October 2009 to September 2010
Economic impact models that are spatially disaggregated are an important methodological tool in regional science. Aggregation to the national level obscures important details and is potentially misleading whenever positive impacts in one place cancel negative impacts in another. It is also likely that aggregated results are of limited interest to policymakers because of most politicians’ keen and logical interest in impacts on local constituencies. Inter-regional and multi-regional input-output models were first suggested over half a century ago to address these problems. In recent years, there had been considerable advancements in the regionalization of national input-output data. Yet, while the available multi-regional models depict trade between regions, the infrastructure over which the trade occurs remains mostly neglected. This is perhaps ironic in the context of inter-industry analysis. In this paper, we present applications of TransNIEMO, a model we developed to address this omission. Our group previously developed and applied the National Interstate Economic Model (NIEMO), a multi-regional input-output model that includes the 50 states and the District of Columbia as well as 47 industrial sectors. TransNIEMO adds the nation’s highway network which accommodates most of the intra- and inter-industry trade that NIEMO estimates. The new model seeks highway network and economic equilibria that are consistent. The U.S. economy is vulnerable to terrorist attacks, and modeling how disruptions at major choke points on the nation’s highways might impact the U.S. the economy on a state-by-state and industry-by-industry basis is of particular interest. We believe that TransNIEMO is the only operational model that can be used for this type of analysis. While this document reports the results of simulated attacks on three major choke-points, disruptions from natural or man-made events on any other vulnerable highway link can easily be modeled by applying TransNIEMO. We find that, as a proportion of the nation’s total output, the losses experienced in all three scenarios are relatively small. We ascribe this result to the high levels of redundancy in the highway network. Our results do show, however, there are significant differences in state-by-state as well as industry-by-industry impacts. Highway infrastructure, especially the bridges or tunnels, are vulnerable to disruption or structural failure from neglect, natural or man-made disasters and the possibility of terrorist attacks, which are now a major national concern. And even absent the terrorist threat, the aging of key infrastructure links has been widely noted. According to the National Bridge Inventory, 12.1 percent of bridges were rated structurally deficient, and 13.3 percent rated functionally obsolete in December of 2007 (USDOT-FHWA 2008a). Although a bridge rated either structurally deficient or functionally obsolete is not unsafe for all vehicles under normal conditions, it is more vulnerable in emergency situations and especially so in cases of old designs that lack modern safety features. The failures of the I-10 Twin Span Bridge during Hurricane Katrina in 2005 and the I-35W highway bridges during rush hour in Minneapolis in 2007 have brought home to planners and policy makers concerns about the safety of highway bridges and the possible effects of bridge collapse on highway performance and the economy at both local and national levels. Xie and Levinson (2009) evaluated the effects of the I-35W bridge collapse on regional traffic patterns and estimated the economic losses imposed on travelers in terms of travel time delay using a regional travel demand and investment model. However, their research focused on regional effects of the bridge collapse and did not examine any ripple effects on traffic with respect to the national highway network. In addition, freight trips were not modeled directly. According to the Federal Highway Administration (FHWA), there are 50 tunnels over 500 meters in length along the 4-million miles of US roadways. Most tunnels are proposed or constructed to improve highway system performance. For example, planners have proposed a longer than 16-mile tunnel, the world’s longest highway tunnel, under Long Island Sound to reduce the traffic congestion in the New York metropolitan area. In comparison to highway bridges, highway tunnels have received less attention even though they are also vulnerable to natural or manmade disasters, such as fire, flooding, tornadoes or pollution. Turning to the demand side, because of rapid population growth and brisk economic development, freight transportation has grown dramatically in recent years. Total U.S. freight ton-miles increased from 2,420 to 3,137 billion between 1993 and 2002, an annual growth rate of 2.9 percent. As the dominant freight mode, trucks moved about 40 percent of total revenue ton-miles of freight, accounting for almost 74.3 percent of the total dollars and 67.2 percent of the total tons of all freight shipments in 2002 (USDOT 2004). The total Vehicle Miles Traveled (VMT) on public roads increased by 39.4 percent (from 2.1 to 3.0 trillion) between 1990 and 2005 while truck VMT grew at a much faster rate by 52.2 percent during the same period. Freight moved by the U.S. transportation system reached 21.0 billion tons in 2006, and was worth about $14.9 billion. Trucks on highways moved about 60.4 percent in tonnage and 65.4 percent of dollar value (USDOT-FHWA 2008a). The FHWA (2002) projected that freight volume will nearly double between 1998 and 2020, increasing from 9 billion tons to about 17 billion tons (USDOT-FHWA 2002). USDOT also estimated that truck volumes on the National Highway System (NHS) will increase 230 percent from 2002 and reach 10,000 trucks per day by 2035 (USDOT-FHWA 2008b). This report cites the relevant preceding literature (Section II), describes how a computable highway network and its constituent parts were assembled (Section III), explains the minimum path algorithm that was used (Section IV), describes network flow results of various simulated disruptions (Section V), and wraps up with conclusions and reflections (Section VI).