## Chapter 2 — Introduction and Main Findings

## Source details

**Canonical URL:** [Chapter 2 — Introduction and Main Findings](https://www.imf.org/-/media/files/publications/reo/apd/2018/apd-reo-april-2018-chapter-2.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/reo/apd/2018/apd-reo-april-2018-chapter-2.pdf.md)
- [Structured JSON version](/-/media/files/publications/reo/apd/2018/apd-reo-april-2018-chapter-2.pdf.json)

---

### Overview and methods
- Global growth in 2017 was the highest since 2011 and is expected to strengthen further in 2018–19, supported by broad-based momentum across countries and fiscal expansion in the United States.
- Headline inflation has been picking up with the upturn in oil prices since September, but core inflation remains surprisingly subdued, especially in advanced economies.
- Methods:
  - Estimation of augmented Phillips curves (time-varying parameter New Keynesian specification).
  - Principal component / latent-factor analysis to separate global factors from country-specific factors.
  - Unobserved-components (trend/transitory) decomposition of inflation.

### Key findings (summary)
- Recent low inflation in Asia has been driven mainly by temporary forces, including imported inflation.
  - Phillips curve estimation indicates weaker import prices, including low commodity prices, contributed to half of the undershooting of inflation targets in advanced Asia and most of the undershooting in emerging Asia in recent years.
  - China played an important role in driving both global and regional inflation.
  - Temporary shocks have accounted for the bulk of the recent reduction in inflation.
- The inflation process has become more backward-looking.
  - Inflation expectations are generally relatively well anchored, especially in advanced Asia and in economies with inflation-targeting frameworks.
  - The importance of expectations in driving inflation has declined; past inflation plays a larger role.
- The sensitivity of inflation to the unemployment gap has declined (flattening Phillips curve).
  - Advanced-Asia median slope declined from about 1 in the 1990s to about 0.6.
  - Other advanced economies had median coefficients of about 0.2 in the 1990s.
  - Emerging-Asia slope has been declining steadily and significantly.
- The coefficient on imported inflation remained broadly constant in both advanced economies and emerging markets.

### Recent inflation dynamics in Asia
- Headline inflation declined sharply during 2012–15 across many advanced, emerging market, and developing economies in Asia.
- Disinflation was broad-based across sectors (food, other goods, services) and across inflation measures (headline, core, wage inflation).
- Food inflation declined the most across advanced economies but generally remained positive, indicating low pass-through from international to domestic food prices.
- Other goods inflation entered negative territory in several advanced Asian economies, reflecting large declines in manufacturing producer prices (PPI).
- Headline inflation started to pick up in 2016; pickup driven primarily by other goods and services inflation and by recovery in the manufacturing PPI.
  - Pickup observed in Australia, Japan, Korea, and some ASEAN economies.
  - In China, PPI pickup did not spill over to CPI.
- Despite the pickup, headline and core inflation remain below inflation targets in most economies.

### Structural drivers — New Keynesian Phillips curve (specification and stylized results)
- Model specification (Annex 2.1):
  - π_t = a_t π^e_{t+10} + (1 − a_t) π̂_{t−1} + β_t u^c_t + γ_t π^m_t + ε_t
    - π_t: headline consumer price inflation.
    - π^e_{t+10}: inflation expectations 10 years ahead (or 5-year WEO forecasts when unavailable).
    - π̂_{t−1}: moving average of previous four quarters of inflation.
    - u^c_t: cyclical unemployment (deviation from NAIRU).
    - π^m_t: imported inflation (import-price deflator relative to GDP deflator).
    - a_t: weight on long-term expectations; β_t: Phillips-curve slope; γ_t: imported-inflation coefficient.
- Main structural findings:
  - The coefficient on inflation expectations has declined in both advanced economies and emerging markets back to levels of the early 2000s, implying inflation is driven more by past inflation than by expectations.
  - In advanced Asia, the role of long-term expectations was less important historically but has converged; since the global financial crisis the coefficient has declined in line with other advanced economies.
  - In emerging Asia, expectations have generally played a larger role than in other emerging markets; adoption of inflation targeting tends to increase the role of expectations, and declines since the global financial crisis have been smaller and slower.
  - The slope of the Phillips curve (β_t) has declined:
    - Advanced Asia median β: about 1 in the 1990s → about 0.6 more recently.
    - Other advanced economies median β in the 1990s: about 0.2.
    - Emerging Asia shows steady flattening across most countries; result robust to alternative slack measures (NAIRU, capacity utilization, output-gap estimates).
  - The coefficient on imported inflation (γ_t) remained broadly constant across advanced and emerging economies.

### Quantified contributions to undershooting inflation targets (regional results)
- Imported inflation emerged as the most important driver of inflation deviations from targets in recent years.
  - Advanced economies: cyclical unemployment was the main driver before the global financial crisis; import prices accounted for about half of the undershooting since 2013.
  - Emerging markets: undershooting since 2014 mainly accounted for by import prices; cyclical unemployment also important.
  - Inflation expectations began contributing to undershooting in 2016–2017: accounted for 35 percent of undershooting in advanced Asia and 40 percent in emerging Asia in 2017:Q1–Q2.
  - Depreciation of Asian currencies in 2015 following China-induced volatility contributed positively to inflation in both advanced and emerging Asia.
- Large unexplained share of undershooting remains, especially in advanced economies (possible causes: mismeasurement of slack, divergence between price-setter expectations and professional forecasters, zero lower bound).

