A High-Resolution Lead-Lag Analysis of US GDP, Employment, and Unemployment 1977–2021: Okun’s Law and the Puzzle of Jobless Recovery
Bright Isaac Ikhenaode
Round 1
Reviewer 1 Report
The topic of this paper is very challenging. However, the authors did not mention at all the purpose or the objectives of their research.
Moreover, the abstract and the introduction part are not adequately conceived. The abstract presents only the results, instead of a brief presentation of the topic, then the goals and only afterwards the results. The introduction should also present the state of art and, starting from the purpose of this paper, its contribution to the scientific and business environments.
The literature review should be substantially developed. Moreover, the authors should conceive the research hypotheses based on the existing literature.
The conclusions should include more policy implications.
Author Response
Thank you for very useful corrections and suggestions.
- R. The topic of this paper is very challenging. However, the authors did not mention at all the purpose or the objectives of their research.
Response. We now have added a new paragraph expressing the purpose of our study in “plain language” ( the term “plain language” is used in some American Society of Meteorology, AMS, journals.) With this introduction, we think it is better, as the referee suggest, to move the hypotheses down to after the literature survey. Lines 26-32
- R. Moreover, the abstract and the introduction part are not adequately conceived. The abstract presents only the results, instead of a brief presentation of the topic, then the goals and only afterwards the results.
Response. The abstract is rewritten and shortened and is now 203 words (lines ≈ 9-21)
- R. The introduction should also present the state of art and, starting from the purpose of this paper, its contribution to the scientific and business environments.
Response. The literature on Okun’s law can be divided into three parts. One part delas with Okun’s law applied to the whole time series and gives β values for a region or a country. The second part deals with Okun’s law related to the business cycle and the third part focuses on unemployment issues. The two last parts are most closely related to the objective of the present study, so we focus on those parts in the literature survey. (lines ≈ 26-135)
- R. The literature review should be substantially developed. Moreover, the authors should conceive the research hypotheses based on the existing literature.
Response. The literature survey has been reorganized and extended. We also moved the hypotheses down to after the literature survey so that the reader can compare our hypotheses to results in other studies. However, literature on short term developments in unemployment and output often give cautious remarks on the interpretation of results. This is partly due to lack of relevant data. (Hypotheses, lines ≈ 122-135)
- R. The conclusions should include more policy implications.
Response. We do suggest some policy implications, but we are a little hesitant to extend policy recommendations beyond those that we already have suggested. Another reviewer recommended us to be more cautious when formulating recommendations. (Lines ≈ 598-625).
Reviewer 2 Report
SUMMARYThis paper refers to the Okun's law in order to study the relation between GDP and employment, as well as to provide plausible explanations about the so-called "jobless recovery puzzle". By estimating the Okun's elasticity coefficient, the authors find that employment led GDP before and lagged after almost all of the recessions that affected the US economy during the period 1977 - 2021, except for the most recent recession that took place in 2020. Moreover, the authors find that low labor productivity is associated with the beginning recessions.
EVALUATION
The question regarding the puzzle of jobless recovery is certainly interesting and of great relevance. However, the research design, questions and methods are not clearly stated and the results are not adequately presented.
MAIN COMMENTS
1) The abstract needs an improvement. In particular, the explanation regarding finding ii) requires a clearer and more concise delivery.
2) The introduction needs a strong improvement. First, all of the research questions of the paper should be presented and discussed in a clearer way. Second, the reader would benefit from learning about the literature survey before reading the hypotheses that the paper aims to verify. By anticipating the literature survey subsection, it becomes easier for the reader to understand why such hypotheses are taken into account in the paper.
Last, in Section 1.2, page 3, you state that "Several studies also offer explanations for the “jobless recoveries” and the mechanisms that could explain jobless recoveries can be divided into five categories [...]". Because the authors propose policy recommendations on how to solve the puzzle of jobless recovery in the US, it would be to focus more on the explanation regarding the jobless recovery puzzle and these "several studies".
