Debt-Growth Nexus in the MENA Region: Evidence from a Panel Threshold Analysis

Round 1

Reviewer 1 Report

Referee report on the paper entitled ’ DEBT-GROWTH NEXUS IN THE MENA REGION: EVIDENCE FROM A PANEL THRESHOLD ANALYSIS The Theory Applications of Finance and Macroeconomics ’ submitted for publication to ’Economies’ (manuscript number economies-975632)

This is a very interesting paper with potentially important policy implications. Nevertheless, I believe the paper needs considerable amendments before publication.

Literature

The literature overview in different parts of the paper seems a little patchy, sometimes overstretched and sometimes its focus could be increased.

The description of the Cobb-Douglas production function is too lengthy. The paper could describe this in two lines, and maybe provide a more focused discussions on the channels through which public debt can have positive and negative effects on economic growth.

The discussion of the earlier literature, focused on large panels of countries, following the 90% threshold claim by Reinhard and Rogoff could be summarised in a table. What threshold for what groups of countries, method, and positive or negative effect. The paper also seems to miss some of the literature saying that the threshold value is very uncertain and that it is considerably lower than 90% or even 60%.

 Baglan and Yoldas (2013) identify a threshold effect of 20%

 Égert (2012) shows thresholds of between 20% and 60% of GDP.

Baglan, D. and E. Yoldas (2013), “Government debt and macroeconomic activity: a predictive analysis for advanced economies”, Federal Reserve Board Finance and Economics Discussion Series No. 2013-05.

Égert, B. (2012), “Public debt, economic growth and nonlinear effects: myth or reality? OECD Economics Department Working Paper No. 993.

Methodology

I have two main reservations:

The paper uses time series cross-country panel data. Estimations carried out on such data shold include country and time fixed effects. Based on Table 4, none of them are used.

All growth regressions are variants of an error correction representation of the level relationship between income and its drivers.

 Level relationship: Y(t)=f(X1(t), X2(t) etc.)

 Growth regression: • dY(t)=f(Y(t-1), X1(t-1), X2(t-1) etc).

Therefore, the actual effect of drivers is on the level of Y and this effect should be determined as coeff estimate on X(t-1) over coeff estimate on Y(t-1) (the error correction term). Any growth effects should be looked at using dynamic terms not included here, but following this form: •

dY(t)=f(Y(t-1), X1(t-1), X2(t-1) etc, dX1(t), dX2(t) ). (richer short-term dynamics can also be added)

The paper assumes constant TFP. I am not sure where this is coming from. The empirical growth literature includes drivers of TFP such as a measure of innovation, policies related to resource reallocation (product and labour market regulation) and framework conditions (quality of institutions).

Some of the countries considered have considerable natural resources, not considered here.

What is the rational for including government spending in a linear fashion on top of public debt? The negative coefficient on government spending implies that government spending is bad at any level. If debt is below a threshold, it is good. But how can it be good if public spending is bad?

Data 

It is not clear whether public debt is central or general government debt Results

The paper has a number of weird results. They might be due to the methodological problems described above:

Openness has a negative coefficient estimate because of barriers to trade. This is stated on page 8. If anything, it might have to do with terms of trade. However, in Table 4, the coeff estimate on openness is a positive one. 

Financial depth has a negative coefficient estimate. There is a large literature on discussion the non-linear relationship between finance and growth. Also, financial depth is measures by money supply. There are better variables capturing financial depth (eg private credit). Money supply is probably correlated with inflation.

Minor things 

First para in the introduction: paper states there is a close link between debt and growth. But it does not tell whether this is a positive or negative link

Table 3: there are double commas in the last line.

Equations 4, 5, and 6 write that the regressions are done for regimes 1.) smaller and equal to lambda and 2.) larger than or equal to lambda. It has to be EITHER smaller than or equal to lambda AND higher than lambda; OR smaller than lambda AND higher than or equal to lambda

Author Response

Attached our response to Reviewer 1.

 

Regards.

