RURAL AMENITY VALUES AND LENGTH

Size: px
Start display at page:

Download "RURAL AMENITY VALUES AND LENGTH"

Transcription

1 RURAL AMENITY VALUES AND LENGTH OF RESIDENCY ROBERT J. JOHNSTON,STEPHEN K. SWALLOW, TIMOTHY J. TYRRELL, ANDDANA MARIE BAUER New residents of rural communities are often assumed to have preferences for development and conservation that differ from those of longer-term residents. However, the literature offers little to quantify presumed preference heterogeneity. This article assesses whether stated preferences differ according to length of residency. Results are based on a conjoint (choice experiment) survey of Rhode Island rural residents. Heterogeneity according to length of town residency is modeled using dummy variables, multiplicative interactions, and Lagrangian interpolation polynomials. Results are compared across the three models, and identify a range of attributes for which willingness to pay depends on length of residency. Key words: contingent valuation, Lagrangian interpolation polynomial, length of residency, rural amenity values. New residents of rural communities are often assumed to have preferences for the development and conservation of rural lands that differ from those of longer-term residents (Kelsey, Spain, Myers). Newer residents are often assumed to have relatively stronger preferences for aesthetic and recreational attributes, and to be more willing to accept associated development restrictions (Dubbink, Healy and Short, Spain). Newer residents may also be willing to incur greater costs of living in return for expanded community services, and may be more (or less) willing to accept wildlife nuisance associated with rural lands (Kelsey, Spain, Zin and Aldelt). Despite the ubiquity of assumptions regarding preference heterogeneity and residence duration, the literature offers little to quantify presumed preference shifts. While the literature addresses heterogeneity in stated preferences for resource policies (e.g., Swallow Robert J. Johnston is associate director of Connecticut Sea Grant and assistant professor, Department of Agricultural and Resource Economics, University of Connecticut; Stephen K. Swallow is professor, Department of Environmental and Natural Resource Economics, University of Rhode Island; Timothy J. Tyrrell is professor, Department of Environmental and Natural Resource Economics, University of Rhode Island; and Dana Marie Bauer is graduate research assistant, Department of Environmental and Natural Resource Economics, University of Rhode Island. This research was funded by the USDA Fund for Rural America No and the Rhode Island Agricultural Experiment Station No Opinions belong solely to the authors and do not imply endorsement by the funding agencies. et al., Loomis) typically as related to demographic indicators few works address preference heterogeneity related to length of residency. Those few works that include residence duration as a determinant of rural preferences (e.g., McLeod, Woirhaye, and Menkhaus; Inman, McLeod, and Menkhaus) focus on broad aspects of rural policy, rather than on detailed land use or amenity trade-offs. While geography, rural sociology, and planning research addresses differences between new and more established residents, these works focus primarily on qualitative findings (Salamon and Tornatore, Dubbink), or on general attitudes toward development and conservation (Spain, Theodori and Luloff, Pendall). Quantification of potential preference heterogeneity associated with length of rural residency may improve understanding of the ways in which development and conservation trade-offs influence the welfare of different groups. To this end, this article examines whether stated preferences for development and conservation differ according to length of residency in selected Rhode Island rural communities. Results are based on preferences estimated from a conjoint survey of rural residents, addressing alternative proposals to develop rural lands for residential purposes. Preference heterogeneity is incorporated using three approaches that allow regression coefficients to shift as a function of length of residency. The first approach uses dummy Amer. J. Agr. Econ. 85(4) (November 2003): Copyright 2003 American Agricultural Economics Association

2 Johnston et al. Rural Amenity Values 1001 variables to allow systematic, discrete shifts in slope and intercept coefficients, based on discrete residence time categories. The second approach appends multiplicative interaction terms to the preference function, modeling the influence of model attributes as a linear function of residence duration. The third approach applies Lagrangian interpolation polynomials (LIPs) to model the influence of model attributes as a higher-order polynomial function of residence time (Tyrrell). 1 The Contingent Choice Model A Random Utility Model with Homogeneous Preferences We begin with a random utility model of rural development and conservation trade-offs, in which preferences are assumed to be homogeneous. To model a respondent s choice, we define a utility function that includes attributes of a rural development or conservation plan and the net cost of the plan to the respondent (Hanemann, McConnell): (1) U( ) = U(X c, Y F c ) = v(x c, Y F c ) + ε c where X c is a vector of variables describing attributes of development or conservation plan c; Y is disposable income of the respondent; F c is the change in mandatory taxes paid by the respondent under plan c; v( ) is a function representing the empirically measurable component of utility; ε c is econometric error. If one compares Plan A (c = A) to Plan B(c = B), the change in utility (du) may be modeled as (2) du = U(X A, Y F A ) U(X B, Y F B ) = [v(x A, Y F A ) v(x B, Y F B )] [ε B ε A ] = dv. The model assumes a respondent assesses the difference between utility under the two plans and indicates the sign of du by either choosing Plan A (du > 0) or Plan B (du < 0). Although the literature offers no firm guidance regarding the choice of specific functional 1 Another approach to modeling heterogeneous preferences is the mixed logit or random parameters model (Train). forms for dv, in practice linear forms are often used. Hence, (3) dv = v(x A, F A ) v(x B, F B ) = x (X A X B ) + f (F B F A ) where x is a conforming vector of coefficients associated with the vector of attribute differences (X A X B ) and f as a scalar coefficient associated with the tax difference (F B F A ). The parameter vector x may be interpreted as the marginal utility of various development or conservation attributes of a development plan, while f represents the marginal utility of income. 2 Alternative Approaches to Modeling Preference Heterogeneity A heterogeneous preferences model allows parameters (marginal utility) to vary across groups or individuals (Swallow et al.). The socioeconomic attribute for which preference heterogeneity is modeled is length of residency in a rural community. Although we discuss heterogeneity solely in terms of this variable, identical models apply to any demographic attribute that may be characterized as a continuous variable (e.g., age). The Dummy Variable Approach Perhaps the most common approach to modeling preference heterogeneity is to allow systematic but discrete variations in slope and intercept coefficients using dummy variables. One might, for example, define a dummy variable to distinguish those with less than ten years of town residency from longer-term residents. This would allow different slope and intercept parameters for the two groups, but would impose constant preferences within each group. Formally, this approach redefines dv in (3) to provide a separate utility estimate for each respondent group. In the simplest two-category case, we define a dummy variable D t to equal one for respondents whose residency is t years or more, and to equal zero otherwise. Accordingly, 2 The absence of income (Y) from (3) reflects the fact that disposable income is assumed unaffected by rural development, aside from direct deductions associated with the cost of each plan, and hence subtracts out of the linear model for dv.

3 1002 November 2003 Amer. J. Agr. Econ. (4) dv = t x (1 D t )(X A X B ) + t f (1 D t )(F B F A ) + >t x D t (X A X B ) + >t f D t (F B F A ) where ( t x, t f ) represents marginal utility parameters for respondents whose residence time is less than or equal to t, and ( >t x, >t f ) represents these marginal utilities for respondents with residence times greater than t. The Multiplicative Interaction Approach This approach appends the linear specification of dv (3) with a vector of multiplicative interactions such that (5) dv = x (X A X B ) + f (F B F A ) + x (X A X B )t + f (F B F A )t where t is the length of residency of the respondent. Here, x and f represent adjustments to the marginal utilities x and f, respectively, associated with the length of residency. The interaction model eliminates the discrete preference shifts characteristic of dummy variable models. While (5) allows marginal utility to vary as a continuous function of t, the effect is linear. The Lagrangian Interpolation Polynomial Approach To relax the linear relationship between the marginal utility of plan attributes and residence duration, one may specify the marginal effect of each attribute as a polynomial or otherwise nonlinear function of residence time. A common means to accomplish this is to append the multiplicative interaction model (5) with higher-order interactions. For example, one might specify an interaction model in which attributes interact with both t and t 2, (6) dv = x (X A X B ) + f (F B F A ) + x (X A X B )t + f (F B F A )t + xx (X A X B )t 2 + ff (F B F A )t 2. Here, xx and ff represent adjustments to the marginal utility associated with t 2. A potential disadvantage of this approach is that model output may be somewhat cumbersome in terms of interpretation and hypothesis testing. For example, marginal utilities are not directly illustrated by estimated parameters; additional calculations are required. Tyrrell suggests an alternative approach, which, while statistically equivalent to (6), offers practical advantages. This approach uses Lagrangian interpolation polynomials (LIPs) to shift regression coefficients. Unlike (6), estimated LIP model results directly reveal the value of parameter estimates for individuals at length of residency reference points chosen by the researcher. This can facilitate both model interpretation and statistical tests designed to assess differences in parameter estimates across reference points; equivalent tests using (6) would generate identical results but would be more cumbersome. In addition, direct access to the value of the polynomial at any reference point allows one to more easily constrain the polynomial function. To illustrate the LIP approach, we first specify the effect of independent variable g as a polynomial function of residence duration, (7) P g (t) = g0+ g1 t + g2 t gn t n where g indexes an independent variable corresponding to one of the columns of matrix (X A X B, F B F A ), P g (t) is the value of the polynomial corresponding to variable g, and n represents the degree of the polynomial chosen by the researcher. The gi are unknown parameters determining the value of the polynomial corresponding to attribute g. We use the polynomials in (7) to estimate an empirical version of model (3) that smoothly modifies the basic utility function for individuals whose residence time is t. Thus, the polynomial-based model causes the analyst to index dv by t, such that (8) dv t = P (t)(x A X B ) + P f (t)(f B F A ) where (t) is a row vector whose elements are defined P by (7), conforming to matrix (X A X B ). In comparing this with the systematically varying slopes approach represented by (4), one sees that model (4) is a special case of model (8) where the polynomial parameters are replaced by constant parameters such that (9) [P (t), P f (t)] [ t x [ >t x, t ] f for t t ] for t > t., >t f If residence time is used to define more than two residency groups in the systematically

