>Usually if official statistics and what you see are misaligned, its the official statistics that are wrong.
Nope, a few people's unmeasured selection bias is no where near as accurate as a professional methodology done like the official numbers.
>there is a huge incentive to keep the official numbers down
Also false. If the numbers were gamed, there would be significant arbitrage employed in markets to extract value from the lie. There are ample papers analyzing this, and no one has found a way to do it, since the numbers are the best possible.
Also the numbers are done in many places, from Goldman Sachs, to the Billion Prices project at MIT, to the Bureau of labor and Statistics, and so on.
If you really think the numbers are gamed, do this experiment (I have, it's easy to do, and then you'll never claim the numbers are gamed again since you methodically checked them).
Take BLS listed inflation for a decent period of time, say 20-30 years. Make a basket of goodssplit by how most people buy: energy, housing, food, eneterainment, etc.
Pick a list of goods, say same sized apartments, houses, foods, etc.
Look at prices now. Write them down. Hide that list so you don't cheat.
Look at ads from the beginning of the period. Get prices.
Compute.
Now add say 1% to BLS annual inflation numbers, and compute.
Viola! BLS and rest are very accurate. People's unmeasured selection bias beliefs are not.