## _wp07224

## Source details

**Canonical URL:** [_wp07224](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07224.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07224.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07224.pdf.json)

---

### Introduction
- Real GDP growth has been over 8 percent for the past four fiscal years; growth accelerated to over 9 percent in the last two years.
- Oxus Investment (2007) suggests India has entered a new growth trajectory of about 9 percent.
- The Planning Commission of India target: GDP growth of 8−10 percent for the next five years (Approach Paper for the 11th Plan).
- Rising demand pressures (inflation, higher capacity utilization, strong non-oil imports) suggest current growth may be above potential and require policy cooling measures.
- Assessments of potential growth rely heavily on judgment because of limited data on the recent pickup.
- Paper objective: clarify sources of differing potential growth estimates for India and provide a medium-term potential growth projection based on existing exercises (Rodrik and Subramanian (2004a); Poddar and Yi (2007); WEO (2006); Bosworth and Collins (2006)).

### Framework and Factor Share
- Production function (Cobb-Douglas growth accounting): y(t) = a(t) + α k(t-1) + (1−α)(l(t)+h(t)), where:
  - y: real GDP growth rate
  - a: total factor productivity (TFP) growth rate
  - k: physical capital stock growth rate (capital stock measured at end-period, implying one lag)
  - l: labor force growth rate
  - h: human capital growth rate
  - α: capital share
- Capital share (α) used across studies: values from 0.3 to 0.4:
  - Rodrik-Subramanian: 0.35
  - Poddar-Yi: 0.33 (0.3-0.4)
  - WEO: 0.35
  - Bosworth-Collins: 0.4
- Gollin (2002): adjusted labor share stabilized in range 0.65−0.80.
- BCV (2007): small differences between capital and labor growth made results relatively insensitive to factor share choice.

### Total Factor Productivity (TFP)
- Recent TFP growth estimates for India: on the order of 3.2–3.5 percent for recent years (Poddar and Yi (2007) and WEO (2007) in Table 1).
- This recent rate is much higher than India’s long-term history and international experience.
- Cross-country/Asian experience:
  - TFP growth can be volatile; a decade-long pickup does not guarantee a permanent shift.
  - China shows possibility of sustained high TFP but is an outlier and its estimates face statistical issues.
- Sources of productivity growth:
  - Cross-sector resource reallocation (labor shift from low-productivity agriculture to nonagriculture).
  - Within-sector productivity improvements.
- India-specific observations:
  - Cross-sector reallocation contributed as much as one quarter of labor productivity gains since 1980 (BCV (2007)).
  - Employment in agriculture: 57 percent of total employment (versus 47 percent in China and 34 percent in Asia).
  - Nonagriculture labor productivity is 4–5 times agricultural productivity.
  - Pace of labor shifting in India has been slower than other Asian countries; service sector employment share remained nearly flat between 1980 and 2000.
  - Industrial growth skewed to capital-intensive sectors.
  - Rodrik, Subramanian, and Trebbi (2002): India’s level of TFP is just 30−40 percent of what it should be given institutional, geographic, and trade openness conditions.
  - Hsieh and Klenow (2007): improving resource misallocation across plants could yield TFP gains of 40−50 percent in Indian manufacturing.
- Baseline judgment: a decade average of 2.5 percent TFP growth per year is a conservative baseline (Rodrik and Subramanian (2004a)).
- Caution: recent few years’ averages over 3 percent may reflect cyclical factors; avoid declaring a structural break.

### Physical Capital Accumulation
- Favorable demographics (declining dependency ratio) cited as basis for higher future potential growth through increased saving.
- Key methodological translation steps:
  - Translating demographic trends into saving ratio.
  - Translating saving ratio into capital stock growth.
- Across methodologies, overall contribution to GDP growth ranges from 0.17 percent to [text truncated in source].
- Note: "0.3 percent per 1 percent increase in the saving ratio" appears as a calibration of saving-to-growth translation in the source.