### Cross-country/global factor analysis (latent-factor / PCA results)
- Three global factors explain 56 percent of the variance in headline inflation in the full cross-country panel.
  - Factor 1: 33% of variance.
  - Factor 2: 15% of variance.
  - Factor 3: 8% of variance.
- Factor 1 correlates with global commodity prices:
  - ρ commodity all = 0.52
  - ρ fuel = 0.56
  - ρ nonfuel = 0.28
- Factor 3 relates to movements in the US-dollar NEER with varying correlations across samples and subperiods (examples preserved from source):
  - AE, EM, LIC sample: ρ NEER = 0.64
  - AE, EM sample: ρ NEER = 0.19
  - AE, EM, LIC sample: ρ NEER = –0.06
  - AE, EM sample: ρ NEER = 0.58
  - 2001Q1 - 2014Q2: 0.52 and 0.58 reported in different contexts
  - 2014Q3 - 2017Q3: –0.28 and 0.91 reported in different contexts
- China’s role:
  - Correlation of Chinese PPI with Factor 1: 0.34.
  - Correlation of Factor 3 (AE and EM sample) with Chinese PPI: 0.59 overall; post-2014 correlation: 0.91.
  - Regional residual PCA: first common regional factor explains 50 percent of residual variation; correlation of this regional factor with Chinese PPI: 0.83.
- Many economies have more than 50 percent of inflation variance explained by common factors; Emerging Asia appears more responsive to Factor 1 than Latin America.

### Trend versus transitory components
- Unobserved-components framework (Annex 2.3): π_t = π^*_t + c_t, with trend π^*_t and deviation c_t.
- Trend inflation has been declining in both advanced and emerging Asia:
  - Came down substantially over the 1990s.
  - Experienced another, milder, decrease after 2011.
- Transitory component:
  - Concentration of transitory disinflationary shocks in the region since 2014; transitory component predominantly negative for most economies since 2014.
  - Positive transitory shocks could lift inflation relatively quickly.
- Nonlinearities: evidence suggests Phillips-curve sensitivity may increase at higher inflation levels, implying the possibility of jumps in inflation when underlying pressures rise.

### Evidence linking Phillips-curve flattening to structural factors (panel regressions)
- Advanced Asia: flattening statistically significantly linked to automation (relative price of investment goods).
  - Reported panel regression: ŷ_t = –1.865*** + Pinv × 3.173***; Time and country fixed effects; R^2 = 0.647.
- Emerging Asia: flattening statistically significantly linked to GVC integration (share of foreign value-added in exports).
  - Reported panel regression: ŷ_t = 1.077*** + Foreign VA × (–0.0267***); Time and country fixed effects; R^2 = 0.66.
- Interpretation: automation and greater GVC integration weaken labor bargaining power and can flatten the Phillips curve; as these factors stabilize, the slope may normalize.

### Inflation expectations: anchoring and effects of inflation targeting
- Two approaches used:
  1. Fraction of time expectations are within inflation targets (Consensus Forecasts by horizon).
  2. Response of expectations to inflation surprises (estimated β_h).
- Key numeric examples:
  - One-year-ahead sensitivity: a 1 percentage point increase in inflation raises one-year-ahead expectations by 0.32 percentage points in emerging markets and 0.26 percentage points in advanced economies.
- Empirical patterns:
  - Expectations are better anchored at longer horizons (one-year to five-year ahead) than for the current year.
  - Inflation expectations have become better anchored over time for both advanced and emerging Asia.
  - Adoption of inflation targeting tends to reduce sensitivity of expectations to inflation shocks; six economies analyzed include New Zealand (March 1990), Australia (July 1993), Korea (January 1998), Philippines (January 2002), Thailand (May 2002), Indonesia (July 2005).
- Tables (selected preserved values):
  - Table 2.1 (Percent of Time Expectations within Target Range, IT start till 2017) — examples:
    - Australia: horizons 0–5 = 47 07 68 28 68 88
    - China: 07 99 29 21 00 96
    - Japan: 25 65 80 80 80 90
    - Korea: 48 86 90 90 80 78
    - Thailand: 50 72 83 83 86 89
  - Table 2.2 (Percent of Time Expectations within Target Range, 2013–17) — examples:
    - Australia: 100 100 100 100 100
    - China: 45 75 100 100 100
    - Japan: 25 65 80 80 80 90
    - New Zealand: 100 100 100 100 100
    - Thailand: 90 100 100 100 100
  - Table 2.3 (Average Absolute Deviation from Target, IT start till 2017) — examples (horizons 0–4):
    - Australia: 0.7 0.5 0.4 0.3 0.2
    - China: 1.3 0.6 0.6 0.5 0.3
    - Korea: 0.7 0.4 0.3 0.3 0.3
    - Thailand: 1.4 0.9 0.9 1.0 1.0
  - Table 2.4 (Average Absolute Deviation from Target, 2013–17) — examples:
    - Australia: 0.4 0.2 0.2 0.1 0.1
    - China: 1.0 0.7 0.5 0.5 0.4 0.4
    - Japan: 1.4 0.7 0.6 0.9 0.6 0.6
    - Thailand: 2.5 1.4 0.7 0.5 0.6 0.4

### Box 2.1 — China: disconnect between PPI and CPI (high-level points)
- 2017: Chinese PPI surged above 5 percent after 54 consecutive months of PPI deflation since 2012; CPI remained around 2 percent.
- CPI composition: final consumption goods ≈ three-quarters of CPI basket; only one-quarter of PPI basket.
- PPI driven by intermediates; CPI driven by final consumption goods.
- Drivers of 2017 PPI reflation:
  - Expansion in domestic infrastructure and real estate investment since early 2016 supported PPI inflation of raw materials and intermediate manufacturing goods.
  - Rebound in advanced-economy demand and global trade in 2017 boosted China’s real exports of goods and their prices (intermediate manufacturing goods important in exports).
  - Supply restrictions in selected industries played a marginal role; broad reflation consistent with investment and foreign-demand shocks.
- Outlook: global environment supportive of Chinese PPI, but less favorable domestic investment and easing of supply restrictions point to gradual unwinding of producer price pressures.
- Data note: China does not publish CPI/PPI composition; figures rely on regression-estimated weights.

