How far ahead should we look when planning economic policy?

A couple decades? 50 years? An entire century?

If job automation will be slow and relentless push, rather than a sudden tsunami-like wave, should we begin training people today for the jobs of tomorrow?

On one hand, it’d be more efficient to educate people once for a job they can work for a lifetime. Employing people now, in jobs doomed for extinction, can only mean periods of widespread unemployment later on. We can avoid repeatedly stretching our social fabric, creating tension and growing distance between the haves and have-nots — what we’re seeing today between the coastal elite and working-class families, struggling to stay afloat in Rust Belt cities and coal-mining counties. We recently elected a populist, malcontent president. We let the fabric rip.┬áCan our democracy handle another?

On the other, implementing a minimum wage or a universal basic income ahead of a wave not only far on the horizon, but speculative in size and strength and speed, might be overkill. There’s overwhelming evidence that raising the minimum wage slows down economic growth. It burdens the everyday consumer as prices skyrocket; purse strings tighten and hiring crawls to a snail’s pace. Should we be lifting our foot off the accelerator and slamming on the brakes when our ramp exit is still 10, 20 miles away? What if China is tailgating us and they ram into our bumper, causing a traffic jam for everyone riding behind us, putting a screeching halt on the global economy? Or worse, what if they pass us — along with everyone else?