6)	(322.9)
County tax rate	12,418.8	590,799.9	526.8 *	347.3	126.3
	(11,065.0)	(514,116.4)	(302.7)	(218.2)	(79.01)
Total sales tax rate	45,845.8 *	2,559,324.0 **	2077.3 ***	1657.2 ***	302.5
	(26,762.1)	(1,254,097.6)	(748.8)	(543.6)	(189.7)
Per capita income	0.182	16.29	0.0000386	0.00319	−0.00350
	(0.586)	(27.31)	(0.0160)	(0.0115)	(0.00416)
Age 20–64	1,408,268.5 ***	62,122,236.0 ***	37,578.4 ***	19,748.9 ***	12,258.2 ***
	(300,611.2)	(13,837,426.9)	(8156.8)	(5690.3)	(2165.5)
Age 65+	846,422.7 ***	34,877,858.6 ***	22,419.4 ***	11,913.0 ***	7292.1 ***
	(213,639.3)	(9,778,151.3)	(5809.3)	(4091.8)	(1520.5)
Nonwhite population	58,764.0	1,064,151.9	1301.2	339.6	779.5
	(87,913.2)	(4,080,100.0)	(2406.7)	(1733.1)	(629.9)
Interactions with post-2005 dummy:
Combined special district rate × post-2005 dummy	23,894.6	1,098,562.4	247.4	136.0	19.03
	(21,443.6)	(996,583.4)	(583.6)	(420.9)	(151.5)
School tax rate × post-2005 dummy	20,341.5	1,004,039.1	728.9	569.1	73.82
	(25,255.4)	(1,175,732.7)	(691.1)	(498.5)	(180.1)
County tax rate × post-2005 dummy	18,904.8	839,230.8	386.9	189.9	118.3
	(13,576.4)	(630,062.2)	(370.3)	(266.9)	(96.88)
Total sales tax rate × post-2005 dummy	−16,052.3	−919,761.5 *	−707.3 **	−643.3 ***	−15.10
	(11,348.7)	(528,610.7)	(311.2)	(225.8)	(80.66)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$.

,

Table 8. IV Regressions with Property Tax Rate Components, the Sales Tax Rate, and Interactions with post-2005 dummy (Direct and Indirect Effects).

	Direct Effects					Indirect Effects
	Employment	Annual Payroll	Number of Establishments	Establishment Size: 1–9	Establishment Size: 10–49	Employment	Annual Payroll	Number of Establishments	Establishment Size: 1–9	Establishment Size: 10–49
Combined special district rate	19,179.0	881,278.3	866.7 **	686.3 ***	117.2	30,203.9	686,340.5	453.9	458.3	−63.80
	(13,901.8)	(641,552.1)	(371.3)	(259.1)	(106.1)	(27,914.7)	(1,287,778.8)	(755.3)	(543.3)	(199.9)
School tax rate	−108,992.0 **	−5,348,799.4 **	−5029.9 ***	−3665.5 ***	−1006.0 **	383.1	3,164,702.3	2334.0	1333.5	770.9
	(52,690.6)	(2,429,599.9)	(1402.2)	(971.9)	(407.9)	(100,839.7)	(4,646,870.1)	(2724.6)	(1945.2)	(732.4)
County tax rate	−18,232.2	−953,465.0 *	−981.6 ***	−781.8 ***	−136.5	36,819.9 *	2,355,122.3 **	1960.7 ***	1359.5 ***	421.9 ***
	(12,186.5)	(561,990.7)	(324.3)	(223.5)	(95.25)	(21,858.0)	(1,010,812.0)	(591.2)	(419.7)	(161.7)
Total sales tax rate	−255,874.6	−12,165,129.6	−13,212.8 ***	−9807.1 ***	−2615.2 *	381,404.4	30,034,668.2 *	22,289.2 **	14,171.9 **	5966.0 **
	(189,048.8)	(8,723,373.1)	(5043.5)	(3495.5)	(1462.9)	(363,778.3)	(16,835,288.4)	(9868.9)	(7052.3)	(2652.8)
Per capita income	0.381	21.96	0.0232 **	0.0182 **	0.00390	−0.356	−36.48	−0.0370	−0.0201	−0.0136 **
	(0.427)	(19.70)	(0.0114)	(0.00793)	(0.00328)	(0.877)	(40.57)	(0.0238)	(0.0170)	(0.00635)
Age 20–64	−81,774.3	−5,545,864.4	−10,361.6	−9147.4 *	−1167.3	1,576,374.0 ***	86,681,902.4 ***	56,495.9 ***	31,956.4 ***	17,218.5 ***
	(255,726.1)	(11,797,959.2)	(6841.6)	(4743.7)	(1972.6)	(524,565.5)	(24,320,513.4)	(14,311.9)	(10,151.4)	(3829.0)
Age 65+	−238,110.9	−13,349,038.4	−16,108.6 **	−12,926.5 ***	−2470.3	1,218,818.0 **	69,844,169.4 ***	49,546.1 ***	29,792.7 ***	14,035.0 ***
	(269,464.5)	(12,429,553.1)	(7195.3)	(4973.9)	(2092.8)	(528,301.9)	(24,454,383.5)	(14,381.9)	(10,221.3)	(3872.3)
Nonwhite population	98,775.8	4,353,402.8	3048.3 *	1940.4 *	805.0 *	−65,511.0	−5,687,519.8	−3340.7	−2570.7	−486.2
	(61,421.2)	(2,830,865.9)	(1627.9)	(1118.5)	(482.7)	(115,388.2)	(5,328,808.1)	(3105.4)	(2198.8)	(857.9)
Interactions with post-2005 dummy
Combined special district rate × post-2005 dummy	−8167.7	−450,139.1	−386.9 *	−312.6 **	−41.49	38,056.4 **	1,945,064.4 **	985.5 **	701.0 **	133.9
	(7957.8)	(367,379.5)	(213.8)	(150.2)	(59.77)	(17,834.2)	(828,056.4)	(485.6)	(350.6)	(126.5)
School tax rate × post-2005 dummy	−1919.5	−75,909.2	44.41	71.32	−22.90	22,354.7	921,179.4	658.9	519.7	63.53
	(8477.4)	(392,181.2)	(228.4)	(161.6)	(62.84)	(19,335.4)	(897,074.5)	(528.1)	(383.5)	(136.4)
County tax rate × post-2005 dummy	9286.1 **	467,853.9 ***	282.6 ***	192.5 ***	48.15 *	6565.3	241,422.0	−50.63	−129.7	47.16
	(3696.1)	(170,651.8)	(99.21)	(70.01)	(27.70)	(11,976.4)	(554,109.6)	(325.0)	(234.2)	(85.94)
Total sales tax rate × post-2005 dummy	−10,593.8 *	−641,768.8 **	−397.2 **	−311.7 ***	−35.15	−9073.4	−635,598.8	−550.6	−479.3 *	−48.63
	(5917.1)	(273,460.7)	(157.8)	(109.2)	(45.88)	(12,563.4)	(582,646.9)	(342.8)	(246.6)	(91.54)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$.

and

Table 9. IV Regressions with Property Tax Rate Components, the Sales Tax Rate, and the Interactions with post-2005 dummy (Total Effects).

	Employment	Annual Payroll	Number of Establishments	Establishment Size: 1–9	Establishment Size: 10–49
Combined special district rate	49,382.9	1,567,618.7	1320.6	1144.6 *	53.44
	(32,637.5)	(1,508,031.7)	(887.4)	(643.3)	(229.3)
School tax rate	−108,608.9	−2,184,097.1	−2695.9	−2332.0	−235.1
	(104,408.9)	(4,817,244.0)	(2841.5)	(2060.1)	(732.2)
County tax rate	18,587.7	1,401,657.3	979.1 *	577.8	285.3 **
	(20,040.5)	(929,890.9)	(550.3)	(397.9)	(142.5)
Total sales tax rate	125,529.7	17,869,538.6	9076.4	4364.9	3350.8
	(387,970.9)	(17,997,254.4)	(10,620.0)	(7679.7)	(2754.1)
Per capita income	0.0244	−14.52	−0.0138	−0.00195	−0.00966
	(0.977)	(45.27)	(0.0266)	(0.0193)	(0.00692)
Age 20–64	1,494,599.8 **	81,136,038.1 ***	46,134.3 ***	22,809.0 **	16,051.2 ***
	(590,256.5)	(27,398,583.0)	(16,225.9)	(11,581.9)	(4238.5)
Age 65+	980,707.1 *	56,495,131.0 **	33,437.5 **	16,866.1	11,564.7 ***
	(549,281.7)	(25,480,915.3)	(15,129.4)	(10,863.5)	(3925.5)
Nonwhite population	33,264.8	−1,334,117.0	−292.4	−630.3	318.7
	(94,745.4)	(4,388,299.8)	(2582.0)	(1870.2)	(670.5)
Interactions with post-2005 dummy:
Combined special district rate × post-2005 dummy	29,888.7	1,494,925.3	598.6	388.4	92.38
	(21,854.0)	(1,014,829.7)	(597.3)	(433.1)	(152.9)
School tax rate × post-2005 dummy	20,435.2	845,270.2	703.3	591.0	40.62
	(25,854.4)	(1,199,528.3)	(705.7)	(511.5)	(182.7)
County tax rate × post-2005 dummy	15,851.4	709,275.8	232.0	62.74	95.31
	(13,747.7)	(636,391.3)	(374.3)	(271.3)	(97.24)
Total sales tax rate × post-2005 dummy	−19,667.3	−1,277,367.7 **	−947.8 ***	−791.1 ***	−83.78
	(12,715.4)	(592,180.0)	(351.1)	(256.3)	(89.60)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$.

, using the SDM specification, with both county- and year-fixed effects incorporated into all regressions.4 The first set of regressions, reported in Table 4, focuses on the total property tax rate and the sales tax rate as the main explanatory variables. In this section, we report the results from instrumental variable (IV) regressions, where the sales tax rate is instrumented using the share of registered Democrats in each county. Additionally, we include Non-IV regressions for comparison purposes.

In this set of regressions, we observe a predominantly positive association between the total property tax rate and the change in business activity variables. The total effect of the property tax rate is consistently positive and statistically significant across all regressions, with the exception of the regression involving the change in the number of establishments with 10–49 employees. These results hold true for both IV and Non-IV specifications. Notably, the significance of direct and indirect effects varies across different regression models. However, the overall relationship between the property tax rate and business activity remains positive in all cases. The indirect effect, which reflects spillover effects from neighboring counties, is positive and statistically significant only at the 10% level, suggesting that property tax increases in one county may have positive spillover effects on neighboring counties’ business activity.

