Seasonality in the Treasury Market

[WITH CODE] A dive into seasonality in the U.S. treasury market

Reviewed and updated 20 July 2026

Hello!

Welcome back to another paid-only post. Today, I want to dive deep into the seasonality of Treasuries from an academic paper standpoint.

All types of financial markets and asset classes exhibit seasonality. We often hear about the “Sell in May and Go Away” effect in equities, but the Treasury market has unique characteristics that affect its behavior, from auction cycles to quarterly refunding announcements.

In this post, I’ll highlight a collection of academic papers that explore these seasonal effects, along with my own analysis and commentary. I’ll also point out some common pitfalls that come up in the interpretation of this kind of research.

Let’s get into it.

---

Introduction

Before we get into the academic papers, I want to highlight some nuances in the bond market. Unlike with equities, the value of a bond can be assessed in multiple ways - price, yield, spread, etc.

It’s important to understand how the value of the Treasury is being measured before drawing conclusions. For example, this analysis by Barchart shows the returns of the 10Y Treasury Note. However, the analysis shows the percent change in yield. Therefore, the “gains” in the chart are actually times when the bond underperforms. For example, it shows the yearly return for the 10Y Note in 2022 was 156%. Yet during this time, rates sold off (yields up, prices down) as the Fed began its hiking cycle.

Additionally, a percentage change in yield should not be confused with a bond return, especially when P&L for many interest-rate positions is more directly connected to DV01 and outright basis-point change. These concepts may be foreign to some of you, and I can follow up in the paid-only chat after this post goes live.

In this post, I will cover academic findings on seasonality in the treasury market and showcase some elementary analysis through python.

---

Academic Findings

Mark Kamstra, Lisa Kramer, and Maurice Levi uncovered an interesting pattern. From their detailed study of U.S. Treasury returns, they found that monthly returns are approximately 80 basis points higher in October than in April. They argue that seasonal mood variation influences risk appetite, driving asset class preferences and term premia across the yield curve. In other words, investors may demand more compensation for interest rate risk in the spring and summer, when optimism (and risk-taking) is higher, leading to higher yields. Their model accounts for over 60% of this seasonal variation, and the results are statistically significant.

During the last trading days of the month, insurers and index-tracking funds tend to increase purchases of Treasury securities (especially those being added to benchmark indices). This predictable demand spike leads to a "richening" of prices: buying just before month-end and selling right after can yield significant excess returns. This effect is widespread across various Treasury maturities and is consistent from 1990 through 2018. The “turn-of-the-month” behavior is likely driven by index rebalancing mechanics and has spillover effects into futures and swaps markets too.

A complementary perspective comes from New York Fed analysis, which finds that trading volume since 2020 was roughly 46% higher on the last trading day of the month, or 51% higher when Decembers are excluded. The pattern appears across security types, including notes issued at month-end and the 10-year note issued mid-month. The authors suggest that the growth of passive, index-driven fund flows concentrated at month-end may help explain it.

Liu and colleagues (2022) charted a pattern across multiple tenors (2Y through 10Y) using rank-based analysis. Unlike traditional measures like yield changes or levels, rank-ordering each month within a year reveals a consistent pattern: yields rank higher from March through August and lower between September and February, regardless of maturity. This suggests a half-yearly cycle embedded in the Treasury curve, and aligns with the results from Kamstra, Kramer, and Levi.

Shifting gears to risk interplay, Rubin and Ruzzi (2020) find that an option-implied measure of left-tail equity risk predicts one-month-ahead Treasury excess returns both in and out of sample. They also document contemporaneous Treasury price increases and flows from equities into bonds when perceived tail risk is higher, consistent with a classic flight-to-safety dynamic.

---

Python Analysis

The code-and-output PDF is included in the research companion below.

In this analysis, I pulled Treasury yield data from FRED with an API. The companion includes the reference summary tables, so API credentials are not needed to reproduce the reference results. I used data from 15 February 1977 through 20 August 2025 and computed summary statistics for each security (2Y, 5Y, 10Y, and 30Y), using the 12,124 daily observations where all four series are present.

For example, we can see the average (mean) monthly change in yield, measured in basis points, for each Treasury security. Remember that, holding other inputs fixed, a decrease in yield means that the Treasury rose in value (price); these yield changes are not total returns or a trading-strategy P&L.