A Comment on the Draft Environmental
Impact Statement for Tusayan Growth
Dennis Foster and John Eastwood
Bureau of Business and Economic Research
College of Business Administration
Northern Arizona University
August 1997
Summary
The Bureau of Business and Economic Research (BBER) has been contracted, in part, to evaluate and analyze the Draft Environmental Impact Statement for Tusayan Growth, issued by the Kaibab National Forest, United States Department of Agriculture. At the core of the draft EIS is a consideration of a proposed exchange of land between the Kaibab National Forest and privately held lands scattered throughout the forest. The proposed alternative, which initiated the EIS process, calls for commercial development on the exchanged land, centered around the construction of 3650 lodging units.
The draft EIS considers five alternatives. Alternative A is referred to as the “no action” alternative, as the current situation is not altered in any dramatic way. Alternative B is the main Canyon Forest Village proposal, which plans for the construction of 3650 rooms. Alternative C is a scaled down version of Alternative B, with Canyon Forest Village building only 2250 rooms. Alternative D is a proposal by the Grand Canyon Improvement Association that calls for Forest Service land to be used for community and infrastructure purposes, but not commercial purposes. Alternative E involves the use of Forest Service land by the National Park Service for housing and as a transportation staging area. The focus of this comment is on Alternatives A, B and C.
BBER has concluded that the economic analysis in the draft EIS suffers from an incomplete accounting of relevant economic impacts, a failure to consider both the seasonality in Grand Canyon visitation and the potential impacts on Tusayan, and a calculation of displacement effects that may emanate from the construction of Canyon Forest Village. Consequently, the resulting impacts detailed in the draft EIS are unreliable and should be re-examined.
In attempting to begin addressing these shortcomings, we have constructed a lodging demand model and used it to generate a likely range of economic impacts. We have found that the potential economic disruptions stemming from either Canyon Forest Village alternative might be quite dramatic for Tusayan area and/or for the surrounding communities of Williams and Flagstaff. This conclusion is in sharp contrast to the results drawn in the draft EIS.
Following our discussion of these impacts, we will turn to a review of the commentary made by BBC Research & Consulting on the draft EIS, look at a host of lesser concerns about the draft EIS and comment on the role of the Forest Service in interacting with a well-established market-place.
Section 1:
An Analysis of the Draft Environmental
Impact Statement for Tusayan Growth
1.1 Overview
This analysis will begin with three specific concerns that are embedded in each draft EIS alternative. We then turn to individual assessments of alternatives A, B and C. Alternatives D and E have the same essential economic impact as Alternative A, and, as in the draft EIS, we will not subject them to separate analysis.
1.2 EIS Economic Analysis Methodological Concerns.
Incomplete accounting of economic impacts. The draft EIS tends to state effects and impacts in a fashion that is, at times, less than revealing, especially insofar as making an informed judgment is concerned. There are two serious problems here. First, results are generated only for some alternatives (notably, only for Alternative B) and only for some growth rates (generally, only 4%). Secondly, the information is often presented only as a snapshot view for the year 2010, while failing to characterize what will happen during the intervening years.
The following table shows what information is presented in the draft EIS only as a snapshot view in the year 2010:
Table 1: Draft EIS Data Shown Only for the Year 2010
GCNP Visitation / Table 5-2; p. 104GCNP/Tusayan Annual Lodging Demand / Table 5-3; p.105
Acreage requirements / Table 5-4; p.106
Lodging employees and housing demand / Table 5-5; p. 107
New retail demand / Table 56; p. 108
Retail employees and housing demand / Table 5-8; p. 110
New food and beverage area / Table 5-9; p. 111
Food and beverage employees and housing demand / Table 5-10; p. 112
Summary of Land Use Demand - 70% Occupancy / Table 5-11; p. 113
Potential Annual Visitor Spending / Table 6-5; p. 128
Annual Displaced Sales Taxes / Table 6-6; p. 128
Potential new lodging units in northern Arizona,
for Alternatives A and C / Table 6-7; p. 129
Community net new rooms / Table 6-8; p. 130
Community net new spending / Table 6-9; p. 131
Tables and page numbers refer to the “Economic Analysis” section of the Appendix of the Draft Environmental Impact Statement for Tusayan Growth.
We would agree that some of this information is most usefully presented in this particular manner. Such would be the case with employees, housing and space-related figures. However, demand, visitation, displacement, spending and tax figures will mask patterns that develop over the intervening years that are potentially critical to appropriate decision-making. For example, the Canyon Forest Village proposals have most lodging being built early on over the 1997-2010 time frame. Thus, we would expect that dislocation and disruption effects, with respect to rooms and spending in outlying areas, to be more severe early on, and somewhat reduced by the year 2010. The presentation of these values only for the year 2010 might easily lead to false inferences about the conditions expected during the in-between years.