A positive association between certain property tax components and business activity is consistent with a “tax–service bundle” view in local public finance: higher effective tax effort may coincide with greater provision of productivity-enhancing local public goods (e.g., infrastructure, public safety, utilities, and other services typically financed at the county or special-district level). Under this channel, taxes need not be purely distortive if they proxy for public inputs that raise local productivity. At the same time, the estimates should not be interpreted as strictly causal effects. Counties with stronger underlying growth prospects may both expand business activity and choose higher tax rates or create special districts to finance capital projects; measurement differences in service quality and local development strategies may also generate positive correlations. We therefore present the results as associations and discuss them alongside plausible competing explanations.

While the sales tax rate is generally associated with weaker business activity in the direct effects, the corresponding total effects are less uniform across outcomes and specifications. In several regressions, estimated spillovers (indirect effects) partially offset the own-county effect, so the total effect is not always statistically distinguishable from zero. However, the sales tax rate does not ‘disappear’ in all cases: for some outcomes—most notably establishment-related measures—the estimated net effect remains negative and, in some specifications, statistically significant. Overall, the pattern is consistent with cross-county reallocation that can attenuate the net effect even when local (direct) impacts are present, and the IV specifications tend to produce more pronounced and more frequently significant estimates. Importantly, significant results for the sales tax rate are more common in the IV regressions, which further highlights the importance of considering the potential endogeneity of the sales tax rate in these analyses. To gauge economic magnitude, consider the IV estimates in Table 4. Because the sales tax rate varies by about 0.32 percentage points (one standard deviation; Table 1), a one-standard-deviation increase corresponds to sizeable changes in outcomes. For example, using the IV direct-effect coefficients, a 0.32 percentage-point increase is associated with approximately 49,600 fewer jobs and about 1870 fewer establishments (in changes).

Other control variables, such as per capita income and demographic factors, also show significant associations with business activity. Specifically, per capita income tends to have a negative association with business activity, while certain demographic variables—particularly those related to the population age structure—show a positive relationship with business activity, particularly in the direct effects regressions.

In the second set of regressions, presented in Table 5, we disaggregate the total property tax rate into its components: the county rate, the school rate, and the combined special district rate. Here, we find a positive association between both the county tax rate and the combined special district rate with changes in business activity, which provides further insight into the positive results observed in the previous set of regressions. However, the school tax rate exhibits a negative association with business activity. For the property tax components, the IV direct-effect estimate for the county tax rate implies that a 0.1 increase is associated with roughly 31 additional establishments (in changes). This magnitude is consistent with the view that the net effect of property taxation may reflect both a cost channel and a public-input channel financed by property tax revenues. This result may seem counterintuitive, but it can be explained by the fact that businesses are likely to perceive greater direct benefits from county-level public services (financed by the county tax rate) and from projects funded through special district taxes. In contrast, the school tax rate, which funds education, may be seen as providing more indirect benefits, especially for businesses themselves, rather than having an immediate impact on their operations.

The pattern across property tax components is consistent with a tax–service bundle interpretation. Positive associations for the county rate and the combined special district rate align with the notion that these revenues finance more visible and business-relevant local public inputs (e.g., local infrastructure, public safety, or project-based services), which can offset the cost channel. In contrast, the negative association for the school tax rate is consistent with a channel in which firms perceive education spending as providing more diffuse or longer-horizon benefits, leading to weaker contemporaneous links with establishment and employment outcomes.

For sales taxes, the negative direct effects are consistent with demand-side mechanisms and border-related reallocation emphasized in the literature: higher sales tax rates can reduce local retail demand and shift transactions and activity toward nearby jurisdictions. Throughout, we interpret these estimates as conditional associations within Nevada’s institutional setting rather than definitive causal effects, given the potential for policy endogeneity and unobserved time-varying local shocks.

Regarding the total effects of these tax components, we find that the county tax rate and the combined special district rate are positively and significantly associated with business activity. However, the school tax rate consistently shows a negative relationship with business activity in most of the Non-IV regression results. The sales tax rate continues to exhibit a negative association with business activity, similar to the findings in the earlier regressions. This negative association is particularly significant in the regressions involving establishments (columns 7–12 in Table 5). Spatial spillovers are heterogeneous. Many indirect effects are imprecisely estimated, but several specifications—especially those involving the post-2005 interactions—indicate economically meaningful spillovers.

In the third set of regression results, displayed in Table 6, Table 7, Table 8 and Table 9, we introduce an interaction term between our key variables and a dummy variable for the introduction of the property tax abatement program in 2005. This allows us to assess whether the policy change had any significant effect on the relationship between our tax variables and business activity. The results from these regressions largely mirror those from the previous models, with a notable exception: the positive association between the county property tax rate and changes in business activity becomes particularly robust and significant after the introduction of the property tax abatement program. This may indicate that the tax abatement policy sent a positive signal to businesses, suggesting that future property tax increases would be constrained, which in turn could coincide with higher measured business activity.

At the same time, we observe a consistently negative and significant relationship between the sales tax rate and business activity after 2005. This pattern is consistent with a substitution-type narrative in which local governments may have relied more heavily on sales tax measures in the post-2005 regime; however, we treat this as an interpretive mechanism rather than a causal claim. Additionally, the school tax rate continues to show a negative and significant association with business activity, especially in the direct effects regressions (first half of Table 8) and total effects regressions (Table 9). The interaction between the school tax rate and the post-2005 dummy is unstable across specifications.

Finally, the combined special district rate remains positively associated with changes in business activity, though the relationship is significant only in some of the indirect effects regressions (second half of Table 8), pointing to potential spillover effects from this tax. The interaction term for the special district rate and the post-2005 dummy is particularly significant in the Non-IV results (second half of Table 6), suggesting that the policy reform may have had a more pronounced impact on business activity in areas funded by special district taxes.

In sum, the results from these regressions highlight the complexity of the relationship between property tax, sales tax, and business activity, with varying effects across different tax components and spatial spillovers. These findings underscore the importance of considering both direct and indirect effects in understanding the broader impact of local taxation on economic activity. We assess the stability of key findings along three dimensions already embedded in our empirical design. First, we compare baseline SDM estimates with specifications that instrument the sales tax rate; the negative association between sales taxes and business activity is most pronounced in the IV specifications, consistent with concerns about policy endogeneity. Second, disaggregating the total property tax rate into county, school, and special district components reveals that the positive association for the aggregate rate is not uniform across components, and is instead concentrated in the county and special district rates, while the school rate is persistently negative. Third, interaction models that allow coefficients to differ after the 2005 property tax abatement reform largely preserve the baseline patterns while highlighting a stronger negative relationship between sales taxes and business activity in the post-2005 period. Taken together, these comparisons suggest that the qualitative conclusions are not driven by a single specification choice and that institutional context (e.g., the post-2005 regime) matters for interpretation.

5. Robustness and Sensitivity Analyses

This section evaluates whether our main conclusions depend on (i) the choice of spatial weights matrix and (ii) the presence of a single economically dominant county. We implement two complementary robustness checks: First, we re-estimate the SDM using an inverse-distance spatial weights matrix in place of the baseline contiguity matrix. Second, we re-estimate the same specifications after excluding Clark County. In all cases, we continue to report SDM impacts (direct, indirect, and total effects) to maintain comparability with the baseline results.

To assess sensitivity to how spatial interaction is operationalized, we re-estimate our main specifications using an inverse-distance matrix (denoted M) instead of the contiguity matrix (denoted W). Whereas contiguity weights primarily reflect border-local mechanisms (e.g., border shopping, commuting, and relocation along adjacent county lines), inverse-distance weights capture broader regional integration that may extend beyond immediate neighbors. Both matrices are row-standardized to facilitate comparisons across specifications and ensure that spillovers are interpretable as relative exposure to neighboring jurisdictions’ policies.

Before turning to the SDM estimates, we note that residual spatial diagnostics (Moran’s I tests on OLS residuals) do not indicate strong residual spatial dependence under either weighting scheme. Nevertheless, the SDM framework remains informative because it explicitly decomposes tax impacts into own-county (direct) and cross-county (indirect) components implied by spatial interaction in policies and outcomes.

Table 10. Robustness: Regressions with Total Property Tax Rate and Sales Tax Rate Across Multiple Outcomes (Inverse-Distance Matrix M).

