Econometric Competencies and Entrepreneurship Development 

Adebayo G ADEBAYO

Department of Accountancy

Rufus Giwa (Formally Ondo State) Polytechnic.

Owo, Ondo State, Nigeria

Abstract

This study is designed specifically to demonstrate the application of econometrics or quantitative analysis especially in helping the numerous small and medium enterprises in Nigeria in critical decision making. An Agro-business labour saving fabricating firm in Nigeria was selected as a case and pilot study for all other small or medium enterprises fabricates labour saving machines such as the cassava frying machine (CFM) and palm-oil extracting machine (PEM). The agro-business firm also fabricate cassava grater (GRT) and appliances (APP). The entrepreneur is worried on some critical problems. First, whether these special offers would affect the sale of the major products CFM and PEM since the same customers buy the major products and the special offers. This implies that money spent on special offers will not be spent on CFM and/or PEM. An increase in the sales of special orders may correspond to a decrease in the sale of CFM and PEM. Second is what would be the total sales taking into consideration especially the seasonal fluctuations and third how to identify existing customers who are most likely to respond to proposed improved offer on CFM and PEM. This is important as the Firm fabricates these products on demand. Three models were designed to find solution to each of the problems. The first model was the two-stage least square regression. Findings revealed that the special offer sales GRATER and APPLIANCES) had no negative effect on the sales of the major products (CDM and PEM). All the products significantly contributed to the total sales. The second model was forecasting. The model satisfied forecasting requirements and was able to forecast total sales from April to December 2016 taking into consideration seasonal variations referred to as fluctuations. The third model is the recency, frequency and monetary (RFM) analysis. This marketing tool is used to classify customers according to how recent they patronize, how often and how much is involved in individual customer’s cumulative patronage. The RFM analysis carried out on the customers identified the firm’s customers that would likely respond to new offer. An RFM score of 300 and above qualified a customer to be selected. Part of the recommendations was that entrepreneurs should avail themselves of the decision making tools for better management of their enterprises.

 Keywords: 

 Introduction

The world is facing many economic challenges and issues. These do not isolate the developed economies. Both are debt ridden, with regional economic imbalances and geo-political challenges. There is general economic meltdown in the world market. The Nigeria economy has been acutely affected because of the fall in the price of crude oil in the world market as a result of these economic imbalances and trade policies that are not conducive to Nigerian oil market. Since revenue from crude oil takes about 90% of the Federal government total revenue, there is the critical need to raise non-oil revenue to ensure fiscal sustainability while maintaining infrastructure and social spending.

The Federal government has taken a bold step towards revamping agriculture and overhauling its solid mineral resources. From an entrepreneurial perspective, the present economic meltdown in Nigeria would eventually be a blessing in disguise. The Central Bank of Nigeria has been instructed by President Buhari to create a “synergy” and organize soft loans to agro-based industries and other export businesses.

The main objectives of the 2016 Buhari government’s budget are to make the “synergy” work among all the different players in the country’s economy. These include the banking system, financial institutions, government entities, regulators and other arms of the government.

  • Entrepreneurial Development.

The credibility of the Buhari government among the Nigerian populace as a result of his zero tolerance to corruption and his “Big Bang” [a rapid reform which is economically necessary as a result of severe macroeconomic imbalances (Gelb, Jefferson and Singh 1993) such as this period of economic crisis in Nigeria] approach to major economic reforms, has spurred many investors into agro-based industries. They believe that the government meant business.

One of these respondents is a medium scale firm in Nigeria that fabricates machines to boost agricultural produce. The Firm has recently reactivated two of its machines-the Cassava Frying Machine (CFM) and the Automated Palm oil Extracting Machine (PEM)

  • Objectives of the Firm

The major objectives of the Firm in fabricating these labour saving machines is the issue of health hazard on the one hand. Many local manual cassava frying and palm oil producers especially women, had been subjected to untimely death due to incessant heat from the local frying system and the palm oil production. These machines would reduce drastically the feminine life wasted on daily basis. On the other hand, these machines will boost agricultural produce of “GARI”(a local name for the fine grain output from the frying process of cassava) and palm oil to contribute to increase the recent low GDP rate in Nigeria. [Economic growth in the last quarter of 2015 was 2.1% while total growth in the year was 2.8%, the slowest since 1999 to date (NBS, 2016). This statistics seems to toe the line of the global GDP growth projection of 2.5% which is 0.3% point less than November 2015 outlook (GEO, 2016)]. Therefore any entrepreneurial effort to boost the Nigerian GDP at this trying period is a right decision in the right direction.

  • The Cassava Frying Machine

The CFM is powered by electricity and is capable of frying about 5-10-kg of already peeled, washed and had been cut into smaller sizes and loaded into the machine drum. The machine grinds the cassava, presses it cause fermentation and fry the cassava into very fine grains. Adjustments by the use of some special appliances sold by the firm to its customers may cause the machine to dry cassava from the normal “GARI” into smoother form or into powder. It is an automated machine with 100% local components. A single CFM costs N45, 000 ($225)

  • The Automatic Palm oil Extracting Machine..

