---
ticker: "KRXMP"
title: "South Korea Exchange Market Pressure Index"
unit: ""
frequency: "Monthly"
source: "Goldberg and Krogstrup (2023)"
release: "Updated Monthly"
category: "Exchange Rates"
country: "KR"
language: "en"
canonical: "https://kred.dev/en/series/KRXMP"
license: "https://creativecommons.org/licenses/by-nc-nd/4.0/"
latest_value: -0.05
latest_date: "2026-08-01"
first_date: "2000-12-01"
observations_total: 309
observations_shown: 120
---

# South Korea Exchange Market Pressure Index

## Overview

Measures the combined pressure on the won that appears in the exchange rate, in reserves, and in the policy rate.

## Key Figures

|  | Value | Date |
|---|---|---|
| Latest | -0.05 | 2026-08-01 |
| Change from previous | +0.02 | 2026-07-01 |
| Change over one year | -0.05 | 2025-08-01 |
| Highest on record | 0.16 | 2008-11-01 |
| Lowest on record | -0.16 | 2008-12-01 |
| Period covered | 2000-12-01 – 2026-08-01 |  |
| Observations | 309 |  |

## Recent observations

| Date | Value | Change |
|---|---|---|
| 2016-09-01 | -0.02 | -0.01 |
| 2016-10-01 | 0.04 | +0.07 |
| 2016-11-01 | 0.02 | -0.02 |
| 2016-12-01 | 0.03 | +0.01 |
| 2017-01-01 | -0.04 | -0.08 |
| 2017-02-01 | -0.02 | +0.02 |
| 2017-03-01 | -0.01 | +0.01 |
| 2017-04-01 | 0.01 | +0.03 |
| 2017-05-01 | -0.01 | -0.02 |
| 2017-06-01 | 0.01 | +0.02 |
| 2017-07-01 | -0.02 | -0.03 |
| 2017-08-01 | 0.00 | +0.02 |
| 2017-09-01 | 0.02 | +0.02 |
| 2017-10-01 | -0.02 | -0.04 |
| 2017-11-01 | -0.04 | -0.02 |
| 2017-12-01 | -0.01 | +0.03 |
| 2018-01-01 | -0.00 | +0.01 |
| 2018-02-01 | -0.00 | 0.00 |
| 2018-03-01 | -0.00 | 0.00 |
| 2018-04-01 | 0.01 | +0.01 |
| 2018-05-01 | 0.00 | 0.00 |
| 2018-06-01 | 0.04 | +0.03 |
| 2018-07-01 | -0.00 | -0.04 |
| 2018-08-01 | -0.01 | 0.00 |
| 2018-09-01 | 0.00 | +0.01 |
| 2018-10-01 | 0.02 | +0.02 |
| 2018-11-01 | -0.02 | -0.04 |
| 2018-12-01 | -0.00 | +0.01 |
| 2019-01-01 | -0.00 | 0.00 |
| 2019-02-01 | 0.00 | 0.00 |
| 2019-03-01 | 0.02 | +0.02 |
| 2019-04-01 | 0.02 | 0.00 |
| 2019-05-01 | 0.03 | +0.01 |
| 2019-06-01 | -0.03 | -0.06 |
| 2019-07-01 | 0.02 | +0.05 |
| 2019-08-01 | 0.03 | +0.01 |
| 2019-09-01 | -0.01 | -0.04 |
| 2019-10-01 | -0.03 | -0.02 |
| 2019-11-01 | 0.01 | +0.04 |
| 2019-12-01 | -0.02 | -0.03 |
| 2020-01-01 | 0.02 | +0.04 |
| 2020-02-01 | 0.03 | +0.01 |
| 2020-03-01 | 0.01 | -0.02 |
