Why are interest rates high and what would make them fall?

Rates are high against 2019, below the tightening peak

The latest policy action was a quarter-point increase on September 16, taking the federal-funds target to 3.75%–4.00%. Comparing target-range midpoints puts policy 225 basis points above December 2019 and 150 basis points below the September 2023 setting. “High” is a useful description only with a benchmark attached. [9] [10] [3]

The latest policy setting sits between late-2019 rates and the cited tightening-era peak.

Sources: dated FOMC statements. Midpoints are the arithmetic average of each target range’s bounds. [9] [10] [3]

We are skeptical that the old low-rate environment should be treated as the automatic destination. Fed participants’ median longer-run nominal-rate projection rose from 2.5% in December 2019 to 3.2% in September 2026. Those conditional judgments suggest a higher resting point, although the neutral rate itself is unobservable and subject to reassessment. [11] [12] [13]

The current policy midpoint exceeds that latest longer-run projection by 67.5 basis points, calculated as 3.875% minus 3.2%. The path matters as much as the destination: participants’ September median projection for the end of 2027 was still 4.1%. A gap to estimated neutral offers no calendar for removing restraint. [3] [12]

Inflation remains the clearest reason for restraint

The Fed explicitly connected September’s increase to elevated inflation and its 2% goal. August PCE inflation, subsequently released by the BEA, was 3.4% over the year; excluding food and energy, it was 3.0%. For perspective, headline PCE inflation was 1.5% in November 2019. The inflation backdrop gives the Fed a clear reason to keep policy above its late-2019 setting. [3] [14] [15]

The difficult question is how much persistence to expect from supply-driven price increases. Higher rates can restrain spending and reduce the risk that a shock spreads, but they cannot directly remove the disruption. Energy CPI rose 16.3% over the year to August—a sharp increase in that component’s prices. Yet Jefferson said on October 1 that broader, persistent spillovers were not yet apparent. We give that qualification substantial weight. [16] [17]

Our conclusion is therefore conditional. Above-target underlying inflation argues for caution; the absence of broad spillovers leaves room for easing if inflation improves. We reject a confident forecast of further hikes. [14] [17]

Mortgage rates follow the cost of long-term money

That explains why the Fed has kept policy elevated. It leaves a separate question: why does a mortgage cost so much more?

A long-term yield incorporates the expected future path of short rates and compensation for holding a long bond. The San Francisco Fed’s framework separates those two elements, with its model residual included in term compensation. The mechanism allows long yields to fall in anticipation of easing—or remain high after a cut if expectations and risk compensation offset it. [5]

Mortgages add an option in the borrower’s favor. When rates fall, borrowers can refinance, returning investors’ money just when reinvestment becomes less attractive. Expectations, interest-rate volatility, refinancing costs and intermediation costs therefore enter mortgage pricing. A 10-year Treasury is a useful, imperfect benchmark for those cash flows. [6]

The selected mortgage and Treasury observations rose by similar amounts, leaving their simple difference near two percentage points.

Sources: Freddie Mac, FRED and Federal Reserve H.15. [1] [2] [7] [8]

Note: Treasury observations use the mortgage survey’s publication dates. Survey windows differ, and Freddie Mac changed methodology in 2022. [18] The displayed differences are mortgage rates minus Treasury yields.

The arithmetic is straightforward: the mortgage increase of 363 basis points equals the Treasury increase of 370 basis points plus a seven-basis-point decline in their difference. That difference was 2.11 percentage points in the selected 2019 observation and 2.04 points in 2026. Given the survey and benchmark limitations, the useful conclusion is its broad stability. A substantially wider mortgage spread is unnecessary to account for the increase across these observations. [1] [2] [7] [8] [18]

We are also skeptical of assigning government borrowing the dominant role in current yields: the evidence available here cannot identify its contribution. Our conclusion rests on the separation between expected rates, term compensation and mortgage pricing, without claiming a numerical allocation among their underlying causes. [5] [6]

Weak hiring is the strongest case for earlier cuts

The strongest challenge to prolonged high rates arrived in the employment data. September payrolls increased just 29,000, against a preceding-12-month average of 45,000 per month. We think that deserves substantial weight when assessing the September policy stance. [19]

The evidence is mixed rather than uniformly weak. Unemployment was 4.2%; earlier activity data showed second-quarter real GDP growth of 2.2% at an annualized rate and a 0.6% monthly increase in real consumer spending in August. These readings cover different periods, leaving the current direction of demand less settled than the weak hiring number alone suggests. [19] [20] [14]

There is a clear precedent for acting before every inflation measure reaches the target. In September 2024, the Fed cut by half a percentage point as Powell explained that inflation risks had diminished and employment risks had increased. The lesson is about the balance of risks: enough employment deterioration can bring easing forward. [4]

Our preferred route to sustainably lower rates is repeated improvement in underlying inflation toward 2%. Persistent weak hiring accompanied by rising unemployment and weakening spending would create a stronger employment-driven case for cuts. Today’s combination of 3.0% core PCE inflation and mixed activity data supports neither a firm cut date nor a confident further-hikes call. [3] [14] [19] [20]

Borrowers and businesses receive relief unevenly

The transmission depends on the contract. An existing fixed-rate mortgage retains its rate while the borrower keeps the loan. Index-linked borrowing depends on its benchmark and reset terms. A lower policy target therefore reaches different balance sheets at different speeds. [18] [21]

For a new borrower, even a partial decline matters. On the same hypothetical $300,000, 30-year mortgage, reducing the rate by one percentage point to 6.28% lowers monthly principal and interest to about $1,853, a saving of roughly $200. These payments use the standard monthly amortization formula, holding principal and term constant and excluding taxes, insurance, fees and refinancing costs. [1] [2]

Lennar shows how affordability pressure reaches company earnings. Fiscal third-quarter home-sales revenue fell to $7.7 billion in 2026 from $8.2 billion a year earlier. Home-sales gross margin declined from 17.5% to 15.8%, a 170-basis-point contraction calculated by subtraction. Management described incentives and base-price adjustments needed to sustain volume. Those disclosures connect affordability to commercial decisions, while leaving other influences on earnings in the picture. [22]

Outcomes still vary across companies. Toll Brothers reported 5% growth in net signed contracts in its separate fiscal third quarter of 2026. At JPMorgan, management attributed part of consumer-segment revenue growth to higher card net interest income, largely associated with greater revolving balances. We would not generalize either experience into a rule that expensive money hurts every builder equally or raises every bank’s profits. [23] [24]