Recreational demand for a gulf coast tourism destination.

Ha, Inhyuck "Steve"

ABSTRACT

Policy makers are often faced with limited resources and continuing demand for public services, and must make difficult decisions about how to allocate the public funds entrusted to them. To assess the economic value of ecosystems, such as beaches, a recreation demand function is estimated using the individual travel cost method (ITCM) for tourist areas in Northwest Florida. Visitor behavior patterns, broken down by the purpose of trip, such as business, vacation, and visits to friends and relatives (VFR), are examined. Survey data provided determinants of length of stay in the recreation area. The empirical results demonstrate the elasticities of income and prices of recreation products. Consumer surplus is also estimated to measure the changes in welfare according to the changes in value of resources.

INTRODUCTION

Policy makers in beach communities are faced with limited resources and continuing demand for public services, and must make difficult decisions about how to allocate the public funds entrusted to them. Those in charge of protecting and managing vital beach resources must justify their decisions in terms of benefits to the natural environment and demonstrate fiscal accountability if they wish to maintain public support. Often they are asked to justify their decisions in terms of the economic value that is generated for the community (Font, 2000). One of the primary economic benefits that these communities enjoy is spending related to beach tourism. Beach related tourist activity in the Northwest Florida area has long been a major source of employment for local residents, sales for local companies, and tax revenues for local government. Tourism's contribution to economic activity in the area is therefore an important consideration in community planning. Economic analyses that provide tangible estimates of these economic interdependencies and a better understanding of the role and importance of tourism in a region's economy are valuable to policy makers.

The purpose of this paper is to estimate a recreation demand function to estimate the economic value of ecosystems, such as beaches, using the individual travel cost method (ITCM) for tourist areas in Northwest Florida. Once the demand curve has been defined and estimated, one can also estimate the average consumer surplus, or economic benefits, for the recreational amenities of the beach. It is often mistakenly assumed that market price is the same as economic value. Actually, the market price represents the minimum amount that someone buying a good is willing to pay for it. People purchase marketed goods only if their willingness to pay is equal to or greater than the price of the good. Many people are actually willing to pay more than the market price for a good, reflecting an economic value greater than the market price. For policy makers to make resource allocation decisions based on economic values, what they need to know is the net economic benefit of a good or service. For individuals, incremental net benefits beyond the price paid are called consumer surplus, and are measured as the difference between the price actually paid for a good, and the maximum amount that an individual is willing to pay for it.

This paper consists of six sections. They are literature review, data, the theoretical model, empirical results, consumer surplus, and conclusion.

LITERATURE REVIEW

Assessing the economic value of ecosystems such as a beach is challenging because the intangible beach amenities that vacationers seek are not bought and sold in markets as are other commercial goods and services (Pendleton, 1999). Thus, determining value requires the estimation of how much money or purchasing power people are willing to give up to avail themselves of all that a particular beach has to offer. For the past several decades, the demand for recreational trips has been estimated using either direct or indirect method 1. In the direct method, vacationers are asked how much they would be willing to pay for an amount of recreation. The contingent valuation method (CVM) is a well-known approach to directly estimate the non-market value of recreational trips. Estimated values of a non-market good can be specified in monetary terms by willingness-to-pay (WTP) or willingness-to-accept (WTA). In the CVM approach, monetary values are based on the hypothetical questions associated with WTP or WTA for non-market goods.

On the other hand, the travel cost method (TCM) is one of the most popular indirect method approaches. Since Hotelling's letter was published in response to a US National Parks solicitation in order to value the economic benefits of National Parks (Hotelling, 1949), the TCM has been one of the useful tools to measure the value of a non-market resource. In the TCM approach, values for non-market goods can be inferred from the relationships between non-market use value and other market goods and services that are purchased as complements to a site visit (Bishop, 1979; Herath, 1999). The observed travel cost is used as a price proxy in this method 2.

