Global Economics Paper No: 218 - ghsl.cn · Global Economics Paper No: 218 Research Report China: More efficient cities key to a brighter growth path A more efficient use of surplus

May 21, 2013

Global Economics Paper No: 218

Research Report

China: More efficient cities key to a brighter growth path

A more efficient use of surplus labor will more than offset the impact on growth from a modest

decline of the working age population. Further urbanization will be part of future growth.

Urbanization process needs to improve, in order to deliver more efficient and balanced growth

through: 1) productivity gains from city agglomeration; 2) a decline in capital intensity as land prices

better reflect market forces; and 3) a moderation in the household savings rate, helped by reforms

to social safety nets.

Using US cities as a benchmark, greater efficiency in Chinese cities has the potential to add 1ppt to

real economic growth over the next decade. Investment-to-GDP share will decline and consumption-

to-GDP share improve meaningfully with reforms. Urban housing construction will benefit from

future income growth.

The recent policy directions are promising and suggest that economic imbalances should gradually

be corrected. A more concentrated urban growth model would be positive for public transportation,

utilities, water management and diversified consumption, and the burden on public finances would

be lower.

Investors should consider this report as only a single factor in making their investment decision. For Reg AC certification and other important disclosures, see the Disclosure Appendix, or go to www.gs.com/research/hedge.html.

Li Cui

+852-2978-0784 [email protected] Goldman Sachs (Asia) L.L.C.

The Goldman Sachs Group, Inc. Goldman Sachs Global Economics, Commodities and Strategy Research

May 21, 2013 Global Economics Paper

Goldman Sachs Global Economics, Commodities and Strategy Research 2

Table of Contents

Summary 3

1. Urbanization has a long way to go 4

2. The multiple links between urbanization and Chinese growth 8

A. Urbanization and high investment share: a more conservative use of land key for bringing down

the investment/GDP share 8

B. Urbanization and productivity gains: Exploring the agglomeration effects of cities 11

C. Urbanization and household savings: Migrant savings key to the rising household savings rate 14

D. Urbanization and housing: Higher income, more housing consumption 16

3. Next stage of urbanization to be driven by key policy changes: Land reforms, city agglomeration, savings 18

A. A more concentrated approach is preferable 18

B. Land policies and better infrastructure are key ingredients for more compact cities 20

C. Comprehensive social security reforms supported by fiscal reforms needed to stem the rise in

household savings 22

4. Growth and sectoral impact under different urbanization scenarios 23

5. Conclusion 26

References 27

Disclosure Appendix 28

Many thanks to Jan Hatzius, Dominic Wilson, Andrew Tilton, Goohoon Kwon and MK Tang for their insightful comments.



Summary The most pressing issues for China’s macro landscape are linked to its medium-term

growth model and whether the country will become stuck in the ‘middle-income trap’.

Investors are concerned about the country’s high investment share, elevated household

savings and constrained consumption growth, as well as the apparent decline in the

economic returns on investment.

In this paper, we argue that many of these imbalances have their roots in the urbanization

process pursued so far. Carefully considered and measured policies can go a long way

towards addressing such imbalances. The government has announced its intention to

tackle several key reforms under a “new urbanization strategy”. We examine some of the

key dimensions of urbanization and assess quantitatively how the growth trajectory is

likely to be affected under different urbanization patterns. Our analysis suggests there is

significant potential to achieve more efficient and balanced growth through a well

calibrated strategy. Land policies, city size, reforming social welfare and housing policies

are key areas that we will be monitoring closely. Although policy details have not yet been

announced, the strong focus on a number of these areas is promising.

1. Land reforms are important: A more economic use of land would lower the intensity of

investment in structures, i.e., construction-related activities such as infrastructure and

buildings for industrial needs, and as a result lower the investment/GDP share. This would

help reverse the recent trend whereby urban land expansion has outpaced population

growth, and in turn would increase the relative share of equipment in total investment. As

the latter is more directly related to future growth, this would help to boost investment

efficiency for the economy as a whole. We estimate that a more modest pace of expansion

in city construction—by 1.5 percentage points (ppt) less each year, or about one-third

slower than the current pace—can reduce the investment/GDP share by 3-6ppt in the

coming decade.

2. The ‘agglomeration effect’ in cities can be a key source of future productivity gains but this

requires careful urban planning. Although urbanization has so far succeeded in shifting

surplus resources into non-agricultural use, efficiency gains through resource

agglomeration have been absent. Such economies of agglomeration have the benefit of

greater economies of scale, a deeper job market, efficiency in land use and a lighter burden

on public finances. We estimate that, over the coming decade, if Chinese cities achieve the

economies of agglomeration present in US cities, there will be an additional gain of 1ppt in

the growth of total factor productivity (TFP) by the end of the period. Land policies and

related urban management are critical to realize such gains. In particular, adequate

infrastructure is needed to support more compact cities in order for the positive impact of

city agglomeration to offset the negative side-effects of big cities. The recent focus on

public transportation in cities is positive in this respect.

3. Hukou and social welfare reforms are critical: Increased savings by migrant workers have

been behind the decline in the consumption share. The urban savings rate is higher and is

increasing faster than in rural areas, driven in part by the high propensity of migrant

workers to save for precautionary reasons. In the absence of any reforms, further urban

migration can increase the average household savings rate each year by 0.1 ppt per year.

We expect hukou and social security reforms, accompanied by fiscal reforms, will be

important to realize the full potential of urbanization for domestic demand. We estimate

that consumption could be boosted by as much as 2-3 ppt of GDP as a result of these

reforms.

4. Housing purchase has been financed mostly by household savings rather than leverage.

Similar to consumption, it will continue to benefit from further income growth. Urban

housing supply has been mostly driven by population and income growth, and residential

housing investment currently accounts for less than 7% of GDP and 10% of total

investment. We forecast that housing investment as a share of GDP will recover somewhat

in the next few years, even under the ongoing policy controls.



1. Urbanization has a long way to go

China’s growth in the past 30 years has gone hand-in-hand with the rise of cities, driven by

rapid industrialization and integration with the world economy, and an urban population

that has more than tripled from 1980. In the past 10 years alone, the number of urban

residents has grown by 20 million each year, accounting for about 30% of the increase in

the world’s urban population. China now has the largest urban population in the world,

even though the urbanization rate is still just above 50% (Exhibits 1-2).

Growth and urbanization have also been closely associated geographically, suggesting that

urbanization is the outcome of growth and job creation. The regions along the coastal

areas have urbanized at a fast pace, with around 70% of their population already urban

residents. In contrast, in western regions such as Sichuan, the share of city dwellers

remains at around 40% of the total population, although these provinces have been

urbanizing rapidly along with brisk income growth in recent years (Exhibit 3).

Cities are likely to remain the engine of growth in China, as has been the case for economic

development in the past (Spencer, 2009). The new Chinese government has identified

urbanization as one of the four pillars of its economic mandate in the coming decade.1

Indeed, rural areas still retain about 30% of the total labor force, with an average labor

productivity that is one-sixth that in urban areas. The United Nations projects that China’s

urban population will grow by another 185 million in the coming decade, bringing the

urbanization rate to 60% by 2020, and sustained non-agricultural job creation will remain

critical.

The large amount of surplus labor is a key counterforce to the overall demographic trend,

and argues against the widely held, yet misplaced, view that China’s growth is destined to

slow as a result of a shrinking labor force. In our view, a more efficient use of the surplus

labor in rural areas and of under-employed workers in urban areas would more than offset

the impact on growth of the slight decline in the overall working age population (about

0.1% each year) in the coming decade (see Asia Economic Analyst: China: Shifting to a

lower gear, September 20 2012.) How best to bring these underutilized resources into more

productive use will therefore be critical and, as in the past, sustained non-agricultural job

creation will be the driver.

Meanwhile, a new strategy is necessary, given the imbalances in the past and more limited

scope for easy productivity gains in the future:

First, urbanization has so far been associated with some of the pressing

imbalances, including the depletion of arable land, environmental damage, air

pollution, income disparity and rising urban congestion, to name a few. This has

raised concerns that the urbanization path pursued so far is not sustainable.

Second, there is a need to search for new sources of efficiency gains as the scope

of easy productivity gains through the ‘catch-up’ effect naturally narrows. In

particular, trade growth is likely to be more modest given China’s sizeable market

share, and the scope of productivity gains through trade is also arguably smaller

than before, as a result of the rising sophistication of Chinese products.

