[{"id":"T1","family":"size_breadth","name":"Trade size","question":"Are the member's trades unusually large, especially just before news?","signature":"Informed trades are large relative to the trader's normal activity, because the profit from an informational edge scales with position size.","statistic":"Lowest-band share; robust size score z; event-linked size gap (event-linked minus other mean z)","formula":"z_i=\\frac{\\ln\\tilde v_i-\\operatorname{med}(\\ln\\tilde v)}{1.4826\\,\\operatorname{MAD}(\\ln\\tilde v)},\\quad \\Delta_z=\\bar z_{E=1}-\\bar z_{E=0}","inference":"Wilson interval for the lowest-band share; within-issuer permutation test for the size gap","decision_rule":"Fits if the size gap is significantly positive after Holm and BH adjustment. Does not fit if the lower bound of the lowest-band share is above one half and no event-linked trade is a size outlier. Otherwise open question.","margin":null,"inputs":["trades","firm events from daily returns"],"limitations":"Bands are wide; wealthy members trade larger amounts for reasons unrelated to information.","doc_section":"sec:t1"},{"id":"T2","family":"size_breadth","name":"Breadth and concentration","question":"Is trading concentrated in a few companies rather than spread across a diversified portfolio?","signature":"Informed trading concentrates in the few issuers tied to the information.","statistic":"Dollar-weighted HHI (identified set), effective number of issuers, normalized sector entropy","formula":"\\mathrm{HHI}=\\sum_j s_j^2,\\quad N_{\\text{eff}}=1/\\mathrm{HHI},\\quad \\tilde H=-\\sum_g q_g\\ln q_g/\\ln G","inference":"Sharp bounds over the amount box; Monte Carlo benchmark drawn from all congressional trades; date-clustered bootstrap","decision_rule":"Does not fit if trade-count concentration is at most 0.10 (ten or more effective issuers) or within the peer benchmark, and sector entropy is above 0.7. Otherwise open question; concentration alone never produces fits.","margin":null,"inputs":["trades","sector map"],"limitations":"A concentrated portfolio is common among members who hold a family business or employer stock.","doc_section":"sec:t2"},{"id":"T3","family":"timing","name":"Cadence and event proximity","question":"Do trades arrive in routine rebalancing batches, or sporadically just before company news?","signature":"Opportunistic trades carry information and routine trades do not.","statistic":"Batch share, index of dispersion, routine share, proximity share to firm events","formula":"\\bar W=\\frac1N\\sum_i\\mathbf 1\\{\\exists e\\in\\mathcal E_{j(i)}:0<e-t_i\\le w\\}","inference":"Chi-square dispersion test; circular-shift permutation of trade dates against event dates","decision_rule":"Fits if the proximity share is above the 95th percentile of its permutation null after adjustment and the excess sits in non-routine trades. Does not fit if proximity is within its null range and most trading days are batches. Otherwise open question.","margin":null,"inputs":["trades","firm events from daily returns"],"limitations":"Firm events are proxied by large abnormal moves when an earnings calendar is not available.","doc_section":"sec:t3"},{"id":"T4","family":"timing","name":"Event timing (abnormal returns)","question":"Do purchases rise and sales fall more than the market in the weeks after the trade?","signature":"Purchases before price rises and sales before price falls.","statistic":"Mean signed cumulative abnormal return over 21 trading days; calendar-time alpha","formula":"\\hat\\theta_4=\\frac1N\\sum_i d_i\\,\\mathrm{CAR}_i(1,H),\\quad t_{KP}=t_{BMP}\\sqrt{\\frac{1-\\bar r}{1+(N-1)\\bar r}}","inference":"Kolari-Pynnonen adjusted BMP test (primary), Patell, Corrado rank, Newey-West calendar-time alpha, date-clustered bootstrap","decision_rule":"Fits if the Kolari-Pynnonen lower bound and the calendar-time lower bound exceed zero after Holm and BH adjustment. Does not fit if both upper bounds are below the margin. Otherwise open question.","margin":{"value":0.005,"unit":"return","label":"0.5% per trade over 21 days"},"inputs":["trades","daily returns","factors"],"limitations":"Range-reported amounts; low power for members with few trades; disclosure dates, not execution