Exactly. Supplying some context to support this: The level of crime is pretty hard to measure. You can measure reported crime, but crimes are reported at diffe

Exactly. Supplying some context to support this: The level of crime is pretty hard to measure. You can measure reported crime, but crimes are reported at different rates in response to complicated incentives. How much the earth is warming depends on what you measure. Do you measure atmospheric temperature? Ocean temperature? And of course how much the world will warm is dependent on complicated models with tons of inputs. How many people live in poverty depends on what your threshold for poverty is. There's a "Federal Poverty Level", but cost of living varies by significant amounts across the country. The rate of inflation is highly dependent on the basket of goods measured and how improvements in goods are measured and so on. There are easily a dozen different measures of "inflation" and they're all reasonable and carefully considered, but none of them is the ground truth. It is of course relatively easy to measure Social Security inflows and outflows, but usually when we talk about the "cost" of programs like this, we mean something like the net cost, which incorporates lots of societal effects. Also the interpretation of the accounting concept of the Social Security Trust Fund, despite being a fairly simple concept, has significant camps with diametrically opposed views.