United StatesDepartment ofAgriculture

Forest Service

Southeastern ForestExperiment Station

Research PaperSE-275

Theory and Techniques for Assessingthe Demand and Supply of OutdoorRecreation in the United States

The Authors:

H. Ken Cordell is Project Leader and National AssessmentSpecialist, Outdoor Recreation and Wilderness Research,Southeastern Forest Experiment Station, USDA Forest Service,Athens, GA. John C. Bergstrom is Assistant Professor,Department of Agricultural Economics, University of Georgia,Athens, GA.

Acknowledgments :

We thank Greg Ashley and Carter Betz of the study team forhighly professional analytical support, Don English for therecreation opportunity indices, and Larry Hartmann forcontributions toward development of the PARVS data set. Weespecially acknowledge the PARVS Working Group which isresponsible for the creation and implementation of the PARVSnational research effort. Very helpful review comments from DanMcCollum (Forest Service, RM), Ted McConnell (University ofMaryland), Dick Walsh (Colorado State University), and DanStynes (Michigan State University) are also gratefullyacknowledged. Any remaining errors or omissions are theresponsibility of the authors. This research represents more than 6years of effort on the part of many. Funding and guidance fromthe Programs and Assessment staff of the Forest Service providedinvaluable support for completion of the project.

September 1989

Southeastern Forest Experiment StationP.O. Box 2660

Asheville, North Carolina 26602

Abstract

Modeling and estimating consumption, demand, supply, values, scarcity, future equilibria,and price changes have been major challenges in comprehensive recreation assessment andplanning. The results of 6 years of work reported in this Paper show research advancementstoward meeting these challenges. Accomplished as the central analysis for the 1989Renewable Resources Planning Act Assessment, a household market model covering 37recreational activities was computed for the United States. Equilibrium consumption andcosts were estimated, as were likely future changes in consumption and costs in response toexpected demand growth and alternative development and access policies. Except forconsumptive wildlife and fishing activities, most outdoor recreation demand and supplyequilibria will grow to much higher levels of consumption in future years. For mostactivities, aggressive expansion of resource access and site development to provide morerecreation opportunities will be necessary if rapid increases of opportunity scarcity are to beavoided.

Keywords: Outdoor recreation, market equilibrium, demand, supply, forecasts, householdproduction theory.

Summary

A conceptual model of outdoor recreation demand,supply, and equilibrium consumption and costs isdescribed in this Paper. This model is based oncommunity-level, household demand and supplyfunctions. The aggregate demand function shows totalannual trips for an activity that a community willdemand at various trip costs. The aggregate supplycurve, based on household production theory, showstotal annual trips for an activity that a community canproduce at various trip costs, given the existingavailability of opportunities or destinations.

Equilibrium trip consumption and costs are definedfrom a community-level household or consumerperspective. Equilibrium trip consumption refers to thenumber of trips consumed when aggregate demand andsupply are equal. Equilibrium trip costs are the costs orprice of a trip where the number of trips demanded by acommunity is equal to the number of trips available to acommunity. At the household equilibrium point, themarginal benefits of trip consumption are equal to themarginal costs to households of trip production.

Equilibrium consumption and costs were estimated forseveral demand/supply scenarios. These scenariosassumed moderate demand growth, combined witheither negative, zero, moderate, or high growth ofavailable facilities and resources. For a typicalcommunity, equilibrium consumption of mostrecreational activities was projected to increase under alldemand/supply scenarios to the year 2040. Notableexceptions included big-game hunting, small-gamehunting, and warm-water fishing. The results, in

general, indicate that equilibrium consumption is quitesensitive to changes in recreational facility and resourcegrowth rates.

The interaction of outdoor recreation demand andsupply over time is summarized in the form of changesin market equilibrium trip costs. Increases inequilibrium trip costs indicate that demand is increasingfaster than supply and that outdoor recreationalopportunities are becoming more scarce. Decreases inequilibrium trip costs indicate that supply is increasingfaster than demand and that outdoor recreationalopportunities are becoming more abundant. Constantequilibrium trip costs indicate that demand and supplyare increasing or decreasing at the same rate and thatthe scarcity of recreational opportunities is remainingabout the same over time.

By summarizing the balance between demand andsupply, market equilibrium trip costs provide a broadindicator of the relative scarcity of recreationalopportunities over space and time. The need for abroad recreation scarcity indicator has also beenadvocated in the recreation literature (Clawson 1984;Cordell and English 1989; Harrington 1987).Equilibrium trip costs were applied in this Paper toassess the impact of public policy pertaining torecreational facility and resource growth on theprovision of recreational opportunities. Results indicatethat for most activities, a moderate to high rate of publicfacility and resource growth will be needed to satisfyfuture demand and to prevent increases in the relativescarcity of recreational opportunities.

Theory and Techniques for Assessingthe Demand and Supply ofOutdoor Recreation in the United States

H. Ken Cordell John C. Bergstrom

CONTENTS

The Purpose for This Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Recreational Trip Demand and Supply . . . . . . . . . . . . . . . . . . . . . . . . 1Background Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1Supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Household Market for Recreational Trips . . . . . . . . . . . . . . . . . . . . . . 4Trip Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Consumption Trend Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . zConsumption Trend Lines Under Alternative Scenarios . . . . . . . . 6Equilibrium Trip Costs or Price . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Empirical Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . aTrip Consumption and Cost Data . . . . . . . . . . . . . . . . . . . . . . . . 8Defining the Representative Communities . . . . . . . . . . . . . . . . . . 9Resource and Facility Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9Other Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9Organization of the Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9Demand Curve Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10Consumption Function Estimation . . . . . . . . . . . . . . . . . . . . . . . . 11Household Market Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Research Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . l.3

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . l3

Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

The Purpose for This Research

The Renewable Resources Planning Act (RPA) of 1974requires that the USDA Forest Service comprehensivelyassess demands for and supplies of forest and rangeresources at lo-year intervals. The next assessmentreport will be published for release in 1989. The RPArequires estimates of demand and supply trends for anumber of recreational settings, resources, and activities.Previous recreation assessments were limited byinadequate data, theories, and methods. This Paperreports advancements in all three of these dimensions ofthe assessment process, leading to a capability to modeland forecast within a market equilibrium framework.This framework provides a conceptually strong, unique,and comprehensive means for resource assessmentconsistent with RPA research, planning, and policythrusts.

The first section provides a discussion of community-level household demand and supply. A conceptualmodel of household equilibrium consumption and costsfor recreational trips is then presented. This conceptualmodel shows that equilibrium consumption and costs canbe estimated from aggregate demand and consumptionfunctions. Empirical estimation of the aggregatedemand functions, aggregate consumption functions,equilibrium consumption, and equilibrium costs are thendiscussed.

Recreational Trip Demand and Supply

Background Considerations

Demand for and supply of outdoor recreation areelusive concepts. Most outdoor recreational activitiesare provided as public goods, for which the marketsituation is quite different from that for other forestproducts. Most other forest products are private goods,like timber. Since market-determined recreationalopportunities as conceptualized by classical economicsare nonexistent, outdoor recreation demand and supplymust be assessed using extra market techniques.Outdoor recreation demand and supply analysis isfurther complicated by two other problems. Fist, it isdifficult to define exactly what is being demanded andhow to measure recreation demand. Second, once ademand quantity measure has been selected, definingtheoretically appropriate demand and supply curvestends to be very problematic.

For the RPA analysis herein reported, recreational tripswere considered to be the most conceptually appropriatemeasure of quantity consumed (McConnell 1975). Atrip is defined as a purposeful commitment of time awayfrom home to travel to a destination(s) for enjoyment of

a particular type of recreational opportunity. Measuringthe demand for and supply of recreational trips isconsistent with the RPA goal of assessing the overallavailability of recreational opportunities to the nation,rather than the demand for a particular site or facility bya particular segment of the population. For the latterpurpose, a visit or an estimate of the amount ofrecreation participation time is the appropriate measure.National data on recreational trip consumption wereavailable for this research from the Public AreaRecreation Visitors Study (PARVS).

Demand

Recreation demand functions are generally of twotypes--aggregate or individual. Aggregate demandfunctions provide the basis for the zonal travel costmethod (ZTCM). In the ZTCM approach, as usuallyapplied, recreationists are grouped into zones around asite. Demand curves are derived by estimating thestatistical relationship between aggregate trips from azone and the distance from the zone to the site(Clawson and Knetsch 1966; Dwyer and others 1977;Ward and Loomis 1986).

A major criticism of the ZTCM is the loss of detail onrecreation demand behavior which results fromaggregating individuals and trips within zones. Amethod developed for retaining detailed information isthe individual travel cost method (ITCM) in which theunit of observation is an individual’s consumption oftrips. ITCM demand curves are derived by estimatingthe statistical relationship between an individual’s tripsand the distance traveled from residence to recreationalsite (Brown and Nawas 1973; Gum and Martin 1975;Ward and Loomis 1986).

The choice between ZTCM or the ITCM is dependenton several considerations. For one, the ZTCM is oftenselected because less data are required. A furtheradvantage is that it adjusts for both probability andfrequency of participation using a single equation thatcan be estimated by ordinary ieast squares. Severalmore complex statistical procedures have beendeveloped as adjustments for both the probability andfrequency of participation when the ITCM is chosen.Much debate remains, however, over the appropriate-ness of these adjustments (Walsh 1986; Ward andLoomis 1986).

There are also several perceived disadvantages of theZTCM which have led to increased use of the ITCM in

recent years. First, when aggregation is selected as theappropriate analysis objective, there is difficulty in

1

reconciling its results with utility theoretic models ofindividual choice. Also, the ZTCM assumes thatindividual characteristics and participation rates areconstant within zones. These assumptions probably arenot realistic (Bockstael and McConnell 1981). Also, theZTCM eliminates within-zone variation insocioeconomic characteristics and trip consumptionbehavior. Hence, it may be difficult to estimatestatistically significant relationships between individualsocioeconomic variables and the quantity of tripsconsumed (Brown and Nawas 1973; Rosenthal 1985;Ward and Loomis 1986).

In general, if the objective of an analysis is to explainindividual consumption behavior, the ITCM is likely tobe more appropriate and efficient than the ZTCM. Byfocusing on individual observations, the ITCM approachallows for more statistically robust and theoreticallyconsistent analyses of individual recreation consumptionbehavior. The primary disadvantages of the ITCM arelarge data requirements and complicated estimationprocedures (Ward and Loomis 1986).

Although consideration of the relationships betweenindividual socioeconomic characteristics and con-sumption behavior is important, the primary objective ofthe RPA Assessment, for which this research was done,is to analyze aggregate relationships and trends innational outdoor recreation demand and supply. Esti-mation of aggregate demand and consumption functionsusing the ZTCM is appropriate to this overall objective.An additional consideration used to select ZTCM as themore appropriate modeling procedure was that the datacollected for the RPA assessment through PARVS aremore compatible with an aggregate model. A finalconsideration was that statistical implementationprocedures, particularly with respect to adjustments forprobability of participation, are well established in theZTCM. The ZTCM has been successfully employed inhundreds of applications over the last three decades(Walsh 1986; Ward and Loomis 1986).

For the RPA Assessment, it was necessary to modify theclassical ZTCM to account for regional recreationalopportunities. This modification was accomplished witha region-based zonal travel cost method for multipledestination opportunities (RZTCM) (Sorg and others1985). RZTCM is designed to model trips for anactivity from a zone to all sites used by the population inthat zone. For the RPA Assessment, zones weredefined as counties, each assumed to be a homogeneouscommunity. The modeling task was therefore to developaggregate demand and consumption functions for tripsto all sites used by a county population for a particularactivity.