| 2020-04-01 | 0.00 | -0.01 |
| 2020-05-01 | 0.01 | +0.01 |
| 2020-06-01 | -0.03 | -0.04 |
| 2020-07-01 | -0.01 | +0.02 |
| 2020-08-01 | -0.01 | 0.00 |
| 2020-09-01 | -0.01 | -0.01 |
| 2020-10-01 | -0.04 | -0.03 |
| 2020-11-01 | -0.03 | +0.01 |
| 2020-12-01 | -0.02 | +0.01 |
| 2021-01-01 | 0.02 | +0.04 |
| 2021-02-01 | -0.01 | -0.03 |
| 2021-03-01 | 0.02 | +0.03 |
| 2021-04-01 | -0.02 | -0.05 |
| 2021-05-01 | 0.01 | +0.03 |
| 2021-06-01 | 0.01 | +0.01 |
| 2021-07-01 | 0.01 | 0.00 |
| 2021-08-01 | 0.01 | 0.00 |
| 2021-09-01 | 0.02 | +0.01 |
| 2021-10-01 | -0.01 | -0.03 |
| 2021-11-01 | 0.02 | +0.03 |
| 2021-12-01 | -0.01 | -0.03 |
| 2022-01-01 | 0.02 | +0.02 |
| 2022-02-01 | 0.00 | -0.01 |
| 2022-03-01 | 0.01 | +0.01 |
| 2022-04-01 | 0.05 | +0.04 |
| 2022-05-01 | -0.02 | -0.07 |
| 2022-06-01 | 0.04 | +0.06 |
| 2022-07-01 | 0.01 | -0.03 |
| 2022-08-01 | 0.03 | +0.02 |
| 2022-09-01 | 0.07 | +0.04 |
| 2022-10-01 | -0.01 | -0.08 |
| 2022-11-01 | -0.06 | -0.05 |
| 2022-12-01 | -0.05 | +0.01 |
| 2023-01-01 | -0.03 | +0.02 |
| 2023-02-01 | 0.07 | +0.10 |
| 2023-03-01 | -0.01 | -0.08 |
| 2023-04-01 | 0.03 | +0.04 |
| 2023-05-01 | -0.01 | -0.04 |
| 2023-06-01 | -0.01 | 0.00 |
| 2023-07-01 | -0.02 | -0.02 |
| 2023-08-01 | 0.03 | +0.06 |
| 2023-09-01 | 0.02 | -0.01 |
| 2023-10-01 | 0.01 | -0.01 |
| 2023-11-01 | -0.05 | -0.05 |
| 2023-12-01 | -0.00 | +0.05 |
| 2024-01-01 | 0.03 | +0.03 |
| 2024-02-01 | 0.00 | -0.03 |
| 2024-03-01 | 0.01 | +0.01 |
| 2024-04-01 | 0.03 | +0.02 |
| 2024-05-01 | -0.00 | -0.03 |
| 2024-06-01 | 0.01 | +0.01 |
| 2024-07-01 | -0.00 | -0.01 |
| 2024-08-01 | -0.04 | -0.03 |
| 2024-09-01 | -0.01 | +0.02 |
| 2024-10-01 | 0.05 | +0.06 |
| 2024-11-01 | 0.01 | -0.04 |
| 2024-12-01 | 0.05 | +0.04 |
| 2025-01-01 | -0.02 | -0.07 |
| 2025-02-01 | 0.01 | +0.03 |
| 2025-03-01 | 0.02 | +0.01 |
| 2025-04-01 | -0.01 | -0.03 |
| 2025-05-01 | -0.04 | -0.03 |
| 2025-06-01 | -0.02 | +0.02 |
| 2025-07-01 | 0.02 | +0.04 |
| 2025-08-01 | 0.00 | -0.02 |
| 2025-09-01 | 0.01 | +0.01 |
| 2025-10-01 | 0.01 | 0.00 |
| 2025-11-01 | 0.03 | +0.02 |
| 2025-12-01 | -0.02 | -0.05 |
| 2026-01-01 | -0.00 | +0.02 |
| 2026-02-01 | -0.00 | 0.00 |
| 2026-03-01 | 0.06 | +0.06 |
| 2026-04-01 | -0.03 | -0.09 |
| 2026-05-01 | 0.02 | +0.05 |
| 2026-06-01 | 0.02 | 0.00 |
| 2026-07-01 | -0.07 | -0.09 |
| 2026-08-01 | -0.05 | +0.02 |