The PEM is an automated machine that produces fine, clear, well heated palm oil. Palm fruits are removed from the bunch after some period for partial fermentation, washed and then loaded into the machine tank or drum to the brim. The machine twists and separate kernels from tissues, extract the paste and heat into fine glossy oil. The slag is released from an outlet. It is also powered by electricity. The machine is made with 90% local components, that is, 10% components have to be imported. A single machine costs N50, 000 ($250), about N5000 ($25) more than CFM probably as a result of the cost of the imported components.

  • Statement of the Problem.

The two machines fabricated by the firm are the Cassava Frying Machine (CFM) and Palm oil Extracting (PEM) and are sold to customers on demand. Apart from these machines, there are two special offers that are also sold. These are grater (GRT) and appliances (APP). The APP is a device to enhance either the CFM or PEM, especially as a power saving device, at the customers’ option and is sold at the rate of N12, 000 per unit. These appliances are capable of making adjustments possible to CFM and PEM and other agro-based machines.  The GRT powered by electricity, the pealed, and washed cassava are loaded into its receptor and it grinds cassava very well. This is sold at the rate of N15, 000 per unit. Every month the Firm makes these special offers to customers who need the APP on previously bought machines or on a proposed purchase of CFM and /or PEM. The GRT is mostly purchased by customers who could not afford CFM. The firm is now concerned about the following problems:

  1. Whether these special offers would affect the sale of the major products CFM and PEM since the same customers buy the major products and the special offers. This implies that money spent on special offers will not be spent on CFM and/or PEM. An increase in the sales of special orders may correspond to a decrease in the sale of CFM and PEM.
  2. How to make a good forecast of the total sales from major products and the special offers.
  3. How to identify existing customers who are most likely to respond to proposed improved offer on CFM and PEM. This is important as the Firm fabricates these products on demand.
  • Objectives of the Study.

The primary objective of this study is to underscore the importance of econometric or quantitative analysis in solving most of the problems of entrepreneurs and hence enhance entrepreneurship development in Nigeria. In realization of this objective, the study had focused on helping to collect all relevant data on the Firm’s customers. These include each customer’s date of transactions with the Firm, amount of purchases each time, total number of transaction and the most recent transaction. There will also be collection of data on total monthly sales on CFM, PEM, APP and GRT for the past 51 months. Other secondary objective, in addition are:

  1. To create compactible models to find solution to each of the Firm’s area of concern.
  2. To discuss the findings and give expert recommendations on the findings.

  1. Review of Related Literature

2.1 The Nigerian Cassava

Cassava is well known as manihot esculenta or manilot utilissima (Yakasi, 2010). In Nigeria, cassava is grown in all the ecological zones andit is planted all the year round on the availability of moisture (Odoemenem and Otanma, 2011). Production is vital to the economy of Nigeria as the country is the world’s largest producer of the commodity. The crop is produced in 24 of the country’s 36 states. In 1999, Nigeria produced 33 million tonnes, while a decade later, it produced approximately 45 million tonnes, which is almost 19% of production in the world. The average yield per hectare is 10.6 tonnes.(Wikipedia, n.d.)

In Nigeria, cassava production is well-developed as an organized agricultural crop. It has well- established multiplication and processing techniques for food products and cattle feed. Cassava is processed in many processing centres and fabricating enterprises set up in the country. Cassava is used in the preparation of several household foods and derivatives such as paste, biscuits, bread sagos and sauce. Its starch is for industrial use such as baby food, jelly, custard poeders and confectioneries (Echebiri and Edeba, 2008).  Roots or leaves are made into flours. Flours are of three types, yellow garri, white garri, or intermediate colour. These varieties are a matter of choice and traditional attachment. Therefore it may be erroneous to classify any type as the best product in Nigeria. Its other products are as dry extraction of starch, glue or adhesives, modified starch in pharmaceutical as dextrines, as processing inputs, as industrial starch for drilling, and processed foods.

 2.2 Palm Oil in Nigeria

 

Palm oil is as old as Nigeria itself and has been an important subsistence, but until recently a supportive factor in the diet of many Nigerians. Palm oil is the world’s largest source of edible oil, accounting for 38.5 million tonnes or 25% of the global edible oil and fat production (MPOC, 2007). Palm oil is a product extracted from the fleshy mesocarp of the palm fruit (Elaeis guineensis). The global demand for palm oil is growing thus, the crop cultivation serves as a means of livelihood for many rural families, and indeed it is in the farming culture of millions of people in the country. Akanbge et al (2011), referred to this product as capable of having multiple values, a feature that underscores its acclaimed economic importance. Eventually oil graduated from domestic use to industrial application which had appreciated its production geometrically (Omereji,2005). Ekine and Onu(2008) estimated palm oil consumption of about two litres per a family of five per week for cooking. Today, consumption must have tripled since Nigerian house hold now uses palm oil beyond normal consumption. Palm oil is also an essential multipurpose raw material for both food and non-food industries (Armstrong, 1998). Palm oil is used in the manufacturing of margarine, soap candle, base for lipstick, waxes and polish bases in a condense form, confectionary (Embrandiri et at., 2011; Aghalino, 2000), and other uses in pharmaceuticals.

