Business Cycles, Climate Anomalies, and Regional Height Cycles in Japan, 1892-1937

Jean-Pascal Bassino a and Ulrich Woitek b

This version: April 2012 Please do not circulate

a IAO, Ecole Normale Supérieure de Lyon, 15 Parvis René Descartes, BP 7000, 69342 Lyon Cedex, France (corresponding author). b Department of Economics, University of Zurich, Zürichbergstrasse 14 8032 Zurich, Switzerland. Business Cycles, Climate Anomalies, and Regional Height Cycles in Japan, 1892-1937

Abstract

Recent research has shown that there is a significant inter-relation between business cycles and height cycles. We analyze patterns in average stature of Japanese conscripts for the period 1892-1937 using regional level annual time series (47 prefectures). Cycles in the 3-5 years range dominate; this is similar to findings for the US and Europe. Human stature in prefectures characterized by product specialization seems to be more sensitive to the dominant business cycle in the 7-10 years range than in others. We find that also that regional summer temperature contributes significantly to the explanation of the height cycle.

Keywords: Business cycles, human stature, height cycles, climate, Japan, spectral analysis 1. Introduction

Most studies analyzing the relation between economic and anthropometric variables point toward a significant and positive relationship between income and stature. It has been well established that long-term trends in average stature in a given population are influenced by changes in economic conditions measured by income; the major explanation is the link between income and nutritional status. Quantitative analyses using international or inter- regional cross-section data highlight the role of other explanatory variables such as availability of health services, urbanisation, and workload of children and pregnant women (Steckel 1995, 2009). Recent research has also shown that there is a significant inter-relation between business cycles, measured by macroeconomic indicators of income, output, or investment, and cycles in human stature (Woitek 2003, Sunder and Woitek 2005).1 These results indicate that short- term variations in economic bell being can have a discernible impact on height. By focusing on deviations from the secular trend, we can study the relation between income and in stature without having to take into consideration explanatory variables such as urbanization, workload, and the provision of health services that tend to evolve at a relatively slow pace. In this paper, we add to the existing knowledge by analyzing height cycles observed on Japanese regional level annual series of average height of conscripts measured between 1892 and 1937 (47 prefectures). In order to relate short-term variations in income to height cycles, we have to identify the period(s) of the life during which the growth of stature is particularly sensitive to economic conditions. The height of an adult reflects the cumulated influences of nutritional status, health conditions, and other environmental variables at different ages, from gestation to early adulthood. For any individual, the potential growth in stature between conception and early adulthood is essentially determined by two factors: genes, and the height of the mother, which constrains the growth of the foetus in utero (Cole 2003). But depending on environmental conditions, a certain percentage of the population will become shorter than the potential level. A sharp degradation in living conditions during infancy and adolescence can result in a slowdown or complete halt of height growth. When the effect is of short duration and moderate intensity, an almost complete recovery is possible the following months or years.

1 These fluctuations in human stature are not harmonic waves with fixed frequency and amplitude, but akin to economic fluctuations that are commonly describe as business cycles. It is therefore acceptable to describe these phenomena as height cycles. However, depending on their severity and duration, adverse conditions could have an irreversible impact. The age at which these events occur is of critical importance: adult final stature is particularly influenced by conditions during gestation and early infancy (ibid.). A sudden decline in disposable income can be expected to induce a decline in the nutritional status of a large share of the population, particularly if the initial level is only slightly above subsistence level. Hence, height cycles observed on series of average stature reported for successive annual cohorts reflect to a large extent the impact of changes in income that occurred during the year of birth. Therefore, average height of individuals aged 20 measured in 1892-1937 can be related to living condition during the period 1872-1917. A major difficulty encountered in the analysis of height cycles is that changes in living conditions are influenced by nation-wide economic conditions, but also by asymmetric shocks taking place at the regional level. Yearly variations of income levels are the manifestation of national business cycles that are essentially related to phases of expansion and recession of industrial production and investment2. In the meantime, local climate anomalies can result in severe crop failures (or exceptional bumper crops), drastically depressing (or strongly increasing) agricultural income3. Hence, height cycles reflect the combined impacts of business cycles in the industrial and financial sectors, and asymmetric shocks on agricultural output. The motivation for using Japanese data in an investigation of the impact of business cycles and regional climate anomalies on height cycles is related to several attractive features of this country for a case study. First, income levels were low by international standards; nutritional status, almost no import of foodstuffs before 1918. Second, as the first non-western industrial nation4; business cycles can be easily identified. Third, we can rely on anthropometric data of an exceptional quality. Fourth, detailed climate data are available for a large sample of prefectures using regional level data is twofold. First, we can expect the results based on these data to be more robust than those obtained by relying on national-level averages, or average for different population sample. Second, and more importantly, using regional-level data allows us to investigate the relation between stochastic shocks originating from exogenous variables (i.e. variables that are not determined by the economic dynamics) and variation in stature. Climate is the most obvious purely exogenous variable and it is

2 The role of monetary variable and linkage with the economic conditions in the rest of the world are discussed in section 2. 3 Market integration; price levels. note: also, in case of imperfect market integration, locally available food supply, or purchasing power; Diseases 4 Equally important is the fact that Japan has been subject to unequal treaties, between 1858 and 1911, and therefore influenced by business cycles in the rest of the world. particularly suitable for our study as it impacts on both economic and biological welfare, particularly in lower income countries. Climate anomalies in rainfall and temperatures (as well as wind) are key determinants in short-term variations in agricultural output volume and therefore influence food supply and agricultural income, which in turn could be expect to explain a large share in short-term deviations in the secular trend in stature. Moreover, as some climate anomalies are mostly regional in their occurrence, resulting in asymmetric shocks that could be expected to exert a strongly negative impact on local economic conditions, we should be able to disentangle two determinants of short-term deviations in the secular trend of stature: on the one hand, the influence of macro-economic variables reflecting the dynamics of the national economy as a whole; on the other hand, the impact of variations in local agricultural income and food supply resulting from regional-level climate anomalies. 5

As the dynamics of both economic and anthropometric variables can be decomposed as a combination of a long-term trend (either linear or non-linear) and of short terms fluctuations (that can be cyclical, i.e. more or less stable in terms of period and magnitude) we can detrend economic and stature series in order to investigate the influence of changes in economic conditions on stature. In addition to the cross-section correlation between levels of income and stature, a positive and significant relation is observed in interwar Japan, using also cross- section regional level data, between short or medium term variation in income (5 or 10 years) and the lagged variation in stature, without any influence of health variables (Bassino 2006). The remaining of the paper is organized into 5 sections. Section 2 considers the impact of exogenous shocks and endogenous economic variables on stature; section 3 presents the data available and offers a summarized view of business cycles in pre-WWII Japan; section 3 describes the method used for analyzing cycles in stature; section 4 presents the results and analyzes the influence of economic fluctuations and climate anomalies; section 5 concludes.

2. Endogenous cycles and exogenous shocks

Although the purpose of this paper is not to revisit business cycles theory, it is worth considering the implications of the availability of two clearly distinct and largely independent

5 Health conditions prévalence of pathologies. Lack of information. Assume random (not sure i twas a white noise). types of indicators.6 We use, on the one hand, economic data capturing the macroeconomic dynamics of the Japanese economy as a whole (actually more the modern sector, due to the very nature of the data); on the other hand, measures of climate anomalies that can be regarded as proxies for the shocks affecting the agricultural sector. As we are concerned with two types of influences, we should acknowledge that this implies either a theoretical framework that considers the business cycles are largely exogenous (climate), albeit also perhaps partly endogenous (economic conditions), or alternatively that there are at least two independent components in the business cycles resulting from stochastic shocks of different natures, and possibly resulting in fluctuations of different frequencies. The alternative could be summarized as follows: is the business cycle a purely of partly stochastic process? By relying on regional level data and macroeconomic-data that are essentially independent, one additional aim of this paper is to address from an empirical side endogeneity, which is the key issue in the theoretical discussions related to business cycles. Since the identification of cycle-like fluctuations in the late 19th century, the debate has opposed the exponents of the theory of the cycle as an endogenous self-perpetuating dynamics to those considering cycle-like fluctuations as resulting from a succession of stochastic shocks gradually absorbed by the economic system. The notion of economic cycle being an endogenous phenomenon appeared after the observation of a certain periodicity in the recurrence of economic recession in the 19th century interpreted as a self perpetuating process in which “the expansion is the sole cause of the recession” (Juglar 1889). Although Juglar’s interpretation relied on the role of monetary variables in particular the role of banks in an accumulation of credit leading to overinvestment7, it does not differ essentially from the theory of the cycle focused on the interaction among real variables (Kalecki 1935, Hicks 1950, Goodwin 1967); the theory of innovation business cycles prposed by Schumpeter (1939) is also comparable despite his attempt to analyze the dynamics as resulting from the interaction of real and monetary variables. Most of these theories refer to a cycle using macroeconomic indicators that has a 9 to 10-year period on average. This fluctuation is labelled as Juglar cycle by Schumpeter (1939), who also refers to business cycles of shorter period, related to variations in inventory, that he designates as Kitchin cycle (Kitchin 1923). For the sake of convenience, we use the terms Kitchin and Juglar cycles in the analysis of the

6 The rest of the section briefly mentions several theoretical approaches that are the most relevant for our purpose; a discussion on the implications in terms of source and propagation mechanism of the business cycles in beyond the scope of this paper. Zarnowitz (1996) provides a detailed presentation of the different theories and techniques used for the identification and measurement of business cycles. 7 A similar approach is adopted by (Hayek 1933). Other monetary theories give more importance to the fluctuation of money supply. relations between economic variable and stature (without implication in terms of adhesion to these theories). The idea that business cycles result from stochastic shocks can be also traced back to the late 19th and early 20th century. For Jevons (1878) and Moore (1926), business cycles are due to the influence of variables that are entirely exogenous. Climate anomalies, due to sunspots and therefore resulting in variations in temperatures for Jevons, or taking the form of variations in rainfall for Moore, are the cause of oscillations affecting the economic system. During the 20th century, reflecting the declining importance of agriculture as a share of national income in the leading economies, it become apparent that climate alone was an implausible determinant of economic fluctuations; still, the concept of stochastic Monte Carlo cycle resulting from exogenous shocks is explicitly a transposition of the impact of climate anomalies, with stochastic shocks caused by a number of factors (Fisher 1925). In search for a more suitable candidate for the origin of exogenous shocks, the attention shifted to the role of sudden changes in the money supply (Fisher 1933, Friedman and Schwarz 1963). The reformulation of this approach by Lucas (1975) takes into account rational expectations and allows considering business cycles as resulting from monetary shocks; other authors also the possibility of real stochastic shocks, for instance Grossman and Weiss (1982). But, using the same kind of micro-macro interaction, others authors have proposed alternative interpretation of endogenous business cycles (Long and Plosser 1983, Grandmont 1985). The role of climate as determinant of economic fluctuations is understandably modest in high income countries (Beaulieu and Miron 1992, Ghosh and Wolf, 1997), is can be assumed that the impact of anomalies in temperatures and rainfall was larger in pre-WWII Japan as about half of the population lived in rural areas with an average per capita income only marginally above subsistence. It is well known that climate anomalies still exert a strong influence in the economic dynamics and biological welfare in present-day developing countries. From this viewpoint, the study of the case of pre-WWII Japan, based on an exceptionally detailed and dependable regional-level data set, provides a useful case study. Although going beyond the scope of the present paper, this could lead to an investigation of the response of economic systems to these stochastic shocks, and ultimately an evaluation of social costs that are suggested by episodes of short-term absolute decline in stature (and more commonly relative decline resulting in a temporary slowdown in the secular trend).

3. Data

In this study, we rely on three types of annual data: estimates of regional-level series of average stature measured at age 20, a selection of macroeconomic indicators constructed retrospectively for pre-WWII, and different measures of climate anomalies based on data collected in a number of cities across Japan from the late 19th century. Average stature annual series were calculated by relying of official reports of conscription of the Japanese army recorded between 1892 and 1937 (data are missing after 1937, except for 1941).8 During that period, the entire male cohort was measured at age 20.9 Nevertheless, averages can be calculated by relying on the number of conscripts for each interval of body height. For each interval and prefecture, the average was calculated for each prefecture using the median value of each interval. For the upper and lower open tails, it is assumed that the difference between the median value of the open tail and the median value of the closest interval is equal to the difference between the medians of adjacent intervals. Before 1898, regional conscription data are for the traditional administrative units (kuni). The conversion of data for kuni to prefectures (fudôken) is based on a comparison of administrative maps. In most cases, the prefecture is equivalent to one or two kuni.10 Although average stature is not measured on the basis of individual data, but estimated using the procedure described above, the data set can be regarded as reflecting accurately the reality. Because the number of conscripts in the open tails and in the intervals close to these open tails is small, this approximation is unlikely to impact on the values of average heights. Furthermore, the percentage of conscript for which the information is reported as missing is extremely low and, for each year and each prefecture, the population of the cohort is normally distributed among the height intervals. Prefecture-level annual series display, in addition to a general upward trend in stature, some fluctuations that are invisible or less visible on national series. This suggests that a number of asymmetric shocks resulted in cycle-like short-term deviation from the trend. The magnitude of the deviation from the trend is particularly large for several rural prefectures, but the tendency is toward a fading of the fluctuation after the turn of the 20th century. This

8 This is probably due to deliberate destruction of military archives by the Japanese authorities during the interval between the capitulation and the arrival of US authorities; for the same reason, most individual data have been lost and the remaining is not accessible for privacy motives. 9 The original sources are Rikugunshô Tokei Nenpô [Statistical Yearbook of the Imperial Army], Nippon Teikoku Tôkei Nenpô [Statistical Yearbook of the Empire of Japan] and Nippon Rikugun Chôhei Tekiyô [Report on the Conscription of the Japanese Army]. These official yearly reports do not provide data for the average body height of the conscripts originating from each region (except for the last years). About 98% of the conscripts were in the Army and about 2% in the Navy. As the Navy conscripts originated from all Japanese prefectures, incompleteness of the coverage is unlikely to result in a selection bias. 10 See Bassino (2006a), appendix 1 for details. appears consistent with the gradual process of market integration observed during the late 19th century that resulted in a decline in short term volatility and a convergence of unit-prices of food items towards Tokyo levels. This implies that the impact of asymmetric shocks related to local climate anomalies (or eventually other change in local conditions) became smaller. Exogenous shocks were still probably affecting agricultural output volume during the interwar, but the lower short-term volatility in the prices of food items suggests that shipment of grain from other regions or from overseas allowed maintaining a relatively stable food supply during the years of crop failure. As business cycle measures, we use real GNP, real per-capita GNP, real gross fixed investment (GFCF), and an index for industrial production (1934-1936=100). All series are on the national level.11 These series have been reconstructed after WWII on the basis of available information; these pre-WWII estimates of Japanese national accounts produced as part of the Long Term Economic Statistics project (LTES)12 are generally regarded as more reliable, by far, than the series reconstructed for other Asian countries. Estimates of agricultural output volume for the late 19th century are based on series recorded by the Ministry of Agricultural and Commerce that tend to underestimate output volume of staple food items other than rice (Bassino 2006b).13 Nevertheless, as we use the macroeconomic indicators in order to identify business cycles, we should not be too much concerned with the underestimation of output; the margin of error affects mostly the level, and therefore the average growth rate over a long period, rather than fluctuations in output. An additional and potentially more serious problem is that output value and value added have been estimated by relying on Tokyo prices, although regional differences in level and short term volatility persisted until the 1890s.14 Still, this is acceptable if we consider that macroeconomic indicators are reflecting movements for the national economy as a whole (or the manufacturing sector in the case of

