Read on Twitter
If you have an opinion about the minimum-wage proposal, I hope you have an opinion about the effects it would have. This is a forecasting problem, but have any minimum-wage experts made explicit forecasts about the effects of Biden's proposal? [thread] [1/6]
For the most part, academic researchers simply don't make explicit forecasts. According to tradition, their job is to publish theories, and do experiments or sophisticated retrospective analysis. These endeavours complement forecasting, they shouldn't substitute for it. [2/6]
Social scientists make vague implicit forecasts all the time when discussing policy, and we'd gain a lot of insight if more of them made these forecasts explicit. Traditional social science research is valuable in its own right, and as an input into explicit forecasting. [3/6]
A conditional forecasts states: "If X1, I expect Y1, if X2, I expect Y2" ... if X1 occurs, we don't get to evaluate the forecast under X2, but accuracy under X1 tells us something about accuracy would have been under X2, and therefore the difference between X1 and X2. [4/6]
Policy changes typically have many effects. People may disagree about the policy because they disagree about the effects or because they value different effects differently. But it's hard to sort these differences out if people aren't explicit about their beliefs and values.[5/6]
In the minimum-wage debates people typically only consider the effect of the policy on wages and unemployment. Forecasting those would be a good start, but why not also study/forecast consider job satisfaction, geographic mobility, prices, education/training, etc. [6/6]
As much as I'd be happy to see a dozen related forecasting questions on @metaculus or @GJ_Open, what I'd really like are for people to generate models that make thousands of forecasts. @scottleibrand
e.g., for every city or town in the U.S., how many people will live there, how many hours/wk will they work at what wage, for every year for the next 10 years? (w/ and w/o $15 minimum wage)
