Discussion Paper

No. 2013-7 | January 23, 2013
The Impact of Financial Openness on the Size of Utility-enhancing Government


This paper studies the impact of financial openness on the size of government, and other key economic variables, such as the consumption-wealth ratio, the growth rate of wealth, and welfare, in a two-country world, based on a portfolio approach, assuming that public spending is utility-enhancing. The model suggests that the size of government, the consumption-wealth ratio, and welfare should be higher in an open economy due to a higher productivity and/or less volatility through risk sharing. The theoretical results for the growth rate depend on differences on productivities and consumption-wealth ratios. The empirical evidence based on a sample of 50 countries for the period 1970-2009 broadly supports the main theoretical results of the model, even though the inclusion of Singapore distorts sometimes the broad picture.

Data Set

JEL Classification:

F41, F43


  • Downloads: 2136


Cite As

Iñaki Erauskin (2013). The Impact of Financial Openness on the Size of Utility-enhancing Government. Economics Discussion Papers, No 2013-7, Kiel Institute for the World Economy. http://www.economics-ejournal.org/economics/discussionpapers/2013-7

Comments and Questions

Anonymous - Referee report
April 03, 2013 - 10:09

I agree with the author that the impact of financial openness on government size has not been studied extensively, and therefore a combined theoretical and empirical paper in this area is certainly timely and welcome. The model considers a real economy only, so it does not allow for nominal shocks ...[more]

... to the economy. I would consider nominal shocks and volatility in international capital flows significant factors affecting the real economy, so at least a bit more discussion of this limitation of the model should be added to the paper.

The model provides a framework with a utility function with constant relative risk aversion, the production function is linear, the stochastic processes are stationary, and hence all the real assets growth at the same rate, and we have constant portfolio and consumption-wealth shares. These seem to be rather standard results, so the contribution to the literature should be outlined more clearly, especially with respect to Turnovsky (1999).

Further, the manuscript reads more like a dissertation rather than a journal submission since many of the sections could be cut out or streamlined to provide a more focused text. For example, section 2.4 on the foreign economy seems entirely redundant since the results are analogous to the domestic economy. Many other parts of the text, in both the theoretical and empirical sections, could be streamlined by removing the repetitive parts and/or consolidating their presentation.

At the same time, I would like to see some additions to provide intuitive explanations for the results obtained rather than having a somewhat mechanical account of them. Relevant examples (not an exhaustive list) include:
- towards the end of p. 28 the sentence “However, this evidence suggests how apparently contradictory results can be reconciled.” – should be expanded to clarify the argument.
- on p. 17 after equation (26) “No clear results can be given on whether the variance of the growth rate is lower for an open economy or not. However, one can think that it will be lower in an open economy since it is more able to diversify away country-specific risks.” I think empirical evidence in this regard is more mixed.
- On p. 21 just before section 5 starts “Thus, the growth rate will be higher in a closed economy than in an open economy for similar productivities, but the opposite may also be true.” This part would also benefit from more explanation.
- On p. 23 at the end of the long paragraph “Domestic wealth, W, is constructed according to equation (1), and we also add the net foreign asset position of the country on empirical grounds.” I think more detail would be helpful in this regard.

The empirical part of the paper also raises several questions. While the coefficient on the measure of financial openness is highly significant in most cases, it is odd that none of the other control variables seem to be statistically significant under any of the specifications. I do not know if this is an omission of the indicators of statistical significance (the stars) because of some pre-occupation with the variable of interest with respect to financial openness, or if indeed none of these control variables exhibit any statistical significance. If it is the former, then a simple correction is in order, but if it is the latter, I think a second look at the data and the methodology is needed since this result is rather unusual compared to other papers in the literature. I do not object in principle to using ratios with respect to wealth rather than GDP, but data on wealth might be more unreliable in practice due to measurement problems, more assumptions employed, etc. Perhaps using logs of some of the control variables might be a better fit as well.

I also have concerns about using simple OLS, no time lags for the explanatory variables, and hence about potential endogeneity problems.

Similarly, using stocks of international assets and liabilities (relative to domestic wealth) as measures of financial openness can conceal a great deal of volatility in capital flows, while volatility in international capital flows can have significant effects in itself. Therefore, a discussion of the potential impact of using stocks rather than flows (such as gross private capital flows) to measure financial openness should be included. But in general, I agree with using de facto measures rather than de jure measures of openness.

On the issue of Singapore: I think Figure I clearly shows that Singapore should be excluded from the regression analysis. It is not uncommon to exclude very small economies from analysis in cases like this. This also leads me to comment on the choice of countries selected for the analysis. At the bottom of p. 22 the author provides the list of countries included and a reference to “following seminal contributions in this area”. I think matching the sample in order to have comparable results is one thing (although I think the results in this case are not directly comparable because of using wealth rather than GDP to benchmark the variables), but larger sample in general should be preferable. I also feel from a pedagogical point of view that it is unhelpful to students to present the choice of sample as a tribute to “seminal works” rather than justifying the selection on methodological grounds. I discourage students from making references to authors’ bios in research seminars (despite their inclination to do so) as it might introduce bias and discourage critical thinking when analysing studies in the literature.

Inaki Erauskin - Reply to the referee
May 16, 2013 - 15:16

Let me thank you for all your comments. Please find enclosed my reply. Best regards,