Marginal Propensity to Save of Workers in Textile Products Industry (TPT)

The purpose of this study was to find out the desire to save the textile industry workers in Semarang Regency and the factors that influence it. This type of research is quantitative descriptive. The population in this study was Labor at 10 TPT Industries totaling 17,757. The sample was determined by 100 respondents using the Slovin formula, and was considered representative. The sampling technique used is simple random. Methods of collecting data used questionnaires, interviews and observations. Data analysis method used multiple linear regression. The variables used in this study are age, income, consumption and education level as independent variables. While the desire to save as a dependent variable. The results showed that the income variable and education level had a positive and not significant effect, the consumption variable had a negative and significant influence, the age variable had a negative and not significant effect on saving desires. Suggestions for optimizing the increase in savings, employees should be able to increase the desire to save by consuming as needed and reduce unnecessary expenses so that the amount of savings can be increased and the benefits of saving can be felt.


INTRODUCTION
Community savings, basically is part of the income received by the community that is not used for consumption or in other words, community savings are the difference between income and consumption (Sumastuti, 2009). The desire to save is only done when consumption and taxes are smaller than income. Given the important role of community savings in sustaining development financing, development economists have sought to find and formulate factors that can influence and encourage the level of community savings.
The existence of the textile industry is very important for the economy because it is one of the biggest contributors of foreign exchange and absorber of labor (Prasetyo, P. Eko, 2015). The textile industry is also a labor intensive industry (Prasetyo, P. Eko, 2015). The textile and textile products industry is more focused on a large number of workers or workers in the construction and operation and contributes a large contribution to employment (Apriliyanto, 2018).
The textile and textile products industry in Semarang Regency is a leading industry or a well-known industry in the area, where the number of workers in the industry is thus categorized as very large. With the large number of workers absorbed in the industry, is the phenomenon of the desire to save labor in the TPT industry high with the average income they get for a full month of work and their average monthly consumption they are still able or not to set aside money for them to save. Graph 1 describes consumption and savings expressed in units of ratios. Income is greater than consumption so that the excess income can be saved, this can be seen from the positive MPS curve. Thus in the opinion of Keynes can be interpreted, if income increases, but consumption decreases, then savings will increase. Conversely, if income increases and consumption increases, savings will decrease. The ratio used is a comparison of the value between the share of total income used for consumption and the part used for savings (Ubaidillah, 2019).

MPS
Based on the above graph, it can be seen that the average MPS of workers in Semarang Regency is 0.11. This number still shows a low tendency to save when compared with the research of Dini Septiani (2013) with the average MPS result of 0.20. The data was obtained from 30 respondents per 10 textile industry companies in Semarang Regency.
The low level of labor savings in the textile industry in Semarang Regency according to the results of a preliminary study with preliminary research on 10 textile industry companies shows that it does not mean that workers cannot afford to save but mostly save in real assets and prefer to save money at home rather than saving at a bank. The success of increasing business in mobilizing savings in the workforce itself so that it can be more absorbed in the banking institutions must have an effort that depends on understanding the characteristics of saving behavior in the workforce itself.
Generally they save 5-10% of the total income earned. Most of these workers do not save for a period of time each month, they save if the expenses made in the month are not too much or only if there is a residual income. Most of them save their money for reuse in the same month, only a few people actually save in the bank for a long time. Consumptive behavior is one of the things that motivates people to spend their income.
Based on the description above, this study is intended to find out the desire to save labor in the TPT industry is influenced by several factors such as age, income, consumption, education level or not. This research was conducted in the TPT Industrial area in Semarang Regency by conducting initial research. And from the explanation above, the research statement in this study can be formulated as follows : 1. Does the age of workforce in the textile industry and textile products in Semarang Regency affect the desire to save? 2. Does the income of the workforce of the textile industry and textile products in Semarang Regency affect the desire to save? 3. Does the labor consumption of the textile industry and textile products in Semarang Regency affect the desire to save? 4. Does the education level of the textile industry and textiles products of Semarang Regency affect the desire to save?

METHOD
This study uses a quantitative approach. with the aim of testing the predetermined hypothesis. (Sugiyono, 2017: 8). Research designs include: population and research samples, research variables, data collection methods, validity and reliability of research instruments, and data analysis. The population used in this study are workers who are still employees of the textile and textile products industry in Semarang Regency. In this study the respondents used were the number of workers of 10 companies with a total of 17,757 workers in the textile industry in Semarang district. Then the tolerance error value used is 10% with a confidence level of 90%. Samples obtained as follows: : 17.

