Post a brief(200 Word APA Style) description of the three crisis-response organizations you selected. Then describe the roles of each organization as well as the relationships between and among them. Be specific and provide examples.

349

Problem: Florida’s 1985 Growth Man- agement Act required the state’s coastal communities to include policies for two types of hurricane hazard zones in their comprehensive plans: to direct populations away from coastal high hazard areas (CHHAs) and to maintain evacuation times within larger hurricane vulnerability zones (HVZs). State law requires local governments to initiate measures to implement these policies within one year of state approval of the local plan. Have communities complied with these state mandates?

Purpose: This research aims to determine the extent to which post-plan residential development intensities within hurricane hazard zones conform to the state’s policy mandate and the degree to which success in this regard can be explained by the quality of local plan maps and policies.

Methods: We conducted graphical analysis of development trends, and undertook quasi- experimental analysis of pre- and post-plan residential development inside and outside CHHAs, as well as analyzing correlations between plan quality and post-plan resi- dential development intensity. We also conducted interviews for case studies.

Results and conclusions: We found residential exposure to hurricane flood hazards to have increased substantially in the majority of 74 municipalities and 15 coastal counties in Florida after the state approved local comprehensive plans. Residential development inside CHHAs did not slow after plans were adopted by most of these coastal communities. We found better maps and stronger policies to be correlated with lower post-plan develop-

The Proof of the Planning Is in the Platting

An Evaluation of Florida’s Hurricane Exposure Mitigation Planning Mandate

Robert E. Deyle, Timothy S. Chapin, and Earl J. Baker

O ver the last several decades, some states have taken back land use management authority they traditionally delegated to local govern- ments. Such state activism aims to protect public welfare by minimiz-

ing spillovers from local land use practices on other communities, protecting natural resources and public health, and minimizing the public costs of impru- dent local decisions. Many states have intervened directly, adopting and enforcing state regulations to control the management of solid waste, waste- water, and storm water, and to minimize development impacts on natural and historic resources. More than a dozen states have employed what Burby et al. (1997) call the “comprehensive planning approach” to achieve such goals. These states have defined overall policy goals and employed a variety of statu-

ment intensity, but the policy quality effect, though not the map quality effect, dis- appeared after controlling for pre-plan development intensities. These results may be due in part to vesting of development approved prior to adopting the plans, pre- existing zoning entitlements, and Florida’s 1995 property rights law.

Takeaway for practice: State planning mandates aimed at managing development in critical areas are likely to have only marginal effects because of prior entitlements and the legal and political inertia of existing local plan policies and land development regulations.

Keywords: comprehensive plan imple- mentation, plan quality, state planning mandates, hazard mitigation policies, Florida

Research support: Research support was received from the National Oceanic and Atmospheric Administration’s Office of Sea Grant, the Florida Department of Commu- nity Affairs, and Florida State University’s DeVoe Moore Center.

About the authors: Robert E. Deyle ([email protected]) is a professor of urban and regional planning at Florida State University. His principal research concerns plan implementation in the context of coastal hazards and adaptive response planning to sea level rise. Timothy S. Chapin ([email protected]) is an associate professor of urban and regional planning at Florida State University. His teaching and research interests are in the areas of growth management, comprehensive planning, and downtown development. Earl J. Baker ([email protected]) is an associate professor of geography at Florida State University. His current research focuses on the effects of hurricane experience on future evacuation response.

Journal of the American Planning Association,

Vol. 74, No. 3, Summer 2008

DOI 10.1080/01944360802229612

© American Planning Association, Chicago, IL.

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 349

tory mandates and incentives to foster and enhance the preparation of local comprehensive plans and their imple- mentation through development controls and other growth management measures. Both the direct intervention and the comprehensive planning approaches have focused on critical areas that provide public benefits or present hazards to public welfare, including wetlands, productive coastal waters, wildlife habitat, scenic areas, historic and archeo- logical resources, aquifer recharge areas, prime agricultural soils, floodplains, and other areas exposed to natural hazards (see, e.g., Washington State Department of Community, Trade and Economic Development, 2003).

Where states follow the comprehensive planning approach, achieving state goals depends on two tiers of implementation: (1) state agencies implementing legislated state goals and objectives and (2) local governments im- plementing state agency directives. As Burby et al. (1997) observe, the outcome of state agency implementation depends both on the level of effort expended by the agency and on the style they employ in interacting with local governments. However, even among the most prescriptive of the states that mandate local comprehensive planning (i.e. Florida, Maine, Oregon, Rhode Island, and Vermont), the reach of state oversight does not extend beyond the content of the plans adopted by local governments. Thus, effective local implementation of state goals is determined by the quality of the plan policies adopted by local gov- ernments, how local officials translate those policies into development controls and other growth management measures, and how they enforce those measures.

We count ourselves among the “logical-positivists” (Gilg & Kelly, 1997) who believe that implementing comprehensive plan policies can have a significant impact on the state of the world and, therefore, that it is useful to evaluate the extent to which planning goals have been achieved (Alexander & Faludi, 1989; Burby et al., 1997; Preece, 1990; Talen, 1996a, 1996b). Toward that end, this article addresses both plan conformity (whether plan implementation behaviors or outcomes conform to state goals), and plan quality (whether better quality plans are associated with better outcomes).

We pursue these two topics by examining the effec- tiveness of Florida’s growth management mandates for mitigating exposure to hurricane hazards by limiting development in hazardous areas. Vulnerability to natural hazards is widely recognized as a significant constraint on the long-term sustainability of human settlements and thus an important concern of comprehensive planning (Burby, 1998; Burby, Deyle, Godschalk, & Olshansky, 2000; Burby et al., 1997; Godschalk, Brower, & Beatley, 1989; May et al., 1996; Mileti, 1999). Continued exponential

growth along the United States’ coasts has increased the numbers of people and amount of property vulnerable to the high winds, waves, and storm surge flooding of cata- strophic coastal storms (Burby, 1998; Godschalk, Beatley, Berke, Brower, & Kaiser, 1999; Pielke & Pielke, 1997; The H. John Heinz III Center, 2000).

Growth management strategies such as land develop- ment regulations, capital facilities policies, land acquisition, and taxation and fiscal policies are frequently advocated as nonstructural means communities can use to reduce their vulnerability to natural hazards (Burby, 1998; Burby et al., 2000; Godschalk et al., 1989; May et al., 1996; Mileti, 1999). However, there have been few rigorous analyses of these strategies. While a number of studies have analyzed the quality of the hazard mitigation policies in local comprehensive plans in Florida and other states (Berke & French, 1994; Berke, Roenigk, Kaiser, & Burby, 1996; Brody, 2003; Burby et al. 1997; Deyle & Smith, 1998), no one has analyzed the extent to which better plans are associated with better development management outcomes.

We employ a quasi-experimental research design to assess whether residential development patterns in desig- nated hurricane hazard zones conformed to state directives requiring Florida coastal communities to reduce the vul- nerability of people and property in such areas. In addition, we use multivariate modeling to measure the extent to which the content of local comprehensive plans explained residential density after localities adopted plans to meet the state mandate. We supplement our empirical analyses with case studies from a subsample of coastal communities as well as interviews with state officials who oversaw review and approval of local comprehensive plan amendments during the study period. These qualitative inquiries help to illuminate the stories behind our observations.

Our findings suggest that existing local policies and land use entitlements fundamentally constrain any alterations new state policies can make to development patterns. While some coastal communities have successfully limited devel- opment in hazardous areas, implementation of the state’s mandate has been blunted by state and local officials who perceive political and legal constraints to altering entitle- ments. Our results do demonstrate that development is more likely to conform to state goals when maps of future land use in local comprehensive plans explicitly show critical areas.

We believe our results apply to a broad range of comprehensive planning objectives because the following factors constrain local government conformity with all state initiatives to influence development management of critical areas (Burby et al., 1997; May et al., 1996; Olshansky & Kartez, 1998): (a) weak or absent local political constituen-

350 Journal of the American Planning Association, Summer 2008, Vol. 74, No. 3

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 350

cies for the state goals; (b) competing local issues of greater perceived priority; (c) short-term political and fiscal costs of implementing the state mandate combined with uncertain future benefits; (d) limited local capacity to implement state directives; and (e) pre-existing local development management measures that are perceived as entitlements to specific land uses and densities.

In the following sections, we briefly summarize both plan conformity and plan quality research approaches, and what studies taking these approaches have learned about implementing hazard mitigation policies. We then present an overview of the Florida growth management and coastal planning context, after which we take up each of our plan implementation questions. In the final section, we explore the underlying phenomena that may explain our findings and their implications for future efforts to change local planning and development management through state mandates.

Plan Conformity Research

One important segment of plan conformity research has examined whether the prescription and persuasion states employ when they mandate planning shapes the content of local comprehensive plans (Burby et al., 1997). In a comparative study of five states, Burby and his col- leagues showed that the natural hazards elements of local plans were higher quality in states that coupled strong mandates with formal state review of local plans and the threat of state sanctions where local governments failed to comply (see especially Berke & French, 1994; Burby & Dalton, 1994; and Berke et al., 1996). Related studies have also shown that state oversight and sanction authority result in better plan quality (May & Burby, 1996; May et al., 1996).

Florida appears from these studies to have the most effective mandate, measured by the comparative quality of the natural hazards elements of local comprehensive plans. A study by Deyle and Smith (1998), however, revealed that there was considerable inconsistency in the quality of local plans, and that many fell short of complying with the substance of the state mandate. They attributed this in- complete and variable conformity to the Florida Depart- ment of Community Affairs (DCA) according a relatively low priority to the natural hazard goals of the state growth management statute, and to differing levels of scrutiny of draft plans under different agency administrations.

