Rep. Mike Kelly

Question
Is trading concentrated in a few companies rather than spread across a diversified portfolio?

Insider signature
Informed trading concentrates in the few issuers tied to the information.

Analyst note
Dollar HHI 0.119 (identified set [0.037, 0.276]); trade-count HHI 0.027 vs. 95th percentile 0.025 of 53 random draws from all congressional trades (p = 0.012). Normalized sector entropy 0.78 [0.69, 0.87]. Top three issuers (GLD, HLMEX, CLF) carry at least 21% of dollar volume.

Evidence

Statistical

Effective number of issuers

8 [4, 27] 90%

Question
Were trades disclosed within the STOCK Act's 45-day limit?

Insider signature
Delayed disclosure can conceal well-timed trades.

Analyst note
Late share 0.065, Wilson [0.043, 0.099]; 90th percentile lag 63 days; longest 771 days.

Evidence

Descriptive

Late share

6.5% [4.3%, 9.9%] 90%

Robust to variants

Yes

Question
Did trading change after press scrutiny or an ethics inquiry?

Insider signature
A member who trades on information has an incentive to keep control of trading after scrutiny.

Analyst note
After scrutiny on 2021-09-29, trading ran at 0.73 times its earlier rate. Change point 2018-01-05 (p = 0.004).

Evidence

Statistical

Rate ratio after/before

0.73 [0.43, 1.22] 90%

Question
When a traded stock made a large unexpected move soon after the trade, was the trade on the right side?

Insider signature
Buying shortly before a large jump or selling shortly before a large drop.

Analyst note
Hit rate 0.67, Clopper-Pearson [0.14, 0.98], exact one-sided p = 0.500; threshold |AR| > 3 sigma.

Evidence

Statistical

Hit rate

66.7% [13.5%, 98.3%] 90%

p-value

0.500

Holm-adjusted p

1.000

BH q-value

1.000

Margin δ

10 points above a coin flip

Question
Are the member's trades unusually large, especially just before news?

Insider signature
Informed trades are large relative to the trader's normal activity, because the profit from an informational edge scales with position size.

Analyst note
Lowest-band share 0.830, 90% Wilson [0.730, 0.898]. Event-linked size gap Delta_z = 0.00 (within-issuer permutation p = 1.000); 0 event-linked trades with z > 3 in bands of $50,001 or more.

Evidence

Statistical

Lowest-band share

83.0% [73.0%, 89.8%] 90%

p-value

1.000

Holm-adjusted p

1.000

BH q-value

1.000

Question
Do trades arrive in routine rebalancing batches, or sporadically just before company news?

Insider signature
Opportunistic trades carry information and routine trades do not.

Analyst note
Batch-day share 8/26, Wilson [0.18, 0.47]; index of dispersion 2.72 (chi-square p = 0); routine share 0.06 (portfolio-routine or Cohen-Malloy-Pomorski two-year rule). Proximity share 0.019 vs. circular-shift null mean 0.059, p = 0.906.

Evidence

Statistical

Batch-day share

30.8% [18.3%, 46.9%] 90%

p-value

0.906

Holm-adjusted p

1.000

BH q-value

1.000

Question
Does the member trade industries their committees oversee, and do those trades do better?

Insider signature
Members may trade on information about industries their committees oversee.

Analyst note
Committees HSWM; jurisdiction sectors Distillers & Vintners, Health Care, Life & Health Insurance, Tobacco. Dollar share 0.154 (identified set [0.029, 0.521], bootstrap 0.054-0.283); benchmarks S&P 500 0.139, congressional peers 0.119.

Evidence

Descriptive

Jurisdiction excess share

1.5% [-8.5%, 14.4%] 90%

Margin δ

0.5% per trade over 21 days

Question
After accounting for noise, does the member's trading beat that of other members?

Insider signature
A member with an information edge stands out from the cross-section of congressional traders.

Analyst note
Empirical Bayes: raw -0.0040, posterior -0.0028 (sd 0.0015); Congress mean -0.0028, between-member variance 0.000000.

Evidence

Statistical

Shrunk signed 21-day CAR

−0.28% [−0.53%, −0.03%] 90%

Margin δ

0.5% per trade over 21 days

Question
Does the member trade the same stock in the same direction as other members at the same time more than chance predicts?

Insider signature
Shared nonpublic information spreads through networks and shows up as clustered trades.

Analyst note
Co-traded share 0.132; circular-shift null mean 0.133, 95th pct 0.245, p = 0.580. Co-traded signed CAR -0.0123, p = 0.599.

Evidence

Statistical

Co-traded share

13.2%

p-value

0.599

Holm-adjusted p

1.000

BH q-value

0.902

Robust to variants

Yes

Question
Does the member use options, and do option trades point the right way?

Insider signature
Informed traders favour options because leverage raises the payoff to private information.

Analyst note
No option trades were reported.

Evidence

Descriptive

Margin δ

0.5% per trade over 21 days

Question
Do purchases rise and sales fall more than the market in the weeks after the trade?

Insider signature
Purchases before price rises and sales before price falls.

Analyst note
Mean signed CAR(1,21) = -0.0040; 90% KP interval [-0.0312, 0.0232]; t_BMP = -0.88, r-bar = 0.022, t_KP = -0.63, Patell Z = -0.87, Corrado Z = -0.43. Dollar-weighted identified set [-0.0899, 0.0236]. Calendar-time alpha 15.68%/yr (NW t = 0.95, 210 days).

Evidence

Statistical

Mean signed 21-day CAR

−0.40% [−3.12%, +2.32%] 90%

p-value

0.735

Holm-adjusted p

1.000

BH q-value

0.983

Margin δ

0.5% per trade over 21 days

Robust to variants

Yes

Question
Before the market's biggest moves, did the member's trades point the right way?

Insider signature
An informed trader's trades just before a market-moving event point in the direction of the coming move.

Analyst note
No market-wide shocks or documented events fell within five trading days after the member's trades.

Evidence

Descriptive

Margin δ

10 points above a coin flip

Question
Does the member direct the trades, and does trading change when the member's access to information changes?

Insider signature
The person holding the information directs the trade.

Analyst note
Trading arrangement: not documented; trading during absence ran at 0.00 times the normal rate.

Evidence

Statistical

Absence rate ratio

0.00 [0.00, 6.25] 90%

Question
Did trades in stocks affected by the member's legislative work come before votes and move the right way?

Insider signature
Trading on knowledge of legislation before it is public.

Analyst note
13 priced trades fell in affected industries in the 30 days before the member's committee events, but on only 9 distinct dates; at least 10 dates are needed for a date-clustered comparison.

Evidence

Descriptive

Margin δ

0.5% per trade

Question
Do trades in a spouse's or dependent's name do better than the member's own?

Insider signature
Routing informed trades through family accounts.

Analyst note
Family signed CAR -0.0054 [-0.0375, 0.0268], t_KP = -0.59; family minus self -0.0042 [-0.0529, 0.0287], p = 0.570.

Evidence

Statistical

Family-account signed CAR

−0.54% [−3.75%, +2.68%] 90%

p-value

0.720

Holm-adjusted p

1.000

BH q-value

0.928

Margin δ

0.5% per trade over 21 days

Question
Does the member buy and sell the same stock within six months, and do those round trips beat the market?

Insider signature
Short-horizon round trips capture a temporary informational edge.

Analyst note
2 round trips but fewer than three priced.

Evidence

Descriptive

Margin δ

0.5% per trade over 21 days