2.3 Forecasting

Because economic and business conditions vary over time, managers must find ways to keep abreast of the effects that such changes will have on their organizations. One technique that can aid in plan­ning for future needs is forecasting. Although numerous forecasting methods have been devised, they all have one common goal—to make predictions of future events so that projections can then be incorporated into the planning and strategy process.

2.3.1 Time-series forecasting meth­ods involve the projection of future values of a variable based entirely on the past and present obser­vations of that variable. Examples of economic or business time series are the monthly publication of the Consumer Price Index, the quarterly statements of gross domestic prod­uct (GDP), and the annually recorded total sales revenues of a particular company.(Levine et al, 2005)

2.3.2 Least-Squares Trend-Fitting and Forecasting. The component factor of a time series most often studied is trend. Trend is studied as an aid in making intermediate and long-range forecasting projections. As depicted in Figure 1 to obtain a visual impression of the overall long­term movements in a time series, a chart is constructed in which the observed data (dependent variable) are plotted on the vertical axis, and the time periods (independent variable) are plotted on the horizontal axis.(See figure 1 below)

The Forecasting add-on module provides two procedures for accomplishing the tasks of creating models and producing forecasts.

The Time Series Modeler procedure estimates exponential smoothing, univariate Autoregressive Integrated Moving Average (ARIMA), and multivariate ARIMA (or transfer function models) models for time series, and produces forecasts. The procedure includes an Expert Modeler that automatically identifies and estimates the best-fitting ARIMA or exponential smoothing model for one or more dependent variable series, thus eliminating the need to identify an appropriate model through trial and error. Alternatively, one can specify a custom ARIMA or exponential smoothing model.

AR(1)        = a first-order autoregressive to correct for residual serial correlation. It is regressing the dependent variable(s) with linear combination of its past values or lagged values.

MA(1)       = a first-order moving  average model ,i.e., regressing the dependent error with linear combination of its past error or lagged error. It also corrects serial correlation.

2.4 Research Questions.

The following research questions are formulated by the researcher

  1. What are the total monthly sales of the Firm for 51 months?
  2. How many are the numbers of customers of the Firm are their variable such as name, address, city, state- province, post code, gsm numbers, country, gender, age category are available?
  3. How can customers that would respond to new offer be identified?
  4. What is the forecast total sales from April to December 2016?

2.5 Research Hypotheses.

The following research hypotheses are formulated by the researcher at 5% level of significance.

  1. There will be no significant relationship between total sales of special offers and CDM sales.
  2. There will be no significant relationship between total sales of special offers and POE sales.
  3. There will be no significant relationship between total sales of special offers and APP sales.
  4. There will be no significant relationship between total sales of special offers and GRTsales.

 

  1. Methodology

 3.1 Data Collection

The data collected for this study is a demonstrative data only on behalf of the firm. It is characteristic of most small and medium enterprises of nearly the same status. This case study Firm is having about 90 customers. Sales records on CFM, PEM, APP and GRT were collected for 51 months (January 2012 to March 2016). In the Table below, the author computed the total sales for CFM, PEM, APP and GRT and special offer total sales (SPSALES). Others computed are discount on CDM at 3.5% from records and discount on PEM at 4.8%. A log discount on CFM and PEM and their lagged variables are automatically computed in the SPSS 21 and E-View 7.1 work files. The data are presented in Table 1 below.