### From the Dependency Ratio to Saving
- Empirical relationships:
  - Poddar and Yi (2007): a 1 percentage point decrease in the dependency ratio adds 0.8 to the national saving ratio.
  - Rodrik and Subramanian (2004a) use Mulheisen (1997) showing a one-for-one relationship between dependency and saving ratios.
- Empirical experience 2000/01–2005/06:
  - A 5.2 percentage points of GDP increase in the domestic private saving ratio corresponded to a 4.4 percentage point decline in the dependency ratio (rough support for near one-for-one relationship).
- Demographic projection implication:
  - Using a one-for-one ratio and U.N. population projection implies about a 10 percent increase in the saving ratio between 2005 and 2025. (Rodrik and Subramanian (2004a) use 14 percent for 2000–25.)
- Caveats:
  - Demographic effects emerge in the long run and should not be incorporated into current-period potential growth estimates.
  - Mechanical exercises assume (a) jobs are created for the larger labor force and income is generated and saved; and (b) financial sector efficiently allocates savings into high-return investments.

### From Saving to Capital Accumulation and Growth
- ICOR approach:
  - Recent ICOR for India about 3.5 implies y = (I/Y)/ICOR and a coefficient of about 0.3 percent per 1 percentage point increase in the investment ratio.
  - This approach may overestimate impact because it implies no diminishing returns to capital.
- Cross-country patterns:
  - Fast-growing Asian countries tend to show rising ICORs in the long run; India’s ICOR has been exceptionally stable over two decades.
- India-specific constraints and risks:
  - Investment ratio remained about 22–24 percent of GDP until 2003/04.
  - Demographic-driven potential increase in saving/investment of 10−15 percentage points of GDP is unprecedented and would require large TFP gains to maintain a constant ICOR.
- Modeling approach:
  - Growth accounting combined with perpetual inventory model: K(t)−K(t-1) = I(t)−dK(t-1).
  - Working-age population growth: 1.9 percent (U.N. medium fertility variant) used; labor input assumed to grow at same rate.
  - Baseline: participation rate and unemployment rate remain at present levels.
  - Human capital assumptions vary across studies:
    - Poddar and Yi (2007): average schooling years growth ~2 percent per year.
    - Bosworth and Collins (2003): human capital growth ~0.7 percent per year.
    - WEO (2006): human capital accumulation similar to Bosworth and Collins (~0.67 percent reported elsewhere).
    - Rodrik and Subramanian (2004): no human capital in baseline.

### Comparative Estimates and Projections (highlights from Table 1)
- TFP projections/estimates:
  - Rodrik-Subramanian: 2.5 percent (projection)
  - Poddar-Yi: 3.3 percent for 2006–20 (based on subperiods with 3.5 for 2003–05, 1.3 for 1990–99, 1.4 for 1980–89, and −2.2 for 1971–90)
  - WEO: 3.2 percent for 2002–05 (up from 1.5 percent during 1982–2001)
  - Bosworth-Collins: 2.3 percent (1993–04; up from 1.1 during 1978–93)
- Capital stock growth projections/estimates:
  - Rodrik-Subramanian: recent 6 percent will increase to 8.3 percent reflecting demography (saving rate increases by 14 percent of GDP between 2000–2025).
  - Poddar-Yi: 8.3 percent for 2010–20 (constant ICOR with 8 percent GDP growth), up from 5.6 percent for 2003–05.
  - WEO: 8.1 percent for 2005 (6.8 percent for 1996–05), extended Nehru-Dhareshwar with 4 percent depreciation; +2.3 percent with improved demography (Rodrik & Subramanian).
  - Bosworth-Collins: 6.4 percent for 1993–04.
- Labor growth:
  - UN working age: 1.9 percent used by multiple studies.
- Human capital:
  - Poddar-Yi: 2.0 percent average education attainment growth for population above 15.
  - Bosworth-Collins (Jorgenson Vu (2005) model): 0.83 percent labor quality growth since 1982 takeoff.
  - WEO: 0.67 percent (1993–04) human capital index growth.
- Potential growth, 2006/07 (author's estimate using each study's TFP, capital share, and population/human capital assumptions; CSO capital stock growth 2005/06 = 8.8% used):
  - Rodrik-Subramanian: 6.8
  - Poddar-Yi: 8.8
  - WEO: 8.1
  - Bosworth-Collins: 7.4
- Potential growth, long run (medium-term 2007/08–2012/13 average based on simulations; low end constant investment ratio as 2006/07; high end assumes demographic gains: 10 percentage points in 20 years => 0.5 percentage points per year increase in medium term plus IMF staff projected fiscal consolidation as of 2007 April WEO; actual growth 2006/07 = 9.4 percent):
  - Rodrik-Subramanian: 7.3-7.6
  - Poddar-Yi: 9.5-9.8
  - WEO: 8.7-9.0
  - Bosworth-Collins: 8.0-8.4
- Note: Table 1 estimates revised upward with CSO capital stock growth data between 2006/07 and 2005/06 editions.

