There's 2 questions - was the initial decision wrong? If so, was it clearly and obviously wrong? You can say 'yes' to the former without saying 'yes' to the latter.
There's a few things to consider. Position of the arm, movement of the arm, time to react being a few things.
The attacker did touch the ball first - but the change to the trajectory of the ball was minimal. So that hasn't changed what the defender had to react to. If, for instance, the ball was going to miss the defender but the path of the ball was significantly changed at the last moment, then you consider that the defender has had no time to react.
The defender also had a view of the ball the entire time it was crossed in - again, if the defender was unsighted until it passed his opponent, then that's in the defender's favour.
So, the defender has had plenty of time to react to the cross. Now, one could argue that the defender expected the attacker to head the ball in a different direction thus didn't position himself for the ball to go backwards like that - but I think that's a very weak argument at the best of times; in this case, given the goal was behind the players, you should expect the ball to go somewhat in that direction. So given that the eventual path of the ball wasn't a completely unpredictable one, that leaves it being the defender's responsibility to do the most he can to ensure the ball is not going to strike his arms.
And that leaves us with the positioning of the arms. I had to look at it a number of times to work through this decision making process - but I was left with the fact that he jumped with his arms right out to the side. Why on earth would any defender do that?
I find myself making similar arguments to that in favour of the handling penalty in the Australia - France match (this time a French player was penalised). You simply can't be jumping with your arms right out to the side - doing so puts the defender in a position where there's a good chance they're blocking the ball. The onus is on the player to take reasonable steps to minimise the chance of handling the ball. And that means keeping the arms closer to the body when jumping as a defender. Not jumping with arms flailing out to the side.
When defenders are standing in front of an attacker with the ball, they know now to stick their arms right out to the side - it's not different when jumping.
So while the contact occurred while the arms were coming back down, I don't think they should have been there in the first place. The defender has jumped in a manner that broadens his field position using the arms, he's had visibility of the ball and nothing happened to change his reaction time. That makes the handling wholly the defender's responsibility - and that makes it a foul for me.
So the 2nd question is - was it a clear and obvious error? Fortunately this question doesn't affect us as grassroots referees - it's purely a discussion piece. It's difficult to answer this without being party to FIFA's training on 'clear and obvious error'.
I don't think taking a while to reach the decision means it isn't clear and obvious. I don't believe that 'immediately apparent' is a prerequisite for 'clear and obvious'.
I can only wonder - under what circumstances would a VAR review ever leave the referee thinking 'it's a foul, but I can see an argument to say it's not, so I won't change it'?
When the VAR is to intervene is still one of the challenges facing football. The implementation at the World Cup seems to have a fairly low bar on 'clear and obvious' - once the referee thinks it's a foul, the decision gets made. So I think it's consistent with the VAR application this world cup, though I still wonder if it's the best way to implement VAR.
I don't know what 'clear and obvious' means, and I'm not sure anybody else - even FIFA - does.
Another point to consider, the time from the sight touch from the player in front of him to the his hand is under 0.05 seconds. I am not an biology expert but my guess is that is not enough time to react to anything.
He didn't have to react to it - the impact on the ball was very slight; had there been no touch it still would have collided with the defender's hand.