Before you begin to conduct input-output analysis it is important to establish “boundaries” that in some sense reflect functioning economic regions. Though at first the concept of “region” or “community” might seem obvious, defining one for economic modeling purposes is not always so easy. This section will therefore serve as an introduction to understanding how to capture your functioning economy.
A. Functional Economic Regions
For the purpose of economic modeling what do we mean by communities or regions? We first might think of political regions such as states and counties. However, when trying to determine functional economies political and administrative boundaries are not sufficient because they are simply drawn to serve some narrow political or administrative task and do not necessarily represent any economic relationship. We might also think of areas that share some similar characteristic (e.g., Silicon Valley, North Carolina Research Triangle, Appalachia, the Rust Belt, or the Corn Belt). This is a little closer, but still not quite what we are looking for.
Basically, when working to accurately portray a regional economy, one must avoid being influenced by political boundaries and/or other shared-feature regions. Instead, the area should reflect a spatial exchange of goods, services, and labor which are, incidentally, all types of exchange tracked in input-output models (Fox and Kumar, 1965). A region defined by such economic exchange is referred to as a functional economic region. The key characteristic of functional economic regions is that they have a cohesive, semi-closed market for goods, services, and labor.
B. Central Place Theory
So what is a “semi-closed market for goods, services and labor?” Perhaps the best way to understand that is through the concept of Central Place Theory (CPT), which provides a description of how communities (e.g. places) are economically linked. This spatial linkage is simple to understand and marks essentially what are functional economic areas. By tracing linkages, CPT also describes the trade hierarchy among the various places in a region. The idea is most easily conveyed with a simple graphic.
In its simplest form, a single “highest-order place” (i.e., Metropolitan areas, which offer the widest array of goods and services) “dominates” some collection of lower-order places. The figure presents a four-order hierarchy, with a single “metro area” economically dominating a pair of “cities,” which in turn dominate “small towns,” which dominate “hamlets.” The analysis is completed by specifying that each place (or “node”) dominates peripheral areas populated with isolated homesteads. Notice that a political region (such as a state or county boundary) may cut through central place regions regardless of trade links. The figure below illustrates this idea and reinforces the difference between functional and non-functional regions.
An alternative way of presenting the concept of CPT is presented in the table below.
Here it is seen that the lowest-order places provide goods that are available everywhere (in our case, post offices and restaurants), and that additional goods and services are added as one ascends the hierarchy. The highest-order place offers a unique subset of goods and services that is only regionally available.
It is important to note that the specific configuration of places and the specific collection of goods and services shown in these figures are illustrative only: real-world central place hierarchies will be far less tidy. Yet the principles remain, in that larger places will exhibit market reach (dominance) over smaller places, and through this dominance the two places are spatially interlinked. As noted above, this interlinking of places provides the basis for the economic cohesion that defines a functional economy, and provides impor-tant context for the community-level modeler.
C. Complicating Factors
Unfortunately for the would-be modeler, many situations in the real world are much more complex than the ideal cases. From the previous discussion, it should be no surprise that nearly all complications will arise from a lack of isolation. In this section, we will discuss two basic situations that could raise serious obstacles to an appropriate community definition: (1) thin economies and overlapping market areas, and (2) interlinked economies.
“Thin Economies” and Overlapping Market Areas
As a general rule, it is fair to say that input-output and community level modeling has different applications based on the difference in the character and population density of your area. The population density west of the Mississippi is, of course, much lower than that of the east, even in rural regions. Rural communities in the West are also much fewer in number and are generally spread further apart than in the East. Moreover, larger cities and metro-places are far less plentiful, meaning that opportunities for shopping and work are less plentiful.
The following simple map shows a large collection of small towns and larger cities around Cornwall, PA.
The circular area in the figure approximates a one-hour commute from Cornwall. Residents of this community will travel in many directions to work and shop. Now picture dozens of circles similar to the one drawn in the figure, each centered on the other nodes on the map, whether large or small. The result would appear as an incomprehensible tangle of overlapping labor markets. Certainly a great many people who work in the circled area (i.e., Cornwall’s commuting field) live elsewhere.
