The poor quality electricity supply has been recorded as a major problem hampering the operations of SMEs in developing countries, and is more prevalence in Sub-Saharan African countries and some part of the Middle-East. In recent times, access to reliable electricity supply and associated high tariffs is creeping to the top spot of SMEs challenges in Ghana, with SMEs in the country losing over US $686.4 million sales annually since the beginning of 2009. Considering the significant contributions SMEs made towards the socio-economy of countries, if care is not taken to assess the effect of the intermittent power outages on the running of businesses so as to create the awareness to policy formulators in other to find a lasting solution to this canker, then, the demise rate of SMEs will be on the higher pedestal. Against this backdrop, the researchers assess the impact of the power insecurity on the growth of SMEs with a particular study on cold-store operators in Asafo Market of Kumasi in Ghana, since previous researchers have not zoom on small sectors and also did not used case study approach. The research findings indicated that, power outages had a negative effect on SMEs growth, while the cost of operating businesses saw a significant increase under the power outages. Cost of alternative sources of power also significantly pushes the operation cost of businesses high.
Access to electricity and its accompanied high tariffs poses a greater challenge to SMEs growth in lower income countries, as compared with those in higher income countries.
This assertion is confirmed in a research by [
Research by [
These have been partly blamed on market and state failures, which have led to the poor electricity supply.
Similar research by [
A previous research on the obstacles to SMEs growth and development in the Sub-Saharan African countries, [
In an attempt to mitigate the hash effect of the poor power supply on their businesses, some SMEs use alternative power generators, reduce hours of work, and vary the working hours [
The absence of competition in the electricity supply business in Ghana has been blamed as a possible cause of poor performance by the state-owned Electricity Company of Ghana (ECG). [
Considering the important role electricity plays in the running and growth of these SMEs (Cold Store Operators), and the fact that there are no prudent alternatives [
A study on the effect of power fluctuation on the profitability and competiveness of SMEs is therefore crucial and a necessary step to influence government action on energy supply.
This research aims at assessing the effect of electricity power insecurity on the growth of SMEs (Cold Store Operators), its accompanied cost implications, and the available alternative power supply. In the end, a suitable solution to the crises will be recommended.
The study will also add to the existing limited literature in the sector, since the electricity crises in Ghana are an on-going phenomenon. The research is organized and presented in four sections, the theoretical background, methodology, research finding and conclusions.
SMEs growth is the wish of both the entrepreneurs and the State, because of the important role SMEs play in the economic development of the individual and the nation at large [
A positive correlation has been established by several researchers [
There is an ample evidence that, poor and high tariff electricity supply affects the growth of firms negatively [
When it come to the size of the firm and the impact of electricity outages on productivity, research indicate that small firms are the hardest hit than their larger counterparts, whiles the shorter reoccurrences of outages have hasher impact than longer durations [
Some researchers are of the view that poor electricity supply generally may not always impact negatively on firms output, as a study by [
Notwithstanding those findings, electricity outages have been proven severally as having negative effect on firms’ productivity, especially in Africa and among small firms whose activities hinges around the availability of electricity power [
Based on the above literatures reviewed, the researcher proposes a hypothesis that H1: Electricity power outages affects the growth of SMEs negatively.
The cost of electricity supply is higher in most developing countries than in developed ones, whiles there are also fewer and costly alternative energy supply in developing countries [
Empirical research by [
We therefore propose a hypothesis based on the literatures reviewed that H2: Electricity power outages increases the cost of operating SMEs.
Report by [
Report by [
In Ghana, some of the cold store operators have adopted to drying and smoking of their fish and meat product as a means of preserving them, but the demand for the type is low and very tedious to practice [
The use of generators as alternative source of electricity is very popular among African SMEs than most regions in the world [
Research by [
From the literatures above, the ensuing hypothesis have been proposed H3: Alternative energy supply has higher cost implications on SMEs operations.
H4: high cost of alternative power supply positively affect the operational cost of SMEs.
H5: High operational cost has negative impact on the SMEs growth.
(POEXP-POWER OUTAGES EXPERIENCE, HCAPS-HIGH COST OF ALTERNATIVE POWER, HOCOST-HIGH OPERATIONAL COST.)