		Employment		Payroll		Establishments		Estab. (1–9)		Estab. (10–49)
		Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)
Direct	Total property tax rate	−2159.2	575.7	−109,392.5	−73,134.5	−51.68	−78.71	−32.27	−64.03	−12.59	−12.82
		(4745.9)	(5159.9)	(220,785.2)	(239,832.7)	(136.1)	(145.5)	(93.65)	(100.4)	(36.94)	(39.72)
	Total sales tax rate	−1018.3	−240,968.2	−29,275.4	−10,694,677.0	−62.56	−9260.4 **	−55.83	−6358.9 **	4.348	−2000.3 *
		(8230.2)	(152,424.1)	(382,339.7)	(7,079,882.9)	(232.2)	(4316.0)	(163.2)	(2992.4)	(61.94)	(1173.3)
	Per capita income	−0.493 **	0.182	−22.54 **	5.395	−0.0130 **	0.00973	−0.00750 *	0.00779	−0.00363 **	0.00145
		(0.223)	(0.457)	(10.36)	(21.32)	(0.00627)	(0.0130)	(0.00440)	(0.00904)	(0.00167)	(0.00349)
	Age 20∼64	80,432.6	−29,224.6	2,132,670.9	−3,490,910.9	1310.8	−4058.5	538.0	−3244.8	552.3	−573.4
		(85,895.7)	(118,479.9)	(3,993,373.1)	(5,524,976.6)	(2439.8)	(3386.2)	(1701.3)	(2343.9)	(655.3)	(920.3)
	Age 65+	51,107.3	−184,798.0	1,000,377.2	−10,460,194.8	619.2	−9933.2 *	−40.34	−7408.4 **	511.0	−1726.7
		(76,255.1)	(181,761.1)	(3,544,276.8)	(8,443,838.0)	(2165.2)	(5165.1)	(1509.9)	(3568.0)	(582.1)	(1409.4)
	nonwhite	−65,804.4 **	−13,600.1	−3,604,590.0 ***	−1,148,439.7	−3159.1 ***	−958.7	−2361.7 ***	−827.7	−530.1 **	−61.92
		(28,440.6)	(45,079.3)	(1,320,483.4)	(2,092,986.4)	(805.7)	(1268.8)	(557.9)	(877.8)	(217.9)	(345.4)
Indirect	Total property tax rate	43,922.8 ***	53,445.2 ***	2,080,416.6 ***	2,242,713.1 ***	1232.7 ***	1189.1 ***	805.7 ***	732.5 **	275.0 ***	281.9 **
		(14,599.7)	(16,348.6)	(677,886.9)	(757,197.7)	(399.8)	(440.7)	(286.5)	(317.1)	(102.4)	(112.9)
	Total sales tax rate	2956.7	−407,798.5	203,572.4	7,475,674.7	980.5	24,337.0	816.5	21,335.5 *	192.2	3432.4
		(37,693.7)	(585,610.7)	(1,751,805.8)	(27,492,174.9)	(1032.8)	(16,528.3)	(755.6)	(12,197.6)	(263.7)	(4176.2)
	Per capita income	−1.610 *	−0.397	−83.16 **	−93.42	−0.0406 *	−0.0924 **	−0.0215	−0.0679 **	−0.0134 **	−0.0202 *
		(0.833)	(1.616)	(38.77)	(77.30)	(0.0224)	(0.0468)	(0.0161)	(0.0344)	(0.00579)	(0.0117)
	Age 20∼64	468,395.5	355,367.1	23,906,265.6	32,118,619.4	18,055.3 *	33,847.8 **	11,142.8	24,394.7 **	5079.6 *	7504.1 **
		(394,754.5)	(494,263.3)	(18,363,390.9)	(23,602,763.0)	(10,834.7)	(14,409.9)	(7820.2)	(10,519.6)	(2815.5)	(3641.1)
	Age 65+	25,016.5	−313,275.8	3,419,904.9	15,294,947.7	4701.3	32,829.1 *	2324.0	26,433.5 *	2060.2	6202.8
		(275,895.6)	(680,660.4)	(12,815,821.1)	(32,108,038.3)	(7530.0)	(19,424.2)	(5461.6)	(14,246.3)	(1958.7)	(4923.2)
	nonwhite	230,263.4	273,489.6	10,630,147.4	9,161,095.7	4923.2	1312.4	2321.9	−778.7	1782.0	1240.3
		(167,576.7)	(185,699.6)	(7,764,574.9)	(8,691,679.5)	(4529.3)	(5103.8)	(3276.8)	(3720.1)	(1177.9)	(1318.5)
Total	Total property tax rate	41,763.7 ***	54,020.9 ***	1,971,024.1 ***	2,169,578.6 ***	1181.0 ***	1110.4 **	773.4 ***	668.4 **	262.4 ***	269.0 **
		(13,299.3)	(16,316.5)	(617,979.9)	(755,010.6)	(359.9)	(434.8)	(259.4)	(314.2)	(90.31)	(110.2)
	Total sales tax rate	1938.3	−648,766.7	174,297.1	−3,219,002.3	917.9	15,076.6	760.6	14,976.6	196.5	1432.1
		(41,003.5)	(578,221.7)	(1,905,296.6)	(27,176,517.5)	(1120.2)	(16,311.8)	(821.9)	(12,168.4)	(285.0)	(4059.6)
	Per capita income	−2.104 **	−0.214	−105.7 **	−88.02	−0.0536 **	−0.0827 *	−0.0290	−0.0601	−0.0171 ***	−0.0187
		(0.949)	(1.722)	(44.15)	(82.49)	(0.0255)	(0.0498)	(0.0183)	(0.0368)	(0.00658)	(0.0124)
	Age 20∼64	548,828.1	326,142.5	26,038,936.5	28,627,708.5	19,366.0 *	29,789.3 **	11,680.8	21,149.9 **	5631.9 *	6930.7 *
		(416,386.6)	(487,805.3)	(19,374,005.5)	(23,428,703.9)	(11,396.3)	(14,351.9)	(8251.8)	(10,554.4)	(2948.9)	(3579.2)
	Age 65+	76,123.8	−498,073.9	4,420,282.1	4,834,752.9	5320.5	22,895.9	2283.6	19,025.1	2571.2	4476.1
		(293,329.5)	(645,637.1)	(13,627,169.8)	(30,560,091.3)	(7974.4)	(18,510.5)	(5808.0)	(13,747.4)	(2064.3)	(4605.1)
	nonwhite	164,459.0	259,889.4	7,025,557.4	8,012,656.0	1764.1	353.7	−39.73	−1606.3	1251.9	1178.4
		(169,870.9)	(186,028.1)	(7,866,630.3)	(8,707,100.9)	(4553.8)	(5073.2)	(3309.3)	(3723.9)	(1178.8)	(1299.9)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$. Results use the inverse-distance weight matrix M as a robustness check.

reports SDM impacts under the inverse-distance matrix for the baseline specification that includes the total property tax rate and the sales tax rate. The core sales tax finding is robust: the IV-identified effects of the sales tax rate remain negative for several business activity margins. Importantly, the inverse-distance specification clarifies that the net (total) effect can differ from the direct effect when spillovers are present. In particular, establishments-related outcomes continue to display the clearest and most consistently negative IV pattern, indicating that exogenous increases in the sales tax rate are associated with reductions in firm entry/expansion even when spatial exposure is defined by distance rather than adjacency.

Table 11. Robustness: Regressions with Property Tax Rate Components and Sales Tax Rate Across Multiple Outcomes (Inverse-Distance Matrix M).

		Employment		Payroll		Establishments		Estab. (1–9)		Estab. (10–49)
		Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)
Direct	Combined special district rate	−4094.5	12,714.3	−207,274.5	566,623.5	−115.2	632.9 *	−61.97	493.1 *	−38.33	90.96
		(8444.0)	(13,064.8)	(395,035.1)	(612,586.9)	(237.9)	(362.3)	(164.8)	(252.2)	(63.93)	(98.19)
	School tax rate	−13,977.9	−98,138.9 **	−605,565.0	−4,773,274.9 **	−425.6	−4357.2 ***	−230.5	−3047.7 ***	−145.3	−922.8 **
		(16,977.0)	(47,705.6)	(795,945.7)	(2,225,543.2)	(475.3)	(1328.8)	(332.4)	(913.0)	(127.8)	(366.3)
	County tax rate	−1898.8	−20,079.0 *	−132,502.7	−1,058,360.2 **	−96.61	−960.4 ***	−90.95	−704.4 ***	1.274	−176.2 **
		(6065.0)	(11,553.3)	(283,740.1)	(538,396.1)	(171.7)	(322.2)	(117.9)	(220.2)	(46.91)	(89.36)
	Total sales tax rate	−1535.6	−315,929.5 *	−30,090.1	−15,411,358.0 **	−115.9	−14,673.6 ***	−110.1	−10,606.5 ***	10.82	−2809.2 **
		(10,063.3)	(167,544.2)	(471,619.4)	(7,819,751.4)	(281.8)	(4663.7)	(197.1)	(3208.1)	(75.66)	(1284.9)
Indirect	Combined special district rate	78,891.3 **	83,261.9	3,638,306.9 **	3,369,710.7	2188.5 **	2085.7	1627.3 **	1704.0 *	301.7	143.5
		(34,987.4)	(51,449.4)	(1,649,832.4)	(2,440,539.9)	(957.9)	(1408.2)	(696.6)	(1026.8)	(243.9)	(360.0)
	School tax rate	24,600.2	104,359.7	1,130,668.9	7,409,419.6	868.7	5982.8	1092.6	4258.8	−285.4	1282.8
		(80,726.0)	(190,481.0)	(3,812,081.6)	(9,047,628.1)	(2212.1)	(5239.8)	(1601.7)	(3774.3)	(576.6)	(1368.8)
	County tax rate	34,148.3 *	49,877.7	1,727,426.4 *	2,897,481.1 *	1063.2 **	2039.6 **	695.2 *	1308.8 *	250.1 *	535.7 **
		(19,202.0)	(34,508.2)	(906,611.9)	(1,640,892.1)	(530.4)	(956.6)	(379.4)	(683.2)	(138.4)	(251.9)
	Total sales tax rate	11,343.1	221,268.6	127,937.5	19,285,799.1	491.3	15,657.4	284.3	9093.9	255.7	5415.8
		(50,393.8)	(645,647.0)	(2,378,230.2)	(30,680,244.0)	(1382.3)	(17,758.6)	(1001.5)	(12,782.7)	(358.1)	(4647.0)
Total	Combined special district rate	74,796.8 **	95,976.2 *	3,431,032.4 *	3,936,334.2	2073.3 **	2718.6 *	1565.4 **	2197.1 *	263.4	234.5
		(37,243.3)	(57,441.6)	(1,757,322.5)	(2,727,519.7)	(1017.2)	(1571.9)	(743.0)	(1147.2)	(256.4)	(400.7)
	School tax rate	10,622.3	6220.8	525,103.9	2,636,144.7	443.1	1625.6	862.2	1211.1	−430.7	360.1
		(89,237.3)	(189,317.8)	(4,215,376.5)	(9,022,305.4)	(2442.6)	(5201.0)	(1771.8)	(3776.0)	(635.2)	(1344.5)
	County tax rate	32,249.5 *	29,798.7	1,594,923.8 *	1,839,120.9	966.6 *	1079.3	604.2	604.3	251.3 *	359.4
		(18,733.4)	(31,652.4)	(887,260.2)	(1,515,188.8)	(515.2)	(878.1)	(371.1)	(633.4)	(133.1)	(228.2)
	Total sales tax rate	9807.5	−94,660.9	97,847.4	3,874,441.1	375.4	983.8	174.3	−1512.6	266.5	2606.6
		(55,260.3)	(652,148.7)	(2,608,546.5)	(31,082,051.5)	(1513.9)	(17,914.5)	(1099.1)	(12,983.2)	(391.1)	(4648.6)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$. Results use the inverse-distance weight matrix M as a robustness check.

replicates the component decomposition of property taxation under the inverse-distance matrix. The decomposition highlights that results for the property tax bundle can be more sensitive to how spatial interaction is defined. Compared with the contiguity-weight specification, some component estimates become less precise, and the allocation between direct and spillover channels can shift. By contrast, the negative IV sales tax pattern for establishments-related outcomes remains present, supporting the interpretation that sales tax changes have economically meaningful implications for local business activity in Nevada even when accounting for broader regional integration.

