The most important preliminary phase or study that precedes the evaluation of an investment project is the analysis and projection of demand. Hence it follows that its study is vital to justify a possible investment since it is closely interrelated with the basic objective of the project, with the determination of its size and with the production assortment.
It is necessary to point out that "the basic or primary objective of any investment is the production of goods and / or services intended to satisfy a certain need, which is expressed in a given demand". Therefore, logically, if there is no such demand, the investment would not be justified.
This that has been mentioned may seem repetitive, and many who have attended economics courses could classify it as elementary and evident when evaluating investment projects; however, practice and experience have shown that the Cuban company suffers today, and still has a pending subject related to the use of quantitative tools, which go beyond simple subjectivity, and that in combination with qualitative elements They offer scientific rigor and much greater precision when making forecasts linked to demand studies. Many times due to lack of knowledge, others due to time constraints,The truth is that almost none of the many quantitative methods that exist and for which others worked hard in the past are used to provide us with tools that allow us to have a minimum margin of error.
The objective of this article is then to put into practice two of these quantitative and qualitative methods in combination, to carry out the projection of the demand for tests by Laboratorio de Cubacontrol SA as part of a study of financial and social evaluation of investments. The methods used will also be briefly described.
DEVELOPING
As pointed out in the introduction, there are different methods to carry out studies related to the projection of demand. Choosing one or the other will depend on the situation you are in, the data you have, and obviously the type of product you want to carry out the study. They are not the same methods in the case of consumer goods, as in the case of intermediate or capital goods. This article will not go into detail about other methods that are not related to the case study, that is, to the demand for tests from the Laboratorio de Cubacontrol SA, and therefore, the first thing the reader should know, to become familiar with the type of study, is the activity of the Laboratory, and what product does it offer.
Brief characterization of laboratory activity.
The Cubacontrol SA Quality Supervision Laboratory is a dependency of Servicios Internacionales de Supervisión Cubacontrol SA It provides independent laboratory services to foreign firms based in Cuba or abroad, and to Cuban entities that request it.
The laboratory assumes the execution of the commercial operations of quality supervision related to the shipments of imported or exported products, both in Cuba and abroad at the orders of its clients, and for this it uses its own specialists, technicians or contracted representatives from other institutions abroad.
Basically it has two main areas.
Food area, with 5 laboratories.
- Oils, fats and dairy products. Meats, fruits and vegetables. Flours, grains, spices and condiments. Microbiology. Sensory
Chemical area, with 4 laboratories.
- Chemicals and fertilizers. Drinks and alcohols. Sugar. Atomic Absorption Spectrometry.
In these laboratories certain types of tests are carried out associated with the type of product in question and, of course, depending on the customer's requirements. In general, the client makes an analytical service request (SSA) that can, depending on the service, carry a certain number of samples associated, which is specified to the clients through a contract. Furthermore, within the same analytical service request, as many tests as required can be performed. In general terms, there is no specific relationship between the number of tests carried out and the number of requests for analytical services, but in 2006, the laboratory, on average, carried out 13 tests per service request, which does not mean that exactly that relationship has to be. Generally,the relationship is that for each analytical service, at least one test is carried out.
customers
In 2006, if an analysis is made of the quantity of services demanded, it can be seen that there is a group of clients that, due to their volume, represented 85% of the total demand for laboratory services, and that even in term of income, they concentrate almost 95% of the company's income. Therefore, in the study of demand, the Pareto criterion will be applied, and the behavior of all other clients will be conditioned by the behavior of the main clients, which are:
Alimport, Propes and CubaExport. Furthermore, it will be assumed that this structure will not vary significantly during the 10 years for which the demand is to be projected.
Theoretical synthesis of the most important elements that must be mastered to carry out a demand study.
There are basically two types of demand. The potential demand that is the quantity of a good or service that consumers would be willing to acquire based on their purchasing power, therefore, the maximum that theoretically could be acquired from a specific good or service in a given market and time.
