The Solow Growth Model in Saudi Arabia
In this research, the Solow Growth model has been employed to assess the steady state growth SSGR) of Saudi Arabia. An extensive volume of work using the Solow model has been cross-country empirical work and has provided insights into the growth policies. However, studies that are specific to a particular country with respective time series data are limited, and close to none when it comes to Saudi Arabia. Therefore, the objective of this research was to develop a simpler and alternative model that would estimate Saudi’s SSGRs. Notably, this study demonstrated that the Solow exogenous model is extendable and used for estimating the SSGR specific to a country that could be utilized for growth policy. The study illustrated that this could be attained by estimating the SSGR with consideration of benefits from two externalities, specifically the learning by doing (LBD) and the openness of trade (O). The study found that externalities have a substantial effect on the country’s SSGR. Notably, while the effects of these two externalities have demonstrated importance in Saudi Arabia, trade openness played a relatively considerable higher function in the country’s growth. The researcher hopes that this approach and the empirical findings are useful for further research and development of the Solow model. The developments will guide in developing better policies that will improve the nation’s growth rates in the long run and permanently.
Introduction
In a global environment made of numerous countries desperately working hard to get richer, the winners will always be the benchmarking models for the rest of the countries. However, the question of what accounts for that success lingers. Notably, Sarel (1996) has indicated that a consensus tends to bulls towards the numerous policies that ambitious countries implement, affecting the global trade patterns, among other major issues. Notably, the use of endogenous growth models (EGMs) has proven fundamental in answering two fundamental policy questions. The first one is the potential factors that impact the long-run economic growth, specifically The Steady-State Growth Rate (SSGR). The second policy question is whether the policy could improve SSGR. The Solow (1956) model entails the SSGR equaling the total factor productivity (TFP) rate from a policy viewpoint. The assumption is that TFP is exogenous or, rather, is trend-dependent. While an extensive volume of cross-country empirical work deals with ENGMs and provides insights into the growth policies, empirical work specific to a particular country with respective time series data is limited, with many studies having used ad hoc specifications.
It is common for empirical studies to regress the growth rate of output on several variables which empirical investigators believed to be fundamental growth-enhancing factors. Notably, the scope for the arbitrary selection of the respective variables is large, with Durlaeuf et al. (2005) indicating that the list could reach over 100. Easterly et al. (2004) also remarked on the ad hoc nature of specifications within the empirical growth literature and asserted that it had the typical limitations of selecting a specification without precise guidelines from the theoretical frameworks. This will hence lead to more feasible specifications than the data points in the specific sample. Subsequently, there is a risk of the studies overestimating the growth effects since the work based on ad hoc specifications fails to provide a clear difference between the short and long-run growth effects. Additionally, another neglected weakness is that it is very difficult to acknowledge that the dependent variable, which is the SSGR pointed out in many of the empirical works, can sufficiently be proxied to the early growth rate of output or rather its averages over shorter durations of between 3 to 5 years. Conceptually, SSGR will be similar to the natural rate of employment that is specific to a country. The two are long-run equilibrium variables and unobservable. They need to be obtained from estimating the non-steady-state models properly through the imposition of steady-state conditions.
Considering that the estimates of country-specific SSGRs and their determinants are fundamental in policymaking, this paper develops a framework that will estimate the SSGR for Saudi Arabia. While it is principally possible to estimate the country-specific SSGR with an ENGM, the econometric challenges in estimating the nonlinear dynamic equations are substantial. According to Greiner et al. (2005), dealing with some of the econometric challenges entails discarding the scale effect in ENGMs, a fundamental property within the models. Rather, some studies chose to use the plausible a priori value for the computation of effects of policies in SSGR via calibration strategies instead of making estimations of the structural parameters of the ENGMs.