The terms of reference for the study was the fish and meat sellers using refrigerators in storing their products within the Asafo market (noted for cold store businesses) since the most significant impact of the electricity outages on SMEs will be much felt in that area since the business survival thrives of electricity.
The study adopted the National Board for Small Scale Industries (NBSSI) of Ghana’s definition of SMEs, for those employing below 250 workers with capital less than US $100,000 [
This research used both qualitative and quantitative approach in obtaining the right data (primary and secondary) for the analysis. A total of 450 cold store operators within the Asafo Market catchment area whose business depends heavily on the electricity supply was select for the study. The sample size was influenced by the population of the cold store operators in the cluster and some common characteristics they exhibited during the
pre-survey exercise.
A purposive sampling techniques was employed in the selection process, since the major targets were those severely affected by the electricity outages. So the area was imaginary divided into four and 60 sample from each group was selected with additional 10 from their union executives. A stratified sampling techniques was used. Because the traders were desperately looking for a solution to the problem, they were willing to give all the needed information, as a result, the researcher was able to retrieve 250 questionnaires from the total questionnaires administered.
The dependent variable is the growth of the SMEs, whiles the independent are power outages experience high cost of alternative power, and high operational cost (the sample questionnaire is attached as Appendix for reference).
The quantitative data obtained was put through several stages such as, data cleaning, normality test, linearity test, EFA, CFA and Structural Equation Model was used in estimating paths coefficients and significant levels whilst Alpha Cronbach’s, discriminant and convergent validity estimate were generated to validate the measurement model.
SPSS version 21 was used in analyzing the data using the SEM to establish the relationships and associations between the dependents and independents variables proposed in this study.
Purposively, 250 questionnaires were sent to selected SMEs located in Ghana with 250 questionnaires repressenting 100% approximately retrieved, involving 124 respondents as females and 126 as males, 131 with ages of 30 years and above and 119 with ages of 30 years and below, 156 as married and 94 as single. Also, 145 of the respondents have experienced formal education and 105 have none of educational experience.
To ensure sufficient usability, reliability, and validity of the data received for further analysis in testing causal relationships formulated in this study, the data was staged in multiples of screening. In Microsoft excel, the compute blank values for all the cases showed values not equal to zero and that none of the cases had missing value to be either eliminated or replaced with the mean. Subsequently, unengaged responses were checked in the data set to establish more decent analysis. This represent cases that respondents answered with either same values (e.g. 3333, 1234 or 1111, 2222, 3333, etc.) in the Likert scale for survey item for all the constructs. Explicitly, this implies that, the research respondents were not engaged in providing indebt views or explained no reasonable variance among constructs in the data to support the causal specifications proposed in the research model. In identification of such cases, standard deviation was computed for individual cases received using excel and given the threshold of SD > 0.5, all cases with SDs less than the threshold were eliminated as the amount of variance explained among the items in such cases are not highly strong enough to predict the effect of one construct on the other (IVs and DV). None of the cases had discrepancies as the SD were all sufficiently greater than 0.5 and that, all the cases were considered sufficient for further analysis. In expedition for more accurate and reliable data set, the distribution, location and variability of the data as responded by the 250 selected sample was verified. Guided by kurtosis, none of the values was greater than or less than +/− 2.00 to be considered an indication for potentially problematic kurtosis, hence incidence of insufficient variance were present in the data set. Successively, multivariate assumption was clarified in the data set. Testing the linear assumptions established in the research model, is the linearity and multi-collinearity tests that exposes the linear concepts in the model to predict the existence of multi-collinearity issues among constructs specified in the model. In Linearity test, curve estimation regression was computed for all direct effects or linear relationships in the model. The results show that the relationships between variables are sufficiently linear with all p-values less than 0.05 and F-value for all linear relationships among variables higher than other curves estimates. Regarding Multi-colli- nearity the Variable Inflation Factor (VIF) for all the independent variables were simultaneously tested. Given the outcome, the VIFs were all less than 2.0, indicating that all the variables are distinct and are of less or no issues of multi-collinearity problem.