3) Although the use of novel techniques to produce some of the graphs are appreciated, the explanation is often insufficient and some Figures might be rather hard for the reader to interpret. For instance, what is the meaning of Figure 1.c? Other figures in the paper could also be adjusted / replaced with a simpler graphical representation and described more clearly.
4) There are no output tables regarding the OLR estimates. Are the estimated coefficients always statistically significant? Are the proposed estimates robust to the inclusion of other variables (such as the ones you described, i.e. US government expenditures, Federal
government tax receipts, etc...)?
5) The authors refer to the Covid 19 recession as just an "exception to the rule". They should at least try to give a plausible explanation about why their hypotheses do not hold true for the most recent and deepest recession.
MINOR COMMENTS
1) Why does the title of the paper refer to the period 1977-2019? The study actually takes into account the US economy during the period 1977 - 2021.
2) In Table 1, it is not clear what the authors refer to as "jobless depth" and "jobless duration". How are those values obtained?
Author Response
Thank you for making an accurate summary of our research. Thank you also for the comments and suggestions that we appreciate very much.
This paper refers to the Okun's law in order to study the relation between GDP and employment, as well as to provide plausible explanations about the so-called "jobless recovery puzzle". By estimating the Okun's elasticity coefficient, the authors find that employment led GDP before and lagged after almost all of the recessions that affected the US economy during the period 1977 - 2021, except for the most recent recession that took place in 2020. Moreover, the authors find that low labor productivity is associated with the beginning recessions.
EVALUATION
R. The question regarding the puzzle of jobless recovery is certainly interesting and of great relevance
Response. Thank you for supporting our choice of study object.
- R. However, the research design, questions and methods are not clearly stated and the results are not adequately presented.
Response. Your critique and suggestions are detailed below, so we address your concerns there.
MAIN COMMENTS
- 1) The abstract needs an improvement. In particular, the explanation regarding finding ii) requires a clearer and more concise delivery
Response. We agree. ii) the β E coefficient measures changes in GDP and employment and show pronounced changes around a recession. It recovers to non-recession values over 2 to 4 years periods. (We explain that we have replaced unemployment, UE, with employment; EM, before this sentence) (Lines ≈ 9-21)
- R. 2) The introduction needs a strong improvement. First, all of the research questions of the paper should be presented and discussed in a clearer way.
Response. We now give a “plain language” description of the research question in the beginning of the introduction. ( the term “plain language” is used in some American Society of Meteorology, AMS, journals.) lines ≈ 26-33
- R. Second, the reader would benefit from learning about the literature survey before reading the hypotheses that the paper aims to verify. By anticipating the literature survey subsection, it becomes easier for the reader to understand why such hypotheses are taken into account in the paper.
Response. After the “plain language” text in the introduction, we follow the normal sequence for articles in economic and give a short summary of the results and what our contribution to the research questions are. (lines ≈ 49-62). Then follows the literature survey and our hypotheses.
R. Last, in Section 1.2, page 3, you state that "Several studies also offer explanations for the “jobless recoveries” and the mechanisms that could explain jobless recoveries can be divided into five categories [...]". Because the authors propose policy recommendations on how to solve the puzzle of jobless recovery in the US, it would be to focus more on the explanation regarding the jobless recovery puzzle and these "several studies".
Response. Thank you and we agree very much. We have now reorganized part of the introduction that delas with “jobless recovery” and added several new references. (Lines ≈ 88-121)
R. 3) Although the use of novel techniques to produce some of the graphs are appreciated, the explanation is often insufficient and some Figures might be rather hard for the reader to interpret. For instance, what is the meaning of Figure 1.c?
Response. We have added more explanation to Figure 1c. Thank you for making us aware of it. (lines ≈ 254-258)
- R. Other figures in the paper could also be adjusted / replaced with a simpler graphical representation and described more clearly.
Response. We have reread the explanations of the figures and tried to improve them. However, it is difficult to make them simpler without loosing important information.
