## 1. The Rise in Corporate Saving in Japan

## Source details

**Canonical URL:** [1. The Rise in Corporate Saving in Japan](https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021026-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2021/english/wpiea2021026-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2021/english/wpiea2021026-print-pdf.pdf.json)

---

### I. Introduction and research questions
- Headline current account balance has fluctuated around 3 percent of GDP (except after the 2011 Fukushima accident).
- Paper objectives:
  - Study main factors driving Japan’s growing income balance.
  - Provide first estimates of the responsiveness of the income balance to the real effective exchange rate (REER).
- Key high-level findings:
  - Japan’s income balance has grown along with increasing corporate saving.
  - Japan’s high income balance reflects asymmetry between income credits and debits, underpinned by high FDI yields on investment abroad, very low FDI liabilities, and low yields on portfolio debt liabilities.
  - Trade and income balances exhibit offsetting patterns; income responses to REER movements partially reinforce trade responses for large net creditor countries such as Japan.

### II. Stylized facts: drivers and magnitudes
- Trade and income trends:
  - Trade balance averaged 1.3 percent of GDP over 1996-2000, but only -0.3 percent over 2014-18.
  - Primary income balance rose from 1.3 percent of GDP (1996-2000 average) to 3.8 percent (2014-18 average).
  - Secondary income balance remained around -0.2 percent of GDP over the whole period.
  - Japan’s NFA reached 60 percent of GDP in 2018; Japan’s present NFA is stated as $3 trillion.
  - Gross income flows relative to gross trade flows: ratio increased from 21 percent in 1996 to 25 percent in 2018.
- Mechanisms and sectoral origins:
  - Increase in the income balance primarily reflects net revenues from an increasingly positive NFA position; primary income is composed almost exclusively of investment income.
  - Asymmetric expansion of income credits versus debits explains the large income balance; income credits exceed debits especially in direct investment income and portfolio income.
  - The increase since the mid-1990s is mostly attributable to the corporate sector and linked to rising corporate saving; household and government net property income remained broadly flat.

### III. The rise in corporate saving: drivers and correlations
- Stylized facts on corporate saving:
  - Corporate net saving (gross saving minus investment) increased significantly after the early-1990s real estate bubble; rise mostly driven by non-financial firms.
  - Japanese banks acted as intermediaries when investing abroad in search for yield.
  - Households exhibited dissaving and government net saving fell due to fiscal stimulus and higher social security spending.
- Main drivers:
  - Rise in net property income (concentrated in the 1990s), related to corporate deleveraging and lower interest rates.
  - Fall in the labor share from the late 1990s to the mid-2000s after 1995-96 labor market reforms expanding non-regular employment.
  - Investment did not contribute markedly to changes in net corporate saving, abstracting from bubble-era cyclical developments.
- Distributional link:
  - Rise in corporate saving is closely linked to top-income inequality; both series rose sharply from the early 1990s to the mid-2000s.
  - Possible explanations for household offsetting behavior vs. Germany include demographics and credit access: credit to the private sector was 107 percent of GDP in 2017 in Japan vs. 77 percent in Germany.

### IV. How Japan compares to peers on income and trade balances
- Cross-country position:
  - Japan’s current account is larger than average mainly because of an exceptionally large primary income balance; its trade balance is close to the median among G7 or a 100-country sample.
  - Japan’s gross income flows (credit + debit) are relatively modest and lower than other G7 economies but highly asymmetric (credits much larger than debits).
  - Japan receives more net income from its NFA than the cross-country NFA–investment income relationship would predict.
- Quantitative comparisons (2015–17 averages where stated):
  - Difference in investment income balance between Japan and G6 equals 3.5 percentage points of GDP.

### V. Decomposition: stocks versus yields and implications
- Decomposition framework (per asset/liability category k):
  - Contribution of stock positions: (S_Japan^k − S_G6^k) × (Y_Japan^k + Y_G6^k)/2
  - Contribution of implicit yields: (Y_Japan^k − Y_G6^k) × (S_Japan^k + S_G6^k)/2
- Key decomposition results (2015–17 average):
  - On the asset side, Japan holds fewer foreign assets than G6, but yields on those assets are higher.
  - Foreign assets explain only 0.3 percentage point of the 3.5 percentage point GDP difference.
  - On the liabilities side, Japan’s smaller stock of foreign liabilities—especially FDI but also portfolio debt—explains 3.2 out of the 3.5 percentage point GDP difference.
- Annex table exact figures (selected rows preserved as in source):
  - Implied yields quoted: 3.8%, 7.5%, 3.4%, 3.7%, 0.9%, 1.6%, 12.9%, 1.9%, 1.0%, 0.5%.
  - Income flows % of GDP (rows shown): 3.7 5.5 2.2 1.0 1.9 0.3 1.8 0.6 0.7 0.3 0.2.
  - Stocks % of GDP (rows shown): 62.1 146.2 29.6 30.5 51.3 34.7 109.5 5.0 35.3 26.2 43.0.
  - Additional table entries preserved exactly as presented in the source:
    - Income flows % of GDP: 0.2 5.2 2.6 1.0 1.0 0.6 5.0 1.6 1.0 1.8 0.6.
    - Stocks % of GDP: -4.5 225.1 60.0 48.9 41.0 75.2 231.8 48.5 34.1 73.4 75.9.
    - Income flows % of GDP: 3.5 0.3 -0.4 0.1 0.9 -0.3 -3.2 -0.9 -0.3 -1.5 -0.4.
    - Stocks % of GDP: 2.2 -2.3 -1.8 -0.5 0.3 -0.4 -4.5 -3.5 0.0 -0.8 -0.2.
    - Implied Yields % of GDP (table row): 1.4 2.7 1.4 0.6 0.6 0.1 1.3 2.6 -0.3 -0.7 -0.2.

