How Much Dark Trading Is Too Much?

Not every trade happens on a lit exchange with a publicly displayed order book. A large share of US equity volume — often estimated at close to half of all trades in some periods — executes in dark pools, which don't display quotes before a trade, or is internalized by wholesale market makers who fill retail orders directly rather than routing them to an exchange. Prices for that trading are typically set by referencing the lit market's quotes, most often the midpoint of the NBBO, which raises a structural question: if a large and growing share of volume free-rides on prices set elsewhere, does the lit market still generate accurate prices for everyone to reference?

Why the concern is real, up to a point

Price discovery — the process by which trading activity reveals what an asset is actually worth — happens primarily where orders compete openly, which is largely the lit, displayed market. If enough volume migrates to dark venues that reference lit prices rather than contributing to setting them, the lit market thins out, in principle making its prices noisier and easier to move, which paradoxically could make the reference price that dark venues rely on less reliable. There's a self-limiting logic to how bad this can get, though: as dark trading share rises and lit liquidity thins, spreads on the lit market would tend to widen, which makes dark execution at the (worse) midpoint less attractive relative to trading on the lit market directly, pulling some volume back.

What the evidence shows

Empirical studies generally find that price discovery holds up reasonably well even at dark trading shares in the 30-40% range typical of many liquid US stocks — lit markets remain the dominant venue for setting prices, and measured price efficiency doesn't degrade sharply until dark share reaches quite extreme levels or concentrates heavily in less-liquid names. The effect is also uneven: it shows up more clearly in smaller, thinner-traded stocks, where losing even a modest share of order flow to dark venues has a proportionally bigger impact on lit-market depth, than in the most liquid large caps.

What this means in practice

Regulators including the SEC and European authorities (through MiFID II's dark trading caps) have implemented rules limiting how much of a given stock's volume can trade dark, precisely as a hedge against the tipping-point risk even though clear evidence of current harm is limited. The practical takeaway is that dark trading share is a metric worth watching per stock rather than market-wide — a healthy aggregate number can still mask individual illiquid names where dark migration has gone far enough to matter.