The Margin of Error

US midterm outcomes may turn on a handful of races, just as market returns are increasingly driven by a handful of stocks. This piece argues that the common risk is concentration, not polarization, and shows how investors can translate political scenarios into portfolio stress tests using volatility, rates, credit, inflation and factor shocks.

When political concentration meets market concentration

Forecasting is an exercise in confidence. Stress testing is an exercise in humility. 

Every two years, Americans vote to elect members of Congress. Midterm elections, held halfway through a presidential term, are often framed as a referendum on a nation growing ever more polarized. Congress is held together by razor-thin majorities, legislation increasingly passes or fails by a handful of votes, and the Supreme Court frequently divides along ideological lines. Judging by Washington alone, it is easy to conclude that America has become two irreconcilable tribes. For investors, however, the more interesting story is not polarization, but concentration. Polarization explains the noise, but concentration explains the risk.

The 2026 midterms are not one election, but actually hundreds taking place simultaneously. All 435 seats in the House of Representatives are up for election, alongside 35 Senate seats. Yet despite the scale of the exercise, control of Congress is expected to be decided by remarkably few contests. In the early 1990s, well over one hundred House districts were considered genuinely competitive. Today, the Cook Political Report (July 2026) classifies just 77 competitive districts (D+5 to R+5), and notes Democrats need 57 of those 77 for a majority. Of those 77, most forecasters identify only 30-40 as genuinely competitive races, with just 18 true toss-ups likely to determine control of the House. In the Senate, where Republicans currently hold only a narrow majority, control of the chamber may hinge on just two or three closely contested races out of only 35. Michigan and Georgia are both must-wins for Democrats if they are to gain control.

In other words, the legislative agenda of more than 343 million Americans may ultimately be determined by the outcome of a few congressional races. Politics, it turns out, is a game of margins. Financial markets have become one too, with the AI trade concentrating index risk in fewer and fewer names.

That is where the political story becomes an investment consideration. Narrow political margins matter because they compress the range of possible policy outcomes into a small number of races. Narrow market leadership matters because it compresses portfolio outcomes into a small number of stocks. In both cases, the issue is the amount of risk now carried by a shrinking number of decision points.

Just as control of Congress may depend on a handful of districts and Senate races, the direction of today's equity indices increasingly depends on a handful of companies. An unprecedented concentration of capital in a single investment theme, Artificial Intelligence, and an even smaller group of hyperscalers has left the performance, and increasingly the risk, of entire indices dependent on remarkably few stocks.

Laws are decided by narrow majorities, elections by narrow margins, and markets by narrow leadership. What ties these seemingly unrelated phenomena together isn't polarization, but instead, fragility.

A system becomes fragile when its outcomes depend on an ever-smaller number of participants or decisions, such as a few of House districts or a handful of hyperscalers. The narrower the leadership, the larger the consequence of being wrong.

That is precisely what makes this year's midterms so difficult to handicap. Investors are naturally tempted to forecast who will win. But when outcomes are this concentrated, the more useful question is not who wins. It is how much portfolio risk is being attached to each possible result.

Rather than asking which party is most likely to control Congress which is already amply covered by polls and prediction markets, we consider four plausible scenarios: three based on electoral outcomes, and one based on what happens if the Executive and Congress actively work against each other. The first three scenarios still assume a recognizably normal political process. The fourth suggests an alternative starting point.

The Trump 2.0 administration has repeatedly tested the limits of executive authority, relied on executive orders, and pushed policy disputes into the courts. Investors should therefore not assume that established institutional constraints will limit the outcome. In this scenario, the stress is both political and institutional. There is no reliable precedent, no stable playbook and no obvious limiting principle. Investors are left pricing a confidence shock in which the unexpected, the extreme and the previously unimaginable become part of the scenario set. 

Four political scenarios

The scenarios below are a framework for translating narrow political margins into portfolio-relevant policy and market risks. 

This is especially true if American forces remain involved in a war in the Gulf, where the chain of command itself may become contested: the President remains Commander in Chief, but only Congress can declare war.

The objective here is not to predict which of these futures will materialize. It is to understand how each could alter the path of monetary policy, fiscal policy, inflation, corporate earnings and, ultimately, portfolio risk. Forecasting is an exercise in confidence. Stress testing is an exercise in humility. In a fragile, concentrated market, the winning portfolio is not the one built for the future everyone expects. It is the one built to survive the future that investors failed to price.

Stress testing the midterms

Forecasting elections is difficult, but forecasting how markets will respond to elections is harder still. The purpose of the scenarios below is therefore not to predict which party will control Congress after the 2026 midterms. It is to identify the economic and market transmission channels that each plausible outcome could trigger, and to understand how those channels might affect portfolio risk.