Two major variants of the TCM are the zonal travel cost method (ZTCM) and the individual travel cost method (ITCM). In the ZTCM, the area surrounding the recreation site is divided into various zones of origin. Each zone has an associated average travel cost to the site (Garrod and Willis, 1999). The visitation rate per zone given time period, which is weighted by the number of visitors and the reverse of the sample size and its population, can be estimated on the average travel cost. According to Herath (1999), visits per thousand residents per year t (Vt)3 can be obtained as follows.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

where [R.sub.it] = the total population of residents in area i in time t; [V.sub.it] = visitors from area i in time t; [n.sub.t] = the sample size in time t; and [N.sub.t] = total number of visitors per week in time t.

Compared to the ZTCM, the estimation using ITCM is relatively straightforward when the individual number of visits correlates with travel cost and other economic and socio-demographic variables (Dobbs, 1993; Smith and Kaoru, 1990; Ward and Loomis, 1986). The Individual Travel Cost Method assumes that the value of the beach or the recreational activities it offers is reflected in how much people are willing to pay to get there. It is referred to as a "revealed preference" method, because it uses actual spending behavior to infer values. The premise of this method is that the time and travel cost expenses that tourists incur to visit a beach represent the recreational value of the beach. The advantages of the Individual Travel Cost Method are that it 1) imitates the conventional methods used by economists to estimate economic values based on market prices; and 2) it is based on what people actually do rather than on what people say they would do in a hypothetical situation (Bell and Leeworthy, 1990).

The Individual Travel Cost Method uses survey data from individual visitors to link the demand for tourism to its determinants. Determinants include how far the tourist must travel to get to the beach, the amount of time spent, travel and on-site expenses, how often they have visited the beach in the past, their income and other socioeconomic characteristics, etc. Because the tourist's costs will vary as the determinants vary, this method allows us to calculate the amount of beach visits "purchased" at different "prices." These values are used to construct the demand function for a beach vacation. The demand function relates price and quantity by illustrating how many units of a good will be purchased at different prices. In general, at higher prices, less will be purchased giving the demand function (the graphical representation of the demand function is referred to as the demand curve) a negative slope. Using survey data and regression analysis, we are able to estimate the demand function for the "average" visitor to the beach. This demand function, or demand curve, allows us to quantify the impact that changes in any of the determinants will have on the revenue generated by the local tourism industry.

Due to the weak theoretical foundation of the behavioral patterns in the aggregate demand models, the ZTCM has been often less preferred to the ITCM. Empirical studies provide mixed results (Cook, 2000; Hellerstein, 1995). The ZTCM is considered more appropriate to estimate consumer surplus when origins are uniformly distributed. The ZTCM is relatively more unsuitable for the case of multiple-destination of the recreational areas because of the difficulty of obtaining the site-specific travel cost estimates. Those difficulties can be overcome by adopting the ITCM, which is used in this study to estimate a recreational demand function for the Pensacola recreation area in the Northwest Florida.

DATA

Visitor data were collected between September 1999 and April 2002 at the four visitor information centers4 in the greater Pensacola area of Northwest Florida. These four visitor centers are located in two counties--Escambia and Santa Rosa in Northwest Florida. Walk-in visitors at each visitor information center filled out surveys in person. There was no respondent-selection procedure. Some people argue that walk-in visitor survey can be age-biased. Younger people are less likely to stop by visitor centers on highways to collect information. However, in the greater Pensacola area, all four visitor information centers are located in the center of each subdivision. Under- or over-representation of a specific group of population might not be significant. Surveys have been conducted year-round during the regular visitor center operation hours. Frequency varies month-to-month, which reflects the monthly variation of visitors. The questionnaire has been attached (See Attachment 2)

The total number of traveler groups included in this analysis is 8,625. 66.7% of respondents can be classified as vacationers. The others are business travelers (15.4%) and those who visited friends and relatives (17.8%). Almost 90 percent of visitors reside outside the local area. Half of the visitors have made multiple visits over the past five years. Top five reasons to visit the area are (1) beaches, (2) natural beauty of area, (3) climate, (4) quiet and relaxing atmosphere, and (5) cleanliness of area.