## Definition

Exchange market pressure is the combined capital-flow pressure on a currency that shows up jointly in the exchange rate, in the official reserves a central bank buys or sells, and in the policy rate it sets, so that a currency defended by intervention or by a rate hike registers pressure that a look at the exchange rate alone would miss (Girton and Roper 1977; Weymark 1995). For the Korean won the index of Goldberg and Krogstrup (2023) expresses this pressure as a single monthly number in exchange-rate-depreciation-equivalent units, formed as

$$EMP_t = \frac{\Delta e_t}{e_{t-1}} + \frac{\pi_i}{\pi_e}\,\Delta i_t - \frac{1}{\pi_e}\,FXI_t$$

where $e_t$ is the won per U.S. dollar, $i_t$ is the domestic policy rate, and $FXI_t$ is official foreign exchange intervention, with the conversion factors $\pi_e$ and $\pi_i$ translating a dollar of intervention and a percentage-point rate change into depreciation-equivalent units.

The sign convention is depreciation-positive, where $\Delta e_t / e_{t-1} > 0$ is a won depreciation, so that a larger value is greater depreciation pressure. Reserve accumulation that leans against appreciation enters with a minus sign and a rate hike that defends the currency against depreciation adds to the measure, so the index reads as a super-exchange-rate, namely the counterfactual depreciation that would have occurred had the pressure not been partly absorbed by intervention and by the rate (Girton and Henderson 1977; Calvo and Reinhart 2002).

What distinguishes this index from the earlier exchange-market-pressure tradition is the conversion factors themselves. Rather than the precision weights of the older literature, which weight each component by the inverse of its sample variance and thereby misassign weight under managed regimes, the factors are derived from a portfolio-balance model of imperfectly substitutable assets, so the weight on intervention and on the rate follows from balance-of-payments quantities rather than from sample volatility (Kouri 1981; Branson and Henderson 1985; Klaassen and Jager 2011). The series is a descriptive gauge of realized pressure rather than a forecast, and it places the won's episodes on the same footing as the cross-country measures built in this tradition (Eichengreen, Rose, and Wyplosz 1995; Aizenman and Hutchison 2012; Patnaik, Felman, and Shah 2017; Frankel and Rose 1996).

## Methodology

The index is built from three additive components and two conversion factors at monthly frequency, following the construction of Goldberg and Krogstrup (2023).

**(1) Headline identity.** The index is the sum

$$EMP_t = \frac{\Delta e_t}{e_{t-1}} + \frac{\pi_i}{\pi_e}\,\Delta i_t - \frac{1}{\pi_e}\,FXI_t$$

of the proportional exchange-rate change, the domestic policy-rate change scaled by $\pi_i / \pi_e$, and official intervention scaled by $1 / \pi_e$ and entered with a minus sign. The exchange-rate change is the monthly log change of the end-of-month won per U.S. dollar, and the rate change is the first difference of Korea's policy rate, with no shadow-rate substitution since Korea lies outside the zero-lower-bound roster for which shadow short rates replace the policy rate (Krippner 2015; Wu and Xia 2016).

**(2) Conversion factors.** The factors $\pi_e$ and $\pi_i$ are country- and time-varying and strictly positive, and they follow from a portfolio-balance model of imperfectly substitutable assets rather than from component variances, extending the original exchange-market-pressure identity onto a balance-of-payments foundation (Girton and Roper 1977; Kouri 1981; Branson and Henderson 1985). Each factor combines gross foreign-currency assets and domestic-currency liabilities with home-bias portfolio shares and with small calibrated interest-rate and exchange-rate elasticities, so that the interest sensitivities are deliberately small and the rate term contributes little (Girton and Henderson 1977; Weymark 1995). The position stocks are the gross external assets and liabilities net of direct investment, the home-bias shares govern the currency split, and both enter as in the definition-consistent measurement tradition (Lane and Milesi-Ferretti 2018; Coeurdacier and Rey 2013; Klaassen and Jager 2011).