Normality Test
According to Imam Ghozali (2013: 110) the purpose of the normality test is to find out whether each data is normally distributed or not. The normality test uses the Normal P-Plot graph, the Histogram and One Graph of the Kolmogorov-smirnov (K-S) sample with a probability value (significance level) of 0.05.
The results of the normality test using the IBM SPSS 23 Statistic. Based on picture 3 above, it is known that the histogram graph provides a distribution pattern that deviates to the right which means that the data is normally distributed.
Picture 3. Histogram Graph Source: Primary Data processed, 2018 Based on graph 4 above, it is known that the dependent variable of desire to save points on normal probability plot images tends to form and follow diagonal lines and spread around diagonal lines, so that it can be seen that the model is normally distributed. Based on the normality test by using one sample kolmogorov-smirnov test, the savings desire variable obtained Kolmogorov-Smirnov Z value and Asymp Sig (2-tailed) value of 0.056> 0.05, which means that Unstandardized Residual data are normally distributed.

Multicollinearity Test
Detecting the presence or absence of multicollinearity in the regression model is to look at tolerance values and Variance Inflation Factor (VIF). If the tolerance value is> 10% and the VIF value is <10, it can be concluded that there is no multicollinearity between the independent variables in the regression model.

Heteroscedasticity Test
Heteroscedasticity tests can be done by observing the scatterplot graph with a pattern of dots that spread above and below the Y axis.
Picture 5. Scatterplot Graph Source: Primary Data processed, 2018 Graph 5 above, shows that on the Scatterplot graph visible dots spread randomly and scattered both above and below the zero on the Y axis and the spread of data points does not form a wavy pattern widened and then narrows and widens again. It can be concluded that for the dependent variable and the independent variable there is no heteroscedasticity in this regression model.

Multiple Linear Regression Analysis
The following are the results of the regression calculation of Ln age, Ln income, Ln consumption, and Ln level of education towards Ln saving desire by using IBM SPSS Statistic 23:  (X1), income (X2), consumption (X3) and education level (X4) is 0, the value of saving desire or marginal propensity to save is 6.714. Age value (X1) is -0,389, which means that every age increases by 1%, the saving desire variable will decrease by 0.342%. The value of income or wages (X2) is 0.397 which means that every increase of 1% in the amount of income, the variable of saving desire will increase by 0.318%. The consumption value (X3) is -1.019 which means that every 1% increase in the amount of consumption expended will reduce the saving desire by 0.198%. The value of education level (X4) is 0.452, which means that every 1% increase in the level of education then the saving desire variable will increase by 0.423%.

T-test
The significance test of individual parameters (statistical test t) is used to determine whether in the regression model, the independent variables partially have a significant effect on the dependent variable. To determine the value of t statistics can be used a significant level of 5%. In accordance with the regression results and variables that affect saving desires can be accepted statistically then the following tests are carried out; If tcount < ttable then partially the independent variable has an effect on the dependent variable and α < 0,05 which means significant.
If tcount < ttable, then the independent variable partially does not affect the dependent variable and α > 0.05, which means it is not significant. Based on table t, then get t table = t (α / 2; n-k-1) = t (0,025; 95) = 1,985. Based on the results of processed data, the results obtained t-count of: 1. Age variable (X1) t count = -1.140 and α = 0.257. Then -1,140 <1,985 and 0.257> 0.05. This means that the H1 hypothesis which states that age has a positive and significant effect on saving desires is rejected. These results indicate that there is no influence of age on saving desires. 2. Income variable (X2) value of t count = 1.248 and α = 0.215 Then 1,248 <1,985 and 0.215> 0.05. This means that the H2 hypothesis which states income has a positive and significant effect on saving desires is rejected. These results indicate that there is no influence of age on saving desires. 3. Consumption variable (X3) value of t count = -4,952 and α = 0,000. Then -5,151> 1,985 and 0,000 <0.05. This means that the H3 hypothesis which states consumption has a negative and significant effect on the desire to save is accepted. 4. Variable level of education (X4) value of t count = 1.285 and α = 0.202. Then 1.067 <1.985 and 0.289> 0.05. This means that the H4 hypothesis which states the level of education has a positive and significant influence on saving desires is rejected.
These results indicate that there is no influence on the level of education on saving desires.