The scenarios should be viewed as a stress-testing suite rather than a set of forecasts. Each combines a political outcome with a plausible fiscal, monetary, inflation and market response based on historical precedent, current policy debates and published research from election forecasters, policy institutions and financial firms. The objective is to determine whether a portfolio is unintentionally exposed to fiscal expansion, tariff inflation, policy paralysis, institutional uncertainty, or some combination of all four.

In Axioma Risk, the objective is to avoid a generic election sell-off and instead model the transmission channel. Fiscal policy changes the Treasury curve. Tariffs change inflation and margins outlook. Policy conflict changes volatility and credit spreads. Institutional uncertainty changes the risk premium investors demand for US assets. In a transitive stress test, the more useful approach is therefore to shock volatility, rates, credit spreads, inflation expectations and currencies directly, choose the appropriate regime to model correlations, then allow the model to infer the associated equity and factor impact.

Implementation principle: do not force a small S&P 500 move into a large-volatility narrative. If the scenario is about uncertainty, shock the VIX first and let the equity loss emerge through the model.

Axioma stress suite summary

The table below replaces direct broad-equity shocks with a VIX-centric framework. Equity-market losses are not removed from the scenario. They are allowed to emerge from the correlation structure once the primary shocks are applied.

Choosing the correlation regime

The choice of calibration period may be as important as the size of the shock itself. A transitive stress test requires two assumptions. The first is which risk factors are shocked and by how much. The second is the historical regime used to estimate how those shocks propagate through the rest of the portfolio. Axioma Risk can use a default recent-history calibration, but these scenarios are explicitly about possible breaks from recent history. If the future stress is not expected to behave like the last year, the correlation window should not be chosen mechanically.

For routine market fluctuations, the trailing one-year window can be a reasonable starting point. For election and institutional scenarios, it may mix unrelated shocks and dilute the investor behavior being modeled. The current or past-year window may include tail events such as Epic Fury or Liberation Day. Those events may be exactly right for the tariff and constitutional-risk variants, but too noisy for scenarios that are closer to orderly rotation or divided-government gridlock. The calibration regime should therefore be treated as part of the initial scenario design.

Practical rule: use at least a 60-day historical window, but choose a period whose investor behavior resembles the scenario: flight to safety, reflation, sector rotation, funding stress or tariff/inflation repricing.

Democratic sweep

Suggested calibration period: November 2022 through March 2023.

This period combines divided government, moderating inflation, slower fiscal expectations, and relatively orderly market behavior. It is useful if the expected response is sector rotation and lower rates rather than wholesale de-risking.

Republican retention

Suggested calibration period: 2017-2018 reflation period.

This period captures tax reform, fiscal expansion, deregulation, rising Treasury yields, and reflation expectations. It is useful when the expected response is risk-taking with rotation toward cyclicals, financials, and domestic exposure.

Split Congress

Suggested calibration period: 2018 government shutdown – the longest in US history (12-22-2018 to 01-25-2019).

This period provides a template for political confrontation without a broader constitutional crisis. It is useful when the expected response is temporary volatility, modest spread widening, and flight-to-quality behavior while investors remain largely invested.

Constitutional Crisis A: institutional confidence shock

Suggested calibration period: March 2023 regional banking crisis.

This period captures confidence shock, liquidity preference, rapid volatility repricing and pressure on financials, and funding-sensitive exposures. It is useful when the expected response is flight to quality.

Constitutional Crisis B: executive conflict and legal challenges

Suggested calibration period: Liberation Day tariff shock period.

This period captures policy uncertainty, tariff repricing, inflation concerns and regulatory unpredictability. It is useful when the expected response is sector rotation, higher volatility and changing inflation expectations rather than a pure financial panic.

Constitutional Crisis C: combined stress

Suggested calibration period: Custom blended regime.

No historical period provides a perfect analogue for a simultaneous confidence shock, executive-legislative conflict, tariff escalation, and institutional uncertainty. Use March 2023-style confidence correlations, then overlay Liberation Day-style tariff, breakeven, USD and long-end term-premium shocks manually.

Monitoring the scenarios over time

The useful exercise is not simply running the stress tests once but instead monitoring whether reality is moving closer to one scenario and further from another. The monitoring framework should combine political indicators, policy indicators and market indicators, because the scenarios will rarely announce themselves cleanly.