### Japan: prolonged low inflation, policy response, and outlook
- Facts:
  - Annual CPI inflation averaged 0.2 percent between 1993 and 2017.
  - GDP deflator contracted by an average of 0.3 percent a year over same period.
- Causes of deflationary episode included asset-bubble collapse, 1997–98 Asian crisis, banking crisis, deregulation increasing nonregular workers, aging labor force.
- Policy evolution:
  - 1999: zero interest rate policy; 2001: quantitative easing; early 2006: exit from QE (later judged premature).
  - Early 2013: Abenomics — explicit 2 percent inflation target and large JGB purchases, flexible fiscal policy, structural reforms.
  - CPI inflation reached 1.6 percent in 2013; later fell due to energy prices, global slowdown, and 2014 consumption tax hike.
  - Oct. 2014: increase in JGB purchase target (Figure note: Target raised to 80 trillion yen).
  - Early 2016: negative interest rate on excess reserves.
  - Sept. 2016: yield curve control (YCC) introduced with two commitments: expand monetary base until inflation overshoots target stably; shape entire yield curve considering financial conditions.
  - YCC allowed reduced JGB purchases and some rise in long-term yields.
- Current conditions and challenges (as of April 2018):
  - Eight consecutive quarters of above-potential growth; unemployment lowest since 1993; CPI headline inflation 1.1 percent in March 2018.
  - Core inflation remains stubbornly low; long-run expectations not convincingly moving to 2 percent; wage growth remains very low for regular workers.
- Policy recommendation: use favorable macro environment to implement comprehensive package including labor-market reforms and coordinated income and demand policies to durably unwind the deflation mindset.

### Policy implications and recommended actions
- Central banks should be vigilant and responsive to early signs of inflation pressure, including from global factors; sudden increases in inflation may persist and disinflating could be costly given a flatter Phillips curve and more backward-looking inflation process.
- Strengthen monetary policy frameworks and improve central bank communications to increase the role of expectations and keep expectations anchored to targets.
- Allow greater exchange-rate flexibility to mitigate imported-inflation transmission.
- Monetary policy response to commodity-price shocks: accommodate first-round CPI effects but not second-round effects; consider underlying as well as headline inflation given temporary nature of commodity shocks.
  - Exception: economies with limited central bank credibility and high commodity shares in CPI may need more aggressive responses when excess demand is high and inflation above target.
- With more backward-looking inflation and a flatter Phillips curve, the output costs of disinflating could be larger, requiring larger domestic-demand adjustments.

*Prepared by IMF staff; Regional Economic Outlook: Asia and Pacific, April 2018.*

### Introduction and Main Findings

### Introduction and Main Findings

### Overview
- Global growth in 2017 was the highest since 2011 and is expected to strengthen further in 2018–19, supported by broad-based momentum across countries and fiscal expansion in the United States.
- Headline inflation has been picking up with the upturn in oil prices since September, but core inflation remains surprisingly subdued, especially in advanced economies.
- Asia has been in a sweet spot of strong growth and benign inflation.
- While GDP growth forecasts for 2017–18 have been repeatedly revised up over the past two years, inflation forecasts have been kept constant or revised down.
- Core inflation remains below inflation targets in many Asian economies.

### Main analytical questions
- Why has inflation been low in Asia recently, and how long will it last?
- What has been the role of import prices and global factors?
- How well anchored are inflation expectations?
- To what extent has inflation become less sensitive to economic slack?
- How do these drivers of inflation in Asia differ from those in other regions?
- What are the key implications for policymakers?

### Methods
- Estimation of augmented Phillips curves.
- Principal component analysis to distinguish global factors from country-specific factors.
- Analysis of trend inflation to assess persistence of low inflation.

### Key findings
- Recent low inflation has been driven mainly by temporary forces, including imported inflation.
  - Phillips curve estimation indicates weaker import prices, including low commodity prices, contributed to half of the undershooting of inflation targets in advanced Asia and most of the undershooting in emerging Asia in recent years.
  - China seems to have played an important role in driving both global and regional inflation.
  - An analysis separating temporary and trend components suggests temporary shocks have accounted for the bulk of the recent reduction in inflation.
- The inflation process has become more backward-looking.
  - Inflation expectations are generally relatively well anchored, especially in advanced Asia and in economies with inflation-targeting frameworks.
  - The importance of expectations in driving inflation has declined in recent years, with past inflation playing a larger role.
- The sensitivity of inflation to the unemployment gap has declined.
  - There appears to be a flattening of the Phillips curve compared with the 1990s in advanced Asia and a similar but more continuous flattening in emerging Asia.
  - Outside Asia, the slope of the Phillips curve seems to have been more stable.
- The coefficient on imported inflation remained broadly constant in both advanced economies and emerging markets (detailed estimates in chapter annexes).

### Recent inflation trends in Asia
- Headline inflation declined sharply during 2012–15 across many advanced, emerging market, and developing economies in Asia.
- Disinflation was broad-based across sectors and inflation measures; decline in inflation was comparable across food, other goods, and services on average.
- Food inflation declined the most across advanced economies but remained generally positive, suggesting low pass-through from international to domestic food prices.
- Other goods inflation entered negative territory in several advanced Asian economies, reflecting large decline in manufacturing producer prices.
- Core inflation and wage inflation also declined widely during the disinflation period.
- Headline inflation started to pick up in 2016, with the share of low-inflation economies dropping from its 2015 peak.
  - Inflation is picking up in Australia, Japan, Korea, and some ASEAN economies.
  - Recent pickup in headline inflation is primarily explained by other goods and services inflation, with the manufacturing producer price index (PPI) recovering strongly.
  - The pickup is in line with other advanced economies and emerging markets, reflecting the recent increase in commodity prices.
  - In China, the pickup in PPI inflation did not spill over to CPI inflation.
- Despite the pickup, the level of inflation is still low in many economies, with headline and core inflation below inflation targets in most economies.

