It would seem that share prices incorporate expectations regarding the availability of willing and able buyers, i.e., of investors with sufficient liquidity. Such expectations are influenced by the price level - it is more difficult to find buyers at higher prices - by the general market sentiment, and by externalities and new information, including new information about earnings.
The capital gain anticipated by a rational investor takes into consideration both the expected discounted earnings of the firm and market volatility - the latter being a measure of the expected distribution of willing and able buyers at any given price. Still, if earnings are retained and not transmitted to the investor as dividends - why should they affect the price of the share, i.e., why should they alter the capital gain? Earnings serve merely as a yardstick, a calibrator, a benchmark figure. Capital gains are, by definition, an increase in the market price of a security. Such an increase is more often than not correlated with the future stream of income to the firm - though not necessarily to the shareholder. Correlation does not always imply causation.
Stronger earnings may not be the cause of the increase in the share price and the resulting capital gain. But whatever the relationship, there is no doubt that earnings are a good proxy to capital gains.
Hence investors' obsession with earnings figures. Higher earnings rarely translate into higher dividends. But earnings - if not fiddled - are an excellent predictor of the future value of the firm and, thus, of expected capital gains. Higher earnings and a higher market valuation of the firm make investors more willing to purchase the stock at a higher price - i.e., to pay a premium which translates into capital gains.
The fundamental determinant of future income from share holding was replaced by the expected value of share- ownership. It is a shift from an efficient market - where all new information is instantaneously available to all rational investors and is immediately incorporated in the price of the share - to an inefficient market where the most critical information is elusive: how many investors are willing and able to buy the share at a given price at a given moment.
A market driven by streams of income from holding securities is "open". It reacts efficiently to new information. But it is also "closed" because it is a zero sum game. One investor's gain is another's loss. The distribution of gains and losses in the long term is pretty even, i.e., random. The price level revolves around an anchor, supposedly the fair value.
A market driven by expected capital gains is also "open" in a way because, much like less reputable pyramid schemes, it depends on new capital and new investors. As long as new money keeps pouring in, capital gains expectations are maintained - though not necessarily realized.
But the amount of new money is finite and, in this sense, this kind of market is essentially a "closed" one. When sources of funding are exhausted, the bubble bursts and prices decline precipitously. This is commonly described as an "asset bubble".
This is why current investment portfolio models (like CAPM) are unlikely to work. Both shares and markets move in tandem (contagion) because they are exclusively swayed by the availability of future buyers at given prices.
This renders diversification inefficacious. As long as considerations of "expected liquidity" do not constitute an explicit part of income-based models, the market will render them increasingly irrelevant.