### Medium-term Simulations and Investment-Ratio Scenarios
- Baseline inputs for simulations starting 2007/08:
  - Actual growth rate 2006/07: 9.4 percent
  - Gross investment ratio 2006/07: 35.3 percent
  - Actual capital stock growth 2005/06 (CSO): 8.8 percent
- Four investment-ratio scenarios:
  1. Constant: investment ratio remains at 2006/07 level.
  2. Fiscal consolidation: investment ratio increases reflecting additional saving from fiscal consolidation (IMF staff projection as of Spring 2007 WEO).
  3. Demographics: investment ratio increases reflecting demographics (0.5 percentage points of GDP per year based on U.N. projection of 10 percent decline in dependency ratio from 2005 to 2025).
  4. Combined: both fiscal consolidation and demographics.
- Example investment-ratio profile (percent of GDP) from Table 3:
  - 2005/06: 33.8
  - 2006/07: 35.3
  - 2007/08: 35.3 (constant); 35.5 (fiscal); 35.8 (demographic); 36.0 (fiscal+demographic)
  - 2008/09: 35.3; 36.4; 36.3; 37.4
  - 2009/10: 35.3; 36.8; 36.8; 38.3
  - 2010/11: 35.3; 37.2; 37.3; 39.2
  - 2011/12: 35.3; 37.6; 37.8; 40.1
  - 2012/13: 35.3; 38.2; 38.3; 41.2
- Medium-term potential growth results:
  - Depending on parametric assumptions, medium-term average potential growth ranges from 7.3 percent to 9.5 percent.
  - Excluding Poddar and Yi (2007) and Rodrik and Subramanian (2004a), WEO (2006) and Bosworth and Collins (2006) assumptions give narrower range: 8.0−9.0 percent.
  - Major difference between WEO and Bosworth-Collins is TFP assumption:
    - WEO (2006): TFP growth 3.2 percent.
    - Bosworth and Collins (2006): TFP growth 2.3 percent.
  - Above-potential growth in 2006/07 (positive TFP shock) has a lasting but diminishing positive impact on medium-term growth via higher saving and capital formation if saving ratio does not fall.
  - Alternative saving ratio assumptions imply about 0.5−0.6 percentage points differences in potential growth at end of medium term across cases despite up to 5 percentage points variation in the investment ratio for 2012/13 — reflecting a larger role for productivity growth than factor accumulation.

### Sources of Uncertainty, Risks, and Constraints
- TFP volatility and uncertainty:
  - TFP growth can reverse sharply; recent high rates may partly reflect cyclical factors.
  - Sustaining high TFP growth likely depends on continued reforms and institutional improvements.
- Investment efficiency and financial-sector constraints:
  - India’s sustained investment ratio of about 35 percent of GDP is high relative to historical experience.
  - Financial-sector issues to address:
    - Large share of state-owned banks.
    - High statutory liquidity requirements compelling banks to set aside a quarter of deposits for government securities.
    - High lending requirements: 40 percent of net lending for the priority sector.
    - Underdeveloped government and corporate bond markets.
- Turning demographics into jobs and saving:
  - An average of 13 million people expected to enter India’s labor force each year for the next four decades on the back of a declining dependency ratio.
  - Creating quality jobs for these entrants is necessary to raise saving; formal-sector job creation has been weak.
  - Employment trends (1999/00–2004/05) presented in source as table entries: 39 74 592.9 (aggregate percent change, annual average); Organized: 28 26 -1.2; Unorganized: 36 94 32.63 3.2.
  - Organized-sector employment declined over past 5 years; unorganized employment rose but with lower productivity.