Table 12. Robustness: Regressions with Property Tax Rate Components, the Sales Tax Rate, and Interactions with post-2005 dummy (Direct and Indirect Effects, Inverse-Distance Matrix M, without IV).

	Direct Effects					Indirect Effects
	Employ.	Annual Payroll	Estab.	Estab. 1–9	Estab. 10–49	Employ.	Annual Payroll	Estab.	Estab. 1–9	Estab. 10–49
County tax rate	−6118.6	−293,045.7	−169.9	−131.9	−14.06	12,766.2	729,413.6	779.4	609.2	142.7
	(6808.7)	(316,317.2)	(189.7)	(129.8)	(53.63)	(23,328.7)	(1,120,008.1)	(655.3)	(472.3)	(178.5)
Combined special district rate	6411.4	384,103.6	339.6	285.8	27.73	63,720.1	3,188,090.3	1906.1	1509.7	197.9
	(10,042.8)	(469,800.0)	(279.8)	(193.7)	(78.26)	(46,036.4)	(2,221,500.9)	(1288.6)	(937.3)	(345.9)
School tax rate	−18,936.1	−955,056.2	−762.0	−503.7	−186.9	123,456.8	5,742,886.1	3060.1	2324.6	315.7
	(17,394.2)	(816,666.9)	(484.3)	(337.2)	(135.2)	(99,038.8)	(4,770,311.5)	(2763.3)	(2002.2)	(750.5)
Total sales tax rate	16,701.0	921,131.2 *	579.2 *	401.9 *	117.9	−26,459.4	−1,587,363.8	82.14	253.4	52.30
	(11,844.6)	(556,214.4)	(329.9)	(229.8)	(92.19)	(55,255.4)	(2,664,572.0)	(1552.7)	(1129.5)	(420.2)
Per capita income	−0.379	−17.10	−0.00712	−0.00277	−0.00305	−0.858	−56.81	−0.0276	−0.0129	−0.0111 *
	(0.268)	(12.55)	(0.00747)	(0.00518)	(0.00209)	(0.790)	(38.37)	(0.0221)	(0.0159)	(0.00610)
Age 20–64	99,589.4	3,761,603.6	3162.7	1771.2	1019.2	812,294.9 *	31,728,296.6	17,631.6	6409.0	8124.9 **
	(113,068.6)	(5,264,744.5)	(3135.6)	(2158.4)	(888.1)	(487,106.2)	(23,281,326.2)	(13,477.8)	(9605.5)	(3795.9)
Age 65+	48,582.6	1,226,177.3	1461.0	497.8	785.6	99,788.8	1,022,644.4	1059.4	−2527.1	3067.4
	(98,220.6)	(4,580,834.6)	(2731.7)	(1883.7)	(770.8)	(365,226.8)	(17,527,519.6)	(10,175.3)	(7342.6)	(2824.9)
Nonwhite population	54.28	−663,645.5	−1509.1	−1257.6 *	−166.2	217,860.1	11,615,338.1	5827.4	3563.9	1754.5
	(40,119.4)	(1,869,835.1)	(1114.6)	(764.4)	(313.4)	(198,066.7)	(9,538,099.6)	(5521.3)	(3965.2)	(1506.6)
Interactions with post-2005 dummy
County tax rate × post-2005 dummy	10,238.8 **	480,506.2 **	280.7 **	184.2 **	55.07 *	32,307.8	1,506,347.2	523.5	264.0	138.9
	(4164.7)	(196,749.2)	(116.0)	(81.46)	(32.21)	(21,811.4)	(1,052,432.2)	(603.5)	(437.1)	(163.3)
Combined special district rate × post-2005 dummy	−5997.1	−434,707.9	−367.0	−299.4 *	−46.67	56,406.4	2,003,025.8	901.1	446.7	228.4
	(8189.2)	(387,000.2)	(228.0)	(160.5)	(63.18)	(44,804.9)	(2,157,001.2)	(1246.4)	(905.0)	(338.1)
School tax rate × post-2005 dummy	5589.5	198,058.7	88.02	31.84	30.83	18,018.5	310,630.3	−150.3	−201.0	−40.44
	(7340.7)	(346,526.4)	(204.9)	(143.8)	(57.00)	(44,209.1)	(2,129,955.8)	(1234.5)	(895.4)	(335.5)
Total sales tax rate × post-2005 dummy	−16,072.4 ***	−814,423.2 ***	−590.9 ***	−427.4 ***	−93.44 **	37,138.7	2,009,040.3	890.8	454.8	281.7
	(5507.1)	(257,719.3)	(153.2)	(105.8)	(42.93)	(34,078.5)	(1,641,486.6)	(951.5)	(685.4)	(260.8)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$. Results use the inverse-distance weight matrix M.

,

Table 13. Robustness: Regressions with Property Tax Rate Components, the Sales Tax Rate, and Interactions with post-2005 dummy (Total Effects, Inverse-Distance Matrix M, without IV).

	Employment	Annual Payroll	Number of Establishments	Establishment Size: 1–9	Establishment Size: 10–49
County tax rate	6647.5	436,368.0	609.5	477.4	128.6
	(22,937.8)	(1,108,718.5)	(646.1)	(470.0)	(174.6)
Combined special district rate	70,131.4	3,572,193.9	2245.7	1795.5 *	225.6
	(48,796.0)	(2,362,069.9)	(1366.6)	(998.6)	(365.0)
School tax rate	104,520.6	4,787,829.9	2298.0	1820.9	128.8
	(105,245.7)	(5,082,791.2)	(2936.9)	(2136.6)	(794.7)
Total sales tax rate	−9758.5	−666,232.6	661.4	655.3	170.2
	(60,543.3)	(2,924,206.0)	(1700.7)	(1239.9)	(459.5)
Per capita income	−1.237	−73.92	−0.0347	−0.0157	−0.0141 **
	(0.929)	(45.07)	(0.0260)	(0.0187)	(0.00714)
Age 20–64	911,884.3 *	35,489,900.2	20,794.3	8180.2	9144.1 **
	(516,931.6)	(24,746,457.7)	(14,284.1)	(10,212.0)	(4023.2)
Age 65+	148,371.3	2,248,821.7	2520.4	−2029.3	3852.9
	(388,502.3)	(18,697,945.4)	(10,822.8)	(7845.3)	(2998.5)
Nonwhite population	217,914.4	10,951,692.6	4318.3	2306.3	1588.3
	(201,611.2)	(9,755,430.2)	(5618.9)	(4057.0)	(1524.2)
Interactions with post-2005 dummy:
County tax rate	42,546.6 *	1,986,853.4 *	804.2	448.2	194.0
	(24,437.4)	(1,179,895.9)	(677.0)	(490.7)	(183.1)
Combined special district rate	50,409.4	1,568,317.9	534.1	147.3	181.7
	(50,207.1)	(2,418,319.5)	(1396.8)	(1014.8)	(378.6)
School tax rate	23,608.1	508,689.0	−62.28	−169.2	−9.614
	(47,963.7)	(2,315,310.2)	(1340.3)	(974.7)	(363.5)
Total sales tax rate	21,066.3	1,194,617.1	299.9	27.42	188.3
	(35,357.6)	(1,709,100.7)	(987.2)	(714.3)	(269.5)

Notes. Standard errors are in parentheses. ** $p<0.05$, * $p<0.1$. Results use the inverse-distance weight matrix M.

,

Table 14. Robustness (Inverse-Distance Matrix M): IV Regressions with Property Tax Components, Sales Tax Rate, and Interactions with post-2005 dummy (Direct and Indirect Effects).

	Direct Effects					Indirect Effects
	Employ.	Annual Payroll	Estab.	Estab. 1–9	Estab. 10–49	Employ.	Annual Payroll	Estab.	Estab. 1–9	Estab. 10–49
County tax rate
Direct	−14,111.6	−683,515.7	−721.2 **	−519.5 **	−150.8	24,329.4	1,911,287.3	1724.5	1232.6	413.6
	(12,491.8)	(580,074.9)	(344.9)	(236.1)	(97.82)	(39,173.2)	(1,880,200.6)	(1103.6)	(799.3)	(297.8)
Combined special district rate
Direct	13,539.4	645,308.5	814.8 *	631.1 **	137.9	58,903.7	2,110,461.4	1401.7	1246.5	9.092
	(15,312.5)	(717,156.4)	(424.1)	(295.1)	(118.4)	(63,217.8)	(3,035,115.5)	(1773.6)	(1294.7)	(473.3)
School tax rate
Direct	−57,069.3	−2,708,093.0	−3368.4 **	−2339.9 **	−827.5 *	195,191.6	13,060,752.4	8901.5	6209.4	1983.1
	(54,657.9)	(2,543,823.9)	(1510.6)	(1038.2)	(426.8)	(217,512.1)	(10,460,894.6)	(6120.0)	(4444.1)	(1646.3)
Total sales tax rate
Direct	−125,148.0	−5,333,034.5	−9065.8 *	−6434.5 *	−2226.4	205,822.6	24,440,207.9	19,487.3	12,887.7	5758.2
	(198,981.5)	(9,271,169.9)	(5499.2)	(3789.4)	(1551.0)	(740,913.7)	(35,617,598.3)	(20,818.5)	(15,121.6)	(5607.2)
Per capita income
Direct	−0.129	−6.050	0.00988	0.00927	0.00109	−1.274	−104.3	−0.0627	−0.0357	−0.0214 *
	(0.441)	(20.58)	(0.0122)	(0.00843)	(0.00343)	(1.581)	(76.49)	(0.0447)	(0.0324)	(0.0121)
Age 20–64
Direct	−71,180.5	−3,817,651.0	−8443.5	−6462.4	−1808.3	1,078,628.8	62,757,383.4	40,288.0	20,983.8	14,827.9 *
	(263,752.8)	(12,295,280.1)	(7291.4)	(5023.2)	(2060.1)	(1,029,696.0)	(49,608,212.5)	(28,986.7)	(20,949.8)	(7863.2)
Age 65+
Direct	−144,781.4	−7,361,513.6	−11,695.1	−8827.3	−2417.1	416,495.3	35,645,256.9	27,344.4	14,596.6	10,744.2
	(286,834.5)	(13,359,956.3)	(7926.5)	(5456.5)	(2239.0)	(1,032,897.6)	(49,662,489.1)	(29,033.6)	(21,032.2)	(7852.8)
Nonwhite population
Direct	37,721.0	1,298,096.0	1130.3	584.3	493.9	150,743.6	6,994,811.1	938.6	148.1	469.0
	(62,355.9)	(2,890,670.0)	(1714.9)	(1170.2)	(486.5)	(218,933.8)	(10,460,021.0)	(6080.4)	(4370.5)	(1654.5)
Interactions with post-2005 dummy
County tax rate × post-2005 dummy
Direct	9763.5 **	465,389.6 **	250.0 **	162.4 *	47.77	34,485.8	1,720,465.4	703.3	386.6	187.8
	(4246.1)	(200,406.4)	(118.0)	(82.97)	(32.66)	(22,523.1)	(1,087,364.5)	(628.7)	(457.2)	(168.1)
Combined special district rate × post-2005 dummy
Direct	−7094.5	−472,665.7	−439.7 *	−355.7 **	−62.45	49,216.5	1,673,974.5	402.8	63.99	115.1
	(8402.4)	(396,800.5)	(233.6)	(164.9)	(64.50)	(45,855.0)	(2,205,528.4)	(1286.7)	(940.3)	(345.2)
School tax rate × post-2005 dummy
Direct	6773.5	238,694.3	165.2	88.44	48.65	18,514.5	152,208.8	−160.3	−188.3	−56.42
	(7608.1)	(358,682.5)	(211.6)	(148.7)	(58.68)	(45,336.6)	(2,178,404.4)	(1270.0)	(925.1)	(341.6)
Total sales tax rate × post-2005 dummy
Direct	−14,562.9 **	−754,370.2 ***	−488.2 ***	−352.9 ***	−69.51	37,346.4	1,755,750.5	853.0	457.1	251.2
	(5959.9)	(278,742.9)	(164.9)	(114.5)	(46.08)	(36,922.0)	(1,774,727.3)	(1033.8)	(749.0)	(280.0)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$. IV specification; results use inverse-distance weight matrix M.

