Extant research reveals the adaption and implementation of pension schemes in jurisdictions such as the United States of America date back to the 1600s, specifically, 1636. However, considerations for and implementation of same in Ghana came into effect during the 1940s under the then Gold Coast, the country’s name prior to the attainment of independence in 1957. The British Colonial Ordinances, Pension Ordinance Number 42, commonly called CAP 30 remained the dominant pension scheme in Ghana in prior years.
However, after independence in 1957, Ghana’s first President, the late Osagyefo Dr. Kwame Nkrumah, instituted measures to ensure the review of existing pension schemes and establishment of a comprehensive pension scheme that would meet the retirement needs of workers during retirement. In 1972, the Social Security and National Insurance Trust (SSNIT) scheme was introduced in Ghana as mandatory and universal pension scheme for all workers while workers enrolled earlier under the CAP 30 scheme were allowed to remain on that plan.
However, the SSNIT pension scheme did not address the concerns of retired workers; workers who retired under the CAP 30 were paid more retirement income than those enrolled under the SSNIT pension scheme. A Presidential Commission on Pensions was established by the Government of Ghana in July 2004 to commence major reforms in Ghana’s Pension System.
The Commission’s work included inter alia, examination of existing pension arrangements; and making appropriate recommendations for the implementation of sustainable pension scheme or schemes that would assure retirement income security for Ghanaian employees, with special emphasis on the public sector.
The Presidential Commission’s final report was submitted in March 2006. Almost all the recommendations contained in the final report were accepted by the government; and a Government White Paper (W.P. No. 1/2006) was issued to that effect in July 2006 (NPRA, 2019a and c).
The Presidential Commission recommended the creation of a novel Three-Tier Pension System for Ghana. The three-tier pension schemes would be funded directly by employers and employees; and these schemes were expected to replace existing parallel pension schemes in Ghana.
The novel three-tier contributory pension scheme comprises two mandatory schemes and a voluntary scheme. The mandatory pension schemes include the Tier One SSNIT benefits scheme and Tier Two Occupational or Work-based pension scheme. The voluntary scheme includes the Provident Fund and Personal Pension schemes. Based on its findings, the Presidential Commission recommended the need for the Government of Ghana to decentralise management of pensions for the public sector; and restructure administrative systems for CAP 30; while it remains operable in the country.
Further, the Commission recommended the need to restructure SSNIT, including an overhaul of its administrative, management, and governance structures; review existing SSNIT Laws; ensure pension coverage for the informal sector of the Ghanaian economy; phase out CAP 30 due to non-sustainability; establish a National Pensions Regulatory Authority (NPRA) to regulate the country’s public and private pension schemes; and the need for the NPRA to ensure unification of all pension schemes in Ghana within five (5) years of coming into effect of the new pension scheme.
Following recommendations of the Presidential Commission, an 8-member Implementation Committee was appointed by the President in October 2006. The mandate of the Committee included the implementation of accepted recommendations on pensions as contained in the Government White Paper No. W.P. No.1/2006 of July 2006; and drafting of a new pension reform bill that would provide legal backing for all the recommendations accepted in the Presidential Commission’s report on pensions (NPRA, 2019a, and b).
National Pensions Act of 2008, Act 766
The National Pensions Act of 2008, Act 766, was passed into law on 4^{th} December, 2008 to give legal backing to the Contributory Three-Tier Pension Scheme and National Pensions Regulatory Authority in Ghana. Act 766 comprises 221 Sections and a Schedule. The Sections in Act 766 are categorised into four parts, parts one through four. Establishment of the Social Security and National Insurance Trust (SSNIT) as a corporate body with perpetual succession and common seal; and with the ability to sue and be sued are stipulated in Section 32 and its related sub-sections. Further, this section states SSNIT’s ability to acquire immovable and movable properties for effective performance of its operations (ILO, 2008; IOP, 2008).
