Round 1

Reviewer 1 Report

The article tackles a very interesting and on fashion topic. Is well structured and written. It clearly explains the data, methodology, results and conclusions.

I have only two minor comments:

1)    Cites [2] and [3] are cited before [1] in p. 1

2)    I suggest to incorporate a scree plots of the eigenvalues in the PCA analysis. Is just to use the command “screeplot”. I also suggest to explain with it the explaining power of the variance of each component.

Author Response

Author’s Response to Reviewer 1

I appreciate the first reviewer’s favorable review and his(her) suggestions for revision.

The following comments are numbered in accord with the reviewer’s suggestions:

The citations are actually ordered properly.   The first citation is in the title of this first Section because it quotes a book title.

The eigenvalues in Table 1 give rather unsightly screeplots.  I prefer to save manuscript space and let the reader view them in this very compact table.

        The new first paragraph in Section 5.1 explains how these eigenvalues give the proportion of variance in all three GDP indicators that is due to their first principal component.  Thank you for this suggestion.

The revisions in this new version run over every page in the manuscript.  On each page, each new passage is flagged in red at its start and finish.       

Reviewer 2 Report

Review on Manuscript ID: economies-248988 entitled Well-Being as Isoelastic GDP

Summary

The paper aims to demonstrate that a nation’s well-being can be computed from its powered GDP.

General comments

The paper is unacceptable in the present version and needs major revisions. My main concerns are on the research questions of the paper and on the interpretation of a nation’s well-being.

What follows are several suggestions for how the paper might be improved.

Specific comments

1-       Firstly, the authors investigate the internal consistency of the indicators comprising GDP by means of the principal component analysis. Albeit I agree that PCA is a valid multivariate method for assessing whether sub indicators are appropriately chosen for constructing a synthetic indicator, I do not understand why we need to investigate whether GDP’s components are internally consistent. In the economic debate the controversial matter is not the internal consistency of the indicators chosen for computing GDP: on the contrary, what is questioned is its appropriateness to measure well-being beyond the economic/income/production dimension of progress. As a matter of fact, GDP is the most (universally) accepted and used indicator of economic activity that economists have ever constructed. Nevertheless, it has been increasingly recognized that many other dimensions of people’s quality of life should be taken into account  in order to understand the different progress performances of different countries. Furthermore, in the last decades, literature has highlighted that well-being we need to take into account, especially for advanced and  developed countries, a number of dimensions that go beyong those included for the calculus of HDI: by neglecting this issue means putting apart most of the recent findings of well-being literature. Thus, in my opinion, the first aim of the paper sounds as a good exercise in principal component analysis, yet it has a very low scientific and academic interest for the debate on well-being measurement.

2-       Secondly, the paper aims to assess whether GDP predicts well-being. However, the existence of a high correlation between GDP and multidimensional measures of well-being is a replicated finding in economic literature (see, among others, Ferrara and Nisticò 2013, 2015; Marchante et al 2006; Brandolini and Vecchi 2011), even if, as recent papers point out, a rise in per capita GDP, which is a production indicator, is not necessarily mirrored in  increasing well-being (Stiglitz et al. 2009; Fleurbaey 2009; Decanq and Schokkaert 2015, among others). An accurate review of the literature on this issue, to be added in the paper, should contribute to better contextualize, in a less simplistic manner than in the current version of the paper, the comparison between GDP and multidimensional well-being indicators.

3-       Even more, the authors consider HDI as the well-being measure suggested by the “Stiglitz Commission”, making a double confusion: the Report of the Stiglitz commission does not propose a synthetic indicator for measuring well-being; secondly, the report indicates a number of domains that a suitable measure of well-being should consider, going beyond both GDP and HDI. Thus, in my opinion, the sentence “The present paper conflates ….of GDP.” (pag. 2, rows45-46) is incorrect. This is a substantial matter, not just a matter of re-writing the above sentence: the effort of measuring multidimensional well-being carried on by international organizations (European commission, OECD), national governments, scholars and policy makers has a rationale in decades of scientific thinking, that should be adequately taken into account. Analogously, I found strongly reductive of the importance of the debate on the well-being measurement the sentence “In view of…global economy” (pag. 2, rows 58-60): the prominence of the Keynesian construct in the global economy regarding the measurement of production does not conflict with the need of measuring multidimensional well-being beyond the economic/productive sphere.