Researchers examining whether plan outcomes conform to plan policies have focused primarily on behavioral measures, including subsequent zoning or plan amendments

(Johnston, Schwartz, & Klinkner, 1978; Johnston, Schwartz, & Tracy, 1984), subdivision approvals (Alterman & Hill, 1978; Johnston et al., 1978), and development permit issuance (Anderson, 1981; Blacksell & Gilg, 1977, 1981; British Tourism and Recreation Research Unit, 1981; Brody & Highfield, 2005; Brotherton, 1992; Curry, 1992; Gregory, 1971; Keyes, 1986; Moore & Nelson, 1993; Nelson & Moore, 1996). Two studies are noteworthy because the authors analyzed actual land use patterns. Talen (1996a) examined changes in community recreation facilities following adoption of the recreation element of the Pueblo, Colorado, comprehensive plan. More recently, Esnard, Brower, and Bortz (2001) assessed differential development patterns inside and outside of flood hazard zones in Nags Head, North Carolina.

One of the principal challenges in conducting plan and policy implementation research is obtaining sufficient data to employ a research design that can both demonstrate that the plan or policy intervention is associated with a change in behavior or the state of the world, and control for unmeasured variables that may affect behavior or the state of the world independent of the plan or policy being evaluated. The ideal research design is a pretest/post-test design with control groups (Cook & Campbell, 1979; Preece, 1990). However, few studies attain this ideal.

Several studies of the impacts of comprehensive plans have used simple posttest only research designs that exam- ine land use changes or permit issuance behavior after plan adoption within a single community or regulatory frame- work (e.g., Alterman & Hill, 1978; Brody & Highfield, 2005; Johnston et al., 1984; Esnard et al., 2001). Many researchers have employed cross-sectional designs that include control groups to compare development activity inside and outside of regulated areas (e.g., Anderson, 1981; Blacksell & Gilg, 1977, 1981; British Tourism and Recre- ation Research Unit, 1981; Burby & French, 1981; Burby et al., 1988; Nelson & Moore, 1996). However, in these studies it is possible that different development patterns observed inside and outside the regulated areas simply continue practices already occurring, independent of the adoption of the new policies. Other researchers have used pretest/posttest designs within single jurisdictions to assess the impact of plan policies (e.g., Johnston et al., 1978; Talen, 1996a). These studies, however, cannot rule out the effects of other factors that may have led to the observed changes over time.

The exceptions are limited. Three studies of permit issuance in Great Britain employed pretest/post-test control group research designs (Brotherton, 1992; Curry, 1992; Keyes, 1986). More recently, Burby, Salvesen, and Creed (2006) employed such a design to examine how adopting

Deyle et al.: The Proof of the Planning Is in the Platting 351

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 351

a rehabilitation building subcode affected the issuance of rehabilitation permits in jurisdictions inside and outside of New Jersey.

To date, studies of whether plan implementation behavior and outcomes conform with hazard mitigation policies have relied principally on subjective data to assess the effectiveness and impacts of local hazard mitigation policies and programs (i.e., interviews or surveys of local planning officials). Findings have been mixed. In a study of development permits issued in 1978 by a sample of communities participating in the National Flood Insurance Program (NFIP), Burby and French (1981) found com- munities that had already substantially developed their floodplains were more likely to participate in the NFIP, but were also more likely to issue higher proportions of their development permits for construction within 100-year floodplains. Yet, in a comparative study of 10 communities, Burby et al. (1988) found evidence that floodplain manage- ment effectively reduced development in the flood-hazard area, as 7% of each community’s land area lay within the 100-year floodplain on average, while an average of only 2.2% of the building permits issued over a 10-year period were granted for new structures within the floodplain. In contrast, Esnard et al. (2001) found that in 1997, 42% of the developed property parcels in the coastal town of Nags Head, North Carolina, lay within the 100-year flood zones designated under the NFIP.

Plan Quality Research

The second line of plan implementation inquiry has attempted to ascertain whether or not differences in plan or land use regulation content are correlated with plan implementation outcomes. These studies have typically employed multivariate models that control for other possible causal variables. Most, however, have used cross- sectional research designs that do not control for conditions that preceded the intervention.

Burby and Dalton (1994) examined an intermediate step in plan implementation, the adoption of land devel- opment regulations based on plan policies. They found that communities with higher quality plan policies limiting development in hazardous areas were more likely to actually adopt growth management measures for imposing such limitations.

Findings from studies that attempt to explain imple- mentation behavior have not provided consistent and compelling evidence of the efficacy of plan policies. Burby and French (1981) tested bivariate correlations between the percentage of building permits issued within floodplains

and the numbers and stringency of floodplain management regulatory measures employed by flood-prone cities. They found only weak evidence that more comprehensive and stronger regulations were associated with less floodplain development. Brody and Highfield (2005) compared wetland development permits issued by the state of Florida and the U.S. Army Corps of Engineers (USACE) within individual watersheds with future land uses designated in local comprehensive plans. Contrary to their expectations, they found that permit issuance was less likely to conform to future land use designations in watersheds where local comprehensive plan environmental policies were stronger. Yet conformity was higher where plan implementation policies were high quality. In a follow-up multivariate analysis of these data, Brody, Highfield, and Thornton (2006) amended these results to support only the link between implementation policy quality and conformity, reporting that the quality of environmental policies in local comprehensive plans had no significant relationship with whether wetland permits conformed to local land use designations.

Berke et al. (2006) found that the design and structure of a plan may mediate the influence of plan policies on development outcomes. In a study of storm water manage- ment permits issued by district councils in New Zealand, they found that permit conditions were more likely to reflect plan policies where the quality of plan design and structure was higher.

Nelson and French (2002), on the other hand, exam- ined the relationship between plan quality and a measure of physical outcomes. They applied a plan quality scoring method to assess the effects of comprehensive plan seismic safety elements on subsequent earthquake damage in California. Using multivariate models that controlled for the exposure and vulnerability of housing stock to earthquake ground shaking and earthquake intensity, they found that levels of residential damage after the 1994 Northridge earthquake were lower in those communities whose seismic safety elements had higher quality scores.

Taken as a whole, this literature has yielded the rather humble conclusion that “Plans can matter!” (Burby & Dalton, 1994; exclamation point in the original). This research suggests that the substantive quality of plan policies has at best a modest influence on the development out- comes they are designed to shape. The recent analysis by Berke et al. (2006), however, suggests that plan design and structure may also affect plan implementation success. We next describe the Florida state planning mandate context within which our study is set.

352 Journal of the American Planning Association, Summer 2008, Vol. 74, No. 3

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 352

The Florida Growth Management and Coastal Planning Context

Florida’s growth management planning mandate (Florida Growth Management Act, 1985a) offers fertile ground for exploring plan conformity and plan quality implementation questions because it is a highly prescriptive, top-down system in which state laws and regulations set out very explicit requirements governing the scope of local comprehensive plan goals, objectives, and policies (Ben- Zadok, 2005; Chapin, Connerly, & Higgins, 2007; Pelham, 2001). Local governments are required both to adopt local plans that conform to state rules and to implement the policies in those plans through local land development regulations. Research on plan implementation in Florida is facilitated by the state’s review and reporting procedures, which have resulted in a central record-keeping system that allows tracking changes in local plan policies over time.

The first comprehensive plans developed under Florida’s 1985 planning mandate were adopted by local governments and approved by the Florida DCA between 1988 and 1991.1 However, most local governments in the state had comprehensive plans in place prior to the 1985 act, because an earlier statute (Local Government Compre- hensive Planning Act, 1975), required local governments to adopt plans in conformance with it by 1979 (Pelham, 2007). Many of these local governments already also had zoning ordinances in place when they crafted their revised comprehensive plans in the late 1980s and early 1990s.

The 1985 act required local governments to enact land development regulations (LDRs) and to take other initia- tives to implement their plan policies (Florida Growth Management Act, 1985b). Where existing LDRs, including zoning ordinances, were at odds with policies in the new plans, local governments were required to amend the LDRs accordingly (Florida Growth Management Act, 1985c). Thus, implementation of new comprehensive plan policies occurred over several years as LDRs were brought into conformance with the plans following approval by DCA. In addition, most local governments have subse- quently made periodic, piecemeal amendments to their comprehensive plans, as well as completing mandated synoptic plan updates on a seven-year cycle (Florida Growth Management Act, 1985d). These amendments are reviewed by DCA for conformity with the state statute and regulations (Florida Growth Management Act, 1985e).

Regulations adopted by DCA (Requirements for Coastal Management Goals, Objectives, and Policies Rule, 2007), require coastal counties and municipalities to include objectives and policies in their comprehensive plans that limit development in and direct populations

away from “coastal high-hazard areas” (CHHAs) and maintain or reduce evacuation times within “hurricane vulnerability zones” (HVZs). The state statute initially allowed local governments flexibility in defining CHHAs (Florida Growth Management Act, 1985a, §7). A subse- quent amendment to the statute (Law of May 11, 1993) changed the minimum CHHA definition to the evacuation area for a Category 1 hurricane.2 An HVZ is defined by state rule as the area that would be evacuated for a Category 3 hurricane (Definitions Rule, 2007).

In the next section, we examine whether or not residen- tial development that has occurred in coastal communities since local plans following the 1985 state law were approved is consistent with these mandates.

Plan Conformity: Has the World Changed as It Should?

We used county property appraiser data to analyze residential land use change at the parcel level inside and outside of hurricane hazard zones in 89 coastal jurisdictions to assess how well post-plan residential development has conformed to this planning mandate. We addressed three plan outcome questions:

1. How did residential land use change within the hurricane hazard zones of coastal communities in Florida following approval of local comprehensive plans by DCA?