Community-level (county) recreation trip-consumptionbehavior is influenced by trip costs, communitypopulation characteristics, and the availability anddiversity of suitable recreational opportunities.Community characteristics, such as the percentage ofhouseholds of middle-income status, affect theproportion of the community population that participates(analogous to probability) and the frequency ofconsumption of outdoor recreational trips. Availabilityof suitable recreational opportunities throughout theregion determines the cost of recreational trips and thequantity of substitutes. Trip costs and substitutes affectnot only the frequency of participation but also theprobability of participation. Community-level recreationdemand behavior was modeled for the RPA Assessmentwith the following specification:

ATRIPSD = f(P, S, SO, Z, H) , (1)

where

ATRIPSD = annual trips demanded for activity k by acommunity,

P = cost or price of trips for activity k,S = suitability of sites used for activity k,

SO = substitute recreation opportunitiesavailable to a community,

Z = community population 12 years old andolder,

H = community characteristics.

Figure 1 illustrates a community-level demand functionfor trips for activity k. This demand function holds sitesuitability (a crude measure of quality), substituterecreational opportunities, county population size, andcounty population characteristics constant. Thus, itshows the total number of trips a community willdemand at various trip costs. For example, at a cost of$60 per trip, 100,000 trips are demanded. Trip costsinclude both out-of-pocket travel expenditures and theopportunity cost of time used for travel for a two-waytrip.

* Includes trovel and time costs for o two-way trip

Figure l-Activity k aggregate demand curve for a typical community.

2

SUPPlY

Outdoor recreation trips cannot be purchased as mostgoods can at a local shopping center. Rather, trips mustbe “produced,” a process carried out by the recreationiststhemselves. Recreationists for example, may combinegasoline, time, equipment, supplies, and recreationalfacilities to produce a trip for the activity of hiking. Theconceptualization of recreationists as producers of tripsis derived from household production theory (Becker1%5; Bockstael and McConnell 1981).

As travel distance or time is increased, morerecreational opportunities are available to a communityand the number of trips which can be produced orsupplied increases. The number of trips that can beproduced at a given trip cost is determined by theavailability of recreational opportunities; knowledge,skills, and abilities (KSAs) of households in thecommunity; and technology available to households, suchas light-weight backpacking gear. In the absence ofmeasures of KSAs, household characteristics are used asproxies. Trip production was modeled for the RPAAssessment with an aggregate household supply functionconceptually specified as:

ATRIPS’ = h(P, RO, H) ,

where

(2)

ATRIPSS = annual number of trips supplied for activityk by a community,

P = cost or price of trips for activity k,RO, = recreational opportunities available to a

community for activity k,H = household characteristics affecting trip

production technology.

Equation (2) is a modified version of Bockstael andMcConnell’s supply or marginal cost function for wildliferecreation which focused on trip production byindividuals. In contrast to Bockstael and McConnell’smodels, equation (2) focuses on total annual tripproduction by a community and is more general in thatit applies to many other activities, in addition to wildliferecreation. This community production perspective isconsistent with the conceptual analysis of aggregateoutdoor recreation supply conducted by Harrington(1987). Equation (2) describes trip supply for a giventime period (e.g., 1 year), during which P and ROk areassumed to be exogenous variables.

In equation (2), recreational opportunities (ROk) referto specific sites and facilities. Amount and location ofthese recreational opportunities are fmed in the shortrun. This fured amount and location of recreational

opportunities are expressed conceptually by theequation:

R”k = &MC, B, M) ,

where

(3)

LMCk = long-rM marginal COStS Of providingrecreational sites used as inputs for activity k,as provided by public and private resourcemanagers,

B = annual recreation budget available toresource managers,

M = current management policies.

Equation (3) implies that availability of recreationalopportunities is determined by the costs of providingrecreational sites (such as expenditures on new publicfacilities), an exogenously determined budget (such ascongressional appropriations), and current attitudes andpolicies (such as attitudes and policies regarding newland acquisitions) (Hof and Kaiser 1983). The variable,RO, , is an important variable. A change in publicmanagement policies, for example, may change RO,. Achange in ROk , in turn, will impact the ability of acommunity to produce recreational trips throughequation (2).

Aggregate household supply functions for activity k areillustrated by the curves labeled S, and S, in figure 2.Each curve represents a different fured amount ofrecreational opportunities for activity k. The curvelabeled St indicates that at an average trip cost of $60, ahypothetical community can produce 75,ooO trips.Suppose a change in government management policiesincreases the availability of recreational opportunities,RO,. This increase will cause the aggregate householdsupply curve to shift from St to S,. Given S,, a typicalcommunity can now produce 150,000 trips at an averagetrip cost of !$60.

v Includes travel ond time costs for o two-way trip

Figure 2-Activity k aggregate supply curves for hvo levels ofavailable opportunities for a typical community.

Derivation of an aggregate household supply curve isdemonstrated further by the diagrams in figure 3. Theblack dot in the center of figure 3a represents thelocation of a community surrounded by six recreationalsites represented by squares numbered l-6. Each site isassumed to have the capacity to accommodate 100 tripsper year for activity k. In order to produce more tripsthan those which the closest site can accommodate,community residents must travel to sites farther away. Ifresidents travel up to 25 miles, they can produce a totalof 200 trips annually (100 trips each to sites 1 and 2). Ifresidents travel up to 50 miles, they are able to producea total of 400 trips annually (100 trips each to sites 1, 2,3, and 4). At distances up to 100 miles, residents canproduce a total of 600 trips annually (100 trips each tosites 1, 2, 3, 4, 5, and 6).

traveling to more distant sites. If a large number ofsites were located at continuous distances from thecommunity, the staircase-shaped curve in figure 3bwould become the smooth, upward sloping aggregatesupply curve shown in figure 2. Changes in the number,capacity, or location of sites shift the aggregate supplyfunction. For example, if additional sites were providedbetween 25 and 100 miles of the community shown infigure 3a, the aggregate supply curve would shift to theright, such as from S, to S, in figure 2. Conversely, ifsites or site capacity were reduced between 25 and 100miles of the community, shown in figure 3a, theaggregate supply curve would shift to the left, such asfrom S, to St in figure 2.

Household Market for Recreational Trips

Trip Consumption

In the context of the Bockstael and McConnell (1981)household production model, equilibrium consumptionand costs are defined by the household market demandand supply curves illustrated in figure 4.

04

Figure 4-Community-level household market for activity k trips.

Figure 3-Derivation of activity k aggregate supply curve for a typicalcommunity.

The numbers of trips that can be produced at variousdistances by the community, shown in figure 3a, areindicated by the staircase-shaped aggregate supply curvein figure 3b. Figure 3b assumes, for convenience, thateach additional mile of travel costs exactly one moredollar. This “staircase” supply function assumes that thecommunity first exhausts capacity at closer sites before

As indicated by the demand curve, at a cost of P pertrip, households in the community desire and wd4consume Q1 trips. As indicated by the aggregate supplycurve, at a cost of P, per trip, households in thecommunity are able to produce Qt trips. Thus, thequantity of trips households in a community would liketo consume is equal to the quantity of trips they canproduce. From the community’s perspective, aggregatedemand and supply are in equilibrium. At theequilibrium number of trips, the marginal costs to thecommunity of producing trips are equal to the marginalbenefits to them of consuming trips. This equilibrium isin the nature of a general equilibrium as the aggregatedemand and supply curves account for overallrecreational opportunities across multiple sites.

4

The intersection of these aggregate demand and supplycurves defines the current number of activity k tripsconsumed, Qr, and the cost of these trips, P, andrepresents a community-level household marketequilibrium.

The existence of a household market equilibriumassumes that the community is not restricted toproducing and consuming trips only at local sites.Restriction to local sites would be the case, for example,if the community were on an island and all feasiblerecreational sites were located on that island. In thisisland situation, the aggregate trip supply curve would betruncated at the point corresponding to the maximumcapacity of sites at any one time. At maximum capacity,the aggregate supply curve effectively turns vertical,indicating that no more trips can be produced using theexisting stock of sites (Cicchetti 1973; Loomis and Hof1985). This situation is illustrated in figure 5.

Figure S-Community-level household market for activity k trips(alternative case) with truncated aggregate supply function.

If the aggregate demand curve intersects the effectiveaggregate supply curve in this vertical segment (point Bin figure 5) and if negligible congestion costs areassumed, a household market disequilibrium results.Because of capacity constraints, trip consumption isrationed at Q1 trips. The marginal cost of producingthese trips is equal to P, per trip. Given the aggregatedemand curve in figure 5 and a marginal cost of P, pertrip, community households would desire to consume Qztrips. Thus, there is a shortage of trips equal to Q2-Q1.

The limited number of trips denoted by Q1 are usuallyrationed by nonprice mechanisms (such as first come,fast served or reservations). If trip costs were increasedto Pz, however, the limited number of trips would berationed by the cost or price of a trip. One way ofincreasing trip costs to Pz is to charge an entrance feeequal to P,-P, per trip. This fee would shift theeffective aggregate supply curve from OAS to OBS, andresult in a new household market equilibrium at point Bin figure 5.

In the more common case where a community is not onan island, local site capacity does not restrictconsumable trips to effectively truncate the aggregatesupply curve. While it is assumed that local site capacityis used up first, recreationists are free to move intoother regions to seek other destinations until the point isreached where the marginal benefits and costs ofadditional trips produced and consumed are equal. Thenumber of trips consumed is illustrated by the householdmarket equilibrium in figure 4. If the cost of producingmore trips using more distant sites increases rapidly, thehousehold market equilibrium may occur at relativelyfew trips and high costs per trip.

Consumption Trend Lines

A further objective of the RPA Assessment is tocompare demand and supply of recreationalopportunities at future times. Comparisons over timeare made using consumption trend lines. The totalnumber of trips consumed by a community isdetermined by the intersection of the aggregate demandcurve shown in equation (1) and figure 1 and theaggregate supply curve in equation (2) and figure 2.This intersection, illustrated in figure 5, represents thesolution to the simultaneous equation system given byequations (1) and (2). The solution to thisdemand/supply simultaneous equation system can beestimated from the reduced form of the system:

ATRIPS = g (SO, Z, S, RO,,H) ,

where

(4)

ATRIPS = annual number of trips for activity kconsumed by a community.

The derivation of a consumption trend line is illustratedgraphically in figure 6. In panel (a) of figure 6, the baselevel of consumption at a defined point in time is shownby the intersection point a on the vertical axis, Q1 trips.Point a corresponds to the household marketequilibrium point A in panel (b). At point A, Q1 tripsare consumed at costs of P, per trip.

Assume next that by the year 2000, demand for trips isexpected to increase from D to D, in panel (b). Giventhis new demand curve, the community will desire toconsume Qz trips at the current trip costs of P,.Consumption level Q2 is referred to as “preferreddemand.” Preferred demand is defined generally as thenumber of outdoor recreational trips a community wouldtake if the cost or price of a trip remained what it is

today. Preferred demand in 2080 is represented bypoint B in panel (b), which corresponds to point b onthe line labeled Lt in panel (a). Given the aggregatetrip supply function labeled S in panel (b), however, onlyQ1 trips can be produced and consumed at a cost of P,per trip. This constant level of trips over time isrepresented by point A in panel (b), which correspondsto point e on the line labeled La in panel (a).

p2

Pl

(a)

Ql Qs Q,Trips

Figure 6-Conceptual derivation of consumption trend lines.