### Conclusion and Policy Implications
- Summary of quantitative estimates:
  - Potential growth in 2006/07: range 6.8–8.8 percent based on four studies' parameters.
  - Excluding two studies with extreme human capital assumptions, 2006/07 range: 7.4−8.1 percent, which is 1.3–2.0 percentage points below actual outturn of 9.4 percent (signaling possible overheating).
  - Medium-term average potential growth: 7.3 percent to 9.5 percent; excluding two extreme studies: 8.0–9.0 percent. Width of latter range mainly reflects differing TFP assumptions.
- Overall assessment:
  - Shift to more investment-led growth and strong productivity gains appear to have raised India’s medium-term potential growth to around 8 percent, but risks remain on both sides.
  - Downside risks: volatile productivity gains; possible decline in investment pace if cyclical forces reverse; challenges maintaining investment efficiency.
  - Upside potential: productivity catch-up could raise medium-term potential growth if realized.
- Policy recommendations to realize demographic and productivity gains:
  - Foster improvements in labor market conditions to enable better job creation, including reforms of labor laws that hamper organized-sector employment.
  - Strengthen financial-sector efficiency to sustain investment effectiveness by addressing:
    - State-owned bank dominance.
    - High statutory liquidity requirements.
    - High priority-sector lending mandates.
    - Underdeveloped bond markets.

*Content unit derived from the IMF working paper text provided in the source PDF.*

### 8.1 percent for 2006/07, and about 8 percent for the medium term. The medium-term

### _wp07224 - 8.1 percent for 2006/07, and about 8 percent for the medium term. The medium-term

### Introduction
- Real GDP growth has been over 8 percent for the past four fiscal years; growth accelerated to over 9 percent in the last two years.
- Oxus Investment (2007) suggests India has entered a new growth trajectory of about 9 percent.
- The Planning Commission of India target: GDP growth of 8−10 percent for the next five years (Approach Paper for the 11th Plan).
- Signs of rising demand pressures (inflation, higher capacity utilization, strong non-oil imports) suggest current growth may be above potential and require policy cooling measures.
- Assessments of potential growth rely heavily on judgment due to limited data on the recent pickup; this explains widely ranging estimates in the literature.
- Paper objective: clarify sources of differing potential growth estimates for India and provide a medium-term potential growth projection based on existing exercises (Rodrik and Subramanian (2004a); Poddar and Yi (2007); WEO (2006); Bosworth and Collins (2006)).

### Framework and Factor Share
- Production function used: Cobb-Douglas growth accounting form y(t) = a(t) + α k(t-1) + (1−α)(l(t)+h(t)), where:
  - y: real GDP growth rate
  - a: total factor productivity (TFP) growth rate
  - k: physical capital stock growth rate (capital stock measured at end-period, implying one lag)
  - l: labor force growth rate
  - h: human capital growth rate
  - α: capital share
- Studies use capital share (α) values from 0.3 to 0.4:
  - Rodrik-Subramanian: 0.35 (Bosworth-Collins, 2003)
  - Poddar-Yi: 0.33 (0.3-0.4)
  - WEO: 0.35 (literature standard)
  - Bosworth-Collins: 0.4
- Gollin (2002) evidence: labor share is fairly constant across countries once self-employment income is properly allocated; adjusted labor share stabilized in range 0.65−0.80.
- Bosworth, Collins, and Virmani (BCV) (2007): small differences between capital and labor growth made growth accounting results for India relatively insensitive to factor share choice.