National Pensions (Amendment) Act of 2014, Act 883
Amendment of Act 766 was deemed necessary to help address issues related to the transfer of contributions; reduce the age at which an employee is exempted from the tier one pension scheme; and to address other significant pension-related issues. Thus, some Sections in the National Pensions Act of 2008, Act 766, were amended to improve the content and implementation of pension schemes in Ghana; and to meet international standards in pension administration.
The foregoing affirms the National Pensions (Amendment) Act of 2014, Act 883, was promulgated in furtherance of the National Pensions Act of 2008, Act 766. Specifically, the following Sections of Act 766 were amended: 55, 60, 64, 77, 81, 83, 91, 94, 100, 108, 109, 120, and 147. The following Section was inserted: 73A (NPRA, 2019b).
Objective
The main objective of this work was to evaluate workers’ contributions to pension schemes operable in Ghana; and the relative effect of those contributions on the nation’s total gross domestic product (GDP). Stated in different terms, the writer was interested in examining the economic significance of pension fund assets to Ghana’s overall gross domestic product.
Methodology
The quantitative approach to scientific inquiry was applied to the work. Specifically, a cross-sectional design, an example of survey design, was adapted and used in the study. This allowed the writer to gather relevant research data over a specific period of time (Ashley, Takyi & Obeng, 2016; Creswell, 2009; Frankfort-Nachmias and Nachmias, 2008).
Data required for the write-up were obtained mainly from secondary sources including text books, peer-reviewed articles published in journals, research papers, newspaper publications; Google Search Engine, financial websites such as Tradingeconomics.com; and electronic databases of the Bank of Ghana (BoG), Ghana Statistical Service (GSS), National Pensions Regulatory Authority, Social Security and National Insurance Trust; the World Bank and World Economic Outlook, among others. Annual data on the World’s total gross domestic products (GDPs); data on pension fund assets values for selected economies; and pension fund assets-to-Ghana’s GDP ratios during the period were computed and used in the research.
Analytical Tools
Regression models and descriptive statistics were used to describe the research variables; and to evaluate their behaviour over the stated time frame within the Ghanaian economy. Measures such as the range, minimum, maximum, skewness, and standard deviation were employed to describe the extent of dispersion about the central tendency (Ashley et al., 2016; Creswell, 2009; Frankfort-Nachmias & Nachmias, 2008). These measures were used to describe trends in pension fund assets for selected economies over a given time period.
Research Variables
The independent research variable was pension fund assets while the dependent research variable was the level of overall performance of the Ghanaian economy, represented by Ghana’s GDP.
Regression Model
Regression statistical model was adapted to measure the effect and level of interaction of pension fund assets on Ghana’s total GDP. The Microsoft Excel analytical software was used in the research. Diagrammes and tables were derived from Microsoft Excel to explain the research data; and to determine fitness of the regression model for the research.
Research Hypothesis
The study tested causal relationship between pension fund assets and gross domestic product, using the following null and research or alternative hypotheses:
Ho: µ1 = µ2; this implies pension fund assets do not have strong effect on Ghana’s GDP
H1: µ1 ≠ µ2; this implies pension fund assets have strong influence on Ghana’s GDP
Descriptive Statistics
Figure 3 presents a statistical description of the pension fund assets values for each of the fourteen global economies included in the analysis. The skewness of the distribution in Figure 3 is 3.654; while the sample variance is 4.386. Data in Figure 3 depict respective standard error and Kurtosis values of 1.770 and 13.521.
The standard error value in Figure 3 tells us the extent to which the coefficients are significantly different from zero; whereas the Kurtosis value affirms the extent to which the tails of the distribution in Figure 3 differ from the tails of a normal distribution.