times.","doc_section":"sec:t4"},{"id":"T5","family":"timing","name":"Direction before market-wide shocks","question":"Before the market's biggest moves, did the member's trades point the right way?","signature":"An informed trader's trades just before a market-moving event point in the direction of the coming move.","statistic":"Hit rate over episodes (net dollar direction in the five trading days before each shock)","formula":"\\hat h=X/n,\\quad X\\sim\\mathrm{Bin}(n,1/2)\\text{ under }H_0","inference":"Exact binomial test, Clopper-Pearson interval, sign-flip permutation of the magnitude-weighted statistic","decision_rule":"Fits if the one-sided exact p-value is significant after Holm and BH adjustment. Does not fit if the hit rate is below one half and T4 does not fit. Otherwise open question.","margin":{"value":0.1,"unit":"share","label":"10 points above a coin flip"},"inputs":["trades","benchmark returns","documented episodes (optional)"],"limitations":"Few market-wide shocks fall within any one member's trading; most members have small episode counts.","doc_section":"sec:t5"},{"id":"T6","family":"behavior","name":"Control and independence","question":"Does the member direct the trades, and does trading change when the member's access to information changes?","signature":"The person holding the information directs the trade.","statistic":"Documentary record of discretion; rate ratio of trading during absence versus presence","formula":"\\hat\\rho_6=\\frac{Y_A/T_A}{Y_{A^c}/T_{A^c}}","inference":"Exact conditional binomial interval for the Poisson rate ratio","decision_rule":"Does not fit if discretion lies with others and the rate ratio interval includes one. Fits if the member directs trades and T4 or T5 fits. Otherwise open question.","margin":null,"inputs":["arrangements","absence periods","trades"],"limitations":"Requires a documentary record that exists for few members; stated arrangements are not audited. Absence periods (runs of two or more voting days with every House roll call missed) are reported as context, but without an arrangement record the test is not assessed.","doc_section":"sec:t6"},{"id":"T7","family":"compliance","name":"Disclosure timeliness","question":"Were trades disclosed within the STOCK Act's 45-day limit?","signature":"Delayed disclosure can conceal well-timed trades.","statistic":"Late share (over 45 days from the trade and not within 30 days of a stated notification date), median and 90th percentile lag; late-minus-timely signed CAR","formula":"\\hat\\pi_L=\\frac1N\\sum_i\\mathbf 1\\{f_i-t_i>45\\ \\wedge\\ \\neg(n_i\\ge t_i \\wedge f_i-n_i\\le 30)\\}","inference":"Wilson interval at trade and filing level; date-clustered bootstrap for the return difference","decision_rule":"Does not fit if no trade was late. Open question if any trade was late. Fits only if late trades earn significantly more than timely ones.","margin":null,"inputs":["trades","filing dates","daily returns (for the return link)"],"limitations":"Late filing is a compliance failure, not by itself evidence of informed trading. Notification dates are self-reported and, on scanned filings, read by OCR.","doc_section":"sec:t7"},{"id":"T8","family":"overlap","name":"Committee overlap","question":"Does the member trade industries their committees oversee, and do those trades do better?","signature":"Members may trade on information about industries their committees oversee.","statistic":"Jurisdiction excess share versus S&P 500 and congressional peers; in-minus-out of jurisdiction signed CAR","formula":"\\Delta_8=s_J-s_J^{*},\\quad d_i\\mathrm{CAR}_i=\\gamma_0+\\gamma_1J_i+\\eta^\\top x_i+e_i","inference":"Identified set and bootstrap for the share; regression with sector and month fixed effects, date-clustered errors","decision_rule":"Fits if the in-jurisdiction return premium is significantly positive after adjustment. Does not fit if there is no jurisdiction tilt and the premium is bounded below the margin. Otherwise open question.","margin":{"value":0.005,"unit":"return","label":"0.5% per trade over 21 days"},"inputs":["trades","committee assignments","jurisdiction map","daily returns"],"limitations":"Committee assignments are current only; the jurisdiction map is at GICS sector or sub-industry