Of course, this is a fairly general definition that, if applied to the company analyzed, would be understood as the demand for tests by all companies that are willing to contract the services provided by Cubacontrol SA
In addition to the previous concept, the effective demand can be analyzed . This is generally defined as the quantity of a good or service that consumers would acquire based on the price level in a given market and given time. However, if we add to this the effect that commercial effort, income level and government policies could have, then the effective demand would be the volume and structure of goods and services required in a given period of time, in depending on their prices, in a market with a potential for purchasing, commercial effort and given political and social situation.
In the specific case of the company Cubacontrol SA, its potential demand would be the quality supervision tests required by all importing and exporting companies in the country, since there are several laws related to the activity of Cuban Foreign Trade that oblige companies to monitor quality.
However, since not all Cuban companies comply with what is established, effective demand is considerably less than potential.
Already entering the use of quantitative tools, as mentioned, there are many that more or less accurately, and more or less scientific, provide estimates of the levels to be demanded by customers. In the specific case of this article, there will be the help of Econometrics, which explains through models the approaches that arise from economic theory, models that can be presented and verified through statistics, in order to make the corresponding forecasts.
A model is the simplified representation of reality, and it is expressed through a fundamental tool in Econometrics that is regression analysis. Because most problems involve more than one related variable, you generally want to achieve a functional expression that expresses that relationship. Therefore, the regression methods are used to determine the best functional relationship between the variables under study.
"If a variable" y "is expressed: called the dependent variable, only as a function of an explanatory variable" x "will one be in the presence of simple regression, and will it be multiple regression if there is more than one independent variable in the relationship,"
Therefore the past is going to be studied to project into the future. There are different types of econometric models that are used: linear models, in terms of parameters and variables, and nonlinear models that can be transformed into linear, such as the Cobb-Douglas models, the exponentials, and the logarithmic ones. In practice, and according to many specialists, after years of discussions, it has been concluded that linear models largely explain most of the problems present in the economy, and that other models, although they add statistical complications, do not they are able to forecast more accurately than linear ones. In this study we will work with linear models then.
The information obtained can be defined as time series, since it is stored over a period of time. In the specific case of the company, said information is related to the tests carried out by the laboratory on the different companies, and data on their imports or exports, collected at regular (annual) periods of time.
As noted before, linear regression is the type of econometric model to be used in this study, and presents a general linear model as follows:
And i = β 1 + β 2 x 2i + β 3 x 3i +… + β k x ki + u i
In this case, the variable Y, is called the endogenous, dependent variable, and Y t, denotes its value at time t, t = 1, 2, T, compared to the variables, x 2t, x 3t, x kt They are called exogenous variables, while they denote the magnitude of the effect that the variables x 2t, x 3t, x kt have on Y t. The u i they are called random disturbances, which collect everything that is not in the model, and that in one way or another affect the dependent variable (Y), but that for one reason or another could not be in the model.
Then the process by which the values of the parameters will be obtained, based on the sample information, is called the Ordinary Least Squares Method, which minimizes the differences between each value of Y i and its estimated value, which are the estimation errors.. For its part, the process by which the regression equation is estimated is called Curve Fit. The statistical package to work with is Eviews VERSION 4.1.
For a model to be valid, the first thing that must be met is that the marginal probability is less than the level of significance assumed by the person carrying out the study (α), which is apparent from the level of reliability with which it is working. For example, if you work with 95% reliability, the significance level will be 5%. It is precisely this level that will be used to perform the regressions.
Any general linear model must also meet certain assumptions that will be set out below.
- One of the assumptions of the classic linear regression model states "that there is no exact linear relationship between the X variables", which indicates that there is no multicollinearity between the explanatory variables included in the model.
Multicollinearity is understood as the effect produced as a consequence of a strong interrelation between two independent variables.If the R 2 is high, and the t tests are not significant, the existence of multicollinearity is suspected, which is reaffirmed with the correlation matrix.
- Normality Assumption Ui ~ N (0; σi). Random disturbances tend to a normal distribution. Normality is detected with the test statistic called Jarque Bera provided by the Eviews.Autocorrelation Assumption. Random shocks cannot be correlated. Autocorrelation can be defined as the correlation that exists between the members of an ordered series in time, or in space.There are different tests to detect it, such as the Durbin-Watson, and the Breusch-Godfrey. As it is a much more powerful test, we will work with the second one, which will be analyzed with the statistical package. Homocedasticity Assumption. The disturbances must be homoscedastic, that is, they all have the same variance. Heteroskedasticity is generally said to exist when the variance of the residuals is not constant. There are different tests to detect it, in this case the statistical package to be used has the White test.
