First, you might find this recent paper by Israeli, Lee and Sridharan (Review of Accounting Studies, forthcoming) interesting. This is the abstract:
We examine whether an increase in ETF ownership is accompanied by a decline in pricing efficiency for the underlying component securities. Our tests show an increase in ETF ownership is associated with: (1) higher trading costs (bid-ask spreads and market liquidity); (2) an increase in “stock return synchronicity”; (3) a decline in “future earnings response coefficients”; and (4) a decline in the number of analysts covering the firm. Collectively, our findings support the view that increased ETF ownership can lead to higher trading costs and lower benefits from information acquisition. This combination results in less informative security prices for the underlying firms.
Second, Seth Klarman from Baupost published some interesting thoughts on this matter in his 2016 letter to investors (not public, but copies surface here and there):
One of the perverse effects of increased indexing and ETF activity is that it will tend to ‘lock in’ today's relative valuations between securities. When money flows into an index fund or index-related ETF, the manager generally buys into the securities in an index in proportion to their current market capitalization (often to the capitalization of only their public float, which interestingly adds a layer of distortion, disfavoring companies with large insider, strategic, or state ownership). Thus today's high-multiple companies are likely to also be tomorrow's, regardless of merit, with less capital in the hands of active managers to potentially correct any mispricings. Conversely with money pouring into market indices, stocks outside the indices may be cast adrift, no longer attached to the valuation grid but increasingly off of it. This should give long-term value investors a distinct advantage. The inherent irony of the efficient market theory is that the more people believe in it and correspondingly shun active management, the more inefficient the market is likely to become.
Third, I agree that there seems to be a contradiction: The "iron law of costs" (no correlation between gross fund performance and fees, hence a negative correlation between net fund performance and fees) means that investors will be worse off if they try to beat the market. Nevertheless, investors trying to beat the market actually contribute to valuable public goods, such as scrutiny of listed companies and the efficiency of capital allocation in the market.
I would argue that passive investing and index investing are not exactly the same thing and that this difference is crucial to understand this contradiction. In order to save costs, investors should abstain from trying to beat the market; the advantage of passive investing is that costs are lower. However, can the cost-saving goal of passive investing only be achieved by EXACTLY replicating an index (i.e., index investing) or could investors just as well simply build their own low-maintenance portfolio of 50-100 big companies? My point is thus that the goals of passive investing could largely be achieved without giving way to the disadvantages of index investing (in short, less efficient capital markets). In this spirit, the third resource that you might find interesting is The Kay Review of UK Equity Markets and Long-Term Decision Making, a report published in 2012 on behalf of the UK government. On this specific topic, Kay writes:
The objectives of ‘passive investment’ strategies may not be best achieved (other than tautologically) by replicating a benchmark index based on all stocks in a particular category, if that category is not defined by any criteria of relevance to the underlying beneficiaries. An indexed portfolio holds a wide range of stocks – the widest possible in any asset class – but holding a wide range of stocks is not the same as holding a diversified portfolio. An index fund is concentrated in a few sectors – in the UK index financials were once heavily represented, now oil and mining are [remember, this was written in 2012]. Telecommunications and pharmaceuticals are over-represented relative to their contribution to economic activity, because these activities are mostly
conducted by large companies. House-building is mostly conducted by small businesses; automobile construction is not included in the UK index because it is mostly conducted by foreign companies. ‘Fundamental indices’ whose weights are related to the economic significance of companies or markets rather than based on market capitalisation are an attempt to devise passive strategies which avoid these defects, and also to escape the losses from long-run mean reversion which arise from the adoption of market weightings in index funds or benchmarks. Index funds are necessarily overweight in overpriced sectors and underweight in underpriced sectors and this is true even if one has no knowledge of which
sectors are over or underpriced.
Should passive strategies be based on indices at all? Interest in passive strategies developed in the 1970s as a result of scepticism about the ability of asset managers to pick stocks successfully and concern about the level of fees and costs of turnover associated with many active management strategies. One large endowment described to us the development of a portfolio of large, global
companies with the intention of building stakes in these companies and holding them indefinitely. Such a strategy may have lower trading costs than even a conventionally passively managed fund. We note that several asset managers are now developing similar investment strategies.
(Of course, one of the reasons why passive investing is often conflated with index investing is that investors are increasingly unwilling to tolerate even short bouts of underperformance versus the index. See p. 40-42 of the Kay Review for more details on these cultural / psychological / institutional mechanisms.)