Figure 3: Descriptive Pension Fund Assets Values for Selected Economies
DESCRIPTIVE STATISTICS FOR PENSION FUND ASSETS VALUES – 2016 | |
Mean | 2.38617E+12 |
Standard Error | 1.77001E+12 |
Median | 2.34977E+11 |
Mode | #N/A |
Standard Deviation | 6.62278E+12 |
Sample Variance | 4.38612E+25 |
Kurtosis | 13.52050609 |
Skewness | 3.654401644 |
Range | 2.5229E+13 |
Minimum | 8821566328 |
Maximum | 2.52379E+13 |
Sum | 3.34063E+13 |
Count | 14 |
It is worth reiterating a significant number of the pension fund assets values presented by the fourteen global economies showed steady increases over the period.
The minimum value in the figure is 8821566328, representing Kenya’s pension fund assets value for 2016 ($8.822 billion); while the maximum value (2.52379) is representative of the total pension fund assets value for the United States during 2016 ($25.238 trillion). The range explains the difference between the maximum and minimum values for the distribution. Value for the range (2.5229) in Figure 3 explains the substantial difference ($25.229 trillion) between the respective pension fund assets values for the United States ($25.238 trillion) and Kenya ($8.822 billion).
RESULTS
Statistics in column 4 in Figure 4 depict the computed pension fund assets (PFAs) ratios for Ghana from 2012 through 2017. Consistent with the World Bank standard, these ratios were expressed in percentage terms. We observe steady increase in Ghana’s pension fund assets to GDP ratios from 2013 through 2016.
Figure 4: Computed Pension Fund Assets to GDP Ratios for Ghana
YEAR | GDP IN BILLIONS (GH¢) | TOTAL PFAs (GH¢) | PFAs TO GDP RATIO (%) | |||
2012* | 67,942,000,000 | 4,824,412,125.84 | 7.10078 | |||
2013 | 123,650,000,000 | 6,084,262,016.88 | 4.92055 | |||
2014 | 155,433,000,000 | 9,526,828,054.82 | 6.12922 | |||
2015 | 180,399,000,000 | 12,627,748,464.24 | 6.9999 | |||
2016** | 215,077,000,000 | 15,793,220,000.00 | 7.34305 | |||
2017*** | 256,671,000,000 | 17,345,700,000 | 6.75795 | |||
*GDP value for 2012 not rebased | ||||||
**The total PFAs value for 2016 equals SSNIT investment portfolio for that year multiplied | ||||||
by 2 = 7,896,610,000 x 2 = 15,793,220,000 | ||||||
***The total PFAs value for 2017 equals SSNIT investment portfolio for 2009 and 2011 | ||||||
multiplied by 3 = 5,751,900,000 x 3 = 17,345,700,000 | ||||||
Test of Hypothesis
The alternative hypothesis formulated for this research sought to test whether or not pension fund assets have strong influence on Ghana’s GDP. Values in columns 2 and 3 in Figure 4 formed the basis of the analysis in this section. A regression analysis was conducted to test the relationship between the independent variable (pension fund assets) and the outcome or dependent variable (Ghana’s GDP); and to test the fitness of the regression model for the research. Outputs from the statistical analysis on the research hypothesis are presented in the following section.
Model Summary
Summary constitutes an integral part of a regression model. Tables 3 through 5 present the regression analysis outputs. An overall description of the regression model is presented in Table 3. Values for R (0.974298839), R² (0.949258228) and adjusted R² (0.936572785) are displayed in Table 3. Value for the multiple correlation coefficients between the independent variable (pension fund assets) and the dependent variable (Ghana’s GDP) is presented in the R row.
Table 3: Summary Output
Regression Statistics | |
Multiple R | 0.974298839 |
R Square | 0.949258228 |
Adjusted R Square | 0.936572785 |
Standard Error | 16840298674 |
Observations | 6 |
The R² value (0.949258228) in Table 3 reveals the extent to which variability in the dependent variable (Ghana’s GDP) is accounted for by the independent variable (pension fund assets). The R² value implies pension fund assets account for about 94.93% (0.949258 x 100% = 94.93%) of the variation in Ghana’s GDP. The results suggest only about 5.07% (100% – 94.93% = 5.07%) of the outcome is explained by external random factors. The number of values observed in the analysis is 6; as shown in Table 3.