### VI. Income balance responsiveness to exchange rates and interplay with trade
- Decomposition of REER effects:
  - Mechanical effect: driven by currency composition of assets/liabilities and related income credits/debits.
  - Economic effect: changes in domestic activity, profits, and outward spillovers altering income credits/debits.
- Empirical strategy:
  - Annual panel for 1999-2018, >40 countries.
  - Novel decomposition separating mechanical from economic income flows using currency-of-investment weighted (CIW) indices and panel estimation (CGER-inspired).
- Baseline elasticities (selected estimates preserved exactly):
  - Income credit elasticities (total): -0.56, -0.52.
  - Economic credit elasticities: -0.20, -0.22.
  - Income debit elasticities (total): -0.28, -0.24.
  - Economic debit elasticities: -0.14, -0.15.
  - Reported Log of REER coefficients include: -0.560***, -0.774***, -0.200**, -0.106, -0.279***, -0.474***, -0.141*, -0.0945.
  - Observations: 964 (or 944 for economic series); R-squared examples: 0.931, 0.932, 0.864, 0.922; RMSE examples: 0.213, 0.212, 0.326, 0.216.
- Interpretation:
  - Elasticities are larger for total income series (around -0.5 credit; -0.3 debit) and smaller when isolating the economic channel (around -0.2 credit; -0.1 debit), indicating a relatively strong mechanical channel.
- Japan-specific semi-elasticities:
  - Using panel elasticities and Japan’s income credit and debit ratios to GDP of 5.5 and 1.8 percent yields an income balance semi-elasticity of -0.026 (calculation preserved: [0.055*(−0.56) − 0.018*(−0.28) = −0.026]).
  - Japan-only panel estimates give a larger semi-elasticity of -0.073 (calculation preserved: [0.055*(−1.97) − 0.018*(−1.95) = −0.073]) but are discounted due to limited sample size (observations: 20).
  - By comparison, trade balance semi-elasticities in Cubeddu et al (2019) are reported as -0.12 to -0.14; hence for Japan the income balance response would marginally amplify the trade balance response to REER changes.

### VII. Annex findings on FDI geography, profit shifting, and measurement
- Geographical allocation of outward FDI (2013-17, ultimate investor, excluding SPEs) — Japan vs. G6 (unweighted average) (exact shares preserved):
  - Japan: Advanced Asia 11.4%, Emerging Asia 29.2%, Latin America & Caribbean 4.1%, CIS, CESEE, MENAP, SSA 2.4%, United States 10.3%, Euro Area 33.3%, Other advanced 9.2%.
  - G6: Advanced Asia 4.5%, Emerging Asia 5.4%, Latin America & Caribbean 7.5%, CIS, CESEE, MENAP, SSA 9.4%, United States 32.8%, Euro Area 24.4%, Other advanced 15.9%.
  - More favorable geographical allocation may explain around one percentage point of the higher yields in Japan than in G6 on average since 2000.
- Profit shifting and measurement:
  - Tørsløv, Wier and Zucman (2018) estimate $28 billion (0.6 percent of GDP) of profits shifted out of Japan in 2015.
  - If profit shifting mainly distorted direct investment income credits (not FDI asset positions), Japan’s “corrected” implied FDI yield would be around 2 percentage points lower.
  - Annex cautions high yields on foreign assets could reflect measurement issues.
- Liabilities and inward FDI:
  - Implicit yields paid on FDI liabilities are reported as very high in Japan (close to 13 percent in the source).
  - Adjusting implicit yields to G6 levels, while leaving FDI income payments unchanged, would change Japan’s stock of FDI liabilities from 5 to 20 percent of GDP—still well below the G6 average (~50 percent of GDP).
  - Low portfolio debt liabilities and low implicit yields reflect accommodative monetary policy, low credit risk, strong home bias, and corporate deleveraging.

### VIII. Interconnectedness of trade, income, and current account balances
- Correlations and mechanisms:
  - Current account (CA) is driven by saving-investment identity; CA highly and positively correlated with trade balance (Japan: 0.76 for 1980-2018).
  - Income balance is generally negatively correlated with trade balance (Japan: -0.58 for both total and primary income balances).
  - Mechanisms: aging, income effects on consumption/imports, market pressure, globalization/offshoring, and profit shifting/transfer pricing.
- Profit-shifting implications for CA-REER semi-elasticities:
  - Profit shifting can inflate trade balance semi-elasticities and alter income balance semi-elasticities; including income semi-elasticities improves reliability for tax-haven-like cases.

### IX. Policy implications and recommendations
- High-level policy implications:
  - Compositional shift toward a larger income balance, driven by corporate saving and asymmetric income flows, affects how Japan’s current account responds to exchange rate movements.
  - For large creditor economies like Japan, income account responses to REER appreciations can modestly reinforce trade responses; policymakers should consider both channels in external adjustment and exchange rate assessments.
- Steps to boost inward FDI and productivity (recommended in the source):
  - Address corporate governance issues (insider system limiting shareholder power).
  - Address regulatory and administrative issues to reduce the cost of doing business.
  - Reform product markets by reducing barriers to entry and accelerating deregulation of agricultural and professional services sectors (consistent with IMF advice referenced).
  - Note: higher inward FDI would likely raise income debits but the net effect on the income balance and current account is ambiguous due to indirect offsetting effects.
- Need for further research:
  - Better quantify profit shifting’s role and measurement distortions to inform taxation and policy design.

*Source – wpiea2021026-print-pdf: "1. The Rise in Corporate Saving in Japan" (excerpts provided).*

### 1. The Rise in Corporate Saving in Japan _________________________________________6

### 1. The Rise in Corporate Saving in Japan

### I. Introduction and research questions
- Headline current account balance has fluctuated around 3 percent of GDP (except after the 2011 Fukushima accident), while its composition shifted: the trade surplus has decreased and the income balance has increased.
- Paper objectives:
  - Study main factors driving Japan’s growing income balance.
  - Provide first estimates of the responsiveness of the income balance to the real effective exchange rate (REER).
- Key high-level findings:
  - Japan’s income balance has grown along with increasing corporate saving.
  - Japan’s high income balance reflects asymmetry between income credits and debits, underpinned by high FDI yields on investment abroad, very low FDI liabilities, and low yields on portfolio debt liabilities.
  - Trade and income balances exhibit offsetting patterns; income responses to REER movements partially reinforce trade responses for large net creditor countries such as Japan.