 Table 1    Company Data for 51 Months – from January 2012 to March 2016

tsales cfm pem spsales grt app discfm dispem
N N N N N N N N
109.50

123.50

130.00

147.00

154.10

174.00

166.50

178.00

166.50

153.20

132.50

132.70

148.80

167.80

180.50

192.50

192.00

202.50

193.50

183.20

196.70

198.80

179.10

169.80

193.10

194.10

213.00

241.40

231.80

226.40

226.50

245.70

231.20

207.50

219.90

214.00

241.30

265.00

267.50

246.90

237.00

234.20

251.20

266.00

270.80

264.80

258.00

243.80

244.00

246.50

263.30

57.50

60.00

62.50

67.50

75.00

78.00

75.00

73.00

58.00

55.00

56.00

57.00

58.00

65.00

84.00

78.00

77.50

74.50

62.50

68.00

83.00

87.50

78.50

73.00

71.00

73.00

83.00

95.00

92.50

82.00

87.00

95.00

100.00

97.00

100.00

95.00

103.00

112.00

113.00

110.00

105.00

97.00

94.00

95.00

94.00

93.00

100.00

103.00

102.00

102.50

103.00

40.00

50.00

52.50

65.00

64.00

60.00

75.00

86.50

87.50

77.50

60.00

52.00

65.00

73.00

67.00

85.00

85.00

97.50

100.50

85.00

83.00

80.00

68.00

60.00

82.50

80.00

90.00

110.00

108.00

105.00

100.00

110.00

90.00

80.00

78.00

72.00

90.00

105.00

107.50

90.00

85.00

90.00

108.00

120.00

125.00

120.00

105.00

85.00

86.00

85.00

100.00

12.00

13.50

15.00

14.50

15.10

16.00

16.50

18.50

21.00

20.70

22.50

13.70

25.50

29.80

29.50

29.50

29.50

30.20

30.20

30.70

30.70

31.30

32.60

36.80

39.60

41.10

40.00

39.40

39.30

39.40

39.50

40.70

41.20

40.50

41.90

47.00

48.30

48.00

47.00

46.90

47.00

47.20

49.20

51.00

51.80

51.80

53.00

55.50

57.00

59.00

60.30

4.00

5.00

6.00

6.00

7.00

8.00

8.50

10.00

12.00

11.50

13.00

14.00

16.00

20.00

19.50

19.00

18.50

18.50

18.50

17.80

18.00

18.80

19.60

20.00

20.80

21.40

22.00

22.40

22.80

23.00

22.50

22.00

21.20

21.50

23.50

26.00

26.30

26.00

26.00

26.10

26.50

27.00

27.20

28.00

27.00

26.80

27.00

29.00

29.00

29.20

29.30

8.00

8.50

9.00

8.50

8.10

8.00

8.00

8.50

9.00

9.20

9.50

9.70

9.80

9.80

10.00

10.50

11.00

12.00

12.00

12.40

12.70

12.50

13.00

16.80

18.80

19.70

18.00

17.00

16.50

16.40

17.00

18.70

20.00

19.00

18.40

21.00

22.00

22.00

21.00

20.80

20.50

20.20

22.00

23.00

24.80

25.00

26.00

28.80

28.00

29.80

31.00

2.01

2.10

2.19

2.36

2.63

2.73

2.63

2.56

2.03

1.93

1.96

1.99

2.03

2.28

2.94

2.73

2.71

2.61

2.19

2.38

2.91

3.06

2.77

2.56

2.49

2.56

2.91

3.33

3.24

2.87

3.04

3.33

3.50

3.40

3.50

3.33

3.61

3.92

3.96

3.85

3.68

3.40

3.29

3.33

3.29

3.26

3.50

3.61

3.57

3.59

3.07

1.92

2.40

2.52

3.12

3.07

2.88

3.60

4.15

4.20

3.72

2.88

2.50

3.12

3.50

3.22

4.08

4.08

4.68

4.82

4.08

3.98

3.84

3.26

2.88

3.41

3.50

4.32

5.28

5.84

5.04

4.80

5.28

4.32

3.84

3.14

3.46

4.32

5.04

5.16

4.32

4.08

4.32

5.18

5.76

6.00

5.76

5.04

4.08

3.57

4.08

4.80

       *The average Sales discount on CFM is 3.5% and on PEM is 4.8%. No discounts for APP and GRT.

            The LAG variable for CFM and LAG for PEM are automatically Created Series in the SPSS 21 and E-                           View 7.1 work files.

    Figure 1    Graph of the Relationship between Total Sales, and the sales from Cassava Frying Machine, Palm oil Extracting Machine, Appliances and Grater

3.2 Models Specification

Three models are specified to the three major problems faced the Firm as stated under the statement of the problem. These are (1) Two-Stage Least Squares Regression (2) Forecasting Models and (3) The RFM Customer Analysis

 

3.2.1 Model 1.Two Stage Least Square Regression

A careful observation of the relationship between CFM, PEM and the special offers shows that there is a feedback loop between the response and the two major products which are predictors. One of the basic assumptions of the ordinary least-squares (OLS) regression model is that the values of the error terms are independent of the values of the predictors. When this “recursivity assumption” is broken, the two-stage least-squares (2SLS) model can help solve these problematic predictors. The 2SLS model assumes that there exist instruments, or secondary predictors, which are correlated with the problematic predictors but not with the error term.

Given the existence of instrument variables, the 2SLS model:

  1. Computes OLS models using the instrument variables as predictors and the problematic predictors as responses.
  2. The model-estimated values from stage 1 are then used in place of the actual values of the problematic predictors to compute an OLS model for the response of interest.

Fifty two months of sales information is collected. The file also includes a variable, special offer, displaying each month’s special offer, which has also been recorded into two indicator variables, Appliances offer and Grater offer that can be used as predictors in the regression procedures. Lastly, the monthly discounts (and log-discounts) offered to customers are also listed. Since the monthly discounts are chosen independent of special offer sales but do not affect CDM and POE sales, they should make good instrument variables. Additionally, the lagged CDM and POE should also make good instruments. The independent, predictors and instrumental variables are in the model description as in Table 2 below.

Table 2: Model Description

   Variables Type of Variable
SpecialOfferSales Dependent
CDM Predictor
POE Predictor
Appliances predictor & instrumental
Grater predictor & instrumental
Logdiccfm Instrumental
Logdispoe Instrumental

Special Offer Sales = α0 + α1CFM + α2POE + α 3APP + α4GRT + ε                                                        Eq 1

where

α 0             =    the intercept or constant term

α      α 4   =    the coefficients of both the predictors and instrumental variables

 ε            =    the stochastic error term

All other variables are as described in the model description in Table 2 above.