Table 1 shows the differences in means for several selected variables by pre-and post-9/11 attack. Vacation trips have significantly decreased from 67.2% to 62.6%. Trips by airplane also have decreased significantly from 12.1% to 10.1% while there is no change in auto trips. Visitors have stayed less nights (from 5.28 to 4.99) and spent less (from $203.55 to $190.68) during their stays. The portion of repeated visitors has increased which was measured by number of visits (from 2.63 to 2.84). It has the negative impact on the international travelers. U.S. citizens increased from 88.7% to 92.0%. The number of children in each travel group has decreased significantly from 0.6 to 0.4 persons.

The distance between origination and destination is calculated by using US Census data, based on the ZIP code information that each respondent provided. ZIP code coordinates, latitude and longitude, were obtained from the US Census STF-3 data sets. Given the latitudes and longitudes of the two points, the great circle distance between them can be calculated by the following formula (Paine, 1981).

d = R x arccos[sin([[pi].sub.1]) x sin([[pi].sub.2] + cos([[pi].sub.1]) x cos([[pi].sub.2] x cos([[gamma].sub.1] - [[gamma].sub.2])] (2)

where d = distance between the two points in km; R = radius of the earth in km, which is 6378.02km; [[pi].sub.1] = latitude of point 1 in radians; [[pi].sub.2] = latitude of point 2 in radians; [[gamma].sub.l] = longitude of point 1 in radians; and [[gamma].sub.2] = longitude of point 2 in radians.

THEORETICAL MODEL

This analysis assumes that a tourist's utility can be described in the following utility function

U = f(V,X) (3)

where V is the number of visits to a specific recreation area over a certain period of time, and X is a vector of all other goods and services. Demand for recreation can be expressed in various ways. One measure can be the nights of spent in a specific area or the length of stay, which is represented by V in this model. In order to differentiate outside visitors from local residents, only those who spent at least one night are considered in the estimation. The budget constraint can be specified as follows:

Y = pX + [alpha]V + [beta]T (4)

where Y = income; p = a vector of prices of other goods and services; X = a vector of other goods and services; [alpha] = price of demand for recreation, which is the actual cost per day; V = number of nights spent in a given period of time; [beta] = total cost per trip; and T = number of trip in a given period. Utility maximization given the budget constraint yields the following demand function for the recreation demand, V.

V = f([alpha],T,Y,X). (5)

Assuming that recreational demand is a normal good, it is hypothesized that V is positively related to T and Y while negatively related to . X consists of demand shifters, which are listed in the table shown in the section of the empirical results.

One of the response variables in X is the number of nights staying in the area. The upper open-ended interval of the range of the variable is '10 nights or more'. It is a very common way to define a variable in this type of survey questionnaires. To avoid the right-hand-side truncation bias, the censored regression model is used to provide more accurate results. The regression is obtained by making the mean in the preceding correspond to a classical ordinary least square model (Greene, 1993).

EMPIRICAL RESULTS

Travel cost is usually assumed to be positively correlated with the length of trip and negatively correlated to the frequency of trips. It has been widely accepted that the length of trip and the frequency of trip are substitutes in a given period of time (Font, 2000). However, for certain destinations or types of travelers in this study, empirical tests show that repeated visitors are likely to spend more days.

Table 2 shows the coefficient estimates and descriptive statistics from the ordinary least estimation for the number of nights for different classes of visitors: business travelers, vacationers, and those visiting friends and relatives (VFR).

Business travelers and vacationers arriving by airplane are more likely to stay longer than visitors using other forms of transportation, reflecting their higher opportunity costs for traveling. The effect of age is also significant, however it is positively related to length of stay for the business traveler, and negatively related to length of stay for the vacationer. U.S. citizenship, on the other hand, increases the length of stay for vacationers, while decreasing it for business travelers.

Distance is another important factor in explaining length of stay for business visitors and vacationers alike. The greater the distance traveled, the longer the stay. The average distance of travel is 1,083.7 km (673.53 miles). The winter dummy variable has significantly positive effects on the length of stay. Many of our winter visitors are known to be "snow birds" who spend their summers in northern states and winters in Florida.