In an effort to reach a household market equilibrium, itis assumed that recreationists will incur increased coststo produce more trips. These trips will be producedalong the aggregate supply curve in figure 6, panel (b).Increased trip production, for instance, may involvetravel to more distant sites and increased time spentsearching for available facilities at more local sites.Increased trip production will continue until themarginal benefits and costs of trip production andconsumption are equal at point F in panel (b). At thishousehold market equilibrium, Q3 trips are consumed atcosts equal to P per trip. Point F in panel (b) corre-sponds to point 1 on the line labeled 5 in panel (a).

In a household market for recreational trips, it isassumed that temporary trip shortages caused byincreases in demand but no increase of sites andfacilities will be eliminated by increased trip productionwith an associated increase in trip costs as more distantsites are accessed. The number of trips consumed aftertrip production and demand have equilibrated at highertrip costs is mapped out by the line labeled % in figure6. Lz 1s an example of a consumption trend hneshowmg the number of recreational trips consumed overtime with community-level household market for tripsadjusting to new equilibria at every future point in time.Notice that there is still a gap between preferreddemand mapped by L and expected trip production andconsumption mapped by b. This gap is given by thedistance Q2 - Q3 in figure 6b and the distance b-f infigure 6a.

Consumption Trend Lines Under Alternative Scenarios

Consumption trend lines under three alternative supplyscenarios are shown in panel (a) of figure 7. Therelationship of these consumption trend lines to thecommunity-level household market for trips is illustratedin panel (b) of figure 7. The base year level of trips(1987 for RPA) is given by point b in panel (a), Qt trips.Point b in panel (a) corresponds to point B in panel (b).At point B, Qt trips are consumed at costs or priceequal to P, per trip. Suppose demand for trips isexpected to increase by the year 2000 from D to D, inpanel (b). Also, assume that available recreationalfacilities and resources (RO) will be held constant.Under this assumption, the aggregate supply curve fortrips is given by the curve labeled S, in panel (b). Withthis aggregate supply curve and the new aggregatedemand curve given by D,, the community will establisha new household market equilibrium at point E in panel(b). At point E, Qz trips are consumed at trip costs orprice equal to P,. Point E in panel (b) corresponds topoint e on the consumption trend line labeled C, inpanel (a).

Next, assume again that demand is expected to shiftfrom D to D, by the year 2000, but that recreationalfacilities and resources available to the public are alsoincreased, from RO, to RO,, causing the aggregatesupply curve to shii out from St to S, in panel (b).Given these new demand and supply curves, thecommunity will establish a new household marketequilibrium at point F in panel (b). At point F, Q, tripsare consumed at trip costs equal to P3. Point F in panel(b) corresponds to point f on the higher consumptiontrend line labeled C, in panel (a).

Finally, assume once more that demand will shift fromD to D, by the year 2000. Suppose, however, thatrecreational facilities and resources will be reduced fromRO, to R03 causing the aggregate supply curve to shiftfrom S, to Ss in panel (b). Given these new demandand supply curves, the community will establish a newhousehold market equilibrium at point G in panel (b).At point G, Q, trips are consumed at trip costs equal toPd. Point G in panel (b) corresponds to point g on thelower consumption trend line labeled C, in panel (a).

TrpeQa

Q2

Q4

Q 1 b

Figure 7-Consumption trend lines under alternative demand/supplyscenarios.

As illustrated in figure 7, different demand or supplyassumptions will result in diierent consumption trendlines. These lines can be used to demonstrate theimpact of alternative policies affecting the amount ofavailable recreational facilities and resources. Forexample, given an expected increase in demand from Dto D,, the consumption trend lines in panel (b) of figure7 illustrate the impact on trip consumption and costs ofthree alternative supply scenarios. Consumption trendline C, shows trip consumption and costs, assuming thatRO

illquantity of recreational facilities and resources are

ava able through the year 2WO. Consumption trend lineC, shows trip consumption and costs at a relatively

higher quantity of recreational facilities and resources,RO,. Consumption trend line C, shows trip consump-tion and costs at a lower quantity of recreationalfacilities and resources, ROs, available without changethrough 2ooo.

Equilibrium Trip Costs or Price

Changes in equilibrium trip costs or price determined bythe household market indicate the relationship betweenchanges in recreational demand and supply over time.Assume that the consumption trend lines shown infigure 7 are derived from equation (4), the consumptionfunction for trips for activity k. Using this consumptionfunction, it is estimated that under a scenario ofincreased demand from D to D, and quantity RO, ofrecreational facilities and resources, consumption of tripsfor activity k will increase from QI in the base year toQ, in the year 2000. Q3 trips are mdicated by point eon consumption trend hne C, in panel (a) of figure 7.

Point e in panel (a) of figure 7 corresponds to house-hold market equilibrium point E in panel (b) of figure 7.Hence, the household market trip costs or priceassociated with Q trips are determined by substitutingQz trips into the 14 emand function for year 2000, derivedfrom equation (l), and then solving for trip costs. Thesolution is trip costs equal to P2. Under a scenario ofincreased demand and a stable quantity of recreationalfacilities and resources into the future of Rot,household market trip costs or price would increasefrom P, to P,

Another possibility is that demand will increase, but thatrecreational facilities and resources will be less in thefuture. Under this scenario, consumption of trips foractivity k will increase to Q4 as shown in panel (b) offigure 7. Q4 trips is indicated by point g on consump-tion trend line C, in panel (a) and the household marketequilibrium point G in panel (b). The equilibrium tripcosts associated with Q4 are estimated by substitutingthis quantity of trips into the demand function for theyear 2000 and solving for trip costs. The solution is tripcosts equal to P,. Thus, if demand increases andavailable recreational facilities and resources decrease toquantity RO, in the future, market trip costs willincrease sharply from P, to Pd.

Next, assume that it is estimated from the consumptionfunction that increased demand and a higher quantity ofrecreational resources and facilities will result inincreased consumption of trips for activity k to Qs in theyear 2000. Q, number of trips is indicated by pomt f onconsumption trend line C, in panel (a) of figure 7, andthe household equilibrium point F in panel (b). House-

7

hold market trip costs associated with QJ would, there-fore, be estimated by substituting Q3 into the year 2000demand function and solving for trip costs. The solutionis market trip costs equal to P3. Thus, if futurerecreational facilities and resources were increased to ahigher level, household market trip costs would beexpected to decrease from P, to P3.

Over several future points in time, an increase inequilibrium trip costs suggests that demand is increasingfaster than supply. A decrease in equilibrium trip costssuggests that supply is increasing faster than demand.Constant equilibrium trip costs over time indicate thatdemand and supply are increasing or decreasing at thesame rate, or that neither is changing. Thus, changes inequilibrium trip costs provide a convenient means forsummarizing the relationships between changes inrecreational demand and supply over time and provide ameasure of adequacy of trends in making resources,facilities and services available for public recreationaluses.

Empirical Approach

To carry out a national analysis of the magnitudeimplied by the preceding model, and to meet the RPAAssessment mandate, extensive data and analysis wereneeded to describe the factors identified in the demand,supply, and consumption equations. The principalsource of the data on number of trips, trip costs, andconsumer characteristics was the Public Area RecreationVisitors Study, PARVS. The National OutdoorRecreation Supply Information System, NORSE,provided data on resources and facilities available forproduction of recreational trips. These two nationaldata sets were each aggregated. to the community(county) level for this analysis, and a sample of 239representative counties was used to construct cross-sectional models.

Trip Consumption and Cost Data

The PARVS data set was developed through a cooper-ative effort of 17 State and Federal agencies from thesummer of 1985 through the fall and winter of 1987.These data were assumed to be homogeneous acrosstime with respect to equilibrium levels of consumption,and therefore were assumed to represent consumptionduring a single year, 1987. PARVS trip consumptiondata were developed through onsite interviews of visitorsat over 280 recreation sites across the country. The totalnumber of interviews conducted during the above periodwas over 32,000, nearly 26,000 of which were usable forthis analysis, having neither recording errors nor missingvalues for questionnaire items needed for modeling.

The PARVS questionnaire contains data identifying therecreational activity that was the main reason for visitingthe destination site, location of origin, trip costs(distance and time), profile of number of trips by activitytaken during the last 12 months prior to the interview(including the current trip), and the respondent’shousehold characteristics. Origins were recorded asboth county name and Zip Code, thus enabling identi-fication of place of residence. From this identification, amap plotting of Zip Codes and longitude/latitudes indi-cated that respondents represented almost 80 percent ofthe counties in the country. Counties not representedwere mostly the very sparsely populated ones in theMidwest and several others which were mostlycomposed of public land, primarily in the West.

The principal dependent variable for empiricalestimation of demand and consumption models wasannual number of trips by residents in the 239 repre-sentative counties. Computation of total trips byresidents of these counties was accomplished by weight-ing each respondent’s reported trips, costs, and char-acteristics such that the percentages of respondentsacross 32 socioeconomic strata (representing age, race,gender, and rural-urban residence) were proportionatelyadjusted to match the socioeconomic makeup of partic-ipant respondents to the 1983 National RecreationSurvey (NRS). The NRS was designed by the Bureau ofCensus and weighted to reflect the most current profileof the United States population and thus could serve asa base for weighting the PARVS data. Thus thePARVS records, each representing individual-level trips,were weighted to proportionately represent the U.S.population’s profile of trips by activity. These weightedindividual trip totals were then aggregated to thecommunity (county) level and multiplied by the ratio ofnumber of participants in outdoor recreation in thecommunity to the weighted number of PARVS respon-dents in the community. This same weighting, and asappropriate, extrapolations, were performed for allvariables in the PARVS data set.

Trip costs, described in more detail later, weredeveloped by combining transportation costs to travelthe distances adjusted to reflect route circuity from eachorigin to each destination with the wage value of traveltime. The total number of origin-by-destinationcombinations with relevant travel distances for theactivities considered was over 7,200. Relevant traveldistances were computed across reported single desti-nation trips as the distance at the 95th percentile,separately developed for each activity. All sites onwhich PARVS interviewing had been conducted andwhich lay within these relevant-travel-distance radiicomposed the set of representative trip destinations forpeople living in the 239 representative origins

8

(communities equivalent to counties). Trips and tripcosts between this set of representative origins anddestinations formed the consumption and priceobservations for this analysis.

Defining the Representative Communities

Respondents to the PARVS survey represented almost80 percent of the counties in the United States, exceptfor those which are very sparsely populated. Subregionswithin the four major assessment regions (North, South,Rocky Mountains, and Pacific Coast) were identified bycombining counties with similar rural-urban proportionsand physiographic characteristics and lying adjacent toone another. To assure sufficiency in number ofPARVS respondents per subregion, counties withPARVS respondents were added until a minimum of 90respondents was achieved. These subregions, therefore,did not overlap, and counties within them were com-bined in such a manner as to result in a block or circleof adjacent counties, rather than a strip or band.Selection of 90 as the minimum number of respondentswas based on a preliminary analysis to identify theminimum sample size needed to achieve reasonable sta-bility of standard deviations about trip consumption andtravel means. Within each of the 239 thusly definedsubregions (excluding Alaska and Hawaii), the onecounty nearest the center which had a high percentageof the subregion’s respondents was selected as the singleorigin to represent the subregion. All respondents, tripconsumption, and travel were assumed to originate fromthis central county (community). The maps in figure 8show the locations of these representative communities.All data aggregations and descriptions of residentpopulations used in this study pertain to these 239counties.