This problem of interpretation is especially apparent for the data on “Potential Annual Visitor Spending” (Table 6-5; p. 128). Here, dollars of additional spending are derived for the year 2010 in various communities (for the Alternative B scenario only, at the “moderate” growth rate of 4%). These values are likely to be quite variable during the in-between years, and it is certain that in the year 2010 these figures will be at their highest.
This same format is also used to show tax revenue generated by Alternative B (again for a growth rate in visitation of 4% per year), and shown as “Annual Displaced Sales Taxes” in Table 6-6 (p. 128). In a reverse subtraction ordering (that is, to get the number in the last row, subtract the number in the first row from the number in the second row) of this data, the additional tax revenue is shown, on an annual basis, but only for the year 2010. Again, this must be the highest value generated over the fourteen year time period.
A failure to consider seasonality of visitation and lodging demand. Although the draft EIS does show the monthly park visitation for 1996, and uses some inferred occupancy rates to derive monthly room demand, explicit consideration of seasonality is excluded from the analysis. To some extent, the annualization of values used in the EIS model is not especially burdened by this omission. However, there are some significant consequences that are overlooked.
Because visitation to the Grand Canyon, and, by inference, the demand for lodging, is sharply peaked during the summer months, it should be recognized that a persistence of the excess demand for rooms, during the summer peak, is unavoidable. Figure1, Grand Canyon/Tusayan Lodging Demand and Supply (1996), shows the daily supply and demand for rooms in the Grand Canyon/Tusayan area. The supply is the currently available 1964 rooms, reported in the EIS. The demand for rooms is derived from the information on visits contained in the EIS, the characteristics of lodging parties, and an assumed “penetration rate” of 30%.[1] The figure shows two obvious aspects of this market--excess supply during the off-peak months and excess demand during the peak months.
The consequence of this seasonality on Tusayan cannot be understated. The EIS has implied that, given a higher cost structure in Tusayan, the average annual occupancy rate that is necessary for these hotels to be profitable is about 70%. The shaping of this average has been seen to come from three component parts--very low occupancies in the offpeak months, moderate to high rates during the shoulder months and very high rates during the peak months. This is shown, in Figure 2, Tusayan Monthly Occupancy Rates - 1996.
While the occupancies enjoyed by Tusayan hotels from May through October are well above the national average cited in the EIS, their annual average is relatively low, dragged down by exceedingly poor performance during the off-peak months of November through February. The addition of between 2000 and 3600 more rooms, as considered in the two Canyon Forest Village proposals, would almost certainly reduce these monthly occupancies in months where there exists excess supply. However, even if we ignore the possible deterioration of occupancy rates during the peak, it is a virtual certainty that Tusayan’s rates will fall during the off-peak. Whether these hotels can survive if they lose these customers is directly related to the issues of seasonality. For example, if the offpeak rates for November through February were to fall to 10%, and the shoulder rates for March, April and October fell to 50%, and all other months remained as shown, the average occupancy would fall to 51%. It is very unlikely that these hotels would stay in business for very long with this low an occupancy rate.
No economic impacts are generated for Tusayan. This omission is one of the most substantial in the draft EIS. The calculation of “displacement” effects due to Canyon Forest Village completely ignores the potential impact on Tusayan. There are two different versions of what may be called a “worst-case” scenario that must be considered here. One is where the impact of Canyon Forest Village falls on Tusayan. The other is where the impact skips Tusayan and is primarily felt by downstream markets. The draft EIS considers only the latter possibility, but not the former. However, it is far more likely that Tusayan would experience the bulk of this “displacement,” and it is incumbent for the EIS to explicitly consider such an outcome.
The displacement issue, and how it relates to Tusayan, comes down to the interrelationship between two factors--whether rooms are/are not substitutable and whether there is much excess demand. As to the first factor, rooms in Canyon Forest Village are expected to attract visitors away from other hotels if they are substitutes for one another, but not if they are complements. If we presume that rooms closer to the Grand Canyon are more desirable than rooms further away, and if the lodging in the canyon is full, then Canyon Forest Village would be expected to draw business away from hotels located farther away. Given that it would directly compete with the hotels in Tusayan, and noting the locational advantages enjoyed by Canyon Forest Village, it would seem most likely that visitors would be quite willing to substitute out of Tusayan and into Canyon Forest Village. On the other hand, rooms in Williams and Flagstaff are relatively less likely to be substitutes for rooms at Canyon Forest Village. That is, it is more likely that an overnight visitor in Flagstaff will not switch because the overnight stay in Flagstaff may be better suited to their travel itinerary.