Once all the assumptions have been fulfilled, it is accepted that the model with which we are working is valid for making the forecasts.
Analysis and Projection of the Demand for Laboratory Tests.
Based on the fact that this study was part of the evaluation of an investment project that was in the first phase of its achievement, that is, a primary stage where what is wanted is to determine the viability of the investment, it became necessary the realization of a previous exploration that consisted of the aggregation of all the existing information and estimates consulted with the laboratory specialists. It is necessary to clarify that there were no previous studies to consult, since until now the similar investments made in the laboratory did not have an analysis of such magnitude, and were carried out due to the need of the country to have the equipment, more than supported by economic-financial criteria. This caused all estimates to start from zero. Of course,there is no market study by specialists that can estimate the demand for laboratory services in the future, however, based on data from previous years regarding the number of tests carried out and the level of imports and exports of the food sector In the country, it is possible to determine whether or not there is a correlation between these, and by means of a regression analysis, to be able to project the future, to estimate demand, which, as has been said, is the most important study of the previous phase.it is possible to determine whether or not there is a correlation between these, and by means of a regression analysis, to be able to project the future, to estimate the demand, which, as has been said, is the most important study of the previous phase.it is possible to determine whether or not there is a correlation between these, and by means of a regression analysis, to be able to project the future, to estimate the demand, which, as has been said, is the most important study of the previous phase.
For the calculation of the regressions, the food sector will be taken, since the activity of the laboratory, at the discretion of experts, is 90% belonging to that sector, for this reason, as with clients, the Pareto criterion will apply for the activity.
Following a logical order of proceeding, a study of the historical background of the demand in question should then be carried out in order to have qualitative elements that constitute a strong basis for the future of the activity to which the demand projection is being made. For reasons of space, this analysis of the historical behavior of demand will not be present in this work, but it could be consulted in the aforementioned diploma thesis, so the methodology used will be explained directly.
Projection of demand.
In order to forecast the tests that the laboratory would have in the future, it was initially considered that, having a database of more than twenty years, it should be more convenient to carry out a simple regression analysis of the tests against time (the latter factor as an independent variable). Starting from 1984 and until 2006, and using a confidence interval of 95%, or what is the same, a significance level α = 5%, it was verified that the model used was valid, since it had a lower marginal probability of the significance level used. However, if the slope of the independent variable is observed, it is verified that it is negative, making the test demand equation decreasing and, of course, in the future, below zero.
It is evident that any analysis related to the tests and the time, for the entire period analyzed, is not valid for the projection of the demand of the laboratory of Cubacontrol SA, especially since if the historical antecedent of the demand is analyzed, it will be verified that there is what is usually known as a change in structure, reflecting past behavior that would skew information for the future.
Therefore, the regression analysis to be used to forecast the company's demand for trials should take into account another variable. In this sense, given that the main clients of the laboratory are engaged in exporting and / or importing food, it would be logical to suppose that if the levels of both export and import of them increased, then the demand for laboratory associated with such levels of activity, and there may therefore be a correlation between the tests, taken as the dependent variable, and exports or imports and time, as independent variables.
In order to verify the veracity or not of this assumption, a multiple regression analysis was performed between the aforementioned variables. In order to access the activity levels of the three main clients, ALIMPORT, CubaExport and Propes, the Statistics Department of the Ministry of Foreign Trade was contacted, and the data was subsequently adjusted in consultation with experts from the companies themselves, where it was obtained information, which, because it was classified, was already provided multiplied by a factor, but which, while maintaining its trend, is valid for this study. It is necessary to clarify that the data obtained in the ministry are only for the period 2000-2006, and the data from the tests carried out on the three most important clients have only existed since 2000,since it was from that moment that the laboratory began to enter them in a database.