R. 4) There are no output tables regarding the OLR estimates. Are the estimated coefficients always statistically significant? Are the proposed estimates robust to the inclusion of other variables (such as the ones you described, i.e. US government expenditures, Federal government tax receipts, etc...)?
Response. With respect to Figure 3, the PCA constructs new variables that are orthogonal to each other. However, pairs of the original variable may be or not be significantly related. Note that the variables are cyclic variables and then PCA results will have another interpretation than for variables with normal distribution (lines ≈ 414-419). With respect to Figure 4, none of the regressions were significant, (lines ≈ 471-476). The results contrast with our expectations but are not discussed further.
5) The authors refer to the Covid 19 recession as just an "exception to the rule". They should at least try to give a plausible explanation about why their hypotheses do not hold true for the most recent and deepest recession.
Response. We suggest an explanation, lines ≈ 578-580
MINOR COMMENTS
R. 1) Why does the title of the paper refer to the period 1977-2019? The study actually takes into account the US economy during the period 1977 - 2021.
Response. Tank you for pointing out this, it was an error.
- R. 2) In Table 1, it is not clear what the authors refer to as "jobless depth" and "jobless duration". How are those values obtained?
Response. They are anomaly values obtained from linearly detrended series, lines 169- 171
Reviewer 3 Report
The subject of the article is important, while the research concept is well thought out. However, I have some comments and suggestions how to improve the paper.
1) It is necessary to note that in times of crisis, a decrease in GDP produced by a single employee is noted first, while an employment may decline with some delay. Usually, the shares of informal employment and informal economic turnover also increase. After that, enterprises usually adapt to new conditions and achieve greater efficiency at the lower level of employment. In the case of the EU, this process is strongly disrupted by the state interventionism. Therefore, the adjustments of enterprises are not always rational. In the United States it is possible to generalize certain dependencies found in the course of research, trusting in their truth throughout the country. However, the EU is not homogeneous. It consists of countries differing in terms of the level and structure of employment and GDP, strongly diversified culturally and governed by political forces with various approaches to economic policy. Their common feature, however, is that each subsequent crisis results in a higher level of intervention and regulations, strengthening the artificially created realities of the functioning of enterprises. One should also take into account the impact of the climate policy on the economy and – what is currently of special importance - energy dependence on Russia, the negative potential of which may be comparable to the crude oil shocks of the 1970s. The measures of unemployment and GDP used are not perfect as well. For these reasons, your results are subject to a great deal of uncertainty. Hence, I believe that you should be more cautious when formulating recommendations for the economic policy. These limitations should be also clearly indicated, preferably by expanding the Discussion section and referring to literature sources. The math approach is always useful, however, you shouldn't blindly trust the numbers.
2) Adapt the text to the journal's formal requirements.
3) Be sure to check the text for editorial purposes.
4) Enrich the literature review. In a present form, the section "A short literature survey" is too short.
Author Response
Thank you for your kind comments.
The subject of the article is important, while the research concept is well thought out. However, I have some comments and suggestions how to improve the paper.
- R. 1) It is necessary to note that in times of crisis, a decrease in GDP produced by a single employee is noted first, while an employment may decline with some delay. Usually, the shares of informal employment and informal economic turnover also increase.
Response Thank you. We agree. We have earlier written an article, Orsi and Seip (2019) discussing the of the underground economy. We add some comments in this respect. (Lines ≈ 543-546)
- R. After that, enterprises usually adapt to new conditions and achieve greater efficiency at the lower level of employment. In the case of the EU, this process is strongly disrupted by the state 537-54 0i nterventionism. Therefore, the adjustments of enterprises are not always rational. In the United States it is possible to generalize certain dependencies found in the course of research, trusting in their truth throughout the country.
Response. What you say, if we interpret you correctly, is that enterprises adjust their hiring according to their requirements, and not related to state interventionism. This is an important point (our lines 92-100) if we have interpreted you correctly.
