Is it too late to copy a congressional trade?
By the time you see a congressional trade, it is between a few days and 45 days old. The obvious worry is that the move has already happened and you are buying the top.
We can measure that directly, and the answer is more interesting than yes or no: the lag is real but it is not what breaks the strategy. Something else does.
What the lag actually costs
If disclosed buys spiked immediately and then flattened, the delay would be fatal — all the return would land in the window you cannot see.
That is not the shape. Across 4,791 disclosed buys:
| After the purchase date | Average return |
|---|---|
| 1 week | +0.29% |
| 1 month | +1.60% |
| 3 months | +5.15% |
Most of the movement happens over months, not days. Roughly a third of the three-month return has accrued by the one-month mark. A follower entering at, say, day 30 misses a meaningful slice — but nothing like all of it.
On the lag question alone, copying looks survivable.
Why that does not rescue it
Now measure the same trades against the sector each stock belongs to, over the identical windows:
| After the purchase date | vs. sector benchmark |
|---|---|
| 1 week | −0.16% |
| 1 month | −0.58% |
| 3 months | −0.32% |
There is very little edge being lost to the delay, because there is very little edge in the first place. The +5.15% was the market doing what the market did over this period; the politician's contribution, net of sector, is approximately nothing.
This reframes the whole question. People ask "am I too late?" assuming there is something to be late for. The full analysis.
The costs a follower pays that the member does not
Even taking the raw figures at face value, a copier's return is worse than the member's for reasons that have nothing to do with skill:
You buy at a different price. Filings never disclose an execution price. You are entering at whatever the stock costs on the day you read the filing, which may be weeks of price movement away from what the member paid.
You cannot size it. Amounts are disclosed as ranges — $1,001–$15,000 and
similar. You know roughly what band, never the figure, so you cannot replicate
the weighting.
You do not see the exits. Sales are disclosed on the same lag. A member who sold three weeks ago is still, as far as your data shows, holding.
You pay to trade. Spreads and commissions apply to every position you open and close. The member's disclosed return carries none of them.
Some of it was never a decision. A meaningful share of disclosed trades are made by a spouse or dependent child, filed under the member's name.
What about the ETFs?
Two funds run this strategy as a product — NANC and KRUZ, split by party. They face the same lag as anyone else, because they read the same filings. What they actually hold.
Their existence is useful evidence in itself: a professionally managed, mechanically executed version of the strategy, running in public, with fees disclosed. Any claim that copying Congress is a reliable edge has to explain why that edge does not show up straightforwardly there.
The honest answer
Not too late, and not worth being early for.
The delay costs you something, but it is not the reason this does not work. The reason is that congressional buys, measured against what their own sectors did, do not outperform — and a strategy with no edge does not become one by being executed faster.
If you want to check that rather than take it from us, the per-trade returns and benchmark comparisons are in the published dataset, and the method is documented.
Nothing here is investment advice. See Terms.