**(3) Intervention leg.** For Korea the intervention series is the official balance-of-payments reserve-asset transaction flow, which is net of valuation changes by construction and is further corrected for the interest receipts that accrue on reserve holdings, with off-balance-sheet derivative positions and monetary gold both excluded from the reserve base. The reference currency is the U.S. dollar, assigned by the monetary-anchor rule that governs the sample (Klein and Shambaugh 2008).

**(4) Segmented construction.** The published series concatenates three segments, namely an overlap segment that adopts the official Goldberg-Krogstrup Korea inputs, an official-tail segment that carries those inputs forward, and a live segment that extends the position stocks from current external accounts while holding the home-bias shares, the currency weights, and the interest-receipt correction fixed at their last official values, in the paper's own frozen-weights spirit (Goldberg and Krogstrup 2023). Reconstructed from the official inputs, the series reproduces the official Korea index with a correlation of 0.999885 over the 241-month validation overlap. Because late external-account and balance-of-payments prints revise recent months, the live segment is revised rather than real-time.

## Applications in Economics

Exchange market pressure was introduced to measure the pressure on a currency when a central bank absorbs part of it through intervention rather than letting the exchange rate move, so that the measure captures both the realized move and the pressure taken onto the official balance sheet (Girton and Roper 1977; Goldberg and Krogstrup 2023). This makes it a natural gauge of capital-flow pressure in economies that manage the exchange rate, where the flows show up partly in reserves and partly in the rate (Forbes and Warnock 2012; Aizenman and Hutchison 2012).

The measure has a long history in the currency-crisis literature, where a pressure index built from the exchange rate and reserves, sometimes with the interest rate, is used to date speculative attacks and to construct early-warning models (Eichengreen, Rose, and Wyplosz 1995; Kaminsky, Lizondo, and Reinhart 1998; Sachs, Tornell, and Velasco 1996). Because a defended currency can be under severe pressure without a large exchange-rate move, the index dates episodes that a look at the rate alone would understate (Frankel and Rose 1996; Kaminsky and Reinhart 1999).

The construction also speaks to fear of floating, the documented reluctance of emerging-market authorities to let the exchange rate move freely, which pushes pressure into reserves and rates (Calvo and Reinhart 2002; Ilzetzki, Reinhart, and Rogoff 2019). The reserve-and-rate legs of the index are exactly the margins through which such absorption operates, so the index measures the pressure that fear-of-floating behavior would otherwise hide (Aizenman and Hutchison 2012).

Finally, the index bears on the global financial cycle, the common component in gross capital flows, asset prices, and leverage that moves emerging-market currencies together and often independently of domestic fundamentals (Rey 2015; Cerutti, Claessens, and Rose 2019). Read in this light the won's pressure is partly a local reflection of that global cycle, and the index is a descriptive gauge for surveillance rather than a forecast, most useful when read with capital-flow and financial-conditions indicators rather than on its own.

## Applications in Financial Markets

For a market audience the index is a monthly read on how much depreciation pressure the won faced and how much of it was absorbed rather than realized, which the exchange rate alone does not convey (Goldberg and Krogstrup 2023). Its reserve leg makes it a lens on official intervention, whose effectiveness and timing are the subject of a large empirical literature and which the reserve-and-rate margins of the index directly summarize (Adler, Lisack, and Mano 2019; Fratzscher et al. 2019). Asymmetric intervention, where authorities resist depreciation more than appreciation, leaves a signature in exactly these legs (Pontines and Rajan 2011; Pontines and Siregar 2012).

