Korea-US Interest Rate Comovement: Why Does Korea Follow the US?

Whenever news breaks that the US has raised interest rates, Korean government bond yields often jump in response.

The primary cause of this Korea-US interest rate comovement is not the US interest rate itself, but the inflation shock that hit both countries simultaneously (41.0% contribution), and this shock is transmitted via market expectations that “the Bank of Korea will eventually follow suit."

This conclusion comes from the report, “Analysis of the Comovement of Long-term Interest Rates in Korea and the US,” published in September by the Bank of Korea’s Institute for Monetary & Economic Research. While news outlets only focused on the 41% figure, the 85-page original document contains much more interesting insights. The research team analyzed Korean government bond yields and US Treasury yields from January 2001 to July 2024 using three methods: daily interest rate fluctuation analysis, a term structure model (GDTSM) that decomposes long-term rates by component, and an event study looking at the days surrounding Federal Reserve announcements.

When Was the Korea-US Interest Rate Comovement Strongest?

It happened twice: immediately after the 2008 Global Financial Crisis and following the 2021 inflation outbreak. The covariance of daily fluctuations in 10-year Korea-US bond yields spiked in December 2008, remained quiet for a long time, and then surged again after 2021. During these periods, the volatility of interest rates in both countries also increased in tandem. The rates in both countries began to resemble each other not only in direction but also in the magnitude of their swings.

This isn’t just a phenomenon between Korea and the US. The report compared long-term interest rates across the US, Germany, the UK, Canada, and Korea; the rates of all five countries, which had been falling together for nearly 20 years, shifted direction and began to rise simultaneously in 2021. In contrast, short-term rates only moved together during crises and otherwise diverged according to each country’s monetary policy. This indicates that long-term interest rates are tied much more tightly.

Exchange rates are also part of the story. Recently, when Korean government bond yields rise—meaning bond prices fall—there has been a tendency for the Korean Won to weaken as well. For foreigners who bought Korean government bonds with dollars, this means bond losses are compounded by exchange rate losses, effectively doubling the volatility.

One thing to note: the phrase “comovement has become structurally stronger” is frequently seen these days, but the structural change tests conducted in the report yielded mixed results (the QLR test and CUSUM test gave different answers). While it is clear that comovement strengthened during crises, there is no evidence yet to conclude that the landscape has changed permanently.

The #1 Factor in Korea-US Long-Term Rate Comovement Was Not US Treasury Yields

Using a term structure model with a 60-month forecasting horizon, the researchers decomposed the factors behind the comovement of Korea-US long-term rates; the figures were reported by Money Today. Inflation shock accounted for 41.0%, direct US long-term rate shock for 22.7%, Fed monetary policy (shadow rate) shock for 18.3%, and US economic (output) shock for 18.0%.

What was somewhat unexpected here was the second-place ranking. The item that most directly represented the claim that “Korean rates follow US rates” is the direct US long-term rate shock, which at 22.7% was only slightly more than half of the inflation shock. Rather than Korea simply copying US rates, it is closer to saying both countries are being hit by the same winds of inflation.

Let’s clarify the terminology. Within the model, this shock is labeled “US expected inflation shock,” using the University of Michigan’s inflation expectations as the US indicator. Because the US economy accounts for such a large share of the global economy, the report interprets this as a global inflation shock. That is where the “global inflation” in the article headlines originated.

Counterfactual analysis shows just how significant this was. By removing the inflation shock from the model since March 2021, when US inflation began to exceed the 2% target, US inflation expectations stopped at around 3.5% instead of the actual peak of 5.4%. Korean inflation expectations would have been up to 1.1%p lower, and the Korean 10-year government bond yield would have been up to 1.5%p lower than it actually was.

Decomposing Long-Term Rates Reveals the Channel of Comovement

The real core of this report is how these shocks were transmitted. To understand that, one must look at what constitutes long-term interest rates.

There are two ways to invest for 10 years: buy a 10-year bond once, or buy a 3-month bond and roll it over 40 times as it matures. Since the returns on both methods should be roughly similar, the 10-year yield includes the average expected short-term rate for the next 40 periods. This is the ‘expectations component.’ However, locking up money for 10 years carries the risk that rates might jump or inflation might surge in the meantime. The premium paid for that risk is the ‘term premium.’

Ultimately, it is the central bank that determines the path of short-term rates. Therefore, the channel through which the expectations component is transmitted is called the ‘policy path,’ and the channel through which the term premium is transmitted is the ‘risk compensation path.’ If a US-originated shock pushed up Korean long-term rates, it could be one of two things: either the market calculated that “the BOK will raise the base rate too,” or investors viewed Korean bonds as riskier and demanded a higher premium.