and

Table 15. Robustness (Inverse-Distance Matrix M): IV Regressions with Property Tax Components, Sales Tax Rate, and Interactions with post-2005 dummy (Total Effects).

	Employment	Annual Payroll	Number of Establishments	Establishment Size: 1–9	Establishment Size: 10–49
County tax rate	10,217.8	1,227,771.6	1003.3	713.1	262.8
	(38,653.9)	(1,868,753.3)	(1094.7)	(800.9)	(292.4)
Combined special district rate	72,443.0	2,755,769.9	2216.5	1877.6	147.0
	(70,064.0)	(3,368,647.3)	(1967.1)	(1438.6)	(523.6)
School tax rate	138,122.3	10,352,659.3	5533.1	3869.6	1155.6
	(221,870.6)	(10,725,797.6)	(6263.0)	(4581.6)	(1671.5)
Total sales tax rate	80,674.6	19,107,173.4	10,421.5	6453.2	3531.7
	(775,903.1)	(37,447,729.7)	(21,849.8)	(15,965.4)	(5850.2)
Per capita income	−1.402	−110.4	−0.0529	−0.0265	−0.0203
	(1.708)	(82.84)	(0.0483)	(0.0352)	(0.0130)
Age 20–64	1,007,448.3	58,939,732.4	31,844.6	14,521.4	13,019.6
	(1,084,205.8)	(52,420,112.3)	(30,582.4)	(22,208.5)	(8266.0)
Age 65+	271,713.9	28,283,743.4	15,649.3	5769.3	8327.1
	(1,079,591.3)	(52,124,044.5)	(30,418.7)	(22,167.3)	(8184.5)
Nonwhite population	188,464.6	8,292,907.1	2068.9	732.5	962.9
	(209,925.2)	(10,098,274.2)	(5831.8)	(4235.4)	(1569.3)
Interactions with post-2005 dummy:
County tax rate × post-2005 dummy	44,249.2 *	2,185,854.9 *	953.4	549.1	235.6
	(25,142.1)	(1,214,554.7)	(702.6)	(511.4)	(187.8)
Combined special district rate × post-2005 dummy	42,122.0	1,201,308.8	−36.94	−291.7	52.62
	(51,530.5)	(2,479,432.7)	(1446.1)	(1057.0)	(387.7)
School tax rate × post-2005 dummy	25,288.0	390,903.0	4.942	−99.83	−7.780
	(49,346.6)	(2,375,286.7)	(1384.1)	(1010.7)	(371.3)
Total sales tax rate × post-2005	22,783.5	1,001,380.3	364.9	104.1	181.7
	(38,670.5)	(1,863,908.3)	(1084.1)	(788.7)	(292.3)

Notes. Standard errors are in parentheses. * $p<0.1$. IV specification; results use inverse-distance weight matrix M.

extend the inverse-distance robustness check to the post-2005 interaction specifications (direct/indirect effects and total effects, respectively). Two results are noteworthy. First, the interaction patterns for the sales tax margin remain economically meaningful, although their precision and the balance between direct and spillover components can change under distance-based exposure. Second, the interaction between the school tax rate and the post-2005 dummy is not uniformly identified: in some specifications it is omitted due to multicollinearity, and where it is identified, estimates are not uniformly significant across outcomes. Accordingly, we interpret these interaction estimates cautiously and emphasize the broader pattern rather than any single coefficient.

Taken together, Table 10, Table 11, Table 12, Table 13, Table 14 and Table 15 indicate that moving from contiguity to inverse-distance weights does not overturn the main conclusions. Instead, it helps distinguish border-local versus broader-market spatial channels: several effects (especially those tied to sales taxation in the IV specifications) remain robust, while some property tax component estimates are more sensitive to the definition of spatial exposure.

Nevada comprises a small number of counties, and economic activity is heavily concentrated in Clark County. To ensure that the baseline results are not mechanically driven by a single large county, we re-estimate the same SDM specifications after excluding Clark County while maintaining the baseline contiguity matrix to preserve the border-local interpretation of spatial interaction.

Table 16. Regressions with Total Property Tax Rate and Sales Tax Rate Across Multiple Outcomes (Contiguity W, Excluding Clark County).

		Employment		Payroll		Establishments		Estab. (1–9)		Estab. (10–49)
		Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)
Direct	Total property tax rate	313.8	594.2	28,229.0	42,080.4	20.77	48.98	−0.277	25.85	14.81	16.88
		(1194.9)	(1348.0)	(49,498.5)	(55,767.7)	(31.68)	(35.29)	(19.16)	(21.24)	(13.60)	(15.24)
	Total sales tax rate	−818.0	−34,133.5	−43,524.0	−1,514,248.8	−56.92	−2258.8 **	−9.269	−1668.3 **	−39.99	−455.7
		(2281.5)	(43,236.3)	(94,099.3)	(1,787,469.6)	(59.42)	(1129.8)	(36.02)	(680.0)	(25.60)	(488.2)
	Per capita income	−0.101 **	−0.0330	−4.113 **	−1.118	−0.00290 **	0.00161	−0.00147 *	0.00194	−0.00104 *	−0.000199
		(0.0485)	(0.101)	(2.005)	(4.168)	(0.00128)	(0.00264)	(0.000775)	(0.00159)	(0.000549)	(0.00114)
	Age 20–64	46,667.1 **	33,056.1	1,783,974.9 *	1,180,809.0	1796.8 ***	888.2	921.4 **	239.0	700.6 ***	528.4
		(22,998.9)	(29,060.0)	(949,332.9)	(1,199,995.5)	(601.4)	(757.0)	(364.8)	(455.6)	(258.7)	(326.9)
	Age 65+	42,803.2 **	11,311.6	1,664,997.6 **	273,581.4	1742.8 ***	−340.3	906.0 ***	−657.7	662.1 ***	265.8
		(20,382.3)	(45,619.2)	(842,325.9)	(1,885,743.1)	(535.2)	(1192.4)	(324.6)	(717.8)	(230.1)	(515.1)
	Nonwhite	20,599.2 *	25,364.4 *	888,523.3 *	1,092,711.2 **	775.5 **	1052.4 ***	407.6 **	603.6 ***	271.7 **	334.2 **
		(12,046.1)	(13,260.4)	(498,406.5)	(548,335.9)	(317.3)	(346.2)	(192.3)	(208.3)	(136.3)	(149.5)
Indirect	Total property tax rate	−4024.9 *	−2854.0	−195,714.0 **	−149,860.7	−175.5 ***	−135.3 **	−94.16 ***	−78.70 **	−61.98 ***	−44.01 *
		(2075.5)	(2433.0)	(85,107.4)	(99,832.5)	(52.76)	(61.56)	(31.98)	(37.04)	(22.86)	(26.62)
	Total sales tax rate	−3153.4	−6538.1	−113,856.8	−165,229.4	−197.1	182.7	−116.7	389.3	−64.64	−149.8
		(4824.0)	(16,257.9)	(196,900.6)	(672,124.4)	(120.3)	(438.7)	(72.39)	(270.3)	(52.57)	(184.4)
	Per capita income	0.117	0.109	3.271	2.831	0.00404	0.00305	0.00208	0.00121	0.00165	0.00155
		(0.103)	(0.105)	(4.224)	(4.286)	(0.00261)	(0.00269)	(0.00157)	(0.00163)	(0.00113)	(0.00115)
	Age 20–64	−225.8	16,034.0	−120,469.8	467,018.6	692.6	946.1	457.1	284.1	298.2	598.5
		(31,996.4)	(40,164.1)	(1,302,119.6)	(1,644,270.2)	(808.2)	(1010.6)	(482.7)	(607.5)	(353.6)	(439.4)
	Age 65+	27,765.5	44,275.7	961,180.7	1,606,644.8	641.6	1205.4	56.15	220.7	461.9 *	746.1 **
		(22,439.6)	(29,717.4)	(915,481.5)	(1,223,210.3)	(577.4)	(764.4)	(340.9)	(462.8)	(251.6)	(332.9)
	Nonwhite	775.2	8519.1	−36,786.2	227,127.7	379.2	408.6	141.4	−36.33	235.4	380.4
		(20,749.3)	(24,552.1)	(848,880.9)	(1,005,828.5)	(526.5)	(621.3)	(316.1)	(372.9)	(229.2)	(269.0)
Total	Total property tax rate	−3711.1 *	−2259.8	−167,485.0 **	−107,780.3	−154.7 ***	−86.34	−94.44 ***	−52.86	−47.17 **	−27.13
		(2081.4)	(2590.3)	(84,990.6)	(105,909.5)	(51.89)	(64.54)	(31.30)	(38.81)	(22.57)	(27.92)
	Total sales tax rate	−3971.4	−40,671.6	−157,380.8	−1,679,478.2	−254.0 *	−2076.1 **	−126.0	−1279.0 **	−104.6	−605.5
		(5950.3)	(40,984.4)	(242,574.8)	(1,676,646.3)	(147.4)	(1032.7)	(88.68)	(622.4)	(64.57)	(445.3)
	Per capita income	0.0164	0.0756	−0.842	1.713	0.00113	0.00467	0.000604	0.00315	0.000609	0.00135
		(0.111)	(0.132)	(4.533)	(5.401)	(0.00277)	(0.00335)	(0.00166)	(0.00202)	(0.00121)	(0.00144)
	Age 20–64	46,441.3	49,090.1	1,663,505.1	1,647,827.6	2489.3 **	1834.3	1378.5 **	523.0	998.8 **	1126.8 **
		(44,237.6)	(51,688.4)	(1,800,271.2)	(2,111,040.8)	(1109.1)	(1290.1)	(665.2)	(774.9)	(484.5)	(560.3)
	Age 65+	70,568.7 **	55,587.3	2,626,178.2 **	1,880,226.2	2384.4 ***	865.2	962.2 **	−436.9	1124.0 ***	1011.9 **
		(31,109.4)	(46,531.8)	(1,265,725.2)	(1,900,515.3)	(783.2)	(1161.3)	(467.1)	(699.7)	(341.3)	(503.8)
	Nonwhite	21,374.3	33,883.5	851,737.1	1,319,838.9	1154.7 **	1461.1 **	548.9	567.3	507.2 **	714.6 **
		(23,278.4)	(27,954.5)	(948,543.7)	(1,142,272.4)	(579.4)	(698.5)	(347.9)	(419.5)	(253.4)	(302.7)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$. Contiguity weight matrix W; Clark County excluded from the estimation sample.