2. How did those land use changes affect the exposure of people and property to hurricane flooding?

3. To what extent did land use changes differ inside and outside the communities’ CHHAs before and after approval of local comprehensive plans?

Three hypotheses inform this research. The first is based on Deyle and Smith’s (1998) finding of significant variation in local governments’ compliance with the state’s mandate to include hazard mitigation policies in local comprehensive plans. That study leads us to expect to find that the exposure of people and property to hurricane flooding increased substantially in some coastal communities after they adopted comprehensive plans, despite the state mandate. Secondly, if the state’s mandate was effective in the aggregate, we would expect to find that residential development intensity decreased in CHHAs after local comprehensive plans were approved. Finally, we would also expect residential growth rates inside CHHAs, where plan policies should limit residential development, to be lower than outside of these zones.

Deyle et al.: The Proof of the Planning Is in the Platting 353

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 353

In the following sections we summarize how we measured our key concepts. For a more detailed account of our methods, see Chapin, Deyle, and Baker (2008).

Measuring Residential Land Use Before and After Plan Adoption

We used county property appraiser tax roll data and digital property parcel polygons to compare the 2002 residential land use within a community’s hurricane hazard zones with that in the same area in the year the local com- prehensive plan was approved by DCA.3 At the time this study was conducted, 2002 was the most recent year for which the necessary data were available.

Communities are required to adopt land development regulations to implement their plan policies within one year of approval of their comprehensive plans by DCA. We therefore defined properties as developed pre-plan if their improvements were built in the year of plan approval or earlier. Parcels with improvements built after the year of plan approval we designated post-plan. This approach accounts for the expected lag in plan implementation after the adoption and approval of the local plan.

Measuring the Exposure of People and Property to Hurricane Flooding

We defined three measures of exposure to hurricane flooding for each jurisdiction: number of residential units, 2002 market value of residential property improvements, and total population.

We counted all parcels for which land use was coded as single family residential, mobile home, condominium, or cooperative as having one residential unit.4 We obtained data on the number of units in multi-family residential structures (e.g., duplexes and apartment complexes), motels, and hotels from the Florida Department of Business and Professional Regulation, which licenses multi-family resi- dential properties in the state. We also counted properties in nonresidential uses if they had homestead exemptions, which Florida grants only to properties occupied by per- manent residents. Thus, our total number of residential units for each municipality was the sum of these three components.

To estimate the approximate market value of property improvements, we used the values assessed by county property appraisers for each parcel, subtracting land value from total just value, with the latter typically being 85% of the assessor’s determination of the true market value of a property. We refer to this remaining value as the just value of improved property.

We defined total population as the sum of tourist and nontourist residents, summing both for each jurisdiction.

We used occupancy rate estimates from regional hurricane evacuation studies to calculate the number of people associated with tourist units (hotels and motels) during the peak season. We applied tract-specific 2000 census occupancy rates to the total of 2002 nontourist residential units in each census tract, and then multiplied by the average household size for the tract to estimate the number of nontourist residents by census tract.

Measuring Land Use Changes Inside and Outside CHHAs

To differentiate residential land use patterns both over time (pre- and post-plan approval) and in space (inside and outside CHHAs), we defined the base year of the pre-plan time period for each jurisdiction so that the post-plan and pre-plan time intervals were of equal length. Thus, for ex- ample, if a community’s comprehensive plan was approved in 1989, the length of the post-plan period would be 13 years (1989–2002). The base year for the corresponding pre-plan period would be 1976. We compared development patterns inside and outside each community’s CHHA because the state plan mandate explicitly calls for limiting development within CHHAs. Once we had determined the above parameters for each jurisdiction, we calculated growth rates inside and outside the CHHA for each exposure measure: number of residential units, just value of improved property, and people.

Sample Selection A total of 35 counties and 158 municipalities in

Florida are required to include specific policies to mitigate hurricane hazards in their comprehensive plans. Our sample includes 15 counties (see Figure 1) and 74 municipalities within those counties. We excluded jurisdictions for which 2002 tax roll or property parcel polygon data were unavail- able. We also excluded jurisdictions with unreliable land use coding, comprehensive plans approved after 1991, extensive coastal redevelopment, or other circumstances that made them atypical, and those that had no vacant land within their CHHAs in 1995. The resulting sample provides good coverage of the geographic and socioeconomic varia- tion among the state’s coastal jurisdictions. The counties we include constitute 36% of the total 2000 population of the 35 coastal counties in Florida and 41% of the area. The 1990–2000 average population growth rate for our sample was 26%, which is nearly identical to the average of 27% for all 35 coastal counties (U.S. Census Bureau, 2001).

354 Journal of the American Planning Association, Summer 2008, Vol. 74, No. 3

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 354

Plan Conformity Results

To discover whether comprehensive plan policies influenced residential development inside hurricane hazard zones, we first look at our sample in the aggregate, and then examine land use change trends and patterns for the individual communities in our sample.

Aggregate Changes in Exposure to Hurricane Flood Risks

As shown in Table 1, despite the state’s hazard miti- gation mandates, there was substantial residential growth within the CHHAs and HVZs of the 89 coastal local governments in our study between when local plans were approved and 2002. We calculated an approximate increase of 153,000 new residential units in both types of hurricane hazard zones combined. If we assume other

Deyle et al.: The Proof of the Planning Is in the Platting 355

Figure 1. Sampled and excluded Florida counties.

Included in sample

Excluded from sample

Noncoastal areas

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 355

coastal counties in the state grew at the same rate, approx- imately 425,000 new residential units would have been built in hurricane hazard zones statewide between the year of plan approval and 2002.5 Based on our method of assigning people to residential units, the associated population increase in both types of hazard zones in the communities in our sample would be more than 345,000 people. The corresponding increase in just value of resi- dential improved property would be nearly $30 billion. Extrapolation to all 35 coastal counties yields estimates of approximately 958,000 new residents and $80 billion in just value of new residential structures in hurricane hazard zones over this time period.

These figures do not suggest significant success in attaining the state’s broad legislative goal of “protect[ing] human life in areas that are subject to destruction by natural disaster” through the implementation of local comprehensive plan policies (Florida Growth Management Act, 1985f). Nonetheless, it is possible that the vulnerability of people and property to hurricane flooding is less than it might otherwise have been in the absence of the state’s hurricane hazard mitigation planning mandate.

Changes in Hurricane Flood Exposure Patterns in Individual Coastal Communities

To further explore this apparent failure in local im- plementation of the state planning mandate, we narrow our focus to examine changes in the temporal and spatial residential development patterns within 58 individual coastal communities in our sample that have land both inside and outside their local CHHAs.6 We do so because of the explicit state mandates to limit development inside CHHAs and direct populations away from those areas. We present here two additional analyses: a comparison of trends in annual development levels, measured as numbers of new residential units inside and outside of CHHAs

between 1960 and 2002, and a two-dimensional comparison of median residential growth rates inside and outside of CHHAs before and after plan approval.

We observed three different patterns in our comparison of annual development trends. First, we found in some cases what appeared to be a rising trend inside the CHHA after plan approval, while the number of new residential units built outside the CHHA began to decline and dropped below the level inside the CHHA (see, e.g., Figure 2). Second, in some cases after plan approval the patterns inside and outside of the CHHA were roughly parallel (see, e.g., Figure 3). Third, we found cases where the trends diverged after plan approval: the numbers of new residential units inside the CHHA declined while the numbers outside increased (see, e.g., Figure 4).

These findings provide additional evidence that the majority of communities did not effectively implement policies to direct residential development away from CHHAs. We classified 23% of the communities as ex- hibiting an increase in new residential units after plan approval that was distinct from the observed pattern outside the CHHA. In these cases it appears that communities have not only failed to direct development away from the CHHA, but they have allowed it to increase relative to areas outside the CHHA. In 57% of our sample, communities exhibit parallel patterns inside and outside the CHHA. In these cases, there is no evidence of differential growth management controls inside and outside the CHHA. We saw evidence in 20% of our sample, however, that com- munities may have implemented policies that resulted in a decline in new residential units inside their CHHAs relative to areas outside the CHHAs.

We turn now to a formal test of whether or not post- plan residential growth rates within CHHAs were signifi- cantly different from those that occurred prior to plan approval in the early 1990s, and whether the pre-plan/ post-plan pattern inside CHHAs differed significantly from that observed within areas outside the CHHA. We employ a Mann-Whitney-Wilcoxon test of differences in medians because none of the variables is normally distributed.7

Table 2 presents the four-way comparison. All else being equal, we would expect slower residential

growth rates within CHHAs after plan approval than before if coastal communities were effectively implementing policies to limit growth within CHHAs. Table 2 shows this to be the case (a drop from 67% to 14%). The Mann- Whitney-Wilcoxon test reveals the difference to be statis- tically significant at better than the 99% confidence level. Further examination of Table 2 shows, however, a similar, statistically significant decline in the median growth rate for numbers of residential units built outside of CHHAs.

356 Journal of the American Planning Association, Summer 2008, Vol. 74, No. 3

Table 1. Aggregate changes in residential hurricane exposure inside CHHAs and HVZs between plan approval year and 2002.

Extrapolation Our sample to entire state

Exposure parameter (n = 89) (n = 193)

Residential units 152,697 425,000 Total population 345,766 958,000 2002 just value of residential

improved property $29 billion $80 billion

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 356

Deyle et al.: The Proof of the Planning Is in the Platting 357

Figure 2. Numbers of residential units built in unincorporated Gulf County by year and location.