"Total per Day Spending" represents the price of recreational services, and is a significant factor in length of stay. The negative coefficient illustrates that higher daily costs result in shorter visits for both business travelers and vacationers. The business travelers and vacationers that stayed the longest were those who planned the vacation at least a month in advance, and those who had visited the area previously. The greater the number of previous visits, the longer the stay.

Surprisingly, annual gross income does not play an important role in this demand model. It was hypothesized that higher-income individuals would spend more nights, but this was not supported by the results.

CONSUMER SURPLUS

Consumer surplus is estimated to measure the changes in welfare according to the changes in value of resources. This is represented graphically as the area under the demand curve and above the market price. When the average individual consumer surplus is multiplied by the total population of beach visitors, an estimate of the total consumer surplus for the beach is obtained. By changing value estimates of the various determinants of the demand function, one can estimate the effect they have on consumer surplus. Changing values generates two different demand curves, one for each level of the determinant. The area between these two curves is the estimate of the change in consumer surplus caused by a change in one of the determinants. This type of analysis allows us to estimate the change in recreational benefits that result from changes in the determinants of visitor spending behaviors.

Consumer surplus is widely accepted as a method to measure the changes in welfare according to the changes in value of resources (Hausman, 1981). However, there is relatively less agreement on how to calculate it (Bell and Leeworthy, 1990). From the above discussion, the demand function can be re-written as follows:

V = [ZIGMA] - [gamma] P + [delta]Y (6)

where N = number of nights, P = price of recreational services, Y = income, [ZIGMA] = sum of all demand shift factors except for Y, multiplied by their corresponding rates of returns, and [gamma] and [delta] are estimated parameters for price and income, respectively. Then the consumer surplus (CS) can be estimated as follows:

CS = 1/2 ([P.sup.*] - [bar.P]) [bar.V] (7)

where [P.sup.*] = intercept, and [bar.P] the corresponding price with mean value of P dependent variable, [bar.V]. To estimate CS, mean values of demand shifters except for V P are plugged into the demand function. This yields

V = 6.5512 - 0.0118 P (8)

Then, the demand equation is obtained as

P = 555.19 - 84.75 V (9)

Plugging the mean value of V, which is 5.0227, into the above equation, then.

The consumer surplus is estimated

CS = ($555.19 - $129.52) (5.0227) (0.5) = $1,069.01. (10)

Then, we can estimate the value of one day spent in the recreation area, which would be $69.9 [= 1069.01 (5.0227 3.0449)] per person when the average size of a travel group is 3.0449. Similarly, we could estimate the value of one day spent in the area for the business travelers and VFR, which are $78.58 and $49.84, respectively.

Using bed tax data for the local area we estimate that approximately 1.8 million tourists visit the Pensacola area each year. Survey responses tell us that our beaches draw tourists to the Pensacola area. Multiplying the number of visitors by the consumer surplus of $69.9 experienced by the average tourist, we estimate the total consumer surplus, or excess recreational value of the area beaches, at $125,820,000.

CONCLUSIONS

A recreation demand function is estimated for tourist areas in northwest Florida. Visitor behavior patterns, broken down by the purpose of trip, such as business, vacation, and visits to friends and relatives (VFR), are examined. Policymakers who need to know that the benefits of beach protection programs are greater than the cost to taxpayers have been provided with calculations of the consumer surplus, or recreational value, of the beaches in the Pensacola area of Northwest Florida. Tourism directors who need to allocate advertising expenditures have been provided with a description of important determinants of visitor length of stay, which is directly related to total visitor spending.

Determinants that are shown to have statistically significant positive impact on length of stay for vacationers include air mode of travel, U.S. citizenship, distance traveled, number of visits in the past five years, and length of time spent planning the vacation. Age was shown to have statistically significant negative impact on length of stay for vacationers. Annual income was found not to play an important role in vacationer's length of stay. Business travelers were shown to differ from vacationers in that older business travelers stayed longer, and non-U.S. citizens here on business had a shorter length of stay. Income has a significant positive impact on length of stay for business travelers.

