Mid-Wilamette Valley Intermodal Center

Legislative Intent: Mid-Willamette Valley

The Oregon Business Development Department (dba Business Oregon) breaks Oregon into economic regions based on economic activity and geography. The Willamette Valley Economic Region is described as Benton, Marion, Polk, Yamhill, Linn, and Lane Counties. The LEDG Project Proposal is centrally located in the Willamette Valley economic region. When the Legislature allocated $25 million for the development of an intermodal facility in the Keep Oregon Moving (HB 2017-A), they did so with the intention that the facility will be centrally located in the Mid- Willamette Valley. We believe the proposed site in Millersburg provides the best location to meet the intent. Joining in this belief is the unanimous support of the county commissions in Benton, Lane, and Linn Counties.

Mid-Willamette Valley Industrial Park

& Intermodal Center Project Site

Proposed Property

The LEDG proposal would include the development

and construction of an intermodal facility at the

former International Paper (IP) Mill site located

along Interstate-5 in Millersburg. The property has direct access to both the Union Pacific Railroad (UPRR) and Portland & Western Railroad (BNSF).

The proposed site has an existing 60,000 square foot warehouse with a rail siding and consists of over 140 acres of industrially zoned land. The site is considered Zone H in the Mid-Willamette Valley Industrial Park - Millersburg which together has over 400 acres of industrially zoned land ready for development (see image below).

Location Comparison Analysis

As part of the LEDG's market feasibility study, LEDG requested ECONorthwest to evaluate the set of economic conditions that vary along the Interstate 5 (I-5) corridor in the Mid-Willamette Valley which may affect the respective geographic market, market share, and public benefits if the proposed intermodal center were to be constructed in a different location than the proposed site in Millersburg, Oregon. Specifically, this analysis evaluates three alternative geographic locations in Oregon (South to North):

Eugene: ~46 miles south of Millersburg,
Lebanon: ~16 miles southeast of Millersburg, and
Brooks: ~28 miles north of Millersburg.

This analysis evaluates the following elements:

anticipated market area within a one-day “turn” of each location,
share of the market likely to use the intermodal center,
relative change in user transportation cost savings of each location, and
relative change in public benefits of each location.

Anticipated Market Area

Utilization of an intermodal center in any location is a function of the demand for transportation services and the availability of substitutes. Currently, Willamette Valley shippers transport agricultural commodities by truck to either their final destination or to intermediate intermodal facilities in Portland. These shippers are spatially distributed throughout the valley, however, all products destined for ports in Seattle and Tacoma pass through Portland on trucks on I-5. Intermodal facilities in Portland are ideally located to serve as a feasible alternative for all containers heading north. The decision to stop in Portland and transfer to rail – as opposed to continuing by truck to the ports – is determined by the marginal cost to truck the container the remaining ~175 miles compared to the cost to transfer the container to rail in Portland, rail charges between Portland and Seattle, and the level of service of rail (e.g. timeliness and reliability).

Figure 1 shows the relative areas within a five-hour drive of each potential location. Locations further south (i.e. Eugene) slightly expand the market area further than Millersburg, while locations further north (i.e. Brooks) slightly reducing it.

Construction of an additional intermodal center south of Portland introduces a new mode alternative but does not fundamentally change the preference structure for shippers. That is, shippers will continue to make decisions based upon the optimal mix of price and quality of service. The new set of transportation options for all shippers located south of the proposed intermodal center is described in Figure 2.

As noted in the main report, the differential price for trucking is calculated by fitting a regression model to a sample of 683 drayage rates from Oregon to Seattle/Tacoma ports from May 2014 to June 2018, obtained from The Drayage Directory. The differential price for rail is generated from two costs: the expected rail transit charges, estimated from the most recent complete version (2016) of the Surface Transportation Board Carload Waybill Sample, plus the projected intermodal center charges identified in the Capital and Operating Cost analysis in the main report. The expected differential prices of all alternatives are calculated for a hypothetical shipper located along the I-5 corridor south of Portland. The expected values for existing transportation alternatives (e.g. truck all the way to Seattle or transfer at a Portland intermodal center) are described in Figure 3. For simplicity, destinations other than the ports of Seattle and Tacoma are excluded from this analysis.

The trucking cost function exhibits both fixed and variable cost components, with decreasing marginal per-mile costs with distance. Additionally, the data indicate that there is a minimum drayage rate for any distance before the per-mile costs begin affecting the total price. This causes a non-linear kink for trucking distances under approximately 90 miles. Shippers with their own trucking resources may be able to achieve lower per-mile costs than the drayage rates observed on the open market. There is an additional non-linear kink that distinguishes a one- day versus multiple-day transit since additional costs are incurred when a driver spends a night away from home. This occurs at approximately the 350-mile point. Including the distance to Seattle, this occurs approximately 175 miles south of Portland, and can be seen in Figure 3. As a result of this expected cost structure, intermodal transfers in Portland are price- competitive with trucking the entire distance to Seattle from all likely locations in the Willamette Valley.