level.","doc_section":"sec:t8"},{"id":"T9","family":"overlap","name":"Policy-topic overlap","question":"Did trades in stocks affected by the member's legislative work come before votes and move the right way?","signature":"Trading on knowledge of legislation before it is public.","statistic":"Signed CAR of trades in affected industries in the 30 days before the member's committee or floor actions, minus that of the member's other trades; pre-trade drift","formula":"\\theta_{9,i}=d_i\\sum_{\\tau=1}^{e-t_i+h}\\mathrm{AR}_{i,\\tau},\\quad \\psi_i=d_i\\sum_{\\tau=-q}^{-1}\\mathrm{AR}_{i,\\tau}","inference":"Date-clustered bootstrap of the difference (BCa); one-sided t test for pre-trade drift","decision_rule":"Fits if pre-event trades beat the member's other trades significantly after Holm and BH adjustment, earn a positive signed CAR, did not significantly follow public price moves in their own direction, and the same industries do not also beat the member's other trades in the month after events (placebo). Does not fit if the difference is bounded below the margin. Otherwise open question. Needs at least 10 trades in each group, on at least 10 distinct dates.","margin":{"value":0.005,"unit":"return","label":"0.5% per trade"},"inputs":["trades","legislative event calendar (committee markups and reports, chamber passage, enactment)","committee assignments","jurisdiction map","daily returns"],"limitations":"Events are bill milestones from GovInfo bill status records, attributed to every committee a bill was referred to; affected industries come from the committee jurisdiction map rather than the bill text, so large committees produce many events. Committee assignments are current only.","doc_section":"sec:t9"},{"id":"T10","family":"behavior","name":"Behavior after scrutiny","question":"Did trading change after press scrutiny or an ethics inquiry?","signature":"A member who trades on information has an incentive to keep control of trading after scrutiny.","statistic":"Rate ratio after versus before a known date; estimated change point","formula":"\\Lambda(\\tau)=2\\Big[Y_{\\le\\tau}\\ln\\tfrac{Y_{\\le\\tau}}{\\tau}+Y_{>\\tau}\\ln\\tfrac{Y_{>\\tau}}{T-\\tau}-Y\\ln\\tfrac YT\\Big]","inference":"Exact rate ratio; permutation test for the maximal likelihood ratio","decision_rule":"Does not fit if trading fell sharply after scrutiny or a blind trust was adopted. Open question if trading continued without a blind trust. Never fits on its own.","margin":null,"inputs":["trades","scrutiny dates","arrangements"],"limitations":"Requires a documented scrutiny date; not assessed otherwise.","doc_section":"sec:t10"},{"id":"T11","family":"peer","name":"Performance relative to Congress","question":"After accounting for noise, does the member's trading beat that of other members?","signature":"A member with an information edge stands out from the cross-section of congressional traders.","statistic":"Empirical Bayes posterior mean of the member's signed 21-day CAR; percentile among members","formula":"\\tilde\\theta_m=B_m\\mu+(1-B_m)\\hat\\theta_m,\\quad B_m=\\frac{s_m^2}{s_m^2+\\tau^2}","inference":"Normal-normal empirical Bayes shrinkage across all members with returns","decision_rule":"Fits if the 95% posterior lower bound is above zero. Does not fit if the 90% posterior upper bound is below the margin. Otherwise open question.","margin":{"value":0.005,"unit":"return","label":"0.5% per trade over 21 days"},"inputs":["T4 estimates for every member"],"limitations":"Shrinkage pulls members with few trades toward the congressional mean.","doc_section":"sec:t11"},{"id":"T12","family":"timing","name":"Anticipation of large stock moves","question":"When a traded stock made a large unexpected move soon after the trade, was the trade on the right side?","signature":"Buying shortly before a large jump or selling shortly before a large drop.","statistic":"Hit rate among trades followed within ten trading days by an abnormal move above three standard deviations","formula":"\\hat h_{12}=\\frac{1}{n}\\sum_{i:|\\mathrm{AR}_{i,\\tau^*}|>3\\sigma_i}\\mathbf 1\\{d_i\\,\\mathrm{AR}_{i,\\tau^*}>0\\}","inference":"Exact binomial test; Clopper-Pearson interval","decision_rule":"Fits if the hit rate is significantly above one half after adjustment. Does