### II. Stylized facts: what has driven the increase in Japan’s income balance?
- Trends and magnitudes:
  - Trade balance averaged 1.3 percent of GDP over 1996-2000, but only -0.3 percent over 2014-18.
  - Primary income balance rose from 1.3 percent of GDP (1996-2000 average) to 3.8 percent (2014-18 average).
  - Secondary income balance remained around -0.2 percent of GDP over the whole period.
  - Japan’s NFA reached 60 percent of GDP in 2018; Japan’s present NFA is stated as $3 trillion.
  - Gross income flows relative to gross trade flows: ratio increased from 21 percent in 1996 to 25 percent in 2018.
- Mechanisms:
  - Income balance increase primarily reflects net revenues from an increasingly positive NFA position; the primary income balance is composed almost exclusively of investment income.
  - Gross income flows (credits + debits) and gross trade flows (exports + imports) have nearly doubled in 20 years, but the asymmetric expansion of income credits versus debits explains the large income balance.
  - Income credits are significantly larger than debits for most components, especially direct investment income and portfolio income.
- Sectoral origins:
  - The increase in the income balance since the mid-1990s is mostly attributable to the corporate sector and linked to rising corporate saving.
  - Corporate net property income and the investment income balance have shown parallel increases since the mid-1990s; household and government net property income have remained broadly flat.
  - Rising corporate net property income contributed to increased corporate saving and, via persistent current account surpluses, to the rise in NFA.

### III. The Rise in Corporate Saving (Box 1): drivers and correlations
- Facts:
  - Corporate net saving (gross saving minus investment) increased significantly after the burst of the real estate bubble in the early 1990s; the rise was mostly driven by non-financial firms.
  - Japanese banks acted as intermediaries when investing abroad in search for yield.
  - Offsetting movements occurred in other sectors: households exhibited dissaving and government net saving fell due to fiscal stimulus and higher social security spending.
- Main drivers of higher corporate saving:
  - Rise in net property income (concentrated already in the 1990s), related to corporate deleveraging and lower interest rates from looser monetary policy.
  - Fall in the labor share from the late 1990s to the mid-2000s following 1995-96 labor market reforms that expanded non-regular employment.
  - Investment did not contribute markedly to changes in net corporate saving, abstracting from bubble-era cyclical developments.
- Distributional link:
  - The rise in corporate saving appears closely linked to top-income inequality; both series show sharp increases from the early 1990s to the mid-2000s.
  - Possible explanations for why households offset higher corporate saving in Japan (but not in Germany) include demographics (more advanced aging in Japan) and easier access to credit in Japan (credit to the private sector was 107 percent of GDP in 2017, compared to 77 percent in Germany).

### IV. How Japan compares to peers on income and trade balances
- Cross-country position:
  - Japan’s current account is larger than average mainly because of an exceptionally large primary income balance; its trade balance is close to median among G7 or a 100-country sample.
  - Japan’s gross income flows (credit + debit) are relatively modest and lower than in other G7 economies, but they are highly asymmetric (credits much larger than debits).
  - Japan receives more net income from its NFA than the cross-country NFA–investment income relationship would predict.
- Stylized quantitative comparisons (2015–17 averages where stated):
  - Difference in investment income balance between Japan and G6 (Canada, France, Germany, Italy, United Kingdom, United States) equals 3.5 percentage points of GDP.
  - Gross income flows size and asymmetry indicators: Japan is identified as having modest size but high asymmetry relative to peers.

### V. Decomposition: stocks versus yields and implications
- Decomposition framework (for each asset/liability category k):
  - Difference in investment income between Japan and G6 is decomposed into:
    - Contribution of stock positions: (S_Japan^k − S_G6^k) × (Y_Japan^k + Y_G6^k)/2
    - Contribution of implicit yields: (Y_Japan^k − Y_G6^k) × (S_Japan^k + S_G6^k)/2
- Key decomposition results (2015–17 average):
  - On the asset (credit) side, Japan holds fewer foreign assets than G6, but yields on those assets are higher.
  - Foreign assets explain a small part of Japan’s higher investment income balance relative to G6: only 0.3 percentage point of the 3.5 percentage point GDP difference.
  - FDI specifics: Japan has relatively low FDI assets compared to G6, but significantly higher implied yields on FDI assets.
- Possible explanations for higher returns on Japanese foreign investment include factors discussed in referenced literature (Rogoff and Tashiro, 2015; Darvas and Hüttl, 2017), noting that Japan enjoys favorable return differentials akin to “exorbitant privilege” described for the United States.

### VI. Income balance responsiveness to exchange rates and interplay with trade
- Existing literature:
  - Most literature focuses on trade response to exchange rate (e.g., Hooper and Marquez (1995) for developed economies; Reinhart (1994) for developing economies); income balance response has drawn less attention (Alberola et al, 2018).
- Empirical findings from this paper:
  - Income credits and debits both tend to decrease with an appreciation in the REER.
  - This decline mostly reflects a mechanical effect due to the currency composition of the net foreign asset position.
  - For large net creditor countries such as Japan, the credit response tends to dominate; thus the income balance response to REER fluctuations reinforces—although only marginally—the traditional trade balance response.
- Methodological contributions:
  - The paper develops a novel framework to separate, within income credit and debit, the mechanical effect (from currency valuation/composition) from the economic response to exchange rate changes.
  - Annexes detail the estimation strategy for income balance semi-elasticities and robustness tests.

### VII. Policy implications (high-level)
- The compositional shift toward a larger income balance, driven by corporate saving and asymmetric income flows, affects how Japan’s current account might respond to exchange rate movements.
- For large creditor economies like Japan, income account responses to REER appreciations can modestly reinforce trade responses, suggesting that policymakers should consider both trade and income-channel effects when assessing external adjustment and exchange rate policy.