0

50

100

150

200

250

300

1960 1965 1970 1975 1980 1985 1990 1995 2000 Year

N u

m b

er o

f U

n it

s

Inside CHHA

Outside CHHA

Year

Figure 3. Numbers of residential units built in the city of Niceville by year and location.

0

25

50

75

100

125

150

175

200

225

1960 1966 1972 1978 1984 1990 1996 2002

Year

N u m

b er

o f

U n it

s

Inside CHHA

Outside CHHA

Year

Comp plan adopted

Comp plan adopted

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 357

This parallel trend both inside and outside of CHHAs may be evidence that the post-plan approval decline within CHHAs was due to broader phenomena such as the overall impacts of comprehensive plan implementation independent of policies directed specifically at CHHAs, or limited supplies of vacant land, reduced rates of population growth, or slower economic growth. It is very likely that the supply of vacant residential land declined over this time period. In addition, the population growth rate in these 15 counties was substantially greater between 1980 and 1990 (45% increase) than between 1990 and 2000 (25%; U.S. Census Bureau, 2001). Thus, the observed declines in numbers of new residential units after plan approval may not reflect the direct impacts of plan implementation.

Table 2 also shows that the median growth rates inside and outside of CHHAs were not significantly different prior to the approval of local comprehensive plans (67% versus 77%). However, the median post-plan growth rate inside CHHAs after plan approval was significantly lower than that outside CHHAs (14% versus 28%). Thus, it appears that some factors above and beyond those affecting the whole jurisdictions may have been at work during the post-plan-approval period. It is possible, for example, that there were greater constraints on the vacant land supply inside the CHHAs than outside, or that other factors may have contributed to these observed differences.

We therefore designed a variable to control for the supply of vacant land at the start of each analysis period: growth density. We defined growth density as the number of new residential units built within an analysis period divided by the acres of vacant land suitable for residential development at the start of the analysis period. We coded parcels as available for residential development if they were designated by the county property appraiser as vacant residential land or unimproved agricultural land with no primary residence. We also assumed that parcels coded as residential in the final year of the pre-plan period or in 2002 were vacant residential land prior to the year in which the current residential structure was built.

Table 3 presents the results of a pre-plan/post-plan comparison of growth densities inside and outside of CHHAs. Here we see a significantly different pattern than in Table 2. Median growth densities inside CHHAs were significantly higher than those outside CHHAs before plan approval. This no doubt reflects the greater development pressure inside CHHAs. However, Table 3 also reveals that this pattern did not change significantly in the post-plan environment when communities should have been reducing development densities inside CHHAs. In fact, the compar- isons show the median growth density inside CHHAs after plan approval to be higher than pre-plan, although the difference just barely fails the significance test at the 90%

358 Journal of the American Planning Association, Summer 2008, Vol. 74, No. 3

Figure 4. Numbers of residential units built in unincorporated Sarasota County by year and location.

0

500

1000

1500

2000

2500

3000

3500

1960 1965 1970 1975 1980 1985 1990 1995 2000

Year

N u

m b

er o

f U

n it

s

Inside CHHA

Outside CHHA

Year

Comp plan adopted

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 358

level. The median post-plan growth density is slightly lower than pre-plan outside CHHAs, but the difference is not statistically significant.

Thus, with a control for the vacant land supply, the story told by simple growth rates is contradicted. It appears that in the aggregate, residential development patterns inside CHHAs after plan approvals were not consistent with the state’s mandate to adopt and implement comprehensive plan policies limiting development in such areas.

Plan Quality: Does Plan Content Make a Difference?

In this section we contribute to the second line of plan implementation research by employing bivariate and multivariate statistical analyses to assess whether or not better local comprehensive plans are associated with more limited residential development of the available vacant land inside CHHAs. The specific hypothesis we test is that communities whose comprehensive plans have the following attributes will exhibit lower post-plan residential growth densities inside their CHHAs than those communities whose plans do not:

1. more restrictive CHHA definitions, 2. maps of the CHHA, and 3. stronger policies limiting vulnerability of populations

inside CHHAs.

Research Design We undertook two levels of statistical analysis to test

the influence of our measures of plan quality on post-plan growth density: (1) correlation analyses using Kendall’s tau-b, and (2) regression analyses based upon the results of the bivariate correlation analyses. In addition to our experimental plan quality variables, we specified a set of control variables for the multivariate analyses.

We employed the bivariate correlation analyses to test the basic relationships between our plan quality variables and post-plan growth density. Given the relatively small number of cases to which the regressions were applied (76), it was important that we be parsimonious with respect to the number of independent variables included in the models. Thus, we also used the correlations to identify the most promising, least-correlated control variables. We utilized the non-parametric Kendall’s tau-b statistic as our correlation measure because our plan quality variables are ordinal and because several of our control variables and our dependent variable exhibit non-normal distributions.8 The ordinal nature of our plan quality variables and concerns about normality also led us to transform the dependent variable from raw growth density values to ranked values.9

Plan Quality Variables. For each of the 76 local jurisdictions in our sample that included land both inside and outside a CHHA, we conducted a content analysis of the coastal element of the comprehensive plan that was approved by DCA in the late 1980s to early 1990s, as well as that which was in effect in 2002.10 Each plan was independently scored by two evaluators who compared scores for each variable and resolved any differences through negotiation.

The content analysis focused on the following three components of the local comprehensive plans: the CHHA definition; mapped representation of the CHHA; and the state’s mandated growth management policies to limit public expenditures within the CHHA, direct populations away from the CHHA, and maintain evacuation times within the HVZ. We scored the CHHA definition and each of the growth management policies on a scale of 0 to 3 as follows:

0 = no definition/policy in evidence; 1 = definition/policy that mirrored the state mandate; 2 = definition/policy that was somewhat more restrictive

than the state mandate; and

Deyle et al.: The Proof of the Planning Is in the Platting 359

Table 2. Median rates of growth in residential units before and after comprehensive plan approval inside and outside of CHHAs (n = 58).

Mann- Before After Whitney- plan plan Wilcoxon

Location approval approval p

Inside CHHA 67% 14% <0.0002 Outside CHHA 77% 28% <0.0002 Mann-Whitney-Wilcoxon p 0.1867 0.0003

Table 3. Median residential growth densities before and after comprehensive plan approval inside and outside of CHHAs (n = 58).

Mann- Before After Whitney- plan plan Wilcoxon

Location approval approval p

Inside CHHA 2.06 2.20 0.1075 Outside CHHA 1.35 1.23 0.2514 Mann-Whitney-Wilcoxon p <0.0002 <0.0002

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 359

3 = definition/policy that was far more restrictive than the state mandate.

For the CHHA map variable, we assigned a score of 0 if the CHHA was not represented on any map in the plan and a score of 1 if it did appear there on a map. We defined a separate dichotomous variable indicating whether the CHHA was depicted on the future land use map (FLUM). In contrast to most states, under Florida’s growth manage- ment law, the FLUM, and accompanying policies in the future land use element of a community’s comprehensive plan legally bound the development rights of property owners.11

Unlike some studies that have investigated the role of plan quality on outcomes resulting from plan implemen- tation over a multi-year study period (Brody & Highfield, 2005; Brody et al., 2006; Nelson & French, 2002), our study investigated the effects of initial, as well as current, plan quality, knowing that plan content can change signifi- cantly over time. We also tested the influence of individual plan components rather than an overall index of plan quality, as other plan implementation studies have done. This allowed us to determine whether specific policies, definitions, and mapping approaches were effective at limiting development within the CHHA. Table 4 presents an overview of these plan quality variables.

Figures 5 and 6 compare the distributions of local plan quality scores for the initial plans approved by DCA and the plans in 2002. Consistent with the findings of Deyle and Smith (1998), these figures demonstrate that initial plan quality was decidedly mixed. However, by 2002, quality had improved for each of the plan content variables, and the number of local governments not meeting the state hazard mitigation requirements had fallen substantially. In 2002, some 10–12 years after initial plan approval, more than 80% surpassed or met the state minimums for the definition requirement and each of the policy mandates.

Control Variables. While the plan quality variables are of primary interest, we also developed a number of variables to control for alternative explanations of observed differences in post-plan growth density (see Table 5). We included two variables to control for the amount of vacant residential land within the CHHA where development might be constrained by state and federal regulatory policies: the percentage of vacant land inside the CHHA in wetlands and the percentage of vacant land inside the CHHA in federal Coastal Barrier Resource System (CBRS) units. We also included two additional variables to control for local planning and fiscal conditions: average annual per capita planning expenditures, which could reflect the ability of local planning efforts to meet state mandates; and the

average percentage of annual local government own-source revenues derived from ad valorem taxes, which could influence the incentive to permit development at higher densities and in areas less suitable for development.

A third set of variables controls for hurricane experience and hurricane response issues. We controlled for recent hurricane histories, both before and after plan adoption, on the assumption that communities with more recent experience with more damaging hurricanes would be more likely to limit residential development within their CHHAs. We also included measures of unmet shelter demand and worst-case evacuation clearance times at the time of com- prehensive plan adoption, theorizing that jurisdictions with greater unmet shelter demand or longer clearance times would be more likely to limit development within hurricane hazard zones.

Our final set of control variables attempts to control for market conditions at the time the original comprehen- sive plan was approved by the state. We used two measures of population growth between 1980 and 1990 as proxies for real estate demand on the assumption that higher demand would stimulate more development pressure inside CHHAs: net county population change and county population growth rate. We also included the pre-plan residential

360 Journal of the American Planning Association, Summer 2008, Vol. 74, No. 3

Table 4. Plan quality variables determined both for the initial plan approved by DCA and for the plan in use in 2002.