11 GNP and GFCF are at 1934-1936 prices. Data source: Japan Statistical Association (1999), Historical Statistics of Japan on CD-ROM (1868-1085). These are the series the most commonly used for the identification of business cycles in pre-WWII Japan. Other monetary and real series are also used by Fujino (1965, 1966), Ohkawa and Rosovsky (1962), Shinohara (1963, 1978), and Kuznets (1968): bank deposits, loans, industrial prices, per capita manufacturing output, per capita consumption, and share of consumption in GDP. The dates of peaks and troughs are only slightly affected by the choice of indicator. 12 Ohkawa, Shinohara and Umemura eds (1965-1983). 13 For the same reasons, national-level series of agricultural output value in constant price are unlikely to provide a better indicator that per capita income or GDCF. In addition, the series of GDCF tend to underestimate investment in the agricultural sector and the coverage of the component for residential construction seems limited to Tokyo. GDCF series therefore do reflect aggregate capital formation but rather the combination of public investment (including military equipment) and private investment in mining, manufactures, non- residential construction and transportation. 14 Unit-prices of food items in Tokyo were almost identical to those observed in Osaka, and very similar to the prices recorded in major industrial cities (Nagoya, Kyoto, Kobe, Hiroshima, or Fukuoka). GDCF and the index of industrial production) driven by monetary or real shocks originating from urban areas and related to the most modern parts or sectors of the Japanese economy.15 Other national-level series are available but do not seem suitable for our purpose. Considering the large body of theoretical literature dealing with monetary cycles (either endogenous or stochastic) the most obvious candidate would be the money supply (or the consumer price index). However, as about half of the Japanese households were still living in rural areas and relied to some extent to self-consumption, it is likely that monetary shocks had only a low impact on their standard of living. In addition, high levels of short-term volatility and large regional differentials in unit-price of food items imply that shocks originating from changes in monetary policy (or in sudden variation in world prices of species and commodities) had not the same impact all over Japan. The same remark applies to variations in the price of food relative to non-food items. Ideally, we would like to use regional-level information on variations in the relative price of food items in order to investigate the role of these asymmetric shocks but the information is available only for a few prefectures and it does not cover the entire period. Considering that regional-level monetary shocks were certainly the result of drought and/or low temperatures, and that agricultural output volume data for food-items other than rice are not very reliable before the 1920s, the best method for analysing the relation between height cycles and the asymmetric shocks affecting agricultural output is to use climate data. Two types of climate anomalies could be expected to have a major impact on agricultural output in all or part of the Japanese archipelago: - First, low summer temperatures, especially in July, as it can impact very of rice output (reducing yields by half or even worse); paddy is blossoming during that period and output is highly sensitive to low temperature in July. The Tohoku region, the northeastern part of Honshu island (the major island of the archipelago), is the region the most affected by natural hazards of this type because average summer temperatures tend to be lower than in Central or Western Japan. Anecdotal evidence suggest that the harshest effects have been concentrated in the region of the Pacific side of the Tohoku that is particularly exposed to Yamase, a wind bringing cold air from the Northern Pacific Ocean. In comparison, to the devastating impact of low temperatures, high summer temperatures are unlikely to influence agricultural output if rainfall is sufficient.

15 Or, if we adopt a different theoretical viewpoint, by endogenous business cycles occurring in urban areas. - Second, low rainfall; droughts occurred in both Eastern and Western Japan, and these natural hazards caused the most severe famines experienced in Japanese history (Saito 2002). This is particularly well documented for the 18th and early 19th century. In most part of Japan, annual rainfall is in a range between 1500 and 2500 mm per year and tends to be concentrated in late spring, summer and early autumn. Agricultural techniques, particularly rice cultivation, are dependent on high annual rainfall. In general, above average rainfall do not cause much problem except when floods occur, in general as a result of a cyclonic event. It is therefore more a problem related to distribution than excess in rainfall. The data set available at the regional does not allow measuring accurately the magnitude of cyclonic events. It should be noted that the region that is the most exposed to cyclonic events, the southern part of Kyushu island, sweet potatoes remained the major staple food until the interwar and this crop if much less sensitive than paddy to heavy rainfall and floods. In order to capture climate anomalies, we measure, at the regional level, two annual indices: for low temperatures, and for drought. Both are calculated as deviation from the average values recorded by the Japanese Meteorological Agency and his forerunners.16 The low temperatures index is based on the ratio of the average temperature of each month of the year to the average value recorded for the entire period (up to 2002). Likewise, the drought index is based on the ratio of annual rainfall of the year to the average for the entire period (up to 2002). The value of the index is the ratio when lower than 1 (i.e. below average) and 0 otherwise. Ideally, we would like to use climate data for all the 47 prefectures but this information is not readily available. The climate data we are currently using covers 16 cities that are prefecture capitals. Considering the location of these cities, it is likely that they have higher average winter temperatures that most other parts of the prefecture; but, in summer, the effect of low elevation on temperature is partially offset by coastal location (all the cities, except Nagano, are either seaports or close to the coast). Thus, these data can be used a proxy for the entire prefecture (see section 4 for a list of cities and prefectures).

4. Method

The method of choice is spectral analysis. This method has the advantage that it allows to derive a measure for synchronization frequency by frequency, which enables us to focus on the relevant business cycle frequency intervals. We focus on three cycle ranges: the 7-10

16 Annual data are available for a number of Japanese cities from the following website: http://www.stat.go.jp/english/data/chouki/index.htm years range (Juglar cycle), the 5-7 years range, and the 3-5 years range (Kitchin cycle). The dominant cycle band is identified by calculating the share of total variance attributable to cycles in these intervals. To address the issue of co-movement, we decompose the variance for each frequency band in the height spectrum into a part explained by the business cycle measure and an unexplained part. In addition, we adopt the dynamic correlation measure suggested by Croux, Forni and Reichlin (2001) to distinguish between in-phase and out-of- phase movements. We proceed as follows: first, we filter the data (height, economic data, climate) to achieve stationarity. In the next step, we fit a vector-autoregressive model (VAR) to the filtered data. Then, we compute the spectral density matrix of this VAR, which provides the business cycle measures discussed below.

Consider two stationary time series, X t representing human stature, and Yt representing a business cycle measure. The spectral density matrix of these series is defined as

17 the Fourier transform of the covariance matrix "xy (!) , ! = 0,±1,±2,…:

1 " F ( ) ( )e#i&'% 0.5,0.5 xy ' = !$xy % , # "[! ]. (1) 2& % =#"

The diagonal elements of the spectral density matrix Fxy (!) are called autospectra. Integrating the autospectra over the frequency band [! 0.5,0.5] gives the variance of the respective series. After dividing the autospectrum by the variance, the contribution of cyclical components in a frequency band [!1,!2 ] can be calculated. The off-diagonal elements or cross-spectra are complex numbers and given by

f xy (!) = cxy (!) " iqxy (!) , # "[! 0.5,0.5], (2)

where cxy (!) is the cospectrum and qxy (!) is the quadrature spectrum. The cospectrum measures the covariance between the ‘in-phase’ components of X t and Yt , whereas the quadrature spectrum measures the covariance between the ‘out-of-phase’ components. Together with the autospectra, the cross spectrum can be used to calculate squared coherency sc(!) , a measure similar to R2 in linear regression analysis:

17 See e.g. Harvey (1993, 175-179), Brockwell and Davis (1991, 434-443), Priestley (1981) and Koopmans (1974, 119-164). 2 f xy (!) sc(!) = , 0 ! sc(") ! 1. (3) f x (!) f y (!)

This measure assesses the degree of linear relationship between two series, frequency by frequency. If we are interested in the extent to which the variance of cyclical components of the height series X t in the frequency band [!1,!2 ] can be attributed to corresponding cyclical components in the business cycle measure Yt , sc(!) can be used to decompose the fraction of overall variance in this interval into an explained and an unexplained part:

"s "s "s f ( )d sc( ) f ( )d f ( )d . (4) ! x " " = ! " x " " + ! u " " "1 $"1 !!#!!" $"1 !#!" "explained" variance "unexplained" variance

As pointed out by Croux, Forni and Reichlin (2001), a measure like the squared coherency presented above is not suited for analyzing the co-movement of time series, because it does not contain information about possible phase shift between cycles in the series

X t and Yt . They propose an alternative measure, the so-called dynamic correlation "(!), which measures the correlation between the ‘in-phase’ components of the two series at a frequency ! :

c (!) "(!) = xy , "1 ! $(#) ! 1. (5) f x (!) f y (!)

Re-writing sc(!) as

2 2 2 f xy (!) c (!) + q (!) sc(!) = = xy xy , (6) f x (!) f y (!) f x (!) f y (!) we can use this idea to further decompose explained variance:

"s "s "s f ( )d sc( ) f ( )d f ( )d ! x " " = ! " x " " + ! u " " "1 $"1 !!#!!" $"1 !#!" "explained" variance "unexplained" variance "s c (")2 + q (")2 "s = xy xy d" + f (")d" (7) ! f (") ! u "1 y "1 "s c (")2 "s q (")2 "s xy d xy d f ( )d . = ! " + ! " + ! u " " f y (") f y (") $"1 !#!" $"1 !!#!!" $"1 !#!" "explained" variance (in-phase) "explained variance" (out-of -phase) "unexplained" variance

Thus, it is possible to distinguish between the ‘in-phase’ component and the ‘out-of- phase’ component of explained variance, adding some information on the importance of the phase shift in a frequency interval to the R2 interpretation of the decomposition in equation (4) above. To estimate the spectra, we fit VAR models to the filtered18 height series and business cycle measures, and calculate the spectra of the estimated models.19 With a VAR model of order p, the spectral density matrix is given by

1 F(#) = A(#)!1"A(#)!* , # "[! 0.5,0.5]. (8) 2$

The error variance-covariance matrix is denoted by ! , and A(!) is the Fourier

2 P 20 transform of the matrix lag polynomial A(L) = I ! A1L ! A2 L !…! AP L .

To assess the robustness of our results, we follow Canova (1998) and compare the outcome for different filters. The filters used in this paper are the Baxter-King filter in the original (Baxter and King 1999) and a modified version (A’Hearn and Woitek 2001), the Hodrick-Prescott filter (Hodrick and Prescott 1997)21, and the difference filter.

5. Results

19 This method is based on the seminal work by Burg (1967), who shows that the resulting spectrum is formally identical to a spectrum derived from Maximum Entropy Principle. For applications to economic time series, see Hillinger and Sebold-Bender (1992) and A’Hearn and Woitek (2001), 20 L is the backshift operator, and the superscript ‘*’ denotes the complex conjugate transpose. 21 The smoothing weight for the Hodrick-Prescott filter is set to 100, as it is common in the literature. In addition, we also test the suggestion of Ravn and Uhlig (2002) and set the weight to 6.25.

The results discussed in the following are based on height series from 47 Japanese prefectures in the period 1892-1937. The regional distribution of the share of total variance in the business cycle intervals 3-5 years, 5-7 years, and 7-10 years is displayed in the boxplot in Figure 1 (modified Baxter-King filter).22 As in the other studies on business cycles and human stature, it turns out that the short 3-5 year cycle dominates the height cycle.

Figure 1: Regional Distribution of the Share of Total Variance

0.7

0.6

0.5

0.4

0.3 Share of Total Variance

0.2

0.1

0.0 7-10 years 5-7 years 3-5 years Business Cycle Range

Notes: The boxplots are based on the results for 47 prefectures in the period 1892-1937 (Hokkaido, Aomori, Iwate, Miyagi, Akita, Yamagata, Fukushima, Ibaraki, Tochigi, Gumma, Saitama, Chiba, Tokyo, Kanagawa, Niigata, Toyama, Ishikawa, Fukui,Yamanashi, Nagano, Gifu, Shizuoka, Aichi, Mie, Shiga, Kyoto, Osaka, Hyogo, Nara, Wakayama, Tottori, Shimane, Okayama, Hiroshima, Yamaguchi, Tokushima, Kagawa, Ehime, Kochi, Okinawa, Fukuoka, Saga, Nagasaki, Kumamoto, Oita, Myazaki, Kagoshima). Filter: modified Baxter-King filter (Baxter and King 1999, A’Hearn and Woitek 2001) The horizontal lines mark the minimum, 25th percentile, median, 75th percentile, and maximum (bottom to top).

22 The results are robust with respect to the choice of the filtering method. In the following, we focus on the modified Baxter-King filter. The results for the other filters are available on request. To get a first impression on the relationship between the national business cycle and the regional cycle in human stature, we look at explained variance. The resulting distributions are displayed in Figure 2, and the medians can be found in Table 1. In all three business cycle frequency ranges, median explained variance is in the range 0.1-0.2. For per-capita GNP and industrial production, the shorter cycles dominate, while it is the longer 7-10 years cycle in gross fixed investment. For the individual results, we see that the prefectures with an explained variance higher than 25 per cent in the 7-10 years range are Hokkaido, Iwate, Chiba, Kanagawa, Yamanashi, Shizuoka, Shiga, Kyoto, Wakayama, Okinawa, Kumamoto, and Kagoshima. A possible interpretation of the relatively high influence of the business cycle in these prefectures is that, as a result of to market integration and specialization - either goods23 or migration (or tourism in Kyoto)24 - local incomes fluctuate more than in other regions. As in Sunder and Woitek (2005), we find that the in-phase share of explained variance is higher in the short 3-5 years range, while the long height cycles are predominantly out-of- phase with the business cycle. This result is robust with respect to the business cycle measure. There is a striking difference between the three measures when it comes to co-movement: for all three measures, explained variance is about the same. The height cycles are much more in- phase with the GNP-cycle than with industrial production or gross fixed investment. In fact, in the shorter cycle range (5-7 and 3-5 years), height cycles and GNP cycles reach the upper and lower turning points almost exactly at the same time (Table 2, Figure 3).

