Abstract
This Selected Issues paper for the Republic of Korea focuses on the role of monetary policy in the current context of slowing growth and rising inflation pressures. Korea has not remained immune to the global slowdown, and with the cycle turning downward, the trade-off between inflation and growth is deteriorating. Subprime-related turbulences in financial markets add an extra element of uncertainty to the economic outlook, and have led to a noticeable increase in Korea’s stock market and exchange rate volatility.
I. Rising Prices, Slowing Growth, and the Implications for Monetary Policy 1
A. Introduction
1. Like many countries, Korea is facing an increasingly challenging environment for conducting monetary policy. Commodity prices are fueling headline inflation and what started largely as a supply-side shock is feeding into higher core inflation. Given the large terms of trade shock—Korea is the fifth largest oil importer in the world—the won has lost 16 percent of its value since mid-2007, further adding to inflationary pressures. In addition, Korea has not remained immune to the global slowdown and with the cycle turning downward, the trade-off between inflation and growth is deteriorating. Finally, subprime-related turbulences in financial markets add an extra element of uncertainty to the economic outlook, and have led to a noticeable increase in Korea’s stock market and exchange rate volatility.
2. In addition, current events pose the first real test to Korea’s inflation targeting framework. With inflation now exceeding its target for several quarters, the main challenge will be to keep inflation expectations well anchored. If inflation expectations get out of hand a wage-price spiral may ensue and make the eventual adjustment more costly in terms of lost output. It is therefore paramount that monetary policy remain ahead of the curve and that the Bank of Korea (BOK) communicate clearly the rationale of its rate decisions.
3. This chapter uses a small structural macro model to analyze the inflation outlook and challenges for monetary policy. The model is the IMF’s forecasting and policy analysis system (FPAS) and is used, in similar forms, by central banks around the world2. Parameter specifications have been chosen such that the model reproduces key characteristics of the Korean economy. The chapter proceeds as follows. The next section reviews Korea’s monetary policy framework, discusses its track record, and describes recent inflationary developments. Section II presents the model and its calibration to Korean circumstances. Section III reports the baseline projection and various shocks to the baseline, including policy responses. Section IV concludes.
B. Background and Recent Developments
4. In 1998, Korea adopted inflation targeting, as financial innovation had made the earlier framework of monetary targeting impractical. For a transitional period the BOK used both systems in parallel, but from 2001 onwards monetary aggregates were dropped as intermediate targets. The Bank of Korea Act stipulates price stability as the purpose of the central bank and every three years the bank sets a medium-term inflation target which it seeks to achieve on average. The central bank targets headline inflation, except for the period 2000-06, when it targeted core inflation, and since 2004 the target band has been 2.5-3.5 percent. Once a month the bank’s monetary policy committee decides on the policy interest rate, which was changed from the overnight call rate to the 7-day repo rate (Base Rate) in March 2008.
5. The inflation targeting framework has served the country well. Between 1998 and late-2007, the year-on-year inflation rate exceeded the upper target band on only one occasion lasting for two months. In addition, Kim and Park (2006) observe that inflation has been lower and less volatile under inflation targeting even after controlling for the size of shocks. Also, inflation expectations seem to be better anchored under the new framework as evidenced by lower inflation persistence and a lower influence of actual inflation on inflation expectations.
6. However, until recently the inflation targeting framework had not been put to a real test. The relative success of inflation targeting may owe much to the special economic circumstances of the last years. For one thing, the sizeable and steady increase in the exchange rate—the won appreciated by 70 percent in NEER terms between 1998 and mid-2007—helped keep inflation at bay. Also, inflation targeting in Korea may have been helped by the integration of China’s and India’s vast labor pool into the global economy and the wage moderation that this induced.
7. Meanwhile global and domestic circumstances have become more challenging (Figure I.1.) Real oil prices are at historical highs at a time when the global economy is slowing and key domestic variables point to a risk of sustained inflation:
Oil price inflation reached 90 percent q/q annualized, in the second quarter of 2008 and is projected to stay above 10 percent through the third quarter. Beyond that oil prices are projected to stay broadly flat.