Name Description

CHHA definition quality CHHA definition present in the comprehensive plan

CHHA map in comp plan Map depicting the CHHA present in the comprehensive plan

CHHA mapped on FLUM CHHA depicted on the future land use map in the comprehensive plan

Limit public expenditures Quality of policies in the comprehensive in CHHA plan that limit public expenditures

subsidizing development inside the CHHA

Direct populations away Quality of policies in the comprehensive from CHHA plan that direct population

concentrations away from the CHHA and/or limit development inside the CHHA

Maintain evacuation times Quality of policies in the comprehensive plan that maintain or reduce hurricane evacuation times inside the HVZ

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 360

growth density inside the CHHA as a measure of pre-plan development conditions.12

Plan Quality Results Our correlation analyses revealed that none of the

2002 plan quality variables was significantly associated with post-plan residential growth density inside CHHAs. However, as shown in Table 6, three of the original plan quality variables were associated with lower post-plan growth densities: existence of a CHHA map in the com- prehensive plan, mapping the CHHA on the future land use map (FLUM), and the strength of policies directing populations away from the CHHA. Only three of our control variables were significantly correlated with post-plan residential growth density, and, therefore, were included in the regression models: percentage of vacant land in the CHHA covered by wetlands (Kendall’s tau-b = −0.25,

p = 0.00), percentage of county shelter demand remaining unmet at time of plan adoption (Kendall’s tau-b = −0.15, p = 0.10), and pre-plan residential growth density inside the CHHA (Kendall’s tau-b = 0.29, p = 0.00).

Not surprisingly, the existence of a CHHA map in the comprehensive plan was highly correlated with the measure of the depiction of the CHHA on the FLUM. We used the latter variable in the regression models because of its stronger correlation with the dependent variable. As shown in Table 7, single-variable models with this variable (Model 1) and the variable measuring the strength of policies directing populations away from the CHHA (Model 2) regressed on post-plan residential growth den- sity rank yielded results comparable to the Kendall’s tau-b correlations. Those local communities that depicted the CHHA on their FLUMs and those with stronger policies for directing populations away from the CHHA had the

Deyle et al.: The Proof of the Planning Is in the Platting 361

Figure 5. Distributions of local plan definition and policy quality variables as initially approved and in 2002.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Pe rc

en t

of c

om m

u n

it ie

s

Initial approved

plan

CHHA definition quality

Definition and policy quality variables

P er

ce n

t o f

co m

m u

n it

ie s

Limit public expenditures in CHHA

Direct populations away from CHHA

Maintain evacuation times

2002 plan

Initial approved

plan

2002 plan

Initial approved

plan

2002 plan

Initial approved

plan

2002 plan

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Exceeds the mandate Meets the mandate Absent

74-3 06 323128 Deyle P 7/11/08 9:04 AM Page 361

desired planning outcome: lower post-plan growth densities inside their CHHAs.13

When pre-plan growth density inside the CHHA is controlled for, the explanatory power of the models in- creases substantially. The adjusted R 2 values increase from 0.05 and 0.03 to 0.22 and 0.15 (see Models 3 and 4). The CHHA map variable remains significant (Model 3) with the addition of the pre-plan growth density control, but the policy quality variable is no longer significant (Model 4). The other two control variables are also significant when added to Models 1 and 2 (not shown), but their effects are subsumed by the pre-plan growth density rank variable in the fully specified model (Model 5).

Discussion

Our findings present a somewhat mixed picture of the impact Florida’s hazard mitigation planning mandate has

had on residential development in areas prone to hurricane flood damage. Our post-plan residential development indicators (Table 1) reveal that residential units, associated population, and the market value of new residential struc- tures increased very substantially within these hazard zones. When we look more closely, we find trends in post-plan residential development in a majority of communities that suggest a lack of constraint on growth inside CHHAs.

We find that the post-plan median growth rate inside CHHAs was significantly less than both the pre-plan median growth rate inside CHHAs and the post-plan median growth rate outside CHHAs (Table 2). However, when we control for the supply of vacant land suitable for residential development at the outset of each analysis period (Table 3), we find additional evidence that coastal communities have failed, in the aggregate, to limit resi- dential development inside their CHHAs after the state approved their comprehensive plans in the late 1980s and early 1990s.

362 Journal of the American Planning Association, Summer 2008, Vol. 74, No. 3

Figure 6. Distributions of local plan CHHA map quality variables as initially approved and in 2002.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Pe rc

en t

of c

om m

u n

it ie

s

Map quality variables

P er

ce n

t o f

co m

m u

n it

ie s

CHHA mapped on FLUM

Initial approved plan

2002 plan

CHHA map in comprehensive plan

Initial approved plan

2002 plan

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Present Absent

74-3 06 323128 Deyle P 7/11/08 9:05 AM Page 362

Our statistical analyses of the influence of plan quality on residential development (Tables 6 and 7) provide evidence that showing CHHA boundaries on the FLUM in a community’s comprehensive plan is associated with lower post-plan residential growth densities inside CHHAs, independent of previous growth patterns, plan policy quality, or state or federal regulatory constraints on the

supply of vacant residential land. However, the plan quality coefficient is no longer significant when the pre-plan growth density control variable is added to the model. One possible interpretation is that communities that were already limiting development within areas susceptible to hurricane flooding prior to the enactment of the state’s 1985 growth management legislation were more likely to adopt stronger policies for limiting development inside CHHAs in their post-1985 comprehensive plans.

These findings invite questions about why the state’s mandate and its implementation by local governments should have had such marginal effects. Here we suggest some possible explanations based on the circumstances faced by many coastal communities, supplemented by insights gleaned from structured telephone interviews with state officials and with local planning officials in a sample of 12 case study communities chosen for variety in com- posite plan quality scores14 and post-plan growth densities. These interviews focused on determining the underlying reasons for the level of post-plan residential development in each community.

Deyle et al.: The Proof of the Planning Is in the Platting 363

Table 5. Overview of control variables.

Name Description Source

Percentage of CHHA vacant land in wetlands Percentage of vacant land inside the CHHA covered by wetlands Calculated by authors

Percentage of CHHA vacant land in CBRS units Percentage of vacant land inside the CHHA covered by Coastal Calculated by authors Barrier Resource System units

Per capita planning expenditures Average annual per capita planning expenditures Florida Department of Financial Services

Percentage of revenue from property taxes Average annual percentage of community’s own-source revenues Florida Department of derived from ad valorem taxes Financial Services

Pre-plan hurricane history Average intensity of all hurricanes to pass within 65 nautical Calculated by authors miles during the pre-plan period weighted by relative damage and number of years prior to plan adoption of most recent hurricane

Post-plan hurricane history Average intensity of all hurricanes to pass within 65 nautical miles Calculated by authors subsequent to plan adoption weighted by relative damage and average of number of years between plan approval and all hurricanes

Percentage of shelter demand unmet at plan Percentage of county shelter demand remaining unmet at time of Regional Hurricane adoption plan adoption Evacuation Studies

Evacuation time at plan adoption Estimate of worst-case evacuation clearance time at time of comp Regional Hurricane plan adoption Evacuation Studies

1980–1990 county population increase Net county population increase between 1980 and 1990 U.S. Census Bureau

1980–1990 county population growth rate Rate of increase in county population between 1980 and 1990 U.S. Census Bureau

CHHA pre-plan growth density Pre-plan residential growth density inside the CHHA Calculated by authors

Table 6. Kendall’s tau-b correlations of post-plan growth densities with original plan quality variables (n = 76).

Plan quality variables Coefficient Significance

CHHA definition quality −0.04 0.64 CHHA map in comp plan −0.17 0.08c

CHHA mapped on FLUM −0.20 0.03* Limit public expenditures in CHHA −0.06 0.47 Direct populations away from CHHA −0.16 0.07c

Maintain evacuation times −0.10 0.23c

cp < 0.10 *p < 0.05

74-3 06 323128 Deyle P 7/11/08 9:05 AM Page 363

Possible Constraints on Limiting Development Inside CHHAs

A literal interpretation of the state’s mandate to include policies to “limit development in” and “direct populations away from” CHHAs could imply stopping development altogether in such areas or, at the very least, reducing the densities allowed in such areas in the future land use element of the plan and the corresponding zoning districts. Almost one quarter (16) of the 77 communities in our sample that have land inside a CHHA have no land outside a CHHA. For these communities to comply with the state’s require- ment to limit development inside CHHAs and direct populations away from these areas would require them to impose moratoria on their own growth.

Marlene Conaway (personal communication, October 2006), former chief of comprehensive planning in the Division of Community Planning of DCA, told us that the agency position was to prevent increases in allowable densities inside CHHAs when reviewing proposed plan amendments. The state was aware that absolute prohibitions on residential development could raise questions of takings in state and federal courts or trigger the state’s property

rights statute, and recognized that it was clearly infeasible for local governments to buy out all undeveloped properties inside CHHAs. This pragmatic approach to reviewing amendments is consistent with the relatively low priority DCA accorded requiring initial plan policies to conform to hurricane hazard mitigation mandates (Deyle & Smith, 1998). However, former and current directors of the Di- vision of Community Planning, Valerie Hubbard (personal communication, February 2006) and Charles Gauthier (personal communication, December 27, 2007), maintain that this approach has prevented substantial increases in residential densities inside CHHAs.

Hubbard noted that many coastal communities whose maximum allowable zoning densities were higher than the as-built densities when their comprehensive plans were approved under the 1985 act experienced pressure to permit higher density residential development inside their CHHAs. In coastal communities with substantial quantities of vacant land at the time their new comprehensive plans were approved, property owners lobbied for increasing the allowable densities. In spite of the state mandate to direct development away from CHHAs, many local governments were hesitant to down-zone property because of the Bert

364 Journal of the American Planning Association, Summer 2008, Vol. 74, No. 3

Table 7. Regression results for post-plan residential growth density rank (n = 76).

