Homeownership
Another important factor which influences the household saving is housing (Lindqvist, 1981), housing plays a rather complex role, for it counts to both consumption and wealth (Yilmazer, 2002). Most empirical studies (at least those mentioned in this paper) considered housing as a component of consumption of durable goods to standardize and simplify the estimation, although durable goods in fact are seen as a significant component of saving (Burbidge and Davis, 1994).
Over the entire life, the purchase of a house may be the most considerable consumption for most households; hence it is reasonable to predict that the decision of house purchasing would have effect on household saving behavior (Horioka, 1988; Takayama and Kitamura, 1994). There is evidence found in Canada and Australia shows that younger households with relative high saving rates, the reason of this phenomenon could lay in paying off the mortgages (Burbidge and Davis, 1994; Harris et al. 2002). The situation of household accommodation has impact on saving as well. The results of analysis suggest that households with occupied-housing save more than those rent a house (Jappelli and Pagano, 1994). Moreover, the value of homeownership is found to have close correlation with the value of financial assets and tend to increase household saving (Börsch-Supan and Essig, 2003; Rha et al. 2006).
An interesting case is found in Japan: until the early 1980s, most households purchase a house with their own income, therefore, paying for the loans could be an important reason for saving; yet after 1986, homeownership has begun to be transferred in the form of bequest (Noguchi, 1990 cited in Takayama and Kitamura, 1994). Even if the households obtain the home from the older generation, their saving is still higher than tenant households (Takayama and Kitamura, 1994).
From the view of macroeconomic situation concerning housing, it could be indicated that the price of housing and the availability of housing finance are likely to do impact on household saving as well as portfolio choices.
Health Care Expenditure
Public health care is to different extent provided in different countries, still there is part of medical care cannot be covered. Therefore, the expenditure for this part becomes one reason for household to save (Jappelli and Pagano, 1994).
Burbidge and Davis (1994) found a hump-shaped age profile of median health expenditure with the Canadian data, which corresponds to the family size. Burbidge and Davis further suggest, although health expenditure might result from precautionary saving motive, yet it takes place along the life cycle and has no significant relation with saving for retirement. In Japan, where a relative complete public health care system is established, households still save for potential expenditure in this aspect; according to Takayama and Kitamura (1994), this could be explained by the anxieties about the unexpected future medical service, therefore, households are driven to save before retirement and even afterwards at older ages. On the contrary, in the case of Italy, the expenditure on health care and effect of health risk could not manage to cause impact on precautionary saving (Jappelli and Pagano, 1994).
Moreover, Le et al. (2010) proposed out of their study that people in poor health status tend to save more than those in average health status, which is also in accordance with the precautionary motive.
- Education and Employment Status
Education
In many empirical studies, the factor of education level appears to be influential on saving behavior. However, like other socio-demographic factors, the effect of education is not simple to be concluded, either.
Lindqvist (1981) argues in his path analysis of saving behavior that the household’s education level has a considerable effect on saving behavior and would directly or indirectly impact on saving; also, the effect of education level could be found in the decision of the types of saving. On the basis of a regression analysis and thereafter hypothetical causal model and path analysis, Lindqvist indicates considerable effects of household income and education level.
One possible explanation could be the following: it is likely to believe that educational attainment could be treated as a strong determinant of future income, thus, because of this expectation of higher future earnings, individuals with higher education levels are to save relative less accordingly (Rha et al., 2006).
Yet, the real situation has not followed the presumption. Solomon (1975) estimates the effect educational attainment of household heads and indicates a positive impact on the likelihood of households to save. He argues the level of education is more relevant to the time preference instead of income, and people with higher education levels tend to pay more attention on the future.
In the study of Bulgaria, Hungary and Poland, Denizer, Wolf and Ying (2002) use the data of the period between 1993 and 1995 and implied the influence of education level results from its role for the future income. Evidence found in New Zealand (Le et al. 2010) also shows a positive relation between educational qualification and saving. Beckmann et al. (2013) added more evidence to the effect of high education level by pointing out that it arouses higher propensity and diversity to save; however, the expectation of higher future income is not the only explanation, also, this presents implications for income effect and points out impact of higher financial literacy.
Nevertheless, the findings of Rha et al. (2006) are somewhat different. They suggest that education does influence saving rates, for example, saving rate is increased by several years of schooling, while household heads with a college or even higher education levels tend to have lower saving propensity than those with high school education. Rha et al. attempted to explain the phenomenon in two aspects: the effect of education could not only exist through the association with permanent income, but it could also concern a lower rate of time preference and possibility to more diversity of investment.
Apart from those connection found between education and saving behavior, Jappelli and Pagano (1994) discovers no significant effect on household saving resulted from education attainment.