The U.S. output gap is estimated to have fallen to a negative 0.4 percent in the second quarter of 2008. By the first quarter of 2009 U.S. GDP is projected to fall 2.2 percent below potential and remain close to this value throughout 2009.
Headline inflation in Korea reached 4.8 percent y/y in the second quarter and 8.2 percent q/q annualized, breaching the Bank of Korea’s target band for the second quarter running. Core inflation has been trending up for some time and stood at 3.9 percent y/y in the second quarter. The gap between producer and consumer price inflation and measures of inflation expectations are also trending upward boding ill for a quick reversal of inflationary trends.
Korea: Recent Inflationary Developments
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Korea: Recent Inflationary Developments
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Korea: Recent Inflationary Developments
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Given the lags in price dynamics, the unfolding slowdown of domestic demand may take some time in providing inflation relief. The estimated output gap for Korea was positive through the second quarter. Other variables that support domestic demand (but are not captured in the model) are rapid money and credit growth of 14 percent and 15 percent, respectively, in June, and an accommodative fiscal stance, with a stimulus projected at 1½ percent of GDP in 2008.
The policy interest rate was kept at 5 percent in the first two quarters of 2008. While this constitutes a 7-year high, the monetary policy stance is accommodative: in the second quarter, the real interest rate was 0.2 percent, well below the 1.5 percent that has, on average, prevailed under the inflation targeting framework and is, hence, assumed to constitute the neutral real rate of interest.
The real exchange rate fell by 9¼ percent over the first two quarters of 2008. The real exchange rate is estimated to be broadly in equilibrium and, therefore, does not add to demand pressures. However, the pass-through of the weaker currency has fueled inflation more directly.3
C. The Model
8. The FPAS model describes a small open economy with an inflation targeting framework. It combines the New Keynesian emphasis on nominal rigidities and the role of domestic demand in output determination with the rational, forward-looking behavior propagated by the real business cycle literature. The model expresses each variable in terms of its deviation from equilibrium and does not attempt to explain the equilibrium values themselves.4 It is a two-country model, in which Korea’s economy is depicted by four key equations, that can be derived from micro foundations (see also Appendix I.1):
IS Curve. This equation describes the interest rate channel of monetary policy. By raising borrowing costs interest rate hikes are assumed to reduce domestic demand and, hence, the output gap. However, significant lags in the monetary transmission mechanism mean that the interest rate coefficient is small relative to the coefficient on the lagged output gap. External demand is assumed to depend on the U.S. output gap and exchange rate misalignment.
Phillips Curve. Underlying this equation is mark-up pricing by enterprises over wage costs, where workers take into account in their wage negotiations the level of unemployment (output gap) and expected inflation. The more backward-looking agents are in forming inflation expectations, the higher the cost of disinflation in terms of lost output. Other determinants of inflation in this equation are oil price inflation and the rate of exchange rate depreciation.
Interest Parity Condition. This equation states that the expected depreciation of the won relative to the dollar is equal to the risk-adjusted interest rate gap over the United States. In projecting the exchange rate, some actors have perfect foresight while others have adaptive expectations. This leads to Dornbusch-like exchange rate overshooting in slow motion.
Monetary Policy Rule. The central bank is assumed to raise nominal interest rates when actual inflation exceeds the target and output exceeds potential. The coefficient on the inflation gap is usually greater than 1 to yield a positive real interest rate response, and greater than the coefficient on the output gap. Moreover, central banks usually give some weight to past policy rates as a smoothed interest rate path is less demanding on the financial sector.5
In sum, the model has four endogenous variables—output gap, inflation gap, real interest rate gap, and exchange rate gap—which equal zero in the steady state. In the event of a positive shock to inflation, interest rate increases lower inflation directly by curbing domestic demand and indirectly through exchange rate appreciation and the dampening effect this has on external demand and pass-through.