*Italic: Source – wpiea2021026-print-pdf: "1. The Rise in Corporate Saving in Japan" (excerpts provided).*

### Annex 1 presents further evidence on the respective role of stocks and yields in explaining Japan’s large

### Annex 1 presents further evidence on the respective role of stocks and yields in explaining Japan’s large investment income balance compared to peers

### Relative contributions of stocks and yields
- The decomposition highlights that both asset and liability positions and their implied yields matter for Japan’s large investment income balance relative to G7 ex-Japan (G6).
- On the liabilities side, Japan’s smaller stock of foreign liabilities—especially FDI but also portfolio debt—explains a large part of the overall difference in investment income balances: 3.2 out of the 3.5 percentage point of GDP difference.
- Implied yields quoted in the source include values such as 3.8%, 7.5%, 3.4%, 3.7%, 0.9%, 1.6%, 12.9%, 1.9%, 1.0%, 0.5% (presented in the Annex table).
- Other entries in the Annex table report implied yields and income flows with the following exact figures in the source (as presented):
  - Income flows % of GDP: 3.7 5.5 2.2 1.0 1.9 0.3 1.8 0.6 0.7 0.3 0.2
  - Stocks % of GDP: 62.1 146.2 29.6 30.5 51.3 34.7 109.5 5.0 35.3 26.2 43.0
  - Implied Yields percent: 3.8% 7.5% 3.4% 3.7% 0.9% 1.6% 12.9% 1.9% 1.0% 0.5%
  - Additional rows in the table present: Income flows % of GDP: 0.2 5.2 2.6 1.0 1.0 0.6 5.0 1.6 1.0 1.8 0.6
  - Stocks % of GDP: -4.5 225.1 60.0 48.9 41.0 75.2 231.8 48.5 34.1 73.4 75.9
  - Implied Yields percent: 2.3% 4.3% 2.0% 2.5% 0.8% 2.1% 3.3% 2.9% 2.4% 0.8%
  - Further table entries: Income flows % of GDP: 3.5 0.3 -0.4 0.1 0.9 -0.3 -3.2 -0.9 -0.3 -1.5 -0.4
  - Stocks % of GDP: 2.2 -2.3 -1.8 -0.5 0.3 -0.4 -4.5 -3.5 0.0 -0.8 -0.2
  - Implied Yields % of GDP: 1.4 2.7 1.4 0.6 0.6 0.1 1.3 2.6 -0.3 -0.7 -0.2

### Geographical allocation of FDI and impact on yields
- Japan has a larger share of FDI in high-growth emerging Asia and the United States compared with G6 (Figure 3, left chart).
- The source quantifies geographical shares for outward FDI (2013-17, ultimate investor, excluding SPEs):
  - Japan: Advanced Asia 11.4%, Emerging Asia 29.2%, Latin America & Caribbean 4.1%, CIS, CESEE, MENAP, SSA 2.4%, United States 10.3%, Euro Area 33.3%, Other advanced 9.2%.
  - G6, unweighted average: Advanced Asia 4.5%, Emerging Asia 5.4%, Latin America & Caribbean 7.5%, CIS, CESEE, MENAP, SSA 9.4%, United States 32.8%, Euro Area 24.4%, Other advanced 15.9%.
- This more favorable geographical allocation may explain around one percentage point of the higher yields in Japan than in G6 on average since 2000 (Figure 3, right chart).
- The source presents a proxy for FDI asset yields using real GDP growth in partner countries (weighted by outward FDI positions).

### Profit shifting, transfer pricing, and measurement considerations
- Profit shifting through transfer pricing may artificially boost investment income credits and suppress export figures.
- The source cites estimates and implications:
  - Tørsløv, Wier and Zucman (2018) estimate that $28 billion (0.6 percent of GDP) of profits were shifted out of Japan in 2015.
  - Assuming profit shifting mainly distorted Japan’s direct investment income credits (and not the FDI asset position), Japan’s “corrected” implied yield on FDI assets would be around 2 percentage points lower than at present.
  - A related point: Japan’s move from a worldwide to a territorial tax system in 2009 may have led to an intensification of transfer mispricing and profit shifting.
- The Annex cautions that high yields on foreign assets could also reflect other potential measurement issues.

### Liabilities: FDI liabilities, portfolio debt, and low inward FDI
- FDI liabilities in Japan are low relative to G6 and contribute substantially to the income-balance difference.
- The source reports that implicit yields paid on FDI liabilities are very high in Japan (close to 13 percent; see Table 1), suggesting possible measurement issues in valuation of FDI liabilities.
- Adjusting implicit yields to the G6 level, while leaving FDI income payments unchanged, would change Japan’s stock of FDI liabilities from 5 to 20 percent of GDP—still well below the G6 average level (about 50 percent of GDP).
- The low level of inward FDI in Japan has been noted in the literature and linked to measurement issues and corporate governance.
- Portfolio debt liabilities and income paid are small in Japan because of:
  - Low portfolio debt liabilities (public and corporate).
  - Low implicit yields driven by extremely accommodative monetary policy and low credit risk in Japan.
- On the public side, foreign borrowing is relatively low due to a large pool of domestic saving and a strong home bias among domestic investors holding Japanese public debt.
- On the corporate side, corporate saving and corporate deleveraging after the early-1990s real estate bubble led to lower corporate bond liabilities. The source notes that while portfolio debt liabilities followed similar trends in Japan and G6 in the 1980s, Japan diverged after the bubble burst.

### Interconnectedness between the trade and income balances
- The Annex emphasizes that analysis of current account developments should take into account the interconnectedness between its components, the trade and income balances.

*Source: wpiea2021026-print-pdf - Annex 1 presents further evidence on the respective role of stocks and yields in explaining Japan’s large*

### Section III studied the income balance through a static comparison across countries, the

### wpiea2021026-print-pdf - Section III studied the income balance through a static comparison across countries, the

### Interconnectedness of trade, income, and current account balances
- The current account (CA) balance is driven by the saving-investment identity and its drivers, rather than any single component in isolation.
- For most countries, including Japan, the CA balance is highly and positively correlated with the trade balance (Japan: 0.76 for period 1980-2018).
- Correlations between the CA balance and the income balance (total, primary, secondary) tend to be much weaker and more heterogeneous across countries.
- The income balance is generally negatively correlated with the trade balance (Japan: -0.58 for both the total and primary income balances), with substantial cross-country heterogeneity.
- Mechanisms behind the negative correlation between income and trade balances include:
  - Aging: accumulation of net external assets in early aging phases (raising income balance), followed by dissaving and rising imports in post-retirement phases.
  - Income effect: higher income balance may raise consumption and imports; effect varies by recipient (likely higher for secondary income; likely lower for primary income).
  - Market pressure: net debtors may run trade surpluses to meet external debt service on foreign borrowing.
  - Globalization and multinational firms:
    - Offshoring: substitution of goods exports by income receipts.
    - Profit shifting and transfer pricing: can affect trade and income balances in opposite directions.