### Structural drivers of inflation
- Framework builds on hybrid new Keynesian Phillips curve (Fuhrer 1995; Galí and Gertler 1999) relating domestic inflation to inflation expectations, cyclical unemployment, and imported inflation.
- Inflation expectations came down substantially in the 1990s for both advanced and emerging Asia and have been on a declining trend since 2011.
- There was some labor market slack after the global financial crisis, but unemployment gaps seem to be closing recently, with some slack still in China, India, Korea, Malaysia, and Singapore.

- First: Inflation has become more sticky (backward-looking) since the global financial crisis.
  - The estimated coefficient on inflation expectations has declined in both advanced economies and emerging markets back to levels of the early 2000s, suggesting inflation is being driven more by past inflation than by expectations.
  - In advanced Asia, the role of long-term inflation expectations was less important than in other advanced economies but has gradually caught up; since the global financial crisis the coefficient has declined in line with other advanced economies.
  - In emerging Asia, the role of inflation expectations in driving inflation has generally been more important than in other emerging markets; the coefficient tends to increase after adoption of inflation targeting, and any decline since the global financial crisis has been smaller and with a longer lag compared with advanced economies.
  - A lower coefficient on expectations implies effects of cyclical unemployment, import prices, and shocks on inflation have become relatively more persistent.
- Second: The slope of the Phillips curve (sensitivity of inflation to labor market slack) has declined in Asia.
  - In Asian advanced economies, median coefficients declined from about 1 in the 1990s to about 0.6, while other advanced economies had median coefficients of about 0.2 in the 1990s.
  - In Asian emerging markets, the slope has been declining steadily; the flattening has been significant and prevalent across most Asian emerging markets.
  - This result is robust to alternative measures of slack (NAIRU, capacity utilization, and some output-gap estimates).
- Third: The coefficient on imported inflation has remained broadly constant in both advanced economies and emerging markets.

### Outlook and implications
- Inflation may rise in Asia as commodity prices and other temporary factors reverse (the April 2018 World Economic Outlook projects a near-term increase in commodity prices).
- Higher inflation in the rest of the world and weaker currencies in the region could pose upside risks to inflation.
- If such risks materialize, higher inflation may persist given the stickiness of the inflation process.
- Given the relative flatness of the Phillips curve, the output costs of disinflating may be high.

### Policy implications
- Central banks should be vigilant in responding to early signs of inflation pressure, including from global factors; a sudden increase in inflation may persist, and disinflating may be costly if sensitivity of inflation to the unemployment gap has declined.
- Strengthen monetary policy frameworks and improve central bank communications to increase the role of expectations in driving inflation and to maintain expectations anchored to targets.
- To mitigate the role of imported inflation, exchange rates should be allowed to adjust more flexibly.
- In principle, the monetary policy response to commodity price shocks should be to accommodate first-round effects but not second-round effects.

*Prepared by Pablo Lopez Murphy (lead), Juan Angel Garcia Morales, Weicheng Lian, Katsiaryna Svirydzenka, Rizki Wimanda, and Qianqian Zhang, under the guidance of Koshy Mathai.*

### 1. Share of Inflation-Targeting Countries

### 1. Share of Inflation-Targeting Countries

### Overview
- Figures show the share of inflation-targeting (IT) countries and the expectations coefficient before and after IT adoption, with regional medians and interquartile ranges for Advanced Asia, Emerging Asia, other advanced economies, and other emerging markets.  
- Source: IMF staff estimates. Note: EM = emerging market.

### Expectations Coefficient: Before and After Inflation Targeting
- Average coefficient in the three years before IT and average coefficient in the three years after IT are presented for individual economies (China, Indonesia, India, Philippines, Thailand) and aggregates (Asian EM, Other EM).  
- Adoption of inflation targeting is associated with a reduction in the sensitivity of inflation expectations to inflation shocks when comparing the three years before and after IT for the sampled economies.

### Contributions to Inflation Deviation from Targets (Figures 2.16 — Advanced Asia and Emerging Asia averages)
- The analysis decomposes inflation deviations from targets into contributions from:
  - Relative inflation of import price in US dollars
  - Inflation expectation minus target
  - Cyclical unemployment
  - Exchange rate
  - Others
- Advanced Asia Average (2000–07; 2008–12; 13–14; 15; 16; 17:Q1–Q2): charts show movements in contributions, with recent undershooting driven largely by import prices and by inflation expectations starting in 2016–2017.
- Emerging Asia Average (same periods): charts show import prices as the main contributor to undershooting since 2014, with cyclical unemployment also an important factor in emerging markets.

### Key quantitative findings on drivers (textual results)
- Imported inflation has been the most important driver of inflation deviations from targets in recent years (Figure 2.16).
  - In advanced economies: cyclical unemployment was the main driver before the global financial crisis; import prices accounted for about half of the undershooting since 2013.
  - In emerging markets: undershooting of inflation targets since 2014 is mainly accounted for by import prices; cyclical unemployment has also contributed to undershooting.
  - Starting in 2016 and 2017, inflation expectations contributed to undershooting: accounting for 35 percent of the undershooting in advanced Asia and 40 percent of the undershooting in emerging Asia in the first two quarters of 2017.
  - The depreciation of Asian currencies in 2015 following China-induced volatility contributed positively to inflation in both advanced and emerging Asia.
- There is a large share of unexplained factors driving inflation undershooting in Asia, especially in advanced economies. Possible explanations include mismeasurement of labor market slack, expectations of actual price setters dropping more than professional forecasters, and the zero lower bound on policy rates.

### Relation to manufacturing slack in major economies (Figure 2.17)
- The undershooting of inflation targets due to import prices seems linked to manufacturing slack in China.
- Country-level regressions show the import price contribution to inflation in Asia is particularly strongly related to slack in China (correlations presented for China slack, Japan slack, United States slack).