4-       In the Discussion sections, the authors should also discuss the implication for multidimensional well-being of sustained GDP growth, taking into consideration the observed contradiction between a fast production growth and the increasing of inequalities, as some countries’ experience, such as India, suggests.

5-       On the methodology content of the paper. In my opinion, the methodological contribution of the paper is low. It applies the PCA considering the three main sub-indicators of GDP in order to assess their internal coherence. This looks  just as an exercise. Anyway, with this aim in mind, along with the percentage of variance accounted for by the first principal component, the authors should show the results of both Bartlett test and the Measure of Sampling Adequacy.

As a consequence of the previous comments I highly suggest to:

1) change the title in “HDI as Isolelastic GDP: evidence from China and Pakistan”, as the findings refer to a precise indicator of human development and to just two countries. 

2) mitigate the sentences and conclusions on the basis of the effective results of the paper.

3) improve the literature review, focusing on the debate on the need of measuring multidimensional well- being

4) specify throughout the paper that the authors refer to HDI, accordingly changing “well-being” word with HDI

5) explain the need to investigate the internal consistency of GDP sub-indicators

6) correct in the discussion section IMF as the acronym of International Monetary Fund

7) move the Stata commands in appendix or in end notes, leaving in a footnote in the main text the source of data.

8) specify the really original contribution of the paper both on the methodological and theoretical ground.

References

Brandolini A.,Vecchi G. (2011), The Well-Being of Italians: A Comparative Historical Approach, Economic History Working Papers, Bank of Italy, n. 19.

Costanza R., Hart M., Posner S., Talberth J. (2009) Beyond GDP: the need for new measures of progress, the pardee papers, 4 (January), Boston University.


Decancq, K., & Schokkaert, E. (2015). Beyond GDP: Using Equivalent Incomes to Measure Well-Being in Europe. Social Indicators Research, 1-35.

Ferrara A.R. and Nisticò R. (2013), Well-Being Indicators and Convergence Across Italian Regions, Applied Research in Quality of Life, 8 (1): 15-44.

Ferrara A.R. and Nisticò R. (2015), Regional Well-Being Indicators and Dispersion from a Multidimensional Perspective: Evidence from Italy, Annals of  Regional Science, 55: 373-420.

Fleurbaey, M. (2009), Beyond GDP: the Quest for a Measure of Social Welfare, Journal of Economic Literature 47(4): 1029–1075.

Marchante A.J., Ortega B., Sanchez J. (2006) The evolution of well-being in Spain (1980–2001): a regional analysis,  Social Indicators Research, 76(2):283–316.

OECD (2017), How’s Life? Measuring Well-being. Paris. http://doi.org/10.1787/9789264201392-en

Stiglitz, J., Sen A., Fitoussi J.P. (2009),  Report by the Commission on the Measurement of Economic Performance and Social Progress, Paris.

Author Response

Author’s Response to Reviewer 2

I appreciate the second reviewer’s extensive and penetrating review and his (her) suggestions for revision.

The following comments are numbered in accord with the reviewer’s suggestions:

 Title is changed as requested.

 Sentences and conclusions are scaled down and data based to the results of this paper.

The expanded literature review of multidimensional well-being measurement is in Section 6.2., which is anticipated at the end of Section 1.  This brief review turns away from subjective and toward objective indicators of well-being.  It concludes with a suggested well-being index that includes HDI.  This is where the internal consistency analysis in Section 4.2 can a) solve the problem of well-being dimensions being measured in different units,  b) provide policy-relevant weights for these dimensions, and c) adjust the single well-being index for a nation’s inequality.

“HDI” replaces “well-being” wherever the present research is discussed.   