From the cleaned field data, the EFA using PCA with Varimax rotation was executed to ascertain the correlations postulated in the model. Several analysis were staged to substantiate the expected loadings of the observed variables and the existence of adequate correlation whilst reliability and validity criteria are all met. The KMO value obtained (0.872) with approximate Chi-Square of 6286.52 at 210 DF, significant at 0.001 level depict that, the correlations in the items of all the constructs are sufficiently large for the PCA. Indeed, significant Bartlett’s test demonstrates that, the inter-correlations in the data set is not from a population in which the correlation is an identity matrix and that, items correlation coefficients different from zero. This is further ascertained by the individual KMO values obtained for each items which were all above 0.5 indicating good sampling adequacy for the research data. Prior to the initial extraction is the factor eigenvalues above 1 as recommended by Kaiser (Kaiser, 1960). The percentage (%) of variance explained by each linear component evinces that, the POEXP factor explains approximately 24.8% of the total variance with 5.2 eigenvalue, the HOCOST factor of 20.6% with 4.3 eigenvalue, the HCAPS factor of 20.6% with 4.3 eigenvalue, and the SMEGROWTH factor of 18.7% with 3.9 eigenvalue. The factor loadings present the correlations of each item with the factors extracted. Explicitly, the observed values for the factor loadings explains that, all the items are highly related to the factors on which they are extracted. Regarding the total proportion of variance in the variable explained in all the constructs by the four factors, is simply the sum of the squared factor loadings which is known as the communality of the variables. In the section labeled Communalities, the values obtained shows that, the four factors are sufficiently defined by the items correlating with them. Communalities ranging from 0.686 to 0.962 explained that, none of the observed variables were out of prediction from the four factors extracted.
The rotated component matrix (factor loading) depicts the correlation of each item with each factor. The items designed for the SMEGROTH variable show a correlation coefficients from 0.884 to 0.977 with internal consistency reliability coefficient of 0.969, the items for HCAPS variable correlate from 0.914 to 0.943 with Cronbach’s alpha value of 0.962, the items for HOCOST variable’s correlation coefficients ranges from 0.908 to 0.952 with reliability alpha of 0.960 and the items for the POEXP variable possess correlation coefficients ranging from 0.816 to 0.944 with a reliability coefficient of 0.931. Given these correlation coefficients and the alpha reliability ratios, all the items are sufficiently correlated with their factors and also internally consistent in obtaining survey data for the study. The factor pattern is clearly defined with no cross loadings meaning the factor structure has no concerns of discriminant validity issues.
Subsequent to the EFA was the Confirmatory Factor Analysis which was conducted to obtain a decent measurement model for the study, utilizing the pattern matrix model builder of the SPSS Amos version. The CFA was computed to reinforce and confirm the factor structure specified in the EFA analysis in view to clarify the relationship between observed measures (indicators) and the latent variables. In confirmation, the modification indices generated examine the predictive ability of the measurement constructs. The generated Chi-square of 412.976 at 183 df is statistically significant at the 0.001 level, meaning, the specified measurement model has less or no discrepancies in predicting the latent variables. All the path estimates were sufficiently higher than 0.5 (as shown in