Besides the general education, some studies concentrates on the influence of financial literacy which is more specific to saving behavior, whose importance is found particularly on planning savings and portfolio decisions (Lusardi and Mitchell, 2007).
In the study of Gale, Harris and Levine (2012), financial literacy is defined as “ the ability to make informed judgments and effective decisions regarding the use and management of money and wealth, as well as the ability and discipline to implement intended or desired saving behavior” (P. 39f.). It is suggested that lower level of financial literacy is to some extent related to the lower level of education (Bucks and Pence 2006), which may lead to poorer financial choices in wealth accumulation (Gale et al. 2012). Therefore, improving financial literacy may contribute in rising personal saving and other financial security for retirement (Lusardi 2008).
Bernheim and Garrett (2003) conduct a study on effect of the financial education for company employees. Results indicate that offer financial education such as retirement seminars in firms could improve the household saving of employees significantly. Similarly, Lusardi (2002) analyzes the data from the University of Michigan’s Health and Retirement, in which increment in wealth due to improvement in financial education is found. Bernheim, Garrett and Maki (2001) find that people who attended high school with mandating financial education are likely to save more in their later life than those without.
Furthermore, Gale et al. (2012) also suggest, besides the relative low level education, low levels of financial literacy are often found in disadvantaged groups with poor financial choices, which could arise uncertainty and insecurity about their economic situation; thus it could contribute to higher saving and better wealth accumulation to improve the level of financial literacy.
As we could see from the empirical studies, the role of educational attainment is often related to the perception of longevity, self-perception as a saver as well as their effect of attitudes towards saving (Ando et al., 1992). In addition, improvement of education contributes to the understanding of information and making more proper financial decisions. Therefore, it is believed that measures on education would be helpful to promote the household saving, for example, certain educational programs could be launched for different groups of individuals (Lusard and Mitchell 2006); also, courses regarding financial knowledge could be added for the general school education.
Employment
According to the life-cycle theory, people earn income mainly during the working age before the retirement; therefore the status of working is expected to have influence on the saving behavior. However, no definitive conclusion can be drawn from the current empirical studies.
Jappelli and Pagano (1994) suggest, due to the income uncertainty and fluctuations, household head with relative unstable or riskier work status are likely to save more in order to prevent possible insecurity, since employment status may reflect earnings. In this way, the higher saving is to anticipate in households with self-employed head than the case of firm employees earning wage or salary. Yet, the Italian data did not confirm the prediction. This result acts in accordance with the study of Skinner (1988), no correlation between saving and employment emerges with other variables controlled. Jappelli and Pagano try to interpret the situation from the psychological aspect: probably less averseness for people with riskier occupation.
In the analysis in Japan by Takayama and Kitamura (1994), it is found saving of company employees higher than that of retirees or unemployed; the highest saving turns out to be observed in the group of self-employed households which at the same time have the highest income level. Households make much more saving when the head of the household is working than the time the head has no work. The exception of increasing positive saving at the time around retirement is quite possibly resulted from the generous lump-sum retirement severance payments.
Evidence in both Australia and other OECD countries shows that households have lower or even no saving when the household head is in unemployment. On this point, the reason could lay more in the impact of “ lower income” due to employment status rather than the “ precautionary saving motive” (Callen and Thimann 1997; Harris et al, 2002).
With the cross-country data in CESEE countries, the similar pattern is found by Beckmann et al. (2013): the self-employed have the highest rates, followed by the employed, and the non-working (retired and unemployed) save the least in the comparison; but they explain rather with precautionary motives. Nevertheless, Denizer et al. (2002) indicate no significant effect of occupation status on saving behavior with the data of Bulgaria, Hungary and Poland. Besides, Obben and Waayer (2011) suggest that unemployment would cause significant precautionary motive, the possible situation could be described as following: household saving is probably driven to rise in risk-averse households, at the time whose one or more members just turn unemployed.
- Psychological Influence
Several decades after the popularity of wealth theories explaining the saving behavior, it is stressed by Katona (1975) in his book of Psychological Economics that psychological factors play a significant role in the process of economic decision making.
The significance of psychological researches for the investigation of saving can be grouped into the following two aspects: on one hand, psychological methods and techniques make it possible to collect subjective and objective data which is difficult to get through pure economic ways; on the other hand, they provide theoretical and empirical support when psychological concepts are used to explain and analyze saving behavior (Garcia, Barros and Silvestre, 2011).
Psychological factors such as saving attitude, expectation, risk aversion and other economy-related attitudes could also impact on household behavior (Lindqvist, 1981). As mentioned in the early part of this paper, there are only a few studies of saving behavior taking psychological influence into account. Attitudes towards economic activities are considered to be strongly influenced and banded with education from early life and life style (Ölander and Seipel 1970; Lindqvist 1981), which takes deep root in awareness and therefore could hardly be changed during a short period. Thus, psychological factor influence rather long-run economic behaviors, and it would be helpful to understand more about this economic phenomenon by adding psychological and sociological considerations (Börsch-Supan and Essig, 2003).