### Total Factor Productivity (TFP)
- Recent TFP growth estimates for India tend to be on the order of 3.2–3.5 percent for recent years (Poddar and Yi (2007) and WEO (2007) in Table 1).
- This recent rate is much higher than India’s long-term history and international experience (Figure 1).
- Cross-country/Asian experience:
  - TFP growth can be volatile; a decade-long pickup does not guarantee a permanent shift.
  - Chinese experience shows sustained high TFP is possible, but China is an outlier and its estimates face statistical issues (Bosworth and Collins, 2003; WEO, 2006).
- Sources of productivity growth:
  - Cross-sector resource reallocation (labor shift from low-productivity agriculture to nonagriculture).
  - Within-sector productivity improvements.
- Empirical drivers from WEO (2006) regressions: trade openness, financial sector development, business environment, institutional quality correlate with productivity growth and labor shifts from agriculture (Table 2).
- India-specific observations and potential:
  - Cross-sector reallocation contributed as much as one quarter of labor productivity gains since 1980 (BCV (2007)).
  - Employment in agriculture remains high at 57 percent of total employment (compared to 47 percent in China and 34 percent in Asia); nonagriculture labor productivity is 4–5 times agricultural productivity.
  - Pace of labor shifting in India has been slower than other Asian countries; service sector employment share remained nearly flat between 1980 and 2000.
  - Industrial growth skewed to capital-intensive sectors.
  - Rodrik, Subramanian, and Trebbi (2002) suggest India’s level of TFP is just 30−40 percent of what it should be given institutional, geographic, and trade openness conditions—implying scope for catch-up.
  - Hsieh and Klenow (2007) estimate improving resource misallocation across plants could yield TFP gains of 40−50 percent in Indian manufacturing.
- Baseline judgment: a decade average of 2.5 percent TFP growth per year is a conservative baseline (Rodrik and Subramanian (2004a)).
- Recent few years’ averages over 3 percent may reflect cyclical factors; caution needed in declaring a structural break.

### Physical Capital Accumulation
- Favorable demographics (declining dependency ratio) are cited as a basis for higher future potential growth through increased saving.
- Key methodological translation steps:
  - Translating demographic trends into saving ratio.
  - Translating saving ratio into capital stock growth.
- Across different methodologies, the overall contribution to GDP growth ranges from 0.17 percent to [text truncated at this point in source].

### Comparative Estimates and Projections (Table 1 highlights)
- TFP projections/estimates from studies:
  - Rodrik-Subramanian: (Underestimated) projection of 2.5 percent (same rate as past two decades)
  - Poddar-Yi: 3.3 percent for 2006–20 (based on estimated 3.5 for 2003–05, 1.3 for 1990–99, 1.4 for 1980–89, and −2.2 for 1971–90)
  - WEO: Estimate: 3.2 percent for 2002–05, up from 1.5 percent during 1982–2001
  - Bosworth-Collins: 2.3 (1993–04), up from 1.1 during 1978–93
- Capital stock growth projections/estimates:
  - Rodrik-Subramanian: Recent 6 percent will increase to 8.3 percent reflecting favorable demography that increases saving rate by 14 percent of GDP between 2000–2025.
  - Poddar-Yi: 8.3 percent for 2010–20 (constant ICOR with 8 percent GDP growth), up from 5.6 percent for 2003–05.
  - WEO: 8.1 percent for 2005 (6.8 percent for 1996–05), extended estimate based on Nehru-Dhareshwar using 4 percent depreciation. +2.3 percent with improved demography (Rodrik & Subramanian).
  - Bosworth-Collins: 6.4 percent for 1993–04.
- Labor growth assumptions:
  - UN working age (1.9 percent) used by multiple studies (Poddar-Yi, WEO, Bosworth-Collins).
  - National household survey historical estimate: 1.9 percent for 1993–04.
- Human capital projections/estimates:
  - Poddar-Yi: 2.0 percent average education attainment for population above 15 based on Barro and Lee (2003) (raw growth rate of average schooling year).
  - Bosworth-Collins (Jorgenson Vu (2005) model): 0.83 percent estimate since 1982 takeoff for labor quality (reflecting education attainment and institutional variables).
  - WEO: 0.67 percent (1993–04) human capital index based on household survey data on schooling.
- Potential growth estimates:
  - Potential growth, 2006/07 (author's estimate based on each underlying study's TFP estimate for the latest year, capital share, and population/human capital assumption; CSO data for capital stock growth in 2005/06 (8.8%) used as historical input):
    - Rodrik-Subramanian: 6.8
    - Poddar-Yi: 8.8
    - WEO: 8.1
    - Bosworth-Collins: 7.4
  - Potential growth, long run (medium-term (2007/08–2012/13) average based on simulation results; low end assumes same constant investment ratio as 2006/07; high end assumes demographic gains for saving and investment ratio (10 percentage points increase in 20 years, translated into 0.5 percentage points per year increase in the medium term) plus fiscal consolidation effects projected by IMF staff as of 2007 April WEO; estimates built upon actual growth for 2006/07 of 9.4 percent):
    - Rodrik-Subramanian: 7.3-7.6
    - Poddar-Yi: 9.5-9.8
    - WEO: 8.7-9.0
    - Bosworth-Collins: 8.0-8.4
- Note: Estimates in Table 1 are higher than Poddar and Yi's original estimates reflecting upward revision with capital stock growth data from CSO between 2006/07 and 2005/06 edition.