9. The model has been parameterized to reproduce key characteristics of the Korean economy (see also Appendix I.1). In a first step, models of similar countries were used to inform the parameter specification; in particular, the Canadian model by Berg and others (2006) was used as a benchmark. In a second step, shocks were applied to the model’s steady state, and the stylized facts generated in this way were compared to what is known about the monetary transmission mechanism in Korea. For example, the model’s sacrifice ratio, the effect of oil price and exchange rate movements on growth and inflation, and the lag in monetary transmission are broadly in line with the Korea-specific literature and accounts by Korean researchers and policy makers. However, the short track record of the inflation targeting framework and the absence of major shocks complicates the parameterization of the model and warrants more than the usual caution in interpreting the results.
D. Simulation Results
10. This section uses the model to analyze the inflation outlook and risks to the outlook, as well as the implications for monetary policy. While the baseline forecast is informed by the model’s predictions it is, in effect, a judgmental forecast that takes into account a much broader set of available data, including short-term indicators, market expectations, and views of policy makers. To analyze key risks to the outlook, residuals in the main equations are chosen such that the model reproduces exactly the baseline forecast. Subsequently, this tuned baseline forecast is subjected to various shocks.
Baseline Forecast
11. In the baseline, monetary tightening and slower growth help contain inflationary pressures (Figure I.2). Using a standard parameterization of the monetary reaction function, as well as parameterizations more in line with past BOK behavior, the model calls for an interest rate hike in the third quarter of 2008. The baseline assumes an interest rate hike by 0.5 percentage points to 5.5 percent. While this move would leave real interest rates in accommodative territory, the resulting appreciation in the currency and, more importantly, the projected economic slowdown should help bring down inflation starting from the first quarter of 2009. The output gap should turn negative from the third quarter onward and growth is expected to remain below potential throughout 2009. Despite the projected moderation of inflationary pressures—oil price inflation is also expected to come down significantly—common measures of inflation persistence suggest that headline inflation would remain elevated for some time and stay above the target band for most of 2009.
Baseline Projections
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Baseline Projections
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Baseline Projections
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Oil price Shock
12. With higher oil prices, monetary tightening would need to be more aggressive(Figure I.3). There is a lot of uncertainty surrounding the oil price baseline, with upside risks from low spare capacity and downside risks from slower global growth. The shock scenario assumes that oil prices reach US$200 per barrel in the fourth quarter of 2008 and return steadily to the baseline by 2012. Under such circumstances, inflation would be about ¾ percentage points higher in 2009 relative to the baseline and return into the target band only by the second quarter of 2010. Policy rates would have to rise to above 6 percent when the shock occurs and remain at that level for three quarters. This would remove most monetary accommodation by mid-2009. Since U.S. and Korean monetary policy react similarly to the oil price shock, the exchange rate would be little affected. The additional monetary tightening in response to the shock, as well as the further slowdown in U.S. growth in response to higher oil prices, would reduce growth by up to 0.3 percentage points and delay the return to full potential by several quarters.
Oil Price Shock
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Oil Price Shock
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Oil Price Shock
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Lower U.S. Growth
13. A deeper than expected U.S. downturn could require some monetary easing (Figure I.4). An end to the U.S. subprime crisis is not in sight. Hence, this scenario assumes that U.S. growth will be 1 percentage point lower in the fourth quarter of 2008 relative to the baseline scenario and remain below the baseline until mid-2010.6 This will reduce growth in Korea by 0.3 percentage points in the quarter of the shock and 0.7 percentage points in the second quarter of 2009. The policy rate should be steadily reduced to a low of 2¾ percent in early 2010, thereby providing monetary stimulus over much of the medium term. Again, the exchange rate would be little affected as U.S. and Korean monetary policy would respond similarly to the demand shock. Despite the monetary accommodation inflation would be lower by up to 0.6 percentage points relative to the baseline, but the inflationary impact of lower growth would occur with a lag of 2 quarters.
Shock to U.S. Growth
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Shock to U.S. Growth
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Shock to U.S. Growth
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Exchange Rate Shock
14. Exchange rate depreciation should lead to monetary tightening if balance sheets remain intact (Figure I.5). The global financial turmoil is far from over and markets remain volatile. In such a situation, another bout of bad news could trigger a substantial depreciation of the Korean won. This scenario assumes that capital outflows lead to a 10 percent depreciation of the won in the last quarter of 2008 without damaging balance sheets and, hence, constraining monetary policy. The weaker currency boosts inflation by 0.3-0.4 percentage points relative to the baseline and the return into the target band is delayed by one year. Monetary policy helps contain the inflationary impact of the devaluation by raising the policy rate by a maximum of 1¾ percentage points relative to baseline. If balance sheet effects can be avoided the weaker currency would give a boost to GDP.