Table 2 demonstrates the need for internal consistency analysis to provide policy-relevant weights, now for GDP indicators measured in the same units.      Unweighted Keynes GDP cannot differentiate Chinese gross domestic saving from Pakistani household expenditure.  Country specificity is buttressed by the new lemma 4 and new Table 4, which justify and demonstrate the value-added properties of weighted indicators (cf. Section 6.1) 

Done

Done

The original theoretical contribution here is the societal data theory in Section 4. Axiom 1 postulates a latent impact of a GDP indicator for every individual in a nation.  Axiom 2  posits a property of the latent population distribution of these impacts.  This property allows the manifestation of the yearly means of the first latent principal component of three latent GDP variables.   Axiom 3 assumes a relation between this first manifest principal component and HDI.

        Methodologically, the original contribution here is the latent 2-level principal components analysis, implied by axioms 1 and 2, that nests individuals within successive years for a given nation.  This analysis is enhanced in this revision by the new Lemma 4 and Table 4 which allow ancillary weighted indicators to help Keynesian GDP emphasize particular aspects of an economy. 

        Finally, the Bartlett test and Measure of Sampling Adequacy are not relevant here because there is no sampling in societal data theory.  This theory allows the axiomatic computation (instead of estimation or significance testing) of population parameters.  In any event, the Bartlett test and Measure of Sampling Adequacy are applied to correlation matrices for factor analysis, where variables are usually measured in different units.  The present principal components analysis uses a covariance matrix, which is needed with GDP indicators measured in the same units (cf. J. Johnston [8] (Appendix A-10)).                     

The revisions in items 1) through 8) run over every page in the manuscript.  On each page, each new passage is flagged in red at its start and finish.        

Round 2

Reviewer 2 Report

Comments to the author’s reviewed version

This revision is certainly an improvement over the last version I have seen time ago. The author addresses the major part of points I have raised in my earlier report. Yet I still have some important recommendations: for some of my concerns the author acknowledged them even though he (she) mainly deals with them in a not accurate way.

First of all, I am not satisfied with the review of the literature on the need of well-being measures beyond GDP  (which was the second of my specific comments the first version of the paper). The review of the literature on this subject still remains quite simplistic. Please, rewrite it.

Secondly, the issue of the “subjective indicators” is completely not pertinent to the issue of this paper. Subjective well-being involves research questions and aims that are not the focus of this study. I suggest to eliminate this portion (rows 245-252) of the new version of the paper and to concentrate the attention on the objective well-being measurement. Regarding this latter strand of literature, I find misleading the space dedicated to Big Data, instead of dealing with the economic and social debate on (objective) well-being measurement.

Anyway, “multidimensional well-being measurement”, even if it is improved, cannot stay in the discussion section! Traditionally, the discussion section is devoted to the discussion of the original results of the paper.

Thirdly, the paper lacks a conclusion section!! Please write it.

Last: regarding the author’s response to the request of showing the results of tests on the internal coherence of the indicators, other than the percentage of variance accounted for by the first principal component, considered for the PCA  I feel that the author cannot (does not want) do much more to address this concern: the tests and specifications I suggested to the author are used not only for sampling data or for principal components extracted by correlation matrix: scatterplots, MSA and Bartlett's tests are usually used also for PC extracted by covariance matrix. All in all, the percentage of variance explained by the first principal component in the case of this paper is so high that all these tests should have the only aim of enriching the methodological content of the analysis.

Author Response

Author’s Response to Reviewer 2 in Round 2

The following comments are numbered in accord with the reviewer’s round 2 suggestions:

The introduction gives a brief economic history of GDP that generated the current debate on well-being measurement beyond GDP.  The review of the latter literature has been upgraded to the mission of the present paper, ie. the demonstration of the value added by GDP’s internal consistency, which is highlighted by GDP’s isoelastic relation to HDI.                

The passages on subjective indicators and Big Data have been deleted.  The discussion section has also been removed, and the review of well-being measurement has been moved to the introduction.

The  conclusion section gives the paper’s contributions to our understanding of GDP’s three classic indicators, along with the added value of latent 2-level principal components analysis for informing national economic policy.     

Thank you for pursuing this point.  Section 5.1 now includes Kaiser-Meyer-Olkin’s Measure of Sampling Adequacy = 1.00 in China and Pakistan.  This enhances the earlier demonstration of internal consistency of the classic GDP indicators.      

Round 3

Reviewer 2 Report

I am satisfied by the author's revision of the paper. I think that it can be accepted in the present form.

Congratulations.