1 Premier Li Ke Qiang laid out several key areas of consideration in promoting future urbanization, in particular the location of urban population growth, land utilization to protect arable land in the process of urbanization, hukou reforms so that rural migrant workers can become more integrated within the cities where they work, energy and environmental implications, etc.



1. China has reached its current income level at a lower urbanization rate than other

countries

Note: Exhibits 1-4 are based on the United Nations Population data. There may be national differences in characteristics that distinguish urban from rural areas.

Source: CEIC, United Nations Population Statistics, World Bank.

2. China has contributed significantly to the world’s

urban growth

3. China's urbanization process closely follows its growth

path

Note: Exhibits 1-4 are based on the United Nations Population data. There may be national differences in characteristics that distinguish urban from rural area.

Source: CEIC, United Nations Statistics.

Source: CEIC, China National Bureau of Statistics.

0

10

20

30

40

50

60

70

80

90

100

5.0 6.0 7.0 8.0 9.0 10.0 11.0

GDP per capita (PPP, constant 2005 int'l dollar, logarithm

term)

Urbanization rate (%)

*1950 to 2010 (China as of 2012):

UK China Korea

India Spain Italy

US Brazil Japan

0

1000

2000

3000

4000

5000

6000

7000

0

10

20

30

40

50

60

70

80

1950 1970 1990 2010 2030 2050

China's contribution to world urbanization rate (left scale)

World urbanization rate (left scale)

World urban population (right scale)

MillionPercentage

China and world urbanization:

20

30

40

50

60

70

80

90

100

0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000

GDP per capita (RMB)

BeijingTianjinGuangdongShandongSichuanHebeiAnhuiHubei

ZhejiangGuangxiYunnan

Urbanization rate (%)



In Section 2, we identify three channels through which the urbanization process helps to

improve growth efficiency and achieve balanced growth, and where urbanization-related

reforms are needed:

Land reforms matter for the intensity of business (non-housing) investment.

City size and infrastructure matter for resource efficiency.

Social welfare reforms matter for household savings and consumption.

To complete the macro picture, we also look at the link between urbanization and housing

investment, although we do not expect a major change in policies here. Hence, the fourth

link we demonstrate quantitatively is:

Housing investment as a result of urban growth and policy controls.

Section 3 highlights policy areas that are important for gauging the future growth path. In

our view, more compact urbanization in the form of large cities or city clusters helps to

facilitate greater ‘economies of agglomeration’. But this requires the support of suitable

infrastructure and land reforms. Adequate social welfare for migrant workers is another

important element of more efficient and balanced growth. Recent policies are encouraging

in that it appears a more efficient strategy will be chosen.

In Section 4, we use the estimates obtained in Section 2 to set out three stylized scenarios,

differentiated by the three reform scenarios highlighted above: the use of land, urban

concentration and the effect of consumption-boosting policies. These reforms are

connected yet different: for instance, land policies help to promote appropriate urban

concentration, but are not sufficient on their own to achieve it. Thus the scenarios

summarize the joint impact from different reform dimensions. To make our analysis

tractable, we assume the same employment (and population) growth across all scenarios.

We also assume that housing investment will be under ongoing policy controls and thus

lower than the market equilibrium across various scenarios. Exhibit 4 summarizes the main

growth variables under these scenarios.

4. Scenarios under different urbanization strategies (annual average)

Source: GS Global ECS Research estimates.

Real growth or share (% or ppt)Baseline scenario

Concentrated and faster reform

scenario

Dispersed and slow reform

scenario

Avg real growth 2013-2016 8.1 8.2 8.1

Avg real growth 2017-2020 7.3 7.7 7.3

Avg investment growth 2013-2016 8.2 7.5 8.4

Avg investment growth 2017-2020 6.8 5.7 7.9

Non-housing investment growth 2013-2016 8.2 7.4 8.4

Non-housing investment growth 2017-2020 6.8 6.1 8.0

Investment/GDP 2013-2016 (share) 54.3 53.6 54.3

Investment/GDP 2017-2020 (share) 54.3 51.6 56.1

Household consumption/GDP 2013-2016 (share) 35.7 36.1 35.7

Household consumption/GDP 2017-2020 (share) 36.9 38.5 36.9

Urban emp growth% 2013-2016 1.7 1.7 1.7

Urban emp growth% 2017-2020 1.4 1.4 1.4



Our baseline model assumes a continuation of the past trend of land expansion and

city creation, and slow reform of social welfare. There is no extra efficiency gain

through city agglomeration. Non-housing investment/GDP would remain elevated

(although moderating slightly for cyclical reasons), while the consumption/GDP share

climbs only slowly.

The concentrated and faster reform scenario assumes that land use becomes more

conservative, that there will be greater efficiency gains through ‘agglomeration’ and

that social welfare reforms have a faster impact on household savings than in the past.

The new scenario would improve growth in three ways: 1) investment/GDP share

declines relative to the baseline; 2) TFP is boosted through better urban agglomeration;

and 3) consumption/GDP share improves faster. This scenario would be the most

positive path for urbanization, in our view, as it improves growth efficiency, reduces

investment intensity and burdens on land and other resources, improves consumption

and alleviates pressure on public finances. Indeed, the positive impact is likely to be

understated as larger cities tend to provide better job prospects, which we have

assumed to be the same across all scenarios.

The dispersed and slow reform scenario assumes that land use will become even more

scattered than in the past, no efficiency gains through city agglomeration and that

social security reforms remain gradual. TFP gains and consumption/GDP shares would

therefore follow the baseline, but the non-housing investment/GDP would rise even

faster. This scenario is the least favorable, in our view, as it implies a worsening of

investment vs. consumption imbalances, an even more significant burden on land and

resources than in the past, overheating pressures in the near term owing to the

significant increase in investment demand, and a deterioration of economic returns on

investment.

The policies discussed in recent months have increased the possibility that the second

scenario may be pursued, although we look for more clarity on the policy front in the

coming months.

Section 5 looks at the implications of our analysis for the economy and associated markets.



2. The multiple links between urbanization and Chinese growth

In this section, we describe several channels through which the shape of urbanization

affects growth balances and efficiency. In particular, we link the urbanization strategy and

policies to non-housing investment, growth efficiency, consumption and savings, and

housing investment. We look at each of these channels separately and estimate the

quantitative linkages between the macro variables and urbanization-related development.

Conscious that urbanization is endogenous to growth, in our estimation we also account

statistically for the potential reverse causality where possible (see, for example, Box 2).

A. Urbanization and high investment share: a more conservative

use of land key for bringing down the investment/GDP share

It is well known that China’s investment as a share of GDP is high by international

standards. Moreover, investment as a share of GDP has risen in recent years, resulting in a

deterioration of economic returns on investment. As an example, the Incremental Capital

Output Ratio (or ICOR), calculated as the incremental change in growth relative to

investment, has deteriorated. While the low cost of funding is one of the key factors

explaining this high investment share [see Aziz and Cui (2007)], it does not explain the

recent increase or the deterioration in investment returns, in our view.

To account for the rising investment share, we believe a more detailed breakdown of

China’s investment composition is necessary. Compared with other countries, it is striking

that investment in structures—related to construction activities such as infrastructure and

buildings for industrial use—is dominant in China. Although the share of investment in

structures within the total declined during 2003-2008, the process has stalled since then

and the share of construction-related investment remains well above 40% (Exhibit 5), and is

higher than in other major economies (Exhibit 6). At the current level, China’s investment

structure remains similar to that of UK in the 1870s at a much lower income level.2

Investment in structures tends to be associated with a lower ICOR, while investment in

equipment is more directly related to technological innovation and productivity gains, and

thus a higher ICOR. De Long and Summers (1991) have demonstrated the link between

investment in equipment and economic growth for a broad group of countries. While many

countries have had high shares of investment in structures at some point during their

development, the share generally trends down over the longer term (Ball, et al, 1996);

indeed, a reduced intensity of investment in structures was related to productivity

enhancement in countries that were industrialized early, such as the UK. Therefore, we

expect a decline in the share of investment in structures to see investment efficiency rising.

2 Over the longer term, the relative share of investment in construction falls relative to equipment investment increases across OECD countries. This is associated with relative price changes: as land prices rise and equipment prices fall, businesses find it more profitable to invest in equipment while preserving the cost of land use. See Ball, et al (2001) for further discussion.