To further establish reliable and valid measurement model, convergent validity with AVE was calculated whilst the square root of the AVE (on the diagonal in the matrix below) were compared to all inter-factor correlations. The results as shown in
From the factor scores, composite variables were created in AMOS. As guided by extant theories, direct paths were added among the variables to achieve excellent fit for the construct model to probably indicate that, the
Kaiser-Meyer-Olkin measure of sampling adequacy. | 0.872 | ||||
---|---|---|---|---|---|
Approx. chi-square | 6286.52 | ||||
Bartlett’s test of sphericity | DF | 210 | |||
SIG. | 0 | ||||
Constructs & items | Items KMO | Factor loadings | Eigenvalues | % of variance | Communalities |
POEXP | 5.215 | 24.835 | |||
Q1 | 0.782a | 0.888 | 0.793 | ||
Q2 | 0.862a | 0.859 | 0.739 | ||
Q3 | 0.872a | 0.816 | 0.686 | ||
Q4 | 0.830a | 0.907 | 0.824 | ||
Q5 | 0.805a | 0.944 | 0.895 | ||
HOCOST | 4.345 | 20.691 | |||
Q6 | 0.932a | 0.908 | 0.827 | ||
Q7 | 0.843a | 0.911 | 0.838 | ||
Q8 | 0.872a | 0.931 | 0.87 | ||
Q9 | 0.880a | 0.952 | 0.91 | ||
Q10 | 0.898a | 0.934 | 0.875 | ||
HCAPS | 4.328 | 20.611 | |||
Q11 | 0.786a | 0.914 | 0.837 | ||
Q12 | 0.902a | 0.931 | 0.868 | ||
Q13 | 0.877a | 0.924 | 0.86 | ||
Q14 | 0.824a | 0.942 | 0.89 | ||
Q15 | 0.903a | 0.943 | 0.891 | ||
SMEGROWTH | 3.942 | 18.773 | |||
Q16 | 0.927a | 0.902 | 0.82 | ||
Q17 | 0.921a | 0.884 | 0.783 | ||
Q18 | 0.833a | 0.977 | 0.962 | ||
Q19 | 0.878a | 0.969 | 0.945 | ||
Q20 | 0.914a | 0.944 | 0.894 | ||
Q21 | 0.950a | 0.902 | 0.822 |
Source: field data.
hypothesized relationships exist through linear relationships among constructs. The goodness of fit indices indicate adequate fit for the construct model at 0.672 probability of 0.179 Chi-square at 1df as shown in
In testing the research hypotheses, standardized path estimates were computed for all the specified direct paths. By inspection, all the estimates both direct and indirect relationships between the predictor variables and the
Items | Component | |||
---|---|---|---|---|
SMEGROTH (F1) | HCAPS (F2) | HOCOST (F3) | POEXP (F4) | |
Chrombach’s alpha | 0.969 | 0.962 | 0.960 | 0.931 |
Q1 | 0.888 | |||
Q2 | 0.859 | |||
Q3 | 0.816 | |||
Q4 | 0.907 | |||
Q5 | 0.944 | |||
Q6 | 0.908 | |||
Q7 | 0.911 | |||
Q8 | 0.931 | |||
Q9 | 0.952 | |||
Q10 | 0.934 | |||
Q11 | 0.914 | |||
Q12 | 0.931 | |||
Q13 | 0.924 | |||
Q14 | 0.942 | |||
Q15 | 0.943 | |||
Q16 | 0.902 | |||
Q17 | 0.884 | |||
Q18 | 0.977 | |||
Q19 | 0.969 | |||
Q20 | 0.944 | |||
Q21 | 0.902 |
Extraction method: Principal component analysis. Rotation method: Varimax with Kaiser normalization. aRotation converged in 5 iterations.
Indeces | ||
---|---|---|
Chi square | 412.976 | |
D/F | 183 | |
Probability | 0.000 | |
Goodness of fit | ||
Metric | Observe values | Recommended thresholds |
Cmin/Dif | 2.257 | Between 1 and 3 |
CFI | 0.963 | >0.950 |
RMSEA | 0.051 | <0.060 |
PCLOSE | 0.331 | >0.050 |
SRMR | 0.072 | <0.090 |
RMR | 0.010 | <0.05 |
PNFI | 0.583 | >0.50 |
GFI | 0.870 | >0.80 |
NFI | 0.936 | >0.90 |
Source: field data.
CR | AVE | MSV | ASV | HOCOST | SMEGROWTH | HCAPS | POEXP | |
---|---|---|---|---|---|---|---|---|
HOCOST | 0.960 | 0.828 | 0.007 | 0.004 | 0.910 | |||
SMEGROWTH | 0.971 | 0.846 | 0.012 | 0.007 | −0.086 | 0.920 | ||
HCAPS | 0.962 | 0.834 | 0.001 | 0.000 | 0.016 | −0.009 | 0.913 | |
POEXP | 0.932 | 0.733 | 0.012 | 0.006 | 0.075 | −0.111 | 0.029 | 0.856 |
Indeces | ||
---|---|---|
Chi square | 0.179 | |
D/F | 1 | |
Probability | 0.672 | |
Goodness of fit | ||
Metric | Observe values | Recommended thresholds |
CMIN/DIF | 0.179 | Less or between 1 and 3 |
CFI | 1.00 | >0.950 |
RMSEA | 0.000 | <0.060 |
PCLOSE | 0.756 | >0.050 |
RMR | 0.007 | <0.05 |
PNFI | 0.166 | >0.50 |
GFI | 1.00 | >0.80 |
NFI | 0.998 | >0.90 |
Source: field data.