This fact constitutes a limitation for the regression projections. When performing the multiple regression for the Alimport client, taking as a dependent variable its demand for tests, and its imports and time as independent variables, the model was valid, with a high determination coefficient (R 2), and some t tests. not significant, reason why it was suspected of the existence of multicollinearity reaffirmed with the correlation matrix. To remedy this situation, a remedial measure was applied, which consisted of performing simple auxiliary regressions for each independent variable, import and time.. When performing these regressions, it was verified that the valid model resulted from the regression tests against imports, therefore a simple regression was performed between imports and time, which turned out to be a valid model, with a high coefficient of determination, with which imports were forecast.
With this forecast, imports were introduced into the simple regression model, tests against imports, and the demand for Alimport tests for the laboratory was forecast until 2016.
A similar methodology was applied for the CubaExport company, performing a multiple regression between trials and exports and time (exports, since it is an exporting company), resulting in a valid model, with a high R 2, but with the non-significant t tests, applying the same remedial measure, auxiliary simple regressions, and contrary to what happened with Alimport, it turned out that the model tests against timeIt was the valid model, so the trials were then forecast until 2016, taking time only as an independent variable, in a simple regression, which also presented a high R 2.
In the case of the client Propes, the multiple regression was carried out, but in this case, taking the tests again as a dependent variable and imports and time as independent, being an importing company. The model turned out to be valid and of a high R 2, however, the t testsThey were again not significant, so again the existence of multicollinearity was suspected, but also when checking if all the other assumptions of the model were met, as had been done for the other two companies, the presence of autocorrelation was detected, therefore, working with this model could be overestimating the value of the determination coefficient, and therefore the goodness of fit. For this reason, auxiliary regressions similar to those performed with Alimport were carried out, with the result that the test against imports model reflected the trend more, since it had a higher coefficient of determination (R 2).
A simple regression between imports and time was then carried out, for the client Propes, which turned out to be a valid model, with a high R 2, with which the imports were forecast, which were later introduced in another model of simple regression trials against imports, with which the trials were finally predicted.
Based on the fact that the predicted trials for the three fundamental clients were already available for the period to be analyzed, that is, until 2016, and bearing in mind that in 2006, the three clients together accounted for 73% of the All the tests, and assuming that this structure would be maintained throughout the period, it is possible to determine the total number of tests that the laboratory would have until 2016. The following table represents them.
Table No. 1. Demand for projected trials
| DEMAND FOR PROJECTED TRIALS. | ||||||||||
| 2007 | 2008 | 2009 | 2010 | 2011 | 2012 | 2013 | 2014 | 2015 | 2016 | |
| ALIMPORT | 18639 | 21547 | 24457 | 27367 | 30277 | 33187 | 36097 | 39007 | 41917 | 44827 |
| CUBAEXPORT | 895 | 985 | 1076 | 1167 | 1258 | 1349 | 1440 | 1531 | 1622 | 1713 |
| PROPES | 837 | 971 | 1105 | 1239 | 1373 | 1507 | 1642 | 1776 | 1910 | 2044 |
| TOTAL | 20371 | 23503 | 26638 | 29773 | 32908 | 36043 | 39179 | 42314 | 45449 | 48584 |
| TOTAL TESTS | 27813 | 32089 | 36370 | 40650 | 44930 | 49211 | 53492 | 57773 | 62053 | 66333 |
Demand projection considering structural changes.
As has been mentioned, this demand projection study is a previous phase of a financial and social analysis that was carried out in the Cuba control laboratory, so it was also necessary to consider the new demand associated with the new investments to be made there.
The new investments will allow, in addition to replacing equipment that becomes obsolete, to carry out new types of tests, which, although required, did not have the necessary equipment to carry them out, which is why it is necessary to add to the historical demand previously projected, the estimate of the new tests. that will be sued in the period. In the opinion of experts, the increase in demand will have its greatest growth in 2008, due to the start-up of a team of which it is already known from experience that the tests it performs will have a significant demand.. To this end, and given the scarcity of statistical records, expert judgment was used using the Delphi method., obtaining the following results as a percentage of the projected historical data.