### Cross-country patterns and Japan-specific observations
- Across countries (average 2015-2017), a negative cross-section relationship is observed between trade balance (% of GDP) and income balance (% of GDP): regression y = -0.7582x -1.6086, R² = 0.6769 (chart depiction in source).
- Countries with large populations living abroad and sizable remittances tend to have high income balances and lower trade balances (upper-left quadrant in scatter), illustrating the income effect.
- Low-tax jurisdictions tend to appear in the lower-right quadrant due to profit shifting and the large role of multinational firms.
- Japan’s increase in the income balance over recent decades has occurred alongside REER depreciation (chart shown: REER, 2010=100; income balance and CA as percent of GDP).

### Currency denomination, mechanical versus economic effects of REER movements
- Decomposition of income balance response to REER changes:
  - Mechanical effect: driven by currency composition of foreign assets/liabilities and related income credits/debits.
    - 2017 shares: foreign assets denominated in foreign currency: United States ~70 percent, Japan 85 percent, median EME nearly 100 percent; median G6 around 50 percent since 1999.
    - 2017 shares: foreign liabilities denominated in domestic currency: United States 85 percent, median G6 82 percent, Japan 67 percent.
    - EME foreign liabilities denominated in foreign currency: 80 percent in 1990, declined to 40 percent in 2017.
  - Economic effect: changes in domestic activity, profits, and outward spillovers can alter income credits/debits (e.g., appreciation may reduce domestic profits and hence income debits if domestic firms are foreign-owned).
- Directional implications:
  - For large net creditor countries (e.g., Japan), REER appreciation likely decreases the income balance if income credits dominate, reinforcing the trade balance response.
  - For large net debtor countries, REER appreciation may increase the income balance if income debits dominate, partially offsetting the trade balance response.
- The overall effect on income balance depends on relative sizes of income credit and debit channels.

### Empirical strategy and key estimation details
- Data and methodology:
  - Annual data for 1999-2018 on a sample of more than 40 countries.
  - Novel decomposition of income flows into mechanical and “economic income flows” (economic income = total income flow minus mechanical effect).
  - Panel-based “CGER-inspired approach” used to estimate income credit and debit elasticities to REER, then combined using country-specific shares (ratios of income credit and debit to GDP) to derive income balance semi-elasticities (Annexes referenced in source).
- Estimation features:
  - Elasticities calculated as sum of coefficients for contemporaneous REER and its lags.
  - Estimations include country and time fixed effects; significance levels indicated as *** p<0.01, ** p<0.05, * p<0.1.

### Empirical findings: elasticities and semi-elasticities
- Baseline elasticities (Table 3; dependent variables: log of Income Credit/Debit ratios to GDP; Economic variants net of mechanical effect):
  - Income credit elasticities (total): -0.56, -0.52 (columns (1) and (2)).
  - Economic credit elasticities: -0.20, -0.22 (columns (3) and (4)).
  - Income debit elasticities (total): -0.28, -0.24 (columns (5) and (6)).
  - Economic debit elasticities: -0.14, -0.15 (columns (7) and (8)).
  - Reported Log of REER coefficients include: -0.560***, -0.774***, -0.200**, -0.106, -0.279***, -0.474***, -0.141*, -0.0945.
  - Observations: 964 (or 944 for economic series); R-squared values reported (e.g., 0.931, 0.932, 0.864, 0.922, 0.885, 0.886); RMSE values reported (e.g., 0.213, 0.212, 0.326, 0.216, 0.268).
- Interpretation:
  - Elasticities are relatively high for total income series (around -0.5 for income credit; -0.3 for income debit) and decrease when isolating economic income (around -0.2 credit; -0.1 debit), indicating a relatively strong mechanical channel.
  - Adding one lag of REER changes individual coefficient significance but does not materially change overall elasticities.
- Robustness and heterogeneity:
  - Distinguishing advanced vs. emerging economies leaves total income credit and debit elasticities broadly unchanged at -0.5 and -0.3 respectively.
  - For emerging economies, “economic credit” and “economic debit” elasticities are negative (not statistically significant); for advanced economies, these economic elasticities are positive and credit is statistically significant—suggesting outward spillovers and advanced-economy investment performance may alter economic credit responses.
  - Five-year averages (periods 1999-2003, 2004-2008, 2009-2013, 2014-2018) reduce mechanical effects; “economic income” elasticities approach total-income elasticities in that specification.
  - Dynamic specifications with lagged dependent variable confirm negative elasticities for total income credit and debit but yield mixed or positive results for economic effects depending on specification.
  - Adding controls for size of foreign assets and liabilities (corrected for valuation effects) yields mixed results; potential measurement and adjustment issues noted.
- Comparative magnitudes:
  - Average absolute magnitude of income flow elasticities ~0.4 (income credit -0.5; income debit -0.3), broadly similar to magnitudes of trade flow elasticities found in literature.
  - Income balance semi-elasticities are smaller than trade balance semi-elasticities because:
    - Income credit and debit elasticities have the same sign and largely offset each other.
    - Income flows as ratios to GDP are much smaller than trade flows.
- Japan-specific semi-elasticity:
  - Using income credit and debit elasticities from Table 3 (columns (1) and (5)) and Japan’s income credit and debit ratios to GDP of 5.5 and 1.8 percent, respectively, yields an income balance semi-elasticity of -0.026 (calculated as [0.055*(−0.56) − 0.018*(−0.28) = −0.026]). A joint test confirms this semi-elasticity is statistically different from zero.
  - Panel-based Japan-specific estimates (using only Japan data) produce larger estimates implying an income balance semi-elasticity of -0.073 (calculation: [0.055*(−1.97) − 0.018*(−1.95) = −0.073]); these are discounted due to limited sample size.
  - By comparison, trade balance semi-elasticities in Cubeddu et al (2019) are reported as -0.12 to -0.14; hence for Japan the income balance response would marginally amplify the trade balance response to REER changes, yielding a small additional CA responsiveness.
- Implication for large creditor vs. debtor countries:
  - Large creditor countries (like Japan): income balance response to REER changes likely amplifies trade balance response (example: Japan income semi-elasticity -0.03 vs. trade -0.12 to -0.14).
  - Large debtor countries: income balance response may offset trade balance response, reducing overall CA responsiveness to REER changes.