Author Response File: Author Response.pdf

Reviewer 2 Report

Debt-Growth Nexus in the MENA Region: Evidence from a Panel Threshold Analysis

 

Summary:

This study tries to find the debt-growth nexus in the MENA Region. Using the threshold estimation technique, this study examines whether a debt-to-GDP threshold exists in the public debt and economic growth relationship for 20 Middle East and North Africa (MENA) countries from 1990 to 2016. The debt-to-GDP threshold value for this region is lower than the debt threshold calculated by Reinhart and Rogoff (2010) and Caner et al. (2010) for developing countries. This paper’s uniqueness is that it used the regression model proposed by Hansen (1999) to provide a reliable, optimum debt-to-GDP threshold. Debt and economic growth have a positive effect only below the debt-to-GDP threshold value. More precisely, debt at a level below 58% of GDP has a positive influence on economic growth. The effect is insignificant when debt exceeds 58% of GDP because the coefficient is insignificant. The findings indicate that the relationship between public debt and economic growth is contingent on the debt-to-GDP ratio. 

 

The research question is interesting; the data and sample used in the study, the econometric approach, seem to be appropriate. Still, there are several issues in this paper:

 

1. The introduction has some redundant information on the debt to GDP ratio for the MENA region. It needs to be updated. 
2. Page-2, footnote, Over the last decade, the average economic growth rates were 3.4% in MENA, 5% in ASEAN-5 and 4.7% in Sub-Saharan Africa. Providing their source would be helpful. 
3. Page-2, line 57-58, Over the last few decades, the debt-to-GDP ratio has held an average of approximately 43.7%, which is 2.1% higher than the average for ASEAN-5 and 9.9% higher than that for Sub-Saharan Africa. Please refer to the source of the data.
4. page-5, line 215, ? = ????1−?, after taking logs for equation (1), the following equation is established: ???? = ⍺ + ?????+ (1-⍺) ????, here, alpha is not the intercept, beta should be the intercept.
5. Page-5, 218, the authors use spline function, it is not clear if it is one dimensional /multi-dimensional/finite dimensional. 
6. Line 248, 264, 306: should be equation 3? 
7. Equation 4 – t subscript is missing in the error term. 
8. Line 307-311: non-significant results are not the same as the ambiguous effect. A better explanation is necessary. 
9. What kind of estimation technique is used to produce the results in Table 4? How does the methodology addresses the issues that arise in panel data estimation?
10. In line 71-74, the paper claims that they select the MENA region because these countries undertook reforms in the institutional and financial sector by decreasing borrowing to reduce their fiscal imbalances and attained economic growth through the gradual removal of trade barriers, which led to strengthened trade relationships. However, in line 317-318, the paper says that trade openness has a significant negative effect on growth, possibly because of trade barriers. These two sentences contradict each other.
11. In line 322, the authors argue that these findings are consistent with the theory. However, all of the variables are not consistent with the theory. The result shows that financial depth has a negative impact on economic growth; it contradicts the existing approach. As they use money supply M2 as a proxy variable, it would be expected that there should be a positive relationship between financial depth and economic growth. This result should be explained.  
12. Government expenditure also contradicts the existing theory as it increases economic growth; however, the current findings tell us a different story. In line 322, the authors claim that their results are consistent with the existing theory. This needs to be explained.
13. In line 361, the paper discusses the productive and unproductive use of GDP. Where does this come from? A better discussion is necessary as these recommendations are not based on the results presented in the paper. 

Author Response

Attached our response to Reviewer 2.

 

Regards.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I am happy with how the authors handled my comments. The paper is ready to go for publication.   Kind regards

Author Response

Many thanks to the reviewer for satisfied our revised version, and recommend the publication.

Reviewer 2 Report

Thank you for addressing my comments. While the Fixed Effect method is a common method of estimation, providing a little more explanation about its advantages would help the readers.  

Author Response

Many thanks to the reviewer.

Herewith is our response.

 

We have explained the Fixed Effect and Random Effect model (static panel) disadvantage and elaborate on the main advantages of using the threshold regression method in our baseline model.

 

 [refer to Page 6, Line: 187-201]