Table 1: Regional Heights and the National Business Cycle – Median Explained Variance 7-10 years 5-7 years 3-5 years Per Capita GNP 0.17 0.19 0.14 Industrial Production 0.19 0.18 0.15 Gross Fixed Investment 0.17 0.15 0.13 Notes: The median explained variance is based on the results for 47 prefectures in the period 1892-1937 (Hokkaido, Aomori, Iwate, Miyagi, Akita, Yamagata, Fukushima, Ibaraki, Tochigi, Gumma, Saitama, Chiba, Tokyo, Kanagawa, Niigata, Toyama, Ishikawa, Fukui,Yamanashi, Nagano, Gifu, Shizuoka, Aichi, Mie, Shiga, Kyoto, Osaka, Hyogo, Nara, Wakayama, Tottori, Shimane, Okayama, Hiroshima, Yamaguchi, Tokushima, Kagawa, Ehime, Kochi, Okinawa, Fukuoka, Saga, Nagasaki, Kumamoto, Oita, Myazaki, Kagoshima). Filter: modified Baxter-King filter (Baxter and King 1999, A’Hearn and Woitek 2001)

23 Examples are silk (Yamanashi), cash crops (Chiba), fish (Hokkaido, Chiba, Wakayama, Kagoshima), handicraft, tea (Shizuoka), or rice (Shiga, Kyoto). 24 Chiba, Kanagawa, Kyoto, Okinawa, Kumamoto, Kagoshima.

In the next step of the analysis, we add climate to the model and estimate a three- variable VAR (height, regional temperature, national business cycle measure). Since the temperature data are available on a monthly base, it is possible to identify the importance of climate in a certain month relative to the economic conditions in the birth year. There are 16 prefectures where we have temperature data in the observation period: Hokkaido, Akita, Niigata, Fukushima, Nagano, Tokyo, Ishikawa, Aichi, Osaka, Hiroshima, Ehime, Kochi, Fukuoka, Nagasaki, Kagoshima, and Okinawa.25 All in all, the number of models underlying the results displayed in Figures 4-6 is 576 (16 prefectures, 12 months, 3 economic indicators).In general, the summer months dominate the influence of temperature on the height cycle, but December temperature is also important, especially for the short 3-5 years cycle. Economic conditions are more important in Winter, Spring, and Autumn. Low summer temperatures (especially in July) occurred frequently before WWI, and induced hardship in the countryside, particularly in the East of Japan. This anecdotic evidence is reflected in our results.

Table 2: Regional Heights and the National Business Cycle – Median Explained Variance (In-Phase) 7-10 years 5-7 years 3-5 years Median Explained Variance (In-Phase) Per Capita GNP 0.05 0.19 0.14 Industrial Production 0.05 0.08 0.06 Gross Fixed Investment 0.06 0.05 0.07 Share of Explained Variance Per Capita GNP 0.30 0.97 0.99 Industrial Production 0.25 0.42 0.41 Gross Fixed Investment 0.35 0.30 0.50 The median explained variance (in-phase) is based on the results for 47 prefectures in the period 1892-1937 (Hokkaido, Aomori, Iwate, Miyagi, Akita, Yamagata, Fukushima, Ibaraki, Tochigi, Gumma, Saitama, Chiba, Tokyo, Kanagawa, Niigata, Toyama, Ishikawa, Fukui,Yamanashi, Nagano, Gifu, Shizuoka, Aichi, Mie, Shiga, Kyoto, Osaka, Hyogo, Nara, Wakayama, Tottori, Shimane, Okayama, Hiroshima, Yamaguchi, Tokushima, Kagawa, Ehime, Kochi, Okinawa, Fukuoka, Saga, Nagasaki, Kumamoto, Oita, Myazaki, Kagoshima). Filter: modified Baxter-King filter (Baxter and King 1999, A’Hearn and Woitek 2001)

25 The weather stations for which data are available are assigned to the following prefectures: Sapporo (Hokkaido), Akita (Akita), Niigata (Niigata), Fukushima (Fukushima), Nagano (Nagano), Tokyo (Tokyo), Kanazawa (Ishikawa), Nagoya (Aichi), Osaka (Osaka), Hiroshima (Hiroshima), Matsuyama (Ehime), Kochi (Kochi), Fukuoka (Fukuoka), Nagasaki (Nagasaki), Kagoshima (Kagoshima), Naha (Okinawa). Figure 2: Regional Heights and Business Cycles – Explained Variance

Gross Fixed Investment (Real)

0.9

0.8

0.7

0.6

0.5

0.4 Explained Variance Explained

0.3

0.2

0.1

0.0 7-10 years 5-7 years 3-5 years Business Cycle Range

Per-Capital Real GNP

0.900

0.800

0.700

0.600

0.500

0.400 Explained Variance Explained

0.300

0.200

0.100

0.000 7-10 years 5-7 years 3-5 years Business Cycle Range

Industrial Production

0.9

0.8

0.7

0.6

0.5

0.4 Explained Variance Explained

0.3

0.2

0.1

0.0 7-10 years 5-7 years 3-5 years Business Cycle Range Notes: The boxplots are based on the results for 47 prefectures in the period 1892-1937 (Hokkaido, Aomori, Iwate, Miyagi, Akita, Yamagata, Fukushima, Ibaraki, Tochigi, Gumma, Saitama, Chiba, Tokyo, Kanagawa, Niigata, Toyama, Ishikawa, Fukui,Yamanashi, Nagano, Gifu, Shizuoka, Aichi, Mie, Shiga, Kyoto, Osaka, Hyogo, Nara, Wakayama, Tottori, Shimane, Okayama, Hiroshima, Yamaguchi, Tokushima, Kagawa, Ehime, Kochi, Okinawa, Fukuoka, Saga, Nagasaki, Kumamoto, Oita, Myazaki, Kagoshima). Filter: modified Baxter-King filter (Baxter and King 1999, A’Hearn and Woitek 2001) The horizontal lines mark the minimum, 25th percentile, median, 75th percentile, and maximum (bottom to top). Figure 3: Regional Heights and Business Cycles – In-Phase Explained Variance

Gross Fixed Investment (Real)

0.9

0.8

0.7

0.6

0.5

0.4

Explained In-Phase Variance In-Phase Explained 0.3

0.2

0.1

0.0 7-10 years 5-7 years 3-5 years Business Cycle Range

Per-Capita Real GNP

0.9

0.8

0.7

0.6

0.5

0.4

Explained In-Phase Variance In-Phase Explained 0.3

0.2

0.1

0.0 7-10 years 5-7 years 3-5 years Business Cycle Range

Industrial Production

0.9

0.8

0.7

0.6

0.5

0.4

Explained In-Phase Variance In-Phase Explained 0.3

0.2

0.1

0.0 7-10 years 5-7 years 3-5 years Business Cycle Range Notes: The boxplots are based on the results for 47 prefectures in the period 1892-1937 (Hokkaido, Aomori, Iwate, Miyagi, Akita, Yamagata, Fukushima, Ibaraki, Tochigi, Gunma, Saitama, Chiba, Tokyo, Kanagawa, Niigata, Toyama, Ishikawa, Fukui, Yamanashi, Nagano, Gifu, Shizuoka, Aichi, Mie, Shiga, Kyoto, Osaka, Hyogo, Nara, Wakayama, Tottori, Shimane, Okayama, Hiroshima, Yamaguchi, Tokushima, Kagawa, Ehime, Kochi, Okinawa, Fukuoka, Saga, Nagasaki, Kumamoto, Oita, Myazaki, Kagoshima). Filter: modified Baxter-King filter (Baxter and King 1999, A’Hearn and Woitek 2001). The horizontal lines mark the minimum, 25th percentile, median, 75th percentile, and maximum (bottom to top). Figure 4: Height, Regional Climate, and the Business Cycle (7-10 Years)

Average Explained Variance, 7-10 Years

0.01

0.009

0.008

0.007

0.006

0.005

0.004 Explained Variance

0.003

0.002

0.001

0 Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Month GFCF Temperature

Average Explained Variance, 7-10 Years

0.012

0.01

0.008

0.006 Explained Variance 0.004

0.002

0 Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Month GNP/POP Temperature

Average Explained Variance, 7-10 Years

0.01

0.009

0.008

0.007

0.006

0.005

0.004 Explained Variance

0.003

0.002

0.001

0 Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Month IP Temperature Notes: The graphs are based on the results for 16 prefectures in the period 1892-1937 (Hokkaido, Akita, Niigata, Fukushima, Nagano, Tokyo, Ishikawa, Aichi, Osaka, Hiroshima, Ehime, Kochi, Fukuoka, Nagasaki, Kagoshima, and Okinawa). Filter: modified Baxter-King filter (Baxter and King 1999, A’Hearn and Woitek 2001)

Figure 5: Height, Regional Climate, and the Business Cycle (5-7 Years)

Average Explained Variance, 5-7 Years

0.03

0.025

0.02

0.015 Explained Variance 0.01

0.005

0 Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Month GFCF Temperature

Average Explained Variance, 5-7 Years

0.035

0.03

0.025

0.02

0.015 Explained Variance

0.01

0.005

0 Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Month GNP/POP Temperature

Average Explained Variance, 5-7 Years

0.035

0.03

0.025

0.02

0.015 Explained Variance

0.01

0.005

0 Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Month IP Temperature Notes: The graphs are based on the results for 16 prefectures in the period 1892-1937 (Hokkaido, Akita, Niigata, Fukushima, Nagano, Tokyo, Ishikawa, Aichi, Osaka, Hiroshima, Ehime, Kochi, Fukuoka, Nagasaki, Kagoshima, and Okinawa). Filter: modified Baxter-King filter (Baxter and King 1999, A’Hearn and Woitek 2001) Figure 5: Height, Regional Climate, and the Business Cycle (3-5 Years)

Average Explained Variance, 3-5 Years

0.14

0.12

0.1

0.08

0.06 Explained Variance

0.04

0.02

0 Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Month GFCF Temperature

Average Explained Variance, 3-5 Years

0.14

0.12

0.1

0.08

0.06 Explained Variance

0.04

0.02

0 Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Month GNP/POP Temperature

Average Explained Variance, 3-5 Years

0.14

0.12

0.1

0.08

0.06 Explained Variance

0.04

0.02

0 Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Month IP Temperature

Notes: The graphs are based on the results for 16 prefectures in the period 1892-1937 (Hokkaido, Akita, Niigata, Fukushima, Nagano, Tokyo, Ishikawa, Aichi, Osaka, Hiroshima, Ehime, Kochi, Fukuoka, Nagasaki, Kagoshima, and Okinawa). Filter: modified Baxter-King filter (Baxter and King 1999, A’Hearn and Woitek 2001)

6. Conclusion

The cyclical structure in regional Japanese average height is dominated by cycles in the 3-5 years range. This is very similar to findings for the US and Europe. There is also a strong correlation with the national business cycle as measured by real per-capita GNP, real fixed investment, and industrial production. Human stature in prefectures characterized by specialization in goods such as silk, cash crops, fish, handicraft, or rice seems to be more sensitive to the dominant business cycle in the 7-10 years range than in others. Apart from specialization, integration of goods and factor markets is also important for the transmission of national business cycles on the regional level and, as a result, for the importance of the national economic cycle for fluctuations in human stature. Climate has also an impact: we find that especially regional summer temperature contributes significantly to the explanation of the height cycle.

