Chapter 7. Origin and Development of the High
7.10 Empirical Study of the Savings Rate
People aged 16–60 years at the starting point of the economic take-off period are called the first generation, and those aged below 16 years and born after the economic take-off are called the second generation.
Each resident is assumed to maintain his/her original consumption inertia, and the savings rate is left unchanged. One year later, the second generation will include people from 0 to 17 years old. Two years later, the second generation will extend to people from 0 to 18 years old. As the years pass, the number of people included in the second generation will consequently increase. For instance, in a family in which the father is 30 years older than his son, the father’s savings rate is 60% and that of his son is 30%. As a discontinuous function in this family, the savings rate suddenly drops by 30%. However, the savings rate of the entire society will not change suddenly but will drop year after year. The savings rate of people who are one year older than their son is 31% and that of people who are two years older than their son is 32%. In other words, the savings rate is a variable that decreases by 1% every year. Based on the supporting data, the range of the decrease in the savings rate every year can be easily obtained, such as 0.9% or 0.8%.39
An overlapping generation model can be constructed based on these assumptions, with the difference equation depicting the change law of the savings rate. People from the first generation maintain their original marginal consumption rate and increase their con- sumption synchronously after their disposable income increases. The consumption of those from the second generation increases as their income and marginal consumption rate increase year after year. Over time, the second generation accounts for a larger proportion in the society. Approximately 25–30 years later, the second generation will
39At present, we have not collected data about the household savings rate. We will conduct a research based on micro statistics.
230 From Trade Surplus to the Dispute over the Exchange Rate
comprise the majority of the population. In the entire process of economic development, income grows faster than consumption in the take-off stage, and the savings rate rises. In the drive to maturity stage, income and consumption increase at the same speed, and the savings rate is maintained at a high level. After entering into the age of high mass consumption, consumption grows faster than income in the entire society; the savings rate drops gradually and finally remains at a certain level. At this time, the country changes from observing the consumption pattern of the poor countries to that of the rich countries.
The best way to study the change rules of the savings rate of China is to refer to the experience of other economies in the world. In the past 50 years, only a few poor economies have become rich. Aside from the petroleum-exporting countries, the most typical examples are the four Asian economic giants. Among them, Hong Kong and Singapore are the urban economies that are incomparable with China. The changes in the savings rate in South Korea and Tai- wan have reference significance. Similar to China, Taiwan and South Korea have a Confucian cultural background. Specifically, Taiwan and Mainland China have the same race and culture and have many similarities in culture and economy. The economic development of Taiwan and South Korea has been ahead of that of Mainland China for a relatively short time. Therefore, research on the changes in the savings rate of Taiwan and South Korea has informative reference value for predicting the savings rate of Mainland China.
The regression model used to verify the savings rate changes in this study is as follows:
SRt=at+a1GPGDPt+a2DM2t+a3DCPIt+a4TOTt +a5APt+a6GPOPt+et,
where SRtis the savings rate, GPGDPtis the per capita GDP growth rate, DM2is the ratio of M2to the GDP, DCPItis the consumer price index change rate, TOTtis the trade change index, APtis the young population dependency ratio, and GPOPt is the population growth rate.
Origin and Development of the High Savings Rate 231
7.10.1. The change rules of the savings rate of Japan The savings rate of Japan was only 20% in 1955, and it reached its peak at 40.4% in 1970. Afterwards, it dropped year after year and was 29.7% in 1995 and 24.4% in 2002. In 2009, the proportion of the total domestic savings in the GDP of Japan dropped to the lowest level of 20.5%. When 25% is taken as a high savings rate, then Japan maintained it for more than 32 years. When 30% is taken as the high savings rate, then Japan maintained it for 25 years.
Before 1970, the savings rate of Japan increased with income (Fig. 7.3). During the economic take-off period (i.e., from 1955 to 1970), its total domestic savings increased by more than 20% every year, which is faster than the growth of labor remuneration (16%).
After the 1970s, Japan was listed as a high-income country, and the growth rate of its domestic savings began to drop sharply. During the 1970s and the 1980s, it dropped to 8.6%. After entering the 1990s, the average annual savings growth rate of Japan was negative, indi- cating the downtrend observed by its savings rate (Table 7.14).
In 1970 and 1990, the gap between the national savings of Japan and its labor remuneration widened further. However, no significant
Figure 7.3: Savings rate of Japan.
Source: Bureau of Statistics, Japan, http://www.stat.go.jp/english/data/chouki/
03.htm.
232 From Trade Surplus to the Dispute over the Exchange Rate
Table 7.14: Actual average annual growth rates of labor remuneration and total savings in Japan (%).
Actual average annual growth rate of labor remuneration
Actual average annual growth rate of total savings
1955–1959 13.7 26.7
1960–1964 17.6 12.8
1965–1969 17.1 25.7
1970–1974 20.7 8.6
1975–1979 9.9 8.9
1980–1984 5.9 5.9
1985–1989 5.8 7.0
1990–1994 3.6 −2.3
Source: Bureau of Statistics, Japan, http://www.stat.go.jp/english/data/chouki/03.
htm.
