What is the True Economic Impact of Data Centers on their Local Communities?
Over the past decade, major corporations such as Google, Facebook and Amazon have invested in billion-dollar data centers and server farms in rural counties across America. While the vast majority of the organizations operations and management take place near cities (namely in Silicon Valley), these data centers represent a significant portion of the corporation’s tangible assets.
These corporations have multiple reasons for building in rural counties. These areas, such as western North Carolina (near the mountains), are plentiful in water and reasonably consistent wind used to cool the server rooms, decreasing costs. As one might expect, the land value in these sparsely populated areas is significantly low to start out with. More important are the politics involved in persuading these corporations to locate in specific counties. Google (and similar corporations) essentially choose the highest bidding county in terms of incentives such as property and income tax breaks. Controversy emerged over the past decade as popular news sources like the Washington Post expressed the rural communities dissenting opinions on the economic profit of these data centers.
While the profit of large scale data centers is still questioned by some today, analysis from Oxford Economics in April 2018, suggests that the overall economic effect on counties hosting these data centers is indeed beneficial, though not immediately so. This is unsurprising as we can find articles from around a decade ago to a few years ago (while this trend was young) expressing the falsehood of corporation’s and politician’s promises of increased economic activity as a result of these deals. As we will discuss shortly, the positive economic externalities involved in these changes take time to build up, as is true in almost any investment situation.
In 2011, the Washington Post released an article titled “Cloud centers bring high-tech flash but not many jobs to beaten-down towns.” A large portion of the article consists of negative accounts of blue collar residents in rural communities with newly build data centers, without giving any authentic or thorough economic analysis. The article, like the community’s residence, only focus on the direct jobs added by these data centers. While examining the direct jobs created is a common way of assessing the effectiveness of such projects, it is a shallow analysis at best. A robust analysis, as found in the Oxford Economics paper, incorporates the direct jobs attributed to the data centers operations, the indirect jobs attributed to the corporation’s supply chain and suppliers, and induced jobs (or the “multiplier effect”) resulting from the consumption of employees and vendors in the local community.
Next City’s article from 2013 titled “Cities keep giving out money for server farms, see very few jobs in return,” inherits from the Washington Post article’s tone and delves into the deal made between Lenoir NC and Google in 2007. Google decided to build one of their six current data centers in Lenoir after the city offered an estimated $254 million dollars in subsidies. The article uses the word subsidies but what it is really referring to is a combination of a $165 million dollar 30-year real estate tax break, and other tax credits for a similarly allotted number of years. When most individuals hear that Lenoir offered Google $254 million in subsidies, their first thought is likely that Lenoir funded a quarter of the cost to construct Googles data center using tax-payers dollars. In reality, the “cost” is actually a loss of potential tax revenue if Google had built their data center in Lenoir without any incentives, which Google would not have done. Lenoir isn’t losing tax revenue in this exchange, its only gaining a billion-dollar investment with multiplier effects rippling throughout the local economy in construction, consumption and education. The subtle choice to use the word “subsidized” instead of “tax breaks” shows a clear bias involved in this news reporting.
Other subtle use of language can also be seen in the Washington Post’s article. When describing the opinion of a hardware store owner from Maiden NC, the author delves into descriptive, sensory language to express traditional and pure associations with the hardware store compared to an imagined cold and calculating data center.
“Samantha Saunders, the longtime owner of a Main Street hardware store, where the old hardwood floors creak and a fresh-paint scent wafts through the cramped aisles, said the only contact she has had with an apple employee is when one came in to make keys for the facility.”
Similar to Next City’s misleading word choice, this use of language is borderline inappropriate. The question of why do (or did) these news organizations want to subtly depict corporations, like Google, as taking advantage of struggling rural counties requires its own attention.
The most important question that the Next City article brings forth is if the local government’s time and energy would be better focused on infrastructure projects. The article specifically brings up the proposed Atlanta-Charlotte high-speed rail line that might benefit economic activity throughout the region. The Next City article writes “anything but a server farm that creates 50 jobs,” while discussing this. While the high-speed rail line would undoubtedly support long term economic growth in the south-east, this comparison is shallow without any reference point to compare each projects costs or benefits.
As already noted above, Oxford Economics 2018 paper titled “Google Data Centers: Economic Impact and Community Benefit,” provides the most current economic analysis on the impact of data centers on the communities they enter. The paper details that the employment impact of Google’s data centers at the state level is widespread and higher than is often supposed. Through their regression analysis (the methodology of which is provided in the appendix of their paper), Oxford Economics found significant local spillover effects (economic externalities) within three years of the data center opening. These benefits include employment gains that extend further than those directly connected with the data center campus (indirect and induced jobs as explained above) as well as an increase in county residence with a college degree.
While you might shrug off the increase in county residence with a college degree as a result of Google relocating educated employees to the county, this overall increase in educated individuals spills over into the local education systems. The primary (non-quantifiable) effect a Google data center has on its local economy is partnering with local education (namely local community colleges) to create new education initiatives that prepare the communities for positions in the increasingly serviced based, technological US economy.
While the Oxford Economics paper may come across as overly optimistic, especially compared to the other articles discussed, it should be noted that these estimates are conservative. Each county examined is matched to control counties and the significant impact of manufacturing of computer, telecom, and other equipment placed into service at the data centers (and constantly updated) is excluded from numerical estimates.
Going back to the Washington Post and Next City’s critical tone, it is easy to think of traditional, rural communities with high unemployment being taken advantage of by corporations like Google. It makes sense that most individuals in these communities would be resentful while watching their manufacturing jobs move overseas to be replaced by data centers as an example of economic opportunities continuing to move out of their grasp. While these now out of place workers are rightly upset, the investment of the world’s dominant technology firms modernizes these communities and helps provide younger generations with opportunities to find a place in the current service based, technological economy. Local politicians choosing to persuade corporations like Google to invest in their counties is a good long-term economic decision.