The results were heavily skewed in one direction. Of the 0.385 covariance created by the inflation shock, 0.361 was the expectations component, while the term premium was only 0.023. For the economic shock (0.223), Fed policy shock (0.261), and US long-term rate shock (0.320), the expectations component also held the largest share, and in these three shocks, the term premium side was actually negative.

This is interesting because previous research analyzing interest rate comovement among developed countries (Jotikasthira et al., 2015) reached the opposite conclusion. Among the US, UK, and Germany, comovement was driven primarily through the risk compensation path of long-term bonds, i.e., the term premium. In Korea’s case, rather than foreign investors being spooked, the calculation that the BOK would follow the same path as the Fed was reflected in long-term rates first.

To What Extent Did Korea-US Rates Comove When the Fed Was Easing?

Beyond the model, they also verified this using actual dates. The research team selected 11 days: 9 days when Quantitative Easing (QE1-QE3) was announced and 2 days when large-scale asset purchases were announced during COVID-19. They isolated the 8 days where the US 10-year yield fell by more than 10bp over a 5-day cumulative period, and found that the Korean 10-year yield fell in tandem on 7 of those occasions.

Korea-US Interest Rate Comovement: 5-Day Cumulative Change Averaged Across 8 QE Announcement Days

Item	Average Change
US 10-Year Yield	-34.2bp
Korean 10-Year Yield	-15.0bp
Korean 2-Year Yield	-14.0bp
KRW/USD Exchange Rate	-1.3% (Won Appreciation)

What is worth noting is the Korean 2-year yield. Even though the Fed was buying US long-term bonds, Korean short-term rates also fell by 14bp. This is evidence that the Fed’s announcement was read as a signal that “the BOK won’t be able to raise rates for a while either.” Decomposing the Korean 10-year yield decline, the expectations component averaged 10.9bp and the term premium 7bp. The report concludes that while the signal component was slightly stronger, both channels functioned significantly.

In terms of cumulative effects, the Fed’s unconventional monetary policy pushed down the Korean 10-year yield by up to 0.6%p and the 3-month yield by up to 0.7%p. There were also events with the opposite direction. During the 2013 ’taper tantrum,’ when Chairman Bernanke signaled the tapering of QE, Korea-US short and long-term rates rose together from April 30 to September 5. During this period, the correlation coefficient of interest rate fluctuations between the two countries was 0.25 for the 2-year and 0.40 for the 10-year, with the rise in the Korean 10-year yield split evenly between the expectations component and the term premium.

Applying the September 2026 Fed Hike to the Korea-US Comovement Formula

There was an event a few days before the report was released. On September 16, the Fed raised the base rate by 0.25%p to 3.75-4.00% per year. According to EBN reports, this was the first hike in 3 years and 2 months since July 2023, driven by rising energy costs due to Middle East conflicts and persistent inflation. On the same day, the US 10-year yield closed at 5.003%, marking its highest closing price since July 2007, and 16 out of 18 Fed officials expected at least one more hike within the year.

Korea reacted immediately. As the Korea-US base rate gap widened back to 1%p (upper bound), pressure on the BOK to hasten its own rate hike grew. The domestic AA- rated 5-year corporate bond yield rose from 3.79% at the beginning of the year to 4.89% on September 21.

Superimposing the report’s framework onto this, two things align: the starting point of this shock is oil-driven inflation—the very type the report identified as the #1 priority—and the rhetoric pulling up domestic rates is the expectation that “the BOK must raise rates too.” Of course, the report’s sample ended in July 2024, so there is no evidence that the 41% ratio applies exactly the same way this time. We can only say that the direction and the channels overlap.

Why the BOK’s Communication Can Reduce Korea-US Comovement

Since external common shocks like inflation are the fundamental cause, some degree of comovement is inevitable. However, if the channel through which that shock is transmitted to Korea is policy rate expectations, then comovement can be reduced to some extent by managing those expectations through market communication. That is the implication of the report.

There is a scene in the report that supports this. In 2023, the US continued to raise its policy rate to 5.25-5.50%, but Korea stopped at 3.5% due to economic concerns. Consequently, the covariance of Korea-US 2-year yields dropped noticeably. As the paths of the two central banks actually diverged and the market believed it, the comovement in the short term was loosened. However, the 10-year correlation coefficient remained at a high level during the same period. That year, the BOK’s decision loosened the ties for short-term rates but could not break the ties for long-term rates, which were linked through inflation.

As a follow-up, the research team proposed indicators to monitor comovement in real-time and provided a pilot calculation in the appendix: a variance ratio measuring the share of US long-term rate volatility within Korean long-term rate volatility.