### Sources of Uncertainty, Risks, and Policy Implications (synthesized from text)
- TFP volatility and uncertainty:
  - TFP growth can reverse sharply; recent high rates may partly reflect cyclical factors.
  - Sustaining high TFP growth likely depends on continued reforms and improvements in policy and institutions.
- Reforms and institutional factors matter:
  - Trade openness, financial sector development, business environment, and institutional quality are correlated with productivity gains and labor reallocation.
  - Further labor market reforms and measures to enhance labor mobility would help realize cross-sector reallocation gains.
  - Addressing resource misallocation within industries could yield large within-sector productivity gains.
- Demographics and capital accumulation:
  - Demographic dividend can raise saving and investment, boosting capital stock growth—but translation from demographics to saving and then to capital stock growth is methodologically sensitive and uncertain.
- Overall projection guidance:
  - Conservative baseline for TFP: 2.5 percent per year (decade average).
  - Medium-term potential growth estimates cluster around “about 8 percent” with study-specific ranges (see Table 1 projections above).
  - Risks on both sides: productivity gains and investment could be volatile, but determined reforms could sustain strong productivity growth.

*Italic line: Content unit derived from the IMF working paper text provided in the source PDF.*

### 0.3 percent per 1 percent increase in the saving ratio. However, it should be noted that these

### III. SIMULATION RESULTS

### From the Dependency Ratio to Saving
- Existing studies on India suggest a near one-for-one relationship between the dependency ratio and national saving.
  - Poddar and Yi (2007) regress the national saving ratio on the lagged dependency ratio and find that a 1 percentage point decrease in the dependency ratio adds 0.8 to the national saving ratio.
  - Rodrik and Subramanian (2004a) use Mulheisen (1997) results showing a one-for-one relationship between dependency and saving ratios.
- Empirical experience over 2000/01–2005/06:
  - A 5.2 percentage points of GDP increase in the domestic private saving ratio corresponded to a 4.4 percentage point decline in the dependency ratio, roughly supporting a negative one-for-one relationship between overall dependency ratio and domestic saving.
- Demographic projection implication:
  - Using a one-for-one ratio and U.N. population projection implies about a 10 percent increase in the saving ratio between 2005 and 2025. (Rodrik and Subramanian (2004a) use 14 percent for 2000–25.)
- Important caveats:
  - Demographic effects emerge only in the long run and should not be incorporated into estimates of potential growth for the current period.
  - Mechanical exercises rely on implicit assumptions: (a) jobs are created for the rapidly increasing labor force and income is generated and saved; and (b) efficiency in the financial sector is improved to allocate savings into high-return investment projects. Both assumptions might not necessarily hold.

