Why 83% × 44% = 44% in Kalshi’s Midterm Markets

Kalshi’s midterm markets offer a useful lesson in correlation, and in why an apparent pricing inconsistency is not automatically an arbitrage opportunity.

On July 24, Kalshi traders gave Democrats an 83% chance of winning the House and a 44% chance of winning the Senate. Yet the platform’s balance-of-power market also priced Democrats winning both chambers at 44%.

Most people would multiply the two standalone probabilities:

83% × 44% = 36.5%

That produces an implied 36.5% probability of Democrats winning both chambers, nearly eight percentage points below the price in Kalshi’s combined market.

But that calculation assumes the House and Senate results are independent events. In the real world however nothing is independent of each other.

The 36.5% Trap

The House and Senate are decided by the same national electorate, under the same presidential approval numbers, economic conditions, turnout patterns and campaign environment.

A strong Democratic election night makes both outcomes more likely. A weak Democratic night pushes both in the opposite direction.

The Senate is the harder leg for Democrats under the current market prices. If the national environment becomes strong enough for Democrats to overcome the Senate map, it would also make a Democratic House victory highly likely.

That is why the probability of Democrats winning both chambers can trade at roughly the same level as their probability of winning the Senate.

The combined contract cannot be worth more than either individual component, but it can approach the price of the less likely component when one outcome is highly likely to accompany the other.

Where the Missing Probability Goes

Correlation does not simply increase the probability of a Democratic sweep. It redistributes probability across all four possible congressional outcomes.

If the House and Senate were independent, the probabilities would look like this:

Vs…

Kalshi’s balance-of-power market instead displays the outcomes at approximately:

The prediction market assigns more probability to the two same-party outcomes and less probability to divided control.

The most revealing difference may be the scenario in which Republicans win the House while Democrats win the Senate. An independence model puts that outcome at 7.5%, while Kalshi displays it at just 1.7%.

Using the combined market as a rough probability map, Democratic Senate control comes almost entirely through a Democratic House. Dividing 44 by the combined 45.7 points assigned to the two Democratic Senate outcomes produces an implied conditional probability of approximately 96.3%.

The displayed contracts do not add to exactly 100% because of spreads, rounding and differences between recent trades. The calculation should therefore be treated as an approximation rather than a perfectly normalized probability.

What the Market Is Really Pricing

Viewed separately, the House and Senate contracts are forecasts. Viewed together, they form a model of the election.

The market’s more substantive claim is not that Democrats are favored to win the House and face a more difficult path in the Senate. It is that the set of political conditions capable of producing a Democratic Senate is narrow enough, and favorable enough, that a Republican House becomes exceedingly unlikely within it. The combined market is therefore pricing a theory about how national political forces move through the electoral system.

For traders, that is the more interesting level of analysis. The obvious question is whether each probability looks reasonable on its own. The harder question is whether the relationships among those probabilities are reasonable. A market can price every individual event plausibly while still mispricing the dependence between them. If there is an edge here, it will not come from multiplying 83% by 44% correctly. It will come from deciding whether traders are too confident that the two outcomes must rise and fall together.