and 

Table 17. Regressions with Property Tax Rate Components and Sales Tax Rate Across Multiple Outcomes (Contiguity W, Excluding Clark County).

		Employment		Payroll		Establishments		Estab. (1–9)		Estab. (10–49)
		Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)	Non-IV	IV (Sales)
Direct	Combined special district rate	203.6	3851.3	30,216.4	198,173.2	7.358	178.1 *	−15.92	100.8 *	13.89	61.47
		(2080.1)	(3711.5)	(85,707.1)	(152,206.1)	(54.06)	(94.96)	(32.43)	(57.00)	(23.17)	(41.00)
	School tax rate	−1601.3	−22,746.7	−98,787.8	−1,079,088.5 *	21.27	−867.2 **	17.31	−521.8 **	5.326	−285.6 *
		(3913.5)	(15,540.4)	(160,813.0)	(639,090.5)	(100.6)	(399.8)	(60.47)	(239.9)	(43.48)	(172.0)
	County tax rate	524.7	−3051.3	49,822.4	−116,558.7	37.11	−112.3	8.689	−80.21 *	20.55	−29.10
		(1573.7)	(2953.4)	(65,009.2)	(121,788.7)	(40.79)	(76.37)	(24.48)	(45.80)	(17.54)	(32.77)
	Total sales tax rate	154.9	−77,390.9	15,357.5	−3,578,134.9	−26.09	−3330.9 **	7.528	−2032.9 **	−28.21	−1088.4 *
		(2621.5)	(56,383.3)	(107,465.5)	(2,318,032.9)	(67.00)	(1449.0)	(40.22)	(869.1)	(29.08)	(624.0)
Indirect	Combined special district rate	−8472.7 **	−5149.2	−388,067.2 **	−240,389.8	−352.2 ***	−260.3 **	−206.7 ***	−175.1 ***	−111.0 **	−62.30
		(3964.4)	(4363.5)	(160,453.0)	(175,587.1)	(98.81)	(107.2)	(59.44)	(64.43)	(43.66)	(47.48)
	School tax rate	−4842.3	−2665.6	−395,274.8	−272,944.7	−114.7	129.3	−35.10	192.2	-49.92	−30.79
		(8572.0)	(10,469.4)	(347,412.3)	(430,740.5)	(210.7)	(266.9)	(125.9)	(159.4)	(94.04)	(115.4)
	County tax rate	−2380.3	675.5	−126,175.9	16,827.5	−87.83	29.60	−32.69	30.16	−42.06	1.556
		(2819.4)	(3387.0)	(114,265.5)	(138,177.1)	(70.07)	(85.64)	(41.90)	(51.45)	(31.06)	(37.20)
	Total sales tax rate	−1864.0	−2030.2	−30,042.4	36,219.0	−217.8	366.3	−133.5	514.6	−71.34	−108.7
		(5630.7)	(20,251.1)	(227,055.0)	(846,961.0)	(138.1)	(557.8)	(82.42)	(333.7)	(61.85)	(231.9)
Total	Combined special district rate	−8269.1 *	−1297.9	−357,850.8 **	−42,216.6	−344.9 ***	−82.20	−222.6 ***	−74.33	−97.10 **	−0.837
		(4338.6)	(6230.1)	(174,515.2)	(250,135.7)	(107.0)	(152.4)	(64.28)	(91.77)	(47.65)	(67.67)
	School tax rate	−6443.6	−25,412.2 *	−494,062.6	−1,352,033.2 **	−93.43	−738.0 **	−17.79	−329.6	−44.59	−316.4 *
		(9899.6)	(15,143.3)	(399,462.7)	(610,995.3)	(240.7)	(371.6)	(143.6)	(223.0)	(108.3)	(165.1)
	County tax rate	−1855.6	−2375.8	−76,353.6	−99,731.1	−50.73	−82.72	−24.01	−50.04	−21.51	−27.54
		(2956.4)	(2968.1)	(119,019.8)	(119,268.4)	(71.95)	(72.56)	(42.92)	(43.66)	(32.38)	(32.20)
	Total sales tax rate	−1709.1	−79,421.1	−14,684.9	−3,541,915.9 *	−243.9	−2964.6 **	−126.0	−1518.3 **	−99.55	−1197.2 **
		(6935.4)	(48,879.6)	(279,067.8)	(1,968,620.8)	(168.9)	(1207.7)	(100.8)	(725.0)	(76.04)	(535.1)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$. County tax rate corresponds to lag_combinedtaxratecol9partb. Estimation uses contiguity weights W and excludes Clark County.

report the baseline and decomposed specifications with Clark County excluded. As expected, standard errors generally increase in this exercise because the sample becomes smaller and less economically concentrated, reducing statistical power. Despite this, the principal sales tax conclusion remains: IV-identified sales tax effects continue to be negative for establishments-related outcomes, particularly for measures capturing smaller establishments. This supports the interpretation that the baseline sales tax findings are not solely driven by the Las Vegas metropolitan area.

For property taxation, the Clark-exclusion exercise is informative in a different way. Several component estimates remain directionally similar, but precision can weaken and the decomposition between direct and spillover channels can change because removing the largest county alters the spatial network of exposure. We therefore emphasize direction and economic plausibility, rather than over-interpreting marginal changes in statistical significance.

Table 18. Regressions with Property Tax Components, Sales Tax Rate, and Interactions with post-2005 dummy (Direct and Indirect Effects; No IV).

	Direct Effects					Indirect Effects
	Employment	Annual Payroll	Number of Establishments	Estab. 1–9	Estab. 10–49	Employment	Annual B	Number of Establishments	Estab. 1–9	Estab. 10–49
County tax rate	1584.7	94,231.0	76.24 *	33.96	34.51 *	−2137.8	−117,925.7	−146.3 *	−89.02 *	−52.52
	(1738.6)	(71,840.2)	(44.46)	(26.17)	(19.69)	(3088.7)	(125,347.7)	(76.58)	(45.70)	(34.64)
Combined special district rate	−191.0	−17,300.0	42.25	27.72	17.09	−9523.1 *	−485,832.9 **	−296.3 **	−110.8	−141.3 **
	(2808.2)	(115,523.0)	(71.44)	(42.17)	(31.65)	(5097.2)	(206,289.9)	(125.3)	(74.75)	(57.06)
School tax rate	−2142.3	−107,483.9	18.40	12.30	3.489	8714.7	119,864.3	208.1	137.7	37.64
	(4291.7)	(176,016.7)	(109.1)	(64.85)	(48.17)	(10,379.9)	(419,118.7)	(255.8)	(153.2)	(116.0)
Total sales tax rate	4648.9	202,004.0	104.5	86.49 *	4.386	249.0	54,968.7	−179.8	−114.4	−71.26
	(3054.3)	(125,181.9)	(77.19)	(45.73)	(34.34)	(5632.4)	(227,429.1)	(137.5)	(82.04)	(62.72)
Per capita income	−0.0835	−3.556	−0.00240 *	−0.00130	−0.000770	0.131	3.762	0.00252	0.000819	0.00152
	(0.0570)	(2.341)	(0.00144)	(0.000856)	(0.000642)	(0.134)	(5.409)	(0.00329)	(0.00196)	(0.00150)
Age 20–64	47,268.5	1,639,285.6	1813.2 **	1047.9 **	651.6 *	−693.3	−662,807.2	443.2	192.2	191.2
	(30,577.9)	(1,252,882.7)	(772.4)	(458.2)	(343.7)	(62,016.4)	(2,500,634.2)	(1522.6)	(909.1)	(694.1)
Age 65+	45,126.6 *	1,769,497.9 *	1768.8 ***	969.4 ***	690.9 **	10,424.2	−50,339.4	−724.0	−1112.2 *	264.8
	(24,591.0)	(1,011,403.8)	(625.2)	(369.8)	(277.1)	(41,244.2)	(1,666,292.2)	(1015.3)	(604.0)	(463.4)
Nonwhite population	27,265.1 *	1,127,736.3 *	897.4 **	437.2 *	355.2 **	35,641.7	1,219,407.1	1466.6 *	997.0 **	316.1
	(15,350.8)	(632,235.1)	(391.3)	(231.5)	(173.1)	(30,325.9)	(1,229,564.7)	(758.3)	(452.1)	(338.3)
Interactions with post-2005 dummy
County tax rate × post-2005 dummy	−302.7	−19,560.4	3.232	7.954	−7.340	539.8	−12,954.3	67.21	76.68 *	−13.13
	(1103.2)	(45,253.4)	(27.95)	(16.53)	(12.40)	(2638.5)	(106,449.3)	(65.48)	(39.67)	(29.38)
Combined special district rate × post-2005 dummy	−842.7	−7376.2	−9.136	−8.993	−10.96	−5675.6	−173,601.9	−280.4 *	−226.1 **	−60.65
	(2090.7)	(85,622.7)	(52.84)	(31.49)	(23.50)	(6462.8)	(261,049.1)	(159.6)	(96.88)	(72.27)
School tax rate × post-2005 dummy	−1029.2	−56,287.6	33.50	33.71	0.473	−8754.6 *	−383,475.2 **	−174.0	−55.36	−93.84 *
	(2142.6)	(87,570.9)	(53.72)	(31.81)	(24.02)	(4716.2)	(190,394.5)	(114.8)	(68.57)	(52.51)
Total sales tax rate × post-2005 dummy	−3002.3 *	−124,166.2 *	−104.0 **	−69.56 **	−20.44	1763.1	85,473.0 *	41.26	14.37	22.22 *
	(1804.1)	(74,237.5)	(46.06)	(27.34)	(20.27)	(1077.5)	(43,804.2)	(27.63)	(16.63)	(11.88)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$. In a few interaction specifications, the optimization routine does not converge for the Estab. (0 and 10–49) outcome. We therefore treat those specific coefficients as descriptive and focus interpretation on outcomes/specifications with stable convergence.

