W/P 25 Oct 21, 2026
Overnight Corporate Bond Trading
Purpose
To set out what issue 25 says, and to trace each figure in it to the paper or to John's post. Commentary on this page is the site's. John's words appear only in the marked quotations.
Source
- Authors
- Farshid Abdi, Marco Rossi, Botao Wu
- Venue
- SSRN working paper
- Version
- This version: September 2026
- Post
- LinkedIn, Oct 21, 2026. Link not added yet.
John's post
Workpapers Meet White Papers - #25
Corporate bonds apparently have a night shift: bigger trades, delayed receipts, and more information per transaction.
In Overnight Corporate Bond Trading (https://lnkd.in/gSMqXr5B), Farshid Abdi, Marco Rossi, and Botao Wu study trades executed between 6:30 p.m. and 8:00 a.m. ET, when public TRACE dissemination waits until morning.
A few findings that stood out:
🔹 The average overnight institutional trade was $1.94 million—45% larger than the $1.34 million intraday average.
🔹 Overnight spreads were about 8 basis points higher than average intraday spreads, but lower than trades requiring immediate daytime execution.
🔹 Overnight trades contributed nearly twice as much to daily price discovery. Order imbalances also predicted returns in other bonds from the same issuer.
🔹 Overnight block trading was concentrated among large dealers, with higher dealer spreads and less favorable outcomes for receiving investors.The economics are more nuanced than a simple liquidity-versus-transparency debate. Overnight trades carried wider spreads than the average daytime transaction, but they were often cheaper than intraday trades that required immediate execution. Temporary opacity appears to create value by allowing large positions to be intermediated without immediately revealing trading intent.
That value, however, is not distributed evenly. The benefits of delayed disclosure were concentrated among large dealers, while investors on the other side of those trades faced less favorable outcomes. The result is a tradeoff between facilitating block liquidity and increasing information asymmetry.
That makes post-trade transparency a market-design question, not a simple choice between “transparent” and “opaque.” Faster disclosure may improve price discovery and reduce information advantages, but it may also make large trades more difficult or expensive to execute. The more useful question is how much delay is appropriate, for which trades, and with what protections for investors receiving the offsetting flow.
How should reporting rules balance timely price discovery with the practical need to execute and intermediate large corporate-bond trades?
#WorkpapersMeetWhitePapers #CorporateBonds #MarketStructure #TRACE #BrokerDealers
Findings
Each finding John lists, beside the paper's matching figure in the paper's own words. Tap a tick to see the exact quote and where it comes from.
| John's post | The paper, in its words |
|---|---|
The average overnight institutional trade was $1.94 million—45% larger than the $1.34 million intraday average. |
1.94 USD millions (mean overnight institutional trade size)
|
Overnight spreads were about 8 basis points higher than average intraday spreads, but lower than trades requiring immediate daytime execution. |
8 basis points (unconditional overnight minus intraday bid-ask spread)
|
Overnight trades contributed nearly twice as much to daily price discovery. Order imbalances also predicted returns in other bonds from the same issuer. |
1.74 WPCT (ratio; 1.0 = proportional to trade share)
0.007 percent return per $1 million sibling-bond overnight imbalance (h = 15)
|
Overnight block trading was concentrated among large dealers, with higher dealer spreads and less favorable outcomes for receiving investors. |
0.268 percent (dealer spread, overnight blocks with at least one offsetting trade)
−0.147 percent (receiver spread, overnight blocks with at least one offsetting trade)
|
Exhibits
Exhibit 1
A block's night
One night for a block trade, from the close of TRACE to the morning report. Clock times come from FINRA rules. The figures at each stage are averages the paper measured across many blocks; the paper does not time the stages of any single trade.
-
6:30 p.m.
1. The overnight window opens
TRACE closes at 6:29:59 p.m. A trade executed from 6:30 p.m. through 11:59:59 p.m. is reported the next business day, after TRACE opens.
-
About $1.695 million larger
Trades from 6:30 p.m. to midnight are larger than intraday trades, in the paper's baseline specification.
-
-
9:00 p.m., an example time
2. A block executes
In this example the block executes at 9:00 p.m. The time is illustrative and does not come from the paper.
The paper defines a block as a trade of $10 million or more.
Net of price impact, the paper finds realized spreads statistically indistinguishable between overnight and intraday blocks.
-
0.275% overnight, 0.150% intraday
Initiating a block costs more overnight. This is the initiator spread for blocks with at least one offsetting trade.
-
0.131 to 0.146 points more price impact overnight
Price impact on blocks is larger overnight than intraday, across the paper's block samples.
-
-
Any time before the report
3. The dealer offsets before the report
While the block is still unreported, the dealer can trade out of part of the position. The paper measures spreads on these pre-report offsetting trades and on all offsetting trades.
The mean share of an overnight block that dealers offset before it is reported ranges from 5.9% to 14.4%.
-
0.181 to 0.204 points before the report, against 0.082 to 0.106 points for all offsetting trades
Dealer spreads on offsetting trades made before the block is reported are higher overnight than intraday, by more than the gap across all offsetting trades.
-
0.268% overnight, 0.186% intraday
For blocks with at least one offsetting trade, dealers earn a wider spread overnight.
-
−0.147% overnight, −0.059% intraday
Receiving investors, on the other side of the dealer's offsetting trades, fare worse overnight.