References

A’Hearn, B., Woitek U., 2001. More international evidence on the historical properties of business cycles, Journal of Monetary Economics, 47, 299-319. Backus, D K., Kehoe P. J., 1992. International evidence on the historical properties of business cycles, American Economic Review, 82, 864-888. Bassino, J-P., 2006a. Regional inequality in Japan, 1892-1941: income, health, and stature, Economics and Human Biology, 4, 62-88. Bassino, J-P., 2006b. Agricultural output and food supply in Meiji Japan: economic miracle or creative destruction? Economic Development and Cultural Change, 54-2, pp. 503- 520. Baxter, M., R. G. King (1999), Measuring business cycles; approximate band-pass filters for economic time series, Review of Economics and Statistics, 81, 575-593. Beaulieu J.J., J.A. Miron (1992), A cross country comparison of seasonal cycles and business cycles. The Economic Journal, 102-413, 772-788. Brabec, M., 2005. Analysis of periodic fluctuations of the height of Swedish soldiers in 18th and 19th centuries, Economics and Human Biology, 3, 1-16. Brockwell, P.J., Davis R.A., 1991. Time Series: Theory and Methods. Springer, Berlin, 2nd ed. Canova, F., 1998. Detrending and business cycle stylized facts, Journal of Monetary Economics, 41, 475-512. Cogley, T., Nason J.M., 1995. Effects of the Hodrick-Prescott filter on trend and difference stationary time series: implications for business cycle research, Journal of Economic Dynamics and Control, 19, 253-278. Croux, T., Forni, M., Reichlin L., 2001. A measure of comovement for economic variables: theory and empirics, Review of Economics and Statistics, 83, 232-241. Da-Rocha, J.M., Restuccia D., 2006. The role of agriculture in aggregate business cycles, Review of Economic Dynamics 9, 3, 455-482 Fisher, I., 1925. Our instable dollar and the so-called business cycles, Journal of American Statistical Association, 20-149, 179-202. Fisher, I. 1933. The debt-deflation theory of the great depression, Econometrica, 1, 337-357. Friedman, M., Schwartz A., 1963. Money and business cycles. Review of Economics and Statistics, 45-1, 32-78. Fujino S., 1965. Nihon no keiki junkan: junkanteki hatten katei no rironteki, tokeiteki rekishiteki bunseki [Business Cycles in Japan: a Theoretical, Statistical and Historical Analysis of the Development of the Cyclical Process]. Keiso Shobo, Tokyo. Fujino S., 1966. Business cycles in Japan 1868-1962. Hitotsubashi Journal of Economics, June, 7, 56-79. Goodwin, R., 1967. The growth cycle, in Feinstein C.H. ed. Socialism, Democracy, and Economic Growth. Cambridge University Press, Cambridge. Grandmont, J-M., 1985. On endogenous competitive business cycles. Econometrica, 53-5, 995-1046. Grossman S. J., L. Weiss L., 1982. Heterogenous information and the theory of the business cycle, Journal of Political Economy, 90-4, 699-727. Harvey, A.C., 1993. Time series models. Harvester Wheatsheaf, 2nd ed. Harvey, A.C., Jaeger A., 1993. Detrending, stylized facts and the business cycle, Journal of Econometrics, 8, 231-247. Hayek, F.A., 1933. Monetary Theory and the Trade Cycle. Jonathan Cape, London. Hicks, J.R., 1950. A Contribution to the Theory of the Trade Cycle. Clarendon, Oxford. Hillinger, C., Sebold-Bender, M. 1992. The stylized facts of macroeconomic fluctuations, in: Hillinger, C. (ed.), Cyclical growth in market and planned economies. Oxford University Press, Oxford, pp. 63-110. Hodrick, R., Prescott E., 1997. Postwar U.S. business cycles: an empirical investigation, Journal of Money, Credit and Banking, 29, 1-16. Jacobs, J., Tassenaar, V. (2004), Height, income, and nutrition in the Netherlands: the second half of the 19th century, Economics and Human Biology, 2004, 2, 181-195. Jevons, W. S., 1878. Commercial crisis and sun-spots, Nature, 19, 33-37. Juglar C., 1889, Des Crises Commerciales et de leur Retour Périodique. Guillaumin, Paris. Kalecki, M., 1935. A Macroeconomic Theory of Business Cycles, Econometrica, 3, 327- 344. King, R.G., S.T. Rebelo, 1993. Low Frequency Filtering and Real Business Cycles, Journal of Economic Dynamics and Control, 17, 207-231. Kitchin J., 1923. Cycles and Trends in Economic Factors, Review of Economics and Statistics, 5-1, 10-16. Koopmans, L.H. (1974), The Spectral Analysis of Time Series. Academic Press, London. Kuznets, S. 1930. Equilibrium economics and business-cycle theory Quarterly Journal of Economics, 44, 3, 381-415. Kuznets S., 1968. Note on Japan’s economic growth, in Klein L., Ohkawa K., eds, Economic Growth: The Japanese Experience since Meiji. Irwin, Homewood, pp. 385- 422. Long, J.B. Jr, Plosser C.I., 1983. Real business cycles. Journal of Political Economy, 91-1, 39-69. Lucas, R.E., 1975. An Equilibrium Model of Business Cycles. Journal of Political Economy, 83-6, 113-144. Moore, H. L., 1926. A Theory of economic oscillations. Quarterly Journal of Economics, 41-1, 1-29. Ohkawa K., Rosovsky H., 1962. Economic fluctuations in prewar Japan: a preliminary analysis of cycles and long swings, Hitotsubashi Journal of Economics, 3-1, 10-33. Ohkawa, K., M. Shinohara, Umemura M., eds, 1965-1983. Nihon Choki Keizai Tôkei [Estimates of Long-Term Economic Statistics of Japan], 13 volumes. Toyo Keizei Shinposha, Tokyo. Priestley, M., 1981. Spectral Analysis and Time Series. Academic Press, London. Ravn, M. O. Uhlig H., 2002. On adjusting the Hodrick-Prescott filter for the frequency of observations, Review of Economics and Statistics, 84, 371-376. Schumpeter J.A., 1939. Business Cycles: A Theoretical, Historical and Statistical Analysis of the Capitalist Process. Mac Graw Hill, New York. Shinohara M., 1963. Growth and Cycles in the Japanese Economy. Kinokuniya, Tokyo. Shinohara M., 1978. Kuznets cycles and Juglar cycles, in Ohkawa K. and Hayami Y., Long term Analysis of Japanese Economy, Tokyo, Nihon Keizai Shinbunsha. Sunder, M., Woitek, U., 2005. Boom, bust, and the human body: further evidence on the relationship between height and business cycles, Economics and Human Biology, 3, 450-466. Ravn, M. O., Uhlig, H., 2002. On adjusting the Hodrick-Prescott filter for the frequency of observations, Review of Economics and Statistics, 84, 371-376. Saito, Osamu (2002).The frequency of famines as demographic correctives in the Japanese past, T. Dyson and C.O.Grada, eds, Famine Demography: Perspectives from the Past and Present, (Oxford University Press, Oxford), pp. 218-239. Steckel, R. H., 1995. Stature and standard of living. Journal of Economic Literature, 33, 1903-1940. Steckel, R. H., 2009. Height and human welfare: recent developments and new directions, Explorations in Economic History, 46, 1-23. Woitek, U., 2003. Height cycles in the 18th and 19th centuries, Economics and Human Biology, 1, 243-257. Zarnowitz, V., 1996. Business Cycles: Theory, History, Indicators, and Forecasting. The University of Chicago Press, Chicago.

Date: 21/03/2012 From: Daniele Brombal To: Carlos Rodrigues; Sandra Bastos Cc: Sandra Bastos; Tiziana Lippiello Subject: EAN Aveiro Workshop, Department of Asian and North African Studies - Ca’ Foscari University of Venice: Presentation Abstract

TITLE

Environmental Crisis in P.R.China: Ca’ Foscari Academic Exchange and Research Projects

ABSTRACT

Environmental crisis is one of the most pressing development issues faced by the current Chinese leadership. Over the last decade, environmental degradation—and relevant damages to human health—has been included among top priorities of government intervention. The recipe to face this crisis is composed by a complex mix of ingredients, including, among others, efforts towards a better management of contaminated sites and the increasing reliance on renewable (clean) energies. Despite political commitment, recurrent local crises and widespread worsening of key health indicators have shown that much remain to be done in order to ensure clean air, soil and water, reducing related health hazards.

Enhancing technical competencies available to relevant Chinese government agencies and research institutions, creating a strong partnership between Europe and China over environmental research, has become a priority for Europe’s strategic projection towards China. The University Ca’ Foscari of Venice is currently involved in two academic exchange and research programs funded under the FP-7 Marie Curie Actions/IRSES, namely the project “Global Partners for the Contaminated Sites Management” (GLOCOM) and the project “Evaluating Policies for Sustainable Energy Investments” (EPSEI). The two projects, which have been initiated in 2011 and will be finalized in 2013-14, are implemented in cooperation with 11 Chinese, Russian and European universities and research institutions.

Projects finance academic exchange activities among partners, providing the chance to conduct cross-competence joint research on issues pertinent to the areas of human and ecological risk assessment and management of contaminated sites, evaluation of policies for sustainable energy investments, and scientific decision making in environmental protection. Personnel of the Department of Asian and North African Studies is currently cooperating with the Department of Environmental Sciences, Informatics and Statistics of Ca’ Foscari University—among the recipients of EC funding under the framework of the above mentioned projects—offering its scientific contribution, analyzing from a social science perspective the institutional determinants of environmental policies and related public health interventions in P.R.China.

China Greening the World

Carlos Frescata BIOSANI, Palmela, Portugal, [email protected], www.carlosfrescata.com

By vastly accelerating the depletion of the planet’s natural resources, which sustain the model of consumption created by the West, China is revealing the eco- hypocrisy prevalent in the world, and rapidly raise the awareness of people all over the world that a new, more sustainable global model is, in fact, vital. The Western world, urbanised, post-industrialised, bourgeois in its consuming comfort, and self-proclaimed “environmentalist”, does not have the legitimacy to advise China not to follow the same path that it has trodden, so as to reach the same benefits currently enjoyed by the West, despite their inevitable short-term durability. There is an impressive capacity of the Chinese people for patiently taking on huge challenges over the long term. When an appeal is made by central government the collective interest prevails over individual interests and their profound ecological awareness can be revealed. China has the potential for becoming a world power for sustainability if, as part of its epic modernisation process, it seizes the strategic opportunity by opting for systems of the future which are exemplary in environmental terms. It can thus play a leading role at the cutting edge of ecologically-sustainable development in the 21st century, rather than simply following out-of-date models deriving from the industrial revolution of the 19th and 20th centuries. The country has people, technicians and activists who know what must be done in order for their country to become such a powerhouse of sustainability. The main hope associated with the China phenomenon lies in the opportunity for reaching out to people on the other side: this is perhaps the first time in the history of humanity that East and West will stand face to face, on an equal footing.

“Climate Change Mitigation: making the case for “free” technology transfer to China”

Abstract

Climate change is happening right now and more is already in the pipeline. If we keep on emitting CO2 at the present rates, we will commit the average surface temperature of the planet, according to the latest science, up to 7º C or more until 2100, with potential cataclismic impacts. Everyone in the planet is set to lose from these impacts, altough many will endure dire consequences. China has become the largest CO2 emitter in the world. In a few decades it will account for more CO2 emissions then the other major industrial countries combined. Coal accounts for 2/3 of China´s electricity production, being the major source of CO2 emissions. Scientists are urging for a moratorium on new coal-fired plants in the developed world and a gradual decommissioning of old and inefficient plants. These efforts may be useless if China keeps on building new coal-fired plants. China refuses to accept binding CO2 emissions reductions, claiming it is not responsible for climate change. Altough China´s responsibility has somehow changed and China itself will suffer much from climate change, there´s a urgent need to think of non-market based instruments for tackling China´s energy technology transfer needs. Market-based instruments such as CDM aren´t enough. In order to involve China, the developed world has to respond favourably to China´s demands to have free access to energy technology patents. The UN declared 2012 International Year of Sustainable Energy for All. Energy technology patents and intelectual property rights have to be seen in similar ways as drugs for AIDS and malaria.

Keywords: climate change; coal; technology transfer; energy

Streamlining Local Behaviour Through Communication, Incentives and Control:

A Case Study of Local Environmental Policies in China

Thomas Heberer and Anja Senz 1

Abstract

This article describes how China uses evaluation ratings and monitoring as incentives in order to foster the implementation of environmental policies at the local level. It is argued that decentralisation in China leaves room for actors at the local levels to manoeuver and bargain with those on higher levels for flexible adjustment of implementation policies according to local conditions. However, decentralisation is accompanied by significant institutional changes in the structure of intergovernmental communication, incentives and control. Accordingly, decentralisation in China exhibits a specific design which leaves space for divergent local environmental policies while also engendering “grass-roots mechanisms”. On the whole, this new institutional setting benefits the implementation of environmental policies.

Introduction

The process of liberalisation in China in the context of the reforms since the 1980s has significantly widened the operative scope of administrative levels below the central government. Competences have been shifted to provinces, municipalities and counties in order to foster economic development in accordance with the needs of each specific locality.

This liberalisation has enhanced the opportunity of actors at the local levels to manoeuver and bargain with superior administrative echelons. However, it has also led to inconsistencies in policy implementation. This leaves the central state in somewhat of a quandary: Should it insist that all localities across China implement a uniform policy, this could cause many difficulties. It could, for instance, negatively affect local economic development, e.g. by forcing polluting enterprises at the local level to close down in accordance with a generalized environmental protection policy. As a result, local unemployment could rise and evoke strong local opposition or even result in local instability.

Thus as part of the reform process after 1978, local governments ceased to be mere agents of the central state and became economic principals and advocates of local interests. They enjoy not only wider leeway in shaping and implementing local policy but are able to focus more on the protection and promotion of local social and economic interests (Gong 2006: 85-102).

In the academic literature, the phenomena of local leeway and divergent local policies have been treated as resulting from decentralisation. Generally, decentralisation can be understood as

the transfer of responsibility for planning, management, and the raising and allocation of resources from the central government and its agencies to field units of government agencies, subordinate units or levels of governments, semi-autonomous public authorities or corporations, area-wide, regional or functional authorities, or non-governmental private or voluntary organizations (Rondinelli, McCullough, and Johnson 1989: 59).

1 This article has been published in the „Journal of Current Chinese Affairs, 3/2011, pp. 77-112. Decentralisation itself encompasses a wide range of mechanisms, including institutional changes with shifts of decision-making authority (deconcentration), transfers of authority to specific authorities outside the regular bureaucratic structure (delegation), and transfers of legislative and executive powers to lower government levels (devolution). The effects of decentralisation, however, depend on the design of the individual decentralisation programme (Litvack, Ahmad, and Bird 1998: 26).

In the recent debate on effects of decentralisation in China, three primary positions can be distinguished: (a) Local development has been triggered less by decentralisation than by factors like competition between factions at the central level (see e.g. Cai and Treisman 2006: 505-535); (b) decentralisation processes have primarily generated only economic development (Cao, Qian, and Weingast 1999: 103-131; Xu and Zhuang 1998: 183-212; Qian and Weingast 1996: 149-185; Feltenstein and Iwata 2005: 481-501; Montinola, Qian, and Weingast 1996: 50-81); and (c) a far-reaching autonomy has been the key not only to economic progress but also to social, political and structural developments at the local level (Laundry 2008; Smoke 2005: 25-28). Moreover, Laundry argues that China is “one of most decentralised countries in the world” (Laundry 2008: 3), and Zheng Yongnian even speaks of a “de facto federalism in China” (Zheng 2007; Tsui and Wang 2004: 71-90). Decentralisation involves social and political costs in the form of possible implementation gaps, uneven development, loss of central control and state capacities, local clientelism or even state capture. However, as Qian, Weingast and others have shown, in China the benefits of decentralisation outweigh such disadvantages (Qian and Weingast 1996: 149-185; Feltenstein and Iwata 2005: 481-501).

With regard to the functions of decentralisation, there are two different opinions on what the functions of decentralisation are: (1) Decentralisation fosters an implementation of central policies in accordance with local conditions and preferences, and (2) decentralisation triggers competition among localities (Campos and Hellman 2005: 237). Both currently seem to be true for China. On the one hand, the local level is endowed with more decision-making power, particularly in terms of policy implementation according to local features. On the other hand, locality rankings have become increasingly prominent. Such rankings are based on three factors: results of policy implementation, successful innovations and policy experiments, and the outcome of evaluations by superior echelons.

This article is not primarily concerned with the costs and benefits of decentralisation but rather with the question of how possible negative effects of decentralisation are mitigated in China. Such negative effects – or “risks of decentralisation” – are: (a) the emergence of strong local-level clienteles where powerful coalitions of local officials directly or indirectly collude with entrepreneurs or certain social groups on specific issues (e.g. preventing the closure of unsafe mines or enterprises which contaminate the environment) and oppose the implementation of central environmental policies in the name of “local stability” or “economic development”; (b) imprecise norms and weak institutions, with overlapping responsibilities and an asymmetric information structure with strong boundaries between administrative entities, making environmental administration and problem-solving extremely complicated; and (c) local-level state capture and corruption due to lack of transparent procedures, (Bardhan and Mookherjee 2000: 135-139; Prud’homme 1995: 201-220). These effects can undermine the general intention of the central state to foster local development, flexible governance, and administrative effectiveness by means of decentralisation.

The risks of decentralisation have been widely discussed in the literature (see Heberer and Senz 2011: 535-546; Hongbin Cai and Daniel Treisman called this type of decentralisation “state-corroding” (2004: 819-843)), however, the mechanisms for minimising them have been somewhat neglected. What is more, most research focuses mainly on a possible loss of control over the local level as a result of decentralisation. The possibility of subtly influencing local behaviour with the existing administrative system is rarely tackled.

Taking environmental policies as an example here, we want to show not only that decentralisation in China fosters development according to local conditions but also that new mechanisms for streamlining local behaviour are required in order to avoid the negative consequences of decentralisation. This article proceeds from the hypothesis that decentralisation in China has both enhanced the discretion of local cadres for action and generated greater incentives for policy implementation by creating new methods of communication and control.

This paper examines in particular the way in which the Chinese system of evaluating policy implementation and local cadre behaviour has altered the basic structures of communication and the incentives for policy implementation. These new incentives for policy implementation according to local conditions are of key importance for a better understanding of the impact of decentralisation and development successes at the local level by virtue of local innovations.

