Tuesday, July 31, 2012

There Will Be Graphs, Part Three of a Series

By Evan D. Robertson, Project Associate.

As a treat, my parents decided to take me to one of my favorite Roswell eateries on the square. For those of you not familiar with Roswell’s square, it’s a historic downtown area with a plethora of local restaurants, antique deals, and art galleries that offers the perceived safety of the suburbs with the urban thrill of the unexpected. As my father drives down Canton Street, we realize that we’ve come at an inopportune time. Every Thursday the square hosts Alive After Five packing the streets with people. This normally wouldn’t be a problem since I am a patient driver especially when pedestrians are involved. My father, who’s behind the wheel, not so much.

At the sight of the innumerable pedestrians thwarting us from even inching forward, my father mutters something under his breath, all I could catch was the word “tourists.” My mother turns to him and says something completely mind blowing to me, “Don’t curse them, they are here to support our local economy. The money they spend here gets recycled around providing jobs for local workers.” She then goes into a further, albeit brief, explanation of economic multipliers and economic base theory. Kudos to whoever has informed my mother about the basic tenants of local economic development as I am sure it wasn’t me. But, with the simple utterance, my mom summarized and condensed my graduate school education (and the debt that came with it). This moment raised a thought: could economic development professionals do a better job of explaining and demonstrating the depth of our economy’s interconnections? Yes, most residents, business leaders, and elected offices intuitively understand the basics of local economic development. But, can we provide both depth and simplicity?

Input-Output analysis defined by the Bureau of Labor Statistics as “an economic tool that measures relationships between various industries in the economy.” The importance of input-output tables cannot be stressed enough, these are the tables used to calculate national gross domestic product, any multiplier you see was derived from an input-output table, and the tables are used to calculate the economic impact of supply-chain disruptions. How will an earthquake in Japan affect the United States automobile industry? First step: grab an input-output table.

The following graphs utilize the 2002 benchmark industry-by-industry total requirements table (see footnote 1). The total requirements table displays the relationship between supply (input) and final use (output). In other words, the industry-by-industry total requirement table displays the total dollar value of input an industry would need to receive in order to produce a dollar of output. For example, electronic computer manufacturing (NAICS 334111) required input of exactly $.0009489 from Breweries (NAICS 312120) and $.001289 from Wineries (NAICS 312130) per dollar of electronic computer manufacturing output. In other words, for every $10,000 of computer and electronic product manufacturing output, the sector required the equivalent of a case of beer (~$9.49) and a bottle or two of wine ($12.89). I’ll let you draw your own conclusions.

In order to simplify the analysis, the following graphs hone in one specific business sector: the Computer and Electronic Product Manufacturing (NAICS 334). The purpose of this blog is simply to demonstrate the ease at which Social Network Analysis can show the interrelations between the business sectors, since input-output tables are nothing more than the monetary relationship between one sector and another, and, more generally, experiment with a new data set. So what did the analysis reveal?

Network Analysis of Computer and Electronic Product Manufacturing




Source: Bureau of Economic Analysis

At first, noise. The red diamonds toward the center of the graph represent the twenty-five six-digit NAICS sectors that compose the Computer and Electronic Product (NAICS 334) sector according to the Bureau of Economic Analysis. Each of the twenty-five sectors is connected to 427 other six-digit sectors composing the remaining industries within the input-output tables. Herein lies the challenge of performing network analysis using input-output tables, every industry is related to every other industry; thus, one of the most importance pieces of information (what isn’t connected, what doesn’t have a relationship) is missing. Thus, this limits the types of conclusions one can draw from the analysis. Nonetheless, network analysis can discern and demonstrate the strength of the relationships between economic sectors.

Strongest Connections in Computer and Electronic Product Manufacturing


Source: Bureau of Economic Analysis

The above graph displays the “strongest” connections within the Computer and Electronic Product Manufacturing (NAICS 334) sector. Strongest is defined as requiring at least $.009 of input from an industry to produce $1.00 of output to less than $.99 input per $1.00 of output. The filter yielded 279 connections within NAICS 334. These 279 sectors were further refined to those sectors that required at least $.05 of input per unit of output, revealing only thirteen connections. The printed circuit assembly (NAICS 33418) sector accounted for the majority of these connections. In fact, of the 13 industries that required at least $.05 of input per $1.00 of output, printed circuit assembly accounted for nine of the 13. This alludes to the sectors’ reliance on outside suppliers, and may indicate a lack of resiliency if its supply chain is disrupted. For example, the printed circuit assembly sector requires nearly $.12 cents of input from the electronic computer manufacturing sector per $1.00 of output. Any supply chain disruption, say an earthquake off Japan’s coast, has the potential to adversely affect the printed circuit assembly sector’s ability to produce circuits (output). This disruption, in turn, ripples throughout other sectors of the economy that cannot obtain circuits from other circuit producers or use substitute goods. Yet, just as the relationship between the sector and its suppliers are important, so too are the relationships within the sector.

Interconnections in the Computer and Electronic Product Manufacturing Sector

Source: Bureau of Economic Analysis

The above graph displays inter-industry relationships that exist in NAICS 334. Due to the technical and highly specialized products produced by the sector that require specific electronic components as input for the final computer and electronic goods, these relationships were by an large fairly important to producing final goods (output). However, in order to judge the most important relationships, inter-industry relationships were filtered eliminating those relationships which required less than $.09 of input per $1.00 of output.

Important Relationships within the Computer and Electronic Product Manufacturing Sector


Source: Bureau of Economic Analysis

Important relationships within the NAICS 334 exist between the semiconductor and related device manufacturing industry (NAICS 334413) and the printed circuit assembly manufacturing industry (NAICS 33418). In order produce $1.00 of semiconductor output; the sector requires $0.25 of input from the printed circuit assembly manufacturing industry (NAICS 33418). Other strong inter-relationships exist between computer storage and device manufacturing (NAICS 334112) and electronic computer manufacturing (NAICS 334111); broadcast and wireless communication equipment (NAICS 334220) and telephone apparatus manufacturing (NAICS 334210); and semiconductor and related device manufacturing and electronic computer manufacturing (NAICS 334111).

So, what can we gather from using network analysis on input-output data? Well, economic development professionals talk quite a bit about the interconnections between business sectors, but we don’t do enough to actually show the strength and importance of certain relationships over others. Moreover, while those outside of the economic development profession intuitively understand multipliers and the interconnectivity of local businesses, more can be done to display those connections and to stress their importance in the local economy. For economic development professionals, input-output tables provide an greater understanding of the relative importance of certain suppliers and other supporting sectors on their targeted industries. It facilitates the identification of important supply chain linkages, serving as a basis for recruitment or expansion efforts. Network analysis provides an expedient means to understand and display relationships between economic sectors.

Footnote (1): Input-output tables are updated and released every 5 years in conjunction with the Census Bureau’s Economic Census. I’m using the 2002 benchmark file because it reports data at the six-digit NAICS level, providing the depth for the analysis and a large data set for experimentation.