4 Johnston et al. Rural Amenity Values 1003 varying slopes model, then (4) (9) would be modified accordingly. The function P g (t) takes values gi at n + 1 unique reference points r i, where i (0, 1, 2,..., n) identifies these reference points. The degree of the estimable polynomial, n, depends on the number of reference points (n + 1) chosen by the researcher. These reference points r i represent residence times t = r i at which coefficients gi will be estimated. That is, P g (r i ) = gi. Reference residence times are chosen to aid in policy analysis or assessment of model implications. It is important to note that once the number of reference points is chosen, the choice of particular reference residence points is a matter of convenience; it does not affect the statistical fit or performance of the model. The choice of reference points simply influences the particular points along the length of residency continuum at which the model will provide coefficient estimates. While the underlying polynomial function is influenced by the number of anchor points chosen by the researcher, it is not altered by the choice of particular reference points. This framework implies a system of equations (10) R n+1 g = g where n+1 is a square (n + 1, n + 1) matrix with rows R [1, r i, r 2 i,..., rn i ] corresponding to the n + 1 reference points; g is a column vector [ g0, g1, g2,..., gn ] ; and g is a column vector consisting of elements { gi } for all i {0, 1, 2,..., n}. Combining (7) and (10) allows a restatement of unknown parameters g (Tyrrell, Shchigolev): (11) g = R 1 n+1 g. By defining T(t) as the vector [1, t, t 2,..., t n ], the polynomial functions then become: (12) P g (t) = T(t) g = T(t) R = L(t) g 1 n+1 g where row vector L(t) contains n + 1 elements based on each of the reference residence times incorporated in R n+1. Equation (12) defines the marginal utility represented by P g (t), for attribute g, at residence time t, as an interpolation of its values at the reference points established in g. This interpolation uses Lagrangian interpolation polynomials (LIPs) as given by Tyrrell: (13) L i (t) = k i (t r k ) (r i r k ) for i = (0, 1, 2,...,n) which defines the n + 1 elements of L(t).The LIPs in (12) take on known values at each of the reference points, or residence times r i.in particular, LIP L j (t) = 1 when t equals reference point r j, while L j (t) = 0 when t equals some other reference point r i, for i j. Fort falling between any two reference points, the LIPs will take on interpolated values, yielding an estimated marginal utility for the attribute through equation (12). Equation (13) also implies that i L i(t) = 1. For example, assume that the parameters fi (i = 0,...,n) represent the marginal utility of income at each r i reference point. The model estimates the value of fi at reference points 0, 1,..., n, generating estimates f 0, f 1,..., fn. 3 These estimates represent the marginal utility of income for residents whose length of residency corresponds exactly to the associated reference point, because L i (t) = 1 when t = r i. For residents whose length of residency does not correspond exactly to one of the reference points, the interpolated coefficient value (i.e., marginal utility of income) is equal to f 0 (L 0 (t)) + f 1 (L 1 (t)) + + fn (L n (t)). The Survey The Rhode Island Rural Land Use survey was designed to assess rural residents tradeoffs among attributes of residential development and conservation (Johnston, Swallow, and Bauer). Respondents from four Rhode Island rural communities (Burrillville, Exeter, West Greenwich, and Coventry) were asked to consider alternative development options for hypothetical tracts of forested land located in their local town. Respondents were provided with two development options, a current development plan and an alternate development plan, where each plan could differ across a set of spatial and nonspatial attributes. These attributes characterized 3 In this case, parameters ( gi ) are estimated using maximum likelihood, more specifically a binary logit model incorporating random effects.

5 1004 November 2003 Amer. J. Agr. Econ. land use features and amenities identified by focus groups and interviews with growth management practitioners. Survey development required approximately eighteen months, and involved background research, and interviews with policy makers and focus groups. Individual and group pretests ensured that survey language and format could be easily understood by respondents, and that respondents shared consistent interpretations of survey scenarios (cf. Johnston et al.). Attributes distinguishing management plans were chosen based on focus groups and interviews, and characterized protected open space (acreage, location, access), residential development (housing density, acreage, location, shape), unprotected undeveloped land, scenic views, wildlife habitat, recreational facilities, traffic, and taxes. Table 1 characterizes attributes distinguishing hypothetical management plans. Prior to presenting respondents with development choices, the survey provided background information on community land use and trade-offs implicit in development choices. Contingent choice instructions and questions were then presented. Each respondent considered three potential pairs of current and alternate plans for the same 400-acre undeveloped site. Respondents were instructed to consider each pair independent of previous choices, and to assume that all choices applied to the same parcel. Respondents were told that if you do not vote for either plan, development will automatically occur as shown by the current development plan, thereby specifying the status quo that would occur if no choice were made (Adamowicz et al.). This framework was chosen to mimic actual community considerations of development proposals, wherein a landowner possesses the property rights necessary to permit development. However, officials may seek to influence the configuration of the development, delaying permits unless changes are made. As a result, officials may exert control over the ultimate form of development. A fractional factorial design was used to construct survey questions with an orthogonal array of attribute levels. 4 All attributes were free to vary over their full range for both the current and alternate plans, with no imposed ordering of attribute levels between the two 4 The statistical design was conducted by Don Anderson of STATdesign, Evergreen CO. plans. This resulted in 128 unique contingent choice questions divided among forty-three different survey booklets (three questions per booklet). Surveys were mailed to 4000 randomly selected residents of the four Rhode Island towns, following the total survey design method (Dillman). Of 3702 deliverable surveys, 2157 were returned, providing 6062 (94% of the potential 6471) complete and usable responses to dichotomous choice questions. Of these, 5774 observations included information regarding length of residency. Resident groups were defined based on responses to the survey question: How long you have lived in your current town? The question was open-ended, and specified that the response be given in years. Approximately 41% of all usable observations indicated ten or fewer years of residency, 35% indicated between eleven and thirty years of residency, and 17% indicated more than thirty years of residency. The Econometric Model Based in part on results of focus groups, reference points for length-of-residency were set at 0, 10, and 30 years. This results in three LIPs, implying a quadratic polynomial function. Based on (12), the three LIPs are given by (14) L 0 (t) = L 1 (t) = L 2 (t) = (t 10)(t 30) (0 10)(0 30) (t 0)(t 30) (10 0)(10 30) (t 0)(t 10) (30 0)(30 10), where t represents the length of residency for each observation. Combined with (13), the terms in (14) provide the basis for empirical estimation of the LIP model. As noted above, these reference points are chosen for convenience of interpretation only; statistical properties of the model are identical regardless of the particular anchor points chosen. Corresponding with the cut-points selected for the LIP model, three length-of-residency categories were selected for the dummy variable (or systematically varying slopes) model. These distinguished residents with (a) less than 10 years of residency; (b) between 10 and 29 years of residency, inclusive; and (c)30 or more years of residency. Following (4), the

6 Johnston et al. Rural Amenity Values 1005 Table 1. Model Variables: Definitions and Summary Statistics Units and Mean Variable Name Description Measurement (Std. Dev.) Adj open Iso open Size dif Dense dif Lg mammal Sm mammal Com bird Uncom bird Wet sp Traf light Taxdif Lowvis Edgearea Develop2 The difference between acres of open space adjacent to developments and roads in the CDP and ADP. The difference between acres of open space not adjacent to developments and roads in the CDP and ADP. The difference between acres of residential development in the CDP and ADP. The difference in housing density in the CDP and ADP. Difference between habitat quality for large mammals in the CDP and ADP. Difference between habitat quality for small mammals in the CDP and ADP. Difference between habitat quality for common birds in the CDP and ADP. Difference between habitat quality for uncommon birds in the CDP and ADP. Difference between habitat quality for wetland species in the CDP and ADP. Difference between dummy variables indicating the presence of a traffic light on the main road, in the CDP and ADP. Difference in additional annual taxes and fees between the CDP and ADP (resulting from management plan). Difference between dummy variables indicating the presence of development either highly screened or not visible from the main road in the CDP and ADP. Survey versions included eight different photographs characterizing different development visibility levels; four of these photographs are characterized as low visibility development. The difference between the edge-area ratio of residential development shown in the CDP and that in ADP. Ratios are calculated as the sum of the perimeter(s) divided by the sum of the area(s) of land highlighted for residential development. Difference between dummy variables indicating the presence of a two-section, fragmented development in the CDP and ADP. In all cases, development sections are rectangular. Acres in CDP minus acres in ADP. (Range: 200 to 200) Acres in CDP minus acres in ADP. (Range: 200 to 200) Acres in CDP minus acres in ADP. (Range: 200 to 200) Houses/acre in CDP minus houses/acre in ADP. (Range: 2 to2) Difference in wildlife habitat quality scale (1 = worst; 5 = best). Difference in wildlife habitat quality scale (1 = worst; 5 = best). Difference in wildlife habitat quality scale (1 = worst; 5 = best). Difference in wildlife habitat quality scale (1 = worst; 5 = best). Difference in wildlife habitat quality scale (1 = worst; 5 = best). Difference between dummy variables for CDP and ADP. Dollars in CDP minus dollars in ADP. (Range: $325 to $325) Difference between dummy variables for CDP and ADP Calculated at a scale of 1 unit = ft. (e.g., a 1 unit 1 unit square is equivalent to 20 acres or 871,180 square feet, with an edge-area ratio of 4). (Range: to 8.5) Difference between dummy variables for CDP and ADP ( ) ( ) ( ) (0.9821) (1.2311) (1.2444) (1.7470) (1.7316) (1.7348) (0.7059) ( ) (0.7004) (3.7059) (0.4273)

