Econometric models and estimations
14 To detect the nonlinear pattern o f household saving in the regressions, a variable o f age squared should be used However, in order to reduce a multicollinearity problem caused by using this variable, the variable o f
age squared is not included in the regression. The experiments conducted with the age squared showed that adding this variable does not influence much the result o f the OLS regression but affects significantly the 2SLS results. Since the study is concerned more with the age effect rather than the pattern, the variable o f age
household and the young household save less by 230 and 190 thousand dong per annum
respectively.
The effect o f education of the household head appeared to be very strong. In all the
regressions, the sign of the coefficient of education was found to be positive, indicating
that the household saving is likely to increase as the education level of the household head
increases. This result supports our hypothesis about the sign o f the coefficient of this factor.
An additional year o f schooling of the household head is found to make the household save
more, from 110 to 150 thousand dong per annum, ceteris paribus. The high level of
significance of the coefficient of the education variable, in addition to the income variable,
is interesting. It could be a reflection o f the fact that income is severely under-reported
among the educated (see Phan Dinh The 1998). But the most likely explanation of the
effect o f education on saving is the presence of a strong precautionary motive for saving.
With liquidity constraints and uncertainty prevalent, the more education the household
head has, the more he or she is capable of organising the household budget and saving to
face unexpected events (sickness, natural disasters). As a consequence, the findings seem
to suggest the presence o f a significant precautionary motive for household saving. The
question of the forms in which households choose to save is considered in more detail in
Chapter 7.
The marital status o f the household head has the expected effect on household saving. The
estimated coefficient of the marital status, MAR, was found to be statistically significant
marriage is likely to shift up the household saving level by 740 to 1130 thousand dong per
annum.
The negative sign of the coefficient of the dummy variable of gender, MALE, also
supported our expectation. The findings showed that households, where the head is male,
tend to save less than households that have a female as household head by a range of from
660 to 1120 thousand dong per annum. Although the magnitude is changed significantly
across models, the consistency of the negative sign of the coefficients supports our
argument about the labour division and the capability of a female household head to
manage and organise household work. The result seems to be reasonable because, as for
many other developing countries, mothers in Vietnam are observed to bear primary
responsibility for nurturing children. Faced with limited income sources o f the household,
the mother may attempt to secure more stable sources o f nurturing for her children. Thus
mothers would manage the household’s spending more carefully, and as a result, saving of
households would tend to be higher if household heads were mothers.
5.4 Conclusions
This chapter has presented an empirical investigation of the saving behaviour of
households in Vietnam using the data set o f the VLSS. Several saving models have been
applied for two types of household - the extended household with old people and the young
household without old people - to analyse the determinants of household saving. The
estimates of the saving determinants are relatively consistent across the models. Except for
empirical results support relatively well our hypotheses about the role, and the signs of the
coefficients o f the determinant variables influencing household saving.
The results suggest that the mps o f households in Vietnam is relatively high compared to
typical levels in developing countries. While the assets of households (non-human capital)
were found to have a negative impact, education (human capital) was found to have a
positive effect on household saving. The fact that the education of the household head
affects household saving positively suggests a possibly important role for precautionary
saving, although this is difficult to test directly at this stage with the available data.
Demographic factors were also found to have a significant effect. An increased number of
children was found to be important in decreasing household saving. This suggests that
while children may actively participate in household work, the cost o f raising them still
dominate the cost-benefit relationship of children and household saving. Marital status and
the gender o f the household head were found to be important in determining household
saving. Households whose heads are married were found to save more than the households
where the household head is single. Households that have female heads on average save
more than households with male heads.