If we compare Cornwall to another example of rural western town (Grangeville, ID) the contrast is stark (below).
Cornwall (approximate population 3,500) and Grangeville (approximate population 3,200) are comparable in size, but Grangeville is isolated while Cornwall is not. Whereas the total population within a one-hour commute of Grangeville is approximately 10,000, that figure might be upwards of 1 million in Cornwall’s radius (these are very rough estimates, but good enough to make the point). The density around Cornwall creates very open markets and a web of “thin” economies with uncertain inter-community/inter-industry linkages. Even if a major event (the opening or closure of a large retail store, for example) occurred in one of these towns, the employment and other market effects would be too fleeting, easily leaking to surrounding areas.
In light of this, it does not make sense to speak of “Cornwall’s labor market” in any closed sense. Workers from Cornwall compete with workers well beyond the commute field shown in map, and similarly, job opportunities within Cornwall’s commute field are affected by business conditions beyond that field. Similar market overlap occurs in consumer and business goods and services markets.
It is therefore difficult to define or otherwise make sense of a “Cornwall economy.” One might draw the commute field shown in the figure and call this “Cornwall’s labor market,” but it would be scarcely useful due to the fact that there are so many overlapping labor markets in this same field. Narrow community boundaries, for instance Cornwall’s city limits or ZIP code, might be proposed, but this is an extremely open economy, with many goods (e.g., perhaps as common as hardware or supermarket) and services available in abundance in nearby communities. Crosshauling, which is the inflow and outflow of the same goods or services, including labor, is doubtless very prevalent.
As a result, it makes more sense to speak of a large-scale regional economy such as “the South Central Pennsylvania economy,” or perhaps a metro-area economy such as “the Harrisburg area economy.” The lesson is that because of complex trade hierarchies input-output modeling is more well suited to either larger multi-county regions around a metropolitan area, or closed economies such as can be found in the rural western states.
In the previous example, we saw that community definition for the sake of modeling purposes, can be impossible due to extreme regional interconnection. For our next case study we turn to a much less radical situation which nonetheless exhibits a complicating lack of isolation. Consider the two communities shown in the following map: Moscow, Idaho, and Pullman, Washington.
Moscow and Pullman are located just 8 miles apart on either side of the Idaho-Washington state line. Excluding each other, these two communities are considerably isolated—the nearest larger cities are Lewiston-Clarkston (35 minutes), Coeur d’Alene-Post Falls (2 hours), and Spokane (2 hours), and few residents commute to or from those places. Thus the case is easily made that Moscow and Pullman represent a single, albeit bi-nodal, functioning economy: there is considerable intercommunity commuting and head-to-head competition among like businesses, e.g., restaurants, retail stores, profes-sional services, consumer services, and so on.
That said, the existence of a state boundary means that these two communities could be modeled as a single functioning economy or as separate, though highly interlinked, communities. Commuting and other trade, which links these two communities, might be tracked using economic base principles, or they might be formally linked as a two-community interregional IO model (beyond the scope of this document). The choice would depend somewhat on the goals of the analysis, and how important political boundaries are in it (if taxes or local government policy are involved, modeling them separately would make more sense).
This chapter has presented some background theory and general advice about defining communities for modeling purposes. The advice can be summarized in three basic points:
- Functional economic areas must take precedence over political regions when attempting to accurately model a regional economy;
- The appropriate scale is that of a city-region; and
- Be aware of complicating factors, which generally arise from lack of isolation.
It is hoped that this chapter provides a solid foundation for the community modeler, but it should be emphasized that community definition, like region definition, is by no means a simple task that can be accomplished with only a handful of all-purpose rules. Rather, the specific geography and economics of a community or area can greatly alter a community’s boundaries for modeling purposes. Moreover, the purpose of the model being built will certainly affect the appropriate scale, and therefore the way boundaries are drawn. A model intended to examine the impact of new road construction in Idaho will likely require a different focus than a model aimed at examining the impact of a mill closure in Wyoming or a commuter rail line in Pennsylvania.