### Flattening of the Phillips Curve: Causes and Evidence (Figures 2.18)
- Inflation has become less sensitive to the unemployment gap; possible channels:
  - Institutional factors: union density, collective bargaining, labor laws, employment regulations.
  - Structural factors: aging, technological progress (automation), global integration, rise of the service economy.
- Empirical links found:
  - In advanced Asia: flattening of the Phillips curve is statistically significantly linked to automation (measured by the relative price of investment goods).
    - Panel regression reported: slope relationship expressed as ŷ_t = –1.865*** + Pinv × 3.173***; Time and country fixed effects; R^2 = 0.647 (standard errors shown in parentheses in source).
  - In emerging Asia: flattening of the Phillips curve is statistically significantly linked to GVC integration (measured by the share of foreign value-added in exports).
    - Panel regression reported: ŷ_t = 1.077*** + Foreign VA × (–0.0267***); Time and country fixed effects; R^2 = 0.66 (standard errors shown in parentheses in source).
- Interpretation:
  - Automation can substitute for routine tasks, weakening workers’ bargaining power and unionization.
  - Greater exposure to international competition and GVCs weakens labor’s bargaining position.
  - The Phillips curve flattened more when Asia experienced rapid increases in GVC integration and automation; as these factors stabilize, the slope may normalize.

### How Well Anchored Are Inflation Expectations?
- Medium-term inflation expectations in Asia have declined since 2011 and started to contribute to undershooting of targets recently.
- Two approaches used to assess anchoring:
  1. Fraction of time expectations are within inflation targets (Consensus Forecasts for different horizons).
  2. Response of expectations to inflation shocks (estimated β_h in equation relating changes in h-period-ahead expectations to inflation surprises).
- Key empirical outcomes:
  - Expectations are generally better anchored at longer horizons (one-year-ahead to five-year-ahead) than for the current year.
  - One-year-ahead sensitivity example: a 1 percentage point increase in inflation leads to a 0.32 percentage point increase in inflation expectations one year ahead in emerging markets, and a 0.26 percentage point increase in advanced economies.
  - Inflation expectations are more sensitive to inflation shocks in emerging markets than in advanced economies, especially over shorter horizons.
  - Inflation expectations have become better anchored over time for both advanced and emerging Asia (Figure 2.20).
  - Adopting inflation targeting tends to reduce the sensitivity of expectations to inflation shocks (Figure 2.21). Six economies analyzed: New Zealand (March 1990), Australia (July 1993), Korea (January 1998), the Philippines (January 2002), Thailand (May 2002), and Indonesia (July 2005). Results suggest smaller sensitivity after adopting IT and a narrower distribution of inflation outcomes.
  - Overall conclusion: no strong evidence that inflation expectations are becoming unanchored; expectations are generally relatively well anchored to targets.

### Tables: Percent of Time Expectations within Target Range and Average Absolute Deviations (Preserved values)
- Table 2.1. Percent of Time Expectations are within Inflation Targeting Range, Inflation Targeting start till 2017
  - Columns: Expectation Horizon (years) 0 1 2 3 4 5
  - Australia (Start 1993, Most Recent Target Band 2–3.5): 47 07 68 28 68 88
  - China (Start 2005, Most Recent Target Band 2–4): 07 99 29 21 00 96
  - Indonesia (Start 2001, Most Recent Target Band 2.5–4.5): 47 54 40 35 54 50
  - India (Start 2016, Most Recent Target Band 2–6): 10 01 00 10 01 00
  - Japan (Start 2013, Most Recent Target Band 1–3): 25 65 80 80 80 90
  - Korea (Start 1998, Most Recent Target Band 1.5–2.5): 48 86 90 90 80 78
  - New Zealand (Start 1990*, Most Recent Target Band 1–3): 73 91 10 01 00 10 0 0 0 0 0 0  (evaluation begins at 1995 due to availability of inflation expectations data)*
  - Philippines (Start 2002**, Most Recent Target Band 2–4): 58 87 73 58 45 35  (evaluation begins at 2008 due to availability of inflation expectations data)**
  - Thailand (Start 2000, Most Recent Target Band 1–4): 50 72 83 83 86 89

- Table 2.2. Percent of Time Expectations are within Inflation Targeting Range, 2013–17
  - Australia: 100 100 100 100 100 (horizons 0–4)
  - China: 45 75 100 100 100
  - Indonesia: 50 65 35 10 25 20
  - India: 10 01 00 10 01 00
  - Japan: 25 65 80 80 80 90
  - Korea: 20 75 80 60 60 80
  - New Zealand: 100 100 100 100 100
  - Philippines: 40 95 75 65 60 60
  - Thailand: 90 100 100 100 100

- Table 2.3. Average Absolute Deviation from Target (from Beginning of Inflation Targeting until 2017)
  - Australia (Start 1993, Point Target 1.5): 0.7 0.5 0.4 0.3 0.2 (horizons 0–4)
  - China (Start 2005, Point Target 3): 1.3 0.6 0.6 0.5 0.3 0.4
  - Indonesia (Start 2001, Point Target 3.5): 1.7 1.1 1.1 1.2 1.1 0.9 1.0
  - India (Start 2016, Point Target 4): 0.9 1.0 1.2 1.0 0.9 0.9
  - Japan (Start 2000, Point Target 2): 1.4 0.7 0.6 0.9 0.6 0.6
  - Korea (Start 1998, Point Target 2): 0.7 0.4 0.3 0.3 0.3 0.4
  - New Zealand (Start 1990*): 0.8 0.4 0.3 0.3 0.3 0.3
  - Philippines (Start 2002**): 0.5 0.4 0.7 0.9 1.0 1.1
  - Thailand (Start 2000, Point Target 2.5): 1.4 0.9 0.9 1.0 1.0 0.9

- Table 2.4. Average Absolute Deviation from Target, 2013–17
  - Australia: 0.4 0.2 0.2 0.1 0.1
  - China: 1.0 0.7 0.5 0.5 0.4 0.4
  - Indonesia: 1.3 1.0 1.0 1.0 1.3 1.3 1.4
  - India: 0.9 1.0 1.2 1.0 0.9 0.9
  - Japan: 1.4 0.7 0.6 0.9 0.6 0.6
  - Korea: 1.1 0.3 0.2 0.5 0.4 0.3
  - New Zealand: 0.7 0.2 0.2 0.2 0.2 0.1
  - Philippines: 0.6 0.4 0.6 0.7 0.7 0.8
  - Thailand: 2.5 1.4 0.7 0.5 0.6 0.4