Table No. 2. Projected test demand considering structural changes (in percent, taking the projected test demand as 100%).
| 2008 | 2009 | 2010 | 2011 | 2012 | 2013 | 2014 | 2015 | 2016 |
| 115.00% | 120.00% | 125.00% | 129.00% | 132.00% | 135.00% | 137.00% | 139.00% | 140.00% |
Therefore, the total demand adjusted for structural changes, when considering the new tests would amount to:
Table No. 3 14. Demand for projected tests considering structural changes (expressed in number of tests).
| 2008 | 2009 | 2010 | 2011 | 2012 | 2013 | 2014 | 2015 | 2016 |
| 36,903 | 43,644 | 50,813 | 57,960 | 64,958 | 72,215 | 79,149 | 86,254 | 92,867 |
In this way, using two tools, quantitative and qualitative, in combination, it is possible to make a projection of demand with a greater scientific basis. The machine outputs that contain all the equations and hypothesis tests and graphs can be consulted in the annexes of the aforementioned diploma thesis.
FINAL CONSIDERATIONS
- Demand analysis and projections constitute the most important preliminary phase to an evaluation of the profitability of an investment, since they largely determine the capacities to be installed, as well as the production assortment and are closely linked to the basic objective of the project. It is evident that, to a large extent, the profitability or not of the project will be given by a good projection of demand, and the methods used must have a solid foundation and provide the minimum margin of error. The Cuban company in general suffers of the use of these tools and projects, to a large extent, with a subjective nature, which has a greater margin of error, and where in conditions of scarce resources, financing must be used in investments that have a solid demand that leads to profitable projects.The use of one method or another will depend on the context in which the study is located, but the combination of quantitative and qualitative techniques that promote a product of scientific rigor, which constitutes a good analysis and projection of demand, is highly recommended.
BIBLIOGRAPHY
- Statistical Yearbook of Cuba, 1980, 1984, 1989, 1994, 1999, 2004, 2006.Baca Urbina, Gabriel. "Project evaluation". McGraw-Hill Publishing House. Fourth Edition. México DF, México, 2004.Bridley, RA; Myers, SC "Fundamentals of business financing". Mc Graw-Hill Publishing. Fourth edition. Mexico City, Mexico. 1993.Bueno Campos, Eduardo; Cruz Roche, Ignacio; Durán Herrera, Juan José: Business economics. Analysis of business decisions. Pyramid editions. Twelfth edition, Madrid, 1989, pp.281-289.Durán Herrera, Juan José: Economy and financial management of the company. Ediciones Pirámide, Madrid, 1992, pp. 418-420.Espallargas Ibarra, Daisy. Econometrics for Accounting and Finance. Power Point Presentation, December, 2004. Rodríguez Mesa, Gonzalo ¨ Demand analysis and projection.Text from the Faculty of Economics of the University of Havana, digital format.
Rodríguez Mesa, Gonzalo. The financial and social evaluation of investment projects. Third Edition, May, 2006. Digital format, Faculty of Economics, University of Havana.
Rodríguez Mesa, Gonzalo ¨ Analysis and projection of demand. Text from the Faculty of Economics of the University of Havana, digital format.
For more information on the methods used depending on the type of asset, consult the book already cited in footnote # 1.
For more information, consult the work on the Financial and Social Evaluation diploma of the modernization of the Cuba control SA laboratory by the authors Damián Echevarria and Iran Miranda, Faculty of Economics, 2007, University of Havana.
Espallargas Ibarra, Daisy. Econometrics for Accounting and Finance. Power Point Presentation, December, 2004.
To know the consequences of not complying with this assumption, it is recommended to consult the conferences already mentioned in footnote # 4.
The R 2 or coefficient of determination indicates the extent to which the independent variable explains the behavior of the dependent. It is generally accepted that more than 75% is high.
Ditto 5.
Ditto 5.
Ditto 5.
Own elaboration. Note that the TOTAL TEST box is calculated by dividing the TOTAL box, which is the sum of the three companies, by 0.73.
This equipment is the atomic absorption spectrophotometer, used to perform tests on metallic contaminants.
Three rounds were performed in this method.
Own elaboration. 14 Idem. 13.
Download the original file