Exchange Rate Shock
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Exchange Rate Shock
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
Exchange Rate Shock
Citation: IMF Staff Country Reports 2008, 296; 10.5089/9781451822236.002.A001
E. Conclusion
15. Current circumstances call for a tightening of monetary policy, despite the projected slowdown of global and domestic demand. Higher oil prices, the weak won, and still-high money and credit growth will continue to exert inflationary pressure over the coming months. This needs to be weighed against the deflationary impact of slower global and domestic growth over the next quarters. Using a standard parameterization of the monetary policy reaction function, as well as parameterizations more in line with past BOK behavior, the model calls for an interest rate hike in the third quarter of 2008.
16. Major risks to the baseline forecast, if realized, call for a significant monetary policy response. Under a fairly common specification of the monetary policy rule, an increase in the oil price to US$200 per barrel should raise the policy rate by 1 percentage point in the first year relative to the baseline; a 1 percentage point shock to U.S. growth should allow a 1 percentage point easing in the policy rate in the first year; and a 10 percent depreciation of the exchange rate would call for a 1¼ percentage point hike of the policy rate in the first year, if balance sheets remain intact.
Appendix I.1. Model Equations and Parameterization
A. Model Equations
IS Curve:
Phillips Curve:
Interest Parity Condition:
Monetary Policy Rule:
B. Variable Definitions
| ygap | output gap, percent |
| ygapUS | U.S. output gap, percent |
| r* | equilibrium real interest rate, percent per annum |
| rUS | U.S. real interest rate, percent per annum |
| z | log of the real exchange rate times hundred (increase implies depreciation) |
| z* | log of the equilibrium real exchange rate times hundred |
| ze | log of the expected real exchange rate times hundred |
| ε | error term (e.g., to calibrate model predictions to actual data) |
| π | quarter-on-quarter CPI inflation, annualized, percent |
| π4 | year-on-year CPI inflation, percent |
| πoil | quarter-on-quarter real oil price inflation, annualized, percent |
| π* | target inflation rate, annual, percent |
| ρ | risk premium, percent per annum |
| i | nominal interest rate, percent per annum |
| ygap | output gap, percent |
| ygapUS | U.S. output gap, percent |
| r* | equilibrium real interest rate, percent per annum |
| rUS | U.S. real interest rate, percent per annum |
| z | log of the real exchange rate times hundred (increase implies depreciation) |
| z* | log of the equilibrium real exchange rate times hundred |
| ze | log of the expected real exchange rate times hundred |
| ε | error term (e.g., to calibrate model predictions to actual data) |
| π | quarter-on-quarter CPI inflation, annualized, percent |
| π4 | year-on-year CPI inflation, percent |
| πoil | quarter-on-quarter real oil price inflation, annualized, percent |
| π* | target inflation rate, annual, percent |
| ρ | risk premium, percent per annum |
| i | nominal interest rate, percent per annum |
C. Parameterization
Parameters were chosen such that the model, when shocked in steady state, reproduces key characteristics of the monetary transmission mechanism in Korea. In particular:
Sacrifice Ratio. A permanent 1 percentage point reduction in the inflation target yields a cumulative negative output gap of 1.3 percent in the first year and 5.3 percent over 5 years. The first sacrifice ratio is in line with the Canada model, while the second falls within the range of 4.6-5.5 percent, the only known estimate for Korea, unfortunately predating the inflation targeting framework (Kim and Shon, 2002).
Monetary Transmission Lags. An interest rate hike has its biggest effect on the output gap 3-4 quarter after the shock, or broadly in line with BOK accounts of actual transmission lags. The maximum effect on inflation takes 8 quarters to materialize, which is at the upper bound of BOK estimates (3-8 quarters).