5. China's investment dominated by structures

6. China’s investment in structures is higher than in other

countries

Source: CEIC, China National Bureau of Statistics. Note: For China, investment in structures (non-housing) is derived by deducting residential

housing investment from total construction. As the residential housing investment includes the purchase of equipment, the investment in structures may be underestimated.

Source: Haver Analytics, OECD, China National Bureau of Statistics.

The strength of investment in structures in China coincides with the remarkable expansion

of the city landscape, which has outpaced urban population growth. We estimate that the

built-up urban area has expanded by near 40% over the last decade, while the population

has increased by less than 30% (Exhibit 7). Urban land expansion has been brisk in the

less-developed Western region, as well as in the export-oriented Eastern regions, while

larger cities such as Beijing saw the slowest urban land expansion (Exhibit 8). Urban

population growth, by contrast, was the fastest in Beijing and Shanghai, and was much

slower in the West and Eastern regions. Thus, much of the city expansion cannot be

explained by population growth or urbanization needs. Indeed, apart from a few big cities,

the increase of urban density has been unremarkable, and the overall density is modest by

international standards.

In our view, both the strength of investment in structures and the rapid expansion of land

use for cities in China reflect current land policies, which allow the conversion of

agricultural land into urban construction land at below-market prices (IMF 2011, Li 2008,

and Deng 2004). Rural land is collectively owned, and local governments have the

discretion to expropriate land at a low price and sell the industrial land use rights for a

period of up to 50 years (IMF, 2011). Systematic data on land expropriation prices is not

available but examples of farmland expropriated at low prices are abundant. One of the key

reforms announced is to grant explicit land ownership and certificates for farmers’

homesteads and contracted land. This year, for the first time, farmers' property rights were

included in the government report3, an encouraging sign that agricultural land

expropriation system is likely to be reformed. A more market-oriented land pricing scheme

is an important first, but arguably not sufficient, step towards more efficient use of land

and less urban sprawl.

Empirically, we estimate the links between the expansion of the urban area and the non-

housing investment share of GDP, based on a panel dataset that covers 31 provinces

during 2003-2011 (Box 1). We find that the expansion of urban land is important in

explaining the rising investment share of the economy. Our estimates suggest that slowing

growth in the amount of land used for construction by 1.5ppt each year would help to

reduce the investment/GDP ratio by 3-6ppt in 10 years, with the impact mostly back-loaded

given the time needed for the effect to accumulate.

3 In China’s No 1 government report in 2013, released in February, the central government requires a complete registration of rural collectively-owned land and that farmers’ contract rights are assigned to the collectively-owned land within five years.

0

10

20

30

40

50

60

0

10

20

30

40

50

60

1995 1998 2001 2004 2007 2010

China: Investment by types(share in total)

Residential housing

Equipment

Others

Non-residential structure

42

36

36

33

23

20

China

Japan

UK

Korea

Germany

US

Structure (non‐housing) Investment as a Share in Total (%)



7. Urban land expansion outpacing population growth 8. Urban land expansion by region

Source: China National Bureau of Statistics, CEIC. Source: CEIC, China National Bureau of Statistics.

Box 1: The impact of urban expansion on capital investment

Chinese cities have been expanding at a faster pace than

the urban population, suggesting that the urbanization

process has led to a high intensity of land use. Not only

has this raised concerns about food supply (although

arable land is now subject to strict policy controls), it also

has an impact on the investment-heavy model of growth.

According to existing laws, individual farmers do not have

ownership rights and agricultural land is collectively

owned by rural cooperatives. Local authorities have the

ability to convert agricultural land into land for urban

development at a below-market price. This has made one

of the key inputs of production—land—too cheap,

spurring investment growth, in particular in construction

activities. This can be seen from the high correlation

between the growth in the investment/GDP ratio and that

of land expansion. To quantify the impact, we link the

land size with non-housing investment growth, using a

panel dataset that covers 31 provinces during 2003-2011.

Exhibit B1: Land expansion positively relates to non-

housing investment

Source: GS Global ECS Research.

The model follows a stylized framework that Li, Yin and

Li (2008) use to analyze the impact of land size on capital

intensity.

ln (I/Y)t = a + b x ln (I/Y)t-1 +c x ln (area) t-1 +e

Where I is non-residential GFCF, Y is output and area is

the size of the land under construction. We use the built-

up area in cities. The first lag of the investment/GDP

ratio is used as an instrumental variable for all other

relevant explanatory variables. The model is then

estimated with generalized method of moments (GMM)

to overcome the statistical issues related to the lagged

dependent variables. Our analysis suggests that land

area has a strong and positive link with non-residential

investment, and that the elasticity of investment share

with respect to the land area has increased in recent

years. The estimated results based on the last five years

of the sample period are presented below (Exhibit B1).

The results suggest that slower land expansion has a

meaningful impact on investment intensity. Land

expansion that is one-third slower than the pace (i.e.

urban land size increases by 3.5% each year compared

to 5% before) could lower the investment-to-GDP ratio

by up to 3-6ppt compared with the baseline scenario

over ten years.

We also use the provincial level of land under

construction as an alternative measure of urban land

areas, and the results are very similar.

100

105

110

115

120

125

130

135

140

100

105

110

115

120

125

130

135

140

145

2003 2004 2005 2006 2007 2008 2009 2010

Construction land and population growth:

Area of built district

Urban population

Index: 2003=100 Index: 2003=100

100

110

120

130

140

150

160

170

180

100

110

120

130

140

150

160

170

180

2003 2004 2005 2006 2007 2008 2009 2010 2011

Urban built areas(2003=100)

Beijing

West and central

NE

East

Index Index

Dependent variable: ln (I/Y)

Variable Coefficient Std. Error t‐Statistic Prob.

Ln (I/Y,‐1) 0.21 0.00 61.88 0.00

Ln (AREA,‐1) 0.59 0.03 22.73 0.00



B. Urbanization and productivity gains: Exploring the

agglomeration effects of cities

Cities play two key functions in a country’s development. The first function is the

shift of surpluses into more productive use. As experienced by many initially agriculture-

based countries, a shift in surplus resources from the agricultural sector into

manufacturing and services with higher productivity helps to enhance factor productivity.

This has been the experience of many developing countries, as highlighted by Lewis (1954).

The second and more lasting advantage of cities is the ‘agglomeration’ effect. As cities

become more densely populated, city dwellers tend to become more productive through

the technology spillover and knowledge sharing permitted by close proximity (Henderson,

2004), but this requires appropriate infrastructure to offset the negative side-effects of big

cities such as traffic congestion and pollution. These agglomeration effects can occur

through businesses or consumers. Business agglomeration increases the productivity of

firms and workers, while consumer agglomeration creates demand for more diversified

consumer needs. One example of such agglomeration is the positive income effect from

larger populations evidenced in US cities (Carlino, 2011).

In China, city growth has so far been mostly about shifting resources away from the agricultural sector; agglomeration effects have been absent.

China’s urbanization has generally succeeded in shifting surplus resources towards more

productive use—including labor and land. Urban job growth has averaged 3.3% per year

over the last decade, creating 146 million urban jobs since the late 1999s, or a 40% increase

in urban employment since then. Since the productivity of rural workers remains about

one-sixth that of urban workers, shifting rural workers into urban areas should continue to

be a key driver of future productivity growth. This is in contrast to a popular view that

China’s labor surplus is now coming to an end, and that China has come close to a “Lewis

turning point” whereby the quantity of labor has become a key constraint.

Such resource shifts have continued to bring productivity gains in recent years, but the pace

has slowed. To see this trend, we rely on a growth model to decompose the contributions of

various production inputs (not including land for simplicity). There has been a marked

increase in the capital contribution to growth, and a reduction in TFP in recent years (Exhibit

9). We then further decomposed the TFP gains into those related to resource shifts

(estimated from urban job growth and urban and rural productivity differentials) and the

residual which represents the efficiency gains from other sources. It is striking that in recent

years the internal shift in resources accounted for most of the TFP gains.

9. TFP gains from resource shifting sectors have continued but other TFP gains diminished

post GFC

Source: CEIC, GS Global ECS Research estimates.

0

1

1

2

2

3

0

2

4

6

8

10

12

14

1997-2003 2004-2007 2008-2012 2013-2015 (forecast)

China's growth accounting: diminishing TFP growth post GFC:

TFP from resource shifting into urban areas

TFP not related to resource shift

Contribution of capital

Percentage points Percentage points



10. Agglomeration economies across US cities…

11. …and the lack thereof in Chinese cities

Source: US Census Bureau.