REFERENCES

Bell, F., & Leeworthy, V. (1990) Recreational Demand by Tourists for Saltwater Beach Days. Journal of Environmental Economics and Management, 18, 189-205.

Bishop, F., & Heberlein, V. (1979) Measuring Values of Extramarket Goods: Are Indirect Measures Biased? American Journal of Agricultural Economics, 71, 926-930.

Cook, A. (2000) Estimation of Community Values of Lakes: A Study of Lake Mokoan in Victoria, Australia: A Comment. Economic Analysis and Policy, 30(1), 99-102.

Dobbs, I. M. (1993) Individual Travel Cost Method: Estimation and Benefit Assessment with a Discrete and Possible Grouped Dependent Variable. American Journal of Agricultural Economics, 75(10), 84-94.

Eberle, W., & Hayden, F. (1991) Critique of Contingent Valuation and Travel Cost Methods for Valuing Natural Resources and Ecosystems. Journal of Economic Issues, 25(3), 649-687.

Feather, P., Hellerstein, D., & Tomasi, T. (1995) A Discrete-Count Model of Recreational Demand. Journal of Environmental Economics and Management, 29, 214-227.

Font, A. (2000) Mass Tourism and the Demand for Protection Natural Areas: A Travel Cost Approach. Journal of Environmental Economics and Management, 39, 97-116.

Garrod, G., & Willis, K. G. (1991) Economic Valuation of the Environment: Methods and Case Studies. Massachusetts: Edward Elgar.

Greene, W. (1993) Econometric Analysis (2nd ed). New York: Macmillan.

Hausman, J. (1981) Exact Consumer's Surplus and Deadweight Loss. American Economic Review, 71(4), 662-676.

Hellerstein, D. (1995) Welfare Estimation Using Aggregate and Individual-Observation Models: A comparison Using Monte Carlo Techniques. American Journal of Agricultural Economics, 77(3), 620-630.

Herath, G. (1999) Estimation of Community Values of Lakes: A Study of Lake Mokoan in Victoria, Australia. Economic Analysis and Policy, 29(1), 31-44.

Hotelling, H. (1949) Letter to the National Park Services: An Economic Study of the Monetary Evaluation of Recreation in the National Parks. Washington D.C.: U.S. Department of the Interior, National Park Service/Recreational Planning Division.

Paine, D. (1981) Aerial Photography and Image Interpretation for Resource Management (1st ed). New York: Wiley.

Parsons, G., & Kealy, M. (1995) A Demand Theory for Number of Trips in a Random Utility Model of Recreation. Journal of Environmental Economics and Management, 29, 357-367.

Pendleton, L. (1999) Reconsidering the Hedonic vs. RUM Debate in the Valuation of Recreational Environment Amenities. Resource and Energy Economics, 21, 167-189.

Pendleton, L., & Mendelsohn, R. (2000) Estimating Recreation Preferences Using Hedonic Travel Cost and Random Utility Models. Environmental and Resource Economics, 17, 89-108.

Randall, A. (1994) A Difficulty with the Travel Cost Method. Land Economics, 70(1), 88-96.

Smith, V., & Kaoru, Y. (1990) What Have We Learned since Hotelling's Letter? A Meta-Analysis. Economics Letters, 32, 267-272.

Ward, F., & Loomis, J. (1986) The Travel Cost Demand Model as an Environmental Policy Assessment Tool: A Review of Literature. Western Journal of Agricultural Economics, 11, 164-178.

Woodward, R. (2001) The Welfare Impacts of Unanticipated Trip Limitations in Travel Cost Models. Land Economics, 77(3), 327-338.

Inhyuck "Steve" Ha, Western Carolina University

ENDNOTES

(1) In the estimation of non-market valuation, as a third category, a discrete-choice modeling approach has been recently recognized and used extensively, which is based on the Random Utility Model (RUM) theory. For more details, see Feather, et al (1995), Parsons and Kealy (1995), Pendleton and Mendelsohn (2000), and Woodward (2001).