Another factor is the currently observed relationship between supply and demand for rooms in Tusayan. As suggested by Figure 2, there is monthly excess supply of lodging rooms in Tusayan, albeit relatively narrow during the super peak months of July and August[2]. If Canyon Forest Village does enjoy a competitive locational advantage, then raising the number of rooms from 931 (the current number in Tusayan) to 3181 (Alternative C; the small version of Canyon Forest Village), or to 4581 (Alternative B; the full version of Canyon Forest Village) would seem most likely to lead to a deterioration of the market share currently flowing to Tusayan.
Incomplete identification of displacement effects. The draft EIS uses the result of a visitor survey, conducted three times from 1994 to 1995. The design of this survey was ill-suited to determining what the displacement effects might be with the addition of Canyon Forest Village in the Tusayan area. BBC Research & Consulting has commented on this problem at some length and that argument will not be repeated here.
We do note, however, the serious conclusion that follows from this critique. The draft EIS uses the results of the visitor survey to infer where visitors may be displaced from and then uses this to estimate the impact on room demand and lodging-related spending in outlying areas, under the Canyon Forest Village alternatives. As presented in the EIS, we believe that these inferences are subject to wide error and are unacceptable. The main problem is that the survey does not really allow us to determine where displacement would occur. If, as the BBC analysis suggests, the displacement will mostly fall on northern Arizona, then the EIS results will not represent, in any meaningful way, a “worst-case” scenario. The importance of this is clear. Where the EIS has concluded, in the “worst-case” outcome, that rooms will be added in surrounding communities, and spending increased by the year 2010, then it becomes easy to conclude that Canyon Forest Village will have no absolutely negative impact. When the “worst-case” scenario is shown to be something far more devastating, as in the BBC comment, this conclusion is much more difficult to make, and the risk of causing serious economic disruptions is much more likely.
We believe that a closer scrutiny of cross-tabulations derived from the EIS visitor survey may allow for a better inference about the displacement effects. Further, we feel that the inclusion of a truly off-peak survey of visitors would go a long way to addressing the issue of where rooms are substitutes for one another and where they are complements. As conducted, the survey actually was done at the beginning, middle and end of peak visitation, especially insofar as lodging demand is concerned. With reference to Figure 1, these surveys were done in late April/early May, late July/early August and in late September/early October.
If a similar survey were to be done during the November to February time frame, visitation patterns could be compared with the peak results. That is, since there is a great deal of excess supply in the Tusayan area during the off-peak, we would expect visitor itineraries to reflect that. That could be contrasted with the peak season visitation pattern to observe what differences there might be. This would help identify to what extent rooms in surrounding communities are substitutes or complements for one another. For example, suppose that an off-peak survey found that the so-called “induced demand” for rooms in Flagstaff was only 30% of canyon visitors. During the peak the draft EIS survey found this proportion to be 39%. Consequently, one may infer that the residual difference constituted some excess demand for rooms in the Tusayan area. This would, then, represent a pool of visitors that might be easily displaced from Flagstaff.
1.3 An Analysis of Alternative A: No Action
The “no-action” alternative under consideration in the draft EIS would not result in any land being traded by the Forest Service. The relevant issues to consider, under this alternative, are:
1. What are the current supply and demand conditions in the Grand Canyon/Tusayan area?
2. How do we expect these conditions to change over the next fourteen years?
The current level of “visits” are identified in the draft EIS, and shown in the form of lodging demand. The current supply of rooms available is shown and some estimation of occupancy rates derived. The general approach is suitable for reaching an understanding of current conditions. We have a number of comments to offer in this regard that we believe can strengthen the EIS analysis.
With respect to future conditions in the Grand Canyon/Tusayan area, there will certainly be much speculation. There are two major sources of disagreement we have with the draft EIS analysis of this alternative--the growth scenarios are not well characterized and the potential for future development is greatly exaggerated.
The use of “visit” data. The draft EIS derives demand by observing current “penetration” rates of visits and applying an average of that to the current visit count. Aside from those who have put together the draft EIS, and those that have studied it closely, this methodology is awkward and has the potential to be easily misinterpreted. The reason for this is that “visits” do not translate on a one-for-one basis into “visitors,” or, more appropriately, into recreational visitors. As the appropriate basis for studying all impacts in the draft EIS, starting with a “recreational visitor” count would greatly enhance a clear understanding of current conditions.