Source: CEIC.

The reduced scope for ‘easy’ productivity gains suggests the need to search for new areas

of efficiency improvements. One such area is the positive efficiency gains through city

agglomeration, which has been largely absent in the urbanization process so far. A simple

(though imperfect) test of such an agglomeration effect is the comparison of per capita

income and population size. If the agglomeration effect is strong, through the positive

externality, larger cities should be associated with higher income. Across US cities, for

instance, a doubling of city population is associated with a 3%-4% increase in the average

wage (Exhibit 10).4

In China, however, the association of population size and per capita income is weak. In

contrast to US cities, where income and population size tend to be positively linked, in

China such an effect is muted (Exhibit 11). In fact, formal tests suggest that when variables

corresponding to regions, administrative levels of cities and initial conditions are controlled

for, income and population tend to be negatively correlated across cities (Box 2). If Chinese

cities exhibit the same ‘agglomeration economies’ as US cities by the end of the coming

decade, the real rate of growth could be boosted by an additional 1ppt relative to the

baseline by the end of the decade. We take this as the potential magnitude of the efficiency

boost through ‘economies of agglomeration’ in section 5.

4 Note that the slope of Exhibit 10 should not be taken as an agglomeration impact itself. If agglomeration effects are present, larger cities should provide higher income. However, there is reverse causality as well– a city with higher income attracts people and leads to the expansion of the city. Once such sources of potential overestimation are controlled for, a 1% larger population is associated with a 0.038% increase in per capita income in US cities (see Carlino [2011]).

2.5

3.0

3.5

4.0

4.5

10 11 12 13 14 15 16 17 18

log (population)

log

(city

inco

me)

US cities: Income and population size

3

4

4

5

5

6

10 11 12 13 14 15 16 17 18

log

(per

cap

ita in

com

e)

log (population)

China: population and income across cities



Box 2: Test for the ‘agglomeration effects’ in Chinese cities

One of the direct tests of the ‘agglomeration effect’ in cities is the positive association between average wages in a city

and the city’s population size. The basic argument is that if agglomeration effects are important, wages and demand for

labor should both reflect the advantages of ‘agglomeration economies’ (Carlino, 2011). The simple test of this

relationship for Chinese cities does not support a positive association of income and population size, in contrast to the

clear and strong positive association in US metropolitan areas. To be sure, such a positive association of city size and

average wage represents the upper bound of the agglomeration effects in cities. If agglomeration effects are important,

workers in large cities should be more productive than those in small cities. At the same time, more productive workers

may be attracted to large cities, exaggerating the agglomeration effects. Once these sources of potential overestimation

of agglomeration effects are accounted for, a doubling of the metropolitan population is associated with 3.8% increase

in average income in the US.

We test the links between the per capita income and population in cities by controlling for the regional income (by

including the provincial average income), the initial condition (the city’s per capita income in 2002), and administrative

importance (a dummy variable corresponding to provincial capitals). In all specifications, the level of city income is

actually negatively related to the population size. This supports the view that agglomeration effects in cities are muted

at best, or even negative (Exhibit B2).

For our simulation exercise, we assume that the impact of city agglomeration on growth efficiency in China would catch

up with that of US cities in 10 years. More specifically, a 1% larger population is associated with a negative 0.1% decline

in average income, when other factors related to the average regional income, the administrative level in the cities and

initial income are controlled for. This compares to a positive income gain of 0.038% from a 1% larger population in US

cities. We use the difference of 0.13ppt in elasticities as the ‘agglomeration effect gap’ between Chinese cities and US

cities. Assuming the same urban population growth as in the baseline, but Chinese cities ‘catch up’ to their US

counterparts in this agglomeration effect over the next decade, the average income growth would be 13% higher than

the baseline. We assume this is achieved in 10 years under the more concentrated model, and this would correspond to

about a 1ppt gain in TFP growth (cumulative) relative to the baseline by the end of the decade, or an additional TFP

growth of 0.1ppt each year.

Exhibit B2: Arable land per capita in China is low

Source: China National Bureau of Statistics, CEIC.

We also augment the analysis by estimating the links between urban population and urban income growth using

provincial level data, where data on urban population is directly available (by contrast, the city population may include

those who live in the surrounding rural areas). We run a regression of the average income growth on urban population

and population density, while controlling for other variables such as initial income, external trade and investment

growth. We find that while the cross-region growth differentials are strongly related to the initial income (via internal

catch-up, and thus a negative association) and investment growth, the urban density tends to contribute little or even

negatively. This is consistent with our findings at the city level.

Dependent variable: ln_cityincome_2010

ln_cityincome_2010 ln_cityincome_2010 ln_cityincome_2010 ln_cityincome_2010

coef p‐value coef p‐value coef p‐value coef p‐value

ln_city popultion ‐0.09 0.08 ‐0.11 0.00 ‐0.12 0.00 ‐0.13 0.00

ln_provinceincome 1.02 0.00 0.25 0.00 0.28 0.00

ln_cityincome2002 0.72 0.00 0.70 0.00

Pro_capital(=1 for provincial capitals) 0.11 0.05



Agglomeration can have positive effects on growth through knowledge spill-over, a deeper

labor market and the diversification of activities. The absence of an agglomeration effect in

urban growth suggests that the potential efficiency gains from bigger cities have been

overshadowed by the negative side-effects, such as transportation costs, congestion and

pollution. We see two likely reasons for this:

The rapid expansion of cities mentioned above: With the exception of Tier 1

cities (which are among the largest cities in the world, see Exhibit 12), for most of

the other large cities, such as Chongqing and Wuhan, the average density remains

quite low (and even declining), given the fast extension of the city limits.

Inadequate infrastructure, in particular public transportation. There is an

undersupply of critical city infrastructure, such as subways and water systems.

Greater agglomeration effects are a critical aspect of why a more concentrated model of

urban development is desirable in the case of China. We discuss this in more detail in

Section 3.

12. Chinese Tier 1 cities are among the largest in the world (2011)

Source: Domographia World Urban Areas.

C. Urbanization and household savings: Migrant savings key to the

rising household savings rate

We find the lack of access by migrant workers to essential city public services to be a main

reason for the rising household savings rate, which has been one of the key challenges to

rebalancing the Chinese economy and promoting domestic consumption. We estimate the

average urban savings rate to be 10ppt higher than that in rural areas (Exhibit 13). We

established a link between the increase in household savings and the lack of access to

adequate public services by migrant workers through an empirical test of the panel dataset

of Chinese provinces (see Emerging Markets Macro Daily: The hukou system: Holding back

China's rebalancing, May 1 2013).

0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

50,000

0

5

10

15

20

25

30

35

40

Density (RHS)

Population

persons per square kilometerMillion



13. Urban household savings rate higher than the rural rate


A simple statistic simulation suggests that in the absence of any change, savings could

continue to drift up along with city migration. For instance, for each 14 million rural

residents moving to urban areas—corresponding to about 1% of the total population

shifting from rural to urban works each year—the average savings rate could increase by

0.1ppt per year.5 Our previous work suggested that removing hukou-related differences in

social services would boost household consumption/GDP share by up to 6ppt over 10 years.

Assuming a slower-paced reform than in our previous work, we estimate that 2-3ppt can be

added to the consumption/GDP share over 10 years.

From a macroeconomic perspective, curbing the rising household savings rate is essential.

All else equal, a rising savings rate is an underlying deflationary source given the demand

‘leakage’. According to some economists, increasing urbanization helps to support

aggregate demand, as it is boosted by spending by migrant workers. However, we think

this argument is dubious, since migrant workers will add to both supply and demand.

Unless migrant workers consume a larger share than they produce, the average savings

rate of the urban household is likely to rise rather than fall.

Therefore, hukou reforms and associated social welfare reforms provided to migrant

workers will be an essential part of reforms for sustained growth.

5 In our forecast, we expect that each year, about 1% of the total population will shift from the rural area into cities. Assuming each one migrant increases her savings rate from the average rural level (25%) to the average urban level (37%). This has the net effect of increasing the average savings rate of households by about 0.1 ppt every year, all else equal.