This research will entail developing a simpler and alternative model that will estimate Saudi’s SSGRs. This approach is primarily an extended version of the Solow model. The two reasons for using the model include the absence of clear-cut evidence that the modern demanding ENGMs will provide better observable facts than the simple Solow model. According to Jones (1995), there is no clear-cut evidence to show that the output growth rate will grow proportionally to the expenditure increase in some of the growth improving factors such as investments and research and development. Another study used the United States time-series data to establish that solely investments in non-military equipment and non-military structural investments demonstrated minimal impacts on the long-run growth rate (Kocherlakota & Yi, 1996). The second reason is that this approach has attracted numerous applied economists to get fast insights into policies that could improve the long-run growth rate. This is because of the relative simplicity in making estimates, especially with country-specific time-series information.
This paper has been organized as follows: Section 2 will present the literature review of studies on steady-state growth rate (SSGR), the typical Solow model, and the extended version of the model needed in developing the study’s specifications. Section 3 presents the study’s research methodology, which will point out the approach taken in obtaining the information and analyzing it. Section 4 will present the empirical results and SSGR estimates for Saudi Arabia. Section 5 discusses the estimates of the SSGRs while considering the alternative assumptions on the share of profits and examining any consistency to previous literature. Finally, the conclusion is to be presented in Section 6.
Literature Review
Various theoretical and empirical frameworks have been used in quantifying the steady-state economic growth rate. Considering the neoclassical growth model, the steady-state refers to the equilibrium that the capital in each effective labor unit converges in the long run. The steady-state ends up being unique and globally stable easier due to the model’s set-up that has the concave production function commonly with the labor-augmented technology. Additionally, other optional models that demonstrate multiple equilibria at the steady-state are also present. One of them is the overlapping generations model posited by Samuelson (1958) that demonstrated age heterogeneity covering all the agents. Diamond (1965)incorporated this model class to prove the inefficiency in accumulating capital in the long run. Matsuyama (2004) also used the model in analyzing the existence of two symmetric steady states, specifically a high level of capital accumulation and a low capital level. The extended version of the neoclassical growth model had the steady-state referring to the long-run equilibrium. The agents have no anticipation of the impact of future shocks. Hence, this steady-state could be handled by assuming the small enough shocks and then considering the first-order Taylor approximation of the perfect-foresight path. Specifically, the risk steady state is then the long-run equilibrium in which the state variables remain constant in the presence of expected future shocks and innovations.
For the shocks turning out to be zero, the affiliated method is a numerical strategy that depends on a system of equations established by the second-order Taylor expansion around the expected future variables. This strategy has been closely affiliated with the model deployed by Tille and Wincoop (2010) and Deveraux and Sutherland (2011). The two deployed models focused on solving the dynamic stochastic general equilibrium (DSGE) model with portfolio choice.
The standard element in these models is the exogenous economic growth rate that is given outside the model. The growth rate can be instigated as the deterministic technology progresses on the neoclassical growth model or the stochastic technology progresses on the DSGE model. At the steady-state, the ratio of endogenous variables including output, consumption, and capital over total efficiency units of labor that grow at a similar rate of technological progress. Conversely, the growth rate of output is equivalent to the growth rate in technological progress. Notably, the source of this technology process has not been identified. The in-determination matter hence underlines the challenge of computing the steady-state economic growth rate (SSGR). This literature hence looks into considering the Solow model in the computation of the SSGR.
An important analysis of the Solow (1956) reading, together with the 1970 one, has clarified that the stylized facts that have been established in the model were not to be defined as universal elements that apply to every nation in the world. Conversely, the present literature imposed robust homogeneity assumptions within the cross-country growth process since each nation is presumed to have a similar and specifically the Cobb-Douglas aggregate production function (Durlauf et al., 2001). This is new considering that the modern growth theory has stipulated that countries need to be defined by different aggregate production functions. This is because new causal growth theories will possibly influence the aggregate production function of the countries instead of constituting the additive components within the growth process. To this study, this could imply that for a spearing growth regression, regardless of the benign found in the Solow model or any other theory, there is a need to explicitly account for the heterogeneity parameter.
In this research, there is a provision of estimates of a local or domestic generalization to the Solow growth model. The local term refers to the notion that this model applies to every nation, but the parameters will differ depending on the country. Specifically, the study permits these parameters to vary depending on the nation’s income. Thus may restrict parameter heterogeneity, but it is an attractive way of generalizing the present empirical practice. Azariadis and Drazeb (1990) suggested that the new growth theories are feasible when the initial conditions could index the countries to produce behaviors that are close to the steady-state are similar to what the Solow model predicts, the approach to be used in this model is providing a simplified strategy of evaluating the goodness of fit of this Solow model.