The results did not support the hypothesis regarding the role of old dependent people for
the case o f the whole sample. However, the significant impact o f this demographic factor
on saving o f households became apparent when the sample was divided into urban and
rural regions, as shown in the next chapter. The break-down of the sample into different
areas and different geographic regions is needed because of differences in natural
Appendix 5.1
Definitions of the variables:
Y : A n n u a l to ta l in c o m e o f h o u s e h o ld
W : A s s e ts o f h o u s e h o ld , in c lu d in g h o u s in g , o th e r p h y s ic a l a n d fin a n c ia l a s s e ts ,
m in u s to ta l d e b ts
S A V IN G : S a v in g o f h o u s e h o ld d u rin g o n e y e a r, c a lc u la te d b y s u b tra c tin g e x p e n d itu r e
fro m th e in c o m e o f th e h o u s e h o ld
A G E : A g e o f h o u s e h o ld h e a d
E D U Y : E d u c a tio n o f h o u s e h o ld h e a d , c a lc u la te d b y y e a rs o f s c h o o lin g
N C H IL D : N u m b e r o f c h ild re n : n u m b e r o f h o u s e h o ld m e m b e r s a g e d u p to 17 y e a rs
N O L D : N u m b e r o f o ld p e o p le : n u m b e r o f h o u s e h o ld m e m b e r s o v e r 6 4 y e a rs
D l : Y o u n g d e p e n d e n c y ra tio , c a lc u la te d a s th e ra tio o f th e n u m b e r o f c h ild re n by th e n u m b e r o f p e o p le o f w o rk in g a g e
D 2 : O ld d e p e n d e n c y ra tio , c a lc u la te d a s th e ra tio o f th e n u m b e r o f o ld p e o p le by
th e n u m b e r o f p e o p le o f w o rk in g a g e
P I : Y o u n g d e p e n d e n t ra tio , c a lc u la te d a s th e ra tio o f th e n u m b e r o f c h ild re n
o v e r th e to ta l n u m b e r o f h o u s e h o ld m e m b e r s P 2: O ld d e p e n d e n t ra tio , c a lc u la te d a s th e ra tio o f th e n u m b e r o f o ld p e o p le by th e to ta l n u m b e r o f h o u s e h o ld m e m b e r s M A R : D u m m y fo r m a rr ia g e ; ta k e s th e v a lu e 1 i f th e h o u s e h o ld h e a d is m a rrie d , a n d 0 i f sin g le M A L E : D u m m y fo r g e n d e r o f h o u s e h o ld h e a d ; e q u a ls 1 i f th e h o u s e h o ld h e a d is m a le , a n d 0 i f fe m a le
Appendix 5.2
Summary statistics of variables
V a r ia b l e s S y m b o l s M e a n S t a n d a r d d e v i a t i o n I n c o m e ( t h o u s a n d d o n g ) Y 7 8 1 7 1 3 2 6 3 S a v i n g ( t h o u s a n d d o n g ) S 2 1 2 9 1 2 2 3 0 A s s e t s ( t h o u s a n d d o n g ) W 9 7 1 6 2 8 7 7 9 A g e o f h o u s e h o l d h e a d s A G E 4 5 . 3 9 1 4 .5 9 E d u c a t io n ( y e a r s ) E D U Y 6 .4 0 4 .3 5 N u m b e r o f c h i l d r e n N C H I L D 1 .9 6 1 .6 7 N u m b e r o f o ld p e o p l e N O L D 0 .2 7 0 .5 6 U r b a n a r e a s U R B A N 0 .2 0 0 .4 0 R e g i o n s ( N o r t h ) N O R T H 0 .5 3 0 .5 0 G e n d e r ( M a le ) M A L E 0 .7 3 0 .4 4 M a r r i a g e M A R 0 .8 1 0 .3 9 Y o u n g d e p e n d e n c y r a ti o ( d e f i n i t i o n 1) D 1 0 .9 2 0 .8 4 O ld d e p e n d e n c y r a tio ( d e f i n i t i o n 1) D 2 0 .1 2 0 .3 0 Y o u n g d e p e n d e n c y r a ti o ( d e f i n i t i o n 2 ) P I 0 .3 6 0 .2 5 O ld d e p e n d e n c y r a tio ( d e f i n i t i o n 2 ) P 2 0 .0 7 0 .1 8
Appendix 5.3
Saving, defined in the empirical study as the difference between total income and total expenditures, is
S = Tw] +1" W2 ~ nci ~ mc4 ~ Xm
or
where
Y = 7wj + /" w2