Don't Like the Numbers? Change 'Em
Politicians and scientists who don't like what their data show lately have simply taken to changing the numbers. They believe that their end—socialism, global climate regulation, health-care legislation, repudiating debt commitments, la gloire française—justifies throwing out even minimum standards of accuracy. It appears that no numbers are immune: not GDP, not inflation, not budget, not job or cost estimates, and certainly not temperature. A CEO or CFO issuing such massaged numbers would land in jail.
...Even more blatant is the numbers game being used to justify health-insurance reform legislation, which claims to greatly expand coverage, decrease health-insurance costs, and reduce the deficit. That magic flows easily from counting 10 years of dubious Medicare "savings" and tax hikes, but only six years of spending; assuming large cuts in doctor reimbursements that later will be cancelled; and making the states (other than Sen. Ben Nelson's Nebraska) pay a big share of the cost by expanding Medicaid eligibility. The Medicare "savings" and payroll tax hikes are counted twice—first to help pay for expanded coverage, and then to claim to extend the life of Medicare.
Seriously, read the whole thing. Great brief on why those of us who trained in empirical economics don't believe a damn thing we hear coming from the mouths of pols and activists until we've reviewed the base assumptions and modeling construction. Because so much of it is complete bullshit.