Actually, the way of undertaking survey on saving habits, attitudes and motives is the one common method for investigation of saving behavior, in which those involved psychological variables have already been mentioned by economists (e. g. saving motive is very frequently discussed in the explanation of household saving) in their studies, because of the unclear complexity of the human behavior (Garcia et al. 2011).
The project of Lindqvist is one of the important empirical studies which pay extra attention to psychological effect, trying to find out the causal influential factors on saving behavior. Attitude and expectation are the main psychological variables to be studied on. For example, in his interviews, respondents are required to do prediction on the trend of living standard, financial development and saving during the next six month in order to measure household’s attitude like optimism or pessimism. And it is found that a positive effect exists between saving-related attitudes and total saving. Also, larger bank savings improve the positive household’s feeling of financial status. However, results from the hypothetical causal model suggest not very significant effects of psychological measures.
For some precautionary theories begin to pay attention to subjective factors and indicate uncertainty or pessimistic attitudes towards future are likely to increase household savings (Juster and Taylor 1975; Carroll and Samwick 1995; Browning and Luscardi 1996), Harris et al. with the Australian data confirm the pattern: possible increment appears, if the individuals anticipate their economic situation may be worse than the current status. Moreover, Beckmann et al. (2013) suggest risk averseness is also to contribute to increase saving in the study of CESEE countries, which could also be explained by the precautionary savings.
Apart from the psychological factors mentioned by Linqvist, self-control is also a substantial component for the interpretation of saving behavior (Warneryd, 1989). Lunt (1996) also emphasizes the role of self-control due to the possible risks in the economic environment which owns higher level of materialism and provides more chances.
Adding question with particular consideration on psychological factors, Rha, Montalto and Hanna (2006) use the data from the Survey of Consumer Finances (SCF) to analyze the effect self-control mechanisms on household saving behavior in the US. With the implications of both life cycle theory and behavioral life cycle theory, several categories of variables are estimated, such as household financial variables, demographic variables, and psychological variables.
“ Specific saving goals”, “ foreseeing future expenses”, and “ saving rules” (P. 6f.) are used as behavioral life cycle (BLC) variables to test the effects of self-control mechanisms. Results show higher probability found in the households saving with some rules; also, households with certain saving goals are more likely to save; whereas foreseeing future expenditure helps only a little on saving. These mentioned outcomes provide confirmation of a positive relation between household saving behavior and self-control. Moreover, both optimistic expectation on growth of future income and the longer planning horizon have positive impact on the likelihood of household saving, whereas the risk averseness has negative effect which is different from some other studies, for example the evidence found in SECEE data (Beckmann et al. 2013).
More recently, Garcia, Barros and Silvestre (2011) conduct a random face-to-face interview in Portugal in order to collect adequate data to investigate the determinants of household saving behavior. With the central concept that saving decision is a combination of attitude and behavior, they set up seven hypotheses:
H1: “ The perception of the importance of saving influences positively and significantly the attitude to savings.”
H2: “ The perception of the replacement rate has a negative and significant effect on the attitude to saving.”
H3: “ The perception of longevity influences significantly and positively the attitude to savings.”
H4: “ The age has a positive and significant influence on the savings attitude and induces saving behavior.”
H5: “ The income level affects negatively and significantly the saving attitude, but has a significant and positive effect on the individual’s saving behavior.”
H6: “ The family size affects negatively and significantly the saving attitude, but has a significant and positive effect on the individual’s saving behavior.”
H7: “ The attitude towards savings has a significant and positive effect on the individual’s saving behavior.” (P. 228f.)
The hypothetical relations among the variables are shown with a structural equation model in Figure A3.
The results of regression analysis and correlation analysis imply a positive and significant effect from perception of the importance of saving and longevity, meanwhile a negative effect from that of replacement on saving attitude, which testifies the validity of H1, H2 and H3. As for H4, it is found that age affects saving attitude positively and significantly, but no significant effect emerges on saving behavior. The result for H5 suggests ambivalent impact from income on saving attitude and saving behavior: negative but insignificant influence on the attitude, whereas positive and significant influence on the behavior. Meanwhile, the no hypothetical effect due to family size is found to validate H6. Moreover, as H7 predicts, attitude towards saving influences saving behavior positively and significantly.
Age has a positive and significant effect on the savings attitude.
These findings provide a more faceted explanation to the saving behavior and could be a substantial addition to the empirical results within the “ pure” economic theoretical frame work.
