To characterize any standard market, generally, we need a model consisting of at least two equations, one of which describes the pattern of demand for the relevant goods or resources, Q^D, and the other the pattern of supply, Q^S. In the simplest case, when all the factors, except price, affecting demand and supply are fixed, and the market is standard, we have the following version of the model:
Q^D=D(P), Q^S=S(P), D^' (P)<0, S^' (P)>0 ,
where D(P) and S(P) functions of demand and supply depend on price. When the specific type of these functions is known, it is easy to determine the equilibrium point of the market (if it exists) using the following equation: D(P)=S(P).
As a rule, in most cases, specific types of D(P) and S(P) functions are unknown, and we have to resort to econometric modeling to estimate them. To accomplish this, we must 1) transform the above-mentioned system into a model comprising of regression equations and 2) obtain Q_i^D, Q_i^S, P_i, i=1,2,⋯,n, data for the variables Q^D, Q^S and P included in the model.
The first requirement, transformation into a regression system, is easy to accomplish by including the random error terms, ε^D and ε^S, with the expected value equal to zero, into the corresponding equations: Q^D=D(P)+ε^D, Q^S=S(P)+ε^S. The second requirement—obtaining the observation data for Q^D, Q^S, and P variables—is much harder to fulfill. The difficulty is that in market model analysis, the separate Q_i^D and Q_i^S values of supply and demand are rarely available, except when research is specifically done to examine market subjects. As a rule, the existing statistical system gives information on price values together with the variable Q-–the volume of actual sales, but not what the demand Q_i^D, and the supply Q_i^Swere under different price conditions. In such cases, if no additional information is available, the demand and supply functions of the market model are considered unidentifiable. The widely used standard approach to solving this problem involves changing the specification of the model by including additional explanatory variables in the equations and making an assumption that observed Q values, together with their respective prices, correspond to the point of market equilibrium, i.e., we have the following equations: Q_i^D=Q_i^S=Q_i, i=1,2,⋯,n.
This article proposes a different method to estimate the market model consisting of regression equations. It takes into account the specifics of the market mechanism and makes it possible to fully or partially identify the functions of supply and demand, even when the only data available is on the price and the volume of sales. Unlike the traditional approach, the proposed method is based on the assumption that not every pair of price and sales volume included in the data sample Q_i, P_i, i=1,2,⋯,n, corresponds to the point of market equilibrium.
The estimation of equations of this model involves a two-step process. First, using the sample of data on price and sales volume, the sales line is estimated, and the point of its maximum, which corresponds to the point of market equilibrium, is defined. Then, given the location of the equilibrium point, using a particular technique, subsamples that define demand and supply are identified from the total sample, and their respective equations are constructed.
The article demonstrates how to apply this method, using the Georgian labor market as an example. The employment line equation estimated using quarterly data on real wages and employment in 2010-2021 showed that the labor market in Georgia was in equilibrium in the third quarter of 2013. On the basis of this information, we used data from the first quarter of 2010 through the third quarter of 2013 to estimate the labor supply function, and the rest of the data was applied to estimate the labor demand function. Based on these estimates, a 100 GEL increase in real wages in Georgia during the analyzed period caused a 5.2% increase in labor supply and a 5.6% decrease in demand for labor.

Keywords: Market model, supply function, demand function, sales line, workforce employment line, equilibrium point.
JEL Codes: D47, D82, J21

References:

• Wooldridge J. (2012). Introductory Econometrics: A Modern Approach. Cengage Learning, 5th Edition, 912 p.
• Greene W. (2012). Econometric Analysis. 7th Edition, Prentice Hal.
• Kane. E. (1977.) Ekonomicheskaya statistika i ekonometriya, vvedeni v kolichestvennyy ekonomicheskiy analiz. [Economic Statistics and Econometrics. An Introduction to Quantitative Economics. Moscow, Statistika, 1st and 2nd Edition.] in Russian.
• Kozletskaia. T. A. (2014). Krivaya prodazh tovara kak funktsiya pokupok mnogimi potrebitelyami. [The Sales Curve as a Function of Purchases Made by Many Consumers Scientific and Technical Statements of the St. Petersburg State Polytechnic University. Economic Sciences, No. 6(209)] in Russian.
• Magnus J. R., Katyshev P. K., Peresetsky A. A. (2005). Ekonometrika. Nachalnyy kurs. [Econometrics: a First Course. Delo, p.504.] in Russian.
• Varshavsky A.E., Kochetkova. E. (2018). Modelirovanie pokazatelej sprosa i predlozhenija inzhenerno-tehnicheskih specialistov. [Analyzing and Modeling the Demand and Supply Indicators for Engineers and Technical Specialists. Economic Analysis: Theory and practice, vol.17, pp. 886-905 ]. in Russian.
• https://www.geostat.ge/ka/modules/categories/683/dasakmeba-umushevroba
• Ananiashvili I. (2018). Dependence of Aggregate Supply on the Main Factors Prices of Production. Bulletin of the Georgian National Academy of sciences. Vol. 12, no. 1