Figure &Location of the 239 representative communities (counties)used estimating national models.

Resource and Facility Data

The NORSIS data set includes primary- and secondary-source data describing over 400 different dimensions ofpublic and private areas, facilities, and services availablefor or servicing public recreational uses. Secondary-source data ranged in vintage from 1982 to 1987, alwaysusing the most recent available source. Primary datawere collected in 1986 and included national studies ofprivate lands and municipal and county governments.The coverage of the NORSIS data set includedwilderness and other remote wild lands to the mosthighly developed and accessible of resorts, theme parks,and high-visitation facilities, including urban parks.

The NORSIS, aggregated at the county level andrepresenting all counties in the United States, provideddata describing opportunities directly relevant to theactivities modeled in this research, as well as an index ofsubstitute opportunities. An example of direct opportu-nities includes Federal, State, local, and private campingsites. The substitute indices were computed usingapproximately the method proposed by Clawson (1984)and described by Cordell and others (in press). Thesesubstitute indices measured comprehensively themagnitude of “other” opportunities available within adistance that community residents were willing to travel,adjusted to reflect (1) how many other people werecompeting for use of these opportunities and (2) howdistant they were from the community’s geographiccenter. This distance was adjusted with a straight-linedistance-decay function to down weight the more distantopportunities (Cordell and others, in press).

Other Data

Data describing household characteristics for each of the239 representative communities were obtained from theBureau of the Census County Data File. These datareflected the most recent census, 1980. Measures of therecreational suitabilities of each destination site wereobtained through a survey of site managers, includingState and national parks, reservoirs, State and nationalforests, and wildlife management areas. Suitabilities foreach of 16 recreational activity categories were scoredon a scale of 0 to 10 by each manager for their own site.In developing models for any one activity, sites rated asnot suitable were excluded as destinations; for example,a State forest with no reservoir or river was rated atzero suitability for water-skiing.

Organization of the Data

All data were organized around the 239 representativeU.S. communities. Trips were aggregated to express

countywide totals of trips for each of the 37 activities forwhich there was a sufficient number of respondentsinterviewed at PARVS sample sites to provide statisti-cally reliable estimates of distance and time of travel.Activity specific recreational opportunities and substituteopportunities for these communities represented the fullset of opportunities within relevant travel radii of eachcommunity centroid, typically including,adjacent, andsometimes quite distant, other counties.

The resulting final data set provided a 239 by 310 matrixof data values. Each community was identified by theFederal Information Processing Standards (FIPS) codeand longitude and latitude of its centroid. All data weretriple checked for accuracy of entry.

Demand Curve Estimation

Trips for activity k are expected to vary acrosscommunities because of differences in costs of producingtrips; in availabilities of opportunities; in suitabilities ofavailable, directly relevant sites; in population size; andin demographic and socioeconomic characteristics. Therelationships between each of these variables and tripsfor activity k were estimated by the general model:

TRIPS& = exp (B,, - B, PRICE . .t B, INC3Q5i + B, BCT18TMDi+ B, CCPOP86, - B, PCTFARM,- B, SUBEROS, t B, SUITkii) , (5)

where

TRIPQ -- annual trips for activity k demandedfrom community i to site j,

PRICEkii -- cost of trips for activity k fromcommunity i to site j,

INC345i = percent of community i populationwith annual income of at least $30,000,

PCT18TMDi = percent of community i population age18 to 32,

CCPOP86t = total community i population (12 yearsold and older),

PCTFARMi = percent of community i populationliving on farms,

SUBEROS, = an index of substitute recreationalopportunities available to community iand which compete with activity k foravailable household time and money,

SUIT& = suitability of site j for activity k.

The dependent variable in equation (5), TRIPS .., wasconstructed in several steps. First, trips per cap1 a for7each of the 239 subregions, covering the entirety of thecoterminous 48 States, was calculated by dividing the

estimate of total number of trips for activity k by therespective population of each subregion (12 years old orolder). As discussed by Walsh (1986), this procedureaccounts for both the probability of participation and thefrequency of participation. Hence, the estimates of tripsper capita account for both participants and nonpartic-ipants in outdoor recreation.

Total trips for activity k from each representativecommunity within each subregion were estimated bymultiplying trips per capita for the subregion by theresident population of the community (12 years old orolder). Because it is a function of trips per capita, thespecification of total trips as the dependent variable forequation (5) implicitly accounts for both participants andnonparticipants (Walsh 1986; Ward and Loomis 1986).Thus, the problem of excluding nonparticipants in thedemand analysis, which is often encountered inapplications of the individual travel cost model, wasavoided (Ward and Loomis 1986).2

In equation (5), site j refers to a specific recreationalsite (e.g., a State park) used by community i for activityk. Sites used by community i for any of the 37 types ofoutdoor recreation modeled were identified fromdestinations reported in the PARVS data.

For each site, the probability that it was visited foractivity k by persons living in community i wascalculated. This probability was a function of thedistance to the site from the geographic center of thecommunity and the suitability of that site for a particularactivity. This estimate of probability that site j wasvisited by community i for activity k served to allocate apercentage of the total trips for activity k fromcommunity i. These allocated trips, or TRIPS&, werethe observations for the dependent variable for equation(5) and contained trips for activity k to each site visitedby people from each of the 239 communities (zones inthe ZTCM context), approximately 7,200 observations.

’ The probability of participation in activity k is given by the totalnumber of participants in community i divided by total population ofcommunity i (participants and nonparticipants). The frequency ofparticipation is given by the total number of activity k trips fromcommunity i/total number of participants in community i. Trips percapita from community i, or TRIP!&, is given by.TRfPSki = narticiuants

total population ’total tliDS total trios

participants ’ total population

2 Results reported later (e.g., consumption and cost indices) can stillonly be extrapolated back to the population of activity k participants,not the general population. Estimates of activity k participants in theUnited States, which were derived from several data sources, areshown in tables 5, 7, 9, and 11.

10

The price variable in equation (5), PRICE .., wasderived by first calculating the straight-line %istancefrom community i to site j and then applying a calculatedcircuity factor.

These circuity factors were computed as the ratio ofreported travel distance to straight-line distance across allPARVS respondents, about 32,008. Travel miles werethen converted to transportation costs, including theopportunity cost of travel time. Transportation cost wasbased on vehicle operating costs as reported by the U.S.Department of Transportation. The assumed vehicleoperating cost was $0.278 per mile, the cost of operatingan intermediate-size automobile. The opportunity costof travel time was valued at one-half the reported wagerate for a community, as recommended by Rosenthal(1985). The average travel speed was calculated bytravel time and distance as reported in the PARVS data.

The substitute variable in equation (5), SUBEROS,,was derived from the effective recreational opportunitysupply data base discussed previously. This substitutevariable combines recreational service, facility, andresource variables, indicating the range of opportunitiesfor activities that compete with activity k for ahousehold’s recreation time and money. Effectiveness ofsubstitute opportunities accounts for distance of theopportunities from community i and number of peoplein other communities who compete for theserecreational opportunities.

The suitability variable in equation (5), SUITu,represents the suitability of site j for activity k. Thesuitability variable is based on responses to thenationwide survey of site managers discussed previously.Socioeconomic variables in equation (5), includingINC345, PCTlSTMD, CCPOP86., and PCTFARM,were obtained from U.S. Census data.

The semilog functional form of equation (5) has beenrecommended in previous studies. This functional formhas been found to be theoretically consistent withrecreation demand behavior and reduces heteroscedas-ticity (Rosenthal 1985; Ward and Loomis 1986; Ziemerand others 1980). Equation (5) was estimated byordinary least squares for 37 recreational activities.These estimated individual community-by-site equationsor demand functions are shown in table 1.

After estimating equation (5) for each of the 37activities, an aggregate or market demand functionacross sites, corresponding to equation (l), was derivedfor each activity. These demand functions were derivedby substituting mean values for all independentvariables, except cost, into equation (8) and solving for acomposite constant term representing the sum of the

products of these means multiplied by their respectivepartial regression coefficients. This simplified equationwas then multiplied by the average number of sites usedfor activity k across all communities so that the equationrepresented total trips consumed by a typical communityat each average trip cost. The results of this operationwere aggregate demand curves for each activity for atypical U.S. community. These aggregate demandfunctions are shown in table 2.

Estimation of the aggregate demand functions depictedby equation (1) assumed that the cross-sectionalmodeling approach correctly identified the community-by-site demand functions. The identification questionrequires consideration of the community-level householdmarket for trips illustrated in figures 4 and 5. In figure5, the observed cost-quantity relationship would be P,and Q1 trips. However, this consumption point is not onthe aggregate demand function. Hence, if observedconsumption data corresponded to the household marketdisequilibrium situation shown in figure 5, the aggregatedemand function would not be identified.

The vertical portion of the effective aggregate supplycurve in figure 5 implies that trip production was limitedto local sites. Such would be the case if each communitywere located on an island. In order to identify theaggregate demand curve, it was assumed that local sitecapacity did not truncate the aggregate supply curve.Rather, it was assumed that recreationists were free tomove into other regions (to other destinations) until thepoint was reached where the marginal benefits and costsof the last trip produced and the last one consumedwere equal. This situation is illustrated in figure 4.

The availability of public recreational facilities andresources varied across the 239 U.S. communities usedin this study. Variations in public recreational facilitiesand resources lead to variations in the aggregate supplyfunction. Other things being equal, changes inconsumption caused by changes in the aggregate supplyfunction identify the aggregate demand function for trips(Cicchetti 1973; Kalter and Gosse 1970).

Consumption Function Estimation

As with the estimation of the aggregate demandfunctions, estimation of consumption functions assumedthat community-level household markets for trips werein equilibrium. Household market equilibriumconsumption is explained generally by equation (4). Forestimation purposes, equation (4) was specified as:

11

TRIPS, = exp (B,, + B, INC345t+ Bz PCT18TMDi + B, OCPOP86,- B4 PCTFARM, - B, SUBEROS,+ B, [FACILITYkt * SUIT,]) , (6)

where

TRIPS, = annual trips for activity k consumed bycommunity i

FACILITYki = quantity of recreational facilitiesdirectly relevant to activity k andavailable to community i.

All other variables are as defined for equation (5).Equation (6) represents the reduced form for therecreation demand/supply system given generally byequation (4).

The calculation of the dependent variable in equation(6) was discussed previously. The variable RO, wasobtained from the extensive recreational resource andfacility data set maintained by the Forest Service’sOutdoor Recreation and Wilderness Assessment Group.This data set enabled computation of the quantity ofvarious types of recreational facilities and resourcesavailable for each of the 37 activities and found withinthe relevant travel radii of each community in theUnited States. Specific facilities and resources which arerelevant to activity k were selected from this data set.As indicated by equation (6), these facilities andresources were then weighted by the average suitabilityof sites used by community i for activity k. The data setonly shows recreational facilities and resources within60- to X&mile travel radii of each community. Thesetravel radii represented the mean plus one standarddeviation of the actual distances reported by PARVSrespondents across 12 categories of resources relevant tothe 37 recreational activities modeled.

Equation (6) was estimated by ordinary least squares for37 recreation activities. In some equations thePCTFARM, variable was deleted because of collinearity.The estimated consumption functions are shown in table3. The recreational resource and facility variables usedin the consumption functions are defined in table 4.Estimated consumption functions were used to projectnumbers of trips for activity k from a typical communityto the year 2040.

