TMCnews Featured Article

March 01, 2011

Author of Rural Broadband Study Reacts to ''Rural Wireless Broadband Creates Jobs, Maybe''

By TMCnet Special Guest
Dr. Raul L. Katz, President, Telecom Advisory Services,

(This letter was written in response to the article "Rural Wireless Broadband Creates Jobs, Maybe" from TMC (News - Alert) Contributing Editor Gary Kim.)

Dear Mr. Kim:

I read your review of my last research on broadband economic impact published in TMC net "Rural Wireless Broadband Creates Jobs, Maybe," and wanted to share with you my observations regarding your conclusions:

1) The article suggests that we are concluding that only wireless, not wireline broadband, can generate jobs. In fact, the study is technologically agnostic. However, the National Broadband Plan has concluded that wireless broadband in the 700 MHz band is the most appropriate technology to cover rural areas.

Quote: "Some might argue the data suggests broadband development of any sort, using fixed or wireless networks, "creates or shifts" jobs."

Response: The study is technologically agnostic, without making any difference between wireline and wireless broadband. That is the only way of assessing the past economic impact of broadband back to 2004 given that the technology is a much recent phenomenon.

On the other hand, as the National Broadband Plan indicates, wireless technology at the 700 MHz band appears to be the most appropriate technology to deliver broadband services in rural geographies. I would refer you to the FCC (News - Alert) report "The broadband availability gap," which is mentioned in the study bibliography. Additionally, it would be useful to examine the spreadsheets included in the National Broadband Plan which quantify the unserved and underserved housing units per county. Column 6 of the data sheets indicates the most appropriate low cost technology on a per county basis. Wireless is the most commonly mentioned platform. Finally, if you want to consider international sources, you might want to consult Germany's National Broadband Strategy, which also mentions wireless in 700 MHz as the most appropriate technology to cover what Germany calls the "white zones".

2) The article suggests that the research denies the shifting of jobs between areas as a result of broadband deployment. However, the study explicitly acknowledges this issue on pages 22 and 23, pointing also that job shifts from overseas represent a net gain for the US.

Quote: "The issue is that much of the measured change could be explained by jobs that are simply shifted from where they are, to new locations, with zero net job creation."

Response: The comment suggests that the research denies the shifting of jobs. However, the study (and our past research) explicitly acknowledges this issue (see page 22). We believe indeed, that some transfers occur. Therefore, rather than arguing against the job shifting point, it might be more appropriate to mention the fact that the authors acknowledge the argument.

However, you fail to mention that some job relocation comes from overseas (as we mention in page 23 and you fail to quote), which is supported by the virtual call center experience. You might want to also read the article "Call Centers Phone (News - Alert) Home --- Small-Town Economics Lure More Companies to Outsource In Remote Corners of the U.S." by Ryan Chittum (9 June 2004, The Wall Street Journal).

In addition, if job transfers occur from the city to rural areas, economists might argue that this is a positive move bringing some equilibrium among labor pools. Finally, sociologists will also argue that job transfers from metro to rural areas help the retention of employment in the latter, which promotes a more viable development path to rural environments threatened by migration and depopulation.

3) The article disputes that jobs can be saved. However, the latest research relying on more detailed datasets substantiate this conclusion.

Quote: "The perhaps more contentious claim is that 78,500 jobs "will be saved" as a result of the combination of economic growth and increased capabilities resulting from the ability to gain access to broadband services. The assertion can be made, but some would argue the claim cannot be validated.

Some would argue that broadband deployment in fact follows development, but does not create it. Job growth and broadband might be correlated, but not causal, in other words."

Response: First, what you call "assertion" and "claim" is supported by the analysis of data described in page 41. In short, we relied on each of the States' Occupational Outlooks which outline the new jobs to be created versus those to be replacement jobs and applied that percentage to our employment impact. Since our total number represents a reliable prediction (more below), breaking it down based on the forecast of each of the state's labor economics unit appears to be a prudent analytical approach.

Second, let us examine your "correlation versus causation" argument. This discussion has been going on ever since researchers at the World Bank issued the first study examining the impact of telephony on economic development (see Jipp, 1963). Luckily enough, research has been adopting more sophisticated techniques to deal with this problem. This was helped obviously by greater availability of data. The primary approach in our study to control for reverse causation was to rely on a panel regression and utilizing lagged variables between 2004 and 2009 (you can read the detailed methodological description in page 29). Furthermore in order to control for fixed effects in the sample, we included seven additional variables ranging from educational level, income, size of labor force to age and ethnicity. It should be noted that our approach is the standard methodology employed in such studies.