The introduction of fixed implementation responsibilities by means of “target responsibility contracts” (agreements between counties and municipalities) in combination with the attempt to tie target fulfilment to career advancement and income have generated new incentive structures for local cadres. (Here, we focus on the hierarchical interaction between two entities (the county and its superior level). We use the term “municipality” when speaking of the administrative level above the county. We are well aware that this superior level sometimes consists of the prefectural level, sometimes of a municipality, and sometimes even of a province.) Additionally, the continuous evaluation of cadre performance has influenced policy implementation and the behaviour of leading local cadres. Concurrently, evaluations have spawned learning effects because the higher echelons draw attention to deficiencies in setting policy targets and the lack of preconditions for policy implementation (such as funding and human resources). Moreover, the pressure on local leading cadres to present visible policy outcomes and innovations has increased.

Up to now, the effects of evaluation on the behaviour of local cadres and on the implementation of environmental policies have been addressed only marginally in the literature (exceptions: Edin 2003; Whiting 2001; Gao 2009; Gong 2009). The same holds true for the role of local policy experiments and modelling in environmental policy implementation and cadre assessment (exceptions: Cao, Qian, and Weingast 1999: 123-124; Xu and Zhuang 1998: 194; Cai and Treisman 2009 Heilmann 2008).

With regard to administration, this article will focus specifically on the county level and to some extent on its interaction with the municipal level as its superior echelon. We take the county level as an entity of research here even though we are fully aware that counties are embedded in the hierarchy of China’s administrative system. However, as mentioned above, counties enjoy a certain discretionary power vis-à-vis the next higher level in shaping policies according to local conditions, and therefore are particularly suitable for our research. Since local cadres at the county and township level are responsible for the “real” work of governing, we focus on their behaviour, motives and interests with regard to environmental issues and ask about the role they are playing in implementing environmental policies. Our paper therefore examines two questions: First, to what degree do local cadres have a “creative space” for shaping policy implementation according to local interests? Second, what incentive and pressure instruments are used by the higher administrative level to ensure the proposed implementation of environmental policies?

After briefly introducing the database and methodology used in our field research, we examine the interactions between the central and the local level in terms of environmental policies and the leeway the local leadership has in shaping policies according to local features. Section 3 discusses Chinese evaluation systems as instruments for communication, incentive and feedback and as a means for exerting pressure. In Section 4, two case studies are analysed to illustrate the interaction between these evaluation systems and the corresponding implementation of environmental policies. Section 5 reflects on model- building as a steering mechanism in environmental policies, and the final section concludes that the institutional setting of local policy shaping and performance control in China effectively fosters environmental policy implementation and innovations according to local conditions.

Database and Methodology

This article is based on extensive field research conducted over a four-year period from 2007 to 2010 in collaboration with the China Centre of Comparative Politics and Economics in Beijing. It relies on the findings of three research projects, two of which compared environmental administration in China and Germany in both urban (Project 1, 2007) and rural (Project 2, 2009) areas. The third dealt with the issue of county and township cadres as strategic actors in the Chinese reform process (Project 3, 2008-2010). Field research was conducted in a total of 12 counties and cities in Xinjiang, Liaoning, Shandong, Fujian, Zhejiang, Sichuan, Jiangxi and Guizhou between 2007 and 2010. The field research comprised the locations and contexts as shown in Table 1. Table 1: Research Focus in Different Locations in China Year Location Rural/Urban Research Focus (R, U) 2007 Shehezi/ R + U Urban environment Xinjiang administration 2007 Xiamen/ U Urban environment Fujian administration 2007 Yingkou/ R + U Urban environment Liaoning administration 2008 Laixi + R Cadre behaviour/New village Qingdao/ construction Shandong 2008 Suining+ R Cadre behaviour/New village Anju construction 2009 Laixi + R + U Cadre behaviour/New village Qingdao/ construction Shandong 2009 Suining+Anju R Cadre behaviour/New village construction 2009 Shouguang/ R Rural environment Shandong administration 2009 Deqing/ R Rural environment Zhejiang administration 2009 Nanfeng/ R Rural environment Jiangxi administration 2010 Meitan/ R Cadre behaviour/New village Guizhou construction 2010 Xifeng/ R Cadre behaviour/New village Guizhou construction

Source: Field research conducted over a four-year period from 2007 to 2010 in collaboration with the China Centre of Comparative Politics and Economics in Beijing.

Qualitative interview techniques were used, for which a series of semi-standardised, open- ended questions were prepared. Projects 1 and 2 focused on issues of environmental administration, evaluations, environmental models, local environmental policies, and agencies. Project 3 concentrated on the role of local cadres and their group behaviour with regard to evaluations, programme implementation, and the creation of models with reference to the target of “constructing new socialist villages” ( Shehuizhuyi xin nongcun jianshe ).

Interview partners were selected from the offices and cadres directly involved in local policy design, implementation and evaluation. Particularly in Projects 1 and 2, this encompassed officials in charge of environmental affairs in cities, counties and townships, environmental experts, administrative personnel in local bureaus concerned with environmental issues, grass-roots officials in townships and villages, entrepreneurs, environmental NGOs, journalists, and ordinary citizens. Altogether more than 100 expert interviews were conducted. Furthermore, local documents on both environmental and evaluation issues were collected, including in particular official documents of local governments, articles from local newspapers, and local Internet resources.

The field sites were selected according to regional distribution and different levels of local development (strongly developed, middle-level development, less developed) and in discussions with each respective Chinese partner institute: Project 1: West (Xinjiang) – Southeast (Fujian) – North (Liaoning)

Project 2: Northeast (Shandong) – East (Zhejiang) – South (Jiangxi)

Project 3: Northeast (Shandong) – Southwest (Sichuan) – South (Guizhou)

Chinese Environmental Policies: Central State Steering and Local Discretion

In comparison to other developing countries, China’s central government started rather early to build up institutions to safeguard the environment and even anchored environmental protection in its constitutions (1978/ 1982) (Jun 2000: 143-145). However, while in the 1980s and 1990s only a handful of laws and guidelines ranging from correct use of different resources to noise emission and hazardous waste management were formulated, a total of 68 laws, regulations and guidelines regarding different environmental issues have been enacted in the decade since 2000 by the central government in Beijing (see MEP 2011a). What is more, the State Environmental Protection Administration (SEPA), which has been responsible for environmental issues since 1998, was made a federal ministry in 2007. This clearly illustrates the central government’s increasing interest in environmental issues.

But the emphasis of the central government does not necessarily translate into successful environmental policies at lower administrative levels. Although environmental policies have been successfully carried out in some areas, this has not always been the case in others, with the effect that the country in general regularly scores rather low in international environmental performance indices (Env. Perf. Index 2010). The 2010 Report on China’s Environmental Situation reveals that in 2010, despite improvements in water and air quality problems in fields like biodiversity, contamination by heavy metals and other toxins became more serious (MEP 2011c).

Currently, the central government employs three main mechanisms for implementing its environmental policies: (a) setting agendas, (b) allowing flexibility in environmental policy implementation, and (c) establishing cadre responsibilities and evaluation systems.

The Centre as Agenda Setter

In principle, the Centre formulates policies in the form of basic environmental ideas, thus putting the focus on issues considered urgent. Instead of precisely specified policies and legal norms, the party-state prefers to set generalised standards which require local interpretation and thus leave room for local modification. To be sure, such abstract standards can result in high administrative costs (Diver 1983: 65-109) inasmuch as unclear norms must be clarified at the lower levels of the political system. But this procedure transfers policy from the central-level actors to local-level actors, that latter of whom can then discuss and implement policy according to their needs, interests and capacities.

The standards of the Centre are fixed in documents which state the intentions of the party- state and its priorities. This kind of documentary politics has been characterised as a functional equivalent to legislation in Western democracies (Wu 1995: 24-38). In this way, and in particular with regard to the environment, the Centre in China can be termed an “agenda setter”, since standards are set out but concurrently leave space for the localities to bargain and modify central policies.

Since the Centre cannot possibly monitor all 2,862 Chinese counties in all the different policy fields, priorities are set by formulating “hard” policies such as economic development, stability, increases of local-level income, and birth control. Compliance with these policies is monitored rather rigorously.

Institutional and legal instruments are sometimes too weak to guarantee attention is paid to “soft issues” – most of which currently involve environmental protection. For this reason, political fragmentation and implementation gaps in several policy fields of minor relevance are accepted by the central government as part of the process of institutional change, pluralisation and flexibility required due to diverse local conditions throughout China. Such a lack of effective regulative institutions represents governance by agenda and policy formulation, with a simultaneous acceptance of apparently diverse interests of the central and local levels and selective Institutional and legal environmental policy implementation at the local level (O’Brien and Li 1999).

Acceptance of Flexibility in Environmental Policy Implementation

Another factor contributing to the Centre being an agenda setter is the perceived need for flexibility in policy implementation at the local level. Zhou Xueguang has discerned three categories of flexibility in policy processes in China:

(a) “flexibility by purposive design”, i.e. flexibility in central policies; (b) “flexibility of unintended design”, i.e. flexible implementation at the local level; and (c) “flexibility by special interests”, which means that the policy of the Centre is sometimes undermined by the interests of local cadres (Zhou 2010: 60).

This flexibility in policy implementation encompasses four features: adaptation to local conditions, implementation primarily of hard targets, local transformation of soft into hard targets (cf. the wuhua case in Laixi in one of the following sections), and model creation as a kind of local policy innovation along with self-recommendation vis-à-vis higher echelons. These will be discussed at greater length below.

Flexibility is reflected by China’s environmental policies, which are formulated and framed rather generally and vaguely – as described above – by setting standards instead of giving clear-cut directives. Although the central government requires implementation in principle, this is also to be attuned to local conditions. In locations with high unemployment, for instance, polluting enterprises are not to be closed down hastily, as this could spawn local protests and affect stability. Concurrently the central state provides funds for technically restructuring outmoded industrial facilities.

Obviously, this flexibility has the advantage of allowing policies to be implemented according to local conditions and the interests of the people concerned; however, it can equally well serve individual interests by leaving room for corruption, bribery, fraud, nepotism and the like.

Locally, we found widely differing conditions for implementing environmental policies and in the cadres’ behaviour in doing so: (a) The environmental behaviour of local cadres tended to differ based on top-down pressure and the existence of incentives for promoting ecological policies. (b) In locations with modern high-tech industries, like in the inner districts of Xiamen or in areas like Shouguang – the latter with its green agriculture and agricultural products processing industry – ecological issues were naturally more prominent than in cities with a strong legacy of heavy industry. (c) In locations where polluting industries are the major source of local revenues and employment and where financial resources for technical upgrading and re-equipment are lacking, the municipalities tended to accept the continued existence of polluting industries in order to avoid local unrest and rising unemployment. (d) Collusion of local cadres in order to conceal the real environmental situation and achieve a better evaluation outcome was prevalent in locations where policy implementation was faulty, targets too high, and resources and human capital for implementation lacking, and in those places unable to effectively mobilise or use resources, where individual or factional interests were being pursued, etc. In particular, less successful county leaderships attempted to hide such problems from their superior administrative bodies. In more well-off areas, on the other hand, local officials sometimes feared that corruption or misuse of funding or resources might be discovered, which sometimes stopped them from hiding problems in the same way as their less well-off counterparts.

Cadre Responsibility and the Evaluation Systems

Since the Chinese party-state lacks the institutions required for fully guaranteeing and monitoring policy implementation at the local level, the Centre needs further mechanisms to influence the behaviour of local officials. The cadre evaluation system referred to in the following section is an important instrument for doing so. It is an example of how the central government strives to improve governance at all administrative levels (Carter and Mol 2007). As early as the 1980s, both a cadre “target responsibility system” and an evaluation system for cadres and enterprises at the local level were established, in which contracts of higher administrative levels with local officials determine the most relevant tasks of policy implementation at the local level. It is argued that this has enhanced environmental governance, too (Lo and Tang 2007: 42-44).

Protecting career positions and being promoted in the hierarchy (Nomenklatura), two salient concerns of local officials, require that they prove their effectiveness by means of successful political and economic performance and that they are evaluated positively. Moreover, they must carry out projects successfully, thus drawing the attention and support of their superiors. Local experiments (“models”) are important in this regard, since they can demonstrate the innovative character of local leadership. The more successfully such experiments are developed, the more likely is the promotion of leading cadres who are involved. Being successful, adapting innovative policies, and being evaluated positively are thus crucial not only to the promotion of leading cadres but also to their applications to higher levels for funding of policy implementation and experimentation.

Environment Performance Evaluations

In the previous section, we argued that evaluation systems determine and structure the interaction between the counties and higher administrative levels in terms of environmental policies. In this section, we will 1) analyse this interaction by portraying the various evaluation mechanisms and their impact on environment policies and 2) specify the place of environmental evaluations in the interactions between the two levels. Finally, we will discuss the political significance of environmental policies in relation to China’s entire policy context. Evaluation Mechanisms

A “target responsibility system” and an evaluation system for cadres and enterprises at the local level were established in the late 1970s (Huang 1995: 828-843; see also Whiting 2001: 101-103; Whiting 2004: 102-105; Renmin Luntan 2009: 120-133). Each year, the prefectural city assigns specific policy targets to its counties (the county leadership). Those targets are stipulated in so-called “responsibility contracts”. The performance and behaviour of local leading cadres in each policy field are evaluated at the end of a given year. By means of these contracts, the municipality communicates the aims of its policies to county leaderships in order to guarantee policy implementation. The performance of leading local cadres at the county level is reviewed and assessed through evaluations and evaluation regulations, but the above-mentioned contracts also serve as incentives for them to comply with implementation policies (see below).

In fact, a multitude of contracts for the various offices and policy fields exists (cf. Edin 2003: 39). Environmental issues are just one item in a broad range of policy domains. The evaluation programme of Qingdao City in 2009, for instance, covered more than 1,000 items.

In principle, two types of evaluation exist: (1) a programme evaluation, which assesses implementation of the various policy targets and programmes and (2) a performance evaluation of leading county cadres by the municipality. In the first, the offices of a municipality responsible for a certain policy field inspect the county offices concerned. At the end of an evaluation process, the counties under the jurisdiction of that municipality are ranked according to the evaluation outcome (one explicit target is to encourage competition among administrative units in the environmental sphere, see Ruan 2011).

A third instrument of local cadre surveillance is the so-called “periodical priorities”, in Chinese: yipiao foujue (one item veto rule). It defines situations which must by all means be avoided or criteria which must be met. The meaning of yipiao foujue is that if one of the items concerned has not been accomplished, all other achievements of local leadership will be negated. If, for instance, birth-planning targets are not achieved, then all other indicators such as GDP growth or preserving social stability are void. Hence, a county or township leadership in this situation cannot be classified as “excellent”, and promotion of the county and township’s party secretary or mayor to a higher position is almost surely precluded.