In tables 5, 7, 9, and 11, future trip consumption indicesrelative to the 1987 base year are projected for fourpossible future recreation opportunity scenarios. Acrossall four scenarios, demand determinants, such as income,were projected to increase the same in the future at alikely, moderate rate as provided by the Bureau of theCensus. Thus, differences in consumption across thefour scenarios are the result only of differences in

growth of availability of opportunities across time, andsimulate the effects of alternative future supply policies,independent of demand growth.

In the low-supply growth scenario, facilities andresources for activity k decrease in the future. In thezero-supply growth scenario, facilities and resources foractivity k remain at current levels into the future.Facilities and resources for activity k increase at amoderate growth rate (about 0.5 percent annually) inthe medium-supply growth scenario. Finally, in thehigh-growth scenario, facilities and resources for activityk increase at a growth rate of about 1 percent annually.The indices in tables 5, 7,9, and 11 indicate that futureconsumption of recreational trips is sensitive to theavailability of recreational facilities and resources formost activities. The strength of recreational facility andresource supply effects varies across the differentrecreational activities.

Household Market Costs

Future household market trip costs were estimated bysubstituting future trip-consumption estimates into thefuture aggregate demand functions and solving for costs.The appropriate future aggregate demand function isdetermined by shifting the current aggregate demandfunction (see table 2) to reflect probable future demandschedules based upon assumed futures for the variablesshown in equation (5). Indices of future householdmarket costs relative to the 1987 base year are shown intables 6,8, 10, and 12. The four demand/supplyscenarios shown correspond to the scenarios explainedabove for the consumption indices.

The cost or price indices (tables 6, 8, 10, and 12)summarize the relationship between outdoor recreationsupply and demand over time. An index greater than100 indicates that trip demand is growing faster thansupply. An index equal to 100 indicates that tripdemand and supply are increasing at about the samerate. An index less than 100 indicates that supply isincreasing faster than demand.

The cost or price indices also summarize the impact onrecreation consumption of public policy on growth ofavailable recreational facilities and resources. Withdecreasing or zero public-facility growth, householdmarket costs would increase for most activities. Suchincreases in household market costs, in general, wouldbe most pronounced for land-based activities. Theimplication of increasing household market costs is thatrecreational opportunities would become more and morescarce. People, therefore, would have to travel greaterdistances or spend increased time searching for availablerecreational opportunities.

12

Under the assumption of medium public-recreational-facility growth, household market costs remain aboutconstant for many activities. The implication is thatgrowth of recreational opportunities at this medium ratewould just maintain a balance between demand andsupply. Thus, for many activities, maintenance of thestatus quo with respect to recreation demand and supplyrequires at least a moderate increase (about l/2 of 1percent per year) in public recreational facilities andresources. For some activities, maintenance of thestatus quo with respect to demand and supply requires a1 percent growth of public recreational facilities andresources.

Household market costs decrease for several activitiesunder the assumption of high public-recreational-facilitygrowth. The implication in this case is that rapid growthof recreational opportunities would cause them tobecome more abundant over time relative to demandgrowth. People would typically be able to travel shorterdistances and spend less time searching for availablerecreational opportunities. Thus, improvements in theavailability of recreational opportunities would require arelatively high rate of public-recreational-facility growthfor a number of activities, adding about 45 percent tothe total stock over the next 50 years.

Research Conclusions

More research is needed to improve estimates ofaggregate demand, supply, equilibrium consumption, andcosts or prices. A particular need is to collect sufficientdata compatible with conceptual models so thatcomplicated, time-consuming, and sometimesquestionable procedures for adjusting data to satisfymodel assumptions can be avoided. For the most part,the abundance of the data used to support the workreported here demonstrates this point. In addition toimproving the quality of available data, additionalresearch is needed to identify the most appropriate andcost-effective statistical procedures for estimatingaggregate demand and supply curves, and for estimatingequilibrium consumption and costs. There is also aspecific need to develop procedures for includingcongestion costs, and the effects of varying quality ofrecreational opportunities. Finally, much moreconceptual and empirical work, and convincing of policymakers, are needed to identify the public welfare andpolicy implications of changes in equilibriumconsumption and costs, not only over time but alsoacross spatial dimensions.

For previous RPA Assessment efforts, the economicanalyses, using existing data, technology, and theory,were unable to produce recreation demand/supply and

price trend comparisons equivalent to those producedfor commodity outputs, such as timber. With theaccomplishment of the research described above,measurable improvements in the appropriateness,accuracy, and believability of aggregate outdoorrecreation demand/supply analyses have been achieved.The theory and techniques described in this Paper, it isargued, provide a stronger foundation for advancedeconomic assessment of the demand and supply situationof outdoor recreation in the United States.

References

Bexker, G.S. 1%5. A theory of the allocation of time. EconomicJournal. 75493-517.

Bockstael, N.E; McConnell, K.E. 1981. Theory and estimation ofthe household production function for wildlife. Journal ofEnvironmental Economics and Management. 8:199-214.

Brown, W.G; Nawas, F. 1973. Impact of aggregation on theestimation of outdoor recreation demand functions. AmericanJournal of Agricultural Economics. 5524649.

Cicchetti, Charles J. 1973. Forecasting recreation in the UnitedStates. Lexington, MA: Lexington Books, D.C. Heath andCompany. 200 pp.

Clawson, M. [Marion]. 1984. Effective acreage for outdoorrecreation. Resources for the Future. 78~2-7.

Clawson, Marion; Knetsck, Jack L 1966. Economics of outdoorrecreation. Baltimore, MD: Johns Hopkins University Press. 328 pp.

Cordell ILK; English, D.B.K. [In press]. Measuring and projectingeffective recreation opportunities in the United States. In:Outdoor recreation benchmark 1988: Proceedings of the nationaloutdoor recreation forum. Gen. Tech. Rep. SE52. Asheville,NC: U.S. Department of Agriculture, Forest Service,Southeastern Forest Experiment Station.

Cordell, H. Ken; Bergstrom, John C; Hartmann, Lawrence k;English, D.B.K. [In press]. An analysis of the outdoor recreationand wilderness situation in the United States: 1989-2040. Atechnical document supporting the 1989 RPA Assessment.Washington, DC: U.S. Government Printing Office.

Dwyer, J.R; Kelly, J.R; Bowes, M.D. 1977. Improved procedures forvaluation of the contribution of recreation to national economicdevelopment. Res. Rep. 128. Urbana, IL: Univemity of Illinois,Water Resources Center. 218 pp.

Gum, RL; Martin, W.E. 1975. Problems and solutions in estimatingdemand for and value of rural outdoor recreation. AmericanJournal of Agricultural Economics. 57588-66.

Harrington, Winston. 1987. Measuring recreation supply.Washington, DC: Resources for the Future. 77 pp.

Hof, J.G; Kaiser, H.F. 1983. Long-term outdoor recreationparticipation projections for public land management agencies.Journal of Leisure Research. 15~1-14.

Kalter, RJ4 Cusse, LE. 1970. Recreation demand functions and theidentification problem. Journal of Leisure Research. 243-53.

13

Loomis, John B; HOT, John G. 1985. Comparability of market andnonmarket valuations of forest and rangeland outputs. Res. NoteRM457. Fort Collins, CO: U.S. Department of Agriculture,Forest Service, Rocky Mountain Forest and Range ExperimentStation. 5 pp.

McConneff, ICE. 1975. Some problems in estimating the demandfor outdoor recreation. American Journal of AgriculturalEconomics. Mayz330-321.

Rosenthal, Donald H. 1985. Representing substitution effects inmodels of recreation demand. Fort Collins, CO: Colorado StateUniversity. 211 pp. Ph.D. dissertation.

Sorg, Cindy F; Loomis, John; Donnelly, Dennis M. 1985. Neteconomic value of cold and warm water fishing in Idaho. Resour.Bull. RM-11. Fort Collins, CO: U.S. Department of Agriculture,Forest Service, Rocky Mountain Forest and Range ExperimentStation. 26 pp.

Walsh, RG. 1986. Recreation economic decisions: comparingbenefits and costs. State College, PA: Venture Publishing, Inc.637 pp.

Ward, F& Loomis, J.B. 1986. The travel cost demand model as anenvironmental policy assessment tool: a review of literature.Western Journal of Agricultural Economics. 11:164-178.

Ziemer, RF; Musser, W.N; Hill, RC. 1980. Recreation demandequations: functional form and consumer surplus. AmericanJournal of Agricultural Economics. 62:136-141.

14

Tables

Table l--Estimated community demand functions for recreational activities

Table 2--Abbreviated demand functions for a typical U.S. community, by activity, 1987

Table 3--Estimated community consumption functions for recreational activities

Table C-Recreational resource and facility variables used in activity consumption functions

Table 5--Future market-clearing trip indices under low public-supply-growth and medium-demand-growthassumptions, by activity

Table 6--Future market-clearing price indices under low public-supply-growth and medium-demand-growthassumptions for land activities

Table 7--Future market-clearing trip indices under zero public-supply-growth and medium-demand-growthassumptions for land activities

Table &-Future market-clearing price indices under zero public-supply-growth and medium-demand-growthassumptions for land activities

Table g--Future market-clearing trip indices under medium public-supply-growth and medium-demand-growth assumptions for land activities

Table lo--Future market-clearing price indices under medium public-supply-growth and medium-demand-growth assumptions for land activities

Table 11--Future market-clearing trip indices under high public-supply-growth and medium-demand-growthassumptions for land activities

Table 12--Future market-clearing price indices under high public-supply-growth and medium-demand-growthassumptions for land activities

15

p?nu!wcy)

9s7

OS

EE

8s

6P

1s

IV

ZS

w

w

IV

9s

OS

ZE

9v

a

SV

8v

6S

zv

s s

0 9

z s

6p’0

WsZlL

o8s1os

682’161

Z8l’li%

6SLWZ

t99v8

9LlPlE

lz45Yl~

lm’zop

9Eo’ozz

9W8S

OWZP

u6’18S

8Sl’L9

6Ez’KV

TEE-ss

P26’9LE

E8LTos

ooo’n9

LWLPE

zo9xa

lEL’K6

PPL’ZZS

LCE’6OS

om

9Mz

LLZZ

ZLZE

9 8 9 1

WI1

8892

9S9Z

8zlE

SC61

6L.S

90s

PESE

08

8662

008

6PLZ

LOEE

OSK

u8z

6LlP

8W

E887.

191s

(zro3 boo-)~181’ l *800�-(no9 boo9*9u l fW-(8lo’) G3oo9l 6Lz *OW-(810’) CL009*9Ll *lzo’-CL109 b.w)*9LZ l 8LO.-G?zo9 (zro�)*VLl *PLO’-CL109 CL009&ZZ l lEO�-klo9 CXN)*ENI(910’) 2%l 861�h-) l El;*SLl’ l *910’-bar) (0109

EL 6%

6%*090’-bo9

*LPI’ *PEO’-(Lid tile?*IL1&lo3 S%*ozl’ *SW-(w) (Zlo’)*98Z *SO’-(Zlo’) boo-).E8I l 190

/:SW ) boo9

*ZPl’ *zxs-(010’) 600-)*SlZ 96&O’-(910’) CL009~651’ tLl0(Zlo’) aoo9*WI’ *OUT-(ore’) bQo9l bOZ l 8ZU-(910’) cxlo9*C60 l LZO.-(910’) cm9l ZZl’O *9Zo’O-

CL009991’-

(Zlo’Yl Slz=bzo9l %l’-ml9el81’

hl-3 EL-E)s68C0lOOO

ho-zI oz-s)l Zloo(nm@O-ZI 6SL)l hlrnGro-z RTS)*‘cl00000c3oCI sv-L)l lU)OoO

cLo-3 sxl>l 01mho-FI ZL’L).ElOOOMlho-~ 98’9)l OlooooOh-LI LE’S)

*ZmoooO630-2 ML)*ZlOOOWm ET)

l ttOOMOO(Lo-a u’l)l SLcmoOO

Gm3 OL-PI*21ooooocL0-a 9vz)l lZcMmo&H3 oc9>*ElooooOh-a 28’1)l 9lculoMl&O-El 6~S)*lmoooOho-EI LLY)*llooooObO5I zl?)