Oil price Shock. A permanent jump in the oil price by 10 percent (i.e., a one-off inflation shock) raises inflation by up to 0.2 percentage points and reduces growth by up to 0.2 percentage points. This is in line with rules-of-thumb cited by academics and policy makers in Korea.
Exchange Rate Shock. A 10 percent depreciation of the won raises inflation by up to 0.5 percentage points. This is equal to the lower bound estimate (0.5-0.8 percentage points) of a recent unpublished study by BOK and somewhat lower than the 0.8 percent estimated by the Korea Development Institute (Dong-Chul and Jun-Hyuk, 2008).
Coefficient of U.S. Output Gap. Based on a VAR and data for 1991 -2007 the IMF estimates that a one percentage point decline in U.S. growth reduces Korean growth by 0.2 percentage points (IMF, 2008). The model parameter of 0.25 is slightly higher, given that global linkages have increased over time.
Monetary Policy Rule. Korea’s monetary policy rule has been estimated, e.g., by Eichengreen (2004) and Kim and Park (2006). While the coefficients on the inflation and output gap are not too dissimilar from the model coefficients, the estimated coefficient on the lagged policy rate is above 0.9. Under current circumstances this degree of smoothing leads to explosive inflation dynamics in the model. Hence, a policy rule was chosen that leads to reasonable results in similar countries.
| α1 =0.1 | β1 =0.2 | δ =0.5 |
| α2 =0.85 | β2 =0.3 | Φ =0.5 |
| α3 =0.1 | β3 =0.1 | γ1 =2 |
| α4 =0.05 | β4 =0.005 | γ2 =0.5 |
| α5 =0.25 | β5 =0.005 |
| α1 =0.1 | β1 =0.2 | δ =0.5 |
| α2 =0.85 | β2 =0.3 | Φ =0.5 |
| α3 =0.1 | β3 =0.1 | γ1 =2 |
| α4 =0.05 | β4 =0.005 | γ2 =0.5 |
| α5 =0.25 | β5 =0.005 |
References
Berg, Andrew, Philippe Karam, and Douglas Laxton, 2006, “Practical Model-Based Monetary Policy Analysis—A How-to Guide,” IMF working paper 06/81 (Washington: International Monetary Fund).
Dongchul, Cho and Song Joonhyuk, 2008, “The Cause and Prospect of Current High Inflation Trends,”(in Korean), Economic Outlook of the Korea Development Institute, pp. 165–73, (Seoul, Korea).
Eichengreen, Barry, 2004, “Monetary and Exchange Rate Policy in Korea: Assessments and Policy Issues,” CEPR Discussion Paper 4676 (London: Center for Economic Policy Research).
IMF, International Monetary Fund, 2008, “Can Asia Decouple? Investigating Spillovers from the United States to Asia,” in Regional Economic Outlook: Asia and Pacific Department, April, pp. 27–40 (Washington: International Monetary Fund).
Kim, Joonwon, and Youngki Shon, 2002, “Structural Estimates of the Sacrifice Ratio: The Case of Korea,” Economic Studies, Vol. 50, No. 2, pp. 185–206 (in Korean).
Kim, Soyoung, and Yung Chul Park, 2006, “Inflation Targeting in Korea: A Model of Success,” BIS Paper No. 31, pp. 140–64 (Basel: Bank of International Settlement).
Prepared by Erik Lueth.
The equilibrium real exchange rate in Figure I.1 has been derived with an augmented HP-filter and is distinct from the equilibrium exchange rate generated by the CGER exercise.
Equilibrium values are derived with the help of an augmented HP-filter that leaves room for value judgments. For example, the smoothing procedure can be programmed to produce a zero output gap in a particular year.
In the Korean context it may also be explained by the strain abrupt interest rate hikes would put on households given the predominance of flexible rate mortgages.
Since the model is a two-country model, U.S. growth is determined endogenously and stays below the baseline for several quarters. The U.S. growth shock does not affect oil prices which are exogenous in the model, hence, simulated effects on Korean growth and inflation should be considered as upper and lower bounds, respectively.