One of the measures that determine the generalisability of the regression model is the adjusted R². Generally, an ideal adjusted R² value is closer to zero or the R² value. The adjusted R² value (0.936572785) in Table 3 is not significantly different from the observed value of R² (0.949258228); implying the cross-validity of this regression model is good; it suggests the model may accurately predict the same dependent variable from the given independent variable in a different group of participants (Field, 2009, p. 221). The R² significance was computed using an F-ratio to cross-validate the F-value (74.8305147) in Table 4. The ideal F-ratio formula for measuring R² significance that was applied to this work is:
F = (N – k – 1) R²
k (1 – R²)
Where:
R² = Unadjusted value
N = Number of cases or participants in the study
k = Number of independent variables in the regression model
Value for the F-ratio was determined as follows:
F = (6 – 1 – 1) 0.949258228
1 (1 – 0.949258228)
= 3.797032912
0.050741772
= 74.8305146
Our computations revealed the change in the amount of variance that can be explained gives rise to an F-ratio of 74.8305146, which is equivalent to the F-value (74.8305147) in Table 4. This F-ratio shows a significant value (p = 0.001, p < 0.01); as illustrated in tables 4 and 5.
ANOVA
The ANOVA helps to determine whether or not regression analysis provides better and significant prediction of the outcome than the mean. Data in Table 4 show degree of freedom (between) of 1 (2 – 1 = 1); degrees of freedom (within) of 4 (6 – 2 = 4); total degrees of freedom (df) of 5 (6 – 1 = 6); and an F-value of 74.8305147.
Table 4: ANOVA
df | SS | MS | F | Significance F | |
Regression | 1 | 2.12216E+22 | 2.122E+22 | 74.8305147 | 0.000982336 |
Residual | 4 | 1.13438E+21 | 2.836E+20 | ||
Total | 5 | 2.2356E+22 |
Data in Table 4 depict the model sum of squares (SSM) value, represented by Regression; the residual sum of squares (SSR) value, represented by Residual; the total sum of squares (SST) value, represented by Total; and the degrees of freedom (df) for each group of squares. The degree of freedom for the SSM is 1, comprising the one independent variable (pension fund assets). The sum of squares divided by the degrees of freedom gives us the mean squares (MS). That is, 2.12216E+22 ÷ 1 = 2.12216E+22. Similarly, 1.13438E+21 ÷ 4 = 2.836E+20.
Model Parameters
Table 5 presents results on parameters of the regression model. Data in Table 5 show the coefficients, standard error, test statistic (t-Stat, t-test), significance; and confidence intervals for the coefficients. The coefficients in Table 5 explain the contribution of the independent variable (pension fund assets) to the regression model. Generally, a positive coefficient connotes a positive relationship between the independent variable and the dependent variable; a negative value symbolises a negative relationship between the two variables.
Results from the statistical analysis and presented in Table 5 show a positive coefficient value (12.759). This affirms the existence of a positive relationship between pension fund assets and Ghana’s GDP (dependent variable). Further, the statistical outputs in Tables 4 and 5 suggest pension fund assets have significant influence on Ghana’s GDP (p = 0.001, p < 0.01).
Table 5: Model Parameters
Coefficients | Standard Error | t Stat | P-value | Lower 95% | Upper 95% | |
Intercept | 25748678816 | 17666857571 | 1.4574566 | 0.21872619 | -23302381409 | 7.48E+10 |
X Variable 1 | 12.75909715 | 1.474961139 | 8.6504633 | 0.00098234 | 8.663948516 | 16.8542458 |
A normal probability plot of the relationship between pension fund assets and Ghana’s total GDP is presented in Figure 5. The figure depicts a steady rise in comparative values over the research period. Specifically, Figure 5 depicts steady increase in values from the 8.33^{rd} percentile (during 2012) through the 41.67^{th} percentile (during 2014) to the 91.67^{th} percentile (during 2017).