6 L o o o o o 0h5I 19’9)

*El000006303 8Lx)*llcKxmc3o-a WE).-

ho-a ZL.9)l tmoooclkm W-P)l lmooo-0

(llo3.m(SKY)*zLo(zzo’)

G%*9Sl’(zzo’)*911’ho9

**lL4l(610.1

GE,,801’cm?l ZZl’(610-1ZW-m9*ZU&09*LLl’(ElO’).lEl’(Lb09

*.*OLo(Llo’).CZl.c?Eo’)

**a&l(910’)l 990(Plo’)*ZlI’(ElO’)l 880(Llo’)*L&oh-r)*lEl’(010’)*180(9lo3l 9El’(Pig’)*880-O

&x0-)l lSOboo-)l 690bXr)l 901�cao-~*6Loboo9l *llO

(010’)110

bxr)&LoCnY~*MObw)

(EL

(ELl EEO(Plo�)l 9tTobaY)l E80(El09l LEl.cm-)l 911’(110’)Cl0

ho-)l 190boo9*zLoho-)l KOc⌧lo9*9LOhxY)*8LlJklo.).6Z4lcxu)*CL0ho9*sLo-0

booO9+zzo’-

CLOOO-1l 620’-(loo’)*ZlO’-

GiooO-1r6ZO’-(loo’)*VW-ca309*DO--(TOO’)*9W-bo?l 6U.P

kiolxs).zzo’-(loo’>l 9ZO.-h⌧r)*pzo�-boo9l ZEO�-ho-)*LZo’-(Plo’)*sEl’-(loo’)*Ku-boo9

tiE-,

g$j

*6Zo’-cmo-~*UO’-(loo’)*9W-

halo-1

&%810’-

(1100’)+oso’-

belo*8loP

(692’).ZZYSWE.)l 618’&@IS).LEZ’E(L6C.1*VW1(EOS)*oovSb3L9l 66Z’L(IEP’)*86VP(S6E’)*llLS(LSE’).L%‘Z(WV)eS6SLbo8-)*9sssb8L9l OZ8’9(OK-)*loo3bol~l>8 1 8 9 - Pcm3*88VEo?w~l LL8’9CLS%-1*sEsscm)*Plz’Ptax)*08L-9kMY).ZLP’S(ELZ.1*ZO6-E63w)*9lo’Lhw)*z88’PbEE’)*EOSP

Table l--Estimated community demand functions for recreational activities--Continued

Activity INI’ERCEFT PRICEijk INC34Si

Parameter estimates (standard error)Adjusted

PCIX3TMDi ccPoP86i PCTFARMi SUBERO$ SUIT, N F-value R2

Pool swimming

Motorized boating

Water-skiing

Rafting/tubing

Canoeing/kayaking

Rowing/other boating

Stream/lake swimming

Saltwater fishing

Warm-water fmhing

Cold-water fishing

Anadromous fishing

Downhill skiing

Crcsscountry skiing

7.76.5*‘f-4g

(lb)

0.059. *

‘:Z(-Q-1

0.001**

(:gg)(Jw

0.3388(.037)

857 131.410 0.43

1537 176.174 .41

15.53 136.448 34

1379 %.158 .29

2381 455.052 53

2413 248.152 .%

2678 521.61 54

664 109.797 .45

2290 316.040 .45

1 2 9 0 132.738 38

9 9 1 31.171 .l5

138 22.706

2656 231.917

0.49

34

*Significant at 0.01 level; **Significant at 0.05 level; *** Significant at 0.10 level.

Table 2--Abbreviated demand functions for a typical U.S. community, by activity, 1987

Parameter estimates

Activity Interceptcost

(standard error)*

Developed campingPicnickingSightseeingFamily gatheringPleasure drivingVisiting historic sitesAttending eventsVisiting museumsOff-road drivingBikingRunning/h&%wakingCutting firewoodCollecting berriesVisit prehistoric sitesPhotographyDay hikingHorseback ridingSmall-game huntingBig-game huntingNature studyBackpackingPrimitive campingWildlife observation

Pool swimmingMotorized boatingWater-skiingRafting/tubingCanoeing/kayakingOther boating/rowingStream/lake swimmingSaltwater fishingWarm-water fishingCold-water fishing

LAND

13.2226 -0.0182 (0.00042)15.1128 -.0499 (0.00110)14.9591 -.0183 (0.00029)14.2297 -.0232 (0.00043)15.1268 -.0359 (0.00109)13.5640 -.0231 (0.00050)13.2127 -.0286 (0.00076)13.3323 -.0227 (0.00073)13.0965 -a437 (0.00380)14.1572 -.0315 (0.00106)15.1549 -.1356 (0.01373)15.1723 -.0271 (0.00066)12.3011 -.0319 (0.00447)11.9465 -.0244 (0.00338)11.2391 -.0261 (0.00140)13.2304 -.0221 (0.ooo9o)13.5763 -.0393 (0.00125)12.5163 -.0459 (0.00142)13.1748 -.0632 (0.00334)13.1341 -.0345 (0.00103)12.6636 -.0289 (0.00093)10.6442 -.0124 (0.00110)12.6221 -.0294 (0.00070)13.7336 -.0220 (O.ooo46)

WATER

14.6800 -0.0364 (0.00208)13.7019 -.0383 (0.00156)12.4289 -.0275 (0.00152)6.0026 -.0326 (0.00204)

12.7674 -4484 (0.00122)11.6530 -.0237 (0.00116)14.9339 -.0341 (0.00073)11.4034 -.0194 (0.00123)14.6418 -.0490 (0.00129)13.1901 -.0272 (0.00151)

SNOW AND ICE

Cross-country skiingDownhill skiing

8.6520 -0.0340 (0.00237)11.2286 -.0308 (0.oo46o)

*All models significant at p c 0.01.

19

s Table 3--Estimated community consumption functions for recreational activities

Activity

Parameter estimates (standard error)Adjusted

INTERCEP INC345 PCIXI’MD CCPOP86 SUBEROS PCTFARM ROl R02 N F-value R2

Developed camping

Picnicking

Sightseeing

Family gatherings

Pleasure driving

Visiting historical sites

Attending events

Visiting museums

Off-toad driving

Biking

Running/jogging

Walking

Cutting firewood

Collecting berries

Visiting prehistoric sites

Photography

Day hiking

Horseback riding

Small-game hunting

Big-game hunting

Nature study

Backpacking

Primitive camping

Wildlife observation

8x3*

6.913*

7.618*

8.873’

E?(:729)

-0.189.(.018)

-.205*(.019)

-.219* l

$gi

(.020)

-.194*(.016)

.ooo59*(.t3803)

800021**C~>

239

239

239

239

239

239

239

239

239

239

239

239

239

239

239

239

239

239

239

239

239

.000000076*** 239(444 E-08) 239

.00677’ 239(.0022)

49.488 0.50

54.607 53

80.838 .67

57.467 54

51.895 52

87.398 .a

56.683 54

57.763 54

14.905 23

53.677 52

25.164 34

58.998 55

53.924 57

55.490 58

77.633 66

59.638 55

97.476 .71

22.482 35

9.984 .16

16544 25

30.993 .43

21.337 34

45.018 A8

49.075 55

Table 3-Estimated community consumption functions for recreational activities--Continued

’ Parameter estimates (standard error)

Activity INTERCEP INC345 PCTlSTMD CCPOP86Adjusted

SUBEROS PCI’FARM ROl R02 N F-value R2

Pool swimming

Motorized boating

Water-skiing

Rafting/tubing

Canoeing/kayaking

Sailing/rowing

Stream/lake swimming

Saltwater fishing

Warm-water fishing

Cold-water fishing

Anadromous fishing

Downhill skiing

Cross-country skiing

3.091*ygz

&’

(:903)-12.7608

‘:.gq$$992

(:919)9.2w

i-2

s::;*p:i

,(yg1

(4.851)

11.45s(2.146)7570**

(4.061)

0.00143*

‘:Ei*(.0003).00144*

(:E&*

‘:Eb

$339**wQ3).000876*

(003).00%09*

(.ow.001225*

‘:x7*WV

.OOOs14*ww

SNOW AND ICE

0.OOOOO11* 0.0013** -0.2.56* .001086*w-73 (-0003)

.OOcQ33*(S.06 E-07) (.000007)

0.109*

‘-g;19**coo@w

.9ss3**

(-gJg*co@w

.ooo622*mw

.00128*~-0oo4)

239

239

239

239

239

239

239

239

239

239

239

239

239

58521 0.59

31.265 A3

29.446 .37

6.431 .12

33.973 .4s

32.137 A0

43.275 52

21.944 .31

32.607 A0

20.027 39

12.451 .19

41.848

23.268

52

.32

*Significant at 0.01 level; **Significant at 0.05 level; ***Significant at 0.10 level. ’ For these activities, age variable was MEDAGE = medium age of central county population.

Table 4--Recreational resource and facility variables used in activity consumption functions

Activity ROl R02

LAND

Developed camping Federal road mileage converted to acres andFederal and State land located within l/2 mile ofa road*

Picnicking

Sightseeing

Family gatherings

Pleasure driving

Federal and State land located within l/2 mile ofa road, and State forest land open to recreation*

Federal road mileage converted to acres, andFederal and State land located within l/2 mile ofa road*

Federal road State, local and privatecampgrounds*

Federal road mileage converted to acres, andFederal and State land located within l/2 mile ofa road

Visiting historical sites Federal road mileage converted to acres, andFederal and State land located with l/2 mile of aroad*

Attending events Federal road mileage converted to acres, andFederal and State land located within l/2 mile ofa road*

Visiting museums

Off-road driving

Federal road mileage converted to acres, andFederal and State land located within l/2 mile ofa road*

Federal road mileage (except for U.S. ArmyCorps of Engineers and Tennessee ValleyAuthority) converted to acres, NationalRecreational Trail mileage open to motorcyclesconverted to acres, and Federal and State landlocated within l/2 mile of a road*

Federal road mileage converted to acres, Federaland State land located within l/2 mile of a road,and State forest acres open to recreation*

Federal road mileage converted to acres, Federaland State land located within l/2 mile of a road,and State forest acres open to recreation*

--

_-

--

--

--

--

__

22

Table 4--Recreational resource and facility variables used in activity consumption functions--Continued

Activity ROl R02

walking

Cutting firewood

Collecting berries

Visiting prehistoric sites

Photography

Day hiking

Horseback riding

Small-game hunting

Federal road mileage converted to acres, andFederal and State land located within l/2 mile ofa road* --

Federal land located within l/2 mile of a road,Federal and State land located within l/2 to 3miles of a road, and acres of nonindustrial forestland open to recreation, both leased andnonleased* --