7 1006 November 2003 Amer. J. Agr. Econ. Table 1. Continued Units and Mean Variable Name Description Measurement (Std. Dev.) Develop4 Develop road Lo inc Age Hi edu Difference between dummy variables indicating the presence of a four- or five-section, fragmented development in the CDP and ADP. In all cases, development sections are rectangular. Difference between dummy variables indicating the presence of developments located adjacent to main roads, in the CDP and ADP. Dummy variable identifying those respondents with reported household income below $40,000 per year. Reported age of survey respondent, in years. Dummy variable identifying those respondents with at least a four-year college education. Note: CDP = Current Development Plan; ADP = Alternate Development Plan. model estimates unique slope and intercept parameters for each of the three categories. The interaction model (5) does not require specification of anchor points or a priori categories. As the final data comprise three responses per respondent, there is a possibility of correlated errors across responses (Alberini, Kanninen, and Carson; Poe, Welsh, and Champ). One common approach to modeling such potential correlation is to split in (2) into two components: that is i.i.d. across all respondents and for each individual respondent, and a random effect h that represents systematic variation related to unobserved characteristics of respondent h (Alberini, Kanninen, and Carson; Hsiao), such that (15) du h = dv h ( + h ) Difference between dummy variables for CDP and ADP. Difference between dummy variables for CDP and ADP (0.6041) (0.7199) Dummy variable (0,1) (0.4045) Years ( ) Dummy variable (0,1) (0.4707) where the subscript h indexes individual respondents. If the h are assumed normally distributed across respondents, and we assume a logistic distribution of, the model is estimated as a random effects logit model (Pendergast et al.). Preliminary random-effects LIP models were estimated to assess whether (13) should be amended to incorporate quadratic interactions with demographic attributes such as age, education, and income. Likelihood ratio tests assessing the joint significance of these appended interactions fail to reject the null hypothesis of zero joint influence for interactions including a respondent s age ( 2 = 50.6, df = 51; p = 0.49) and a dummy variable indicating respondents with at least a four-year college education ( 2 = 54.5, df = 51; p = 0.34). Likelihood ratio tests reject the null hypothesis of zero joint influence for interactions involving a dummy variable identifying respondents with income below $40 k ( 2 = 77.81, df = 51; p = 0.01). However, analogous tests fail to reject the null hypothesis of zero joint influence for interactions between the low-income dummy variable and all model attributes except the payment vehicle ( 2 = 48.42, df = 48; p = 0.46). Based on these results, the final models include quadratic interactions between the low-income dummy and the tax change (i.e., the payment vehicle), but exclude other demographic interactions. Empirical Findings Table 2 illustrates the results for the LIP and dummy variable models; table 3 illustrates the results for the interaction model. For the LIP model, parameter estimates prefixed with L 0, L 1, and L 2 correspond to estimated parameter values at the reference points of 0, 10, and 30 years of residency, respectively. For the dummy variable model, the same prefixes correspond to parameter values for those with 0 9 years, years, and >30 years of residency, respectively.

8 Johnston et al. Rural Amenity Values 1007 Table 2. Results for Lagrangian Interpolation Polynomial and Dummy Variable Models Lagrangian Interpolation Polynomial Model Dummy Variable Model Variable Reference Parameter Std. Reference Parameter Std. Name Point Estimate Error Point Estimate Error Intercept L 0 (0 yrs.) L 0 (0 9 yrs.) Intercept L 1 (10 yrs.) L 1 (10 29 yrs.) Intercept L 2 (30 yrs.) L 2 (>30 yrs.) Edgearea L L Edgearea L L Edgearea L L Develop2 L L Develop2 L L Develop2 L L Develop4 L L Develop4 L L Develop4 L L Iso open L L Iso open L L Iso open L L Adj open L L Adj open L L Adj open L L Develop road L L Develop road L L Develop road L L Lg mammal L L Lg mammal L L Lg mammal L L Sm mammal L L Sm mammal L L Sm mammal L L Com bird L L Com bird L L Com bird L L Uncom bird L L Uncom bird L L Uncom bird L L Wet sp L L Wet sp L L Wet sp L L Dense dif L L Dense dif L L Dense dif L L Size dif L L Size dif L L Size dif L L Traf light L L Traf light L L Traf light L L Lowvis L L Lowvis L L Lowvis L L Taxdif L L Taxdif L L Taxdif L L Taxdif lo inc L L Taxdif lo inc L L Taxdif lo inc L L ln( ) ln( ) Log Likelihood (Likelihood Ratio Test) p < 0.10; p < 0.05; p < 0.01.

9 1008 November 2003 Amer. J. Agr. Econ. All models are statistically significant at p < 0.01, based on likelihood ratio tests at the appropriate degrees of freedom (tables 2, 3). Likelihood ratio tests (of the three unrestricted models versus a restricted model in which homogeneous preferences are imposed) reveal that restrictions imposing homogeneous preferences have a statistically significant impact at p < 0.01 for the LIP model ( 2 = 61.02, df = 36) and at p < 0.05 for the interaction model ( 2 = 29.98, df = 17). However, for the dummy variable model we cannot reject homogeneity in preferences at p < 0.10 ( 2 = 41.59, df = 36, p < 0.25). That is, where both the LIP and interaction models are able to detect joint, statistically significant preference heterogeneity, this heterogeneity cannot be established by the dummy variable model. Given the inability of the dummy variable model to identify statistically significant heterogeneity in preferences, subsequent discussions are based primarily on the results of the LIP and interaction models. Although likelihood ratio tests indicate that length of residency has a significant effect on both the LIP and interaction models, results for individual attributes may vary. Hence, for each individual model attribute, we test for statistically significant differences in parameter estimates between the reference points of 0 and 10 years, 10 and 30 years, and 0 and 30 years. The results are shown in table 4. To reduce table size, the test results are shown only for variables for which homogeneity in preferences can be rejected at p < 0.10, based on asymptotic Wald tests (Judge et al.). The results for other attributes are suppressed. Wald tests reject the null hypothesis of preference homogeneity for seven of sixteen attributes (table 4). These results indicate that residence duration may or may not influence the marginal utility of rural development or conservation outcomes, depending on the particular attributes involved. While the magnitude and sign of the estimated parameter shifts are often similar across the interaction and LIP models, the attributes and reference points over which the null hypothesis is rejected are not identical. For example, change in the parameter estimate associated with adj open (open space adjacent to developments) is similar in both models. However, while the interaction model rejects the null hypothesis of preference homogeneity between the reference points of 0 and 10 years, 10 and 30 years, and 0 and 30 years, the LIP model only rejects the null hypothesis between 10 and 30 years. Table 3. Results for the Interaction Model Parameter Std. Variable Name Estimate Error Intercept Intercept years Edgearea Edgearea years Develop Develop2 years Develop Develop4 years Iso open Iso open years Adj open Adj open years Develop road Develop road years Lg mammal Lg mammal years Sm mammal Sm mammal years Com bird Com bird years Uncom bird Uncom bird years Wet sp Wet sp years Dense dif Dense dif years Size dif Size dif years Traf light Traf light years Lowvis Lowvis years Taxdif Taxdif years Taxdif lo inc Taxdif lo inc years ln( ) Log Likelihood (Likelihood Ratio Test) p < 0.10; p < 0.05; p < Identical results would be generated were one to compare the (first-order) interaction model with an analogous model that also incorporated quadratic interactions (i.e., (6)), given the statistical equivalence of a secondorder LIP and quadratic interaction model. Hence, the comparison between interaction and LIP model results may also be interpreted as illustrating the implications of modeling marginal utilities as a higher-order

10 Johnston et al. Rural Amenity Values 1009 Table 4. Model Attributes With Significant Differences in Estimated Coefficients Estimated at 0, 10 and 30 Years of Residency (Wald Test Results) Lagrangian Interpolation Polynomial Model Interaction Model Parameter Estimated 2 for Null Estimated 2 for Null Estimates Coefficient Hypothesis Coefficient Hypothesis Attributes Compared Difference (Diff. = 0) Difference (Diff. = 0) Develop Develop road Sm mammal Traf-light Adj open Dense dif Taxdif Lo inc p < 0.10; p < 0.05; p < polynomial function of residence duration. In this case, the introduction of nonlinearity into the relationship between length of residency and marginal utility allows the model to identify additional attributes for which length of residency may influence marginal preferences (e.g., devroad, taxdif lo inc). However, for other attributes (e.g., develop2, dense dif ) a statistically significant length-of-residency effect is found in the simpler interaction model but disappears when one models marginal utilities as a higher-order polynomial function of residence duration. Length of Residency and Rural Amenity Preferences Effects related to length of residency notwithstanding, the signs of parameter estimates correspond with prior expectations derived from focus groups, where prior expectations exist. For example, respondents preferred development plans characterized by: (a) larger areas of preserved open space, both isolated from and adjacent to roads and developments (iso open; adj open); (b) smaller areas of developed land (size dif ); (c) lower housing densities (dense dif ); (d) improved habitat for large mammals (lg mammal), birds (com bird; uncom bird), and wetland species (wet sp); (e) low visibility development (lowvis); and (f ) lower taxes (taxdif ). As shown in table 4, the attributes for which homogeneity in preferences may be rejected by either the LIP or interaction models include develop2 (indicating developments in two-part clusters); develop road (indicating developments adjacent to main roads); sm mammal (habitat quality for small mammals); traf light (indicating the presence of traffic lights); adj open (acres of open space adjacent to developments and roads); dense dif (number of houses per acre in developed areas); and taxdif lo inc (interaction between the payment vehicle and a dummy variable identifying households with annual income <$40,000). Implications for each attribute are summarized below, along with the models in which significant heterogeneity is identified.