Yipiao foujue targets (such as birth control or preserving social stability) are assigned by the Centre, the provinces, or by a county according to local requirements. It is therefore a must for local leaders to meet these targets, since failure to do so will negatively affect their evaluation and their individual career advancement and promotion. Yipiao foujue is not a part of the regular evaluation system, even though it is managed by the local programme evaluation bureau. Rather, a check is carried out at the final stage of an entire evaluation process to see whether there has been any violation against this “one item veto rule” (interview, Organisation Department, Meitan, 30 August 2010).

The evaluations described above are clearly a major control instrument used by higher-level administrative echelons to bring the lower levels into line. But they are more than a mere instrument of control. As the following section shows, such evaluations are a multifaceted tool.

Incentives, Control and Communication A pivotal question concerning the interaction between the municipalities and its counties is: what key instruments can the former use to encourage the latter to design and successfully implement environmental policies? It is certain that incentives are required to promote such behaviour. As mentioned above, one of the key interests of leading local cadres is to safeguard their current status and position while promoting their career advancement within the formal administrative hierarchy. This increasingly requires successful – even innovative – policy implementation and performance in line with current national policy trends. Here, the evaluations play a specific role.

However, incentives alone are not sufficient for steering the behaviour of local cadres. We therefore argue that evaluations have four major functions that can affect the interaction between a municipality and its counties. This holds for all policy fields and not merely for environmental policies.

Evaluations function as

(1) an instrument of political communication between higher and lower administrative levels by means of the aforementioned responsibility contracts. In those contracts, the higher level communicates its expectations in terms of environmental policy implementation to the county leadership and its sub-agents (township and village cadres). Environmental policies set out in the contract elucidate the significance assigned to environmental policies by the municipality leadership within the wide array of policy fields.

However, the contracts not only convey information from the municipality to the county but also vice versa: from the villages and townships to the county and from the county to the municipality. In the political domain, the information transmitted by the municipality encompasses the political agenda, policy priorities, methods for implementation, and targets to be fulfilled by subordinate entities. However, the aim is not to punish officials who fail to meet the targets but rather to encourage them to modify their thinking and behaviour and to improve their performance. This is to be achieved by educational means, like attendance at training courses at local party schools or face-to-face conversations with local leaders. These practices serve as a further tool of communication between local authorities in order to achieve compliance with policy contents and policy implementation.

(2) an incentive and steering system. Local officials must meet targets prescribed by higher authorities. The latter even expect that cadres conduct policy experiments, create models (for instance, in terms of environmental policies) and advance policy innovations in order to be positively evaluated and promoted. This is a major reason leading local cadres attempt to design specific policies and models according to local peculiarities and to implement priority tasks assigned by higher echelons. Leading local officials must meet the targets if they wish to advance up the career ladder and increase their income (through bonus payments for target fulfilment).

(3) a control and pressure system in the form of performance ratings (the party secretary of Wuhan city argued that cadres need to sense a certain uncertainty in the policy process, see Ruan 2011). As a rule, only cadres consecutively evaluated as “excellent” can – idealtypically – gain promotion. In a survey of 100 leading local grassroots cadres, over 30 per cent complained that the evaluations put enormous pressure on them (Zhu and Ye 2009). Our interviews with leading local cadres and Chinese scholars confirmed this. A major reason is that the evaluations take place frequently and regularly, can entail massive negative effects for career advancement, and thus make leading cadres feel insecure (many county and township cadres complained about this (interview Xiaozhaiba township, 4 September 2010). Moreover, the contracts include both rewards for target fulfilment and punishment in the case of non- fulfilment.

(4) a feedback system regarding policy implementation. Other feedback institutions in China are weak or even non-existent. Thus the evaluations are an important instrument by which the higher administrative echelons can measure whether leading local cadres have successfully implemented a policy.

Effects of the Environmental Evaluations

As far as environmental issues are concerned, the local Environmental Protection Bureaus (EPBs) are in charge of annually evaluating the environmental performance of local enterprises. However, our investigations revealed that no genuine evaluation process existed in the localities examined by us in 2007, primarily because the responsible departments were understaffed; evaluation in many locations was therefore based for the most part on reports by the enterprises, offices or officials themselves. Furthermore, environmental issues in those locations were regarded as “soft” targets, while economic (GDP) development and the preservation of social stability were the major “hard” objectives.

In Shihezi and Yingkou in 2007, for instance, environmental issues were not even mentioned in the official evaluation documents of those cities. Instead, officials of the local Organisation Departments argued that environmental protection was subsumed under the construction of a “harmonious society” and the “theory of scientific development” (interview, Shihezi, 8 September 2007). And an official responsible for evaluating the environmental behaviour of enterprises in an urban district of Xiamen noted that evaluations were not really significant (interview, Xiamen, 16 September 2007).

From 2008 to 2010, on the other hand, the situation in the locations examined changed noticeably: Local evaluation documents began to itemise environmental issues as a specific category, and environmental evaluations became more standardised. There were probably two major reasons for this. First, in 2008 the State Environmental Protection Administration was upgraded and became the Ministry of Environmental Protection – a clear indication that environmental policies were becoming more prominent; and second, the political centre increasingly emphasised the importance of environmental improvement and enhanced the catalogue of incentives to local governments (in the form of funding for energy-saving systems, waste incineration plants, waste water clarification plants, etc.). Assessments of environmental behaviour drawn up by those assessed themselves were no longer to be found. To be sure, some counties and cities handle environmental evaluations more strictly than others, and some even seek to find ways around the evaluations.

The items of the target responsibility contracts can generally be divided into “hard” issues – for instance, those related to current political priorities of the Centre such as economic growth, social stability, birth planning, etc. – and “soft” issues in other fields, such as the environment, social security, education, etc. The former are strongly linked to an official’s career advancement and rewards and are to be implemented under any and all circumstances; the latter, although mandatory, are not necessarily related to promotion and remuneration, so that pressure to implement them is weak. (In recent years, however, academics, journalists and even the party’s media have begun demanding that the environment should be a “hard” evaluation issue, compare e.g. Sun (2010)). However, the Laixi case in Section 4 below showed that a local environmental policy can become a “hard item” if a county gives priority to it.

In the following section, we will show how environmental policies currently oscillate between being perceived as minor (“soft”) and major (“hard”).

Soft and Hard Environmental Policies

For years, the Centre demanded priority for GDP growth and assigned a subordinate role to environmental protection. This stemmed from the conviction that economic growth would bolster the legitimacy of the party-state. Although this lopsidedness is now increasingly criticised, and policy priorities increasingly include non-economic factors (for example, the environment) (Zhu and Ye 2009), this change continues rather slowly at the local level. For instance, our investigations revealed that the 2008 and 2009 evaluation regulations for the townships in Anju district of Suining City listed environmental protection and energy saving merely as “common work targets” ( gongtong gongzuo mubiao ) and focused evaluation in those areas on the growth of GDP and fixed assets, local budget income, net peasant income, attraction of investments, and increased foreign trade. Concrete environmental projects were itemised only in the appendix to the regulations.

The contractual target responsibilities in this case specified that the Environmental Protection Bureau was to perform the following tasks: construct a wastewater treatment factory, reduce waste emissions and handle concrete environmental projects. A total of 110 points was allowed cumulatively for all fields of evaluation. The regulations ascribed 21 points for attraction of investments and 14 each for GDP growth, an increase of fixed assets, and improvement of the living standard of peasants. Thirty points in total were allowed for 23 items (one of which was environmental protection) under the umbrella term “common work targets”. Although specific points were not allotted for these 23 “general targets”, failure to meet them led to point deductions. One point was deducted, for instance, if serious or dangerous environmental pollution occurred and was not remedied within a prescribed time despite admonitions of the local EPB, and 0.5 points were deducted if reports were not submitted in time. Such “slaps on the wrist” illustrate that environmental policies in Suining City were not regarded as key issues (Suining shi Anju qu mubiao guanli 2009).

On the other hand, the environment is by no means a “soft” issue for some locations in China. Particularly in areas approved as environmental models or those intending to become such models, the environment has become a “hard” issue with political leverage. Central policies also encourage a diversification of policy priorities. In Sichuan, for instance, only four priority targets, one of which includes industrial energy consumption and waste discharge, are set and evaluated at the end of each year (Renmin Ribao 21 Sep. 2010; Sichuan Government 2010). In 2011, Shaanxi Province determined similarly that five priority fields would be evaluated in the future: GDP growth, financial income, income of the urban and rural population, energy saving/ emission reduction, and harmony/ stability (Dang 2011). Sichuan and Shaanxi were among the first provinces to make environmental standards equal in importance to other standards in the implementation and evaluation process.

We now take up two case studies (the wuhua project in Laixi county und the “circular” development in Shouguang County) to substantiate our argument that contracts which include evaluations and target responsibilities have a positive effect on environmental policy implementation. Case Studies: Implementation and Evaluation of local Environmental Programmes

The wuhua Programme in Laixi

Evaluation of the “five changes” ( wuhua ) programme initiated by Qingdao for improving the rural infrastructure and environment in Laixi is taken here as a case study to illustrate the interaction of Laixi’s leadership with both its superordinate level (Qingdao City) and its sub- agents (townships and villages). The wuhua programme encompassed the hardening of rural roads ( yinghua ) and the beautification ( meihua ), greening ( lühua ), illumination ( lianghua ) and cleaning up (jinghua ) of villages. We will begin by outlining why and how this programme is to be regarded as part of an environmental improvement programme.

In a Western context, the term “environment” is strongly related to clean air, water and soil, along with waste management and the protection of natural resources and wildlife (Umweltlexikon 2003). In other words, it involves issues of environmental protection in post- industrial societies. Originally, however, the term referred simply to surroundings in general, and the current Chinese term for “environment” ( huanjing ) corresponds to this in that it encompasses the natural environment, the social dimension (quality of life), and the preservation of natural resources (air, water, soil, etc.) (Baidu baike 2011).

Moreover, with regard to environmental policies in China we must distinguish between rural and urban development. While in urban areas a considerate use of natural resources in the sense of environmental protection is regarded as crucial, environmental issues in rural areas are connected first and foremost with a modernising concept of development. As our field studies showed, basic infrastructural aspects like the paving of rural roads, electrification, waste disposal, presence of sanitary equipment, supply of water, etc., were covered in many counties by the “constructing new villages” ( xin nongcun jianshe ) programme, which also encompasses the greening, illumination and “beautification” of the surrounding areas. In this sense, therefore, environmental policies in China often overlap with aspects of local development and are distinguishable only at a later stage from the general improvement of local living conditions, when they become clear policies for preventing environmental degradation.

We argue that wuhua constitutes the first step in an environmental improvement programme inasmuch as it aims to “civilise” villages and improve both the natural and social environment by so-called “civilisational projects” ( nongcun wenminghua ). In a follow-up step, such a “civilisational project” might easily shift its focus to genuine ecological issues such as enhancing the quality of water, air, and soil, organising waste management, caring for natural resources, etc. In light of our research focus on local environmental policies, it seems relevant to keep in mind that environmental issues in China are often linked closely with efforts to trigger civilisational processes and improve the living standard of the rural population. We therefore view this specific “civilisationalising” wuhua programme as a major step in building up an ecological civilisation. This three-year programme (2008–2010) for improving the rural infrastructure and village environment in Laixi may serve as a good example of the first step in this “civilisationalising” project.

In the following section, we wish to show first that interactions between municipalities and counties in terms of implementing and evaluating this programme have contributed to its success; secondly, we will show that there is a specific rationale behind the idea of using more developed areas as a basis for triggering local development. The wuhua programme was initiated by Qingdao to improve environmental conditions in its rural counties. To do so, the wuhua programme accredited county leaders with the following discretionary powers: a) to make implementation decisions according to specific conditions in the townships and villages; b) to select priorities within sub-domains of the wuhua programme; c) to create distinct models; d) to select locations (townships, villages) for implementation; e) to tap additional resources for implementation and f) to urge offices and enterprises, including private businesses, to take over obligations for poorer villages by supporting wuhua implementation not only financially but also by providing resources and know-how. Although the programme was imposed top-down by Qingdao, its concrete implementation was in the hands of the Laixi leadership and the townships.

Of the county’s 871 villages, 301 were ordered to implement wuhua in 2008, another 301 in 2009, and the remaining 269 in 2010. The county leadership held the townships responsible for ensuring that Qingdao City’s evaluation of implementation by the villages and townships would bring to light no major problems.

As a means of controlling the process, the municipality imposed the following condition: Should Laixi fail to produce a positive evaluation of wuhua in the first year, only half of Qingdao’s promised 34 million CNY would be paid out. Moreover, Qingdao might reduce or even annul its funding in the following year. This would have had severe consequences for the evaluation of the entire county and its leadership (interview, Qingdao, 1 September 2009).

The county leadership attempted to guarantee fulfilment of the targets by means of regular meetings with the township and village leaders and, when necessary, it lowered the quality requirements of a measure (for example, concerning road construction) in order to ensure at least formal implementation (interview, Laixi, 21 August 2009). Along with the evaluation system, funding provided by Qingdao ensured motivation and guided the activities of the leading local cadres. In order to fulfil the above-mentioned wuhua targets and keep expenditures low, the local government began by selecting more developed and wealthy villages close to highways for the first year. In the second year, the county requested additional funding from Qingdao City and implemented the programme in rather average villages. Concurrently, the county leadership instructed local enterprises and government departments to take over responsibility for specific villages and to contribute to the programme financially at the village level. The county also redirected financial means from various funds (for example, for poverty alleviation) to the implementation of the programme.

Accordingly, implementation in the first year was rather easy and successful and was positively evaluated by Qingdao. In the second year (2009), the county leadership selected villages that could contribute a share to the funding or that had good connections to local enterprises. In the final year (2010), the county focused on villages with only marginal funding resources, believing that poor and remote townships and villages would be unable to implement the wuhua programme without external support.

This latter step was not without its problems. As a leading official of the county government recalls, due to the consequences of the financial crisis Qingdao had already reduced its funding twice, in 2009 and 2010. As a consequence, implementation of wuhua in the remaining villages had to be completely supported by public funding. About 20 per cent of Laixi’s villages belonged to this category, including some without even an existing village administration and/ or party committees. Given the lack of adequate funding for 2010, Laixi City saw itself constrained to enter into a bargaining process with Qingdao so as to ensure programme fulfilment. The Laixi leadership had good arguments at its disposal in this regard: a) It had fulfilled the targets of the first two years; b) it had problems funding the poorer villages, particularly because income from the local export industry had declined sharply due to the global financial crisis; c) due to that crisis, most enterprises refused to contribute to further funding; and d) peasants in the poorer areas were neither willing nor able to contribute financially to the implementation of wuhua (interview, Qingdao, 1 September 2009). Moreover, due to the effects of the worldwide financial crisis and the reduction of funding provided by Qingdao, the leadership of Laixi had already decided to shift the focus for 2009 from road hardening (which is more expensive) to the much cheaper target of cleanliness of villages, a change approved by the Qingdao government.

