Expanded summary

This article discusses findings of the empirical research of the five prominent Investments models on the example of the USA, conducted by R. Kopke. Moreover, on the example of Georgian economy I statistical adequacy of the above mentioned models are empirically tested and the results are compared to outcomes of R. Kopke’s study.

The paper reveals the tendency of peaks and falls of the factor variables coefficients and prevalence of capital adjustment rate over capital depreciation rate, which are similar to the empirically tested accelerator models on the basis of the USA’s and Georgia’s economies. Major difference between empirical models is the existence of autocorrelation in residual members. According to the model of R. Kopke, autocorrelation is observed in the residual members, while the autocorrelation is not detected in the empirical model based on Georgia’s economy. The latter adequately reflects the country’s economic reality.

On the basis of Georgian economy example, Paper reveals the seasonal component in time series analysis of the investments. In order to increase statistical significance, the research needs to exclude seasonality from the original data. After seasonal adjustment, the timeline series are checked about stationarity by Dickey-Fuller test.  Investments belongs I(1) type, because using first order differences leads to the stationarity.

Akaike and Schwarz Criteria is Applied in order to define investments autoregressive model type on the example of Georgia. The highest values of determination and corrected determination (0.31; 0.25) are observed in ARIMA (2.1.2). Additionally, the model F>F_Cr. and all variable’s coefficients are statistical significant except constant member’s. By excluding it from the model, determination coefficient value decreases from 0.31 to 0.3, when coefficient of corrected determination minimally increases; this indicates exclusion of constant variable from the model.

Breush-Godfrey null hypothesis is accepted with 99% confidence interval, which means non-existence of autocorrelation.

If we continue inserting lag variables of autoregressive and moving average, every added lag reduces values of corrected determination coefficient. Also, probability of acceptance null hypothesis of Breush-Godfrey’s test decreases. Coefficients are becoming statistically insignificant, which indicates the deterioration of the results.

Paper reveals that empirically verified autoregressive models on the example of two countries similarly result in statistically non-significant result. The difference between the models is the existence of autocorrelation in residual members. Difference results about existence of autocorrelation in residual members might be caused by using different methods in empirical realization process (making time series stationary on example of Georgia’s economy), also by the difference in countries’ economics.

Autocorrelation is not found in residual members in empirically estimated Investments cash flow model on the example of Georgia’s economy and total capital’s coefficient is not equal to zero unlike R. Kopke’s study. Factorial variable coefficients create a saw-like tendency and determination level of the model is low.

It should be mentioned that empirical realization of Tobin’s investments model on Georgian economy case faces following problems:  1) there is no information about Tobin’s coefficient – q; 2) there is no empirical data on total capital. The later can be fixed by using the result from investments accelerator model, which estimates total capital values. However, because of the absence of Tobin’s coefficient values statistically valid model cannot be built on Georgian case.

Calculation of Tobin’s coefficient or replacing it with economically similar factor is not possible, because there are no economic indicators that could be used for fixing stated problem. Also, considering that technical estimation problem in the model should be fixed by assuming that Tobin’s coefficient is constant through time and creating new coefficients  has no valid economic logic, because above mentioned fact confronts Tobin’s theory.

Like the research made with Tobin’s model on USA’s example, using it on Georgia’s example is statistically incorrect. Additionally, we get different results by using Neoclassical models. On USA’s example model has statistically useful values, when on Georgia’s example statistically non reliable results is received.