As a last point, we recognize that broadband alone cannot create jobs. This is exactly what we say in page 57; "78,453 jobs will be saved as a result of the combined impact of economic growth and enhanced capabilities that will be provided to those workers as a result of wireless broadband."

4) The article suggests the study does not address job losses resulting from the productivity impact of broadband. However, job losses are explicitly factored into the methodology, using historical data that measures net job creation.

Quote: "Perhaps oddly, the study does not talk much about jobs that likely will be lost, despite the fact that one of the researchers who wrote the latest paper also wrote an earlier paper that said job gains would be balanced by some number of lost jobs, precisely because of the new broadband facilities".

Response: The reason why we did not mention the lost jobs is simply because our estimates, which are based on the analysis of historical data are net (in other words, gross gains-losses).

5) The article suggests findings of this research are inconsistent with prior research. However, the conclusions are fairly consistent. The 2009 study estimated 136,000 jobs after gains and losses for 18 states. This study estimates 117,000 jobs for 19 states. We believe the difference reflects the precision warranted by better information, which enabled more precise analytic techniques.

Quote: "Raul Katz, adjunct professor at the Columbia Business School, has argued in the past that it isn't entirely clear that new rural broadband networks necessarily will create net new jobs. Research on the productivity impact of broadband indicates the potential for capital-labor substitution and consequently, the likelihood of job destruction resulting from broadband deployment, as well as some incremental job creation. So the issue is whether net job creation exceeds net job destruction, and by how much ."

Response: In fact, the 2009 study did estimate a net job creation in the same range (see figure 22 in page 25 of the 2009 study). The 2009 study estimated in its baseline job creation scenario that 402,000 jobs will be gained (this is what you call "some"), while 266,000 will be lost due to productivity effect. The net gain would be 136,000. This number, calculated for the 18 states with lowest broadband penetration (which would benefit from a broadband stimulus program) is not dramatically off for the 116,000 estimate developed in this study.

One last point needs to be made on this issue. The earlier study was conducted in February 2009. The estimates of broadband economic impact were generated based on a forecasting technique called "chain ratio" analysis which is the only possible approach when one lacks historical data sets upon which one can estimate future impact based on the real past experience. This is why we say in the study that "narrowing th (e) range (of estimates) will require additional research based on the specific geographic areas of program implementation." This is what we were able to do this time around based on the survey data generated between 2004 and 2009 (and published in 2010) by the state of Kentucky.

6) The article incorrectly cites the evidence of sector focus of job creation.

Quote: "The latest study does, however, note some evidence that broadband deployment was the primary driver of employment in only two areas: the information sector, and the “administrative, support, waste management, and remediation” sector."  

Response: This is wrong. This finding was reported not by us but by Shideler et al. in their Kentucky study of 2007 (see our page 21).

7) The article does not quote material according to professional standards.

I would suggest that your article follows conventional rules of quotation. The following sentences are copied and pasted directly from our paper without quotation marks:

--"found no statistically significant economic benefit to telecommuting as a whole besides the consumer surplus of being able to work from home" (p. 22);

--"growth in employment in businesses enabled by broadband may just as easily be “cannibalism” of jobs from elsewhere in the state or in the country instead of the creation of truly “new” rural jobs" (p.22);

--"research on the productivity impact of broadband indicates the potential for capital-labor substitution and consequently, the likelihood of job destruction resulting from broadband deployment" (p.2 of my 2009 paper);

--"The research results do indicate a positive contribution that broadband makes to economic growth and job creation in rural areas. The effects appear to be most significant in the rural peripheries of metropolitan areas, where broadband operates as an enabler of spatial spill-over, resulting in an expansion of labor markets. However, it is important to emphasize that job creation in the rural peripheries might result from labor displacement from either the metropolitan areas or other regions" (p.24);

--"the technology facilitates the redeployment of industries to the rural peripheries to gain access to lower real estate costs, and better link to transportation networks" (p.24)

In sum, I would suggest that you research your concerns more comprehensively before publishing them so these critical policy issues can be addressed without confusion, and policy makers can be more effective in achieving social and economic impact.

TMCnet publishes expert commentary on various telecommunications, IT, call center, CRM and other technology-related topics. Are you an expert in one of these fields, and interested in having your perspective published on a site that gets several million unique visitors each month? Get in touch.

Edited by Janice McDuffee