Did Facebook actually do what the prediction describes?

No. In 2020 Facebook:

• Banned new political ads only in the week before Election Day, while allowing existing political ads to keep running and be retargeted. (cnbc.com)
• Temporarily halted all political and social-issue ads in the U.S. only after polls closed on November 3, 2020, and then extended that post‑election pause for at least a month. (cnbc.com)

That is significantly narrower than “bans political ads for the rest of 2020.” The exact scenario in the prediction never occurred, so we can’t directly observe its effect.

What does the best available evidence say about banning Facebook political ads and partisan advantage?

A large-scale randomized experiment conducted by academic researchers and Meta after the 2020 election removed political ads from Facebook and Instagram feeds for six weeks before Election Day for tens of thousands of users. The study found no detectable effects of removing political ads on:

• Political knowledge, polarization, perceived election legitimacy, political participation, or turnout; and
• Support for either candidate, including vote choice for Trump vs. Biden, for Democrats vs. Republicans, or for different demographic and partisan subgroups. (nber.org)

Researchers conclude that, in 2020, Facebook/Instagram political ads had at most very small effects on political outcomes, and they did not find evidence of a systematic partisan tilt from removing them.

Putting it together

• The prediction is a conditional counterfactual about a policy that did not actually happen in the form stated (a blanket, rest‑of‑year ban).
• The closest high‑quality causal evidence we have—randomly removing political ads for a substantial pre‑election period—finds no significant advantage for either party from removing Facebook and Instagram political ads.

Because (a) the precise policy scenario never occurred, and (b) the best available evidence suggests little to no partisan effect from removing such ads rather than a clear pro‑Biden benefit, we cannot definitively determine whether Friedberg’s claim about a full‑year ban “obviously” favoring Biden is correct or not.

Hence the outcome is ambiguous rather than clearly right or wrong.