Figure 5: Normal Probability Plot
The magnitude of the t-test (p = 0.001, p < 0.01) in Table 5 tells us the independent variable (pension fund assets) has a strong impact on the dependent variable (Ghana’s GDP). A standard error is identified with the coefficients in the table. The standard error indicates the extent to which the coefficients would vary in different research samples (Field, 2009). The probability that a parameter would fall between a pair of values around the mean is measured by the confidence interval. Stated in other terms, confidence interval values affirm the extent or level of uncertainty; or certainty in a method of sampling (Hayes, 2021). The respective upper and lower 95% values for the Intercept (74799739041.6818 and -23302381409) and X Variable 1 (16.8542457867549 and 8.663948516) are displayed in Table 5.
Test of Assumptions
Statistical tests were conducted to determine linearity of the relationship between the independent variable (pension fund assets) and the dependent variable (Ghana’s GDP); and to measure the variance in residual values. The statistical outputs are presented in Figure 6 and Table 6. The scatter plots in Figure 6 are on a straight line. This affirms relationship between the independent variable and dependent variable is linear; it implies the chosen model fits the analysis.
Figure 6: Linear Relationship between X and Y Variables
The residual values in Table 6 allow us to test homoscedasticity of the model. That is, these data facilitate our determination of whether or not the residual values at each level of the independent variable depict constant variance. Residuals in Table 6 show constant variance values; this implies the assumption of homoscedasticity is met.
Moreover, data in Figure 4 indicate relationship between the X and Y variables were measured at the interval level and beyond; while variability of the dependent variable (GDP) was not constrained. The foregoing analysis indicates most of the assumptions have been met; this renders the regression model fit and appropriate for the research.
Table 6: Predicted Y Values and Residual Values for Variable X
Predicted Y | Residuals | Standard Residuals |
87303821828 | -19361821828 | -1.285438893 |
1.03378E+11 | 20271631016 | 1.34584148 |
1.47302E+11 | 8130596488 | 0.539793468 |
1.86867E+11 | -6468348274 | -0.429436162 |
2.27256E+11 | -12178907129 | -0.808562388 |
2.47064E+11 | 9606849726 | 0.637802495 |
Report on P -Value and Confidence Interval
Table 5 depicts P value of 0.00098234 and positive coefficient value of 12.75909715. These values are significant at Alpha level ɑ = 0.05. The table further shows a confidence interval of 8.663948516 and 16.8542458. The Alpha level, a priori, for this study is ɑ = 0.05. The implication is there is 5 per cent probability that we would be wrong.
Further, there is 5 per cent likelihood the population mean would not fall within the interval (Bowerman & O’Connell, 1990; Frankfort-Nachmias and Nachmias, 2008). However, we are 95% certain our conclusions would be right. Again, the Microsoft Excel output in Table 4 shows degree of freedom (between) of 1 (2 – 1 = 1); degrees of freedom (within) of 4 (6 – 2 = 4); total degrees of freedom (df) of 5 (6 – 1 = 5); and an F-ratio of 74.8305147. These values could be interpreted as:
F (1, 4) = 74.8305147, p < 0.01, two-tailed.
Interpretation and Rejection of Null Hypothesis
The foregoing statistical results and related analysis indicate pension fund assets have significant effect on Ghana’s GDP. Therefore, we reject the null hypothesis (Ho: µ1 = µ2) which posits, pension fund assets do not have strong effect on Ghana’s GDP; and accept the alternative hypothesis (H1: µ1 ≠ µ2) which states, pension fund assets have strong influence on Ghana’s GDP.
Conclusion
The analysis revealed valuable contribution of pension fund assets to measurement of Ghana’s GDP. This affirms significant contribution of pension scheme activities and operations; and pension funds management to stimulation of the Ghanaian economy. Further, it underscores the need for policies and programmes geared toward efficient and effective implementation of the various pension schemes to be considered and followed with rapt attention.