In entering into a bargaining process with Qingdao, Laixi aimed to achieve the following: 1) an adjustment of evaluation targets so as to better reflect the financial capacities of its townships and villages, 2) an increase in subsidies, and 3) an extension of the deadline for implementing the programme. As a leading official of Laixi’s Agricultural Commission stated, Qingdao had an obligation to increase its funding of the programme because – as mentioned above – the peasants in the poorer villages were either unwilling or unable to pay for it (interview, Laixi, 21 August 2009). Thus if Qingdao wanted to have the programme fully implemented, it needed to provide additional funding (interview with the person responsible for agriculture, Agricultural Commission, Qingdao, 8 September 2008).

To convince Qingdao of this, the Laixi leadership took the evaluators not only to advanced models but also to remote villages so as to show both sides of the implementation process – the advancements and the constraints. The local government believed that this would have a positive impact upon the evaluation outcome (interview, Laixi, 21 August 2009).

For the year-end evaluation of wuhua , Qingdao City drew up evaluation guidelines for the counties and urban districts under its jurisdiction. The guidelines communicated in detail which policy fields were to be evaluated and what were the maximum points assigned to full task fulfilment in each field. Laixi adopted the Qingdao evaluation criteria for evaluating its townships. The 2008 regulations for programme evaluation of the townships of Laixi stipulated development in 19 fields such as local revenues, large-scale industries, exports, peasant income, local GDP development, etc. Land resources and environmental protection as well as the construction of a new countryside constituted two separate items (Laixi Government 2008a). The evaluation of the so-called “key programmes” was weighted as follows, whereby the percentages give some indication of the significance and weight of each item: economic construction: 50 per cent; social fields (preserving stability, developing social insurance systems, building new houses for villagers, insuring employment and public security, etc.): 15 per cent; constructing rural party organisations: 14 per cent; political construction (improving village elections, increasing the effectiveness of local people’s congresses and political consultative conferences, successfully combating corruption, etc.): 8 per cent; cultural construction: 8 per cent; other items (including environmental issues and wuhua ): 5 per cent (interview, Party School Qingdao, 30 August 2009). These guidelines were intended to communicate to leading local cadres which policy fields were important and which were less or not at all so.

On the surface, the outcome of the entire rural infrastructural programme had little effect on careers, income or bonuses. The percentages showed no prominent role of environmental policies in the policy evaluation process of Qingdao and its counties. Thus wuhua seemed to belong rather to the “soft” evaluation categories. In reality, however, the wuhua programme as a civilising project was politically important for local development and was therefore highly rated by higher authorities who considered its implementation vital for developing the rural areas and their environment and for improving the living conditions of villagers. Moreover, Qingdao authorities allocated large amounts of funding to this project. Accordingly, wuhua turned into a hard programme which by all means had to be realised, and its implementation was evaluated rather strictly (Laixi Government 2008b).

To sum up, what was the major incentive for Laixi’s leadership to implement the wuhua programme successfully? We argue that its evaluation was all the more crucial inasmuch as wuhua was in the focus of Qingdao’s evaluation at the end of each year. The Laixi leadership was fully aware of what might happen if they were believed to be incapable of bringing about major changes in the rural areas’ environment and infrastructure within a certain time frame. This might have even led to the end of their career prospects. Therefore, the sense of foreboding surrounding evaluation played a significant role as a strong incentive for policy implementation. On the other hand, the wuhua case proves that local “soft” policies may become “hard” if local development requires it.

Circle Development in Shouguang

We (the authors) at times got the impression that Chinese environmental policies in the end are no more than a kind of showcase politics. In Shouguang, for instance, specific environmental attention was directed in 2009 towards villages located on highways. The party secretary of a village some distance away from a highway complained that evaluation inspection teams only visited villages close to highways, since more distant villages were too difficult to drive to. Accordingly, it was argued, the responsible township governments had little incentive to pave roads, to provide environmental facilities, or even to set up waste containers and dispose of waste (interview, Shouguang County, 25 February 2009. Similar remarks could be heard in townships in Sunining City and Nanfeng County).

An informal talk with township cadres painted a somewhat different picture. They admitted that the villages close to highways had the best chance of receiving subsidies for paving roads and were provided with waste containers. It was also the case, however, that improvements in the infrastructure of villages took place with financial means made available by both the prefectural city (Weifang) and Shouguang County, both of which demanded that funding had to be matched by the villages either in the form of additional funding or manual labour (like road construction). Since the more well-off villages could contribute more funding, investments there would not only be less costly but also rapidly show results. Since local funds are often scarce, the more developed villages close to highways tended to receive support first. Since almost no funds were available in the less developed villages, investments in them were much costlier (interview, Shouguang County, 25 February 2009).

Behind such deliberations like those in Laixi and Shouguang stands the logic of a civilising development by concentric circles: First, support the cluster of more well-off villages close to highways. This may lead to a positive evaluation and therefore improve the preconditions for bargaining with the prefectural city for additional funding and for progressing to the second circle. As a consequence, a second circle could be developed, then a third, etc. The most remote and poor villages would then certainly be the last ones to be developed. However, the township officials explained that conditions of bargaining with the prefectural city improved notably whenever the programme successfully met the criteria of the first circle. This made it much easier to acquire additional funding from the higher level for the development of the poorer villages. The above description by township cadres in Shouguang illustrates that there is a rationale behind the development concepts, albeit the peasants in the less-developed villages might complain of being the last to receive government support. The conjunction of target contracts with communication of goals, funding by the municipality, and control by means of evaluations proved to be a strong stimulus for wuhua policy implementation. Moreover, there is even a certain flexibility in the interactions between county and municipality: If a county runs into (financial) problems, it can enter into a further bargaining process with the municipality.

The next section is concerned with local policy innovations. County leaderships attempt to develop specific experimental “models” to prove that they are innovative, thus enhancing their standing vis-à-vis superior bodies and improving their chances of achieving a better evaluation rating and further prospects of career advancement. The option and choices of creating policy innovations are not only part of a “Chinese” decentralisation process but are also directly tied to evaluation processes.

Local Policy Innovations: Environmental Modelling

In recent years, the Centre has fostered the creation of “ecological models” and other environmental experiments in order to trigger environmental improvement and development. Primarily counties without major heavy industry or high tech industries are eager to acquire the status of such a model, since this not only improves the outcome of performance evaluations but can also attract additional funding by higher echelons. Here we wish to illustrate how and why cities and counties make efforts to become environmental or ecological models and how such models operate.

Since 2011, the Ministry of Environmental Protection has imposed new criteria for the nomination of a city or county as an ecological model city on the national level. Among them are: (a) economic and social criteria, like the income of the urban population, investment in environmental protection, reduced energy and water consumption, and a reduction of emissions several years running; (b) quality of environment (meeting specific environment standards set by the ministry for 85 per cent of the days in a given year), improving the quality of air and water, noise reduction; (c) environmental construction: 35 per cent of new public space must be green; 80 per cent of wastewater must be purified; increased savings in energy; improved garbage disposal; etc.; (d) environmental administration: drawing up an environmental protection and improvement plan; surveys attesting that more than 80 per cent of the population are satisfied with the local environment situation; establishment of environmental education programmes in schools; etc. (MEP 2011b; Wu et al. 2005). Specific incentives for environmental experimenting and modelling were also initiated.

If the Ministry of Environmental Protection approves a city or county’s application to become a national model, that city or county finds it much easier to acquire funding of environmental and ecological improvement projects (for energy-saving equipment, wastewater treatment plants, waste incineration plants, etc.) from its province and the Centre. The province, the prefectural city and the county must also match central funding. In case a city or county becomes a provincial model, those models are financially supported by the provincial and prefectural level. Acquiring the status of a model concurrently enhances the reputation of a location and its leadership. However, when a location is accepted as a model, the superior echelons impose strict conditions regarding the establishment of new enterprises, wastewater treatment facilities, waste emission limits, further savings of energy and water, the insulation of buildings, and the improvement of water and air quality. Compliance with the criteria for being a national “model” is strictly monitored by the Ministry of Environmental Protection.

As a rule, such an application starts out with a decision at the prefectural, provincial or national level that a given county should be acknowledged as an ecological model in order to boost the local economy (green agriculture, tourism, attraction of investments in environmentally friendly industries, etc.). The decision is followed by submission of an application for funds to the higher levels. Applications are then forwarded to the respective branch of the Development and Reform Commission ( Fagaiwei : a commission on the prefectural city, province or national level).

“Model” status is not without a time limit. Higher echelons regularly check whether a “model” is abiding by the regulations. In 2010, for instance, the central government decided that all cities and counties which had obtained model status prior to 2008 had to re-apply before the end of 2011 (Hebei EPB 2010).

Creating environmental models is not only part of the entire policy incentive system but also helps to give environmental issues more weight. The broad diversity of ecological models and their increasing prominence in policy documents and the media highlight the increasing role of ecological policies.

Since 2007, for example, Shouguang – a city on county level – has been since both a “national ecological city” ( guojia ji shengtai shi ) – a designated national model for environmental (MEP 2009) – and a model for “green” agriculture. Further improving the county’s environment long ago became a pivotal issue within the local cadres’ responsibility system (interview, Bureau for Development and Reform ( Fagaiju ), Shouguang, 24 February 2009). Suining, for example, aimed to become a “green city” in order to develop tourism. Here, energy saving and the reduction of industrial emissions were crucial and were included among the “periodical priorities” ( yipiao foujue ). Similarly, Deqing, a “national ecological county” aimed not only to develop eco-tourism but also to persuade people from the nearby provincial capital of Hangzhou to purchase much cheaper condominium flats in that “clean” county and to move to Deqing. Laixi, a “national ecological model” since 2005, ran into problems in 2008 when higher authorities discovered that a dozen or so enterprises there were major polluters. Accordingly, at the behest of the Qingdao government (also a national ecological model since 2000), Laixi was criticised for neglecting environment policies and had to make environmental protection part of yipiao foujue in order to keep its model status (the county government promised to “establish a soft environment”). Xiamen in turn is a national model for energy saving and renewable energies. Nanfeng County, one model level below, was accepted in 2008 as a “national ecological model area” ( guojia ji shengtai shifan qu ) and gave priority to environmental issues in order to be classified as a national ecological model county. Meitan County has been a “national ecological model area” since 2010. The county subsumed environmental issues under the category of “production safety” as part of the yipiao foujue items. Shihezi has striven to become a model for recycling and to develop the corresponding industry. Finally, Xifeng County has targeted becoming a model county for “constructing an ecological culture” ( shengtai wenming xian jianshe ), something strongly related to attracting ecological enterprises, general improvement of the environment, and increased environmental awareness on the part of local officials and the population.

The examples provided by these cities and counties illustrate that a broad variety of ecological models, backgrounds and contents exists in China. Some have been chosen by national authorities to become models (Xiamen, Qingdao), others by provincial authorities (Deqing, Nanfeng). In the case of Nanfeng, Jiangxi Province has aimed to enhance its environmental status and to create a new national ecological model on the county level. In Xifeng, the party secretary, a graduate of Qinghua University in Beijing with expertise in environmental sciences, wanted to send a clear political message to the higher authorities of Guiyang City, to which the county belonged.

Shouguang (green vegetable), Meitan (green tea and vegetable) and Suining (green tourism) have aimed to become ecological models in order to gain further advantages in market competition. Laixi in turn was pressured by Qingdao City to improve its environmental standards. Additionally, it wanted to attract more investments from high-tech and environmental technology enterprises.

These examples illustrate how economic development based on specific local conditions plays a major role in the model-building process in China. Holding one’s own against competitors and defining one’s own local direction of development, particularly if other resources (financially means, external investments, mineral deposits, etc.) are lacking, are major driving forces behind the choice of an ecological development path. Either the county leadership itself puts forward such a proposal and communicates it to higher authorities, or the initiative comes from the municipality in a top-down manner. Several prerequisites are decisive for determining the path to be chosen by a county: directives from the superior echelon; incentives provided by higher echelons (funding, prestige, becoming a model); local particularities (for example, preconditions for developing eco-tourism); and individual connections to higher echelons.

Undoubtedly, successful and efficient reduction of environmental pollution and contamination do not take place overnight, but rather develop in each case as a process. In the counties we examined, this process was indeed progressing. However, this does not mean that in model counties no problems exist at all.

It is also the case that deficiencies of environmental degradation are sometimes concealed, and fake data are sometimes prepared by the counties so they can acquire the status of an ecological model or achieve a better evaluation. In order to fulfil the criteria for model counties, for instance, Nanfeng County was officially reported to be 71 per cent forested. This high percentage was achieved by designating the vast number of tangerine trees, the major product of this county, as “forest area”. This violated China’s regulations on statistical reporting (interview Nanfeng Forest Bureau, 10 March 2009).

Our own observations in Nanfeng County showed that the preconditions for becoming an ecological model were still widely lacking. In many places in the county, wild rubbish dumps were found for which – according to local authorities – the peasants were responsible. With the exception of a handful of model villages, garbage collection in the rural areas was non- existent. For their part, the local leaders shifted the blame for these deficits, which they discussed openly, elsewhere: In a speech on “constructing a new socialist countryside” in 2008, Nanfeng County’s party secretary blamed not the local administration for environmental pollution but rather the peasants. The rubbish issue, he stated, was primarily a peasant issue: The peasants were defecating and relieving themselves everywhere, so the thinking of the peasants must be altered (Online: (12 March 2009)). According to him, environmental pollution had to do with the “quality” ( suzhi ) of the peasants and was therefore an ideological issue. In contrast to the party secretary, the mayor of this county took a more pragmatic view. In his report to the local People’s Congress in February 2008, he specified the key environmental problem as the destruction of natural forests by uncontrolled expansion of tangerine cultivation. Cultivation on mountain slopes had led to waterlogged depressions and erosion. The overuse of chemical fertilisers and pesticides in agriculture and the enormous contamination by a chemical and a concrete factory were identified as further key problems (Jiangxi Survey 2008). In Nanfeng County in 2009, the evaluations still focused on GDP development, investments, social stability, revenues, etc., as “hard” categories, while environmental issues played a minor role, not negatively affecting the local leadership. In awarding the title of an experimental ecological model to Nanfeng County, the higher echelons aimed to provide incentives to it for developing into a real model. All the more so inasmuch as this county had good preconditions for ecological development (few heavy industrial enterprises, a beautiful landscape, cultural attractions, etc.) and could easily develop into a location for ecological tourism.