15

20

25

30

35

40

15

20

25

30

35

40

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

Household savings rate (% of disposable income):

Rural savings rate

Urban savings rate

Percentage Percentage



D. Urbanization and housing: Higher income, more housing

consumption

We also link urbanization to China’s housing supply. Our estimation based on a cross-

province panel data suggests that housing investment is driven mainly by income growth

and demographics, which implies that housing supply is broadly consistent with the rising

consumption needs of households (Box 3). These factors suggest that housing investment

as a share of GDP should hold up, even with the ongoing policy control:

Housing is part of consumption. As of 2012, the urban housing stock stood at

around 20 million square meters, or roughly 27 square meters per person. We

estimate that about 45% of urban housing was built after 2000 (housing reform

started in 1998, since when more modern residential housing has been

constructed). The newer houses are larger in size, even though household sizes

have been declining (Exhibits 14 and 15). From a growth perspective, while not

contributing to the country’s productive capacity itself, housing investment has

met the needs of urban demand.

Compared with other countries, this consumption is not leveraged…

Households have generally not borrowed to finance their homes. An estimated

20% of households borrow from banks to finance their houses. A lack of

alternative investment of household wealth is another factor driving housing

demand.

…And it will continue to benefit from urban income and population growth. Housing investment is now about 7% of national GDP in real terms. We

estimate that housing investment is likely to stay strong before its share in GDP

declines after 2015. However, the current policy control is likely to keep housing

investment more subdued, in our view.

14. Housing stock driven by rising consumption

15. Newer flats are larger

Source: CEIC, China Population Census.

Source: CEIC, China Population Census.

22

23

24

25

26

27

28

29

0

5000

10000

15000

20000

25000

1995 1997 1999 2001 2003 2005 2007 2009 2011

Urban housing stock and size per person

Urban housing stock (mln sqrm)

urbanhouse/person (sqrm, RHS)

0

10

20

30

40

50

0

25

50

75

100

built in 1970s built in 1980s built in 1990s built in 2000s

City houses by construction time

avg size per flat

share in total housing stock(RHS)



Box 3: What is driving housing investment in China?

China’s housing investment has increased substantially in

recent years, from less than 3% of total GDP in 1996 to

nearly 8% of GDP in 2012, before cooling down somewhat

following government controls introduced last year. The

strong investment growth was partly due to the low

starting point and housing reforms. New and more

modern commercial housing in cities started to replace

the dormitories prevalent until the late 1990s when China

started its housing reform, and households have been

upgrading their housing consumption.

As of 2010, about 40% of the total housing stock had been

built after 2000. The newer houses are larger, even though

household sizes have been shrinking, implying an

upgrade in terms of living space.

Across regions, the housing investment/GDP ratio has

generally climbed over the past decade, except for a

decline in Beijing and Shanghai during 2003 and 2009,

before a strong rebound occurred there too (Exhibit B3).

In general, housing investment as a share of GDP tends to

be higher in areas where population growth is faster but

the correlation is not one-on-one (Exhibit B4). This is

because income growth is just as important a driver of

housing investment.

We estimate the drivers of economic fundamentals using

a panel dataset for China’s provinces in the last 6 years.

Income level and population growth are significant,

explaining the cross-regional differences in housing

investment/GDP ratios. In particular, we estimate the

following equations:

ln(HI/Y)it =a+ b x ln(HI/Y)it-1 +c x ln(GDPPC)it-1 +d x

urbanpop_git-1 +e

Exhibit B3: The share of housing investment in GDP

varies by region


Exhibit B4: Population growth displays diverging trends

across regions


Where HI, Y, GDPPC, urbanpop_g represent real housing

investment and GDP, per capital income and

urbanization population growth, respectively. Again, a

GMM estimation method is employed to account for the

statistical issues introduced by the lagged dependent

variables on the RHS. The estimation results are

presented in Exhibit B5. Using the estimated

coefficients, and expected urban income and population

growth, we are also able to project the housing

investment/GDP ratio in coming year. However, in the

next few years, how housing investment/GDP develops

will also depend critically on the government’s policy

aimed at curbing property price increases (see Riding

the Chinese construction waves, October 10, 2012). The

scenarios, with or without government controls, are

depicted in Exhibit 26.

Exhibit B5: Housing investment explained by income

and population growth

Source: GS Global ECS Research.

2

4

6

8

10

12

14

16

2

4

6

8

10

12

14

16

1999 2001 2003 2005 2007 2009 2011

East ex BJ SH

Central

West

BJ & SH

China regional residential investment as share of GDP:

% share of GDP % share of GDP

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Capital area East Central West

Urban pop growth (avg %)2005-2008

2009-2011 (exclude2010)

yoy% yoy%

Variable Coefficient Std. Error t‐Statistic Prob.

LN(HI/Y,‐1) 0.88 0.01 60.90 0.00

LN(GDPPC,‐1) ‐0.09 0.03 ‐3.28 0.00

Urbanpop_g(‐1) 0.005 0.000 11.90 0.00

Dependent variable: ln (HI/Y)



3. Next stage of urbanization to be driven by key policy changes:

Land reforms, city agglomeration, savings

A. A more concentrated approach is preferable

The debate is ongoing as to whether future urbanization should take the shape of city

clusters around large regional centers, mid-sized cities or a large number of scattered

townships. One popular view is that farmers should be ‘urbanized locally’, which means

absorbing the surplus labor into non-agricultural jobs, without further concentration of the

population. In our view, however, a dispersed model implies duplicate investment in

infrastructure and inefficient use of land and other resources. We see a number of reasons

why a more concentrated model is a preferred approach.

1. More diversified domestic demand, benefiting job growth

As a diversified and dynamic services sector will be the main driver of job growth in

coming years, the deeper labor market in larger cities or city clusters are better placed to

create jobs for both skilled and unskilled workers.

Surplus labor is still abundant. We estimate that rural workers still account for about 30%

of the total working age population, and labor productivity in rural areas remains one-sixth

that in the cities. Therefore, shifting the surplus rural labor into non-agricultural jobs

continues to be an important source of productivity gains. The higher rural birth rate (the

average birth rate in rural areas relative to cities is about 1.2:1, and this is expected to rise

given the demographics) also suggests that the natural growth of the rural population

could potentially outpace that of the urban population in coming years.

Meanwhile, the ability of the manufacturing sector to absorb surplus labor has become

more constrained than in the past, in light of the reduced role of the tradable sector in the

economy. Future surplus labor will therefore increasingly be absorbed by the services

sector. The incremental growth of manufacturing jobs has been strongly linked to export

performance in the past decade. Based on the historical relationship, and assuming exports

grow at 5%-10% a year (given the large size of China’s exports and the high market

penetration in some areas), manufacturing jobs are likely to rise by 0.5%-1% a year. This

means about 1 million new jobs can be created in the manufacturing sector (compared

with 5-7 million a year before the global financial crisis [GFC]), and the rest will need to be

created in the construction sector and other services. Thus, we expect the migrant

population to average 14 million and workers to average 9 million a year, compared with

21 million and 15 million in the past five years (Exhibit 16).

The continued need for cities to absorb surplus labor from rural areas, mostly at the lower

end of the job ladder, while creating more skilled jobs for college graduates can be better

met by a more diversified job market in clusters of larger cities.



16. We expect the urban population and employment growth to moderate in coming years

Source: CEIC, GS Global ECS Research estimates.

2. Helping to improve investment efficiency

Strong investment growth in the past few years and the slowdown in trend growth have

meant that the capital-to-output ratio, which remained largely steady during most of the

2000s, has risen sharply in the last few years, suggesting a deterioration in investment

efficiency.6 The scope to urbanize through more intensive investment has become even

more constrained. As shown in Section 2, increased land use has been associated with the

high investment intensity of growth, suggesting a more conservative use of land would

help to boost investment efficiency.

3. Exploring the agglomeration effect of cities

As illustrated earlier, although productivity gains have continued through shifting

resources into cities, productivity gains associated with other sources have been limited in

recent years. The strong advantage of cities relative to rural areas lies in the ability of cities

to organize resources, both for producers and consumers.7 This is particularly important for

China as its vast population contrasts with its limited arable land, and the scarcity of water

and energy.8 A more concentrated model can balance the manageability of cities and the

efficiency gains, and help to preserve resources.

