Reluctance of Markets to efficiently price-in information related to LFHI events

EMH (Samuelson/Fama 1970) says that at any given time, prices of various securities, tradable financial products and currencies fully reflect all available information "known to market"; if the market is sufficiently large and liquid. The nature of information in this hypothesis is not specified nor limited, i.e. it takes news from the realm of politics, social events, economy and finance and combines it with perceptions of investors on the market to form prices of all commodities and financial instruments on the market. Also, by increasing the speed of dissemination of information, information technology increases the overall efficiency of market(s) and synchronises movements of prices in different markets (Authers).

Despite the decades of controversial lack of consensus in academia and industry about the overall validity of EMH, some aspects of this hypothesis were proved to be empirically true (weak efficiency) while others (strong efficiency) were not (Lo). By and large, most studies conclude that the major financial markets are efficient to some extent, i.e. prices of financial instruments traded there reflect, to some extent at least, all available information.

But there is one another dimension of the market efficiency, and that is time. We know that by trading upon the information of HF (i.e. high frequency) signals short term traders frequently execute transactions, thus supplying HF information to the market. Does that mean that traders with long term horizons (pension funds for example) tend to trade upon information of LF (i.e. low frequency) signals, thus providing the market with the LF information? If so, the adaptive dynamics of prices of tradable commodities and financial instruments should reflect behaviour of the agents that dominate the market, regardless of their characteristic time horizon.

Regarding the time dimension, some people have noticed that markets (starting with those in oil and gold) have very bad track record, historically speaking, in pricing in low frequency (meaning low probability) risks, even if they are of high intensity. The effect is sometimes observed as plain indifference of markets (particularly those in the West) to this type of risks. At other times, when there is a observable negative effects in local markets (in countries affected by a particular local crisis with possible wider consequences) there is a noticeable lack of an efficient transmission mechanism from the local market to the global one. Looking at this matter (in both cases) from the position of market efficiency we can say that it's quite a paradoxical phenomenon.

Perhaps the explanation of the aforementioned market response to LFHI risks lies in the fact that the probability of materialisation these tail-risks materialising is (by the very nature of these things) very difficult to estimate, so therefore, as long as this uncertainty about the value of these risks (i.e. their probability and intensity) remains relatively high, the market (i.e. market participants) simply don't know how to include them in prices of commodities and financial instruments, so it disregard them temporarily as incomplete information until they become "priceable". As a conclusion, we can say that LFHI risks eventually gets priced in (with some time delay), but sometimes only after the fact, rendering the markets' ability to correctly signal LFHI risks (in the prices of commodities and financial instruments) ineffective.

It can be argued that such a behaviour of the market (i.e. prevailing perception of its participants) in certain situation is quite a rational one, but also, just wrong in others (when it's not able to assess the systemic importance of the interaction between local and global risk factors).

...I hope we haven't already forgotten that even the markets can sometimes be "wrong", and that we still do not sufficiently understand the market dynamics.