There are four main reasons economic data have become less reliable over a long period of time.
Much of economic data is survey-based, and response rates have collapsed (a trend that started before 2020 and accelerated in the pandemic). If response rates are low, they may not be representative (with replies biased to the negative). Political bias can also motivate people to answer surveys inaccurately.
The rapidly changing world means that new forms of economic activity may not be captured. And underfunding of statistical agencies weakens the quality of data—funding cuts forced the US to reduce measurement scope of consumer price inflation—a number that directly determines government spending and interest rate payments on some government debt. Dodgy data could cost bondholders money.
While lower quality, the data that is produced is still politically independent. The risk of adding political bias to economic data is that the results become a politician’s fantasy world, not economic reality. That increases the risk of policy error. Privately generated data would lack a reliable benchmark (making it impossible to assess the usefulness of those numbers). And individuals are more likely to suffer confirmation bias—relying on unreliable anecdotes to confirm their own perception of the economy, in the absence of factual evidence.