3.2.2 Model 2. Forecasting Models

The Model of Forecasting from an Equation, can be dynamic or Static. The static model is chosen because the static forecasting model performs a series of one-step ahead forecasts of the dependent variable:

For each observation in the forecast sample:

 yg+k = c(l) + c(2)xs+k+c(3)zs+k+c(4)ys+k1                                                         Eq 2

Such equation is always using the actual value of the lagged endogenous variable. This is translated into sales forecast of the Firm:

Total Sales    = βo + β1Total Sales (-1) + β2Pm Ar(1) + µi                                                 Eq 3

Where:

Total Sales        =     the total sales from CFM, POE, APP and GRT for 51 months. This is the dependent variable.

Total Sales (-1) = a lagged variable of the dependent variable. This is a one step ahead static forecasts that makes the static forecast more accurate than the dynamic forecast since, each period, the actual value of Total Sales(-1) is used in forming the forecast of Total Sales.

Pmi                  =     the period expressed in months with a total of 51 months. This is the independent variable.

  β 0                            =     the constant term or the model intercept.

  β1                   =     the coefficient of the lagged variable.

  β2                    =     the coefficient of the independent variable.

 µi                    =     the stochastic or error term

The tolal sales above is transformed into:

tsalesFp+k = Ф0 + Ф1 tsalesFp+k-1 2PMp+k + AR(1) +  ε                                                    Eq 4

after being subjected to the static forecasting model

where

tsalesF = total sales forecast

 tsalesFp+k-1 =   the previous month’s (lagged) sales to be added to current sales forecast.

Ф I              = constant and coefficients.

p                 = the base period (month) of start of forecast.

k                 = any month from the forecasting period (April to December 2016)

AR(1)        = a first-order autoregressive to correct for residual serial correlation. It is regressing the dependent variable(s) with linear combination of its past values or lagged values.  .

ε i               = the error term.

For seasonal adjustment of the forecast (fitted) sales.

tsalesfSA  = ⨍i(tsalesFi)                                                                                                     Eq 5

where

⨍i = multiplicative scoring factor for a 12 month period.

tsalesFi = total sales forecast for April to December.

The tsalesfSA  = the seasonally adjusted tsalesF for months (p52-60) i.e. 52, 53, 54, 55, 56, 57, 58, 59 and 60 for April, May, June, July, August, September, October, November and December.

The actual trend base is P51. In Table 1 the total sales corresponding to P51 is N263300. A forecast for p52-60 is required; i.e. April – December 2016.

Figure 3.  Graphical Relationship between Total Sales (tsales) and Fitted Sales (tsalesF)

3.2.3 Model 3.  RFM Analysis         

RFM (Recency, Frequency and Monetary) analysis is a direct marketing option that provides a set of tools designed to improve the results of direct marketing campaigns by identifying demographic, purchasing, and other characteristics that define various groups of consumers and targeting specific groups to maximize positive response rates.

It is a technique used to identify existing customers who are most likely to respond to a new offer. This technique is commonly used in direct marketing. RFM analysis is based on the following simple theory:

The most important factor in identifying customers who are likely to respond to a new offer is RECENCY. Customers who purchased more recently are more likely to purchase again than are customers who purchased further in the past.’The second most important factor is FREQUENCY. Customers who have made more purchases in the past are more likely to respond than are those who have made fewer purchases.

The third most important factor is total amount spent, which is referred to as MONETARY. Customers who have spent more (in total for all purchases) in the past are more likely to respond than those who have spent less.

How RFM Analysis Works

Customers are assigned a recency score based on date of most recent purchase or time interval since most recent purchase. This score is based on a simple ranking of recency values into a small number of categories. For example, if you use five categories, the customers with the most recent purchase dates receive a recency ranking of 5, and those with purchase dates furthest in the past receive a recency ranking of 1. The recency ranking for this Firm is based on the past 20 months as below:

Month Interval Ranking
Jan to Apr 2016 5
Sept to Dec 2015 4
May to Aug 2015 3
Jan to Apr 2015 2
Sept to Dec 2014 1

In a similar fashion, customers are then assigned a frequency ranking, with higher values representing a higher frequency of purchases. For example, in a five category ranking scheme, customers who purchase most often receive a frequency ranking of 5.

 

The number of times a customer made purchases up to a maximum of 5,(or simply, the Transaction Counts) represent frequency ranking for the Firm.

 

Finally, customers are ranked by monetary value, with the highest monetary values receiving the highest ranking. Continuing the five- category example, customers who have spent the most would receive a monetary ranking of 5. The monetary ranking for the firm is:

Naira Value of Purchases Ranking
N120,000 and above 5
N100,000 to N120,000 4
N75,000 to N100,000 3
N40,000 to N75,000 2
Less than 40,000 1

The result is four scores for each customer: recency, frequency, monetary, and combined RFM score, which is simply the three individual scores concatenated into a single value. The “best” customers (those most likely to respond to an offer) are those with the highest combined RFM scores. For example, in a five-category ranking, there is a total of 125 possible combined RFM scores, and the highest combined RFM score is 555.

