Earlier this year a gap opened up between what Americans were reporting in economic confidence surveys and the official stats. Historically they've tracked very tightly, almost in lockstep, so that gives some confidence in the official data. Then suddenly Biden was touting an amazing economy and the job numbers departed from what people were reporting as their on-the-ground impression. The cause is probably double counting due to remote work, zombie job ads and so on: lots of jobs that are being advertised but are duplicates or don't really exist, caused by a combination of the pandemic and new job board websites. But if you're on Team Biden are you going to prioritize figuring out why the gap between your data and public sentiment has appeared? Maybe making yourself look bad in the process? Or will you decide that there are more important priorities elsewhere? It's probably the latter and won't even feel political, just like it's the right thing to do. Of course because these are unarticulated decisions not policies, there's always a risk of some other process or person coming in and rearranging priorities for you, which maybe is what happened here.
For another example look at the UK. Their population statistics are a mess and have been for many years. At one point the ONS even reclassified them as experimental. If you're a strongly pro-mass migration politician or cabinet, are you going to be down at the ONS giving them hell about this every day? Probably not. It's not that the numbers are cooked, but more that problems arise that nobody cares to fix because the fix would create new problems for them.