Because the index is defined identically across countries, it lets a desk compare the won's pressure against other emerging and advanced markets on a common scale, which is the cross-country comparability that motivated the conversion-factor construction (Patnaik, Felman, and Shah 2017; Aizenman and Hutchison 2012). The same comparability places the won's episodes alongside the surge-and-stop dynamics of global capital flows that drive emerging-market currency risk (Forbes and Warnock 2012; Bluedorn et al. 2013).

The reserve leg also connects the index to reserve adequacy. A currency under sustained depreciation pressure that is met by reserve sales runs its buffers down, so the pressure series read alongside the level of reserves informs the standard reserve-adequacy and external-vulnerability assessments (Obstfeld, Shambaugh, and Taylor 2010; Ghosh, Ostry, and Qureshi 2018). This is the risk-monitoring use for which such measures are best suited, complementing rather than replacing the currency-pricing signals that carry across the cross-section of exchange rates (Habib and Stracca 2012; Sarno and Schmeling 2014).

Read this way the index is a descriptive input to foreign-exchange risk monitoring and intervention analysis rather than a tradable signal, and it is most informative alongside the platform's cross-currency-basis, reserves, and financial-conditions indicators.

## Statistical Tests

Over 305 observations from 2000-12-01 to 2026-04-01 the exchange market pressure index is a clean I(0), since on the level with a constant the augmented Dickey and Fuller (1979) test in the lag augmentation of Said and Dickey (1984) and the Phillips and Perron (1988) test with a Newey and West (1987) long-run variance both decisively reject the unit root, with statistics of −17.248 and −17.2756, while the Kwiatkowski et al. (1992) test fails to reject stationarity at a statistic of 0.196366 with p above 0.10, so the three tests agree on stationarity. The strongly negative statistics are expected, since each component of the index is a first difference, namely the log exchange-rate change, the policy-rate change, and the intervention flow, so the composite inherits near-serially-uncorrelated behavior from its ingredients (Hamilton 2018).

On the level the first-order autocorrelation is 0.0105878 with an implied half-life of 0.152405 months, a descriptive persistence summary rather than an integration claim, and the Ljung and Box (1978) portmanteau finds no serial correlation, with Q(12) = 9.3149 at p = 0.675825 and Q(24) = 23.8552 at p = 0.469903. Because the time-varying conversion factors grow with the gross external positions, they impart a scale-drift and heteroskedasticity caveat to these persistence and portmanteau readings, which are therefore read as descriptive summaries rather than as tests of a homoskedastic I(0) data-generating process, with the robust portmanteau of Escanciano and Lobato (2009) and the bounded-sample cautions of Cavaliere (2005) as the appropriate companions.

The Bai and Perron (1998) procedure, by the Bai and Perron (2003) algorithm, finds no break in the level, and the single-break sup-Wald statistic of Andrews (1993), with the corrigendum of Andrews (2003), is 3.53745 with an argmax at 2007-11-01, below the critical value of 8.68, so it fails to reject parameter constancy in agreement with the no-break finding. The individual crises appear as amplitude spikes rather than as level shifts. The published index concatenates an overlap segment on the official inputs, an official-tail segment, and a live segment whose position stocks are extended at frozen currency shares and whose intervention correction is held at a frozen constant, and late external-account and balance-of-payments prints revise its recent months under a full-replace loader, so the series is revised rather than real-time and is disclosed as a vintage caveat.

## Frequently Asked Questions

### What is the exchange market pressure index?

It uses the fact that capital-flow pressure on a currency shows up not only in the exchange rate but also in official reserve sales and purchases and in the policy rate, and combines the three channels into a single depreciation-equivalent reading.

### Why does exchange market pressure look beyond the exchange rate?

Episodes in which the authorities declined to let the currency weaken and defended it through intervention or rate increases instead. The rate looks calm in those periods while the pressure sits in reserves and rates.

### Are the exchange market pressure components published?

Only the headline index is published. Attributing pressure across the three channels is far more sensitive to the weighting scheme and the assumed elasticities than the composite itself, so it is kept outside the published surface.