The efficient market hypothesis (EMH) cannot explain economic bubbles because, strictly speaking, the EMH would argue that economic bubbles don't really exist. The hypothesis's reliance on assumptions about information and pricing are fundamentally at odds with the mispricing that drives economic bubbles.
Economic bubbles occur when asset prices rise far above their true economic value and then fall rapidly. The EMH states that asset prices reflect true economic value because information is shared among market participants and rapidly incorporated into the stock price. Under the EMH, there are no other factors underlying price changes, such as irrationality or behavioral biases. In essence, then, the market price is an accurate reflection of value, and a bubble is simply a notable change in the fundamental expectations about an asset's returns.
For example, one of the pioneers of EMH, economist Eugene Fama, argued that the financial crisis, in which credit markets froze and asset prices dropped precipitously, was a result of the onset of a recession rather than the burst of a credit bubble. The change in asset prices reflected updated information about economic prospects.
Fama has said that for a bubble to exist, it would have to be predictable, which would mean that some market participants would have to see the mispricing ahead of time. He argues that there is no consistent way to predict bubbles. Because bubbles can only be identified in hindsight, they cannot be said to reflect anything more than rapid changes in expectations based on new market information.
Whether bubbles are predictable is subject to debate. Behavioral finance, a field that attempts to identify and examine financial decision making, has uncovered several biases in investment decision making, both on an individual and market level. Inattentiveness to certain kinds of information, confirmation bias and herding behavior are a few examples that might be related to economic bubbles. These biases have been shown to exist, but determining the incidence and level of a particular bias at a particular time is not so straightforward.
Of course, it should be noted that the EMH doesn't demand that all market participants are right all the time. However, one of the theory's core tenets revolves around the idea of market efficiency. Given that market participants share the same information, the consensus price should accurately reflect an asset's fair value because those who are wrong transact with those who are right. Behavioral finance, on the other hand, argues that the consensus can be wrong.
The definition of a bubble is that the right price is fundamentally different from the market price, meaning that the consensus price is wrong. Whether it is predictable or not, some have argued that the financial crisis represented a serious blow to the EMH because of the depth and magnitude of the mispricings that preceded it. However, Fama would likely disagree.