11 1010 November 2003 Amer. J. Agr. Econ. 1. Develop road (LIP model): Residents prefer developments to be isolated from main roads. This effect diminishes as length of residency increases 5 ; the placement of developments adjacent to main roads has a more negative marginal utility for new residents. 2. Develop2 (interaction model): Respondents with five or fewer years of residency have a positive marginal utility associated with the division of developments into two distinct subclusters. As length of residency increases marginal utility declines, such that respondents with six or more years of residency have a negative marginal utility for this attribute. 3. Sm mammal (LIP and interaction models): A positive marginal utility for improvements in small mammal habitat at zero years of residency declines with increases in residence time. Marginal utility for this attribute becomes negative at approximately ten years of residency. 4. Traf light (LIP and interaction models): All groups have positive preferences for the presence of traffic lights. As length of residency increases, positive marginal utility increases. 5. Adj open (LIP and interaction models): All groups have positive marginal utility associated with open space preserved adjacent to residential developments or roads; marginal utility declines as length of residency increases. 6. Dense dif (interaction model): Respondents have a negative marginal utility for housing density that increases in magnitude with length of residency. 7. Taxdif (LIP model): All groups have negative preferences for tax increases, and the coefficient estimate associated with linear (i.e., not interacted with income) tax increases cannot be shown to differ as a function of length of residency. However, the effect of low income (i.e., those with incomes less than $40 k; table 1) on the marginal utility of taxes differs as a function of length of residency. Tax changes have a greater impact on utility for lowincome individuals as length of residency increases. Welfare Implications While the valuation literature places a dominant concern on estimating WTP and welfare changes, we caution that the policy relevance of preference differences may hinge directly on marginal utilities. In some contexts, such as voting on growth control policies or voicing concerns in public hearings, it may be differences in marginal utilities that drive the dynamics of public input. Moreover, differences in marginal utilities may imply that differences in WTP are relevant, even if WTP differences fail to meet preset criteria for statistical significance. Despite these potential caveats, estimated changes in marginal WTP remain an important means of policy assessment. Accordingly, marginal WTP estimates are calculated for all model attributes over the reference points of 0, 10, and 30 years of residency; the resulting differences are tested for statistical significance. For the LIP model, WTP is calculated following Hanemann; mean WTP for a marginal change in the gth attribute is equal to g / f, where g is the parameter estimate corresponding to the gth attribute, and f is the parameter estimate corresponding to cost (taxdif ). This approach is extended to incorporate the interaction between taxdif and the low-income indicator variable (lo inc), such that WTP differences between length-ofresidency reference points for the gth attribute are given by (16) g,li f,li + ( f lo inc,li lo inc) g,lj f,lj + ( f lo inc,lj lo inc) for i j. Here, i,j = (0, 1, 2) correspond to the three length-of-residency reference points, f lo inc represents the parameter associated with taxdif lo inc, and WTP is calculated based on the sample mean value for lo inc (table 1). For the interaction model, WTP at any particular length-of-residency reference point, for the gth attribute is calculated 5 The LIP model indicates that the parameter estimate for develop road is negative over the entire range of length of residency. The interaction model indicates a negative parameter estimate for all residents with less than 65 years of residency. (17) ( g + ( g t t)) ( f + ( f t t)) + ( f lo inc + ( f lo inc t t))

12 Johnston et al. Rural Amenity Values 1011 Table 5. Differences in Estimated WTP at 0, 10 and 30 Years of Residency Lagrangian Interpolation Polynomial Model Interaction Model t-statistic t-statistic Estimated for Null Estimated for Null WTP WTP Hypothesis Coefficient Hypothesis Attributes Comparison at: a Difference (Diff. = 0) Difference (Diff. = 0) Develop2 0 vs. 10 yrs vs. 30 yrs vs. 30 yrs Develop road 0 vs. 10 yrs vs. 30 yrs vs. 30 yrs Sm mammal 0 vs. 10 yrs vs. 30 yrs vs. 30 yrs Traf light 0 vs. 10 yrs vs. 30 yrs vs. 30 yrs Adj open 0 vs. 10 yrs vs. 30 yrs vs. 30 yrs a WTP differences are calculated by subtracting WTP at the second reference point from WTP at the first reference point. p < 0.10; p < 0.05; p <0.01. where t represents length of residency, g t represents the parameter estimate associated with the quadratic interaction of the gth attribute and t, and f t represents the parameter estimate associated with the quadratic interaction of taxdif and t. The results are shown in table 5; only those attributes supporting at least one significant WTP difference are shown. Standard errors for estimated WTP are generated following Park, Loomis, and Creel, and Krinsky and Robb. Significant (p < 0.10) WTP differences are found for five attributes in the LIP model, and for three attributes in the interaction model. Despite differences in statistical significance, the relationship between WTP and length of residency is similar across both models (table 5). 6 Compared to new residents, longer-term residents are willing to pay (a) less to prevent the placement of developments adjacent to main roads; (b) more for the addition of traffic signals; and (c) less for open space adjacent to developments and roads. Longerterm residents have a negative WTP for small mammal habitat quality; WTP is positive for new residents. 6 The one notable exception is the attribute develop2, which manifests different WTP patterns across the two models (table 5). In addition to the results shown in table 5, both the LIP and interaction models allow calculation of estimated marginal WTP for the full range of residency durations. These results are illustrated graphically for sm mammal and adj open, two attributes for which both the LIP and interaction models indicate statistically significant changes in WTP associated with residence duration (figures 1 and 2). WTP change is also illustrated for develop road, an attribute for which the LIP model indicates a statistically significant change in WTP associated with residence duration, but for which the interaction model does not indicate a statistically significant WTP difference (figure 3). All figures illustrate estimated WTP, as estimated by both models, over the range of zero to forty years. As shown in figures 1 3, changes predicted by the LIP and interaction models are similar, even for the case in which only one model predicts a statistically significant change in WTP (i.e., develop road in figure 3). In the case of adj open, the two models predict virtually identical changes in WTP. Similar results hold for a wide range of model attributes; despite differences in functional form and the resulting statistical significance of WTP differences (table 5), point estimates of WTP as influenced by length of residency are similar across both models.

13 1012 November 2003 Amer. J. Agr. Econ. $20.00 Marginal WTP: sm_mammal... $15.00 $10.00 $5.00 $0.00 -$5.00 -$ $ $ Years of Residency LIP Interact Figure 1. Marginal willingness to pay for habitat improvements for small mammals (sm mammal; 5-point habitat index) Discussion For attributes whose parameter estimates reveal statistically significant changes associated with residence duration, WTP changes match those suggested by prior research. For example, the negative influence of residence time on WTP for small mammal habitat suggests that certain species may be increasingly viewed as nuisance species as residence time in rural communities increases. This matches similar findings from the wildlife management literature, which suggests that negative reactions to certain small mammals may increase with long-term proximity (Zin and Andelt). The results suggest that preferences for scenic and related attributes of residential developments differ according to length of residency, again supporting common intuition (Spain). According to survey results (table 5, figure 3) newer residents have a higher WTP to prevent the location of new developments in a more visible location adjacent to main roads. 7 Similarly, compared to longer-term residents, newer residents have a higher marginal WTP for open space located adjacent to roads and developments (table 5, figure 2). Despite these findings, preferences for low-visibility development (lowvis) do not differ across groups. Hence, although results reveal evidence of differences in preferences for scenic attributes, these findings are not universal. Model results also indicate that longer-term residents are less willing to support develop- 7 Estimated WTP for develop road is negative; hence respondents are willing to pay to prevent the presence of developments adjacent to main roads. ments characterized by multiple clusters of developed land; this again corresponds with prior evidence in the literature. Both models indicate a generally negative effect of length of residency on WTP to split residential developments into two equal parts (develop2), compared to development of equal area in a single contiguous unit. 8 This corresponds with prior findings of McLeod, Woirhaye, and Menkhaus, who find a negative influence of residence duration (in Sublette County, Wyoming) on the probability of endorsing cluster development. Preferences for traffic control reveal a significant and intuitive increase in WTP among those with longer residency. This may indicate that new residents have moved to rural communities in part to escape traffic congestion that they associate with traffic controls, and hence maintain a lower WTP for such attributes. The results may also indicate a relatively stronger preference for more rapid travel with fewer traffic controls among newer residents. The latter explanation would fit the common characterization of new rural residents as commuters who rely on rapid automobile transportation to reach places of employment in central cities or suburbs (Dubbink). Finally, while neither model predicts welfare changes that are contrary to intuition, preferences for many attributes addressed by the survey cannot be shown to differ according to 8 In the LIP model, WTP is negative, with the magnitude of negative WTP increasing with the length of residency. In the interaction model, WTP is positive at zero years of residency, declining with length of residency to become negative at approximately six years of residency.