6 To measure the capital intensity of the economy, we make an important yet often ignored adjustment to the method of estimating capital stock. Some of the increase in the recent investment growth was ‘non-productive’ in the sense that residential investment as a share of the total has increased substantially. We estimate it to account for more than 7% of national GDP, compared with about 2% in 2004. Estimating the capital stock in the economy thus needs to exclude this part of investment, which is accrued to households rather than the corporate sector. Once this adjustment is made, the rise in capital intensity and the decline in investment efficiency (suggested by ICOR) are less pronounced. Nevertheless, it remains the case that capital intensity is now rising along with the investment share, in contrast to the early 2000s when the rise in the investment share led to strong GDP growth, leaving investment returns largely unchanged. Thus, at least from a cyclical perspective, investment efficiency has declined. In particular, the K/Y ratio, which has stayed steady at around 2.5 for many years, has risen above 3. 7 Greenstone, Hornbeck, and Moretti (2010) attempt to measure the impact of agglomeration on firms’ productivity, and find that opening a new plant boosts the existing firms’ productivity by 12% in cities where the new plant is located relative to other cities. 8 The amount of arable land per capita in China is quite low by international standards. China has 9% of the world’s total arable land and 20% of its total population. China’s per capita arable land is one of the lowest globally. Urbanization in the past 20 years has reduced the amount of arable land, for two reasons. First, arable land has been converted into construction sites. Between 1993 and 2004, total arable land declined by near 10%, largely due to this land conversion. More than half of

0

1

2

3

4

5

6

7

0

1

2

3

4

5

6

7

1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020

Urban employment growth

Urban population growth

year-over-year percentage change year-over-year percentage change

China urban population and employment:



4. Lower burden on public finances

A further increase in debt financing by government and government-related entities is

likely to be unavoidable in future urban growth, given that:

Urban development so far has been financed both by debt borrowing and land

sales. Land reforms should in principle limit the revenue source from land sales,

with the explicit ownership rights granted to farmers. This increases the need to

rely on debt financing compared with in the past.

Although the envisaged greater private-sector participation should help to bring in

private-sector expertise and discipline, public-sector guarantees or participation is

perhaps still needed given the inherent riskiness of infrastructure projects.

This comes at a time when local government finances are already under heavy scrutiny. A

more concentrated model is likely to help reduce the gap in public finances given the lower

investment intensity.

B. Land policies and better infrastructure are key ingredients for

more compact cities

1. Reforms to ensure a more efficient use of land

The recognition of farmers’ land rights is the first step towards more discipline and greater

efficiency in land use. A scheme being tried out is direct market-based pricing and

transaction, such as the auction scheme in Chongqing, where ‘land certificates’ are backed

by underutilized rural land. In principle, this allows for a better relationship between the

price of rural land and the marginal returns from converting it into urban land, thereby

reducing excessive appropriation for city development.

In our view, such a market-based system is important, but it may not be sufficient to curb

urban sprawl on its own. In fact, the system has been designed to free up more rural land

for urban use (but in a more equitable manner), rather than to limit urban land growth itself.

Containing urban expansion will require an explicit set of policies in land management,

reuse of vacant brown-field areas, and a greater focus on public transport.

2. Recent infrastructure plans are right for more compact and efficient cities

Better managed cities can deliver more visible efficiency gains and overcome the negative

side-effects of large cities. This also implies shifting needs for infrastructure:

There is a lack of city infrastructure such as water treatment plants (Exhibit 17) and

utilities, as well as environmental protection in China. There is a critical need to

correct this.

Transportation needs are changing. The length of city subways in the newly

announced metro plans is closely related to the population size in surrounding

areas, and is in line with international experience (Exhibit 18). By contrast, roads

have already been built to meet the population distribution, even though for the

nation as a whole road density remains low (Exhibits 19-20). More compact cities

suggest a preference for public transportation, rather than roads.

the land used for construction in cities (including property, transportation and water systems) during 1998-2004 was formerly arable land. Since then, the government has implemented strict rules against the improper conversion of arable land, and the encroachment of arable land has moderated. Second, the quality of the remaining agricultural land may have declined, since land used for construction is often the best arable land close to city centers.



17. China’s wastewater treatment still has a long way to

go

18. Expected length of subways closely correlated to the

expected population size

Source: CEIC, World Bank.

Source: CEIC, China NBS, and various national agencies of statistics.

19. China’s road density is mostly compatible with

where the population is located…

20. …even though for a nation as a whole road density is

low relative to other countries


Source: CEIC, World Bank.

30

40

50

60

70

80

90

100

30

40

50

60

70

80

90

100

Korea Singapore OECD Malaysia Thailand Philippines China BRICS Indonesia India

% of Population with Access to Improved Sanitation Facilities

2010

% of population % of population

Beijing

Tianjin

Guangzhou

Nanjing

Chengdu

Shanghai

Shenzhen

Seoul

New YorkLondon

TokyoMoscow

Chicago

Frankfurt

0

200

400

600

800

0 5000 10000 15000 20000 25000 30000 35000 40000

Population including those in the suburbs (000s)

Expected length of subway by 2017 (kms)

Henan

Shandong

Jiangsu

HubeiHunan

Anhui

Guangdong

Zhejiang

Guizhou

Jiangxi

Shanxi

0

20

40

60

80

0 50 100 150 200

population and road density by province

Road density: km per 100 square km(2011)

Population density: mn person per 100 square km

UK

Italy

Spain

India

Korea

Japan

US

China

Brazil

0

10

20

30

40

50

60

0 50 100 150 200

Population and road density by country:

Population density: thous person per 100 square km

Road density: km per 100 square km



C. Comprehensive social security reforms supported by fiscal

reforms needed to stem the rise in household savings

Since last year, policies have been put in place to enable migrant workers to obtain city

residence permits, or hukou, based on certain thresholds that vary by the size and

administrative levels of cities. Other social security services—the health care, education,

and pension systems—are also being reformed to reduce the level of precautionary

savings across urban and rural areas (Exhibit 21).

The hukou reform needs to be supplemented by fiscal reforms, in our view. Local

governments are concerned about the extra public spending needed to pay for the cost of

incorporating migrant workers into cities. Although such spending is necessary to correct

the previous misrepresentation of the true costs of city and company growth, the resulting

fiscal gap will require adjustments in the tax system, for instance by extending the

introduction of property taxes to more cities.

21. Recent social security

Source: Xinhua.

Recent changes Impact TimelineSocial welfare reforms

Hukou

Granting hukou depending on city levelsTownships: those with stable jobs and accommodation qualify immediately Small cities(<1m population): those with stable jobs/accommodation for at least three years and at least one year of social security contribution qualify Large cities: continue to restrict population

Expanding hukou coverage Feb-12

Rural medical systemNew Rural Cooperative Medical System (NRCMS);- In 2013, China raised the annual NRCMS compensation per capita from Rnb240 to Rmb280 (Rmb30 in 2003)

More social insurance to rural area residents Effective in 2013

Health care-drug pricingEssential Drug List (EDL) expansion: According to Ministry of Health (MoH), EDL is to be expanded to include more than 500 drugs from 307

Lower the drug costs for patients Dec-2012

Health care-severe illness insuranceExpansion of the coverage of healthcare insurance system to include the treatment of critical illnesses

Enhance social safety net Aug-2012

Urban pension reform

A new urban pension program was launched in July 2011. The program aims to help create full coverage for China's basic social pension system. Insured residents over the age of 60 will be able to receive a basic monthly pension of 55 yuan

Enhance social safety net

60% urban area covered in 2011; 100% urban area covered by end of 2012



4. Growth and sectoral impact under different urbanization

scenarios

In the above discussions, we estimated the impact of the urbanization process on non-

residential investment, residential investment and household savings, based on the

historical data. We also estimated the ‘agglomeration effect’ in China compared with that

of the US. We use these frameworks to simulate the future likely path of these variables

under different scenarios.

Impact of land use on investment (non-housing) intensity of growth: We rely on

the model estimated in Box 1, which links the non-housing investment/GDP to the pace of

land use. The trend model assumes the size of the cities will expand at about 5% a year, a

similar rate to that of recent years. The concentrated model assumes cities expand at a

pace that is one-third slower, or about 3.5% a year, and the dispersed model assumes an

even more rapid expansion of urban land growth of about 6% a year. These scenarios are

mapped into the paths depicted in Exhibits 22 and 23.

Impact of agglomeration on growth efficiency or TFP: We assume that under the

more concentrated model there will be additional average TFP growth of about 0.1ppt a

year, with an accumulated TFP gain of 1ppt in 10 years. This would allow the

agglomeration effect in Chinese cities to catch up with that of the US, as described in Box 2

and Footnote 4. The effect is also assumed to be back-loaded as the impact accumulates,

implying additional efficiency gains from agglomeration of about 13% in a decade. Such

effects on growth itself are mapped into the paths in Exhibit 24, and we can then derive the

investment growth scenarios in Exhibit 25.