### Main conclusions and caveats
- The gradual increase in Japan’s external income balance primarily reflects net revenues from an increasingly positive NFA position; Japan’s high income balance is the result of a highly asymmetric income account.
- Empirical evidence points to a relatively strong total/mechanical effect of REER movements on income credits and debits, with mixed evidence for the economic channel.
- Measurement and identification caveats:
  - Potential under- or over-adjustment in the calculated mechanical effect due to limited official data on currency composition of income flows and imputed values in the IIP currency composition dataset.
  - Offsetting credit and debit responses can mask significant separate effects on income credits and debits when looking only at net income balances—this can explain differing results in the literature.
  - Country heterogeneity and sample-size limitations (especially for country-specific regressions) imply caution in interpreting some country-level estimates.

*Italic line: Source: Section content from wpiea2021026-print-pdf (IMF working paper content provided in the prompt).*

### Annex 5 for a more detailed discussion.

### wpiea2021026-print-pdf - Annex 5 for a more detailed discussion.

### Japan’s investment income and drivers of a high income balance
- Japan’s relatively high income balance is attributed to:
  - higher yields on investment abroad, especially on FDI (from “better” geographical positioning and possibly profit shifting), which more than offset a somewhat lower stock of foreign assets;
  - much lower FDI liabilities (due to possible measurement issues, corporate governance, or regulatory and administrative issues) and lower portfolio debt liabilities (due to strong home bias and corporate deleveraging); and
  - lower yields on portfolio debt liabilities (linked to extremely accommodative monetary policy and low credit risk in Japan).
- The compositional change in Japan’s current account balance over recent decades—declining trade balance offset by rising income balance—does not fundamentally alter external adjustment via the exchange rate, which operates mainly through the trade balance. However, income balance responses may amplify the trade response in large net creditor countries such as Japan, and dampen it in large debtor countries.
- Continued efforts to promote inward FDI could:
  - support productivity growth;
  - contribute to boosting income debits;
  - have unclear net effects on Japan’s income balance and current account, because higher inward FDI could induce indirect offsetting effects (full general equilibrium analysis is beyond the scope of the paper).

### Trade vs. income flows and interconnectedness
- Gross trade flows remain significantly larger than gross income flows in Japan and across countries (charting and cross-country comparisons use the 2015–17 average).
- The income balance is negatively correlated with the trade balance, possibly reflecting mechanisms including:
  - aging;
  - an income effect;
  - market pressures;
  - offshoring; and
  - profit shifting through transfer pricing.
- For Japan, the estimated income balance response to changes in the real exchange rate (REER) tends to reinforce the trade balance response, although only marginally.

### Policy implications and recommendations
- To boost inward FDI and potential productivity gains, suggested steps include:
  - addressing corporate governance issues (Japan’s corporate governance described as an “insider system” with limited power for shareholders, potentially discouraging foreign ownership);
  - addressing regulatory and administrative issues, including reducing the cost of doing business;
  - reforming product markets by reducing barriers to entry in some industries and accelerating deregulation of agricultural and professional services sectors (consistent with IMF advice referenced in IMF, 2018, 2020a, 2020b).
- The paper notes that while higher FDI inflows would likely raise income debits, the overall impact on the income balance and current account is unknown given possible indirect offsetting effects.

### Need for further research: profit shifting and measurement
- Further research is needed to better understand the role and quantitative importance of profit shifting in shaping Japan’s income balance.
- Existing cross-country studies (e.g. Tørsløv et al, 2018) suggest that profit shifting out of Japan may be non-negligible, but the size of related distortions to the composition of Japan’s current account balance remains unclear.
- Additional analysis would inform debates on the costs of profit shifting to the Japanese economy and the design of an optimal taxation regime.

### Methodological contribution: disentangling mechanical and economic effects on income credit and debit
- The paper decomposes income credit and debit into:
  - a mechanical effect: a multiplicative scaling factor capturing direct changes arising from exchange rate movements given the currency composition of assets (for credit) or liabilities (for debit); and
  - an economic effect: the residual after removing the mechanical effect.
- Step 1: Construct currency-of-investment weighted (CIW) exchange rate indices (퐸퐸푅푅_CIW) as geometric averages of bilateral exchange rate ratios vis-à-vis the US dollar (rebased to be equal to one in 2010) using currency shares from Bénétrix et al (2019). For missing years (i.e., 2018), weights are assumed constant at the level of the closest available observation (2017).
- Step 2: Define mechanical and economic effects:
  - Mechanical scaling factor equals the ratio of the CIW exchange rate index for assets to that for liabilities (with scaling factor equal to one in 2010).
  - Economic effect equals observed income credit (or debit) scaled by the inverse of the mechanical factor (with economic effect equal to income credit in 2010).
- Intuition and empirical findings (illustrated for Japan, China, Indonesia):
  - For Japan, income credit as a percent of GDP decreased markedly during the GFC alongside yen appreciation; later increases during yen depreciation are largely accounted for by the mechanical effect, leaving a much smoother economic income credit series.
  - For Japan, the mechanical effect corrects income credit fluctuations more than income debit fluctuations, potentially reflecting an overestimated share of domestic currency (JPY) in foreign liabilities in the underlying dataset, which reduces the mechanical smoothing for debit.
- The decomposition allows separation of purely exchange-rate driven (mechanical) changes from economically driven changes in investment income flows.