Industrial and nonindustrial forest lands*

Federal road mileage converted to acres, Federaland State land located within l/2 mile of a road,Federal and State land located within l/2 to 3miles of a road, and rural transportation useacres*

Federal and State land located within l/2 mile ofa road, and State Forest acres open to recreation* - -

Federal and State land located within l/2 mile ofa road, Federal and State land located l/2 to 3miles of a road, and Federal wilderness* - -

Federal and State land located within l/2 to 3miles of a road, and nonwilderness land morethan 3 miles from a road

Federal land located within l/2 mile of a road,Federal land located within l/2 to 3 miles of aroad, industrial forest-land acres, andnonindustrial private-land acres open torecreation, both leased and not leased*

Miles of NationalRecreational Trails opento horseback riding

__

Federal land located within l/2 mile of a road,Federal land located l/2 to 3 miles of a road,industrial forest-land acres, and nonindustrialprivate-land acres open to recreation, both leasedand not leased

Big-game hunting

23

Table 4--Recreational resource and facility variables used in activity consumption functions--Continued

Activity ROl R02

Nature Study Acres of water in river/streams up to 660 feetwide, and acres of flat-water bodies

Backpacking Federal and State land located within l/2 to 3miles of a road, nonwilderness land located over 3miles from a road, and Federal wilderness acres*

Primitive camping Federal and State land located within l/2 mile ofa road, and State forest acres open to recreation* --

Wildlife observation Federal and State land located within l/2 mile of Acres of water ina road, Federal and State land located with l/2 to rivers/streams up to 6603 miles from a road, nonwilderness land located feet wide, acres of flat-more than 3 miles from a road, Federal water bodies, and acreswilderness acres, The Nature Conservancy acres, of Federal water bodiesand State fBh and game land* open to recreation*

Pool swimming

Motorized boating

Water-skiing

Rafting/tubing

WATER

Public and private swimming pools, State parkswith some swimming facilities, and touristaccommodations

Acres of flatwater bodies and acres of Federalwater open to recreation

Acres of flatwater bodies and acres of Federalwater open to recreation*

Miles of Federal wild and scenic rivers, miles ofrivers designated by States as being significant forhistoric, cultural, scenic or recreational reasons,and miles of Bureau of Land Managementrecreational rivers*

Canoeing/kayaking Acres of flatwater bodies, and acres of water inriver/streams up to 660 feet wide*

Sailing/rowing Acres of flatwater bodies, acres of water inrivers/streams up to 660 feet wide, and acres ofFederal water bodies open to recreation*

Federal and State landlocated within l/2 mileto 3 miles of a road,nonwilderness landlocated over 3 milesfrom a road, and Federalwilderness acres*

National RecreationTrail State park trailmiles*

--

Number of boat ramps*

--

Indicator variable forpresence of mountains(0 = no mountains;1 = mountains)*

Canoe rental firms andcanoe outfitters*

24

Table 4--Recreational resource and facility variables used in activity consumption functions--Continued

Activity ROl RO2

Stream/lake/oceanswimming

Saltwater fishing

Warm-water fishing

Cold-water fishing

Anadromous fishing

Downhill skiing

Cross-country skiing

Federal developed swimming areas* Miles of public oceanbeach*

Miles of public ocean beach*

Acres of flatwater bodies, acres of water inrivers/streams up to 660 feet wide, and acres ofFederal water bodies open to recreation*

Acres of water in rivers/streams up to 660 feetwide, and acres of Federal water bodies open torecreation*

Miles of Federal wild and scenic rivers, and milesof rivers designated by States as being significantrecreational resources* a

SNOW AND ICE

Daily ski-lift capacity

Federal and State lands located within l/2 mile ofa road, Federal and State lands located within l/2to 3 miles of a road, and acres of ruraltransportation use*

*Resource and facility variables are weighted by the average suitability of sites used by a community for anactivity. Average suitability was derived from responses to a survey sent to site managers.

__

Indicator variable forpresence of mountains(0 = no mountains;1 = mountains)*

a States where no anadromous fishing occurs were assigned an average suitability of zero.

25

Table J--Future market-clearing trip indices under low public-supply-growth and medium-demand-growthassumptions, by activity

(1987 base level = 100)

A c t i v i t y

Baselineconsumption

i n U . S .(mi l l ion)

Y e a r

2010 2020 2030 2040

Developed campingPicnickingSightseeingFamily gatheringPleasure drivingVisi t ing his tor ic s i tesAttending eventsVisi t ing museumsOff-road drivingBikingRunning/joggingwalkingCutting firewoodCollecting berriesVisiting prehistoric sitesP h o t o g r a p h yDay hikingHorseback ridingSmall-game huntingBig-game huntingNature studyBackpackingPrimitive campingWildlife observation

Pool swimming 221.0 1 3 0 155 1 8 1 208 228Motorized boating 219.5 1 0 3 107 111 116 119Water-skiing 107.5 1 0 9 117 1 2 5 1 3 3 139Rafting/tubing 8 . 9 9 8 1 0 6 1 1 3 1 3 1 1 4 3Canoeing/kayaking 39.8 107 115 123 1 3 3 1 4 0Rowing/paddling/other boating 61.8 110 120 130 142 1 5 0Stream/lake swimming 238.8 102 106 110 114 1 1 8Saltwater fishing 77.3 103 109 116 127 134Warm-water fishing 239.5 9 0 8 7 8 4 8 4 8 3Cold-water fishing 83.8 1 1 1 1 2 1 130 141 1 4 8

Downhill skiing 64.3 1 4 1 173 205 238 260Cross-country skiing 9.7 1 3 3 152 164 171 1 5 8

LAND

60.6 116 130 144 158 168262.0 105 1 1 1 118 125 130292.7 115 129 146 167 1 9 0

74.4 112 123 134 146 154421.6 112 122 133 145 1 5 3

73.1 118 135 1 5 5 182 21073.7 1 1 1 1 2 1 132 144 152

9.7 1 1 5 129 1 4 3 160 1 7 180.2 104 109 114 120 124

114.6 1 2 1 140 1 6 0 182 19883.7 126 149 1 7 3 200 220

266.5 113 1 2 5 137 1 5 1 1 6 130.3 110 119 1 3 1 147 1 6 519.0 110 1 2 1 134 153 1 7 516.7 128 1 5 0 177 2 0 9 24442.0 1 1 8 1 3 5 152 170 18391.0 1 2 5 149 177 211 24763.2 116 1 2 9 142 154 16258.6 9 0 8 4 8 0 7 8 7 655.2 9 3 9 0 8 8 8 8 8 870.8 102 107 112 120 1 2 526.0 128 152 1 7 8 2 0 5 2 2 438.1 1 1 0 120 1 2 9 139 14669.5 1 1 1 122 133 145 153

WATER

SNOW AND ICE

2 6

Table 6--Future market-clearing price indices under low public-supply-growth and medium-demand-growthassumptions, by activity

(1987 base level = 100)

Activity

Baselinecost per

day (1987)

Year

2010 2020 2030 2040

Developed campingPicnickingSightseeingFamily gatheringPleasure drivingVisiting historic sitesAttending eventsVisiting museumsOff-road drivingBikingRunning/joggingwalkingCutting firewoodCollecting berriesVisiting prehistoric sitesPhotographyDay hikingHorseback ridingSmall-game huntingBig-game huntingNature studyBackpackingPrimitive campingWildlife observation

Pool swimmingMotorized boatingWater-skiingRafting/tubingCanoeing/kayakingRowing/paddling/other boatingStream/lake swimmingSaltwater fishingWarm-water fishingCold-water fishing

Downhill skiingCross-country skiing

LAND

$21.3939.6965.6167.52

2;54:4957.0834.0440.8212.0044.9126.0930.0614.1131.4232.3533.1632.8747.7735.1729.2623.6141.47

WATER

102101101103101102102102101102103102102103102103103103102102102105102103

$46.5239.8541.0446.9138.6139.7957.2888.3840.5840.97

103101101107102101101101101101

SNOW AND ICE

27.81 10642.59 107

104102102105102103103103101103104103104105103105105104103103103108103104

105102102114104102102103101102

109110

105103103106103104104104102105106104105107104107106106104103104111104106

106103103119105103102104102103

112113

106103105108104106105105102106108105106108106108108107105104105114105107

108104104126107104103105102104

115114

107104106109105107106106103107108105108110107109110108105104106115106108

109104105130108105104106102104

116114

27

Table ‘I-Future market-clearing trip indices under zero public-supply-growth and medium-demand-growthassumptions, by activity

(1987 base level= 100)

Activity

Baselineconsumption

in U.S.(million)

Year

2010 2020 2030 2fl48

LAND

Developed camping 68.6Picnicking 262.0Sightseeing 292.7Family gathering 74.4Pleasure driving 421.6Visiting historic sites 73.1Attending events 73.7Visiting museums 9.7Off-road driving 80.2Biking 114.6Running/jogging 83.7walking 266.5Cutting firewood 30.3Collecting berries 19.0Visiting prehistoric sites 16.7Photography 42.0Day hiking 91.0Horseback riding 63.2Small-game hunting 58.6Big-game hunting 55.2Nature study 70.8Backpacking 26.0Primitive camping 38.1Wildlife observation 69.5

Pool swimming 221.0 131 158 186 215 236Motorized boating 219.5 105 110 115 120 124Water-skiing 107.5 110 118 127 135 142Rafting/tubing 8.9 108 124 141 169 190Canoeing/kayaking 39.8 110 128 131 144 153Rowing/paddling/other boating 61.8 111 121 132 144 152Stream/lake swimming 238.8 104 109 114 119 124Saltwater fishing 77.3 104 112 120 133 141Warm-water fEhing 239.5 91 88 85 85 85Cold-water fishing 83.8 111 121 132 142 150

WATER

117107116116113118112116105122127114111112129120127119

E105l31112114

SNOW AND ICE

Downhill skiing 64.3 146 182 219 258 283Cross-country skiing 9.7 137 160 176 185 173

132 147 162114 121 130130 148 169130 144 158124 136 148136 157 184123 135 148130 146 162110 115 121142 162 186152 178 207127 140 155122 134 151124 139 159152 179 212139 157 178152 181 217133 148 16188 85 8395 95 96

111 117 126157 186 215122 133 144127 139 153

173135193

:“:2131571741252022281651701822481922551718297

132237151163

28

Table &-Future market-clearing price indices under zero public-supply-grow& and medium-demand-growthassumptions, by activity

(1987 base level = 100)

A c t i v i t y

Baselinecost per

day (1987)

Year

2010 2020 2030 2040

Developed camping $21.39 101Picnicking 39.69 1 0 1Sightseeing 65.61 101Family gathering 67.52 1 0 1Pleasure driving 40.26 1 0 1Visi t ing his tor ic s i tes 59.97 1 0 1Attending events 54.49 1 0 1Visi t ing museums 57.08 101Off-road driving 34.04 100Biking 40.82 102Running/jot3&3 12.00 102walking 44.91 1 0 1Cutting firewood 26.09 1 0 1Collecting berries 30.06 1 0 1Visiting prehistoric sites 14.11 1 0 3P h o t o g r a p h y 31.42 102Day hiking 32.35 102Horseback riding 33.16 102Small-game hunting 32.87 1 0 1Big-game hunting 47.77 1 0 0Nature study 35.17 100Backpacking 29.26 1 0 3Primitive camping 23.61 1 0 1Wildlife observation 41.47 1 0 1