14 Johnston et al. Rural Amenity Values 1013 $1.40 Marginal WTP: adj_open... $1.20 $1.00 $0.80 $0.60 $0.40 $0.20 LIP Interact $ Years of Residency Figure 2. Marginal willingness to pay for preserved open space adjacent to roads or developments (adj open; dollars per acre) length of residency. These include attributes characterizing both residential development and open space (e.g., size dif, iso open; table 1). Other attributes, such as dense dif (housing density), reveal statistically significant changes in parameter estimates that do not translate into significant changes in WTP. Hence, while results suggest statistically significant differences in development and conservation preferences between newer and more established rural residents, for many attributes the null hypothesis of preference homogeneity is not rejected. Conclusion This article explores whether stated preferences for development and conservation Marginal WTP: develop_road $0.00 -$ $ $ $ $ $60.00 trade-offs differ according to length of residency. We find that residence duration is associated with statistically significant changes in marginal WTP for attributes of residential developments, preserved open space, traffic infrastructure, and wildlife habitat. Where statistically significant welfare effects are found, changes in WTP match common intuition. However, as might be expected, statistically significant changes in marginal WTP only apply to a subset of land use attributes included in the survey. Analysis of the change in marginal WTP for rural amenities reveals that the LIP and interaction models provide similar forecasts, despite differences in the models ability to identify statistically significant changes in WTP across length-of-residency reference points. These findings suggest that results indicating LIP Interact Figure 3. -$ Years of Residency Marginal willingness to pay to locate development on main roads (develop road)

Context Similarity and the Validity of Benefits Transfer: Is the Common Wisdom Correct?

Context Similarity and the Validity of Benefits Transfer: Is the Common Wisdom Correct? Context Similarity and the Validity of Benefits Transfer: Is the Common Wisdom Correct? Robert J. Johnston Department of Agricultural and Resource Economics University of Connecticut April 6, 2006 Contact

More information

Multidimensional Spatial Heterogeneity in Ecosystem Service Values: Advancing the Frontier

Multidimensional Spatial Heterogeneity in Ecosystem Service Values: Advancing the Frontier Multidimensional Spatial Heterogeneity in Ecosystem Service Values: Advancing the Frontier Robert J. Johnston 1 Benedict M. Holland 2 1 George Perkins Marsh Institute and Department of Economics, Clark

More information

Enhancing the Geospatial Validity of Meta- Analysis to Support Ecosystem Service Benefit Transfer

Enhancing the Geospatial Validity of Meta- Analysis to Support Ecosystem Service Benefit Transfer Enhancing the Geospatial Validity of Meta- Analysis to Support Ecosystem Service Benefit Transfer Robert J. Johnston Clark University, USA Elena Besedin Abt Associates, Inc. Ryan Stapler Abt Associates,

More information

The effects of impact fees on urban form and congestion in Florida

The effects of impact fees on urban form and congestion in Florida The effects of impact fees on urban form and congestion in Florida Principal Investigators: Andres G. Blanco Ruth Steiner Presenters: Hyungchul Chung Jeongseob Kim Urban and Regional Planning Contents

More information

Neighborhood social characteristics and chronic disease outcomes: does the geographic scale of neighborhood matter? Malia Jones

Neighborhood social characteristics and chronic disease outcomes: does the geographic scale of neighborhood matter? Malia Jones Neighborhood social characteristics and chronic disease outcomes: does the geographic scale of neighborhood matter? Malia Jones Prepared for consideration for PAA 2013 Short Abstract Empirical research

More information

HORIZON 2030: Land Use & Transportation November 2005

HORIZON 2030: Land Use & Transportation November 2005 PROJECTS Land Use An important component of the Horizon transportation planning process involved reviewing the area s comprehensive land use plans to ensure consistency between them and the longrange transportation

More information

INTRODUCTION TO TRANSPORTATION SYSTEMS

INTRODUCTION TO TRANSPORTATION SYSTEMS INTRODUCTION TO TRANSPORTATION SYSTEMS Lectures 5/6: Modeling/Equilibrium/Demand 1 OUTLINE 1. Conceptual view of TSA 2. Models: different roles and different types 3. Equilibrium 4. Demand Modeling References:

More information

Developed new methodologies for mapping and characterizing suburban sprawl in the Northeastern Forests

Developed new methodologies for mapping and characterizing suburban sprawl in the Northeastern Forests Development of Functional Ecological Indicators of Suburban Sprawl for the Northeastern Forest Landscape Principal Investigator: Austin Troy UVM, Rubenstein School of Environment and Natural Resources

More information

Figure 8.2a Variation of suburban character, transit access and pedestrian accessibility by TAZ label in the study area

Figure 8.2a Variation of suburban character, transit access and pedestrian accessibility by TAZ label in the study area Figure 8.2a Variation of suburban character, transit access and pedestrian accessibility by TAZ label in the study area Figure 8.2b Variation of suburban character, commercial residential balance and mix

More information

Typical information required from the data collection can be grouped into four categories, enumerated as below.

Typical information required from the data collection can be grouped into four categories, enumerated as below. Chapter 6 Data Collection 6.1 Overview The four-stage modeling, an important tool for forecasting future demand and performance of a transportation system, was developed for evaluating large-scale infrastructure

More information

Chapter 7. Testing Linear Restrictions on Regression Coefficients

Chapter 7. Testing Linear Restrictions on Regression Coefficients Chapter 7 Testing Linear Restrictions on Regression Coefficients 1.F-tests versus t-tests In the previous chapter we discussed several applications of the t-distribution to testing hypotheses in the linear

More information

Introduction to Econometrics. Regression with Panel Data

Introduction to Econometrics. Regression with Panel Data Introduction to Econometrics The statistical analysis of economic (and related) data STATS301 Regression with Panel Data Titulaire: Christopher Bruffaerts Assistant: Lorenzo Ricci 1 Regression with Panel

More information

Introduction to Econometrics

Introduction to Econometrics Introduction to Econometrics STAT-S-301 Panel Data (2016/2017) Lecturer: Yves Dominicy Teaching Assistant: Elise Petit 1 Regression with Panel Data A panel dataset contains observations on multiple entities

More information

Neighborhood Locations and Amenities

Neighborhood Locations and Amenities University of Maryland School of Architecture, Planning and Preservation Fall, 2014 Neighborhood Locations and Amenities Authors: Cole Greene Jacob Johnson Maha Tariq Under the Supervision of: Dr. Chao

More information

Chapter 9: The Regression Model with Qualitative Information: Binary Variables (Dummies)

Chapter 9: The Regression Model with Qualitative Information: Binary Variables (Dummies) Chapter 9: The Regression Model with Qualitative Information: Binary Variables (Dummies) Statistics and Introduction to Econometrics M. Angeles Carnero Departamento de Fundamentos del Análisis Económico

More information

The impact of residential density on vehicle usage and fuel consumption*

The impact of residential density on vehicle usage and fuel consumption* The impact of residential density on vehicle usage and fuel consumption* Jinwon Kim and David Brownstone Dept. of Economics 3151 SSPA University of California Irvine, CA 92697-5100 Email: dbrownst@uci.edu

More information

Lecture-20: Discrete Choice Modeling-I

Lecture-20: Discrete Choice Modeling-I Lecture-20: Discrete Choice Modeling-I 1 In Today s Class Introduction to discrete choice models General formulation Binary choice models Specification Model estimation Application Case Study 2 Discrete

More information

AGEC 603. Stylized Cited Assumptions. Urban Density. Urban Density and Structures. q = a constant density the same throughout

AGEC 603. Stylized Cited Assumptions. Urban Density. Urban Density and Structures. q = a constant density the same throughout AGEC 603 Urban Density and Structures Stylized Cited Assumptions q = a constant density the same throughout c = constant structures the same throughout Reality housing is very heterogeneous Lot size varies

More information

Data Collection. Lecture Notes in Transportation Systems Engineering. Prof. Tom V. Mathew. 1 Overview 1

Data Collection. Lecture Notes in Transportation Systems Engineering. Prof. Tom V. Mathew. 1 Overview 1 Data Collection Lecture Notes in Transportation Systems Engineering Prof. Tom V. Mathew Contents 1 Overview 1 2 Survey design 2 2.1 Information needed................................. 2 2.2 Study area.....................................