### Data, panel estimation, and sample
- Panel regression details:
  - Frequency: annual;
  - Period: 1999-2018;
  - Country sample: Argentina, Australia, Austria, Belgium, Brazil, Canada, Chile, Colombia, Czech Republic, Denmark, Egypt, Finland, France, Germany, Greece, Guatemala, Hungary, India, Indonesia, Ireland, Israel, Italy, Japan, Korea, Malaysia, Mexico, Morocco, Netherlands, New Zealand, Norway, Pakistan, Peru, Philippines, Poland, Portugal, Russia, South Africa, Spain, Sri Lanka, Sweden, Switzerland, Thailand, Tunisia, Turkey, United Kingdom, United States and Uruguay.
- Data sources:
  - Foreign assets & liabilities: External Wealth of Nations database (Lane and Milesi-Ferretti, 2018).
  - Income Balance, GDP, nominal exchange rates: IMF’s WEO or IFS datasets.
  - Currency weights: from Bénétrix et al (2019), estimating shares of USD, EUR, GBP, JPY, CNY, domestic currency (DC) and other foreign currency (OFC) in assets and liabilities for an EBA sample of 50 countries over 1990–2017. For 2018, weights are assumed constant at 2017 levels.
- Additional notes:
  - Implied yields are obtained by dividing investment income gross flows (credit, debit) by corresponding stock positions (assets, liabilities).
  - Cross-country comparisons use the 2015–17 average and present medians, unweighted averages, 25–75 percentiles, and 10–90 percentiles where applicable.

*Source: wpiea2021026-print-pdf - Annex 5 for a more detailed discussion.*

### Annex 5.   Income Balance Elasticities –

### Annex 5.   Income Balance Elasticities –

### Empirical Strategy
- The semi-elasticity of the current-account-to-GDP ratio with respect to the real effective exchange rate is written as:
  - ∆(CA/GDP) / ∆(ER) * ER = η_TT + η_IT
  - where η_TLY = ∆(TLY/GDP) / ∆(ER) * ER with IBBB = TTI, IIII is the semi-elasticity of the nominal (trade, income) balance-to-GDP ratio, and ER is a measure of the exchange rate.

### Using the IMF CGER-Inspired Approach for the Income Balance
- Methodology extension:
  - Decompose income balance semi-elasticities into (i) responsiveness of income flows to exchange rate; and (ii) the size of income flows.
- Semi-elasticities expressed as:
  - η_TT = η_X s_X − η_M s_M
  - η_IT = η_IC s_IC − η_IG s_IG
  - where X, M, IC, IG are nominal flows: exports, imports, income credits, income debits.
  - η_FYCC = ∆(F Y C C / GDP) / (∆(ER) * ER) is the elasticity of nominal flow-to-GDP ratio.
  - s_FYCC = F Y C C / GDP is the ratio of nominal flow to GDP.
- General specifications for income flows (baseline specifications are a simpler version):
  - B_l(I C / GDP)_ct = δ1_IC B_l(I C / GDP)_{c,t−1} + Σ_{i=0}^{C} β_{i}_IC B_l(ER_{c,t−i}) + γ1_IC B_l(F_C_{c,t−1}/GDP_{c,t−1}) + α_c + λ_t + ε_{ct}
  - B_l(I G / GDP)_ct = δ1_IG B_l(I G / GDP)_{c,t−1} + Σ_{i=0}^{C} β_{i}_IG B_l(ER_{c,t−i}) + γ1_IG B_l(F_F_{c,t−1}/GDP_{c,t−1}) + α_c + λ_t + ε_{ct}
- Assumption:
  - The CA response to the exchange rate equals the sum of separately estimated trade balance response and income balance response (parallel to CGER’s indirect method).

### Definitions and Estimation Elements
- Variables:
  - I C_{it} : Nominal income, credit of country i at time t (in current US dollar)
  - I U_{it} : Nominal income, debit of country i at time t (in current US dollar)
  - G U G_{it} : Nominal GDP of country i at time t (in current US dollar)
  - E R_{it} : Exchange rate index, for country i at time t (base year = 2010)
  - F A_{it} : Nominal foreign assets of country i at time t (in current US dollar)
  - F L_{it} : Nominal foreign liabilities of country i at time t (in current US dollar)
  - α_i : Country fixed effects
  - λ_t : Year fixed effects
- Elasticities of income credits and income debits:
  - η_FYCC = Σ_j β_j_FFFFF / (1 − δ1_FFFFF)  with FBFF w = IC, IG
- Long-run semi-elasticity of the income balance (ratio to GDP):
  - η_IT = η_IC s_IC − η_IG s_IG
- The same specifications apply to economic credit/debit by replacing total income credit/debit with economic credit/debit.

### Extra: Profit-Shifting and Current Account Elasticities
- Framework for a tax haven (notations):
  - CA = TTI + IIII
  - CA = CA* + CA_ps
  - TTI = TTI* + TTI_ps
  - IIII = IIII* + IIII_ps
  - CA_ps = TTI_ps + IIII_ps = 0 (what comes in as transfer mispricing ultimately comes out as repatriated profits)
- CA-REER semi-elasticity decomposition:
  - η_CA = η_TT + η_IT
  - η_CA = (η_X s_X − η_M s_M) + (η_IC s_IC − η_IG s_IG)
  - Expand with “true” (*) and profit-shifting (ps) components:
    - η_CA = η_X (s_X* + s_X_pppp) − η_M (s_M* + s_M_pppp) + η_IC (s_IC* + s_IC_pppp) − η_IG (s_IG* + s_IG_pppp)
- Panel-setting equalities in the IMF CGER-inspired approach:
  - η_X = η_X* = η_X_pppp
  - η_M = η_M* = η_M_pppp
  - η_IC = η_IC* = η_IC_pppp
  - η_IG = η_IG* = η_IG_pppp
- Final decomposition:
  - η_CA = η_TT* + η_TT_pppp + η_IT* + η_IT_pppp = η_CA* + η_CA_pppp
- Implications for tax havens:
  - Tax havens are highly open economies; profit shifting inflates trade in services.
  - η_TT_pppp may introduce an upward bias in the trade balance semi-elasticity estimated from the panel CGER-inspired approach.
  - Profit shifting also increases trade balance and decreases income balance; countries with large negative income balances tend to have trade and income semi-elasticities of opposite signs.
  - Therefore, η_IT_pppp would likely help offset the upward bias from η_TT_pppp.
  - Conclusion: including income balance semi-elasticities in the CGER-inspired approach would presumably lead to lower and more reliable estimates of CA-REER semi-elasticities for tax havens.
  - Note: This conclusion is qualitative; the quantitative extent of bias and offset depends on relative values of elasticity estimates for exports, imports, income credits and income debits (quantitative assessment is beyond the scope of the paper).