Screen Shot 2018-09-29 at 1.29.50 PM.png

Thus, the transportation costs of using any new intermodal center south of Portland should be compared to the cost of transferring at existing facilities in Portland. As in the example above, all shippers located south of Portland and the new intermodal center incur costs to a) truck to the intermodal center, b) transfer to rail, and c) transit on rail for the remaining distance.

However, a subset of shippers located south of Portland but north of the new intermodal center must decide to either truck south to the new intermodal center or truck north directly to Portland. Transportation options for these shippers are displayed in Figure 4.

Comparison of expected differential prices for shippers is performed using the same cost functions described above for possible facilities located in either Brooks, Millersburg, Lebanon, or Eugene, Oregon. These cost functions are displayed in Figure 5. The dots identify the location of each potential intermodal center.

Using the expected cost functions and a set of general assumptions, an approximation can be made of the market area (by the distance from Seattle) for each location. The specific assumptions are: 1) Only one new intermodal center will be constructed, 2) intermodal center access costs are equal across all potential locations, 3) railcar and container availability is equal across all potential locations, and 4) shippers located between the potential location and Portland chose either intermodal center with equal probability unless a clear price differential exists. This results in three choices displayed in Figure 6 below. One is the market area that is most likely to use existing facilities in Portland. Second is the market area that is closer to the potential location, however, Portland remains a price-competitive alternative. The third is the potential market area within a one-day drive that observes lower costs through the new intermodal center.

A clear tradeoff exists between locations. Potential locations further north capture a greater portion of the market in the northern Willamette Valley, however, they also compete with existing intermodal facilities in Portland. Potential locations further south capture a greater portion of the rural southern and eastern market within a one- day drive of the intermodal center, however fewer trucks from northern portions of the Willamette Valley will ship south to the intermodal center and be diverted from roadways in Portland.

Relative Anticipated Market Share

Using the distance bands identified by the process above with the estimate of containers transiting through the Millersburg location as a baseline, a rough approximation can be made of the volume of containers passing through each intermodal center and thus diverted from roadways in Portland. Additionally, the volume of likely products in each distance band will also determine the number of containers. As described in the full report, the major products likely to pass through an intermodal center in the valley destined for international export are low-value, high volume products such as hay and straw while the largest domestic product share is expected to be grass seed. The approximate distance bands overlaid with county-level production estimates of hay and grass seed can serve as a proxy and benchmark for the volume of containers expected to pass through each potential location. Figure 7 shows the total and share of hay and grass seed produced by each county in the expanded study area, which also includes Yamhill and Clackamas counties.

Each of the production estimates for hay and grass seed is assigned to respective facilities to represent the universe of goods likely to use the intermodal center. The baseline estimate for Millersburg includes all counties from Marion through Klamath but excludes Yamhill and Clackamas counties, as these latter two counties are further north and shippers may be more likely to use existing facilities in Portland as opposed to trucking south. The Brooks estimate includes all counties, adding Yamhill and Clackamas. Although much of these area in these two counties is still north of the potential facility location in Brooks, their closer proximity makes it a more feasible alternative. Shippers in Yamhill and Clackamas counties may opt to turn south to the Brooks location as opposed to shipping to Portland but are less likely to do so for the Millersburg location. The Eugene and Lebanon locations only include Lane through Klamath counties, as shipper further north are outside the market geography for these two sites.

By applying level adjustments to the mode choice model estimated for the Millersburg intermodal center in the main report, the broad set of shipping decisions including price, timeliness, and reliability are already captured. These adjustments only reflect differences in the extent of the market due to the respective location. By being able to capture additional volume originating in Yamhill and Clackamas counties, the Brooks location may observe an 11% higher volume of containers. Since this new area overlaps with the region that is already price competitive with existing facilities in Portland, any increase is likely to be diverted from existing intermodal transfers in Portland. The Lebanon and Eugene locations may observe 34% lower volumes due to their inability to capture products originating further north.

Relative Transportation Cost Savings

Private Benefits

As in the main analysis, the volume of exported and domestic containers will replace either the trucking of loaded containers all the way to ports in Seattle and Tacoma or trucking loaded containers to an intermodal center in Portland. As in the main analysis, there may be cost savings incurred by transporting containers by rail from a proposed intermodal center to the Seattle/Tacoma area as opposed to trucking the entire way. However, to generate conservative estimates of benefits, only truck transportation cost savings from each location to Portland are included. Since options to ship via rail from Portland to Seattle/Tacoma already exist and the price implications of additional competition from the proposed facility are uncertain, savings on this stretch are excluded from the calculation.