More information

The Built Environment, Car Ownership, and Travel Behavior in Seoul

The Built Environment, Car Ownership, and Travel Behavior in Seoul The Built Environment, Car Ownership, and Travel Behavior in Seoul Sang-Kyu Cho, Ph D. Candidate So-Ra Baek, Master Course Student Seoul National University Abstract Although the idea of integrating land

More information

Applied Economics. Panel Data. Department of Economics Universidad Carlos III de Madrid

Applied Economics. Panel Data. Department of Economics Universidad Carlos III de Madrid Applied Economics Panel Data Department of Economics Universidad Carlos III de Madrid See also Wooldridge (chapter 13), and Stock and Watson (chapter 10) 1 / 38 Panel Data vs Repeated Cross-sections In

More information

Forecasts for the Reston/Dulles Rail Corridor and Route 28 Corridor 2010 to 2050

Forecasts for the Reston/Dulles Rail Corridor and Route 28 Corridor 2010 to 2050 George Mason University Center for Regional Analysis Forecasts for the Reston/Dulles Rail Corridor and Route 28 Corridor 21 to 25 Prepared for the Fairfax County Department of Planning and Zoning Lisa

More information

A generic marginal value function for natural areas

A generic marginal value function for natural areas Ann Reg Sci (2017) 58:159 179 DOI 10.1007/s00168-016-0795-0 ORIGINAL PAPER A generic marginal value function for natural areas Mark J. Koetse 1 Erik T. Verhoef 2,3 Luke M. Brander 1,4 Received: 21 August

More information

BROOKINGS May

BROOKINGS May Appendix 1. Technical Methodology This study combines detailed data on transit systems, demographics, and employment to determine the accessibility of jobs via transit within and across the country s 100

More information

Child Maturation, Time-Invariant, and Time-Varying Inputs: their Relative Importance in Production of Child Human Capital

Child Maturation, Time-Invariant, and Time-Varying Inputs: their Relative Importance in Production of Child Human Capital Child Maturation, Time-Invariant, and Time-Varying Inputs: their Relative Importance in Production of Child Human Capital Mark D. Agee Scott E. Atkinson Tom D. Crocker Department of Economics: Penn State

More information

The Economics of Ecosystems and Biodiversity on Bonaire. The value of citizens in the Netherlands for nature in the Caribbean

The Economics of Ecosystems and Biodiversity on Bonaire. The value of citizens in the Netherlands for nature in the Caribbean The Economics of Ecosystems and Biodiversity on Bonaire The value of citizens in the Netherlands for nature in the Caribbean 2 The Economics of Ecosystems and Biodiversity on Bonaire The value of citizens

More information

King City URA 6D Concept Plan

King City URA 6D Concept Plan King City URA 6D Concept Plan King City s Evolution Among the fastest growing cities 2000 Census 1,949 2010 Census 3,111 60% increase 2016 Census estimate 3,817 23% increase Average annual rate 4.3% Surpassing

More information

Appendixx C Travel Demand Model Development and Forecasting Lubbock Outer Route Study June 2014

Appendixx C Travel Demand Model Development and Forecasting Lubbock Outer Route Study June 2014 Appendix C Travel Demand Model Development and Forecasting Lubbock Outer Route Study June 2014 CONTENTS List of Figures-... 3 List of Tables... 4 Introduction... 1 Application of the Lubbock Travel Demand

More information

FINAL PROJECT PLAN TAX INCREMENT DISTRICT #66 MORNINGSTAR. Prepared by the

FINAL PROJECT PLAN TAX INCREMENT DISTRICT #66 MORNINGSTAR. Prepared by the PROJECT PLAN TAX INCREMENT DISTRICT #66 MORNINGSTAR Prepared by the Rapid City Growth Management Department January 17, 2008 INTRODUCTION Tax Increment Financing is a method of financing improvements and

More information

Marketing Research Session 10 Hypothesis Testing with Simple Random samples (Chapter 12)

Marketing Research Session 10 Hypothesis Testing with Simple Random samples (Chapter 12) Marketing Research Session 10 Hypothesis Testing with Simple Random samples (Chapter 12) Remember: Z.05 = 1.645, Z.01 = 2.33 We will only cover one-sided hypothesis testing (cases 12.3, 12.4.2, 12.5.2,

More information

Do not copy, post, or distribute

Do not copy, post, or distribute 14 CORRELATION ANALYSIS AND LINEAR REGRESSION Assessing the Covariability of Two Quantitative Properties 14.0 LEARNING OBJECTIVES In this chapter, we discuss two related techniques for assessing a possible

More information

:: STUDENTS SUPPORTED

:: STUDENTS SUPPORTED PROGRESS REPORT REPORT PERIOD: 2/1/2014-1/31/2015 PROJECT NO: 2014-R/P-NERR-14-2-REG TITLE: Coastal Hazards and Northeast Housing Values: Comparative Implications for Climate Change Adaptation and Community

More information

Measuring Agglomeration Economies The Agglomeration Index:

Measuring Agglomeration Economies The Agglomeration Index: Measuring Agglomeration Economies The Agglomeration Index: A Regional Classification Based on Agglomeration Economies J. P. Han Dieperink and Peter Nijkamp Free University, The Netherlands* Urban agglomerations

More information

Lecture 1. Behavioral Models Multinomial Logit: Power and limitations. Cinzia Cirillo

Lecture 1. Behavioral Models Multinomial Logit: Power and limitations. Cinzia Cirillo Lecture 1 Behavioral Models Multinomial Logit: Power and limitations Cinzia Cirillo 1 Overview 1. Choice Probabilities 2. Power and Limitations of Logit 1. Taste variation 2. Substitution patterns 3. Repeated

More information

Modeling and Predicting of Future Urban Growth in the Charleston, South Carolina Area

Modeling and Predicting of Future Urban Growth in the Charleston, South Carolina Area Modeling and Predicting of Future Urban Growth in the Charleston, South Carolina Area Jeffery Allen Testimony presented to the U.S. Commission on Ocean Policy January 15, 2002 Charleston, South Carolina

More information

NEW YORK DEPARTMENT OF SANITATION. Spatial Analysis of Complaints

NEW YORK DEPARTMENT OF SANITATION. Spatial Analysis of Complaints NEW YORK DEPARTMENT OF SANITATION Spatial Analysis of Complaints Spatial Information Design Lab Columbia University Graduate School of Architecture, Planning and Preservation November 2007 Title New York

More information

How wrong can you be? Implications of incorrect utility function specification for welfare measurement in choice experiments

How wrong can you be? Implications of incorrect utility function specification for welfare measurement in choice experiments How wrong can you be? Implications of incorrect utility function for welfare measurement in choice experiments Cati Torres Nick Hanley Antoni Riera Stirling Economics Discussion Paper 200-2 November 200

More information

Analyzing Suitability of Land for Affordable Housing

Analyzing Suitability of Land for Affordable Housing Analyzing Suitability of Land for Affordable Housing Vern C. Svatos Jarrod S. Doucette Abstract: This paper explains the use of a geographic information system (GIS) to distinguish areas that might have

More information

Fig 1. Steps in the EcoValue Project

Fig 1. Steps in the EcoValue Project Assessing the Social and Economic Value of Ecosystem Services in the Northern Forest Region: A Geographic Information System (GIS) Approach to Landscape Valuation Principal Investigator(s): Dr. Matthew

More information

CIV3703 Transport Engineering. Module 2 Transport Modelling

CIV3703 Transport Engineering. Module 2 Transport Modelling CIV3703 Transport Engineering Module Transport Modelling Objectives Upon successful completion of this module you should be able to: carry out trip generation calculations using linear regression and category

More information

The Stated Preference Approach to Environmental Valuation

The Stated Preference Approach to Environmental Valuation The Stated Preference Approach to Environmental Valuation Volume I: Foundations, Initial Development, Statistical Approaches Edited by Richard T. Carson University of California, San Diego, USA ASHGATE

More information

The Impact of Residential Density on Vehicle Usage and Fuel Consumption: Evidence from National Samples

The Impact of Residential Density on Vehicle Usage and Fuel Consumption: Evidence from National Samples The Impact of Residential Density on Vehicle Usage and Fuel Consumption: Evidence from National Samples Jinwon Kim Department of Transport, Technical University of Denmark and David Brownstone 1 Department

More information

Regional Performance Measures

Regional Performance Measures G Performance Measures Regional Performance Measures Introduction This appendix highlights the performance of the MTP/SCS for 2035. The performance of the Revenue Constrained network also is compared to

More information

Regional Performance Measures

Regional Performance Measures G Performance Measures Regional Performance Measures Introduction This appendix highlights the performance of the MTP/SCS for 2035. The performance of the Revenue Constrained network also is compared to

More information

Modeling Land Use Change Using an Eigenvector Spatial Filtering Model Specification for Discrete Response

Modeling Land Use Change Using an Eigenvector Spatial Filtering Model Specification for Discrete Response Modeling Land Use Change Using an Eigenvector Spatial Filtering Model Specification for Discrete Response Parmanand Sinha The University of Tennessee, Knoxville 304 Burchfiel Geography Building 1000 Phillip

More information

Three-Level Modeling for Factorial Experiments With Experimentally Induced Clustering

Three-Level Modeling for Factorial Experiments With Experimentally Induced Clustering Three-Level Modeling for Factorial Experiments With Experimentally Induced Clustering John J. Dziak The Pennsylvania State University Inbal Nahum-Shani The University of Michigan Copyright 016, Penn State.

More information

Report on Wiarton Keppel International Airport for Georgian Bluffs Prepared by: Alec Dare, BA, OCGC Research Analyst, Centre for Applied Research and

Report on Wiarton Keppel International Airport for Georgian Bluffs Prepared by: Alec Dare, BA, OCGC Research Analyst, Centre for Applied Research and Report on Wiarton Keppel International Airport for Georgian Bluffs Prepared by: Alec Dare, BA, OCGC Research Analyst, Centre for Applied Research and Innovation Georgian College March 2018 Contents 1.