### Robustness Tests and Additional Estimates (summary of reported tables)
- Table A2 / A3 (Income Credit and Debit Elasticities to REER: Advanced / Emerging Economies)
  - Reported elasticities (examples preserved as in tables):
    - Elasticity (Advanced): -0.49, -0.47, 0.29, 0.31, -0.26, -0.26, 0.08, 0.08
    - Log of REER coefficients (Advanced): -0.493***, -0.662***, 0.289***, 0.178, -0.263**, -0.280, 0.0844, 0.104
    - Observations: 444 (various columns)
    - R-squared: 0.949, 0.949, 0.947, 0.947, 0.953, 0.953, 0.952, 0.952
    - RMSE: 0.165, 0.165, 0.175, 0.175, 0.165, 0.165, 0.169, 0.169
  - Emerging economies reported elasticities and statistics:
    - Elasticity (Emerging): -0.53, -0.49, -0.17, -0.20, -0.32, -0.29, -0.17, -0.21
    - Log of REER coefficients (Emerging): -0.530***, -0.710***, -0.165, 0.0113, -0.318***, -0.456***, -0.174, -0.0265
    - Observations: 520 (various columns)
    - R-squared: 0.894, 0.895, 0.804, 0.804, 0.835, 0.835, 0.753, 0.753
    - RMSE: 0.246, 0.246, 0.386, 0.386, 0.249, 0.249, 0.318, 0.318
- Table A4 (Using 5-year Averages)
  - Elasticities: -0.54, -0.47, -0.12, -0.21
  - Log of REER, 5 year average coefficients: -0.541***, -0.470*, -0.123, -0.207
  - Observations: 190, 186, 190, 186
  - R-squared: 0.949, 0.875, 0.938, 0.898
  - RMSE: 0.203, 0.343, 0.212, 0.277
- Table A5 (With Lagged Dependent Variable)
  - Elasticities reported: 0.67, -0.23, 0.66, 0.35, -0.48, -0.11, -0.07, -0.08 (dependent on specification)
  - Dependent Variable, 1 lag coefficients: 0.754***, 0.778***, 0.863***, 0.866***, 0.761***, 0.778***, 0.830***, 0.830***
  - Log of REER coefficients and lags reported (examples): -0.165***, -0.679***, 0.0902**, 0.345***, -0.114***, -0.552***, -0.0115, -0.000839
  - Observations: 959, 959, 896, 896, 959, 959, 896, 896
  - R-squared: 0.975, 0.977, 0.978, 0.979, 0.970, 0.972, 0.974, 0.974
  - RMSE: 0.129, 0.123, 0.131, 0.129, 0.134, 0.130, 0.128, 0.128
- Table A6 (Estimations in First Differences)
  - Elasticities: -0.68, -0.65, 0.39, 0.40, -0.59, -0.53, 0.04, 0.02
  - Change in REER coefficients: -0.684***, -0.684***, 0.391***, 0.391***, -0.585***, -0.586***, 0.0415, 0.0430
  - Observations: 959, 959, 896, 896, 959, 959, 896, 896
  - R-squared: 0.266, 0.267, 0.273, 0.273, 0.180, 0.180, 0.185, 0.185
  - RMSE: 0.131, 0.131, 0.141, 0.141, 0.137, 0.137, 0.136, 0.136
- Table A7 (With Control Variables)
  - Elasticities: -0.38, -0.36, -0.62, -0.61, 0.10, 0.10, -0.06, -0.07
  - Log of REER coefficients: -0.378***, -0.546***, -0.623***, -0.695***, 0.0987, 0.123, -0.0570, 0.0515
  - Controls included:
    - Log of ER-corrected Foreign Assets, ratio to GDP: 0.136***, 0.128***, -0.353***, -0.356***
    - Log of ER-corrected Foreign Liabilities, ratio to GDP: 0.448***, 0.450***, 0.113**, 0.121**
  - Observations: 941 (all columns)
  - R-squared: 0.933, 0.934, 0.872, 0.872, 0.937, 0.937, 0.887, 0.887
  - RMSE: 0.211, 0.211, 0.316, 0.316, 0.196, 0.196, 0.267, 0.267
- Table A8 (Japan-specific Estimates)
  - Elasticities (Japan): -1.97, -2.05, -1.83, -1.94, -1.95, -2.09, -1.91, -2.07
  - Log of REER coefficients (Japan): -1.972***, -1.593***, -1.828***, -1.272**, -1.946***, -1.208**, -1.905***, -1.091**
  - Log of REER, 1 lag coefficients (examples): -0.453, -0.665, -0.884*, -0.975*
  - Observations: 20 (all columns)
  - R-squared: 0.884, 0.895, 0.705, 0.727, 0.770, 0.807, 0.751, 0.797
  - RMSE: 0.125, 0.123, 0.208, 0.206, 0.187, 0.176, 0.193, 0.179
- Notes across tables:
  - Elasticities are calculated as the sum of coefficients for the REER and its lags (or divided by one minus the DV coefficient where specified).
  - Estimations include country and time fixed effects (not shown).
  - Significance: *** p<0.01, ** p<0.05, * p<0.1

*Source: wpiea2021026-print-pdf - Annex 5.   Income Balance Elasticities –*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021026-print-pdf.pdf_