Variables Model 1 Model 2 Model 3 Model 4 Model 5

Percentage of CHHA vacant land in wetlands −0.00 Std. error 0.00 Sig. (2-tail) 0.12

Percentage of shelter demand unmet at plan adoption −11.42 Std. error 8.28 Sig. (2-tail) 0.17

CHHA mapped on FLUM −24.01* −25.32* −25.89* Std. error 11.74 9.97 10.19 Sig. (2-tail) 0.03 0.13 0.01

Direct populations away from CHHA −3.87c −1.64 −0.22 Std. error 2.15 2.14 2.10 Sig. (2-tail) 0.07 0.45 0.92

CHHA pre-plan growth density rank 0.43 ** 0.40 ** 0.35** Std. error 0.11 0.12 0.12 Sig. (2-tail) 0.00 0.00 0.00

Adj. R 2 0.05 0.03 0.22 0.15 0.23 F 0.03 0.08 0.00 0.00 0.00

cp < 0.10 *p < 0.05 **p < 0.01

74-3 06 323128 Deyle P 7/11/08 9:05 AM Page 364

Harris Act (1995), the state’s property rights law (Weaver, 2000). The Bert Harris Act requires local governments to negotiate with and, if necessary, compensate landowners where regulations noticed for adoption or adopted after May 11, 1995, “inordinately burden” private property.

As of 2004, two years past the end of our study period, numerous claims under the act had been settled out of court under a required pre-suit procedure, but few cases had been litigated in the state’s circuit courts of appeal (University of Florida College of Law Conservation Clinic, 2004). Thus, there remained considerable uncertainty about the true legal constraints and potential costs the act imposed on local governments (Weaver, 2000). Three local planning officials we interviewed mentioned this as influencing their communities’ decisions about residential development densities inside the CHHA.

Even where communities attempted to follow both the letter and the spirit of the mandate to mitigate hurricane hazards by reducing allowable densities inside CHHAs on their FLUMs, the state’s vesting policy may have limited their abilities. Under the vesting provisions of the state growth management statute (Florida Growth Management Act, 1985g) once a development “agreement” is issued by a local government, that development is governed by the local laws and policies in force at the time the agreement is issued. Subsequently adopted laws and policies can only be applied to a vested development where “[t]hey are not in conflict with the laws and policies governing the de- velopment agreement and do not prevent development of the land uses, intensities, or densities in the develop- ment agreement,” unless the local government makes an affirmative finding after a public hearing that one of several circumstances applies.

The ramifications of this vesting provision are influ- enced by the structure of the development industry in the state. In many parts of Florida, residential development is undertaken by large real estate firms that build large-scale projects in bursts of activity that respond to market con- ditions. As a result, where a developer holds vested land entitlements on a property, a community can add hundreds or thousands of new residential units within a very short period of time, generating spikes in local development activity that are linked to land use approvals made many years previously.

The state does not maintain a central database of vested developments, nor do most communities have readily available records of such information. Our interviews revealed, however, that vesting was a significant factor in dictating post-plan densities inside the CHHAs of 7 of the 12 case study communities. In an effort to assess the rep- resentativeness of the survey findings, we examined the

residential development trend graphs to look for evidence that vesting may have played a role in determining the residential development patterns that occurred after plan approval. Spikes in the numbers of new residential units that deviate substantially from the general development trend after plan adoption may indicate construction in such vested developments. Figure 7 illustrates such a pattern. Following this approach, we found evidence of possible vesting impacts within the CHHAs of 16 (26%) of the 61 communities in our sample that have land both inside and outside of a CHHA.15 This proportion, which is only half what our interviews suggested, is likely a lower bound for vesting, because in some cases vested develop- ments are built out more slowly, producing a less obvious spike in the development trend graph.

What the Planners Tell Us The case study interviews, as well as correspondence

and interviews with DCA Community Planning Division staff (W. Banning, personal communication, October 2, 2006; M. Conaway, personal communication, October 2006; V. Hubbard, personal communication, February 2006), revealed that most communities had zoning ordi- nances in effect before local adoption and state approval of comprehensive plans pursuant to the 1985 mandate. With few exceptions, the planners we interviewed reported that the densities allowed in those land development regulations were incorporated in the future land use and coastal elements of their post-1985 comprehensive plans. In three cases, Palm Beach Shores, Brevard County, and Hernando County, allowable densities were low to begin with. In several other communities, including Cape Canaveral, Panama City Beach, Venice, and Vero Beach, allowable densities were fairly high at the outset and have remained so.

However, several of the case study communities employed growth management strategies that contributed to lower post-plan growth densities inside their CHHAs. For example, Hernando County engaged in significant efforts to acquire land for conservation purposes, much of which was inside the CHHA, while Pasco County bought land inside the CHHA with the explicit intention of limiting development there. Brevard County down-zoned property within a significant portion of its CHHA because of evacuation clearance time concerns unrelated to specific state planning mandates. DCA staff report that several other communities not included in our sample intentionally reduced allowable densities inside their CHHAs at the time they fashioned their original post-1985 plans (C. Gauthier, personal communication, December 27, 2007).

In contrast, other local governments started with relatively weak coastal elements. The St. Johns County

Deyle et al.: The Proof of the Planning Is in the Platting 365

74-3 06 323128 Deyle P 7/11/08 9:05 AM Page 365

1990 plan lacked a policy explicitly directing population concentrations away from the CHHA, and pre-plan allow- able densities inside the CHHA were as high as 15 units per acre. A plan amendment adopted in 2000 reduced those densities to as little as 8 units per acre; however, the county has only recently begun to revise its zoning ordinance to reflect the lower densities in their amended future land use element. The City of Melbourne has very high plan quality scores and, while the 1988 comprehensive plan reportedly reduced the maximum allowable densities to some degree, they remain high at 10 units per acre.

The story is somewhat more complicated in other places. Growth densities inside Okaloosa County’s CHHA increased substantially between the pre-plan period (1.72 units per acre) and the post-plan period (4.22 units per acre). Much of the county’s CHHA lies along the eastern end of Santa Rosa Island, an area referred to locally as Okaloosa Island, where the land was deeded to the county by the federal government in the 1940s with the stipulation that it be used for public purposes. As a result, until re- cently, all private development in the Okaloosa Island area has been done under long-term leases governed by a set of covenants. These covenants pre-date the county’s 1990 comprehensive plan and were reflected in the maximum

allowable densities contained in the plan. The remaining portion of the county’s CHHA lies east of the City of Destin. In this area, maximum allowable densities range from 4 to 25 units per acre. These are largely consistent with zoning that was in place at the time the plan was adopted, although the county did down-zone some areas in 2000.

Rural, conservative Gulf County’s plan quality is largely the result of a stipulation agreement imposed by DCA in 1992 that mandated stronger plan policies. How- ever, the county’s very low growth densities appear to be a function of relatively little development pressure prior to 2002 and the fact that roughly 90% of the county’s vacant residential land inside the CHHA lies within wetlands.

Conclusions

Taken as a whole, our findings suggest that local implementation of Florida’s planning mandate to limit development within hurricane hazard areas has had limited impact on coastal residential development densities. We find that in the aggregate, residential development inside CHHAs has continued at roughly the same intensity as

366 Journal of the American Planning Association, Summer 2008, Vol. 74, No. 3

Figure 7. Numbers of residential units built in City of Destin by year and location.

0

100

200

300

400

500

600

700

800

1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 Year

N u

m b

er o

f U

n it

s

Inside CHHA

Outside CHHA

Year

Comp plan adopted

74-3 06 323128 Deyle P 7/11/08 9:05 AM Page 366

before the approval of plans adopted under the state’s 1985 growth management legislation. Our findings also indicate that plan quality has had only marginal impacts on devel- opment, with little evidence that higher quality compre- hensive plan policies generated more desirable planning outcomes. The one exception, however, is our finding that showing the critical area boundary, the CHHA, on the future land use map, is associated with lower residential development densities inside the CHHA after plan approval.

Our statistical models and our case studies suggest that in general, higher quality comprehensive plan policies reflected existing practice at the time plans were adopted pursuant to the 1985 state mandate rather than signaling change. Local comprehensive plans generally served to reinforce development patterns that existed at the time of plan preparation rather than altering them. Interviews with state officials suggest that the state continued to give relatively low priority to enforcing the state mandate, as documented by Deyle and Smith (1998), only holding the line at the status quo when reviewing subsequent comprehensive plan amendments. DCA officials maintain, however, that the agency has staved off substantial increases in CHHA density that have been sought through plan amendments. Our analysis of plan quality shows that the content of coastal elements improved over time, but that the higher quality plans in effect in 2002 do not explain the post-plan residential development patterns observed up to that year.

The bald evidence of substantial increases in the num- bers of people and the amount of property at risk within Florida’s hurricane hazard zones suggests a general failure of local and state implementation. However, a more nuanced view seems appropriate when these findings are placed in the context of state vesting policies, the fact that allowable densi- ties had been established in many communities prior to the adoption of the post-1985 comprehensive plans, and the political and legal hurdles to altering these entitlements.