The Solow Model
This model was established by Robert Solow and Trevor Swan in 1956 and has been termed to provide the most fundamental contributions to the economic growth theory. According to Acemoglu (2008), this approach has presented a simplified image of the economy in its entirety and aids in getting insights on the causes of economic growth and also explanations of income differences between countries. The two developers, Solow and Swan made an assumption of the saving rate, the population growth rate and the technological progress rates to be the primary determinants of the economic growth (Karabona & Koutun, 2013).
The Solow model is a dynamic approach that is founded in the neoclassical aggregate production function:
where Y(1) denotes the aggregate output of real income at time t measured as the real GDP. The total output is represented nyu a function of the capital input at time t, K(t), labour input at time t, L(t), and the measure of productivity or the level of technology at time t, A(t). In comparison to the model’s predecessors, the neoclassical character of the Solow model is in the changing proportion between the capital and labour inputs, following the neoclassical theory, this proportion can alter due to the alternating pieces in the factors of production.
Similar to other macroeconomic models, the Solow model tackles the assumption that the production happening is only on a single good, the economy is considered to be in perfect competition in regards to the market of this one good and the market of its factors of production. This means that the economy is at a competitive general equilibrium status. The assumption of this status implies that the supplies of the production factors and of the final goods being equivalent to the respective demand, the model take upon the assumption that there are two types of actors within the market that are the firms and the households. Simplym, the households and firms are termed to be homogenous.
Households are owners to the factors of production (Karabona & Koutun, 2013). They will supply labour inelastically and the national labour supply is to be denoted by the country’s population and consequently the labour force will increase at an unfaltering exogenously given rate n. The households are also owners to the capital and rent in these firms. Capital ownership is determined through the process of savings and investments made by the households in every unit of time. It is hence assumed by Solow (1956) the households will save an unchanging fraction of their income (sY) in every duration. This fraction of income which is saved is considered the savings rate and denoted (s). According to Jones (2002), the economy that is characterized by the Solow model is assumed to be closed, which means that the amount saved (sY) is equivalent to the amount invested (I) in each unit of time. The capital rented by the households to the firms will depreciate at a constant rate (δ). The equation describing the net change in the capital is hence:
with the ‘dot’ on top K signifying the time derivative.
According to Acemoglu (2008), firms within the Solow model are focussed on profit maximization. The homogeneity of these firms implies that all of them in this economy deal with a similar production function hence the study could take up the assumption that it is a representative ‘aggregate’ production function. The technology factor A brings in the effectiveness concept with which the production factors get transformed to the final output. Graphically, the technology factor is one that will shift the production function. Within the Solow model, it is assumed that the factor of the technological change is free since technology is non-rival and non-excludable (Karabona & Koutun, 2013). The non-rival trait is in the sense that the use of the outcomes from the technological change by one person will not deprive the other persons the opportunity of using it. It is also impossible to exclude another individual from the use of the outcomes coming from technological change. Therefore, technology is extensively and freely available to all organizations. The technological level A will grow at a constant exogenous rate g.
Considering the assumptions taken up in the Solow model, then labour and the technology factor will grow at the constant exogenous rates n and g respectively:
The neoclassical aggregate production function that is utilized within the model is characterized with the constant returns to scale and the reducing marginal returns to the capital and labour. When this function demonstrates the constant returns to scale, the increase within the factors of production by an equivalent proportion will increase the ultimate output with a similar proportion. An increase in one of the factors of production will have the total output increasing but the increase happens each period in a smaller amount, this means that the factor of production is characterized by the diminishing marginal returns. Therefore, Weil (2013) indicated that the neoclassical production function to the model would be represented following the Cobb-Douglas production function:
where a denoted the output share paid to capital and(1-a) is the output share paid to labour.