 Results and Discussion

The results of the models had produced numerous tables in their outputs. The option of  table by table explanation and discussion had been taken.. Where appropriate, results had been discussed. Only the summary points need be discussed further. .Summary of the findings revealed that the special offer sales did not affect the sales of CFM and PEM. The assumption that there exist instruments, or secondary predictors, which are correlated with the problematic predictors but not with the error term may not hold. When CFM, PEM GRT AND APP were regressed on the total sales, they were all significant at less than 5%.

A multiplicative scoring factor to produce the total sales forecast with seasonal adjustments from April to December 2016.

The RFM analysis produced a combined RFM score for each customer at the concatenation of the three individual scores, computed as (recencyx100) + (frequencyx10) _ monetary. The recency, frequency, monetary score is 5 5 5 respectively. A customer must score at least 3 points for recency to qualify for customers that are likely to respond to new offer on CFM and PEM.

 

Conclusion and Recommendations

This study had been able to highlight the importance of econometric applicaton in business decision making especially for small and medium scale enterprises. Decision making is a prerogative of the entrepreneur but using the tools would help the effective management of their enterprises. I is important to add a caveat that professionals should be used and that the outcome of decision making tools are to assist entrepreneur only and should not be forced on them.

Notwithstanding, entrepreneurs are advised to avail themselves of the decision making tools from professionals with econometric competences. The present day business environment needs the sixth sense which business decision making tools afford entrepreneurs.

 

APPENDIX: (Al, A2 and A3 for Models 1, 2 and 3 respectively.)

 MODEL 1: TWO-STAGE LEAST SQUARES REGRESSION*

 

Table Al-1 The Lagged Variable Infusion into the Model

Series Name Case Number of Non-Missing No of Valid Creating
Values Cases Function
First Last
1 cfm_1 2 51 50 LAGS(cfm,1)
2 pem_1 2 51 50 LAGS(pem,1)

 

 

Table A1-2 Model Description

Type of Variable
                          Spsales Dependent
                                Cfm Predictor
                                Pem Predictor
Equation 1              Grt predictor & instrumental
                               App predictor & instrumenta  l
                       Logdiscfm instrumental
                      Loadispem instrumental

The model description table gives a summary of the model being fit. Variables specified as predictor will be regressed on the instrumental variables, and the model- estimate4d values will then be used in place of the actual values of these problematic predictors when computing the model for the dependent

 

Table A1-3 Model Summary

                                          Multiple R .995
                                           R Square .990
Equation 1
                             Adjusted R Square .989
                    Std. Error of the Estimate 1.410

The model summary table reports the strength of the relationship between the model and the dependent variable.

Multiple R, the multiple correlation coefficient, is the linear correlation between the observed and model- predicted values of the dependent variable. Its relatively high value indicates a very strong relationship.

R Square, the coefficient of determination, is the squared value of the multiple correlation coefficients. It shows that about 99 percent of the variation in Special offer sales is explained by the model.

Adjusted R Square is an r-squared statistic that is “corrected” for the complexity of the model, and is useful for comparing competing models. The larger values of the statistic indicate better models. It showed that the predictors were able to explain 98.9% of the variations in special offer sales.

Std. Error of the Estimate is the standard error in estimates of Special offer sales based on the model. The value this standard error of estimate compare to the standard deviation of 13.649 of Special offer sales shows that the model has reduced the uncertainty in the “best guess” for next month’s sales.

TableA1-4 ANOVA

Sum of Squares Df Mean Square F Sig
Regression 9221.545 4 2305.386 1160.333 0.000
Equation 1 Residual 91.394 46 1.987
Total 9312.940 50

The ANOVA table tests the acceptability of the model from a statistical perspective

The Regression row displays information about the variation accounted for by the model

The Residual row displays information about the variation that is not accounted for by the model

The regression sum of squares is considerably higher than the residual sum of squares, which indicates that about most of the variation in Special offer sales is explained by the model

The significance value of the F statistic is less than 0.05, which means that the variation that is explained by the model is not simply due to chance.

While the ANOVA table is a useful test of the model’s ability to explain any variation in the dependent variable, it does not directly address the strength of that relationship

Table A1-5 Coefficients

Unstandardized Coefficients Beta T Sig.
B Std. Error
(Constant) -3.237 8.202 -.395 .695
cfm .032 .202 .039 .158 .875
Equation 1 pern .026 .030 .036 .857 .396
grt .916 .146 .484 6.276 .000
app .988 .254 .481 3.892 .000

Dependent Variable  Special Sales

This table shows the coefficients of the regression line. It states that the expected special offer sales is equal to : -3.237 + 0.032CFM + 0.026PEM = 0.916GRT + 0.988APP.

The significance value for GRT and APP are less than 0.05, indicating that the effect of GRT and APP distinguishable from sales of CFM and PEM. In other words, the sales of GRT and APP may not be affecting sales from CFM and PEM. When CFM, PEM GRT and APP are regressed against Total Sales, in Table A1-6, all the variables are significant at 1% level. This shows that they all contribute and no “special sales’’ is limiting the sales of CFM and PEM.