,

Table 19. Regressions with Property Tax Components, Sales Tax Rate, and Interactions with post-2005 dummy (Total Effects; No IV).

	Employment	Annual Payroll	Number of Establishments	Estab. 1–9	Estab. 10–49
County tax rate	−553.1	−23,694.7	−70.05	−55.06	−18.01
	(3013.8)	(121,399.4)	(73.57)	(44.07)	(33.66)
Combined special district rate	−9714.1	−503,132.9 **	−254.1 *	−83.09	−124.2 *
	(5994.7)	(241,930.8)	(146.1)	(87.18)	(67.10)
School tax rate	6572.4	12,380.4	226.5	150.0	41.13
	(11,694.3)	(469,803.7)	(286.2)	(172.3)	(130.4)
Total sales tax rate	4897.9	256,972.7	−75.28	−27.90	−66.88
	(7114.8)	(286,595.0)	(173.1)	(103.5)	(79.32)
Per capita income	0.0473	0.206	0.000127	−0.000477	0.000746
	(0.153)	(6.157)	(0.00373)	(0.00223)	(0.00171)
Age 20–64	46,575.1	976,478.4	2256.4	1240.2	842.9
	(76,436.7)	(3,074,451.8)	(1866.4)	(1116.6)	(854.0)
Age 65+	55,550.8	1,719,158.6	1044.8	−142.8	955.8 *
	(48,904.5)	(1,966,588.1)	(1191.4)	(711.6)	(546.9)
Nonwhite population	62,906.8 **	2,347,143.4 *	2364.0 ***	1434.2 ***	671.3 *
	(32,087.9)	(1,290,553.4)	(791.0)	(475.2)	(356.5)
Interactions with post-2005 dummy
County tax rate × post-2005 dummy	237.1	−32,514.8	70.45	84.63 *	−20.47
	(2957.4)	(118,699.5)	(72.64)	(43.94)	(32.84)
Combined special district rate × post-2005 dummy	−6518.3	−180,978.1	−289.5	−235.1 **	−71.62
	(7250.4)	(291,756.9)	(177.8)	(108.4)	(80.97)
School tax rate × post-2005 dummy	−9783.8	−439,762.8 *	−140.5	−21.65	−93.37
	(6073.9)	(245,014.7)	(147.5)	(88.11)	(67.56)
Total sales tax rate × post-2005 dummy	−1239.1	−38,693.3	−62.75	−55.18 **	1.785
	(1821.9)	(73,356.1)	(44.57)	(26.61)	(20.33)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$. The model reports converged = 0 for the Estab. (10–49) column; interpret that column with caution.

,

Table 20. IV Regressions with Property Tax Rate Components, Sales Tax Rate, and Interactions with post-2005 dummy (Direct and Indirect Effects; Contiguity W).

	Direct Effects					Indirect Effects
	Employment	Annual Payroll	Number of Establishments	Estab. 1–9	Estab. 10–49	Employment	Annual Payroll	Number of Establishments	Estab. 1–9	Estab. 10–49
County tax rate	155.4	−1521.3	19.13	26.28	−12.95	−992.1	−38,358.9	−103.4	−89.82	−11.32
	(3458.2)	(142,789.6)	(88.38)	(51.81)	(38.88)	(3930.4)	(160,571.9)	(99.08)	(58.72)	(43.73)
Combined special district rate	1379.5	81,694.6	110.7	51.28	62.10	−9239.9 *	−454,254.3 **	−305.1 **	−148.3 *	−113.9 *
	(4321.1)	(177,229.5)	(109.5)	(64.54)	(48.24)	(5376.1)	(216,584.7)	(132.6)	(79.33)	(58.99)
School tax rate	−10,573.6	−664,870.0	−322.8	−48.81	−269.0	11,272.0	224,490.3	407.1	353.4 *	31.64
	(18,123.6)	(745,034.3)	(460.8)	(271.2)	(202.8)	(13,847.5)	(562,658.9)	(348.4)	(209.6)	(153.4)
Total sales tax rate	−27,326.0	−1,887,715.0	−1217.4	−209.0	−1000.3	6808.0	358,605.4	314.8	375.5	−52.96
	(66,743.3)	(2,744,338.8)	(1696.6)	(997.9)	(747.3)	(22,348.3)	(937,383.0)	(584.6)	(340.3)	(260.1)
Per capita income	−0.0376	−0.592	−0.000449	−0.000767	0.000622	0.109	2.241	0.00168	0.000797	0.000733
	(0.111)	(4.584)	(0.00283)	(0.00167)	(0.00125)	(0.142)	(5.731)	(0.00352)	(0.00210)	(0.00156)
Age 20–64	15,435.7	−440,723.8	503.7	762.2	−349.4	1385.2	−414,796.2	496.3	48.99	355.5
	(73,076.1)	(2,999,864.1)	(1853.3)	(1091.5)	(816.9)	(62,529.9)	(2,528,305.1)	(1544.0)	(923.6)	(688.3)
Age 65+	8894.1	−599,673.0	271.5	636.5	−448.1	18,838.2	607,612.9	−415.2	−1210.4 *	627.0
	(79,447.6)	(3,267,447.7)	(2020.1)	(1188.0)	(889.7)	(45,371.6)	(1,859,513.8)	(1146.5)	(676.7)	(506.3)
Nonwhite population	32,837.2 *	1,497,541.9 *	1123.5 **	478.4 *	537.3 **	32,506.1	1,133,414.4	1197.4	647.7	371.5
	(19,261.8)	(792,819.5)	(490.3)	(288.0)	(215.8)	(33,537.3)	(1,356,435.8)	(837.6)	(498.1)	(370.9)
Interactions with post-2005 dummy
County tax rate × post-2005 dummy	−400.8	−25,701.9	−0.868	6.609	−10.20	298.6	−33,759.3	64.79	90.18 **	−27.22
	(1120.9)	(45,895.4)	(28.35)	(16.71)	(12.49)	(2749.4)	(110,408.4)	(68.59)	(41.90)	(29.99)
Combined special district rate × post-2005 dummy	−842.7	−8999.5	−8.296	−6.611	−12.62	−4844.1	−119,517.2	−249.6	−225.0 **	−33.86
	(2093.4)	(85,682.7)	(52.83)	(31.39)	(23.33)	(6686.3)	(269,302.3)	(165.7)	(100.9)	(73.48)
School tax rate × post-2005 dummy	−908.3	−48,192.8	38.05	34.41	4.536	−8610.5 *	−364,806.9 *	−179.4	−78.27	−77.95
	(2156.1)	(87,909.3)	(54.10)	(32.00)	(23.89)	(4829.6)	(194,326.3)	(118.2)	(70.97)	(52.56)
Total sales tax rate × post-2005 dummy	−2759.3	−108,538.4	−93.44 *	−66.36 **	−13.01	1707.4	82,094.5 *	37.65	12.00	20.91 *
	(1873.0)	(77,043.2)	(47.74)	(28.16)	(20.94)	(1081.4)	(43,895.1)	(27.72)	(16.69)	(11.79)

Notes. Standard errors are in parentheses. ** $p<0.05$, * $p<0.1$.

and

Table 21. IV Regressions with Property Tax Rate Components, Sales Tax Rate, and Interactions with post-2005 dummy (Total Effects; Contiguity W).

	Employment	Annual Payroll	Number of Establishments	Estab. 1–9	Estab. 10–49
County tax rate	−836.7	−39,880.2	−84.26	−63.53	−24.27
	(3075.8)	(123,573.1)	(75.74)	(45.61)	(33.66)
Combined special district rate	−7860.4	−372,559.6	−194.4	−97.01	−51.76
	(7245.3)	(290,697.1)	(177.7)	(106.7)	(79.16)
School tax rate	698.4	−440,379.8	84.38	304.6	−237.4
	(18,924.5)	(756,992.7)	(466.0)	(283.4)	(206.5)
Total sales tax rate	−20,517.9	−1,529,109.6	−902.6	166.5	−1053.3 *
	(56,406.2)	(2,265,688.9)	(1387.6)	(833.0)	(621.5)
Per capita income	0.0716	1.649	0.00123	0.0000296	0.00135
	(0.162)	(6.483)	(0.00397)	(0.00238)	(0.00177)
Age 20–64	16,820.9	−855,520.0	1000.0	811.2	6.047
	(98,512.0)	(3,961,283.5)	(2419.4)	(1452.5)	(1080.2)
Age 65+	27,732.4	7939.9	−143.7	−573.9	178.8
	(76,031.4)	(3,058,323.0)	(1871.9)	(1122.3)	(834.6)
Nonwhite population	65,343.3 *	2,630,956.3 *	2320.9 ***	1126.0 **	908.8 **
	(35,167.6)	(1,410,181.5)	(867.3)	(520.1)	(385.7)
Interactions with post-2005 dummy
County tax rate × post-2005 dummy	−102.2	−59,461.3	63.93	96.79 **	−37.42
	(3094.1)	(123,523.4)	(76.51)	(46.80)	(33.56)
Combined special district rate × post-2005 dummy	−5686.8	−128,516.8	−257.9	−231.6 **	−46.48
	(7447.0)	(298,702.5)	(183.4)	(112.4)	(81.52)
School tax rate × post-2005 dummy	−9518.8	−412,999.7 *	−141.3	−43.86	−73.41
	(6184.4)	(248,465.6)	(151.1)	(90.76)	(67.20)
Total sales tax rate × post-2005 dummy	−1051.9	−26,443.9	−55.80	−54.36 **	7.905
	(1861.0)	(74,636.9)	(45.77)	(27.47)	(20.35)

Notes. Standard errors are in parentheses. *** $p<0.01$, ** $p<0.05$, * $p<0.1$.

extend the Clark-exclusion analysis to the post-2005 interaction specifications. These models are parameter-rich and thus constitute a stringent stress test in a smaller sample. The interaction patterns for sales taxation remain economically meaningful, but—consistent with the reduced sample size—some estimates are less precisely measured. In addition, the school–tax interaction is not consistently identified in all outcomes, reflecting multicollinearity concerns similar to those observed in the full-sample interaction models. Overall, the Clark-exclusion results reinforce that the main conclusions are not artifacts of a single county while highlighting the natural trade-off between model richness and precision in a small-N spatial setting.