### Policy-relevant implications and conclusions
- Low inflation in Asia has been driven mainly by sluggish import prices and inflation expectations being below targets.
- The Phillips curve fits Asian inflation data but has flattened: the sensitivity of inflation to economic slack has declined.
- Links between manufacturing slack in China and import-price-driven inflation undershooting suggest that stronger growth in China in 2018 and 2019 could put upward pressure on regional inflation.
- Flattening of the Phillips curve is associated with automation in advanced Asia and GVC integration in emerging Asia; as automation and GVC integration stabilize, the Phillips curve slope may normalize.
- Inflation expectations remain relatively well anchored overall; adoption of inflation targeting appears to improve anchoring when coupled with increased central bank transparency.
- Remaining risks: large unexplained factors in inflation undershooting, potential mismeasurement of labor market slack, differences between expectations of price setters and professional forecasters, and constraints such as the zero lower bound.

*Source: IMF staff estimates, Regional Economic Outlook: Asia and Pacific, April 2018.*

### 1. Advanced Asia: Time-Varying Parameter 3–5 years2. Emerging Asia: Time-Varying Parameter 3–5 years

### apd-reo-april-2018-chapter-2 - 1. Advanced Asia: Time-Varying Parameter 3–5 years2. Emerging Asia: Time-Varying Parameter 3–5 years

### Global Factors
- A latent factor model applied to cross-country inflation data over the past 15 years finds three global factors explain 56 percent of the variance in headline inflation in the full cross-country panel sample.
- Share of total variance explained by factors:
  - Factor 1: 33%
  - Factor 2: 15%
  - Factor 3: 8%
- Factor 1 correlates with global commodity prices:
  - ρ commodity all = 0.52
  - ρ fuel = 0.56
  - ρ nonfuel = 0.28
- Factor 3 appears associated with movements in the nominal effective exchange rate (NEER) of the US dollar:
  - AE, EM, LIC sample: ρ NEER = 0.64
  - AE, EM sample: ρ NEER = 0.19
  - AE, EM, LIC sample: ρ NEER = –0.06 (noted elsewhere)
  - AE, EM sample: ρ NEER = 0.58
  - 2001Q1 - 2014Q2: 0.52 (context for a correlation)
  - 2014Q3 - 2017Q3: –0.28 (context for breakdown)
  - 2001Q1 - 2014Q2: 0.58 (context for another correlation)
  - 2014Q3 - 2017Q3: 0.91 (post-2014 increase in correlation with Chinese PPI)

### Role of China in Global and Regional Inflation
- Correlations with Chinese PPI:
  - Correlation of Chinese PPI with Factor 1: 0.34
  - Correlation of Factor 3 in the AE and EM sample with Chinese PPI: 0.59 (overall)
  - Correlation of Factor 3 with Chinese PPI in the post-2014 period: 0.91
- Regional residual analysis for Asian economies:
  - A second principal component analysis on residuals yields a first common regional factor that explains 50 percent of the residual variation in inflation.
  - Correlation of this regional factor with Chinese PPI: 0.83
- Implication: China could affect global inflation indirectly via its impact on commodity prices and plays a significant role in driving inflation at the regional level.

### Common Factors and Cross-Country Loadings
- In many economies, more than 50 percent of the variation in inflation is explained by common factors.
- Advanced Asia is broadly in line with advanced economies in other regions in terms of factor loadings; Emerging Asia seems more responsive to Factor 1 than Latin America.
- Note: AE = advanced economy; EM = emerging market economy; LIC = low-income country; NEER = nominal effective exchange rate; PCA = principal components analysis; PPI = producer price index.

### Trend Inflation and Transitory Components
- Univariate time series decomposition: inflation = permanent component (trend) + temporary component.
- Trend inflation trajectory:
  - Trend inflation has been declining in both advanced and emerging Asia:
    - Came down substantially over the 1990s.
    - Experienced another decrease, although milder, after 2011.
- Transitory component:
  - The decomposition shows a concentration of transitory disinflationary shocks in the region over the past few years.
  - The transitory component of inflation has been predominantly negative for most economies since 2014.
  - Going forward, positive transitory shocks could lift inflation more quickly than expected.
- Nonlinearities in the Phillips curve could lead to a jump in inflation at higher inflation levels (reference to Guimaraes and Papi (2016) finding greater sensitivity of inflation to the output gap at higher inflation levels).

### Policy Conclusions and Implications
- The decline during 2012–15 and recent uptick in Asian inflation largely followed the global pattern and reflects exposure to commodity price cycles and global competition; inflation fluctuations are mainly driven by prices of imported goods.
- Risks and policy recommendations:
  - Asia is likely to see rising inflation when inflation in the United States and commodity prices rise.
  - The inflation process has become more backward-looking since the global financial crisis, implying a risk of persistent deviations from targets and benefits from pursuing a clear, well-communicated policy reaction function.
  - Evidence suggests inflation expectations are better anchored in Asian economies that adopted inflation-targeting frameworks.
  - Phillips curves in Asia have flattened, implying a higher real cost for reducing inflation.
  - Central banks should be vigilant about imported inflation when setting policy reaction functions to avoid sustained deviations from inflation targets.
  - With more backward-looking inflation and a flatter Phillips curve, the costs of disinflating could be larger, requiring larger changes in domestic demand to bring inflation back to target.
  - Higher exchange rate flexibility will better insulate domestic inflation from imported inflation; an appreciation of Asian currencies vis-à-vis the US dollar will help preserve monetary policy autonomy and keep higher US-origin inflation pressures under control.

