We have written before that the bid-ask spread has a floor, because US stocks move in one-cent
increments and a spread cannot be narrower than a tick. That is true above a dollar. Below a dollar
it is not, and the exception turns out to be more interesting than the rule.
Stocks priced under $1 quote in increments of $0.0001, a hundredth of a cent. The grid they trade on
is a hundred times finer than the one everyone else uses. If a fine grid made markets tight, these
would be the tightest quotes on the exchange.
A note on the words. Penny stock is a regulatory term covering everything under $5, which lumps
together two groups that quote on different grids. This study uses the line that actually changes
the rules, which is a dollar, and compares what sits on either side of it.
1. The grid is real, not theoretical
A rule on paper is not the same as behaviour in the book, so the first thing to check is whether
sub-dollar quotes actually use the extra digits.
They do. Across 276,345 recorded seconds of sub-dollar quotes, 76.7% of best bids were not sitting
on a whole cent. Only 23.3% landed on a round cent, which is roughly what the extra digits would
produce on their own.
So these are not cent-priced stocks with two decorative zeros. Their quotes carry four decimal
places, and the tick that separates one price from the next is a hundredth of a cent.
2. The spread does not follow the grid
| Group | 25th percentile | Median | 75th | 90th |
| Under $1 | 0.87% | 1.81% | 2.84% | 4.69% |
| $1 to $5 | 0.44% | 0.72% | 1.15% | 2.56% |
The group with the hundred-times-finer grid quotes a spread 2.5 times wider at the median, and it is
wider at every percentile in the table.
Measured in each group's own ticks the gap is enormous in the other direction.
| Group | Median spread in ticks | Seconds sitting on the minimum tick |
| Under $1 | 96 | 2.4% |
| $1 to $5 | 1 | 60.4% |
A $1 to $5 stock spends most of its life at the narrowest spread its grid permits. The grid is the
binding constraint, and the floor we described earlier is a real floor: sixty percent of the time
the quote is pressed against it.
A sub-dollar stock is almost never at its minimum. It sits 96 ticks off the floor at the median,
which means the floor is not doing anything at all. Something else is setting that spread.
3. What is setting it, if not the tick
The tick can only ever be a lower bound. When it is large relative to the price it binds, and the
spread stops there. When it is small it stops binding, and the spread goes wherever the risk of
quoting takes it.
On a $2 stock a cent is fifty basis points, so one tick is already a meaningful width and market
makers have nowhere tighter to go. On a $0.50 stock a hundredth of a cent is two basis points, so a
maker who wants real protection simply quotes many ticks wide, ninety-six of them at the median, and
nothing in the rulebook objects.
That is the whole result. The tick size question and the spread question are only the same question
while the tick is binding. Below a dollar it stops binding and the two come apart.
4. Does it hold across names
Pooled numbers can be one loud ticker, so we measured each symbol-day separately and kept every one
with at least three hundred recorded seconds.
| Symbol-days measured | Median off-cent share | At or above 80% | At or above 50% | Below 10% |
| 89 | 78.9% | 41 | 84 | 1 |
Eighty-four of eighty-nine used the fine grid for at least half their quoted seconds. The spread
result travels as well: the median symbol-day quoted 2.37% and 78 of 89 were above 1.00%.
The single exception is worth naming rather than hiding. One symbol-day sat off the cent only 0.7%
of the time, quoting a sub-dollar stock almost entirely on whole cents while running a wide spread
anyway. Being allowed to use a finer grid does not oblige anyone to use it.
The spread of outcomes is very wide, from 0.25% to 12.77% at the symbol-day median. This is not a
population where an average describes anybody.
5. What this changes for a limit order
Price improvement is cheap here and nearly free in percentage terms. Beating the best bid by one
tick on a $0.50 stock costs two basis points, against fifty basis points to do the same thing on a
$2 stock. If you are working a limit order, the sub-dollar grid is the one place where stepping in
front of the queue barely costs anything.
The flip side is that everyone else can do it to you just as cheaply, so a resting order can be
stepped in front of repeatedly for almost nothing. Queue priority is worth less on a fine grid,
because priority is only valuable when improving the price is expensive.
And none of it makes the round trip cheap. A median spread of 1.81% is the toll, and crossing it
twice is what a trade has to beat before anything else is considered.
6. A correction to our own guide
Our guide on the bid-ask spread states that a US equity spread cannot be narrower than one cent. In
that guide the claim is used to explain why cheap stocks are expensive to trade, and for the stocks
it was measured on the arithmetic holds.
Stated without a price qualifier it is wrong, and this study is the evidence. We have added the
qualifier there and are noting it here rather than quietly editing, because a reader who learned the
rule from us deserves to know which part changed.
7. What this does not measure
The sample is thin compared with our other studies. It rests on 109 sub-dollar symbol-days, of which
89 cleared the three-hundred-second bar, and sub-dollar names in a day trading scan are unusual by
construction rather than typical of all cheap stocks.
We measured what was quoted, not what filled. Whether a fine grid means your order actually gets
done at a better price is a different question and this study does not answer it.
We did not measure returns. Nothing here says whether sub-dollar stocks are worth trading.
8. Sample accounting
The sample is 38 sessions drawn evenly from 264 recorded sessions between 2025-04-02 and 2026-07-27,
covering 276,345 quoted seconds under $1 and 2,047,344 between $1 and $5, regular session only.
Spreads are measured against the midpoint and reported as medians and quartiles rather than means,
since the tail in this population is long enough to move an average anywhere. Tick counts use each
group's own regulatory minimum, $0.0001 below a dollar and $0.01 above it.
Seconds were used only where both sides of the quote were present. The universe is the stocks our
own scan surfaced each day, which is why sub-dollar names appear at all.