### From Saving to Capital Accumulation and Growth
- Simple ICOR approach:
  - With a recent ICOR for India at about 3.5, the relationship y = (I/Y)/ICOR implies a coefficient of about 0.3 to a unit increase in the investment ratio (i.e., 0.3 percent per 1 percentage point increase in investment ratio).
  - This approach may overestimate impact because it implies simultaneous increases in productivity and/or labor that prevent diminishing returns to capital.
- Cross-country and historical patterns:
  - Fast-growing Asian countries tend to show rising ICORs in the long run; early starters such as Japan and Korea exhibit this trend.
  - By contrast, India's ICOR has been exceptionally stable over the past two decades.
- India-specific constraints and risks:
  - India’s investment ratio remained about 22–24 percent of GDP until 2003/04 and has not previously been accompanied by a strong investment ratio of over 30 percent of GDP.
  - The demographic-driven potential increase in saving and investment is unprecedented (adding 10−15 percentage points of GDP), and the required TFP gains to maintain a constant ICOR might be massive.
- Modeling approach used:
  - Growth accounting equation combined with a perpetual inventory model for capital accumulation: K(t)−K(t-1) = I(t)−dK(t-1), with representation dividing by K(t-1) as shown in the source.
  - Working-age population growth of 1.9 percent (U.N. medium fertility variant) is used; labor input is assumed to grow at the same rate as working-age population.
  - Baseline projections assume participation rate and unemployment rate remain at present levels (no explicit changes).
  - Human capital assumptions vary across studies:
    - Poddar and Yi (2007) use average schooling years growth of about 2 percent per year as human capital growth.
    - Bosworth and Collins (2003) compute human capital stock assuming a 7 percent return to average schooling year, yielding about 0.7 percent human capital growth per year.
    - WEO (2006) yields a similar accumulation rate to Bosworth and Collins (2003) via a different econometric approach.
    - Rodrik and Subramanian (2004) did not incorporate human capital in their baseline projection.

### Potential Growth Estimates for 2006/07 and Medium Term
- Potential growth in 2006/07 (using parametric assumptions from four studies and actual capital stock growth in 2005/06):
  - Range: 6.8 percent to 8.8 percent (Table 1).
  - Narrower range using WEO (2006) and Bosworth and Collins (2006) assumptions: 7.4−8.1 percent.
  - These narrower-range estimates are 1.3–2 percentage points below the outturn of 9.4 percent.
- Medium-term simulations (starting 2007/08) build on equations (1) and (2) and use:
  - Actual growth rate of 9.4 percent in 2006/07.
  - Gross investment ratio of 35.3 percent.
  - Actual capital stock growth of 8.8 percent in 2005/06 (CSO estimate).
- Four simulated scenarios for the investment ratio:
  1. Constant: investment ratio remains at the same level as 2006/07.
  2. Fiscal consolidation: investment ratio increases reflecting additional saving from fiscal consolidation under the Fiscal Responsibility and Budget Management framework (using IMF staff projection as of Spring 2007 WEO).
  3. Demographics: investment ratio increases reflecting demographics (0.5 percentage points of GDP per year, based on the U.N. projection discussed above).
  4. Combined: investment ratio increases reflecting both fiscal consolidation and demographics.
- Assumed demographic gain:
  - 0.5 percentage points GDP gain in saving and investment ratio based on U.N. projection of 10 percent decline in dependency ratio from 2005 to 2025.
- Table 3 (Investment Ratio assumptions, in percent of GDP) highlights example profile:
  - 2005/06: 33.8
  - 2006/07: 35.3
  - 2007/08: 35.3 (constant); 35.5 (fiscal); 35.8 (demographic); 36.0 (fiscal+demographic)
  - 2008/09: 35.3; 36.4; 36.3; 37.4
  - 2009/10: 35.3; 36.8; 36.8; 38.3
  - 2010/11: 35.3; 37.2; 37.3; 39.2
  - 2011/12: 35.3; 37.6; 37.8; 40.1
  - 2012/13: 35.3; 38.2; 38.3; 41.2
- Medium-term potential growth results:
  - Depending on parametric assumptions, the medium-term average potential growth rate ranges from 7.3 percent to 9.5 percent.
  - Excluding Poddar and Yi (2007) and Rodrik and Subramanian (2004a), WEO (2006) and Bosworth and Collins (2006) assumptions give a narrower range of 8.0−9.0 percent.
  - Major difference between WEO (2006) and Bosworth and Collins (2006) is TFP growth assumption:
    - WEO (2006): TFP growth of 3.2 percent.
    - Bosworth and Collins (2006): TFP growth of 2.3 percent.
  - Above-potential growth in 2006/07 (additional positive TFP shock) will have a lasting but diminishing positive impact on medium-term growth, because higher income translates into larger saving and capital formation if saving ratio does not fall.
  - Alternative saving ratio assumptions imply about 0.5−0.6 percentage points differences in potential growth at the end of the medium term across the four cases despite up to 5 percentage points variation in the investment ratio for 2012/13 — reflecting a larger role for productivity growth than factor accumulation in driving potential growth in India.