### Response to Commodity Price Shocks
- Appropriate central bank response to a commodity price shock:
  - Accommodate the first-round effects on the CPI but not the second-round effects on other CPI components.
  - This approach lowers output volatility (Chapter 3 of the September 2011 World Economic Outlook referenced).
  - Since commodity price shocks are typically temporary, central banks should consider underlying as well as headline inflation in policymaking.
  - Exception: In economies with limited central bank credibility and a high share of commodity prices in the CPI, commodity price shocks are likely to have larger second-round effects and may require a more aggressive policy response when excess demand pressures are high and inflation is running above target.

### Box 2.1 — The Disconnect between the Consumer and Producer Price Indices in China
- Observation:
  - In 2017, Chinese producer prices (PPI) soared while consumer prices (CPI) remained subdued.
  - PPI had been in deflation for 54 consecutive months starting in 2012, while CPI hovered around 2 percent; PPI rebounded since November 2016 and remained above 5 percent.
- CPI vs PPI composition:
  - Final consumption goods make up three-quarters of the CPI basket and only one-quarter of the PPI basket.
  - For PPI, intermediates drive the index; for CPI, final consumption goods drive the index.
- Drivers of PPI inflation in 2017:
  - Domestic investment demand:
    - Expansion in domestic infrastructure and real estate investment since early 2016 supported growth and co-moved with PPI inflation of raw materials and intermediate manufacturing goods.
  - Rebound in advanced economies’ demand:
    - Global trade rebounded strongly in 2017, led by final demand in advanced economies.
    - China’s real exports of goods surged and their prices rose; intermediate manufacturing goods are important Chinese exports and their prices co-move strongly with overall goods export prices.
  - Supply restrictions in selected industries in China did not play a lead role in broad-based reflation of producer prices.
- Outlook for Chinese PPI:
  - While the global environment will remain supportive of Chinese producer prices, a less favorable outlook for domestic investment and less stringent supply restrictions will lead to a gradual unwinding of producer price pressures.
- Note on data:
  - China does not publish the composition of its CPI and PPI baskets; figures rely on regression-estimated weights.
  - Within intermediates for PPI there are three categories: mining and quarrying goods, raw materials, and intermediate manufacturing goods.

*Source: IMF staff estimates. International Monetary Fund | April 2018*

### Box 2.1 (continued)Box 2.1 (continued)

### Box 2.1 (continued)

### PPI inflation dynamics and China’s role
- Higher domestic investment and foreign demand led to surging imports of raw materials, which further supported PPI inflation.
- China depends on imports of raw materials to satisfy its investment demand and as the input into its manufacturing exports; price pressures in investment and intermediate manufacturing goods for export translate into higher prices for raw materials and overall import prices.
- The energy component of the CPI is tightly linked with prices of raw materials, but its weight is small, and thus this link is not discernible in the headline CPI numbers.
- Some of the increase in commodity prices cannot be attributed solely to international and Chinese demand; markets such as that for crude oil are also heavily influenced by supply constraints.

### Supply-side adjustments in China’s upstream industries
- China started a campaign in 2016 to shut down overcapacity in the steel and coal industries.
- These supply-side measures may have played a marginal role in supporting PPI inflation but cannot explain the broad-based reflation that is more consistent with the investment and foreign-demand shocks.
- Actual production of steel throughout the period increased.
- Coal production declined in 2016, then was on the rise in 2017 until the government initiated an intensified crackdown on polluting industries ahead of the 2017–18 winter.
- The 2017–18 intensified campaign may have contributed to sustaining price pressures, although it is hard to assess its ultimate effects because the authorities suspended production of intermediate goods while also suspending construction projects that demand the same input.

### Japan’s prolonged low inflation: facts and background
- Annual consumer price index (CPI) inflation averaged 0.2 percent between 1993 and 2017.
- The GDP deflator contracted by an average of 0.3 percent a year during the same period.
- Multiple demand shocks in the 1990s contributed to Japan’s transition into deflation:
  - The collapse of the asset price bubble in the early 1990s led to deleveraging by households, banks, and businesses, slowing the real economy, raising unemployment, and lowering inflation.
  - The 1997–98 Asian crisis further weakened demand.
  - High levels of nonperforming loans resulted in a banking crisis that pushed the economy into deflation in 1998–99.
- Supply-side factors exacerbated deflation pressure:
  - Deregulation of the labor market in the mid-1990s and early 2000s led to a sharp rise in the share of nonregular workers, lowering labor’s bargaining power and exerting downward pressure on prices by lowering unit labor costs.
  - Aging and shrinking of Japan’s labor force intensified in the 1990s, adversely affecting potential growth and fiscal sustainability, negatively affecting permanent income and potentially boosting precautionary savings.

### Policy response and the emergence of the deflation mindset
- Monetary policy response was characterized as “too little, too late”; fiscal policy proved ineffective in stimulating growth.
- Timeline of key Bank of Japan actions:
  - 1999: adopted its “zero interest rate policy.”
  - 2001: switched to quantitative easing.
  - Early 2006: exit from quantitative easing and increase in the policy rate from zero (later judged probably premature).
- Fiscal policy remained broadly accommodative but periodic attempts at consolidation led to stop-and-go implementation, reducing effectiveness.
- The prolonged period of low inflation resulted in a gradual decline in inflation expectations and the emergence of the “deflation mindset.”
  - With the nominal interest rate at the zero lower bound, the Bank of Japan’s ability to lower the real interest rate and generate demand-driven inflation was significantly constrained.
  - Low inflation expectations contributed to less ambitious wage demands by unions; firms became reluctant to adjust prices in response to economic conditions.

