Structural Transformation Path Across Countries: Is It Systematic Pattern

The structural transformation across the economic sectors is one of the prominent features that go together with economic development. The paper scrutinizes whether developing and low-income countries follow the similar path and pattern of structural transformation by which the developed countries are following or followed this threshold and are now experiencing a shift from the industrial sector to the service sector. The structural transformation paths of developed countries are almost identical, but the pattern of sectoral output shares varies from that of developing countries. The research reveals a fascinating finding i.e., low-income countries outperform middle-income countries and some major countries in terms of the pace of structural transformation from agriculture to service sector.


Introduction
Economists have explored a statistical connection between a country's production and occupation systems and its level of economic growth. In his seminal work "The Conditions of Economic Progress," Colin Clark writes: "as time goes on and communities become more economically advanced, the numbers engaged in agriculture tend to decline relative to the numbers in manufacture, which, in turn, decline relative to the numbers engaged in services" (C Clark, 1940). Clark developed the fundamental theoretical relationships that would later become the cornerstone of structural change theory (Syrquin, 1988). Simon Kuznets explored the Clark-Fisher theory empirically in his document and analyzed that structural transformation plays an integral role in the process of modern economic growth among developing countries.
In one of the characteristics of modern economic growth, Kuznets summarized that "the rate of structural transformation of the economy is high" (Kuznets, 1973). For Kuznets, and more generally in economic history and development, growth and structural transition are inextricably linked. Major aspects of structural change of an economy take place mainly along two dimensions, the reallocation of transformation from agricultural to non-agricultural stalking and industry to services.
Economies evolve, not only in terms of growth but also in terms of structure. While economic theory recognizes the connection between economic growth and structural changes, the issue of whether economic growth induces structural changes or structural changes trigger aggregate growth. (Olczyk & Kordalska, 2018) found that a strong causal relationship occurs between economic growth and structural changes among developed countries by using panel Grangercausality analysis of eight transition countries. In contrast to developed countries, emerging countries take very different institutional transition directions. Asia is on a glide path that is most similar to that of developed countries (Bah, 2016). Due to economic reforms, the economies of India and China are increasing rapidly among Asian countries. During the period compares the structural transformation process among the selected countries, whether developing and low-income countries follow the similar path and pattern of structural transformation by which the developed countries have crossed this threshold and are now experiencing a shift from industry to service sector. The reallocation of resources across sectors is one of the prominent features that go together with development. The paper will address the key issues of structural change that varies across different country groups. The paper is divided into five sections, section I includes the brief introduction and literature of this study. Section II includes the data and methodology using in this study. Section III deals with the paths of structural transformation using the Michaely Index. Polynomial functions of sectoral output shares are presenting in section IV of the study and section V is providing the overview findings of the study.

Data and Methodology
We evaluate the relative performance of selected countries on structural change and economic growth and examine whether developing and low-income countries follow the same pattern of economic transformation over the period. The study uses time-series data at the national level of each country groups collected from World Development Indicators base at (2010) constant prices in dollar terms. To condense the complexity of data we first start the conversion of actual data into billions for further analyses in a smooth way. The pattern of structural change and economic growth is studied by examining the log of GDP and sectoral output shares (Agricultural, Industry and Service) across country groups from 1991 to 2018.
For a structural change, the Norm of Absolute Values (NAV) is estimated, also known as Structural Change Index (SCI) or Michaely Index (MI) across the country groups (Michaely, 1962) and (Cortuk & Singh, 2010). Multi polynomial regression is used to fit the relationship between sectoral output shares and Gross Domestic Product (GDP) for all countries and country groups. The degree of the polynomial is determined by the goodness of fit. To reduce the likelihood of getting a warning message covering the collinearity among predictors, the mean centre is estimated from the independent variable (log GDP). The mean centre is the actual deviation which is taken from the actual mean of all country groups, expressed as: Where is the value of an independent variable (log GDP) of country 'i', at time 't', and ̅ is the actual mean of country groups.
Starting from a linear polynomial with log GDP, and mean-centred, variables at power 2 and 3 as independent variables, the degree was increased by 1 and continues the process until we find the suitable model which fits the data. Both the linear and quadratic polynomial function with independent variables (log GDP and log GDP mean-centred variable at power 2) were found statistically insignificant and doesn't fits the data. So, the log GDP mean-centred variable at power 3 (loggdpmcube) became the new independent variable that survived on the Durbin-Watson test.
For each sector, we estimated the following regression equation: Where is the sectoral output share and is the log GDP mean cube (log GDP meancentred at power 3) for country 'i' in period 't', and ′ ′ are the regression coefficients.