For practitioners, our findings suggest that state mandates designed to alter development practices in critical areas through adopting and implementing local compre- hensive plan policies that depart from existing practice may have marginal effects because of the legal and political inertia of previously adopted plans, policies, and regulations. While the planning process established by the 1985 Florida Growth Management Act called for the development of local comprehensive plans that would shape land develop- ment regulations, in practice the reverse often occurred. In almost all of our case study communities, the allowable densities in the future land use element of the compre- hensive plan simply reflected the densities of the zoning ordinance at the time the plan was created. Most coastal

communities made very few changes to their allowable development densities, even given the state mandate to adopt policies limiting development inside CHHAs.

The news is not all bad for proponents of compre- hensive planning, however. Our case studies and interviews with state officials revealed that some communities sur- mounted these obstacles to reduce development intensity inside their CHHAs through down-zoning and land acquisition. The state’s plan amendment review process also may have contributed to the progressive improvement of plan quality. While many of the initial comprehensive plans had key shortcomings in their hurricane hazard definitions and mitigation policies, we found substantial improvements in these provisions by 2002. Given time for their implementation, these changes are likely to be mani- fest in better conformity with the state’s goals of reducing vulnerability to hurricane flooding. We also found that communities whose comprehensive plans depict the CHHA on the future land use map were more likely to have limited residential development within that critical area. This finding underscores the importance of maps as instruments for displaying and implementing plan policies, particularly those showing critical areas. Clearly, more research is required to investigate and document the utility of maps as a tool for implementing plan policies, but these findings suggest that visual representations of policies were more effective at yielding desirable development outcomes than relying solely on descriptive policies.

For planning scholars, our findings draw attention to the importance of controlling for pre-existing conditions when testing the impacts of planning interventions on plan outcomes. We show that for some plan quality variables, apparent correlations between plan content and post-plan land use trends did not persist after controlling for pre-plan land use trends. We also show that convenient assumptions about plan content stability may be flawed. We found substantial changes in plan quality between initial plan approvals in the late 1980s and early 1990s and the year 2002. We also found that policies in the initial plans pro- vided better explanations of post-plan approval development patterns than did policies in place at the end of the post- plan approval study period. Unraveling changes in plan content is tedious, but in Florida it can be done relatively easily because of the requirement for state review of major comprehensive plan amendments. An interesting question that invites further research is the extent to which the Florida DCA has effectively prevented increases in the vulnerability of coastal populations and property to hurri- cane flooding by challenging proposed comprehensive plan amendments that would be inconsistent with the required policies to direct development away from CHHAs.

Deyle et al.: The Proof of the Planning Is in the Platting 367

74-3 06 323128 Deyle P 7/11/08 9:05 AM Page 367

Whereas Burby and Dalton (1994) concluded that “Plans can matter!” our work, along with that of Berke et al. (2006), suggests some modification to this conclusion, namely that well-designed plans can matter. However, this study also highlights the challenges of inducing changes in longstanding local development management practices, especially where pre-existing policies and regulations are perceived as having conferred development entitlements on property owners.

Acknowledgments This research was conducted under the auspices of the Florida Sea Grant College Program with support from the National Oceanic and Atmospheric Administration, Office of Sea Grant, U.S. Department of Commerce, Grant No. R/C-P-26. Additional funding was provided by the Florida Department of Community Affairs and the DeVoe Moore Center at Florida State University. We are deeply indebted to four cohorts of graduate research assistants whose enthusiasm and diligence made this work possible: Michelle Freeman, Daniel Harris, Alex Joyce- Peickert, Ramona Madhosingh, Nicholas Martinez, Brandie Miklus, John Richardson, Audrey Smith, Preeti Solanki, Lara Mae Webster, and Chris Whittaker. We also wish to acknowledge the helpful suggestions of six anonymous reviewers.

Notes 1. The Department of Community Affairs is the state agency that was tasked with reviewing and certifying that all local comprehensive plans meet the state’s minimum standards. This process began in 1988, although most local comprehensive plans were certified in 1990 and 1991. 2. Hurricane intensity is typically measured using the five-level Saffir- Simpson scale based on wind speed and storm surge height. Category 1 hurricanes are characterized by wind speeds between 74 and 95 miles per hour and storm surge heights of 4 to 5 feet above mean sea level. A category 5 hurricane has wind speeds in excess of 155 mph and storm surge heights in excess of 18 feet. 3. Land use codes reported by county property appraisers in Florida for vacant parcels reflect the appraiser’s judgment of highest and best use. Thus, they are influenced by the zoning of the parcel as well as uses of adjacent parcels. 4. In most Florida counties property appraiser parcel maps include separate “parcel” polygons for each condominium or cooperative unit to provide a geographic reference for their unique ownership. 5. This estimate is based on the proportion of total 2000 population in the 35 coastal counties that is accounted for by the 15 counties in our sample (36%). This extrapolation implicitly assumes that our sample is also proportionately representative of the spatial extent of CHHAs and HVZs throughout the state. 6. A total of 61 jurisdictions from our sample of 89 coastal communities have land both inside and outside a CHHA. We could not construct complete pre-plan data for 3 of those communities, thus the sample analyzed here is 58. An additional 16 communities are completely contained within their CHHAs, that is, they have no land outside their CHHAs. The remaining 12 communities have no land inside a CHHA, but have some land inside an HVZ that would be flooded by a category 3 hurricane.

7. The Mann-Whitney-Wilcoxon test (or simply the Mann-Whitney test or Wilcoxon test) is used with ordinal data, or data that are not normally distributed, to determine whether or not two independent samples can be assumed to come from the same population (Conover, 1980; Garson, 2008). Ranks are assigned to each observation from the combined samples. The statistic tests whether the ranks of the individual observations are randomly distributed between the two samples. Shapiro-Wilk tests for each of the variables are significant at the 99.9% level, indicating that their values are not normally distributed. 8. Shapiro-Wilk test results were significant at the 99.9% level. 9. This transformation captures rank-order statistical relationships comparable to those employed by the Kendall’s tau-b test. Although the rank transformation sacrifices information about the variable, loss of statistical power in regression is not usually a consequence, and the transformation generally ensures a normal distribution of error terms (Hettmansperger, 1978; Iman & Conover, 1979). 10. Copies of original comprehensive plans were obtained from the archives of the Florida Department of Community Affairs. We ascer- tained what the plan content had been in 2002 by comparing current comprehensive plan content with the original plans and, where the current plans contained different language, using the DCA library of comprehensive plan amendments to determine when relevant plan content was amended. 11. A Florida community’s FLUM details the jurisdiction’s desired land use mix and land use intensities over the planning horizon of the com- prehensive plan. All local land development regulations, including zoning, as well as individual development orders, must be consistent with the land uses, densities, and intensities set forth in the comprehensive plan (Florida Growth Management Act, 1985c). While the comprehensive plan does not establish entitlements in the manner of the zoning ordinance (Board of County Commissioners of Brevard County v. Snyder, 1993), the future land use element and FLUM define the parameters that must govern the zoning ordinance and other land development regulations. 12. We recognize that these models are not fully specified, as evidenced by the low adjusted R2 values reported in Table 7. Although we have included constraints upon developable land (acreage in wetlands and CBRS units) and accounted for the amount of developable land in our dependent variable, ideally we would also control for variations in development costs (e.g., land costs and construction costs). However, these data were not readily available for the counties included in the analysis. We also acknowledge that our development demand proxies would not capture shifts in the location of development demand that may have occurred in the early 1990s. 13. All else equal, a community with the CHHA shown on their FLUM would exhibit a growth density rank 24 positions lower than one without it, while a difference of one point on the quality of the local policy directing populations away from the CHHA would lower a community’s rank by roughly four positions. 14. Composite plan quality score was computed by averaging the plan scores for the original comprehensive plan and the 2000 comprehensive plan for each jurisdiction. These scores were derived from an aggregate of the assigned values for four plan items: CHHA definition, CHHA map, a policy for limiting public expenditures in the CHHA, and a policy for directing populations away from the CHHA. 15. We were able to assess development patterns after plan approval for evidence of vesting in the full set of 61 sampled jurisdictions with land inside and outside of a CHHA.