Ultimately, the Solow’s steady state level of output will explain the differences in the living standards between countries. The developed nations demonstrate higher stocks of capital that come from higher investment rates and low population growth rates with the other variables remaining constant. According to Jones (2002), the driver to the long run economic growth will hence be the rate of technological change g that is considered to be exogenous. The MRW (1992) study led to the model being modified by introducing the concept of human capita. The new new production function would be:
With H representing human capital stock and B representing the fraction of income that is paid to human capital.
Human capital is primarily the qualities which increase a worker’s productivity and will generally constitute parameters such as health and education (Weil, 2013). The assumption is that physical and human capitals are alike in different ways. Some of them include both of them being productive since they increase the produced output level. The two are also produced through a savings and investments process that is out of real income. The owners of both capital types will earn a return but the difference is that the human capital owners earn the return by working (Weil, 2013). The physical capital owners are not required to put in a lot of their time to earn the returns.
The MRW research contributed to the Solow model by making the conclusion that the nations are richer when they exhibit high levels of both human and physical capital. The higher accumulation of physical capital will lead to a higher level of total output which in turn leads to a higher human capital level and a further higher output level. Also, the greater the income share that is paid to physical capital owners (a), the more fundamental is the contribution to income levels made by human capital. The primary detriment of the Solow model is the assumption of the exogenous technological growth rate that is termed to be the primary driver of economic growth in the long run. The Solow model will not specify the influential factors of technological growth and the speed of technological progress.
Research Methodology
The data collected on Saudi Arabia’s economic variables was between 1980 to 2014, which was used to estimate the country’s SSGR. The specific variables to be used in the study include Y to represent the real GDP, with the 1990 prices being at a constant (in a million riyals). This information was obtained from the National accounts Database. L represented the labor force or the population under the working-age group category of between 18 to 60 years, depending on the available information. This information was from the World Development Indicator Online Website. K represented the real capital stock that was estimated using the perpetual inventory approach. This estimation took up the assumption that a 4% depreciation rate prevailed. The original capital stock was 12.5times the real GDP in 1969 (in a million riyals). The investment information constituted the total investment on fixed capital from the national accounts. The information was obtained from the United Nations National accounts database. Finally, O was computed as the ratio of exports and imports in GDP with the data obtained from the UN’s national accounts.
Notably, the endogenous variable bias would be minimized through the LSE-Hendry GETS techniques in conjunction with the nonlinear two-stage instrumental variable approach. This approach was also effective in utilizing the parameter restrictions. Cointegration tests followed the Ericsson and MacKinnon (2002) test devoted to the study as the EM test. It is important to note that for every equation to be established, a respective numeric number is assigned for easier pointing out in the discussion of the results section.
Model Specification
In extending the Solow model, this study will take up the assumption that the total factor productivity (TFP) relies on two fundamental externalities which substantially simplifies the estimations. The externalities are not dependent ibn additional investments by the firms. Rebelo (1991) also discussed similar externalities, specifically the learning by doing (LBD) and the openness of trade (O). These two externalities are specifically vital for the newly industrializing nations since they can increase TFP (thus the SSGR in the Solow model through an increase in the assimilation of present technologies from the industrialized countries without the prerequisite of firm incurring additional R&D expenditure. The accumulation of capital will possibly have considerable impacts on TFP and SSGR via LBD.
A vital element in this approach is that while there are constant returns to scale at the firm level, there is a possibility of increasing returns at the aggregate level. This maintains the perfect competition assumption that is in the product markets. Considering the Cobb-Douglas production function with the constant returns for a representative firm i and with the assumption that the TFP at the firm level relies on the aggregate capital stock to be:
(Equation 1 and 2 in that order)
where Y represents output, K represents capital, L represents employments and E is the error term such that . B represents the stock of knowledge that is reliant on autonomous factors. To this effect, ΔlnB is the rate of growth of autonomous TFP. B could be assumed to be the constant (ΔlnB = 0) or to grow at a constant autonomous rate of g that is: (Equation 3). Here B0 represents the initial stock of knowledge. ΔlnB hence picks up the impacts of other missing but trended variables that affect A and those that are similar to A in the Solow model. The substitution of the previous equation for A in the second equation via aggregate will have the aggregate production function as:
.(Equation 4)
The alternative assumptions about A are potentially plausible. For instance, in case A relies on different factors with externalities apart from K, these factors will also be encompassed. For instance, trade openness could have an externality that is fundamental to African nations, the A is to be denoted as: (Equation 5) or
(Equation 6).