Results (Michaely Index):
This section analyzes the pattern of structural change among the country groups, whether developing countries have a similar structural transformation path followed by developed or high-income countries. The analysis covers the ten countries based on the economic transition away from the primary sector to the industrial sector and later to the service sector. The yearly estimated Norm of Absolute Values (NAV) is categorized in panel data set by averaging to obtain time series of five intervals from 1992 to 2018 consisting of ten country groups into six cross-sections. The whole data set was incomprehensible to interpret on a single desk or slot.
For its better understanding, the countries were restructured into three groups. The group first It has been generalized that the economic transition path followed by selected country groups were invariantly changing over the period after globalization. The mean of the norm of absolute values of five intervals is not consistent throughout the country groups so that we can say the developing, middle-and low-income countries are following the similar pattern of structural transformation process by which developed ones have crossed or crossing. It needs another study that can identify, how far the economic transitions differ or similar among the economies. To

Results (Polynomial Anaylsis)
The economic transformation of sectoral output shares is a crucial consistency of the data across the countries. Even though the rate of structural change varies by countries and country groups, they all share the same characteristics: the share of agriculture in output declines, with the increase in GDP, the share of industry increases initially and subsequently decreases and the share of services increases steadily. Polynomial regression is estimated to fit the relationship between the sectoral output and GDP across the country and country groups. As mentioned above in the methodology section, the degree of the polynomial is determined by the goodness of fit, so adjusted R squared has to be taken into consideration instead of R squared. Adjusted R squared is a better model evaluator and can correlate the variables more efficiently than R squared. Table 1 and 2 represents the regression results across the countries and country groups with adjusted R squared. For all the countries and country groups, thirddegree polynomial functions are best fitted across the sectors (agriculture, industry and services) except Japan in the industry sector as the change in the adjusted R square is (-0.02) from quadratic to cubic model (insignificant at 5% & significant at 10% level of significance) and high-income countries in agriculture and service sector as the change in adjusted R 2 is (-0.01) from log GDP 2 to log GDP 3 (both are significant at 5%). So for them, the quadratic polynomials are best fitted with an adjusted R 2 equals 0.55 and 0.83 (agriculture sector) & 0.88 (service sector) respectively. The rate of structural transformation from the agriculture sector to the service sector among the individual country group is highest in Japan reported in table 1. The regression coefficient of Japan at the third-degree polynomial (log GDP 3 ) in the agriculture sector is negative (-294.91) and   (1 & 2). The analysis above shows that structural transformation in developing countries is differing which is followed by the developed countries, but the path among some countries is similar.

Conclusion and Discussions
The paper analyzed the structural transformation pattern among the countries and country Low-income countries and was found statistically significant while using Levene's and Post-Hoc tests.
The service sector is a major contributor to the economic structural change, and it grew faster in Japan, Arab World and High-income countries among the countries and country groups.
Among all the countries, Pakistan and China are the only countries where structural transformation is moving more towards the industrial sector than the service sector. Lowincome countries and Indonesia are on the same path on structural transformation rate. The main finding of the paper is, not all developing countries have the same path of structural transformation by which developed countries are passing. But the pattern of sectoral output shares or the shift from the agriculture sector to the industrial sector and from the former to the service sector is almost similar among the developing countries.

Availability of data and materials
Data anaylsed in this study is available online from world development indicators: https://databank.worldbank.org/source/world-development-indicators

Ethics approval and consent to participate
All the participants that helped in this study did so willingly.

Consent for publication
The authors agreed to have this paper published.