22. Investment efficiency is higher under a more

concentrated model…

23. …as investment/GDP share moderates

Source: CEIC, GS Global ECS Research.


0

1

2

3

4

5

6

7

8

0

1

2

3

4

5

6

7

8

1996 1999 2002 2005 2008 2011 2014 2017 2020

ICOR (investment efficiency)

Baseline

Disbursed

Concentrated

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.0

0.1

0.2

0.3

0.4

0.5

0.6

1996 1999 2002 2005 2008 2011 2014 2017 2020

Non-residential investment/GDP (%)

BaselineDisbursedConcentrated



24. More concentrated model associates higher

growth…

25. …with less investment



Impact of urbanization and policy controls on the share of housing investment: We also forecast the share of housing investment in GDP in the coming years, based on the

historical relationships between the residential investment/GDP ratio, income and

population growth estimated in Box 3. However, as government policy controls will remain

in place and these have, in our view, drawn a wedge between the model-simulated

housing investment/GDP rate and the actual rate, the forecast housing investment/GDP

share is adjusted downwards, assuming the policy impact will be most significant in the

near term (Exhibit 26).

Impact of urbanization and reforms on the share of consumption in GDP: The

household consumption/GDP rate rose in 2012 (to 35.3% from 34.9% in 2011) after a

persistent decline over more than a decade. The improvement was both cyclical (as

investment slowed more than consumption) and structural (as government policies have

supported household income growth, offsetting some of the impact of rising savings on

the consumption/GDP share). Our baseline model assumes such an improvement will

continue in the coming year as a result of policies aimed at boosting household income

growth, but the high and rising savings rate is likely to result in a modest pace of

improvement (with the consumption/GDP share rising by about 2ppt in 10 years). However,

strong reform efforts could add as much as 2-3ppt to the consumption/GDP share over 10

years (this is a more conservative estimate than our earlier work, as the reforms are likely

to be implemented gradually) (Exhibit 27).

0

2

4

6

8

10

12

14

16

0

2

4

6

8

10

12

14

16

1996 1999 2002 2005 2008 2011 2014 2017 2020

Real growth (%)

BaselineDisbursedConcentrated

0

5

10

15

20

25

0

5

10

15

20

25

1996 1999 2002 2005 2008 2011 2014 2017 2020

Investment growth (yoy, %)

Baseline

Disbursed

Concentrated

yoy% yoy%



26. Housing investment more subdued under ongoing

policy controls

27. Reforms needed to boost consumption



We then combine these variables to project three urbanization paths. We assume the same

job creation and housing policies in all scenarios, and focus on the impact of land and

social security reforms (Exhibit 5).

Our baseline model assumes a continuation of the past trend of land expansion and

city creation, and a slow pace of social security reform. There is no extra efficiency

gain through city agglomeration. Non-housing investment/GDP remains elevated

(although moderating slightly for cyclical reasons), while the consumption/GDP share

climbs only slowly.

The concentrated and faster reform scenario assumes that land use becomes more

conservative, that there are greater efficiency gains through agglomeration and social

security reforms deliver a faster impact on household savings than in the past. The

new scenario improves growth in three ways: 1) the investment/GDP share declines

relative to the baseline; 2) TFP is boosted through better urban agglomeration; and 3)

the consumption/GDP share improves faster. This scenario is the most positive path

for urbanization, in our view, as it improves growth efficiency, reduces investment

intensity and the burdens on land and other resources, improves consumption and

alleviates pressure on public finances. The positive impact would likely be understated

as larger cities tend to provide better job prospects, which we have assumed to be the

same across all scenarios.

The dispersed and slow reform scenario assumes that land use will become even more

dispersed than in the past, that no efficiency gains are made through city

agglomeration and that social security reforms remain gradual. TFP gains and

consumption/GDP shares would therefore follow the baseline, but the non-housing

investment/GDP would rise even faster. This scenario is the least favored case, as it

implies a worsening of investment vs. consumption imbalances, even more significant

burdens on land and resources than in the past, overheating pressures in the near term

owing to the significant increase in investment demand and a deterioration in

economic returns on investment.

3

4

5

6

7

8

9

3

4

5

6

7

8

9

2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020

Housing investment/GDP

Housing investment/GDP under on-going policy controls

Housing investment/GDP without policy controls

30

32

34

36

38

40

42

44

46

48

30

32

34

36

38

40

42

44

46

48

2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020

Household consumption/GDP (%)

Reform

Baseline



5. Conclusion

The urbanization process has underpinned the transformation of the Chinese economy in

the past 30 years. The shape of future urbanization, along with the necessary adjustment of

factor costs (in particular land) will be important in affecting the efficiency of resource

allocation for the economy as a whole.

Growth and inflation

Urbanization will continue to be a source of productivity growth. Besides continued labor

migration into urban activities, the ‘agglomeration effect’ in cities is another important area

of such productivity gains. The latter is an area where easy gains can be made given the

absence of such an effect so far. Meanwhile, social security reforms can deliver meaningful

benefits which could help reduce the savings of migrant workers, removing one of the

deflationary forces for longer-term growth.

Commodities and sectional implications

A more concentrated urban growth model—say, developing city clusters—combined with

the conservative use of land would reduce the resource intensity of urban growth. This

would be negative for some commodities related to infrastructure investment, but positive

for goods and services associated with public transportation, utilities, water management

and diversified consumption. This is in contrast to a more dispersed urban growth model

(say, by developing townships or small cities) where resource intensity stays high or rises,

with near-term strong demand for commodities. Policy discussions so far, including the

emphasis on land reforms, suggest a more concentrated model is likely to be chosen.

Housing-related sectors. Housing purchases have been financed mostly by household

savings rather than leverage. Urban housing supply has been mostly driven by population

and income growth across regions. Such factors will continue to support housing demand,

notwithstanding the policy controls including the introduction of property taxes. We

therefore expect housing-related sectors to hold up in coming years.

Government finance

Although we do not expect the consolidated government balance sheet to come under

stress in the coming years, concerns related to the micro risks of local government

investment arms are likely to remain. As we have argued elsewhere, government

investment in local infrastructure is a major reason for the rising credit intensity of the

economy in recent years. This is because, compared with other types of investment,

infrastructure investment relies more on debt financing (owing to its longer construction

cycles and more delayed cash flows), and the economic benefits materialize with a time lag.

In our view, the returns on local investment would improve, and concerns for local finances

and debt burdens would ease considerably, under a more concentrated urbanization model.



References

Aziz, Jahangir and Li Cui (2007), “Explaining China’s low consumption: the neglected role of household income,” IMF

Working Paper.

Ball, Michael, Tanya Morrison and Andrew Wood (1996), “Structures Investment and Economic growth: A long-term

International Comparison,” Urban Studies.

Barras, Richard (2001), “Building investment is a diminishing source of economic growth,” Journal of Property Research.

Carlino, Gerald (2011), “Three Keys to the City: Resources, Agglomeration Economies, and Sorting,” Business Review,

Philadelphia Fed, Q3 2011.

De Long, J. Bradford and Lawrence H. Summers (1992), “Equipment investment and economic growth: How strong is

the nexus?” Brookings Papers on Economic Activity.

Deng, Frederic, Youqin Huang (2004), “Uneven land reform and urban sprawl in China: the case of Beijing,” Progress in

Planning.

Greenstone, Hornbeck, and Moretti (2010), “Identifying agglomeration spillovers: evidence from winners and losers of

large plant openings," Journal of Political Economy, 2010, 118 (3): 536-598.

Henderson , J. Vernon (2004), “Urbanization and growth,” Handbook of Economic growth.

Li, HuiZhong, Feng Yin, Jianlun Li (2008), “China’s construction land expansion and economic growth: a capital-output

ratio based analysis,” China & World Economy.

Lewis, W. Arthur (1954), "Economic Development with Unlimited Supplies of Labor," Manchester School of Economic

and Social Studies, Vol. 22, pp. 139-91.

Urbanization and growth (2009), ed. by Michael Spencer, Patricia Annez, and Robert Buckley, IBRD.



Disclosure Appendix

Reg AC

I, Li Cui, hereby certify that all of the views expressed in this report accurately reflect my personal views, which have not been influenced by

considerations of the firm's business or client relationships.