not fit if the upper bound is below 0.6. Otherwise open question.","margin":{"value":0.1,"unit":"share","label":"10 points above a coin flip"},"inputs":["trades","daily returns"],"limitations":"Large moves are rare; most members have few qualifying trades.","doc_section":"sec:t12"},{"id":"T13","family":"peer","name":"Co-trading with other members","question":"Does the member trade the same stock in the same direction as other members at the same time more than chance predicts?","signature":"Shared nonpublic information spreads through networks and shows up as clustered trades.","statistic":"Share of trades with another member trading the same issuer the same way within five business days; excess over the circular-shift null","formula":"c_m=\\frac1N\\sum_i\\mathbf 1\\{\\exists k\\ne m,\\ i'\\in k: j(i')=j(i),\\ d_{i'}=d_i,\\ |t_{i'}-t_i|\\le5\\}","inference":"Circular-shift permutation of the member's trade dates; signed CAR of co-traded trades","decision_rule":"Fits if co-trading is in excess and co-traded trades earn significantly positive signed returns. Open question if co-trading is in excess. Does not fit if co-trading is at or below its null mean.","margin":null,"inputs":["all members' trades","daily returns (optional)"],"limitations":"Public news makes many members trade the same stock on the same day; excess co-trading is a pattern, not a channel.","doc_section":"sec:t13"},{"id":"T14","family":"behavior","name":"Family-account trades","question":"Do trades in a spouse's or dependent's name do better than the member's own?","signature":"Routing informed trades through family accounts.","statistic":"Family-account signed CAR and its difference from self-account trades","formula":"\\Delta_{14}=\\theta_4^{\\text{family}}-\\theta_4^{\\text{self}}","inference":"Kolari-Pynnonen test on family trades; date-clustered bootstrap of the difference","decision_rule":"Fits if family-account returns are significantly positive and above self-account returns. Does not fit if family-account returns are bounded below the margin. Otherwise open question.","margin":{"value":0.005,"unit":"return","label":"0.5% per trade over 21 days"},"inputs":["trades with owner","daily returns"],"limitations":"Many spouses are professionals who trade on their own account.","doc_section":"sec:t14"},{"id":"T15","family":"behavior","name":"Short-swing round trips","question":"Does the member buy and sell the same stock within six months, and do those round trips beat the market?","signature":"Short-horizon round trips capture a temporary informational edge.","statistic":"Round-trip share; signed 21-day CAR of the opening trades (realized round-trip return shown for context)","formula":"\\hat\\theta_{15}=\\frac1K\\sum_k d_{\\text{open},k}\\,\\mathrm{CAR}_{\\text{open},k}(1,21)","inference":"Kolari-Pynnonen adjusted test on the opening trades","decision_rule":"Fits if opening trades earn significantly positive signed returns after adjustment. Does not fit if there are no round trips or the upper bound is below the margin. Otherwise open question.","margin":{"value":0.005,"unit":"return","label":"0.5% per trade over 21 days"},"inputs":["trades","daily returns"],"limitations":"Amounts are bands, so partial sales cannot be matched to purchases exactly.","doc_section":"sec:t15"},{"id":"T16","family":"behavior","name":"Options and leverage","question":"Does the member use options, and do option trades point the right way?","signature":"Informed traders favour options because leverage raises the payoff to private information.","statistic":"Option share of trades; signed CAR of the underlying after option trades","formula":"\\hat\\theta_{16}=\\frac1{N_o}\\sum_{i\\in\\text{options}} d_i\\,\\mathrm{CAR}_i(1,21)","inference":"Kolari-Pynnonen adjusted test on option trades","decision_rule":"Does not fit if no option trades. Fits if option trades earn significantly positive signed returns after adjustment. Otherwise open question.","margin":{"value":0.005,"unit":"return","label":"0.5% per trade over 21 days"},"inputs":["trades with asset type","daily returns of the underlying"],"limitations":"Disclosures rarely give strike, expiry or whether an option was written or bought, so direction is approximate.","doc_section":"sec:t16"}]