In Equation 6, O will have increased the growth rate permanently while in Equation 5, it only brings in permanently level effects. Considering the aforementioned procedures, Equation 5 provides the following production function:
(Equation 7).
Equation 6 implies the production function as:
(Equation 8). This is actually similar to the fourth equation except that g has been calculated as s (g1+ g2O). All the derivations that will later be developed based on the fourth equation will hence golf for Equation 6. The production functions also demonstrate implied parameter restrictions.
Steady State Output And Growth Rate
In deriving the steady state output and t6he growth rate which is the SSGR then the aggregate production function from the extension doine on the solow model will be used. A steady state solution solely happens when ϕ<1. If ϕ>1 then no steady state happens since there are no diminishing returns ion K (capital) and the respective change does not become zero which actually provides the definition to a steady state. To this effect, these deprivations will take up the assumption that ϕ<1. Considering that B is similar to A within the typical Solow model, the division of Y and K with L and B gives y ̃=(Y/BL) and k̃ = (K/BL) is done.
(Equation 9)
How the capital stocks evolves is also similar to the Solow model and hence will expressed as:
(Equation 10) where s represents the saving rate, g represents the growth rate of autonomous technical progress, n represents the growth rate of labour and δ represents the depreciation rate. The equilibrium is set at (Δk̃/k̃ )= 0. Thus, the solution to the equilibrium value of k and the substitution of the production function in the previous equation will lead to the following steady state output:
(Equation 11)
It is prudent to note that when ϕ=0 , then the equation will reduce to the standard solution of the Solow model. Solving the steady state growth rate of income for every worker with a note that (Δy/y)≡(Δỹ/ỹ)+g, where g is the autonomous growth rate of B.
.(Equation 12)
In case ϕ=0 that there is an absence of externalities, the aforementioned growth rate will decrease to the exogenous SSGR if g stipulated within the Solow model. The steaststate output and the growth equations when the TFP is reliant on output as shown in , then the two are similar except that g=g1+g2O. Conversely, in case the externality due to O shoes solely level effects as in , then the SSGR will be denoted as .(Equation 13). The θ will represent the growth rate of O.
Considering the previous equations such as the last two are steady state equations, they can be estimated from the cross-section data, specifically 20 to 30 year average values to the variables utilized, these are credible proxies of SSGR compared to considering the yearly average rates of growth. The country specific annual time series data may not be proper to estimate the steady state growth equations. This is because the duration of one year is insufficient in having a country’s economy attain a steady state. Nonetheless, the yearly time series will be useful in estimating the production functions in the long run through the time series models or panel data models for a set of nations and the SSGRs in the last equations could be obtained using the estimated parameters of the production functions.
This research is to use the country specific time series data and methods, in estimation, it is appropriate that the production functions are rearranged accordingly. To this effect, they will be
(Equation 14, 15 and 16 in that numeric order) where y= (Y/L), k= (K/L).
It is also important to note that in this empirical research the specification in the final equation where O demonstrates a permanent growth effect is found to be the best for Saudi Arabia.
Results and Discussion for Estimation SSGR Results In Saudi Arabia
Table 1: Saudi Arabia’s Externalities
Note: The t-ratios (White adjusted) are below the coefficients. The p-values are below the χ2 tests statistics. 5% and 10% significance are represented by ⁎ and ⁎⁎, respectively. Constrained estimate is indicated with ©.
In the above table, the three optional estimates for the production function for Saudi Arabia have been provided. Saudi Arabia demonstrated robust son=integration test results. All the variables were pretested for the country via unit root tests. Apart from the log to the capital per worker (Ink), the rest of the variables were found to be I(1) in levels and I(0) in the initial differences. The research incorporated two alternate unit root tests and the null hypothesis was that the variable was stationary and following the ESR test which is actually more powerful compared to the unit root null. The two tests demonstrate that Ink is I(1) in levels and I(0) within the initial differences.