criterion variable are significantly different from zero (all p-values <0.001 and 0.05 as shown in
The research findings indicated that, power outage experience has a negative impact on SMEs growth, and this is in line with the findings by [
The research further shows that, the major alternative power supply is generator, but the cost of using the generator was very high and was increasing the operational cost and reducing the firm’s growth.
Considering the significant role play by SMEs and the limitations posed by poor electricity supply on their
Paths | Unstandardized | Standardized | |||||||
---|---|---|---|---|---|---|---|---|---|
Estimate | Estimates | S.E. | C.R. | P | Label | Remarks | |||
HOCOST | <--- | HCAPS | 0.262 | 0.267 | 0.059 | 4.419 | *** | H4 | Accept |
HOCOST | <--- | POEXP | 0.128 | 0.146 | 0.053 | 2.412 | 0.016 | H3 | Accept |
SMEGROWTH | <--- | POEXP | −0.41 | −0.263 | 0.089 | −4.604 | *** | H1 | Accept |
SMEGROWTH | <--- | HCAPS | −0.416 | −0.24 | 0.102 | −4.08 | *** | H2 | Accept |
SMEGROWTH | <--- | HOCOST | −0.34 | −0.193 | 0.105 | −3.244 | 0.001 | H5 | Accept |
performances, it is not in bad taste to see governments from most developing countries, coming out with policies and programs to mitigate the poor electricity supply [
The most obvious area for action is to improve the reliability of the electricity supply, which needs to be measured and monitored. This may require short-term action to reduce technical faults, for example, through maintenance of the transmission and distribution infrastructure, or it may require longer-term interventions to expand generating capacity.
As the case in Ghana, proper load shedding exercise for equitable distribution and future planning by SMEs are also encouraged as a short term remedy, for the long term solution, [
The research findings indicated that, the cost of alternative power supply was high comparing with that of the electricity as a result additional cost incurred by the State in improving the grid power can be transferred to consumers [
The sharing of generators by the SMEs in a common vicinity is also encouraged to reduce the operational cost.
In conclusion, the power outages are having hasher effect on the activities of the SMEs. Hence pragmatic actions should be taken to implement the recommendations of this research to reduce the problems of these entrepreneurs so as to improve growth and reduce the demise rate of the businesses.
Solomon KwartengForkuoh,YaoLi, (2015) Electricity Power Insecurity and SMEs Growth: A Case Study of the Cold Store Operators in the Asafo Market Area of the Kumasi Metro in Ghana. Open Journal of Business and Management,03,312-325. doi: 10.4236/ojbm.2015.33031
This research work is undertaking to have an insight on the effect of the current electricity power outages on the growth of SMEs (cold store operators) in the Asafo District of Kumasi in Ghana.
All information you provide will be treated as confidential and anonymous, and also will be used for academic research only. Thank you.
Please tick (√) the most appropriate that describe your business since 2009 to date.
1. Strongly disagree 2. Disagree 3. Neither agree nor disagree 4. Agree 5. Strongly agree
General questions
Please provide one answer to each of the following general questions on your business.
25. Gender? 1. Male 2. Female
26. What is the highest educational level of the owner or general manager of the business?
1: Less than high school diploma 2: High school diploma 3: A bachelor’s degree 4: A master’s degree or above 5: Others
27. Which of the following best describes your position in this business? Are you...
1: The sole owner of this business 2: A partner in this business 3: The person in charge of finance in this business 4: Occupying another position in this business 5: Others. Is your business...?
28. How long has your business been in existence?
1: Up to 2 years 2: From 2 - 5 3: From 5 - 9 4: From 9 - 13 5: From 13 and above