 

 Table A1-6 Coefficientsa
Model Unstandardized Coefficients Standardized Coefficients T Sig.
B Std. Error Beta
1 (Constant) -.193 3.180 -.061 .952
cfm 1.054 .053 .404 19.745 .000
pem .991 .036 .437 27.209 .000
grt .812 .173 .133 4.702 .000
app 1.000 .176 .151 5.683 .000
a. Dependent Variable: tsales

Table A1-7 Coefficient Correlation

cfm Pem Grt App
Equation 1 Correlations cfm 1.000
pem -.866 1.000
grt -.885 .661 1.000
app -.966 .843 .757 1.000

*SPSS 21 OUTPUT

 

 

MODEL 2: FORECASTING MODELS**

TableA2-1 Sales Forecast and Seasonally Adjusted Trend

S/N p MONTHS Tsalesf (rounded to a whole number) N’000 Multiplicative scoring Factor(𝒇i) TsalesfSA (rounded to a vhole number) N’000
1 0.89
2 0.99
3 51 MARCH 1.02
4 52 APRIL 269 1.05 282
5 53 MAY 273 1.06 289
6 54 JUNE 275 1.02 281
7 55 JULY 278 1.02 284
8 56 AUGUST 280 1.01 283
9 57 SEPTEMBER 283 1.05 297
10 58 OCTOBER 286 1.02 292
11 59 NOVEMBER 288 0.96 276
12 60 DECEMBER 291 0.91 265

The model equation is: tsalesF = 93.79 + 0.341tsalesf (-1) + 1.64pm +[(AR(1)= 0.233]

  For example April Forecast: tsalesF 52 = 93.79 + 0.341tsalesF51 + 1.64(Mouth 52) + 0.233      (See Eq 4)

                                                     = 93.79 + 0.341(263.3 = March Total Sales) + 1.64(52) + 0.233

                                                     = 93.79 + 89.78 + 85.28 + 0.233

                                                     = 269.09

                                    tsalesF53  = 93.79 + 0.341(269.08) + 1.64(53) + 0.233

                                                    = 272.7  e.t.c

Table A2-2     Augumented Dickey-Fuller Unit Root Test
Test Statistics Coefficients
Variable Intercept Intercept &Trend None Intercept Intercept& Trend None
At Level Tsalesf(Total Sales Forecast) -1.53 -4.152 1.183 -0.065 -0.577 0.0095
1st Diff -6.59 -6.557 -6.385 -0.982 -0.989-1 -0.939
2ndDiff -7.34 -7.253 -7.426 -1.835 -1.835 -1.834
Critical

Value

1%    -3.59

 

5%    -2.93

 

10%    -2.6

1%     – 4.18

 

5%     -3.51

 

10%    -3.19

1%        -2.62

 

5%        -1.95

 

10%       -1.61

TsalesF is stationary at First Difference in Table A2-2 above. This satisfies its expected forecasting property.

   Table A2-3 Model is Stationary at the First and Second difference:

 

 

 

AR Root(s)

  0.233330  0.233330
 No root lies outside the unit circle.
 ARMA model is stationary.
MA Root(s) Modulus Cycle
 -0.236191  0.236191
 No root lies outside the unit circle.
 ARMA model is invertible.

 

 

      **E-View 7.1 OUTPUT

MODEL 3: THE RFM ANALISIS

 

The Firm is having a total of about 90 customers on records. A random sample of 14 customers were made using Random Numbers, but customer ID-01 was purposive. RFM Analysis was limited to customers who made purchases between September 2014 and March 2016.

Table A3-1 Transaction Data

ID DATE AMT N’000 ID DATA AMT N ‘000
01 08/24/2014 212 16 03/15/2015 15
16 09/09/2014 50 83 03/20/2015 105
75 09/30/2014 12 30 04/12/2015 15
75 10/15/2014 15 01 04/14/2015 50
39 10/20/2014 50 72 04/15/2015 30
72 11/15/2014 15 35 05/18/2015 15
28 11/20/2014 50 16 05/25/2015 12
24 11/22/2014 12 89 07/16/2015 45
28 11/25/2014 12 16 07/18/2015 45
24 12/03/2014 15 49 07/25/2015 195
28 12/14/2014 15 08 08/10/2015 12
77 12/18/2014 15 01 18/18/2015 15
16 12/20/2014 15 72 08/20/2015 30
75 12/22/2014 12 35 08/22/2015 30
77 01/15/2015 15 81 09/14/2015 27
02 01/19/2015 62 08 10/20/2015 12
02 03/04/2015 69 08 01/13/2016 12
30 03/10/2015 15 89 01/20/2016 12

When you compute RFM scores from transaction data, a new dataset is created that includes the new RFM scores.

By default, the dataset includes the following information for each customer:

  • Customer ID variable(s)
  • Date of most recent transaction
  • Total number of transactions
  • Summary transaction amount (the default is total)
  • Recency, Frequency, Monetary, and combined RFM scores

The new dataset contains only one row (record) for each customer. The original transaction data has been aggregated by values of the customer identifier variables. The identifier variables are always included in the new dataset; otherwise you would have no way of matching the RFM scores to the customers.