- R. However, the EU is not homogeneous. It consists of countries differing in terms of the level and structure of employment and GDP, strongly diversified culturally and governed by political forces with various approaches to economic policy. Their common feature, however, is that each subsequent crisis results in a higher level of intervention and regulations, strengthening the artificially created realities of the functioning of enterprises.
Response. We now include a reference that support the statement that “labor regulations have been strengthened in the EU since 1970” lines ≈ 93-95
- R. One should also take into account the impact of the climate policy on the economy
Response. Yes, but that issue will be outside the scope of the present study.
- R. and – what is currently of special importance - energy dependence on Russia, the negative potential of which may be comparable to the crude oil shocks of the 1970s. The measures of unemployment and GDP used are not perfect as well.
Response. Yes, although little is perfect, we are still wondering about the difference between employment and unemployment. We know the plain “facts”, but see the two figures in Appendix A a and -b.
- R. For these reasons, your results are subject to a great deal of uncertainty. Hence, I believe that you should be more cautious when formulating recommendations for the economic policy. These limitations should be also clearly indicated, preferably by expanding the Discussion section and referring to literature sources.
Response. We agree. Another referee recommend that we include more policy recommendations, but we responded by saying that we were hesitant to do that. We are a little cautious about our recommendation: “.. is a multicriteria question” lines ≈ 623-625, but as you indicate, numbers may give a wrong message as well.
The math approach is always useful, however, you shouldn't blindly trust the numbers.
Response. An appropriate warning. Again, why do employment and unemployment tell different stories? We know that unemployment do not people that permanently drop out of the workforce, but it may not be sufficient to explain the difference. Lines ≈645-647.
- R. 2) Adapt the text to the journal's formal requirements.
Response. Done, the use of a normal text file was due to a misunderstanding and error from our side.
3) Be sure to check the text for editorial purposes.
- R. 4) Enrich the literature review. In a present form, the section "A short literature survey" is too short.
Response. We have enriched the literature survey, but searching Web-of-science, it is difficult to find literature that addresses short term variabilities. For example, Gimbel and Sinclair (2020) states that:” the cyclical component to mismatch,..… (could potentially be used) as an additional indicator of slack .. if we knew.. “. The “if we knew” suggest that at least some of the present knowledge is limited and uncertain.
Gimbel, M. E. and T. M. Sinclair (2020). Mismatch in Online Job Search. LERA 2020 Meetings, Sa Diego, Labor and employment relations assocciation (LERA).
Orsi, R. and K. L. Seip (2019). "Effects of tax changes on GDP and underground economy: Italy 1982 to 2006." Submitted.
Round 2
Reviewer 1 Report
The authors have made most of the suggested changes. However, in the conclusions they should include more policy implications. Meanwhile, the paper needs proof-reading.
Author Response
- R. The authors have made most of the suggested changes. However, in the conclusions they should include more policy implications. Meanwhile, the paper needs proof-reading.
Response. Thank you for your suggestion. We very much agree with you that studies benefit having conclusions that include policy implications - and preferably in “plain language”. The present paper has two: i) Inspecting time series for GDP and employment may be used to support forecasts of coming recessions, and ii) ”, to avoid “jobless recovery”, secure that employment does not increase more rapidly than the real economy.
However, another reviewer advised us to not stress our conclusions too far. Nevertheless, we added a sentence at the end of the conclusion section.
We will ask the editorial office to check that the text is satisfactory language wise.
Reviewer 2 Report
I am satisfied with the author's revision of the paper.
Author Response
- R. I am satisfied with the author's revision of the paper.
Response. Thank you. Your advice was very helpful
Reviewer 3 Report
The improved paper looks fine. The corrections made raised its scientific quality and soundness. It can be accepted for publishing.
Author Response
- R. The improved paper looks fine. The corrections made raised its scientific quality and soundness. It can be accepted for publishing.
Response. Thank you. Your comments were very helpful, and also gave us food for thought.