(2) For critiques of TCM and CVM, see Eberle and Hayden (1991), and Randall (1994).

(3) In the estimation, the visit rate, i.e., participation rate of each zone, is estimated in the ZTCM while the actual number of visits is estimated in the ITCM.

(4) Four visitor information centers (VIC's) are (1) Pensacola VIC, (2) Perdido Key VIC, (3) Pensacola Beach VIC, and (4) Navarre VIC.

Table 1:. Mean Difference Tests: Before and After 9/11

 Before 9/11

 N Mean

Purpose of Trip
 Business 6518 15.1%
 Vacation 6518 67.2%
 VFR 6518 17.5%
Type of Transportation
 Airplane 7048 12.1%
 Auto 7048 82.6%
 Other Vehicle 7048 5.4%
Visiting Patterns
 Number of Visits 6867 2.63
 Number of Nights 7142 5.28
Spending Patterns
 Per Day Spending on Lodging 4426 $88.71
 Per Day Spending on Grocery 3772 $26.87
 Per Day Spending on Restaurants 4879 $43.04
 Per Day Spending on Entertainment 3438 $32.69
 Per Day Spending on Shopping 3767 $44.73
 Per Day Spending on Others 2439 $37.43
 Total Per Day Spending 5305 $203.55
Tourism Destinations
 No Other Destinations 7142 42.9%
 Mississippi Casinos 7142 10.7%
 New Orleans Area 7142 15.3%
 Orlando Area 7142 7.7%
 Ft. Walton Beach/Destin Area 7142 18.3%
 Mobile Area 7142 12.5%
 Orange Beach/Gulf Shores Area 7142 10.5%
 Panama City Area 7142 13.3%
 Other 7142 17.7%
Demographic Information
 Age 6654 47.6
 Married 7000 71.6%
 White 6933 87.9%
 US Citizen 6989 88.7%
 Number of Children in the Household 6896 0.64
 Number of Children in Travel Group 6826 0.60
 Number of Adults in Travel Group 6958 2.53
Economic Information
 Annual Household Gross Income 5865 59,531

 After 9/11

 N Mean

Purpose of Trip
 Business 1379 16.6%
 Vacation 1379 62.6%
 VFR 1379 18.9%
Type of Transportation
 Airplane 1471 10.1%
 Auto 1471 82.8%
 Other Vehicle 1471 7.1%
Visiting Patterns
 Number of Visits 1399 2.84
 Number of Nights 1483 4.99
Spending Patterns
 Per Day Spending on Lodging 671 $96.49
 Per Day Spending on Grocery 583 $22.48
 Per Day Spending on Restaurants 833 $42.13
 Per Day Spending on Entertainment 513 $32.97
 Per Day Spending on Shopping 595 $43.09
 Per Day Spending on Others 402 $37.26
 Total Per Day Spending 894 $190.68
Tourism Destinations
 No Other Destinations 1483 33.2%
 Mississippi Casinos 1483 11.1%
 New Orleans Area 1483 15.0%
 Orlando Area 1483 5.9%
 Ft. Walton Beach/Destin Area 1483 22.0%
 Mobile Area 1483 13.2%
 Orange Beach/Gulf Shores Area 1483 12.7%
 Panama City Area 1483 13.1%
 Other 1483 16.2%
Demographic Information
 Age 1289 49.3
 Married 1441 75.7%
 White 1440 83.3%
 US Citizen 1470 92.0%
 Number of Children in the Household 1421 0.47
 Number of Children in Travel Group 1387 0.40
 Number of Adults in Travel Group 1432 2.55
Economic Information
 Annual Household Gross Income 1167 $62,378

