"I'm fascinated with the problem of why really smart people have such a hard time predicting the future. It's mostly because the future is more random than we think. But it's also because future performance (like earnings and economic growth) doesn't tell you half of what you need to know to predict future outcomes. Outcomes are determined by performance within the context of expectations, with importance heavily weighted toward the latter. And if predicting future performance is hard, calibrating them against expectations is close to sorcery. Even if you accurately forecast future performance, predicting the outcome from that performance requires answering two questions: * Are current expectations reasonable? * What will future expectations be? The first is possible to answer, but pretty hard. We can look at sentiment and valuations to gauge, based on past trends, whether the current mood seems reasonable. But disagreeing with popular sentiment is easier said than done. Few things feel better than fitting in, and having a viewpoint that goes against everyone around you is a mental battle not one person in a ten can maintain for long. Rather than identifying extremes in current sentiment, it's easier to justify the market's mood no matter what it is. [..] The second – predicting future expectations – is even harder. To know where stock prices will be in five years, I have to know what mood people will be in five years from now. Which is as ambitious as it sounds. Ask yourself what kind of mood you yourself will be in in April 2021, and you'll shake your head in laughter. Ask what kind of mood seven billion strangers will be in in April 2021 – that mood, of course, will determine stock prices in five years -- and it's hard to keep a straight face. This is where the disconnect between performance and outcomes occurs: Accurately predicting five years of economic growth might not do much for the stock market if, five years from now, people are worried about the future five years from then."