3. Review of Literature

The literature review covers the various research studies on the link between transportation investment and economic development. This literature review is intended to draw implications regarding the type, structure, and level of detail of transportation investment data that are needed to support transportation investment decisions. The objective here is not to critique the literature or provide an exhaustive review of all research on the subject; rather its purpose is to find out the data needs of current research efforts in order to narrow our data compilation effort towards meeting these and other related data needs.

Largely because of the work by Aschauer (1989), academics have undertaken numerous studies that explore the relationship between transportation investment and economic performance. Economic performance can be measured by output, value-added (GDP), productivity and employment (e.g., Demetriades and Mamuneas, 2000). Some studies have explored the impact of infrastructure investment on technical efficiency that enters the production function (e.g., Delorme, Thompson, and Warren, 1999). Similar studies have also been undertaken based on foreign data (e.g., Kim, Koo, and Lee for South Korea, Rioja for seven Latin American countries, Everaert and Heylen for Belgium, and Sturm, Jacobs, and Groote for the Netherlands). Three major literature reviews, which were conducted in the late 1990s, cover major literature and most research projects on this subject in the U.S. up to early 2000. As pointed out above, by summarizing these literature reviews and some recent academic studies, we will draw observations that shed some light for our data work.

According to Bell and McGuire (1997), a major finding from studies up to early 1994 is that "a positive statistically significant but small effect of public capital on output has been confirmed by many." As pointed out in the report, however, structural changes in relation to infrastructure investment have received insufficient attention in the literature. These changes include "differentiated economic linkage between specific industry and specific type of infrastructure," the role of investment flows in examining "the derived demand for public infrastructure by private sector," and "the relationship between infrastructure types (e.g., complementary vs. substitutable)" (Bell and McGuire 1997, p. 9-10). The review indicates that research on the question of the relative productivity of different types of public infrastructure, which have been constrained by data limitations, requires data by infrastructure type or mode of transportation.

The subsequent literature review by Apogee Research and Greenhorne & O'Mara (1998) focused on evaluating the merits and limitations of techniques used in studies linking transportation investment and economic performance. This review divides the existing studies into macro- and microanalysis, with the former employing mainly production function and cost function methods and the latter cost-benefit analysis and case studies. In an attempt to find the impact of investment on productivity, the macroanalysis methods compare national trends in economic activity with levels of total public infrastructure investment. According to the authors' view, these methods fail to address the causal relationship; instead they simply show how various data series are related to each other. The microanalysis methods, by focusing on the economic effects of a particular project, provide insights into how the private sector reacts to changes in transportation. The review points out that both methods lack "a solid understanding of the mechanisms by which transportation investment influence structural changes in a developed economy" (Apogee Research and Greenhorne & O'Mara 1998, p. 26).

A recent synthesis of practices for assessing economic development impacts from transportation investments (Weisbrod, 2000) covers studies by both academia and practitioners up to early 2000. The synthesis, aimed at providing sub-national (local, regional, and state) planners with analytical tools, defines the economic impact by using regional rather than national indicators (e.g., using gross regional product rather than gross domestic product). In addition to a literature review, the synthesis is primarily based on a survey of 75 transportation planning agencies (in the U.S., Canada and the United Kingdom) on their current research and practice in evaluating the relationship between transportation and economic development. It pointed out that the evaluation of economic impact of public investment in transportation infrastructure has to date "focused primarily on highway spending," implying the need to broaden the study to cover more transportation modes.

Among all the studies of the economic impact of transportation investment, Nadiri and Mamunees (1996) appeared to be the most influential. The study used data on public highway capital, which was developed by Apogee Research, Inc., based on Federal Highway Administration capital outlay data. One of the findings that are most relevant to the design of our data work is that the economic impact of highway capital at the national level differs from that at the industry level and varies by industry. Furthermore, highway capital seems to have a substitution effect on private capital for most industries and is complementary for the rest. One comment on this ongoing study is that it overlooked the welfare benefits of highway capital to consumers that are "likely to be significant." (Eno, 1999, Appendix A).

Among other recent studies, Chandra and Thompson (2000) found that highways not only have "a differential impact across industries" but also "affect the spatial allocation of economic activity." Demetriades and Mamuneas (2000), using an inter-temporal optimization framework, found that in all 12 OECD countries, the magnitude of positive economic effect of public capital changes over time depending on the under-investment gap in infrastructure, which was wider during the 1970s and 1980s but narrowed down significantly by the early 1990s. Everaert and Heylen (2001), using single-equation co-integration analysis based on annual data, found a strong positive relationship with causality running from public capital (including roads, buildings, educational facilities, etc.) to multifactor productivity in Belgium for the period of 1953 to1996.

In summary, the above literature review tells us that past studies were mostly focused on the economic impact of government transportation investment. Most of the studies were concentrated on impacts of highway capital. Little attention has been paid to business transportation investment, its economic impact, and its interaction with government investment; and other modes. Furthermore, as pointed out by the aforementioned literature reviewers, more work needs to be done to explore the structural impact of characteristically different transportation investments on the economy. Structural impacts refer to those impacts that may be classified by industrial, geographical, and even demographic changes within an economy. As for characteristically different transportation investments, one may include a broad range of characteristics that can differentiate one type of investment from the other in terms of transportation mode, asset type, sector of investors, funding source, technology embedded, and the nature of investment (i.e., public goods, club goods, or private goods). Finally, few studies raise the issue of how economic development impacts transportation investment. That is, how economic development (e.g., economic growth and industrial restructuring) stimulates transportation investment in terms of both growing demand and increased funding sources for transportation services.

Ideally, new data development initiatives should be broader and more detailed to support a wide variety of future research undertakings. This report, which provides transportation investment by sector of investors, types of assets, and transportation modes, will fill some of the data gaps mentioned above, thus providing the basic data that will support policy decisions, and researches on transportation investment and economic development. Using the dataset developed by this report, researchers will be able to conduct those studies that they could not conduct in the past due to lack of data. For example, this report provides the infrastructure investment data by mode called for by Bell and McGuire. This is the first step towards the breakdown of infrastructure by type that they also suggest. Similarly, by providing comparable data across all modes, this paper emphasizes the need to move beyond the primary focus on highway spending that was pointed out by Wiesbrod. By bringing together data for infrastructure and equipment, and for government, industry and household sectors, it supports the kind of investigation of substitution and complentarity effects performed by Nadiri and Mamunees. Not only will this data begin to address some of the specific gaps mentioned above, but it may stimulate further related research.

See Bell and McGuire (1997), Apogee Research, Inc. and Greenhorne & O'Mara (1998), and Weisbrod (2000).

Refer to Eno (1996, 1999).