 t-stat. p-value

Purpose of Trip
 Business 1.377 0.169
 Vacation -3.224 0.001 ***
 VFR 1.142 0.254
Type of Transportation
 Airplane -2.268 0.023 ***
 Auto 0.219 0.826
 Other Vehicle 2.414 0.016 **
Visiting Patterns
 Number of Visits 2.871 0.004 ***
 Number of Nights -3.169 0.002 ***
Spending Patterns
 Per Day Spending on Lodging 2.300 0.021 **
 Per Day Spending on Grocery -4.273 0.000 ***
 Per Day Spending on Restaurants -0.588 0.557
 Per Day Spending on Entertainment 0.135 0.893
 Per Day Spending on Shopping -0.576 0.565
 Per Day Spending on Others -0.055 0.956
 Total Per Day Spending -2.006 0.045 **
Tourism Destinations
 No Other Destinations -7.202 0.000 ***
 Mississippi Casinos 0.517 0.605
 New Orleans Area -0.287 0.774
 Orlando Area -2.600 0.009 ***
 Ft. Walton Beach/Destin Area 3.138 0.002 ***
 Mobile Area 0.726 0.468
 Orange Beach/Gulf Shores Area 2.343 0.019 **
 Panama City Area -0.198 0.843
 Other -1.419 0.156
Demographic Information
 Age 3.444 0.001 ***
 Married 3.272 0.001 ***
 White -4.274 0.000 ***
 US Citizen 4.191 0.000 ***
 Number of Children in the Household -5.373 0.000 ***
 Number of Children in Travel Group -6.205 0.000 ***
 Number of Adults in Travel Group 0.712 0.477
Economic Information
 Annual Household Gross Income 2.311 0.021 *

Source: VISIT System Data, April 2003

Note: Only overnight visitors are included.

*** significant at 99%, ** at 95, and * 90% levels

Table 2: Coefficient Estimates of Linear Regression
Model: Number of Nights

 Business
 Overall Only

Constant 4.1060 *** 3.5256 ***
Business 0.3164 --
Vacation -0.0788 ** --
Airplane 0.6497 *** 1.3440 ***
Automobile -0.3319 *** 0.3894
Age -0.0210 *** 0.0409 ***
Age-squared 0.0003 *** -0.0007 ***
US citizen 0.1470 -0.6020 ***
Annual Income -2.63E-06 *** 0.0000 ***
Number of Visits 0.2691 *** 0.2266 ***
Distance 0.0006 *** 6.96E-04 ***
Spring -0.5176 *** -0.0910
Summer -0.2159 ** 0.8121 ***
Fall -0.6083 *** 0.4588 *
Planned at least a month ago 1.1346 *** 1.3512 ***
Pensacola Area Only 0.3137 *** 0.2219
Total per day spendings -0.0125 *** -0.0114 ***
Total spendings 0.0019 *** 0.00162 ***

Mean of Dependent 5.1475 5.4617
Number of Observations 5614 823
p-value <0.0001 <0.0001
Adjusted R-square 0.4482 0.3768

 Vacation Visit Friends
 Only or Relatives

Constant 3.7380 *** 5.5610 ***
Business -- --
Vacation -- --
Airplane 0.5072 ** 0.1387
Automobile -0.2914 *** -0.8371 ***
Age -0.0318 *** -0.0121
Age-squared 0.0005 *** 0.0001
US citizen 0.4829 *** -0.1889
Annual Income 2.01E-07 -3.98E-06 **
Number of Visits 0.2443 *** 0.2852 ***
Distance 0.0005 *** 0.0005 **
Spring -0.6386 *** -0.2137
Summer -0.5218 *** 0.1225
Fall -0.8170 *** -0.5328 ***
Planned at least a month ago 1.1792 *** 0.3832 ***
Pensacola Area Only 0.4019 *** 0.0633
Total per day spendings -0.0118 *** -0.0184 ***
Total spendings 0.00185 *** 0.00272 ***

Mean of Dependent 5.0227 5.3868
Number of Observations 3833 954
p-value <0.0001 <0.0001
Adjusted R-square 0.5067 0.3948

*** significant at 99%, ** at 95, and * 90% levels

Source: VISIT System Data, April 2003