Comment by dgoldstein0

so every inflation number has to be understood by following (a) when is it measuring and (b) what is it measuring. For when: a lot of economic data is lagging indicators, e.g. last quarter - and inflation is usually % more year over year, whereas a lot of people seem to care about inflation on the 2-5 year time frame instead of just 1 year. For the what - we'd have to dig into whether it's national averages, state averages, or local; what percentage of the measurement is rent vs housing prices vs groceries (and what grocery items) vs clothing vs computers vs utilities etc etc. It's very likely that the idealized basket of goods that they are measuring the cost of doesn't actually match your expenses or even the average household expenses for your area. Or possibly even, for the whole country.

The meta problem is that price data - assuming we can even reliably observe it - is super high dimensional, and we're trying to reduce it all to a single number.