The New Empirics of Economic Growth

The New Empirics of Economic Growth

(Critical Review)

?The New Empirics of Economic Growth by Steven N. Durlauf and Danny T. Quah University of Wisconsin, Madison and LSE January 1998, presented an explanation of the latest empirical research on patterns of cross-country economic growth. These new empirical findings are different from earlier ones (the well-known Kaldor stylized facts). The new research does not consider the production function as a central part of the analysis. It also explains that Why do some countries grow at a faster rate than others. Hence the latest empirical research go beyond the neoclassical growth model. Researchers of new empirics of growth proved their studies and findings on the subject through some of the key concepts like classification, convergence, cross-section regression, distribution dynamics, endogenous growth, neoclassical growth, regression tree, threshold, time series, and panel data.

Generally, economists research growth across different countries for three reasons. First, to understand the causes of unique patterns of growth. Continuous variations in aggregate growth rates across different countries, with the passage of time, have led to vast differences in welfare. Second, due to high rewards for intellectual work. Economic growth’s theoretical hypotheses are broad in scale and scope. Third, the new empirical growth analyses made strong and controversial claims that provoked the latest ways of cross-country income dynamics analysis. These new techniques are developing fresh stylized facts on economic growth with significant implications for theory. This paper is based on an overview of the current state of macroeconomists’ knowledge on cross-country growth. Since a number of excellent summaries on this subject already exist so it is useful to clarify how this presentation differs.

First, this paper’s emphasis is on empirical analysis. They presented different growth models based on their practical implications for cross-country income data. To bring out key ideas, they avoided overly-restrictive and detailed parametric assumptions on the theoretical models. They considered those restrictions on data that follow from a general class of models. At the same time, they show that it is relatively easy to specialize from their analysis to the various empirical specifications that have become standard in the literature. Hence they have assessed the generality and robustness of earlier empirical studies. Second, this paper provides an organizing framework for several econometric approaches—time-series, cross-section, panel-data, and distribution dynamics— used by researchers. They explained the links between alternative econometric specifications used in the theory and several practical implications of growth models. The questions raised in the new empirical growth research differ from those in earlier empirical work embodying Kaldor’s stylized facts or those in a production function.

The new empirics of economic growth considering the cross-country patterns of income, not as stability within a single economy of factor ratios (the ratio of output to capital, consumption, or investment) or growth exclusively in terms of factor inputs, rather considered all kinds of additional or explanatory factors. Hence, no longer considering the production function as a primary part of the economic growth analysis, as it was done in past research.

Sections 3 and 4 presented some theoretical models that they used to describe the subsequent analysis of empirical results and models. They consider a number of growth models in the literature and study how they limit observations on growth dynamics.

1.????????The neoclassical model “One capital good, exogenous technical progress”. The first specific structure they consider is the neoclassical growth model. They argued that the key empirical implications of the neoclassical model depend solely on the assumed production function. However, some quantitative elements of the dynamics are based on choices. To clarify those, they presented a general equilibrium formulation here.

2.????????The neoclassical model of multiple capital goods A well-known model due to Mankiw, Romer, and Weil (MRW) adds human capital to the Solow-Swan model and develops empirics that potentially describe the cross-country income data than models that are responsible only for physical capital accumulation under Solow’s original work. In both the MRW and traditional neoclassical models the levels of balanced growth income time paths can vary with the parameters of preferences and technology (τ, ρ, θ, and α).

3.????????Endogenous growth: Asymptotically linear technology. They consider here a range of models that generate long-run growth from other factors than an exogenous technical change.

4.????????No convexities and poverty trap an alternative class of models has focused on specific nonconvexities in the aggregate production function. This research paper has described the implications of nonconvexities for the relation between initial conditions and the steady-state behavior of aggregate output. The neoclassical model differs from Models with no convexities, leading to long-term dependence in the time-series properties of aggregate output. Nonconvex models may show poverty traps, where countries with low initial incomes or capital stocks converge to one steady-state level of per capita output, on the contrary, countries with high initial incomes or capital stocks converge to a different steady-state level. This is particularly convenient for explaining the empirical variations between Nonconvexities and the neoclassical approach.