-
-
8:00:00 a.m.
4. TRACE open
TRACE opens at 8:00:00 a.m. The overnight block must be reported within 15 minutes after the open, and FINRA publishes each report as soon as it receives it. Reading of Rules 6730 and 6750
-
1.74 overnight, 0.99 intraday
Per trade, the average overnight trade contributes more to daily price discovery than the average intraday trade.
-
About 11 basis points the same day, about 19 after 15 trading days
A $10 million overnight order imbalance in a bond has a same-day price impact, and a larger one 15 trading days later.
-
Exhibit 2
The 2003 natural experiment
The two bars measure different things. The first is a level, the overnight share of dollar volume before the event. The second is a change in that share, relative to matched bonds. They share one scale but are not a before and after pair.
Show as a table
| Measure | Paper figure |
|---|---|
| Overnight share of dollar volume for the newly disseminated bonds, before April 14, 2003 | 0.029% |
| Rise in that share after daytime dissemination began, relative to matched bonds not yet disseminated | 0.806 points |
Over the same event, the overnight share of price discovery for these bonds rose too. 1.041 points
The comparison covers 90 newly disseminated bonds and 270 matched bonds.
Exhibit 3
Rules behind the overnight window
Rules as of Oct 25, 2026. Each line is the site's summary of a rule item, with the rule text behind the tick.
-
TRACE is open from 8:00:00 a.m. to 6:29:59 p.m. Eastern Time on business days.
-
During system hours, a corporate bond trade must be reported as soon as practicable and no later than 15 minutes after execution.
-
The next-day rule covers trades executed at or after 6:30 p.m. through 11:59:59 p.m. Eastern Time.
-
A trade done from 6:30 p.m. to 11:59:59 p.m., or on a weekend or holiday, is reported the next business day within 15 minutes after TRACE opens, marked as/of.
-
A trade done between midnight and 7:59:59 a.m. on a business day is reported the same day, within 15 minutes after TRACE opens.
-
FINRA publishes each trade report as soon as it receives it, so an overnight trade reaches the public tape only after it is reported once TRACE opens.
-
Public TRACE prints cap trade size at $5 million for investment-grade bonds (shown as 5MM+) and $1 million for non-investment-grade bonds (shown as 1MM+).
-
The SEC approved SR-FINRA-2024-004 on September 20, 2024 (Release No. 101121). Its cuts to TRACE reporting timeframes under Rule 6730(a)(1) were rescinded by SR-FINRA-2025-008 before they took effect.
-
In Regulatory Notice 25-17, FINRA said it is not moving forward with cutting the 15-minute TRACE reporting limit to one minute. The SEC had approved the cut, but it had not taken effect.
-
The SEC found that under the 15-minute rule, 82.9% of TRACE trades were reported within one minute and 97.6% within five minutes.
-
FINRA is widening its affiliate principal-transaction flag to cover trades with member affiliates, so the duplicate same-price prints between a dealer and its affiliate are not shown on the tape.
-
FINRA's Trade Reporting Facilities extend their hours starting December 6, 2026, to run from 9:00 p.m. Sunday to 8:00 p.m. Friday, with a one-hour pause from 8:00 p.m. to 9:00 p.m. Monday through Thursday.
-
The overnight extension covers over-the-counter trades in NMS stocks only, so bonds keep TRACE's 8:00 a.m. to 6:29:59 p.m. day.
Exhibit 4
Questions an auditor might ask
Questions to consider, not guidance. Each names the rule it rests on. Rule text as of Oct 25, 2026 and Oct 26, 2026.
-
How does the firm show that prices on overnight block trades are fair when its offsetting trades take place before the block is public?
- FINRA Rule 2121, Fair Prices and Commissions
- FINRA Rule 6730(a)(1)(D)
-
How does the firm compare the spreads it earns on overnight blocks with those it earns on intraday blocks in its fair-pricing review?
- FINRA Rule 2121, Fair Prices and Commissions
-
What evidence shows that the firm's best-execution reviews cover customer orders filled while TRACE is closed?
- FINRA Rule 5310, Best Execution and Interpositioning
- FINRA Rule 6710(t), TRACE System Hours
-
If a small number of dealers accounts for a large share of overnight customer volume, how does the firm weigh that in where it routes overnight orders?
- FINRA Rule 5310, Best Execution and Interpositioning
-
For overnight trades with non-institutional customers, how does the firm set the price it measures the mark-up or mark-down against on the confirmation, when overnight trades do not print until TRACE opens?
- FINRA Rule 2232(c), Customer Confirmations
- FINRA Rule 6730(a)(1)(D)
-
Are trades executed after the evening close reported within the deadline after the next TRACE open, and how are late reports justified?
- FINRA Rule 6730(a)(1)(D)
- FINRA Regulatory Notice 25-17
Conclusion
The more useful question is how much delay is appropriate, for which trades, and with what protections for investors receiving the offsetting flow.
How should reporting rules balance timely price discovery with the practical need to execute and intermediate large corporate-bond trades?
Cross-references
No other workpapers are cross-referenced yet.
Rules this site lists for the issue
- FINRA Rule 6730
- FINRA Rule 6750
- Agreed to the paper, with page
- Taken from John's post
- Agreed to rule text