0 50 100 150 200 250 300
1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997
labor compensation Gross saving rate
Figure 7.4: Total savings and labor remuneration of Japan in 1955 and 1998 (calculated based on the constant price in 1990, trillion yen).
Source: Bureau of Statistics, Japan, http://www.stat.go.jp/english/data/chouki/
03.htm.
correlation exists between them based on statistics. In other words, the growth of the Japanese residents’ income cannot explain the change in the savings rate. This period can be referred to as the drive to maturity stage (Figure 7.4).
Origin and Development of the High Savings Rate 233
Since 1991, a significant positive correlation has been observed between the savings rate and the per capita income (Table 7.15).
During this period, the savings rate of Japan was negatively corre- lated with the per capita GDP growth rate. The savings rate declined with the increase in the per capita income, indicating that Japan had entered into the stage of high mass consumption after the 1990s (Table 7.16).
Table 7.15: Regression results of Japan in the first stage.
Sample Range: 1970–1990.
Variable Coefficient T-Statistic
GPGDP 0.04 1.28
DM2 −0.05 −1.29
DCPI −0.07 −1.19
TOT 0.00 0.112
AP −1.13∗∗∗ −5.29
GPOP −0.18 −0.16
C 97.41∗∗∗ 8.99
R-squared 0.948
Adjusted R-squared 0.925
Note: ** means that the coefficient satisfies the 5% level of significance, and *** means that the coefficient satisfies the 1% level of significance.
Table 7.16: Regression results of Japan in the second stage.
Sample Range: 1991–2009.
Variable Coefficient T-Statistic
C 151.63∗∗∗ 7.133
GPGDP 0.08∗∗ 2.99
DM2 0.002 0.11
DCPI −0.50∗∗ −2.18
TOT 0.00∗∗∗ −5.19
AP −1.57∗∗∗ −12.44
GPOP −5.87∗∗ −2.09
R-squared 0.98
Adjusted R-squared 0.97
Note: ** means that the coefficient satisfies the 5% level of significance, and *** means that the coefficient satisfies the 1% level of significance.
234 From Trade Surplus to the Dispute over the Exchange Rate
0 5 10 15 20 25 30 35 40 45
1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Figure 7.5: Savings rate of Taiwan.
Source: The data for 1958 and 2003 are from the World Bank Data,40 and the data for 2004 and 2010 are from the ADB.41
7.10.2. The change rules of the savings rate of Taiwan The savings rate of Taiwan was over 25% in 1970 and 30% in 1972.
In 1987, its savings rate reached its peak (38.5%). Since then, the savings rate began to drop: 30% in 1989 and below 25% in 2001.
When 25% is taken as the high savings rate, then Taiwan spent 16 years from stepping into the high savings rate to reaching the peak.
Taiwan maintained the high savings rate for nearly 30 years. When 30% is taken as the high savings rate, Taiwan maintained it for 24 years (Fig. 7.5).
The regression analysis on the data of Taiwan shows that its savings rate increased with income before 1970. In 1970 and 1989, R2 = 0.26; the variables above could not explain the change in the savings rate (Table 7.17). Nearly all the coefficients had no statisti- cal significance. In other words, during this period, the change in the savings rate had no relationship with the per capita national income.
Therefore, this period can be called the drive to maturity stage.
40See World Bank website at http://www.worldbank.org/data/.
41http://beta.adb.org/data/statistics.
Origin and Development of the High Savings Rate 235
Table 7.17: Regression results for Taiwan in the first stage.
Sample Range: 1970–1989
Variable Coefficient Prob.
GPGDP −6.69 0.38
DM2 0.10 0.52
DCPI 0.04 0.55
TOT −0.01 0.82
AP 2.53 0.23
GPOP −3.51 0.42
C −73.4 0.33
R-squared 0.496
AdjustedR-squared 0.263
Table 7.18: Regression results for Taiwan in the second stage.
Sample Range: 1990–2003
Variable Coefficient Prob.
C −38.1 0.344
GPGDP −9.84 0.023
DM2 0.176 0.024
DCPI 0.247 0.399
TOT −0.664 0.130
AP 2.138 0.052
GPOP 8.979 0.004
R-squared 0.951
Adjusted R-squared 0.909
In 1990 and 2003, an obvious correlation existed between the sav- ings rate and the per capita income. During this period, the savings rate of Taiwan was negatively correlated with the per capita GDP growth rate. The savings rate dropped with the increase in the per capita income, indicating that Taiwan had entered into the age of high mass consumption (Table 7.18).
236 From Trade Surplus to the Dispute over the Exchange Rate
0 5 10 15 20 25 30 35 40
1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012
Figure 7.6: Savings rate of South Korea.
Source: World Bank, WDI database.
7.10.3. The change rules of the savings rate of South Korea
The savings ratio of South Korea was up to 27.9% in 1977 and was greater than 30% for the first time in 1984. It reached its peak of 38.5% in 1988. Afterward, it declined year after year and dropped below 30% in 2008 (specifically, 29.8%). Taiwan maintained the high savings rate of over 30% for 26 years (Fig. 7.6).
The results of the regression analysis on the data of South Korea from 1984 to 2000 are listed in Table 7.19.