In the three specifications within the 14,15, and 16th equations, the specification within Equation 18 where O has permanent effect of growth, it was established to be the most effective and provided more plausible outcomes, in the conservation of space, only the estimates of the Equation 16and the variants have been reported in table 1. First, the typical CD production function without any externalities was estimated for Saudi Arabia with the outcomes denoted in Column 1 of Table 1 similar to Equation 1. The whole GETS specification of Equation 1 is:
This specification also noted the error correction term (ECM) within the square brackets and λ represents the adjustment speed for the error correction process.
Equation 1 served as the baseline equation where the comparisons would be carried out. The estimates for it were satisfactory considering that the different coefficients were appropriately signed and demonstrated statistical significance at 0.05 except for Δkt−1 which was not shown in the table. Its significance was at 10%. The summary of the χ2 tests demonstrated an absence of significance at the 0.05 level for the serial correlation, functional form misspecification and the non-normality within the residuals. The trend’s coefficient that is the SSGR of the Solow model was about 4% which was seemingly a bit high. The profits share (a) at 0.21 appeared very low. Nonetheless, none of the estimates would be termed implausible.
The estimate of specifications in Equation 14 with no consideration of the externalities due to O and equation 15 in which the two have capital and O demonstrated externalities. This was done at the level effects only. The outcome was disappointing since the estimated profit share ended up to be low at about 0.1 and insignificant. To this effect, the grid search procedure that is used for the share of profit within the range of 0.2 to 0.5 is utilized, within the search procedures, the estimates on equation 16 are considered to be satisfactory. In equation 14 and 15, there is one or another externality that is established as negative and/or insignificant, estimates of the specification in Equation 16 with the constraints that a=0.24 brought forth the best outcome and have been depicted in the second column as equation 2. Finally, considering that the trend was found to be insignificant in equation 2, the equation needed to be estimated again through constraints specifically that the autonomous growth rate was 0 and provided in Equation 3. The summary results from the statistics of these two equations were plausible.
Comparing the R^2 of the typical Solow equation of an estimated 0.5 with the other 2 equations of an estimated 0.6 demonstrated that the extended equations provide a better first of 18%. Additionally, the EM cointegration test illustrated that the null of no cointegration could be rejected at the 0.05 level for the two extended equations. The sample size that was adjusted 5% absolute critical value for (II) is 4.27 and the test statistics provided by the absolute t-ratios surpassed the CV. However the null of cointegration could also be rejected for equation 1 solely at the 10% level. To this effect, it can be stipulated that the two equations specifically 2 and 3 that have no externalities are actually preferred to the Solow equation 1.
The estimate of equation three would actually be marginally better in consideration of the two equations due to the better t-ratios of the coefficients. This came from the minimal increase in the freedom degrees; the estimates in this equation implied that the externalities from O and LBD are considerable for Saudi Arabia. Saudi’s SSGR was calculated with the average values of O and the rate of growth of employment which was 3.3%. It is prudent to note that the autonomous growth rate (g1) is zero. The findings are the opposite of the well-known results that indicated that the country’s TRP and SSGR was a negligent value of 0.2% in the second half of 20th century. This could have happened due to some externalities being neglected and the use of greater sample durations for the estimates. The estimate of ϕ for Saudi Arabia which is the measure of externality used in regards to LBD was very close to the estimate of a similar effect in other studies.
From the below figure in the actual rate of growth in each worker’s output, it was evident that the country demonstrated a mild upward trend until the 1996 to 1997 financial crisis. As the country has been evolving from solely an oil-producing country into an industrialized country venturing into other production sectors, the SSGR demonstrated to be improving marginally with a trend of 0.0006 annually.
Figure 1: Steady State Growth in Saudi Arabia
Conclusion
This research shows that the Solow exogenous model is extendable and used for estimating the SSGR specific to a country that could be utilized for growth policy. The study demonstrated that this could be attained by estimating the SSGR with consideration of benefits from two externalities, LBD and O.Tthe study’s findings demonstrate that the externalities are considerable in the country. Notably, while the effects of these two externalities have demonstrated importance in Saudi Arabia, trade openness played a relatively considerable function in the country’s growth.