Goldman Sachs Disclosures

Global product; distributing entities

The Global Investment Research Division of Goldman Sachs produces and distributes research products for clients of Goldman Sachs on a global

basis. Analysts based in Goldman Sachs offices around the world produce equity research on industries and companies, and research on

macroeconomics, currencies, commodities and portfolio strategy. This research is disseminated in Australia by Goldman Sachs Australia Pty Ltd

(ABN 21 006 797 897); in Brazil by Goldman Sachs do Brasil Corretora de Títulos e Valores Mobiliários S.A.; in Canada by Goldman, Sachs & Co.

regarding Canadian equities and by Goldman, Sachs & Co. (all other research); in Hong Kong by Goldman Sachs (Asia) L.L.C.; in India by Goldman

Sachs (India) Securities Private Ltd.; in Japan by Goldman Sachs Japan Co., Ltd.; in the Republic of Korea by Goldman Sachs (Asia) L.L.C., Seoul

Branch; in New Zealand by Goldman Sachs New Zealand Limited; in Russia by OOO Goldman Sachs; in Singapore by Goldman Sachs (Singapore)

Pte. (Company Number: 198602165W); and in the United States of America by Goldman, Sachs & Co. Goldman Sachs International has approved this

research in connection with its distribution in the United Kingdom and European Union.

European Union: Goldman Sachs International, authorized and regulated by the Financial Services Authority, has approved this research in

connection with its distribution in the European Union and United Kingdom; Goldman Sachs AG and Goldman Sachs International

Zweigniederlassung Frankfurt, regulated by the Bundesanstalt für Finanzdienstleistungsaufsicht, may also distribute research in Germany.

General disclosures

This research is for our clients only. Other than disclosures relating to Goldman Sachs, this research is based on current public information that we

consider reliable, but we do not represent it is accurate or complete, and it should not be relied on as such. We seek to update our research as

appropriate, but various regulations may prevent us from doing so. Other than certain industry reports published on a periodic basis, the large

majority of reports are published at irregular intervals as appropriate in the analyst's judgment.

Goldman Sachs conducts a global full-service, integrated investment banking, investment management, and brokerage business. We have

investment banking and other business relationships with a substantial percentage of the companies covered by our Global Investment Research

Division. Goldman, Sachs & Co., the United States broker dealer, is a member of SIPC (http://www.sipc.org).

Our salespeople, traders, and other professionals may provide oral or written market commentary or trading strategies to our clients and our

proprietary trading desks that reflect opinions that are contrary to the opinions expressed in this research. Our asset management area, our

proprietary trading desks and investing businesses may make investment decisions that are inconsistent with the recommendations or views

expressed in this research.

The analysts named in this report may have from time to time discussed with our clients, including Goldman Sachs salespersons and traders, or may

discuss in this report, trading strategies that reference catalysts or events that may have a near-term impact on the market price of the equity

securities discussed in this report, which impact may be directionally counter to the analysts' published price target expectations for such stocks. Any

such trading strategies are distinct from and do not affect the analysts' fundamental equity rating for such stocks, which rating reflects a stock's

return potential relative to its coverage group as described herein.

We and our affiliates, officers, directors, and employees, excluding equity and credit analysts, will from time to time have long or short positions in,

act as principal in, and buy or sell, the securities or derivatives, if any, referred to in this research.

This research is not an offer to sell or the solicitation of an offer to buy any security in any jurisdiction where such an offer or solicitation would be

illegal. It does not constitute a personal recommendation or take into account the particular investment objectives, financial situations, or needs of

individual clients. Clients should consider whether any advice or recommendation in this research is suitable for their particular circumstances and, if

appropriate, seek professional advice, including tax advice. The price and value of investments referred to in this research and the income from them

may fluctuate. Past performance is not a guide to future performance, future returns are not guaranteed, and a loss of original capital may occur.

Fluctuations in exchange rates could have adverse effects on the value or price of, or income derived from, certain investments.

Certain transactions, including those involving futures, options, and other derivatives, give rise to substantial risk and are not suitable for all investors.

Investors should review current options disclosure documents which are available from Goldman Sachs sales representatives or at

http://www.theocc.com/about/publications/character-risks.jsp. Transaction costs may be significant in option strategies calling for multiple purchase

and sales of options such as spreads. Supporting documentation will be supplied upon request.

In producing research reports, members of the Global Investment Research Division of Goldman Sachs Australia may attend site visits and other

meetings hosted by the issuers the subject of its research reports. In some instances the costs of such site visits or meetings may be met in part or in

whole by the issuers concerned if Goldman Sachs Australia considers it is appropriate and reasonable in the specific circumstances relating to the

site visit or meeting.

All research reports are disseminated and available to all clients simultaneously through electronic publication to our internal client websites. Not all

research content is redistributed to our clients or available to third-party aggregators, nor is Goldman Sachs responsible for the redistribution of our

research by third party aggregators. For all research available on a particular stock, please contact your sales representative or go to

http://360.gs.com.

Disclosure information is also available at http://www.gs.com/research/hedge.html or from Research Compliance, 200 West Street, New York, NY

10282.

© 2013 Goldman Sachs.

No part of this material may be (i) copied, photocopied or duplicated in any form by any means or (ii) redistributed without the prior written consent of The Goldman Sachs Group, Inc.



Gao Hua Securities Disclosures

General disclosures

This research is disseminated in China by Gao Hua Securities. Gao Hua Securities is licensed to conduct Securities Investment Consultancy.

This research is for our clients only. This research is based on current public information that we consider reliable, but we do not represent it is

accurate or complete, and it should not be relied on as such. We seek to update our research as appropriate, but various regulations may prevent us

from doing so. Other than certain industry reports published on a periodic basis, the large majority of reports are published at irregular intervals as

appropriate in the analyst's judgment.

Goldman Sachs Gao Hua, an affiliate of Gao Hua Securities, conducts an investment banking business. Gao Hua Securities, Goldman Sachs Gao Hua

and their affiliates have investment banking and other business relationships with a substantial percentage of the companies referred to in this

document.

Our salespeople, traders, and other professionals may provide oral or written market commentary or trading strategies to our clients and our

proprietary trading desks that reflect opinions that are contrary to the opinions expressed in this research. Our proprietary trading desks and

investing businesses may make investment decisions that are inconsistent with the recommendations or views expressed in this research.

The analysts named in this report may have from time to time discussed with our clients, including Gao Hua Securities salespersons and traders, or

may discuss in this report, trading strategies that reference catalysts or events that may have a near-term impact on the market price of the equity

securities discussed in this report, which impact may be directionally counter to the analysts' published price target expectations for such stocks. Any

such trading strategies are distinct from and do not affect the analysts' fundamental equity rating for such stocks, which rating reflects a stock's

return potential relative to its coverage group as described herein.

Gao Hua Securities and its affiliates, officers, directors, and employees, excluding equity and credit analysts, will from time to time have long or short

positions in, act as principal in, and buy or sell, the securities or derivatives, if any, referred to in this research.

This research is not an offer to sell or the solicitation of an offer to buy any security in any jurisdiction where such an offer or solicitation would be

illegal. It does not constitute a personal recommendation or take into account the particular investment objectives, financial situations, or needs of

individual clients. Clients should consider whether any advice or recommendation in this research is suitable for their particular circumstances and, if

appropriate, seek professional advice, including tax advice. The price and value of investments referred to in this research and the income from them

may fluctuate. Past performance is not a guide to future performance, future returns are not guaranteed, and a loss of original capital may occur.

Fluctuations in exchange rates could have adverse effects on the value or price of, or income derived from, certain investments.

Certain transactions, including those involving futures, options, and other derivatives, give rise to substantial risk and are not suitable for all investors.

Investors should review current options disclosure documents which are available from Gao Hua sales representatives or at

http://www.theocc.com/about/publications/character-risks.jsp. Transaction costs may be significant in option strategies calling for multiple purchase

and sales of options such as spreads. Supporting documentation will be supplied upon request.

© 2013 Beijing Gao Hua Securities Company Limited

No part of this material may be (i) copied, photocopied or duplicated in any form by any means or (ii) redistributed without the prior written consent of Beijing Gao Hua Securities Company Limited.

Global Economics Paper No: 218 - ghsl.cn · Global Economics Paper No: 218 Research Report China: More efficient cities key to a brighter growth path A more efficient use of surplus

Documents