5.????????Growth with cross-country interactions presents a growth model with empirical implications that differ markedly from those considered above. The model describes findings on patterns of growth, even when one considers fixed and quite standard production technologies.

Section 5 sketched some Empirical techniques and described a variety of approaches for growth analysis. Their aim was to elaborate a great deal of theoretical perspectives in comparison with different empirical growth studies. Augmented cross-section regression empirical growth studies have explained international differences and suggested some extensions of the neoclassical growth model, we find problematic the lessons drawn from some of the empirical findings. Many studies fail to explain whether the regressions they consider can be interpreted within some economic model. Linear regressions are unable to get at the features of interest. It is unclear what procedure a researcher conducts by adding a particular control variable, even when the variable is motivated by a particular economic theory. The basic Solow-Swan model admits an immense range of extensions through factors such as inequality, political regime, or trade openness. These are often highly correlated with one another and are neither mutually exclusive nor prioritized as possible explanations of growth. Hence, it is difficult to assign much import to the statistical significance of an arbitrarily chosen subset of possible controls. We, therefore, find claims that these regressions are able to identify the economic structure. Sala-i-Martin has attempted to deal with this limitation by calling “robust” only those variables found statistically significant in 95% of a group of regressions in a wide range of possible combinations of controls (variables). This work finds that many more variables appear to be robust. These variables fall into 9 categories: 1) region (dummy variables for Sub-Saharan Africa and Latin America), 2) political structure (measures of rule of law, civil liberties, and political instability), 3) religion, 4) market distortions (measured with reference to official and black market exchange rates), 5) equipment investment, 6) natural resource production, 7) trade openness, 8) degree of capitalism, and 9) former Spanish colonies.

Section 6 of this paper is about their conclusions. They have presented an overview of the latest empirical work on patterns of cross-country growth and how to match those empirical patterns to theoretical models. In Section 2 we described some of the new stylized facts on growth—they differ from Kaldor’s original set. Due to this difference, they go beyond the original neoclassical growth model. Neither the latest empirical nor theoretical research has focused on retaining the stability of the “great ratios” or factor prices. Instead, attention has shifted to a more basic set of questions: Why do some countries grow faster than others? What makes some countries prosper while others lag behind? Sections 3 and 4 discussed a great deal of theoretical growth models and their empirical implications. Distribution-dynamics models make this particularly clear. An appropriate empirical analysis for all the different possibilities has been outlined that remains under study. Section 5 described a spectrum of empirical methods and findings related to studying patterns of cross-country growth. The range is extensive and continues to grow as researchers understand more about both the facts surrounding growth across countries and the novel difficulties in carrying out empirical analyses in this research area. At the same time, the new empirical growth literature remains in its infancy. While the literature has shown that the Solow model has substantial statistical power in explaining cross-country growth variation, sufficiently many problems exist with this work that the causal significance of the model is still far from clear. Further, the new stylized facts of growth, as embodied in nonlinearities and distributional dynamics have yet to be integrated into the full structural econometric analysis.

Sections 7 and 8 related to the Technical and Data Appendices. For a better understanding of cross-country growth behavior, one must know the properties of the cross-country distribution of growth characteristics. Thus, while the overall spread of incomes across countries increased over this 25 year period, that rise was far from uniform. Fig. 1 shows the empirical regularities. The figure shows the distribution of income across national economies at different points in time. Fig. 1 also describes on scale, the past experiences of some relative growth successes and failures. Singapore and South Korea experienced high growth as compared to the world average, Venezuela the opposite. The above constitutes an initial set of stylized facts around which they organize their discussion of economic growth in this paper. They focus on the dynamics of per capita incomes as against which to assess alternative empirical analyses on growth. In this they depart from, say, Kaldor’s stylized facts—the stability of factor shares, the variability of factor input quantities, the stability of time-averaged growth rates in income and in physical capital investment, and so on. Latest empirical analyses of growth and convergence study show alternative conditioning economic variables or different economic hypotheses imply differing behavior for time paths of per capita incomes. Therefore, they focus on exactly those dynamics.