The national broadband plan drafted by Federal Communications Commission staff has a lot of goals in it. Goals for broadband infrastructure deployment include:
- Make broadband with 4 Mbps download speeds available to every American
- Over the long term, have broadband with 100 Mbps download and 50 Mbps upload speeds available to 100 million American homes, with 50 Mbps downloads available to 100 million homes by 2015
- Have the fastest and most extensive wireless broadband networks in the world
- Ensure that no state lags significantly behind in 3G wireless coverage
- Ensure that every community has access to 1 Gbps broadband service in institutions like schools, libraries, and hospitals
The plan also outlines a number of policy steps that the FCC and other federal agencies could take to help accomplish these goals.
So far, so good. But to truly hold federal agencies accountable for achieving these objectives, we need more than goals, measures, and a list of policy proposals. We also need a realistic baseline that tells us how the market is likely to progress toward these goals in the absence of new federal action, and some way to determine how much the specific policy initiatives affect the amount of the goal achieved.
Here’s what will happen in the absence of a well-defined baseline and analysis that shows how much improvement in the goals is actually caused by federal policies: The broadband plan announces goals. The government will take some actions. Measurement will show that broadband deployment improved, moving the nation closer to achieving the goals. The FCC and other decisionmakers will then claim that their chosen policies have succeeded, because broadband deployment improved.
But in the absence of proof that the policies cause a measurable change in outcomes, this is like the rooster claiming that his crowing makes the sun rise. Scientists call this the “post hoc, ergo propter hoc” fallacy: “B happened after A, therefore A must have caused B.” (Brush up on your Latin a little more, and you’ll even find out what Mercatus means. But I digress.)
Enough abstractions. Let me give a few examples.
The first goal listed above is to ensure that all Americans have access to broadband with 4 Mbps download speeds. In his second comment on my March 17 “Broadband Funding Gap” post, James Riso notes that the plan acknowledges that 5 out of the 7 million households that currently lack access to 4 Mbps broadband will soon be covered by 4th generation wireless. That means coverage for 83 percent of the households that lack 4 Mbps broadband is already “baked into the cake.”
Accurate accountability must avoid giving future policy changes credit for this increase in deployment, because it was going to happen anyway. (Of course, policymakers need to avoid taking steps that would discourage this deployment, such as levying the 15 percent universal service fee on 4th generation wireless.) The relevant question for evaluating future policy changes is, “How do they affect deployment to the remaining 2 million households?”
Similarly, the goal of 50 Mbps to 100 million households by 2015 seems to have been chosen because cable and fiber broadband providers indicate that they plan to cover more than that many homes by 2013 with broadband capable of delivering those speeds (pp. 21-22). Future policy initiatives should get zero credit for contributing toward this goal unless analysis demonstrates that the initiatives increased deployment of very high speed broadband over and above what the companies were already planning.
If you think this point is so basic that it’s not worth mentioning, you haven’t read enough government reports. Post hoc, ergo propter hoc is endemic, and not just on technology-related topics. For example, both sides regularly display this fallacy whenever the unemployment figures get released: “Unemployment increased after Obama’s election, therefore his administration caused the unemployment.” “The recession started when Bush was president, therefore his administration caused the unemployment.” These are at best hypotheses whose truth, untruth, and quantititive significance needs to be established by analysis that controls for other factors affecting the results.
Just take this as an advance warning on reporting results of the national broadband plan: Tone down the triumphalism.
Note: For those of you who just can’t get enough discussion of the national broadband plan, Jerry Brito and I will have a dialog on other aspects of the plan in a future podcast that will be available here on Surprisingilyfree.com.
Ed Roberts, designer of the first commercially successful personal computer, died yesterday in Georgia at the age of 68.
The Technology Liberation Front is the tech policy blog dedicated to keeping politicians' hands off the 'net and everything else related to technology.