It is prudent to note that the study had its limitations. One of them is that the model’s structure was quite basic as it failed to account for other factors that may bring in considerable externalities and establish SSGRs. The second limit is that there are optional proxies for LBD and trade openness, and hence it is prudent that they are incorporated in examining the sensitivity of the findings. Also, the study cannot assert that it captures all the fundamental externalities. Nonetheless, considering that the trend’s coefficient was insignificant in all the derived equations, the study could make a modest claim that this approach did sufficiently capture the growth impacts of any trend growth variables missing.
There is still a wide scope for improving SSGRs estimations in different countries, but it was beyond the objective of this research. This research will need to consider the different variables and externalities, including those that need additional resources for improving TFP. Nonetheless, the investments of the government on infrastructure were considered in the estimate that encompassed capital stock. However, the effects of externalities arising from research and development are perhaps not significant for the low-income countries since they can take advantage of the extensive amount of existing technology in the developed nations. Maybe the development policymakers could concentrate on the factors impeding the use of prevailing technologies. The researcher hopes that this approach and the empirical findings are useful for further research and development of the Solow model. The developments will guide in developing better policies that will improve the nation’s growth rates in the long run and permanently.
References
Acemoglu, D. (2008). Introduction to Modern Economic Growth. Princeton: Princeton
Azariadis, C., & Drazen, A. (1990). Threshold externalities in economic development. The quarterly journal of economics, 105(2), 501-526.
Burnside and Dollar’s “Aid, Policies, and Growth. American Economic Review 94 (3), 774–780.
Devereux, M. B., & Sutherland, A. (2011). Country portfolios in open economy macro-models. Journal of the European Economic Association, 9(2), 337-369.
Diamond, P. A. (1965). National debt in a neoclassical growth model. The American Economic Review, 55(5), 1126-1150.
Durlauf, S., Johnson, P., Temple, J., 2005. Growth econometrics. In: Aghion, P., Durlauf, S. (Eds.), Handbook of Economic Growth, vol. 1, pp. 555–677. Chapter 8
Durlauf, S. N., Kourtellos, A., & Minkin, A. (2001). The local Solow growth model. European Economic Review, 45(4-6), 928-940.
Easterly, W., Levine, R., Roodman, D., 2004. New data, new doubts: A Comment on
Ericsson, N. R., & MacKinnon, J. G. (2002). Distributions of error correction tests for cointegration. The Econometrics Journal, 5(2), 285-318.
Greiner, A., Semmler, W., & Gong, G. (2005). The forces of economic growth: A time series perspective. Princeton University Press.
Jones, C. (2002). introduction to economic growth 2nd edition. New York, WW Norton & Company.
Jones, C. I. (1995). Time series tests of endogenous growth models. The Quarterly Journal of Economics, 110(2), 495-525.
Karabona, P., & Koutun, A. (2013). An Empirical Study of the Solow Growth Model.
Kocherlakota, N. R., & Yi, K. M. (1996). A simple time series test of endogenous vs. exogenous growth models: An application to the United States. The Review of Economics and Statistics, 126-134.
Mankiw, N. G., Romer, D., & Weil, D. N. (1992). A contribution to the empirics of economic growth. The quarterly journal of economics, 107(2), 407-437.
Matsuyama, K. (2004). Financial market globalization, symmetry‐breaking, and endogenous inequality of nations. Econometrica, 72(3), 853-884.
Samuelson, P. A. (1958). An exact consumption-loan model of interest with or without the social contrivance of money. Journal of political economy, 66(6), 467-482.
Sarel, M. (1996). Growth in East Asia: What we can and what we cannot infer. In Economic Issues.
Solow, R. M. (1956). A contribution to the theory of economic growth. The quarterly journal of economics, 70(1), 65-94.
Tille, C., & Van Wincoop, E. (2010). International capital flows. Journal of international Economics, 80(2), 157-175.
University Press.
Weil, D. N. (2013). Economic Growth (3 ed.). Edinburgh: Pearson Education.