For a Democratic sweep, monitor House and Senate race ratings, special-election results, presidential approval, the congressional generic ballot and signs of stronger oversight or budget confrontation. Market confirmation would include lower Treasury yields, wider bank spreads, weaker cyclicals and relative strength in quality defensives.

For Republican retention, monitor tax-policy negotiations, deficit projections, Treasury issuance forecasts, tariff announcements and inflation expectations. Market confirmation would include rising BEI yields, a steeper Treasury curve, stronger domestic cyclicals and small-cap outperformance, with more mixed signals from import-heavy consumer names.

For Split Congress, monitor the probability of House control shifting while Senate control remains Republican. Appropriations deadlines, debt-ceiling negotiations, shutdown risk, congressional investigations and litigation over executive authority become more important than legislation itself. Market confirmation would include episodic volatility spikes, wider funding spreads and higher policy-event risk without a wholesale change in long-term growth expectations.

For the Constitutional Crisis variants, monitor institutional behavior more than polling. The most useful indicators are executive orders, emergency declarations, Supreme Court decisions, congressional subpoenas, federal-agency legal challenges, tariff litigation, government shutdown risk, debt-ceiling risk, policy-uncertainty indices, consumer confidence and business-confidence surveys. These indicators can deteriorate even if the election result itself is fully anticipated.

Cross-scenario factor exposures

Across all scenarios, the recurring vulnerabilities are factor concentrations: high market beta, small size, leverage, weak profitability, residual volatility, liquidity risk, crowded momentum and exposure to long-duration cash flows. The recurring defensive exposures are high quality, high profitability, low leverage, low volatility, low liquidity and high earnings visibility. The scenarios differ mainly in the rates and inflation sign. Democratic sweep and Split Congress lean toward lower growth and lower front-end yields. Republican retention leans toward reflation and higher yields. Constitutional Crisis B and C introduce the hardest combination: weaker growth, wider spreads, higher inflation compensation and weaker USD.

How to read the results

The most useful result is the attribution. Which positions explain the loss? Which sectors amplify it? Which style factors are doing the damage? Does the portfolio lose money because it owns a political loser or because it owns long-duration growth, leverage, import-cost exposure, crowded momentum and weak liquidity under another name?

The most informative output may not be the portfolio’s stressed return, but its sensitivity to volatility. Many of the scenarios described here are likely to manifest first through higher uncertainty, wider risk premia and elevated volatility rather than through an immediate collapse in equity prices. Investors may therefore prefer to shock the VIX, credit spreads, Treasury yields, inflation expectations and currencies directly, allowing Axioma Risk’s transitive framework to determine the implied impact on equities and other portfolio exposures.

Choosing the right context for the shocks is important because it determines how investors are assumed to respond when uncertainty rises. A VIX shock calibrated on a quiet rotation regime will not produce the same result as the same VIX shock calibrated on a banking-crisis regime.

Taken together, the suite should be read in three layers: the electoral scenarios test fiscal direction, regulatory risk and gridlock. The constitutional-crisis variants test confidence, legal authority and institutional risk. The correlation-regime framework determines whether the stress behaves like orderly rebalancing, flight to quality, reflation, tariff repricing or a blended tail. Investors do not need to decide who wins the midterms before they run the stress. They need to know how much of their portfolio depends on fiscal expansion continuing, tariffs not becoming inflationary, courts remaining predictable, and investors retaining confidence in the US policy framework.

Elections are ultimately decided by a small number of districts, just as market returns increasingly depend on a small number of securities. The purpose of stress testing is therefore not prediction, but preparation. By mapping political outcomes into fiscal, monetary, inflation, credit, volatility and factor shocks, investors can identify where portfolios are vulnerable before those vulnerabilities become visible in market prices. In a concentrated market, the most dangerous risk is often not the scenario investors expect, but the one they never considered.

Research inputs

Election scenario framing: Cook Political Report race ratings and comparable nonpartisan election forecasters, including House and Senate competitiveness work used to frame the 2026 control scenarios.

Fiscal and macro-policy framing: Congressional Budget Office, Yale Budget Lab, Tax Foundation and Brookings analysis of the fiscal, debt, Treasury-yield and long-run growth implications of Trump-era tax and spending legislation.

Tariff and inflation framing: Yale Budget Lab tariff tracker, Federal Reserve and regional Federal Reserve analysis, and published research from major financial institutions including Goldman Sachs and Morgan Stanley on tariff pass-through, margins, inflation expectations and Federal Reserve reaction functions.

Institutional-risk framing: legal and policy research on executive authority, emergency powers, tariff litigation and election administration from Congressional Research Service, Brennan Center, Protect Democracy and related democracy-policy institutions.