### Discussion: Risks, Constraints, and Implementation Challenges
- Productivity importance and volatility:
  - The exercise highlights productivity as a key contributor to potential growth historically and prospectively.
  - There is no guarantee that current pace of productivity growth will be sustained; productivity could be experiencing a cyclical upswing.
- Investment efficiency and financial-sector constraints:
  - India’s investment ratio of about 35 percent of GDP over the medium term is already high and comparable to Asian countries pre-crisis.
  - Proactive financial-sector policies may be required to maintain investment efficiency.
  - Key financial-sector issues to address include:
    - Large share of state-owned banks.
    - High statutory liquidity requirements compelling banks to set aside a quarter of deposits for government securities.
    - High lending requirements (40 percent of net lending) for the priority sector (including agriculture, small-scale business, and education).
    - Underdeveloped government and corporate bond markets.
- Turning demographics into jobs and saving:
  - An average of 13 million people is expected to enter India’s labor force each year for the next four decades on the back of a declining dependency ratio.
  - Increasing saving requires creating quality jobs for these entrants, which is challenging given weak formal-sector job creation.
  - Employment trends:
    - Employment (percent change, annual average) 1999/00–2004/05: 39 74 592.9 (table entries as presented in source).
    - Organized: 28 26 -1.2
    - Unorganized: 36 94 32.63 3.2
    - (Table 4 presents employment situation in millions; sources: National Sample Survey Organization; Ministry of Labor; Ministry of Finance.)
  - Organized-sector employment has declined over the past 5 years; unorganized (informal) sector employment has risen but with lower productivity than organized sector.
  - Policy implication: Reform of labor laws that hamper creation of organized-sector jobs might be needed to reap demographic benefits.

### Conclusion and Policy Implications
- Summary of estimates:
  - Potential growth in 2006/07 is estimated to be in the 6.8–8.8 percent range based on parameters from four studies.
  - Excluding two studies with extreme human capital assumptions, the range becomes 7.4−8.1 percent, which is 1.3–2.0 percentage points below the 2006/07 outturn of 9.4 percent, indicating overheating.
  - Medium-term average potential growth ranges from 7.3 percent to 9.5 percent; excluding the two extreme studies, the range is 8.0–9.0 percent. The width of the latter range mainly reflects differing TFP assumptions rather than saving/investment.
- Overall assessment:
  - The shift to more investment-led growth and strong productivity gains appear to have raised India’s medium-term potential growth to around 8 percent, but risks remain on both sides.
  - Downside risks: volatile productivity gains; possible decline in investment pace if cyclical forces reverse; challenges maintaining investment efficiency.
  - Upside potential: room for productivity catch-up could raise medium-term potential growth if realized.
- Policy recommendations to realize demographic and productivity gains:
  - Foster improvements in labor market conditions to enable better job creation (including reforms of labor laws that hamper organized-sector employment).
  - Strengthen financial-sector efficiency to sustain investment effectiveness (address state-owned bank dominance, high statutory liquidity requirements, high priority-sector lending mandates, and underdeveloped bond markets).

*Source: Excerpt from IMF working paper (simulation and discussion sections provided in the supplied content).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07224.pdf_