368 Journal of the American Planning Association, Summer 2008, Vol. 74, No. 3

74-3 06 323128 Deyle P 7/11/08 9:05 AM Page 368

References Alexander, E., & Faludi, A. (1989). Planning and plan implementation: Notes on evaluation criteria. Environment and Planning B: Planning and Design, 16 (2), 127–140. Alterman, R., & Hill, M. (1978). Implementation of urban land use plans. Journal of the American Institute of Planners, 44 (3), 274–285. Anderson, M. A. (1981). Planning policies and development control in the Sussex Downs AOB. Town Planning Review, 52 (1), 5–25. Ben-Zadok, E. (2005). Consistency, concurrency, and compact devel- opment: Three faces of growth management implementation in Florida. Urban Studies, 42 (12), 2167–2190. Berke, P. R., Backhurst, M., Day, M., Ericksen, N., Laurian, L., Craw- ford, J., et al. (2006). What makes plan implementation successful? An evaluation of local plans and implementation practices in New Zealand. Environment and Planning B: Planning and Design, 33 (4), 581–600. Berke, P. R., & French, S. P. (1994). The influence of state planning mandates on local plan quality. Journal of Planning Education and Research, 13 (4), 237–250. Berke, P. R., Roenigk, D. J., Kaiser, E. J., & Burby, R. J. (1996). Enhancing plan quality: Evaluating the role of state planning mandates for natural hazard mitigation. Journal of Environmental Planning and Management, 37 (2), 155–169. Bert Harris Act. (1995). Fla. Stat. § 70.001 (2007). Blacksell, M., & Gilg, A. (1977). Planning control in an Area of Outstanding Natural Beauty. Social and Economic Administration, 11 (3), 206–215. Blacksell, M., & Gilg, A. (1981). The countryside: Planning and chal- lenge. London: George Allen and Unwin. Board of County Commissioners of Brevard County v. Snyder, 627 So. 2d 469 (Fla. 1993). British Tourism and Recreation Research Unit. (1981). The economy of rural communities in the national parks of England and Wales. Edinburgh, UK: Author. Brody, S. D. (2003). Are we learning to make better plans? A longitudi- nal analysis of plan quality associated with natural hazards. Journal of Planning Education and Research, 23 (2), 191–201. Brody, S. D., & Highfield, W. E. (2005). Does planning work? Testing the implementation of local environmental planning in Florida. Journal of the American Planning Association, 71 (2), 159–175. Brody S. D., Highfield, W. E., & Thornton S. (2006). Planning at the urban fringe: An examination of the factors influencing nonconforming development patterns in southern Florida. Environment and Planning B: Planning and Design, 33 (1), 75–96. Brotherton, I. (1992). On the quantity and quality of permit applications. Environment and Planning B: Planning and Design, 19 (4), 465–478. Burby, R. J. (Ed.). (1998). Cooperating with nature: Confronting natural hazards with land-use planning for sustainable communities. Washington, DC: Joseph Henry Press. Burby, R. J., Bollens, S. A., Holway, J. M., Kaiser, E. J., Mullan, D., & Schaeffer, J. R. (1988). Cities under water: A comparative evaluation of ten cities’ efforts to manage floodplain use. Boulder: Institute of Behavioral Science, University of Colorado. Burby, R. J., & Dalton, L. C. (1994). Plans can matter! The role of land use plans and state planning mandates in limiting the development of hazardous areas. Public Administration Review, 54 (3), 229–238. Burby, R. J., Deyle, R. E., Godschalk, D. R., & Olshansky, R. B. (2000). Creating hazard resilient communities through land-use planning. Natural Hazards Review, 1 (2), 99–106. Burby, R. J., & French, S. P. (1981). Coping with floods. Journal of the American Planning Association, 47 (3), 289–300.

Burby, R. J., May. P. J., Berke, P. R., Dalton, L. C., French, S. P., & Kaiser, E. J. (1997). Making governments plan. Baltimore, MD: Johns Hopkins University Press. Burby, R. J., Salvesen, D., & Creed, M. (2006). Encouraging residential rehabilitation with building codes: New Jersey’s experience. Journal of the American Planning Association, 72 (2), 183–196. Chapin, T. S., Connerly, C. E., & Higgins, H. T. (Eds.). (2007). Growth management in Florida: Planning for paradise. London: Ashgate Press. Chapin, T. S., Deyle, R. E., & Baker, E. J. (2008). A parcel-based GIS method for evaluating conformance of local land-use planning with a state mandate to reduce exposure to hurricane flooding. Environment and Planning B: Planning and Design, 35 (2), 261–279. Conover, W. J. (1980.) Practical nonparametric statistics (2nd ed.). New York: Wiley. Cook, T. D., & Campbell, D. T. (1979). Quasi-experimental design and analysis issues for field settings. Boston: Houghton Mifflin. Curry, N. (1992). Controlling development in the national parks of England and Wales. Town Planning Review, 63 (2), 107–121. Definitions Rule, Fla. Admin. Code 9J-5.003(57) (2007). Deyle, R. E., & Smith, R. A. (1998). Local government compliance with state planning mandates: The effects of state implementation in Florida. Journal of the American Planning Association 64 (4), 457–469. Esnard, A. M., Brower, D., & Bortz, B. (2001). Coastal hazards and the built environment on barrier islands: A retrospective view of Nags Head in the late 1990s. Coastal Management, 29 (1), 53–72. Florida Growth Management Act. (1985a). 1985 Fla. Laws ch. 55. Florida Growth Management Act. (1985b). Fla. Stat. § 163.3167 (2007). Florida Growth Management Act. (1985c). Fla. Stat. § 163.3194(1)(b) and (3)(a) (2007). Florida Growth Management Act. (1985d). Fla. Stat. § 163.3191 (2007). Florida Growth Management Act. (1985e). Fla. Stat. § 163.3184 (2007). Florida Growth Management Act. (1985f). Fla. Stat. § 163.3178.1 (2007). Florida Growth Management Act. (1985g). Fla. Stat. § 163.3233 (2007). Garson, G. D. (2008). Tests for two independent samples. In Statnotes: Topics in multivariate analysis. Retrieved April 5, 2008, from http:// www2.chass.ncsu.edu/garson/pa765/statnote.htm Gilg, A., & Kelly, M. (1997). The delivery of planning policy in Great Britain: Explaining the implementation gap new evidence from a case study of rural England. Environment and Planning C: Government and Policy, 15 (1), 19–36. Godschalk, D. R., Beatley, T., Berke, P., Brower, D. J., & Kaiser, E. J. (1999). Natural hazard mitigation: Recasting disaster policy and planning. Washington, DC: Island Press. Godschalk, D. R., Brower, D. J., & Beatley, T. (1989). Catastrophic coastal storms. Durham, NC: Duke University Press. Gregory, D. (1971). Green belts and development control: A case study of the West Midlands. Birmingham, UK: Centre for Urban and Regional Studies, University of Birmingham. Hettmansperger, T. P. (1978). Statistical inference based on ranks. Psychometrika, 43 (1), 69–79. Iman, R., & Conover, W. J. (1979). The use of the rank transformation in regression. Technometrics, 21 (4), 499–509. Johnston, R., Schwartz, S., & Klinkner, T. (1978). Successful plan implementation: The growth phasing program of Sacramento County. Journal of the American Institute of Planners, 44 (4), 412–423.

Deyle et al.: The Proof of the Planning Is in the Platting 369

74-3 06 323128 Deyle P 7/11/08 9:05 AM Page 369

Johnston, R., Schwartz, S., & Tracy, S. (1984). Growth phasing and resistance to infill development in Sacramento County. Journal of the American Planning Association, 50 (4), 434–446. Keyes, J. (1986). Controlling residential development in the green belt: A case study. The Planner, 72 (11), 18–20. Law of May 11, 1993. (1993). Fla. Laws ch. 206 § 7. Local Government Comprehensive Planning Act. 1975 Fla. Laws ch. 257 §§ 1–19. May, P. J., & Burby, R. J. (1996). Coercive versus cooperative policies: Comparing intergovernmental mandate performance. Journal of Policy Analysis and Management, 15 (2), 171–201. May, P. J., Burby, R. J., Ericksen, N. J., Handmer, J. W., Dixon, J. E., Michaels, S., & Smith, D. I. (1996). Environmental management and governance: Intergovernmental approaches to hazards and sustainability. New York: Routledge. Mileti, D. S. (1999). Disasters by design: A reassessment of natural hazards in the United States. Washington, DC: Joseph Henry Press. Moore, T., & Nelson, A. (1993). Case study of the effectiveness of coastal growth management in a growth management state. Coastal Management, 21 (2), 197–208. Nelson, A., & French, S. (2002). Plan quality and mitigating damage from natural disasters: A case study of the Northridge earthquake with planning policy considerations. Journal of the American Planning Association, 68 (2), 194–207. Nelson, A., & Moore, T. (1996). Assessing growth management policy implementation. Land Use Policy, 13 (4), 241–259. Olshansky, R. B., & Kartez, J. D. (1998). Managing land use to build resistance. In R. J. Burby (Ed.), Cooperating with nature: Confronting natural hazards with land-use planning for sustainable communities (pp. 167–201). Washington, DC: Joseph Henry Press. Pelham, T. G. (2001). Restructuring Florida’s growth management system: Alternative approaches to plan implementation and concurrency. Florida Journal of Law and Public Policy, 12 (2), 299–310.

Pelham, T. G. (2007). A historical perspective for evaluating Florida’s evolving growth management process. In T. S. Chapin, C. E. Connerly, & H. T. Higgins (Eds.), Growth management in Florida: Planning for paradise, (pp. 7–19). London: Ashgate Press. Pielke, R. A. Jr., & Pielke, R. A. Sr. (1997). Hurricanes: Their nature and impacts on society. New York: Wiley. Preece, R. (1990). Development control studies: Scientific method and policy analysis. Town Planning Review, 61 (1), 59–73. Requirements for Coastal Management Goals, Objectives, and Policies Rule, Fla. Admin. Code 9J-5.012(3)(b)(5) and (6) (2007). Talen, E. (1996a). After the plans: Methods to evaluate the imple- mentation success of plans. Journal of Planning Education and Research, 16 (2), 79–91. Talen, E. (1996b). Do plans get implemented? A review of evaluation in planning. Journal of Planning Literature, 10 (3), 248–259. The H. John Heinz III Center. (2000). The hidden costs of coastal hazards: Implications for risk management and mitigation. Washington, DC: Island Press. United States Census Bureau. (2001). Census 2000 redistricting data (P.L. 94-171): Summary file and 1990 Census. Washington, DC: Author. Retrieved November 17, 2007, from http://quickfacts.census .gov/qfd/states/12000.html University of Florida College of Law Conservation Clinic. (2004). Implementation of the model land development code for Florida springs protection. Gainesville, FL: Author. Washington State Department of Community, Trade and Economic Development. (2003). Critical areas assistance handbook: Protecting critical areas within the framework of the Washington Growth Management Act. Olympia, WA: Author. Weaver, R. L. (2000). Bert Harris Act. Reder’s Digest. Retrieved November 17, 2007, from http://www.redersdigest.com/index.cfm?nav =articles&articleid=6

370 Journal of the American Planning Association, Summer 2008, Vol. 74, No. 3

74-3 06 323128 Deyle P 7/11/08 9:05 AM Page 370