More information

Regression Analysis Tutorial 77 LECTURE /DISCUSSION. Specification of the OLS Regression Model

Regression Analysis Tutorial 77 LECTURE /DISCUSSION. Specification of the OLS Regression Model Regression Analysis Tutorial 77 LECTURE /DISCUSSION Specification of the OLS Regression Model Regression Analysis Tutorial 78 The Specification The specification is the selection of explanatory variables

More information

EPSY 905: Fundamentals of Multivariate Modeling Online Lecture #7

EPSY 905: Fundamentals of Multivariate Modeling Online Lecture #7 Introduction to Generalized Univariate Models: Models for Binary Outcomes EPSY 905: Fundamentals of Multivariate Modeling Online Lecture #7 EPSY 905: Intro to Generalized In This Lecture A short review

More information

Forecasts from the Strategy Planning Model

Forecasts from the Strategy Planning Model Forecasts from the Strategy Planning Model Appendix A A12.1 As reported in Chapter 4, we used the Greater Manchester Strategy Planning Model (SPM) to test our long-term transport strategy. A12.2 The origins

More information

Infill and the microstructure of urban expansion

Infill and the microstructure of urban expansion Infill and the microstructure of urban expansion Stephen Sheppard Williams College Homer Hoyt Advanced Studies Institute January 12, 2007 Presentations and papers available at http://www.williams.edu/economics/urbangrowth/homepage.htm

More information

Chapter 10 Nonlinear Models

Chapter 10 Nonlinear Models Chapter 10 Nonlinear Models Nonlinear models can be classified into two categories. In the first category are models that are nonlinear in the variables, but still linear in terms of the unknown parameters.

More information

Chapter 9. Dummy (Binary) Variables. 9.1 Introduction The multiple regression model (9.1.1) Assumption MR1 is

Chapter 9. Dummy (Binary) Variables. 9.1 Introduction The multiple regression model (9.1.1) Assumption MR1 is Chapter 9 Dummy (Binary) Variables 9.1 Introduction The multiple regression model y = β+β x +β x + +β x + e (9.1.1) t 1 2 t2 3 t3 K tk t Assumption MR1 is 1. yt =β 1+β 2xt2 + L+β KxtK + et, t = 1, K, T

More information

Travel behavior of low-income residents: Studying two contrasting locations in the city of Chennai, India

Travel behavior of low-income residents: Studying two contrasting locations in the city of Chennai, India Travel behavior of low-income residents: Studying two contrasting locations in the city of Chennai, India Sumeeta Srinivasan Peter Rogers TRB Annual Meet, Washington D.C. January 2003 Environmental Systems,

More information

CORPORATION OF THE CITY OF COURTENAY COUNCIL MEETING AGENDA

CORPORATION OF THE CITY OF COURTENAY COUNCIL MEETING AGENDA CORPORATION OF THE CITY OF COURTENAY COUNCIL MEETING AGENDA DATE: Monday, January 14, 2013 PLACE: City Hall Council Chambers TIME: 4:00 p.m. 1.00 ADOPTION OF MINUTES 1. Adopt January 7, 2013 Regular Council

More information

Summary and Implications for Policy

Summary and Implications for Policy Summary and Implications for Policy 1 Introduction This is the report on a background study for the National Spatial Strategy (NSS) regarding the Irish Rural Structure. The main objective of the study

More information

CLAREMONT MASTER PLAN 2017: LAND USE COMMUNITY INPUT

CLAREMONT MASTER PLAN 2017: LAND USE COMMUNITY INPUT Planning and Development Department 14 North Street Claremont, New Hampshire 03743 Ph: (603) 542-7008 Fax: (603) 542-7033 Email: cityplanner@claremontnh.com www.claremontnh.com CLAREMONT MASTER PLAN 2017:

More information

LAND CHANGE MODELER SOFTWARE FOR ARCGIS

LAND CHANGE MODELER SOFTWARE FOR ARCGIS LAND CHANGE MODELER SOFTWARE FOR ARCGIS The Land Change Modeler is revolutionary land cover change analysis and prediction software which also incorporates tools that allow you to analyze, measure and

More information

Impact of Metropolitan-level Built Environment on Travel Behavior

Impact of Metropolitan-level Built Environment on Travel Behavior Impact of Metropolitan-level Built Environment on Travel Behavior Arefeh Nasri 1 and Lei Zhang 2,* 1. Graduate Research Assistant; 2. Assistant Professor (*Corresponding Author) Department of Civil and

More information

Mapping Accessibility Over Time

Mapping Accessibility Over Time Journal of Maps, 2006, 76-87 Mapping Accessibility Over Time AHMED EL-GENEIDY and DAVID LEVINSON University of Minnesota, 500 Pillsbury Drive S.E., Minneapolis, MN 55455, USA; geneidy@umn.edu (Received

More information

THE VALUE OF CULTURAL HERITAGE AN ECONOMIC ANALYSIS OF CULTURAL HERITAGE AND CULTURAL ENVIRONEMENTS

THE VALUE OF CULTURAL HERITAGE AN ECONOMIC ANALYSIS OF CULTURAL HERITAGE AND CULTURAL ENVIRONEMENTS ENGLISH SUMMARY THE VALUE OF CULTURAL HERITAGE AN ECONOMIC ANALYSIS OF CULTURAL HERITAGE AND CULTURAL ENVIRONEMENTS MENON PUBLICATION NR. 72/2017 By Caroline Wang Gierløff, Kristin Magnussen, Lars Stemland

More information

Chapter 3 Multiple Regression Complete Example

Chapter 3 Multiple Regression Complete Example Department of Quantitative Methods & Information Systems ECON 504 Chapter 3 Multiple Regression Complete Example Spring 2013 Dr. Mohammad Zainal Review Goals After completing this lecture, you should be

More information

Incorporating GIS into Hedonic Pricing Models

Incorporating GIS into Hedonic Pricing Models Blanchette 1 Alex Blanchette NRS 509 12/10/2016 Incorporating GIS into Hedonic Pricing Models One of the central themes in environmental economics is determining the value individuals derive from various

More information

Urban Revival in America

Urban Revival in America Urban Revival in America Victor Couture 1 Jessie Handbury 2 1 University of California, Berkeley 2 University of Pennsylvania and NBER May 2016 1 / 23 Objectives 1. Document the recent revival of America

More information

A Micro-Analysis of Accessibility and Travel Behavior of a Small Sized Indian City: A Case Study of Agartala

A Micro-Analysis of Accessibility and Travel Behavior of a Small Sized Indian City: A Case Study of Agartala A Micro-Analysis of Accessibility and Travel Behavior of a Small Sized Indian City: A Case Study of Agartala Moumita Saha #1, ParthaPratim Sarkar #2,Joyanta Pal #3 #1 Ex-Post graduate student, Department

More information

I. M. Schoeman North West University, South Africa. Abstract

I. M. Schoeman North West University, South Africa. Abstract Urban Transport XX 607 Land use and transportation integration within the greater area of the North West University (Potchefstroom Campus), South Africa: problems, prospects and solutions I. M. Schoeman

More information

Lecture-1: Introduction to Econometrics

Lecture-1: Introduction to Econometrics Lecture-1: Introduction to Econometrics 1 Definition Econometrics may be defined as 2 the science in which the tools of economic theory, mathematics and statistical inference is applied to the analysis

More information

Land Use Modeling at ABAG. Mike Reilly October 3, 2011

Land Use Modeling at ABAG. Mike Reilly October 3, 2011 Land Use Modeling at ABAG Mike Reilly michaelr@abag.ca.gov October 3, 2011 Overview What and Why Details Integration Use Visualization Questions What is a Land Use Model? Statistical relationships between

More information

1Department of Demography and Organization Studies, University of Texas at San Antonio, One UTSA Circle, San Antonio, TX

1Department of Demography and Organization Studies, University of Texas at San Antonio, One UTSA Circle, San Antonio, TX Well, it depends on where you're born: A practical application of geographically weighted regression to the study of infant mortality in the U.S. P. Johnelle Sparks and Corey S. Sparks 1 Introduction Infant

More information

Chapter 6 Stochastic Regressors

Chapter 6 Stochastic Regressors Chapter 6 Stochastic Regressors 6. Stochastic regressors in non-longitudinal settings 6.2 Stochastic regressors in longitudinal settings 6.3 Longitudinal data models with heterogeneity terms and sequentially

More information

Wooldridge, Introductory Econometrics, 3d ed. Chapter 16: Simultaneous equations models. An obvious reason for the endogeneity of explanatory

Wooldridge, Introductory Econometrics, 3d ed. Chapter 16: Simultaneous equations models. An obvious reason for the endogeneity of explanatory Wooldridge, Introductory Econometrics, 3d ed. Chapter 16: Simultaneous equations models An obvious reason for the endogeneity of explanatory variables in a regression model is simultaneity: that is, one

More information

Population Trends Along the Coastal United States:

Population Trends Along the Coastal United States: Coastal Trends Report Series Population Trends Along the Coastal United States: 1980-2008 U.S. Department of Commerce National Oceanic and Atmospheric Administration National Ocean Service Assessing the

More information

Course Econometrics I

Course Econometrics I Course Econometrics I 3. Multiple Regression Analysis: Binary Variables Martin Halla Johannes Kepler University of Linz Department of Economics Last update: April 29, 2014 Martin Halla CS Econometrics

More information

Analysis and Design of Urban Transportation Network for Pyi Gyi Ta Gon Township PHOO PWINT ZAN 1, DR. NILAR AYE 2

Analysis and Design of Urban Transportation Network for Pyi Gyi Ta Gon Township PHOO PWINT ZAN 1, DR. NILAR AYE 2 www.semargroup.org, www.ijsetr.com ISSN 2319-8885 Vol.03,Issue.10 May-2014, Pages:2058-2063 Analysis and Design of Urban Transportation Network for Pyi Gyi Ta Gon Township PHOO PWINT ZAN 1, DR. NILAR AYE

More information

Mathematics Standards for High School Financial Algebra A and Financial Algebra B