Truck transportation cost savings are reduced by the additional rail transit costs to cover the distance from the intermodal center to Portland. Both elements are calculated using the truck and round-trip marginal rail cost structures used above. Intermodal center charges are likely to be roughly equivalent at all facilities and are excluded from the estimate of private transportation cost savings. Additionally, any rail savings costs from Portland to Seattle are likely to be equivalent across all locations and are also excluded. The resulting calculation is as follows:

Potential value of private transportation cost savings:

Private Transportation Cost Savings = (Marginal cost to ship to Portland by Truck – Marginal cost to ship to Portland by Rail) * (Distance) * (Containers)

Both the trucking and rail marginal cost functions are observed to decrease over distance, however, as shown in Figure 9 average trucking costs per mile are relatively constant over short distances.

Thus, the marginal rail cost per mile (calculated based on the entire distance to Seattle/Tacoma) decreases over the respective site locations, while trucking costs do not. Calculations for each location are shown in Figure 10 below. Annual savings estimates are rounded to the nearest thousand. When evaluated over a twenty-year timeframe—from 2020 to 2040—at a 3% and 7% discount rate, these savings amount to the values by location listed in Figure 11. These transportation cost savings are likely to be gained in the private market by the growers, shippers, the intermodal center operator, or the rail operator.

Public Benefits

Public benefits accrue from reduced pollution, congestion, highway wear and tear, and fewer accidents. The categories calculated include: 1) Highway Safety, 2) Air Pollution and Greenhouse Gas Reduction, 3) Reduced Highway Maintenance Costs, and 4) Congestion Reduction. The first three categories accrue based on the total highway miles driven, while congestion reduction occurs primarily in the Portland metro area and is mostly independent of the miles of highway driven outside of this region. Thus, to calculate a relative change in public benefits across locations, some estimates are scaled by both projected volume and miles driven, while congestion is scaled only by volume. Figure 12 lists the estimated public benefits for the Millersburg location from the main report. Figure 13 lists the adjustment factors relative to Millersburg, for each location. The congestion adjustment reflects the change in volume, while the distance & volume adjustment is the product of distance and volume.

Applying these adjustment factors to the categories of public benefits calculated for the Millersburg intermodal center baseline produces the following estimates for each location, displayed in Figure 14 below. Values are rounded to the nearest thousand.

When evaluated over a twenty-year timeframe – from 2020 to 2040 – at a 3% and 7% discount rate, these savings amount to the values by location listed in Figure 15.

Public benefits accrue based on the volume of trucks removed from highways in the Portland area (i.e. congestion benefits) plus the per-mile benefits of removing trucks from highways throughout the State of Oregon (i.e. safety, emissions, and maintenance benefits). Locations that have higher utilization generate greater congestion benefits, while locations further south generate large per-truck safety, emissions, and maintenance benefits. When combining all of these elements the Millersburg location generates the largest public benefits, followed by Eugene, Brooks, and Lebanon. Millersburg and Eugene are the only locations that generate a greater than one-to-one return on investment for the State of Oregon at the cumulative low estimate calculated at a 7% discount rate. On average, the Millersburg location generates 2% greater public benefits than Eugene, 25% more than Brooks, and 29% more than Lebanon.

Conclusion

The optimal location for a publicly funded intermodal center is a function of feasibility by volume, private transportation cost savings, and public benefits. These various factors should be weighted differently depending on the goals of the State of Oregon. If the primary goal is to maximize volume, then the northernmost location in Brooks would accomplish this goal, with an expected volume 11% higher than the next location south in Millersburg. However, while a location in Brooks may maximize volume, it would deliver the fewest private transportation cost savings due to the shorter distance that trucks are removed from Oregon highways. The estimates of private transportation cost savings at Brooks are 63% lower than Millersburg and 79% lower than Eugene. If the State’s goal is to maximize public benefits to residents through reduced pollution, highway wear and tear, and fewer accidents, then the optimal location would be Millersburg, followed closely by Eugene. The estimates can be spatially delineated as well. Benefits to the Portland area in terms of reduced congestion are maximized at the Brooks location, while benefits to the remainder of the state are maximized in Eugene, with values nearly twice as large.

Ultimately, the priorities of the State must be considered when determining where the intermodal center should be built. The Millersburg location captures a sizeable share of the market, removes a large number of trucks from roadways in the state, generates generous private transportation cost savings, and maximizes public benefits.