Reiteratively, the Social Security and National Insurance Trust (SSNIT), trustees, pension fund managers, and custodians of pension fund assets are entreated to consider strongly, alternative investments as strategic and viable source of increasing returns on pension investments to help increase the superannuation incomes and other benefits of contributors during retirement.
Nominal returns and real returns on investments play pivotal roles in the measurement and determination of value for total pension fund assets; and in the realisation of the afore-mentioned objective. This notwithstanding, volatilities have been identified as unavoidable companions in the financial markets. To this end, it is imperative for SSNIT and other pension fund managers to ensure constant monitoring and exploration of the local and international financial markets for the best financial investment deals to assure higher real returns on pension funds’ investments. Finally, it is incumbent on pension fund managers to ensure their investment decisions are prudent; and ably guided by the principles of security, profitability and liquidity.
Author’s Note
The above write-up was extracted from an earlier publication titled: “Effect of Pension Fund Assets on Ghana’s GDP” by Ashley et al. (2019) in the International Journal of Business and Management.
The writer is a Chartered Economist/Business Consultant.
List of References
Ashley, E. M., Takyi, H., & Obeng, B. (2016). Research Methods: Quantitative and
Qualitative Approaches to Scientific Inquiry. Accra: The Advent Press.
Bank of Ghana. (2013). Statistical bulletin: March 2013. Retrieved from
https://www.bog.gov.gh/privatecontent/Publications/Statistical_Bulletin/2013/StatBulMarch2013c.pdf.
Bowerman, B. L., & O’Connell, R.T. (1990). Linear Statistical Models: An Applied
Approach (2nd ed.). Belmont, CA: Duxbury.
Creswell, J. W. (2009). Research Design: Qualitative, Quantitative, and Mixed Methods Approaches (3^{rd} ed.). Thousand Oaks, CA: Sage Publications.
Field, A. (2009). Discovering Statistics Using SPSS (3rd ed.). London: Sage.
Frankfort-Nacmias, C. & Nachmias, D. (2008). Research Methods in the Social Sciences (7^{th} ed.). USA.: Worth Publishers.
Ghana Statistical Service. (2019). Rebased 2013-2018 annual gross domestic product. April 2019 Edition. Retrieved from http://www.statsghana.gov.gh/gssmain/storage/img/marqueeupdater/Annual_2013_2018_GDP_April%202019%20Edition.pdf.
Hayes, A. (2021). Confidence interval. Retrieved from
https://www.investopedia.com/terms/c/confidenceinterval.asp.
ILO. (2008). Ghana National Pensions Act, 2008 (Act 766). Retrieved from
http://www.ilo.org/dyn/natlex/docs/ELECTRONIC/88539/101275/F563187713/GHA88539.pdf.
IOP. (2008). National Pensions Act 2008 – International Organisation of Pension. Retrieved from http://www.iopsweb.org/resources/48749441.pdf.
NPRA. (2019a). History/background of the 3-tier scheme. Retrieved from http://npra.gov.gh/index.php/about-us/history-background-of-the-3-tier-scheme.
NPRA. (2019b). National Pensions (Amendment) Act, 2014 Act 883. Retrieved from
http://npra.gov.gh/pdf/National_Pensions%20(Amendment)_Act,2014_Act%20_.pdf.
NPRA. (2019c). National Pensions Regulatory Authority. Retrieved from http://npra.gov.gh/
SSNIT.Org. (n.d.). Act 883 – SSNIT. Retrieved from https://www.ssnit.org.gh/about-us/act-
883/
World Bank. (2019a). Databank. Retrieved from
https://databank.worldbank.org/reports.aspx?source=2&series=NY.GDP.MKTP.CD&country=#
World Bank. (2019b). Ghana. Retrieved from https://data.worldbank.org/country/ghana.
World Bank. (2019c). The global economy: Heightened by tensions, subdued growth.
Retrieved from http://www.worldbank.org/