 

Table A3-2 Customers with Transaction Summaries

ID DATE AMT N’000 ID DATE AMT N’000
01 08/24/2014 12 30 10/202014 50
01 04/14/2015 50 30 03/10/2015 15
01 08/18/2015 15 30 04/12/2015 15
02 01/19/2015 62 35 05/18/2015 15
02 03/04/2015 69 35 08/22/2015 30
08 08/10/2015 12 49 07/25/2015 195
08 10/20/2015 12 72 11/15/2014 15
08 01/13/2016 12 72 04/15/2015 30
16 09/09/2014 50 72 08/20/2015 30
16 12/20/2014 15 75 09/30/2014 12
16 03/15/2015 15 75 10/15/2014 15
16 05/25/2015 12 75 22/12/2014 12
16 07/18/2015 45 77 12/18/2014 15
24 11/22/2014 12 77 01/15/2015 15
24 12/03/2014 15 81 09/14/2015 27
28 11/20/2014 50 83 03/20/2015 105
28 11/25/2014 12 89 07/16/2015 45
28 12/14/2014 15 89 01/20/2016 12

The dataset must contain variables that contain the following information:

  • A variable or combination of variables that identify each case (customer).
  • A variable with the date of each transaction.
  • A variable with the monetary value of each transaction.

Table A3-3 RFM Analysis from Transaction Data -The New Data Set.

ID Date Most Recent Transa

ction

Counts

Amount

N’000

Recency Frequency Monetary

Score

RFM

Score

Customer with High Response to an Offer
01 08/18/2015 3 77 3 3 3 333 ü
02 03/04/2015 2 131 2 2 5 225
08 01/13/2016 3 36 5 3 1 531 ü
16 07/18/2015 5 137 3 5 5 355 ü
24 12/03/2014 2 27 1 2 1 121
28 12/12/2014 3 77 1 3 3 133
30 04/12/2015 3 80 2 3 3 233
35 08/22/2015 2 45 3 2 2 322 ü
49 07/25/2015 1 105 3 1 5 315 ü
72 08/20/2015 3 75 3 3 3 333 ü
75 12/22/2014 3 39 1 3 1 131
77 01/14/2015 2 30 2 2 1 221
81 09/14/2015 1 27 4 1 1 411 ü
83 03/20/2015 1 105 2 1 4 214
89 01/20/2016 2 57 5 2 2 522 ü

The combined RFM score for each customer is simply the concatenation of the three individual scores, computed as: (recency x 100) + (frequency x 10) + monetary.

For example, the RFM score for ID-16 is (3×100 + 5×10 + 5) = 300 + 50 + 5 = 355

The marked customers are those that are likely to respond to new offer on CFM and PEM. To qualify, a customer must score at least 3 points for recency.

 

References

Aghalino, S.O. (2000). British Colonial Policies and the Oil Palm Industry in the Niger Delta

        Region of Nigeria, 1900- 1960. African Study Monographs 21(1): 19-33.

Akangbe, J. A., Adesiji, G. B., Fakayode, S. B. and Aderibigbe, Y. 0. (2011). Towards Palm Oil

      Self-sufficiency in Nigeria: Constraints and Training needs Nexus of Palm Oil Extractors. J. Hum. Ecol. 33(2): 139-145.

Echebiri R.N. and Edeba, M.E.I. (2008) Production and Utilisation of Cassava in Nigeria:

      prospects for Food   Security and Infant Nutrition; PAT:Production Agriculture and

     Technology; 4 (1); 38-52

Embrandiri, A., Singh, R.P., Ibrahim, H. M. and Ramli, A.A.(2011). Land application of

      biomass residue generated from palm oil processing: its potential benefits and threats.

     Springer Science. Environmentalist, DOI 10.1007/s 10669-011-9367-0.

Gelb, A.; Jefferson, G. and Singh, I. (1993) Can Communist Economies Transform                         Incrementally? China’s Experiwnce: World Bank Working Papers WP 1189, October.

Levine, D.M.; Stephan, D.; Krehbiel, T.C. AND Berenson, M.L.(2005) Statistics for Managers,

      Upper-Saddle River, NJ. Pearson Educational Inc.

NBS (National Bureau of Statistics)(2015) Growth of Gross Domestic Product in Nigeria .

       Annual Abstract of Statistics. Federal Republic of Nigeria.

Odoemenem, I.U.,and Otanma, L.B. (2011) Economic Analysis of Cassava Production in Benue

        State, Nigeria. Current Research Journal of Social Sciences; 3(5), 406-411

Omereji, G.O. (2005) The Oil Palm Industry in Nigeria: cultivation, Processing and Trade.

       Mindex  Publishers, Benin City. Pp. 131-156

Cassava Production in Nigeria (n.d.) in Wikipedia. Reviewed April 10, 2016 from

          http://en.wikipedia.org/wiki/cassava production

Yakasi, M.T. (2010) Economic Contribution of Cassava Production (A Case Study of Kuje Area

        Council, Federal Capital Territory , Abuja Nigeria; Bayero Journal of Pure and Applied

       Sciences; 3 (1) ; 215-219.

 

Advertisements

Your views and comments are most welcome

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s