Mathematics Standards for High School Financial Algebra A and Financial Algebra B Mathematics Standards for High School Financial Algebra A and Financial Algebra B Financial Algebra A and B are two semester courses that may be taken in either order or one taken without the other; both

More information

INFERENCE APPROACHES FOR INSTRUMENTAL VARIABLE QUANTILE REGRESSION. 1. Introduction

INFERENCE APPROACHES FOR INSTRUMENTAL VARIABLE QUANTILE REGRESSION. 1. Introduction INFERENCE APPROACHES FOR INSTRUMENTAL VARIABLE QUANTILE REGRESSION VICTOR CHERNOZHUKOV CHRISTIAN HANSEN MICHAEL JANSSON Abstract. We consider asymptotic and finite-sample confidence bounds in instrumental

More information

Use of Dummy (Indicator) Variables in Applied Econometrics

Use of Dummy (Indicator) Variables in Applied Econometrics Chapter 5 Use of Dummy (Indicator) Variables in Applied Econometrics Section 5.1 Introduction Use of Dummy (Indicator) Variables Model specifications in applied econometrics often necessitate the use of

More information

Tourism in Peripheral Areas - A Case of Three Turkish Towns

Tourism in Peripheral Areas - A Case of Three Turkish Towns Turgut Var Department of Recreation, Park and Tourism Sciences Ozlem Unal Urban planner Derya Guven Akleman Department of Statistics Tourism in Peripheral Areas - A Case of Three Turkish Towns The objective

More information

Linear Regression With Special Variables

Linear Regression With Special Variables Linear Regression With Special Variables Junhui Qian December 21, 2014 Outline Standardized Scores Quadratic Terms Interaction Terms Binary Explanatory Variables Binary Choice Models Standardized Scores:

More information

Regional Growth Strategy Work Session Growth Management Policy Board

Regional Growth Strategy Work Session Growth Management Policy Board Regional Growth Strategy Work Session Growth Management Policy Board September 6, 2018 1 Overview Recap June GMPB work session Objectives and outcomes Regional geographies Growth scenarios Breakout Discussion:

More information

Local Area Key Issues Paper No. 13: Southern Hinterland townships growth opportunities

Local Area Key Issues Paper No. 13: Southern Hinterland townships growth opportunities Draft Sunshine Coast Planning Scheme Review of Submissions Local Area Key Issues Paper No. 13: Southern Hinterland townships growth opportunities Key Issue: Growth opportunities for Southern Hinterland

More information

Econometrics Lecture 5: Limited Dependent Variable Models: Logit and Probit

Econometrics Lecture 5: Limited Dependent Variable Models: Logit and Probit Econometrics Lecture 5: Limited Dependent Variable Models: Logit and Probit R. G. Pierse 1 Introduction In lecture 5 of last semester s course, we looked at the reasons for including dichotomous variables

More information

Subject: Note on spatial issues in Urban South Africa From: Alain Bertaud Date: Oct 7, A. Spatial issues

Subject: Note on spatial issues in Urban South Africa From: Alain Bertaud Date: Oct 7, A. Spatial issues Page 1 of 6 Subject: Note on spatial issues in Urban South Africa From: Alain Bertaud Date: Oct 7, 2009 A. Spatial issues 1. Spatial issues and the South African economy Spatial concentration of economic

More information

Does agglomeration explain regional income inequalities?

Does agglomeration explain regional income inequalities? Does agglomeration explain regional income inequalities? Karen Helene Midelfart Norwegian School of Economics and Business Administration and CEPR August 31, 2004 First draft Abstract This paper seeks

More information

MS&E 226: Small Data

MS&E 226: Small Data MS&E 226: Small Data Lecture 15: Examples of hypothesis tests (v5) Ramesh Johari ramesh.johari@stanford.edu 1 / 32 The recipe 2 / 32 The hypothesis testing recipe In this lecture we repeatedly apply the

More information

Introducing GIS analysis

Introducing GIS analysis 1 Introducing GIS analysis GIS analysis lets you see patterns and relationships in your geographic data. The results of your analysis will give you insight into a place, help you focus your actions, or

More information

Key words: choice experiments, experimental design, non-market valuation.

Key words: choice experiments, experimental design, non-market valuation. A CAUTIONARY NOTE ON DESIGNING DISCRETE CHOICE EXPERIMENTS: A COMMENT ON LUSK AND NORWOOD S EFFECT OF EXPERIMENT DESIGN ON CHOICE-BASED CONJOINT VALUATION ESTIMATES RICHARD T. CARSON, JORDAN J. LOUVIERE,

More information

Improving the Model s Sensitivity to Land Use Policies and Nonmotorized Travel

Improving the Model s Sensitivity to Land Use Policies and Nonmotorized Travel Improving the Model s Sensitivity to Land Use Policies and Nonmotorized Travel presented to MWCOG/NCRTPB Travel Forecasting Subcommittee presented by John (Jay) Evans, P.E., AICP Cambridge Systematics,

More information

Question 1 carries a weight of 25%; Question 2 carries 20%; Question 3 carries 20%; Question 4 carries 35%.

Question 1 carries a weight of 25%; Question 2 carries 20%; Question 3 carries 20%; Question 4 carries 35%. UNIVERSITY OF EAST ANGLIA School of Economics Main Series PGT Examination 017-18 ECONOMETRIC METHODS ECO-7000A Time allowed: hours Answer ALL FOUR Questions. Question 1 carries a weight of 5%; Question

More information

R E SEARCH HIGHLIGHTS

R E SEARCH HIGHLIGHTS Canada Research Chair in Urban Change and Adaptation R E SEARCH HIGHLIGHTS Research Highlight No.8 November 2006 THE IMPACT OF ECONOMIC RESTRUCTURING ON INNER CITY WINNIPEG Introduction This research highlight

More information

Creating a Pavement Management System Using GIS

Creating a Pavement Management System Using GIS Christopher Castruita PPD 631 Prof. Barry Waite & Prof. Bonnie Shrewsbury Creating a Pavement Management System Using GIS Problem Definition As is the case with many cities, the city of La Cañada Flintridge

More information

Wooldridge, Introductory Econometrics, 4th ed. Chapter 6: Multiple regression analysis: Further issues

Wooldridge, Introductory Econometrics, 4th ed. Chapter 6: Multiple regression analysis: Further issues Wooldridge, Introductory Econometrics, 4th ed. Chapter 6: Multiple regression analysis: Further issues What effects will the scale of the X and y variables have upon multiple regression? The coefficients

More information

Chapter 11. Regression with a Binary Dependent Variable

Chapter 11. Regression with a Binary Dependent Variable Chapter 11 Regression with a Binary Dependent Variable 2 Regression with a Binary Dependent Variable (SW Chapter 11) So far the dependent variable (Y) has been continuous: district-wide average test score

More information

East Bay BRT. Planning for Bus Rapid Transit

East Bay BRT. Planning for Bus Rapid Transit East Bay BRT Planning for Bus Rapid Transit Regional Vision Draper Prison The Bottleneck is a State-Level issue, Salt Lake County 2050 Population: 1.5M Draper Prison hopefully with some State-Level funding!

More information

Contents. Ipswich City Council Ipswich Adopted Infrastructure Charges Resolution (No. 1) Page

Contents. Ipswich City Council Ipswich Adopted Infrastructure Charges Resolution (No. 1) Page Ipswich City Council Ipswich Adopted Infrastructure Charges Resolution (No. 1) 2014 Contents Page Part 1 Introduction 3 1. Short title 3 2. Commencement 3 3. Sustainable Planning Act 2009 3 4. Purpose

More information

Linear Models in Econometrics

Linear Models in Econometrics Linear Models in Econometrics Nicky Grant At the most fundamental level econometrics is the development of statistical techniques suited primarily to answering economic questions and testing economic theories.

More information

Firm-sponsored Training and Poaching Externalities in Regional Labor Markets

Firm-sponsored Training and Poaching Externalities in Regional Labor Markets Firm-sponsored Training and Poaching Externalities in Regional Labor Markets Samuel Muehlemann University of Berne & IZA Bonn Intl. Conference on Economics of Education, Firm Behaviour and Training Policies

More information

ADDRESSING TITLE VI AND ENVIRONMENTAL JUSTICE IN LONG-RANGE TRANSPORTATION PLANS

ADDRESSING TITLE VI AND ENVIRONMENTAL JUSTICE IN LONG-RANGE TRANSPORTATION PLANS ADDRESSING TITLE VI AND ENVIRONMENTAL JUSTICE IN LONG-RANGE TRANSPORTATION PLANS Activities from the National Capital Region Transportation Planning Board Sergio Ritacco Transportation Planner 2017 Association

More information

TRANSPORT MODE CHOICE AND COMMUTING TO UNIVERSITY: A MULTINOMIAL APPROACH

TRANSPORT MODE CHOICE AND COMMUTING TO UNIVERSITY: A MULTINOMIAL APPROACH TRANSPORT MODE CHOICE AND COMMUTING TO UNIVERSITY: A MULTINOMIAL APPROACH Daniele Grechi grechi.daniele@uninsubria.it Elena Maggi elena.maggi@uninsubria.it Daniele Crotti daniele.crotti@uninsubria.it SIET

More information

Electronic Research Archive of Blekinge Institute of Technology

Electronic Research Archive of Blekinge Institute of Technology Electronic Research Archive of Blekinge Institute of Technology http://www.bth.se/fou/ This is an author produced version of a ournal paper. The paper has been peer-reviewed but may not include the final

More information