Based on the regression results, most of the coefficients were not statistically significant in 1984 and 2000. South Korea was in the drive to maturity stage and always maintained a high savings ratio.
During this stage, its savings rate was not at par with other economic variables, and the per capita income growth could not explain the change in the savings rate. However, when the economy maintained the high savings rate for nearly 30 years, South Korea entered into the age of high mass consumption, during which the savings rate began to decline year after year, and the savings ratio was negatively correlated with the per capita income. In other words, the savings ratio gradually reduced when the per capita income increased.
Origin and Development of the High Savings Rate 237
Table 7.19: Regression results for the savings rate of South Korea.
Sample Range: 1984–2000
Variable Coefficient T-Statistic
GPGDP −0.09 −0.79
GPOP −2.67 −1.05
DCPI −0.38∗∗ −3.24
DM2 0.21 1.59
TOT 0.00 2.08
AP −0.68∗ −2.49
C 77.39∗∗ 4.85
R-squared 0.77
Adjusted R-squared 0.63
Note: **means that the coefficient satisfies the 5% level of significance, and ***means that the coefficient satisfies the 1% level of significance.
7.10.4. The change rules of the savings rate of China The regression results for the data of China in 1979 and 2009 using the same model were the same as those of Japan in 1970 and 1990 and those of South Korea in 1984 and 2000.
In 1979 and 1986, the savings rate of China increased with the increase in per capita income. After 1987, the change in the sav- ings rate had no statistically significant correlation with other vari- ables. The per capita income and residents’ consumption growth rate increased rapidly during this stage. Regardless of how other variables changed, the savings rate of China remained at a relatively stable high level, indicating that China was still in the drive to maturity stage. Therefore, 1987 can be reasonably selected as the starting point of the drive to maturity stage.
According to the world economic outlook published by the IMF in September 2014, the actual average annual growth rate of the per capita GDP of China was expected to drop from 10.6% in the 11th Five-year Plan period (2006–2010) to 8.8% in the 12th Five-year Plan period (2011–2016). With the significant correlation between savings and income, according to the forecast of the IMF, the proportion of the total savings in the GDP of China will decline to less than 50%
after 2016.
238 From Trade Surplus to the Dispute over the Exchange Rate
China began to implement the reform and opening up policy in 1979. Since then, the per capita national income has been consis- tently growing, and the people’s living standards have improved sig- nificantly. In 1979 and 1987, the special consumption distribution mode under the planned economy, such as providing food and cloth- ing coupons, continued, and the market mechanism remained in the process of formation. Consumption and income grew simultaneously, but the latter grew faster than the former. Therefore, the savings rate of China was on the rise, and the first-generation effect occu- pied the mainstream. After 1987, the market mechanism gradually occupied the mainstream in the relationship between consumption and income. Thus, considering the national conditions, 1987 can be rationally selected as the starting point of the drive to maturity stage (Table 7.20).42
As only a few poor economies in the world become rich, sam- ples necessary to perform a quantitative analysis are insufficient.
The periods of maintaining the high savings will most likely vary because of the different cultural backgrounds in different economies.
Table 7.20: Regression analysis of the data of mainland China from 1987 to 2001.
Sample Range: 1987–2009
Variable Coefficient T-Statistic
GPGDP 0.41 2.29
DCPI 0.18∗∗ 3.36
DM2 −0.15 −1.76
TOT 0.00 0.77
AP −1.86∗∗ −4.03
GPOP 8.18 −0.98
C 121.73∗∗ 3.37
R-squared 0.92
Adjusted R-squared 0.89
Note: **means that the coefficient satisfies the 5%
level of significance.
42The taking of 1987 as the starting point for China to enter the drive to maturity stage can be further discussed.
Origin and Development of the High Savings Rate 239
0 2 4 6 8 10 12 14 16
0 10 20 30 40 50 60
1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012
Gross savings (% of GDP)
GDP per capita growth (annual %, left)
Figure 7.7: Changes in China’s total savings rate and the per capita GDP growth rate of China.
Source: World Economic Outlook, WEO, 2014, IMF.
However, one conclusion is very clear: maintaining the high savings rate for more than a generation is difficult. When the first generation is still the main body of the social labor force, income and consump- tion increase proportionally, and the savings rate fluctuates at a high level. Over time, an increasing number of people in the second gen- eration will begin to work. With the increasing proportion of the population of the second generation in the labor force (16–60 years old), the consumption behavior of the entire society will also change gradually. Finally, consumption is bound to grow faster than income, and the savings rate will have a down trend (Fig. 7.7).
Both South Korea and Taiwan maintained the high savings rate for nearly 25 years. Considering that government savings still occupy a large proportion in the economic transformation process, the high savings rate of Mainland China may last for 28–32 years. If 1987 is regarded as the starting point of the drive to maturity stage, by 2015 and 2020, the savings rate of Mainland China will significantly decline. As the dropping savings rate is likely to produce a significant effect on the stability of the financial system in China, the reform of the financial system should be promoted when the savings ratio is still at a high level.
240 From Trade Surplus to the Dispute over the Exchange Rate