### Abenomics and subsequent monetary innovations
- Early 2013: Japanese authorities shifted to decisively lift the economy out of deflation—Bank of Japan announced an explicit inflation target of 2 percent and significantly ramped up Japanese government bond (JGB) purchases; these actions were complemented by flexible fiscal policy and a commitment to structural reforms.
- Initial impact:
  - CPI inflation reached 1.6 percent in 2013.
  - Inflation expectations started to gradually rise.
- Later developments:
  - Sharp drop in energy prices, slowdown in global growth, and the 2014 consumption tax hike caused Japanese inflation to drop again.
  - October 2014: Bank of Japan provided additional stimulus by boosting JGB purchases (Figure notes: Oct. 2014: Target raised to 80 trillion yen).
  - Early 2016: Bank of Japan implemented a negative interest rate on excess marginal reserves.
  - Concerns emerged that the Bank of Japan would run out of JGBs to purchase and that flattening of the yield curve could impair financial intermediation by depressing profitability.
- September 2016: Bank of Japan introduced its yield curve control (YCC) framework with two main commitments:
  1. To keep expanding the monetary base until inflation overshoots its target in a stable manner.
  2. Appropriately shaping the entire yield curve to achieve price stability while considering financial conditions.
- Effects of YCC:
  - By deemphasizing quantities in favor of a yield target, the Bank of Japan has been able to reduce JGB purchases and alleviate some concerns over policy sustainability.
  - Increased long-term yields have helped alleviate pressures on institutional investors (for example, pension funds and insurers).

### Looking forward: prospects for reflation and policy implications
- Recent macroeconomic indicators:
  - The Japanese economy experienced eight consecutive quarters of above-potential growth.
  - The unemployment rate is at its lowest level since 1993.
  - CPI headline inflation reached 1.1 percent in March 2018.
- Remaining challenges:
  - Core inflation remains stubbornly, and crucially low.
  - Long-run inflation expectations have yet to display convincing signs of moving toward the 2 percent target.
  - Wage growth remains very low, particularly for regular workers.
  - Anecdotal evidence suggests firms are reluctant to pass higher labor costs on to prices, instead resorting to labor rationing and investment in labor-saving technologies.
- Bank of Japan stance in 2018:
  - Taking a patient approach, viewing that labor shortages will build enough pressure to force firms to increase prices at a higher rate.
  - With inflation expectations slow to adjust, this process may take time, and the monetary policy cost-benefit trade-off may change as other major central banks begin to normalize policy.
- Policy recommendation:
  - It is imperative to take advantage of the current favorable macroeconomic environment to implement a comprehensive policy package that exploits complementarities between labor market reforms and coordinated income and demand policies, so that the Japanese deflation mind-set can be durably unwound.

### Annex 2.1: New Keynesian Phillips Curve (model specification)
- The new Keynesian Phillips curve relates domestic inflation to inflation expectations, cyclical unemployment, and imported inflation:
  π_t = a_t π^e_{t+10} + (1 − a_t) π̂_{t−1} + β_t u^c_t + γ_t π^m_t + ε_t
- Definitions and notes:
  - π_t is headline consumer price inflation.
  - π^e_{t+10} is inflation expectations 10 years ahead. (When such measures are not available, five-year World Economic Outlook forecasts of headline inflation are used.)
  - π̂_{t−1} is the moving average of inflation in the previous four quarters.
  - u^c_t is cyclical unemployment, estimated as the deviation of the unemployment rate from the nonaccelerating inflation rate of unemployment.
  - π^m_t is inflation in the relative price of imports (the import-price deflator relative to the GDP deflator).
  - ε_t is the impact of other factors, including measurement error and supply shocks.
  - a_t captures the degree to which inflation is driven by long-term inflation expectations as opposed to lagged inflation.
  - β_t denotes the slope of the Phillips curve (sensitivity of inflation to cyclical unemployment).
  - γ_t captures the impact of imported inflation.
- Estimation:
  - Equation (A2.1.1) is estimated at the country level for 44 advanced and emerging market economies, including 13 Asian economies, using quarterly data for the first quarter of 1990 to the second quarter of 2017.
  - The estimation method is maximum likelihood based on a constrained nonlinear Kalman filter that allows for time variation in the regression coefficients.

### Annex 2.2: Common Factors (latent-factor model)
- Purpose: gauge the role of global factors in driving inflation dynamics using a latent-factor model applied to the inflation rates of 136 advanced, emerging market, and low-income economies from the first quarter of 2001 to the third quarter of 2017.
- Model structure:
  - Each economy’s demeaned total inflation (π_t − π̄_t) can be decomposed into a common component (β f_t) and an idiosyncratic component (e_t): π_t − π̄_t = β f_t + e_t.
  - f is a (K × 1) vector of latent factors; β is an (N × K) matrix of loading coefficients.
  - Inflation is measured as the year-over-year percent change in the headline consumer price index.
  - The model is estimated with principal component analysis.
- Interpretation:
  - A higher loading factor means that the country’s inflation is affected more strongly by that common factor.
  - Cross-country differences in inflation reflect country-specific factors and different sensitivity to global factors (different loading factors).

### Annex 2.3: Trend Inflation (unobserved-components framework)
- Decomposition:
  - π_t = π^*_t + c_t, where π^*_t is trend inflation and c_t is deviation from trend.
- Identification assumptions:
  - lim_{j→∞} E_t(π_{t+1j}) = E_t(π^*_{t+1j}) = π^*_t
  - lim_{j→∞} E_t(c_{t+1j}) = 0
- Trend inflation reflects the most likely inflation rate once transitory influences die off.
- Estimation:
  - Trend inflation is estimated within an unobserved component and stochastic volatility model with data on headline inflation for six advanced economies (Australia, Hong Kong SAR, Japan, Korea, New Zealand, Taiwan Province of China) and six emerging markets (China, India, Indonesia, Malaysia, Philippines, Thailand).
  - Survey measures of long-term inflation expectations (6–10 years ahead) from the Consensus Forecasts are used as additional information to improve estimation.

*Prepared by IMF staff; Box prepared by Niklas Johan Westelius.*

---


_Source: https://www.imf.org/-/media/files/publications/reo/apd/2018/apd-reo-april-2018-chapter-2.pdf_