Across both robustness checks—(i) replacing the contiguity matrix with inverse-distance weights and (ii) excluding Clark County under contiguity weights—the evidence supports three broad conclusions. First, sales tax effects in the IV specifications remain consistently negative for establishments-related outcomes, indicating that the core sales tax finding is not driven by a particular spatial-weight choice or by a single large county. Second, several property tax component estimates are more sensitive to how spatial exposure is defined, suggesting that property tax relationships may depend on whether the relevant mechanism is border-local or region-wide. Third, interaction (post-2005) estimates remain informative but should be interpreted cautiously due to multicollinearity in some interaction terms and reduced precision in smaller samples. These robustness exercises strengthen our interpretation that Nevada’s local tax–business relationships operate through multiple spatial channels rather than a single mechanical spillover structure.

6. Discussion and Policy Implications

Because our empirical approach is designed to quantify both own-county and cross-county associations in a spatial setting, the discussion emphasizes policy trade-offs and potential spillovers rather than treating each county as an isolated unit. Throughout, we interpret the findings as conditional associations within Nevada’s institutional environment.

Local governments face a central trade-off when financing services that support economic activity: generating sufficient revenue while limiting distortions to business outcomes. Our results suggest that the economic correlates of revenue instruments can differ meaningfully across tax types. The negative associations between sales tax rates and measures of business activity are consistent with demand-side and mobility channels emphasized in the literature, whereby higher sales taxes may reduce local retail demand and encourage cross-border substitution (Rohlin et al., 2014; Rohlin & Thompson, 2018; Shon, 2017). From a policy perspective, these patterns imply that sales-tax-based revenue increases may partly reallocate activity across nearby jurisdictions rather than solely reducing it within the taxing county.

In contrast, the heterogeneous patterns across property tax components are consistent with a fiscal package interpretation in which property taxes may reflect both a cost channel and a public-input channel. Positive associations for certain components (e.g., county and special-district rates) are consistent with the idea that these revenues may finance more visible or business-relevant local public inputs, which can partially offset tax costs. At the same time, negative associations for other components (e.g., school rates) are consistent with benefits that may be more diffuse or realized over longer horizons, yielding weaker contemporaneous links with business activity. Together, these findings highlight that policymakers should consider not only the overall level of local taxation but also the composition of the tax bundle and the spending programs tied to each revenue source.

A key implication of the spatial framework is that local fiscal decisions can generate inter-county externalities. When consumers and firms are mobile within a regional market, tax changes in one county may affect neighboring counties through redirected spending, establishment location adjustments, and commuting-related activity. Accordingly, policy evaluation should account for both direct effects within the implementing county and indirect (spillover) effects on nearby counties. From a governance standpoint, these spillovers suggest potential gains from coordination or information sharing among neighboring jurisdictions, particularly when local sales taxes are adjusted in competitive border environments.

The policy implications of our study should be interpreted in light of institutional features that are specific to Nevada. First, the county structure is small and economic activity is highly concentrated, particularly in Clark County, which can amplify the role of large metropolitan areas in statewide patterns. Second, Nevada’s local revenue system during our study period relied heavily on property and local sales taxes, making it well suited for analyzing the tax mix within a common statewide legal and administrative setting, but also limiting direct generalization to states with different intergovernmental grant structures, different local tax authority, or a substantially different business tax portfolio. Third, the spatial transmission mechanisms emphasized here may differ in states with more numerous counties, different degrees of urban contiguity, or alternative institutional arrangements governing local sales taxes and special districts.

For these reasons, we view the results as informative about how the composition of the local tax bundle and spatial spillovers are associated with business activity within Nevada’s institutional environment, rather than as a direct forecast of policy effects in other states. Future research that extends the analysis to cross-state settings or alternative institutional regimes would be valuable for assessing the extent to which these patterns generalize.

7. Summary and Conclusions

This paper examines the relationship between local taxation and business activity in Nevada counties from 1999 to 2014, focusing on two primary local revenue instruments: property taxes and local sales taxes. We employ multiple county-level indicators of business activity, including employment, annual payroll, total establishments, and establishments by size (1–9 employees and 10–49 employees). We further disaggregate the property tax rate into the county, school, and special-district components, and we evaluate local sales taxes in relation to voter-approved changes and the stated uses of revenue. Because Nevada’s business tax bundle changed materially after 2014, our conclusions are intentionally framed around the pre-2015 regime and are not intended as direct estimates for the post-2015 policy environment.

Within Nevada’s 1999–2014 institutional regime, we find: (i) a robust negative association between the sales tax rate and multiple measures of business activity, especially after 2005; (ii) heterogeneous associations across property tax components, with county and special-district rates positively associated with business activity while school rates are negatively associated; and (iii) evidence that spatial spillovers matter for interpretation, implying that local tax changes can reallocate activity across nearby counties.

Our contribution to the literature is threefold: First, we shift attention from state-level tax policy to local taxation within a single institutional setting, where counties rely heavily on both property and sales taxes and where cross-county interactions are plausible. Second, rather than treating local taxation as a single lever, we analyze the combined effects of property and sales taxes and explicitly decompose property taxation into salient components (county, school, and special districts), which helps clarify why prior empirical findings may vary across contexts. Third, we incorporate spatial dependence and spillovers by estimating a Spatial Durbin Model and reporting direct and spillover effects, and we complement baseline estimates with an instrumental-variable specification that uses voter registration data to instrument for the sales tax rate.

Our findings suggest a positive association between the total property tax rate and business activity, but disaggregating property taxes reveals important heterogeneity. In particular, the county tax rate and the combined special-district tax rate are positively correlated with changes in business activity, while the school tax rate is negatively associated with these outcomes. The heterogeneous correlations across property tax components are consistent with differences in the perceived and actual incidence of local public spending. County and special-district levies often finance services and projects with relatively immediate links to the local business environment, whereas school-district spending may be viewed as providing longer-horizon or more indirect benefits. Importantly, these patterns may also reflect unobserved heterogeneity in service quality, local economic development strategies, or growth-driven fiscal capacity—factors that can generate positive correlations even absent a direct causal effect. We therefore interpret the estimates as informative associations within the 1999–2014 institutional regime and avoid strong causal claims.

Regarding the sales tax rate, we find a negative association with changes in business activity, and this relationship is particularly strong after 2005. One interpretation is a substitution mechanism: limits on property taxation may have increased reliance on sales taxation, which in turn is more closely linked to demand-side and border-reallocation channels that can reduce local business activity. The post-2005 positive association between the county property tax rate and business activity may reflect the institutional environment created by the abatement policy, which could have altered expectations about the extent of future property tax burdens. At the same time, higher sales tax reliance may have distributional consequences and can raise equity concerns, particularly if it increases the overall regressivity of local revenue systems. These findings reinforce the policy relevance of considering both the composition of the local tax bundle and the spatial transmission of fiscal changes.

From a real-estate finance perspective, the heterogeneous associations across property tax components suggest that the incidence of taxation depends on how revenues translate into locally valued services and infrastructure, which can be reflected in commercial property demand and valuation. The consistently adverse sales tax effects on establishment-based measures under the IV specifications highlight a potential trade-off between short-run revenue generation and local economic base growth, with implications for the volatility of municipal revenue streams.

Several limitations should be acknowledged. First, the analysis is confined to 1999–2014 and therefore reflects a pre-2015 institutional regime in Nevada. Second, although we mitigate spatial dependence using the SDM and address potential endogeneity in sales tax rates through an instrumental-variable approach, endogeneity concerns cannot be fully eliminated for all local fiscal variables; results should therefore be interpreted as conditional associations. Finally, estimated spillovers are necessarily tied to the choice of spatial weights, and external validity beyond Nevada should be assessed with caution.

Building on the insights presented here, future research will seek to extend our analysis by focusing on how specific sectors respond to changes in tax rates. This will allow for a deeper understanding of which industries are most sensitive to changes in local taxation. Additionally, we plan to examine the intended use of new local sales tax revenues, utilizing data from sales tax ballots that specify the purposes for which these funds are earmarked. This will help us understand whether businesses perceive different types of tax expenditures (e.g., for infrastructure, education, or public services) as beneficial or detrimental to their operations. Finally, we aim to expand our dataset to include more recent data, enabling us to capture the effects of more recent tax rate changes and provide a more up-to-date perspective on the evolving relationship between local taxes and business activity.

Beyond within-state extensions, two additional avenues are especially important: First, cross-state comparisons that exploit variation in institutional constraints (e.g., property tax limits, local option authority, and intergovernmental transfers) would help assess external validity and identify which patterns generalize. Second, incorporating alternative local tax instruments and revenue sources—such as gross-receipts-style business taxes, business license fees, or other locally administered charges where relevant—would provide a more complete account of how the broader local revenue system relates to business activity.

Through these future extensions, we aim to provide a more complete account of how local tax composition and earmarked spending relate to business activity while explicitly accounting for institutional changes in the post-2014 period. This evidence would help inform ongoing discussions of local revenue design and potential trade-offs, rather than serving as a direct forecast of policy effects across regimes.