To sum up, model status in China fosters the orientation of localities toward environmental and ecological issues and serves as an incentive to streamline local behaviour in the direction of environmental protection. More and more counties are focusing on such issues, particularly if other resources for development are not available. Acquiring the status of an ecological model can not only fuel economic development (e.g. ecological tourism) but can also attract additional funding from higher administrative echelons. It is also important for positive evaluations and thus for the career advancement of a county’s leading cadres.

Conclusion

Decentralisation has endowed the counties in China with greater decision-making power. Even though the localities remain embedded in a hierarchical system, they now have leeway to define their key areas of development within the focal policies of the central government. To be sure, environmental policies are increasingly playing a more prominent national role. Nevertheless, the Centre gives the counties some leeway to decide whether or not environmental issues are to be regarded as “hard” policies. The natural and financial endowment of a given locality and its legacies (a heavy industrial or agricultural focus) play a salient role here.

Still, decentralisation alone cannot solve all problems in environmental governance. A combination of policy communication between higher and lower echelons, incentives and pressure tools are required. The evaluation systems and the mechanism of creating models foster innovative policies (which can enhance evaluation outcomes). They are strongly tied to the career advancement of leading local cadres and have thus developed into major policy implementation incentives.

To be sure, areas lacking resources for environmental policy implementation and those with a poor policy implementation record, or those or those whose leading cadres are simply predatory , can turn to collusion in order to present mere showcases or enable rent-seeking. Incentives may play a minor role where corruption is rampant.

However, our own field research revealed that the local level is not always a “malign” state without development impetus. The Centre has established a system of strong incentives for county leaderships. It has institutionalised evaluation processes and actively provides funds for which a county may apply in order to improve its environmental and ecological situation. We may even argue that deception and showcase politics by local leading cadres have become increasingly difficult. Pressure from the higher level to make progress and to provide innovation (models) in terms of environmental protection, and to implement environmental policies, is increasing. On an annual basis, the county leaderships have to prove developments and progress vis-à-vis the superior echelons. For evaluators it is increasingly unacceptable to be shown the same environmental “models” every year. As a former township cadre leader in a county of Qingdao noted accurately:

Evaluators from Qingdao want to see fresh outcomes. They will criticise you if you have nothing new to present. In the latter case they may classify your performance as “not fulfilled” and you may get a negative evaluation or a point deduction which might negatively affect your further career prospects (interview, Qingdao, 2 September 2008).

To summarise, evaluation processes in China show that not only do the counties have room to manoeuver in terms of policy priorities but also that these processes offer incentives for policy implementation, particularly for implementing environmental policies. In China, decentralisation is a gateway for improving policy implementation, even in the domain of environment. The incentive and communication system that has been described here is a relevant factor in the specific design of decentralisation; otherwise all benefits of decentralisation – not just those in the domain of environmental policies – would have come to naught.

References

Baidu baike ( 百 度 百 科 , Baidu encyclopedia) (2011), 环 境 (huanjing , Environment), online: (24 February 2011).

Bardhan, Pranab and Dilip Mookherjee (2000), Capture and Governance at Local and National Levels, in: American Economic Review , 2, 135-139.

Cai, Hongbin and Daniel Treisman (2009), Political Decentralisation and Policy Experimentation, in: Quarterly Journal of Political Science , 1, 35-58.

Cai, Hongbin and Daniel Treisman (2006), Did Government Decentralisation Cause China’s Economic Miracle?, in: World Politics , 58, 4, 505-535.

Cai, Hongbin and Daniel Treisman (2004), State-corroding Federalism, in: Journal of Public Economics , 88, 3-4, 819-843.

Campos, Jose E. and Joel S. Hellman (2005), Governance Gone Local: Does Decentralization Imporve Accountability?, in: World Bank (ed.), East Asia Decentralizes. Making Local Government Work , Washington, DC: World Bank, 237-252.

Cao, Yuanzheng, Yingyi Qian, and Barry R. Weingast (1999), From Federalism, Chinese Style to Privatization, Chinese Style, in: Economics of Transition , 7, 1, 103-131.

Carter, Neil T. and Arthur P. J. Mol (eds.) (2007), Environmental Governance in China , London/ New York: Routledge.

Dang, Yanwen (2011), Yong kexue kaohe tuidong kexue fazhan (Foster Scientific Development by Scientific Evaluation), Renmin Ribao , 26 January.

Diver, Colin S. (1983), The Optimal Precision of Administrative Rules, in: Yale Law Journal , 93, 65- 109. Edin, Maria (2003), State Capacity and Local Agent Control in China: CCP Cadre Management from a Township Perspective, in: China Quarterly , 173, March, 35-52.

Environmental Performance Index (2010) , Yale Center for Environmental Law & Policy, Yale University , online: (22 May 2011).

Env. Perf. Index see Environmental Performance Index

Feltenstein, Andrew and Shigeru Iwata (2005), Decentralisation and Macroeconomic Performance in China: Regional Autonomy has Its Costs, in: Journal of Development Economics , 76, 481-501.

Gao, Jie (2009), Governing by Goals and Numbers: A Case Study in the Use of Performance Measurement to Build State Capacity in China, in: Public Administration and Development , 29, 21- 31.

Gong, Ting (2009), Institutional Learning and Adaptation: Developing State Audit Capacity in China, in: Public Administration Development , 29, 33-41.

Gong, Ting (2006), Corruption and Local Governance: The Double Identity of Chinese Local Governments in Market Reform, in: The Pacific Review , 1, 85-102.

Hebei Environmental Protection Bureau (2010), Guanyu jin yibu zuohao chuangjian guojia huanjing baohu mofan chengshi fuhe he zai pinggu gongzuo de tongzhi [Notification of the rechecking and reassessment of the status of an environmental model cities], online: (18 March 2011).

Hebei EPB see Hebei Environmental Protection Bureau

Heberer, Thomas and Anja Senz (2011), Impact of Environmental Change on Stability and Conflict Potentials in China, in: Hans Günter Brauch et al. (eds.), Coping with Global Environmental Change,Disasters and Security , Berlin, New York: Springer-Verlag, 535-546.

Heilmann, Sebastian (2008), From Local Experiments to National Policy, in: The China Journal , 59, 1-30.

Huang, Yasheng (1995), Administrative Monitoring in China, in: The China Quarterly , 143, 828-843.

Jiangxi Survey (2008), online: (10 May 2009).

Jun, Jing (2000), Environmental Protest in Rural China, in: Elisabeth Perry and Marc Selden (eds.), Chinese Society: Change, Conflict and Resistance , London/ New York: Routledge, 143-145.

Laixi Government (2008a ), Laixi shi gongzuo mubiao guanli kaohelingdao xiaozu bangongshi wenjian: Guanyu yinfa 2008 niandu zhen (jiedao) mubiaojixiao kaohe shishi xice de tongzhi [Notification on the Document of the Office of the Leading Group for Performance Evaluation of Laixi City on the Detailed Regulations of Implementing Target Performance Evaluation], Laixi.

Laixi Government (2008b), Laixi shi xinnongcun wuhua gongcheng jianshe qingkuang huibao [Report on the situation of the wuhua project of a new countryside in Laixi], online: (21 December 2010).

Laundry, Pierre F. (2008), Decentralized Authoritarianism in China: The Communist Party’s Control of Local Elites in the post-Mao-era , Cambridge: Cambridge University Press. Litvack, Jennie, Junaid Ahmad, and Richard Bird (1998), Rethinking Decentralisation in Developing Countries , Washington, DC: World Bank.

Lo, Carlos Wing-Hung and Shui-Yan Tang (2007), Institutional Reform, Economic Changes, and Local Environmental Management in China: The Case of Guangdong Province, in: Neil T. Carter and Arthur P. J. Mol (eds.) (2007), Environmental Governance in China , London/ New York: Routledge, 42-44.

MEP see Ministry of Environmental Protection

Ministry of Environmental Protection (2011a), online: (15 April 2011).

Ministry of Environmental Protection (2011b), Guanyu yinfa “guojia huanjing baohu mofan chengshi chuangjian yu guanli” gongzuo banfa [Regulations for “Creating and Administering National Environmental Model Cities], online: (1 March 2011).

Ministry of Environmental Protection (ed.) (2011c ), 2010 Zhongguo huanjing zhuangkuang gongbao [Report on China’s environment situation 2010], Beijing.

Ministry of Environmental Protection (2009), online: (25 February 2011).

Montinola, Gabriella, Yingyi Qian, and Barry R. Weingast (1996), Federalism, China Style: The Political Basis for Economic Success, in: World Politics , 48, 50-81.

Nanfeng County Government (2008), online: (12 March 2009).

O’Brien, Kevin J. and Lianjiang Li (1999), Selective Policy Implementation in Rural China, in: Comparative Politics , January, 167-186.

Prud’homme, Rémy (1995), The Dangers of Decentralisation, in: World Bank Research Observer , 2, 201-220.

Qian, Yingyi and Barry R. Weingast (1996), China’s Transition to Markets: Market-Preserving Federalism, Chinese Style, in: Journal of Policy Reform , 1, 149-185.

Renmin Luntan (ed.) (2009), Juece neican. Da guo “gan kao” [Decision-making, Internal Reference. ‘Imperial Examinations’ of a Big Country], Beijing: Huawen chubanshe, 130-133.

Rondinelli, Dennis A., James S. McCullough, and Ronald W. Johnson (1989), Analyzing Decentralisation Policies in Developing Countries: A Political-Economy Framework, in: Development and Change , 20, 59.

Ruan, Chengfa (2011), Zerenzhi jiu yao rang ganbu “ziwei” (The Responsibility System Should Lead to Uncertainty Among Cadres), in: Renmin Ribao , 18 April.

Sichuan Government (2010): Sichuan sheng renmin zhengfu guanyu yinfa Sichuan sheng renmin zhengfu shi (zhou) mubiao guanli banfa (shixing) de tongzhi [Circular of the People’s Government of Sichuan on the Provisional Methods of Target Management of Cities and Prefectures], online: (11 May 2011).

Smoke, Paul (2005), The Rules of the Intergovernmental Game in East Asia: Decentralisation Frameworks and Processes, in: World Bank (ed.), East Asia Decentralizes. Making Local Government Work , Washington, DC: World Bank, 25-28. Suining shi Anju qu mubiao guanli 2009 [Target administration of Anju district, Suining city], 2009.

Sun, Xiuyan (2010), Rang huanbao kaohe chengwei ‘ying ganggang’ [Let Environment Evaluation Become a ‘Hard Criterion’], in: Renmin Ribao , 1 July.

Tsui, Kai-yuen and Youqiang Wang (2004), Between Separate Stoves and A Single Menu: Fiscal Decentralisation in China, in: The China Quarterly , 177, 71-90.

Umweltlexikon (2003), Umwelt , online: (24 February 2011).

Whiting, Susan H. (2004), The Cadre Evaluation System at the Grass Roots: The Paradox of Party Rule, in: Barry J. Naughton and Dali L. Yang (eds.), Holding China Together. Diversity and National Integration in the Post-Deng Era , Cambridge: Cambridge University Press, 102-105.

Whiting, Susan H. (2001), Power and Wealth in Rural China: The Political Economy of Institutional Change , New York: Cambridge University Press.

Wu, Guogang (1995), ‘Documentary Politics’: Hypotheses, Process, and Case Studies, in: Carol Lee Hamrin and Suizheng Zhao (eds.), Decision-Making in Deng’s China, Perspectives from Insiders , Armonk: M. E. Sharpe, 24-38.

Wu, Qiong, Rusong Wang, Hongqing Li, and Xiaobo Xu (2005), Shengtai chengshi zhibiao tixi yu pingjia fangfa [System of criteria for ecological cities and evaluation methods], in: Shengtai Xuebao [Acta Ecologica Sinica], 8, 2090-2095, online: (1 March 2011).

Xu, Chenggang and Juzhong Zhuang (1998), Why China Grew: The Role of Decentralisation, in: Peter Boone, Stanislaw Gomulka, and Richard Layard (eds.), Emerging from Communism: Lessons from Russia, China, and Eastern Europe , Cambridge, MA: MIT Press, 183-212.

Zheng, Yongnian (2007), De Facto Federalism in China: Reforms and Dynamics of Central-Local Relations , New Jersey: World Scientific Publishing.

Zhou, Xueguang (2010), The Institutional Logic of Collusion among Local Governments in China, in: Modern China , 36, 1, 60.

Zhu, Peixian and Fan Ye (2009), Jianli cujin kexue fazhan de ganbu kaohe pingjia jizhi [Establish cadre evaluation mechanisms fostering scientific development], in: Renmin Ribao , 23 November.

Shifts in multi-level climate change governance in China – an example of ‘hollowing out’ the state?

Literature on domestic climate change governance in China mainly focuses on hierarchical governing modes and the central government’s top-down approach to steer and coordinate climate change policies. Yet, analysis on multi-level climate change governance shows that there do exist certain exceptions to such an approach: Firstly, NGOs are increasingly granted freedom in contributing to climate change policies. As an example, WWF China is implementing local campaigns to increase energy efficiency. Secondly, since 2007 provinces, local authorities and cities have initiated their own action plans and have thus been given autonomy in steering climate change policies. Thirdly, local authorities entered voluntary agreements with state-owned companies to decrease CO 2 emissions. Fourthly, trans-national networks, such as the Cities for Climate Protection Program (CCP) or the C40 Cities Climate Leadership Group have entered the arena in urban climate change governance and deploy 'soft' mechanisms through information- sharing, implementation and capacity-building as well as regulation. This presentation aims at introducing examples of such shifts in climate change governance and examining to what extent and why these shifts occur. Do they really lead to increased dispersion of authority? If yes, are these shifts the result of ‘hollowing out’ the state or part of an intentional strategy to even strengthen state capacity? This presentation is part of a PhD project on multi- level climate change governance in Japan, China and South Korea and the role of the central state concerning shifts in climate change governance. Paper proposed for 9th EastAsiaNet Research Workshop Aveiro, Portugal, 19-21 April 2012

Smart communities in Japan: testing next generation energy Yveline LECLER Lyons Institute of East Asian Studies

Sustainable development, energy saving, reduction of green gazes’ emissions, have influenced public policies in recent years in most advanced economies. It is also the case of Japan which counts on innovation and has placed green innovation at the core of its future growth dynamics as testified by the importance given to green technologies in the 4 th science and technology basic plan for example. The development of an environmental consciousness of the population, accelerated by the nuclear accident of Fukushima, also seems opening new perspectives in terms of acceptability of changes. To promote the experimentations of next generation energy systems in cities, the METI has launched in 2010 a program on smart communities which aims at testing, on a real scale, how could be coordinated environmentally friendly transportation systems, energy management that make efficient the use of renewable energy resources etc. Twenty cities applied to this program, but only 4 areas have been selected for such experimentations in Japan, while NEDO is engaging in international cooperation of such smart communities abroad. The paper proposed for the workshop will first present the policy framework launched by METI and the characteristics of each area selected before discussing the issue of smart communities implementation through a case study in the city of Kyoto.