Adjusted R Square
Adjusted R square is used to measure the magnitude of the contribution of independent variables to the dependent variable. To measure how much the independent contribution (age, income, consumption and level of education) to the dependent variable (saving desire) is used adjusted R square. Based on the results of calculations using IBM Statistic SPSS 23 in Appendix 10, the results of the adjusted R Square calculation are 0.278 or 27.8%. This means that 27% of saving desires are caused by variables of age, wages, consumption and education level, while the remaining 73% are influenced by other variables not examined. According to Ghozali (2009) the determinant of efficiency for crossdata is relatively low due to the large variations between each observation.
Based on table 3 the results of the significant value of age variables are 0.257 which is greater than the value of 0.05. This means that age variables have no significant effect on saving desires. This result is not in accordance with the initial hypothesis which states that age has a positive and significant influence on saving desires. The argument is that they assume that as they grow older, their needs and desires increase. This can be seen from their consumptive lifestyle, judging from their income they tend to be used for everyday life, if there are new residues used to save. While the negative influence is that in this study it was found that the research subjects who were younger (age range 20-30 years) amounted to more than the study subjects who were older (age range 31-40 years). This result is supported by the study of Lea, Tarpy & Webley (1987) who said one of the factors that increases saving behavior is age. Where older individuals will increasingly improve their saving behavior in preparation for the future, in contrast to individuals who are still young. Young people tend to borrow more money than to save if they have not reached the peak of their income.
Based on table 3 the results of the significant value of the income variable is equal to 0.215 which is greater than the value of 0.05. This means that income variables have no significant effect on saving desires. This result is not in accordance with the initial hypothesis which states that age has a positive and significant influence on saving desires. From these results it can be interpreted that the income earned does not affect someone's desire to save. This is because the income they earn is prioritized by workers for consumption needs, so that if there is any leftover money they will just save. The results of the study are in line with the research of Ubaidillah (2019) which states that the non-influential causes of income towards saving desires because someone who has more income will be more able to be used to meet current needs or daily needs. someone, the higher the desire to consume.
Based on table 3 the results of the significant value of the consumption variable is 0.000 which is smaller than the value of 0.05. This means that consumption variables have a significant influence on saving desires. These results are in accordance with the initial hypothesis which states that age has a negative and significant influence on saving desires. This result is seen from the suitability of the sign, in accordance with the initial hypothesis and in line with the simple economic thinking of a consumer household as an economic actor, where they will set aside their income or save if their daily needs are met. The argument in this study is that someone's consumption expenditure is influenced by his income, and tends to fulfill or meet his daily needs so that they only save if there is a residual consumption from the income they receive. has an influence on MPS. According to Sabarina, a consumptive lifestyle and citizens must be able to allocate income in accordance with the needs, prioritizing needs rather than desires.
Based on table 3 the results of the significant value of the education level variable is equal to 0.289 which is greater than the value of 0.05. This means that education level variables have no significant effect on saving desires. This result is not in accordance with the initial hypothesis which states that the level of education has a positive and significant influence on the desire to save. The argument in this study is that because of their education which is majority of high school graduates and unmarried and do not have the desire to continue their education, they have the idea that the remaining income they save or the tube includes everything and is not grouped. Of the 100 respondents who filled out the questionnaire, only around 20% had special savings for education, most of them were married and the savings were specifically for their children. This is in accordance with the opinion of Gedela (2012) which reveals that one's level of education will not affect the desire to save individual households. Because, regardless of the level of education that someone has achieved, both high and low, someone's education will not affect his saving behavior, because he will still think about the future.

CONCLUSION
Based on data analysis and discussion of the results of the study, it can be concluded that, income and education level have a positive and not significant influence on saving desires, consumption has a negative and significant influence on saving desires. While age has a negative and not significant effect on the desire to save the textile industry labor force in Semarang Regency.
Based on the results of data analysis and discussion of the results of research, it can be suggested that optimizing the increase in labor savings in the textile industry can be done by consuming it as needed and reducing expenditures that are considered less important or reducing consumptive behavior by making monthly planning so that consumption is controlled and the amount of savings can be increased so that it is more intensive about the benefits of saving for all layers of labor and the benefits of saving can be felt. And for the next researcher it is suggested to analyze other variables that have not been done in this research. For example gender, wealth owned, occupation, tastes or desires, marital status, number of family dependents, family circumstances, education, just in case or frugality, and interest rates.