Pool swimming $46.52 102Motorized boating 39.85 1 0 0Water-skiing 41.04 1 0 1Rafting/tubing 46.91 102Canoeing/kayaking 38.61 1 0 1Rowing/paddling/other boating 39.79 1 0 1Stream/lake swimming 57.28 100Saltwater ftshing 88.38 1 0 0Warm-water fishing 40.58 1 0 1Cold-water fishing 40.97 1 0 1

Downhill skiing 27.81 104Cross-country skiing 42.59 104

LAND

WATER

SNOW AND ICE

1021 0 11021021021031021021 0 11031031021021021041031041031 0 11 0 01 0 1105102102

1041 0 11021051021021 0 11011 0 1102

1 0 61 0 7

103 104 105102 102 102103 104 1 0 5103 104 104102 103 103104 105 1 0 6103 103 104103 104 1051 0 1 102 102104 105 106105 106 107103 103 104103 104 1 0 51 0 3 104 105106 108 1 0 9104 105 105105 107 1 0 8104 1 0 5 1 0 51 0 1 102 102100 100 1001 0 1 102 103106 1 0 8 109103 103 1041 0 3 104 104

1 0 51011021 0 81031031 0 11021 0 1102

1071021031161041041021041 0 11 0 3

108108

1061021031131041 0 31 0 11031 0 1103

1101 0 9

1 1 11 0 8

29

Table 9--Future market-clearing trip indices under medium public-supply-growth and medium-demand-growth assumptions, by activity

(1987 base level = 100)

Activity

Baselineconsumption

in U.S.(million)

Year

2010 2020 2030 2040

Developed camping 60.6 120 137 155 173 186Picnicking 262.0 110 119 129 139 147Sightseeing 292.7 117 132 151 174 198Family gathering 74.4 120 137 155 174 187Pleasure driving 421.6 115 128 141 155 165Visiting historic sites 73.1 119 138 160 189 219Attending events 73.7 115 128 142 157 167Visiting museums 9.7 118 134 150 169 182Off-road driving 80.2 105 111 117 124 128Biking 114.6 124 145 168 193 211R~dbgg;lg 83.7 131 158 188 221 246walking 266.5 116 131 146 163 176Cutting firewood 30.3 112 124 138 157 176Collecting berries 19.0 114 128 144 167 192Visiting prehistoric sites 16.7 131 155 184 219 258Photography 42.0 125 147 170 195 214Day hiking 91.0 130 157 190 231 273Horseback riding 63.2 122 139 157 174 186Small-game hunting 58.6 95 93 91 90 89Big-game hunting 55.2 100 102 104 106 108Nature study 70.8 109 119 129 141 149Backpacking 26.0 136 167 202 238 265Primitive camping 38.1 115 127 140 154 163Wildlife observation 69.5 120 137 155 174 188

Pool swimming 221.0 133Motorized boating 219.5 107Water-skiing 107.5 111Rat&g/tubing 8.9 134Canoeing/kayaking 39.8 116Rowing/paddling/other boating 61.8 112Stream/lake swimming 238.8 107Saltwater fishing 77.3 108Warm-water fuhing 239.5 92Cold-water f=hing 83.8 113

Downhill skiing 64.3 151Cross-country skiing 9.7 147

LAND

WATER

SNOW AND ICE

10211412117513112311311990

123

192179

193 224 246120 127 132131 141 148225 298 355147 165 177134 147 156120 128 133130 146 15788 88 88

134 146 154

237205

283223

314212

30

Table lo--Future market-clearing price indices under medium public-supply-growth and medium-demand-growth assumptions, by activity

(1987 base level = 100)

Activity

Baselinecost per

day (1987)

Year

2010 2024) 2030 2040

Developed camping $21.39 100 100Picnicking 39.69 9 9 9 9Sightseeing 65.61 101 101Family gathering 67.52 9 9 9 9Pleasure driving 40.26 100 100Visiting historic sites 59.97 101 102Attending events 54.49 100 100Visiting museums 57.08 100 101Off-road driving 34.04 100 100Biking 40.82 101 101Running/hw3h3 12.00 101 101walking 44.91 100 100Cutting firewood 26.09 100 100Collecting berries 30.06 9 9 9 9Visiting prehistoric sites 14.11 102 103Photography 31.42 9 9 9 8Day hiking 32.35 101 101Horseback riding 33.16 100 100Small-game hunting 32.87 9 9 9 9Big-game hunting 47.77 9 9 9 8Nature study 35.17 98 9 7Backpacking 29.26 9 8 9 7Primitive camping 23.61 100 100Wildlife observations 41.47 9 8 9 7

Pool swimming !§46.52 101 102 103 104 104Motorized boating 39.85 9 9 9 9 9 8 9 8 9 8Water-skiing 41.04 100 100 100 100 100Rafting/tubing 46.91 9 1 8 7 8 4 8 4 8 3Canoeing/kayaking 38.61 9 9 9 8 98 98 9 8Rowing/paddling/other boating 39.79 100 101 101 101 102Stream/lake swimming 57.28 9 9 99 9 9 9 9 9 9Saltwater fishing 88.38 9 9 99 9 8 9 9 9 9Warm-water fishing 40.58 100 100 100 100 100Cold-water fishing 40.97 100 101 101 101 102

Downhill skiing 27.81 102 103 104 105 105Cross-country skiing 42.59 100 99 9 8 97 95

LAND

WATER

SNOW AND ICE

1009 9

1029 9

1001031001011001021021001009 9

1049 7

1021019 99 7

iz100%

100 1009 9 9 9

103 1049 9 9 9

100 101104 105100 100102 102101 101103 103103 103100 100100 1019 9 100

105 1069 7 9 7

103 104101 1019 8 9 8% %% 9 5% %

100 100% 9 5

31

Table 11--Future market-clearing trip indices under high public-supply-growth and medium-demand-growthassumptions, by activity

(1987 base level = 100)

Activity

Baselineconsumption

in U.S.(million)

Year

2010 2020 2030 2040

Developed camping 60.6 123 142 163Picnicking 262.0 113 125 137Sightseeing 292.7 118 134 154Family gathering 74.4 125 146 168Pleasure driving 421.6 117 131 147\Usiting historic sites 73.1 121 140 163Attending events 73.7 117 132 149Visiting museums 9.7 119 137 155Off-road driving 80.2 106 113 119Biking 114.6 125 148 173Running/jogging 83.7 134 165 199walking 266.5 118 135 153Cutting firewood 30.3 114 127 142Collecting berries 19.0 116 131 150Visiting prehistoric sites 16.7 132 158 189Photography 42.0 129 156 184Day hiking 91.0 133 163 200Horseback riding 63.2 126 146 168Small-game hunting 58.6 98 97 97Big-game hunting 55.2 104 108 113Nature study 70.8 114 127 141Backpacking 26.0 141 178 220Primitive camping 38.1 118 133 149Wildlife observation 69.5 126 148 172

Pool swimming 221.0 135 166 199 233 258Motorized boating 219.5 110 118 126 134 140Water-skiing 107.5 113 124 135 146 154Rafting/tubing 8.9 167 248 358 522 665Canoeing/kayaking 39.8 122 143 164 189 206Rowing/paddling/other boating 61.8 113 125 137 151 160Stream/lake swimming 238.8 109 118 127 136 143Saltwater fshing 77.3 112 126 141 161 175Warm-water fishing 239.5 93 91 90 91 92Cold-water fuhing 83.8 114 125 137 150 158

Downhill skiing 64.3 157 204 256 310 348Cross-country skiing 9.7 158 200 238 267 260

LAND

WATER

SNOW AND ICE

184150178191163194166175126

23717216217522721524518897118157267165198

199159204207174225179189131220265186183202268237292202981221692 %176217

32

Table 12~-Future market-clearing price indices under high public-supply-growth and medium-demand-growth assumptions, by activity

(1987 base level= 100)

Activity

Baselinecost per

day (1987)

Year

2010 2820 2030 2040

Developed camping $21.39 98 97Picnicking 39.69 98 97Sightseeing 65.61 100 100Family gathering 67.52 98 96Pleasure driving 40.26 99 99Visiting historic sites 59.97 100 101Attending events 54.49 98 98Visiting museums 57.08 100 99Off-road driving 34.04 100 99Biking 40.82 100 100Running/jogging 12.00 99 99walking 44.91 99 98Cutting fuewood 26.09 98 98Collecting berries 30.06 97 96Visiting prehistoric sites 14.11 100 101Photography 31.42 % 93Day hiking 32.35 99 99Horseback riding 33.16 99 98Small-game hunting 32.87 98 97Big-game hunting 47.77 97 95Nature study 35.17 95 93Backpacking 29.26 94 90Primitive camping 23.61 99 98Wildlife observation 41.47 95 92

Pool swimming $46.52 101Motorized boating 39.85 98Water-skiing 41.84 99Rafting/tubing 46.91 80Canoeing/kayaking 38.61 97Rowing/paddling/other boating 39.79 99Stream/lake swimming 57.28 98Saltwater fishing 88.38 97Warm-water fishing 48.58 100Cold-water fuhing 40.97 100

WATER

SNOW/ICE

Downhill skiing 27.81 100 100 100Cross-country skiing 42.59 95 92 88

101979870959997

z100

%97

101959810297999910099989795

101919998%939187

zi

102%98609499979599

100

%%

1019498

10397

10099

10099979794

1028910097959289849788

1029597559399%9499100

9985

%%

102%98

10496

10099

101100979794

10388

1009794918882

z

10294975192

100%9498

100

9981

33

L

Cordell, H. Ken; Rergstrom, John C. 1989. lluxuy and techniquesfor assessing the demand and supply of outdoor recreation in theUnited States. Res. Pap. SE-275. Athens, Gk U.S. Department ofAgriculture, Forest Service, Southeastern Forest Experiment Station.33 PP.

As the central analysis for the 1989 Renewable Resources planningAct Assessment, a household market model covering 37 recreationalactivities was computed for the United States. Equilibriumconsumption and costs were estimated, as uwe likely future changesin consumption and costs in response to expected demand growthand alternative development and access policies.

Keywords: outdoor recreation, market equilibrium, demand, supply,forecasts, household production theory.

Cordell, H. Ken; Retgstrom, John C. 1989. Theory and techniquesfor assessing the demand and supply of outdoor recreation in theUnited States. Res. pap. SE-275. Athens, GA: US. Department ofAgriculture, Forest Service, Southeastern Forest Experiment Station.33 pp. i

As the central analysis for the 1989 Renewable Resources planningAct Assessment, a household market model covering 37 recreationalactivities was computed for the United States. Equilibriumconsumption and costs were estimated, as were likely future changesin consumption and costs in response to expected demand growthand aitemative development and access policies.

Keywords: Outdoor recreation, market equilibrium, demand, supply,forecasts, household production theory.

.

The Forest Service, U.S. Department ofAgriculture, is dedicated to the principle of

multiple use management of the Nation’s forest resourcesfor sustained yields of wood, water, forage, wildlife, andrecreation. Through forestry research, cooperation with theStates and private forest owners, and management of theNational Forests and National Grasslands, it strives-asdirected by Congress-to provide increasingly greater

. service to a growing Nation.

USDA policy prohibits discrimination because of race,color, national origin, sex, age, religion, or handicappingcondition. Any person who believes he or she has beendiscriminated against in any USDA-related actixity shouldimmediately contact the Secretary of Agriculture,Washington, DC 20250.

: .