Domestic formal and informal institutions: their substitutability and comparative advantage

Abstract

This paper empirically examines how country-specific formal and informal institutions affect export patterns. The index for formal institutional quality evaluates electoral rules, judicial independence, and other constraints on executives. The index for informal institutional quality comprises degrees of trust, control, and obedience. Using the revealed comparative advantage index, I find that countries with high-quality formal and informal institutions tend to have institutional comparative advantage. Results also suggest that formal and informal institutions substitute for one another in generating institutional quality. I find robust results even when controlling for an important potential source of reverse causality.

Appendices

Appendix A Map** methodologies

1.1 A.1 Map** between the I–O classification and SITC rev. 2

1.1.1 A.1.1. Baseline methodology

To construct a map** between the I–O classification and the Standard International Trade Classification (SITC) rev. 2, I first use the concordance between the 10-digit Harmonized System (HS10) and SITC rev. 2, constructed by Feenstra (1996). The 5-digit SITC codes in this concordance are truncated to the 4-digit codes, which is further mapped to the I–O classification by the Bureau of Economic Analysis (BEA)’s concordance of HS10 and I–O classification. Then, each HS10 has a matching I–O and SITC code, which allows for counting the number of HS10 codes for each I–O code. For each SITC category, I choose one I–O code with the highest number of HS codes.

1.1.2 A.1.2. Additional map** for robustness check

For a robustness check, I consider the I–O level industry shares for each SITC category, rather than choosing one dominant I–O code. Specifically, I calculate the shares of I–O codes according to their matching HS10 numbers for each SITC level. Then, Nunn’s contract intensity measure for each I–O classification is multiplied by its corresponding industry share. Finally, by adding up these weighted intensities for each SITC code, the institutional intensity measure is listed by the SITC. The Herfindahl index (HI) and raw material intensity listed in the I–O classification are also reorganized by the SITC levels using the same methodology.

1.2 A.2 Map** between the I–O classification and 1987 SIC

The skill and capital intensities, TFP growth, and value added are organized by the 1987 4-digit Standard Industrial Classification (SIC) in Becker et al. (2013). The SIC is linked to the I–O classification by Feenstra (1997)’s concordance between the HS10 and 1987 SIC and the BEA’s concordance between the I–O codes and HS10. For each I–O code, I choose one SIC code that has the highest number of HS10. When an I–O code is matched to two or more SIC codes that have the same number of HS10, I use the average value of the intensities in the SIC codes. Choosing one dominant SIC code for each I–O level is used as a baseline map** methodology.

1.3 A.3 Map** between the I–O classification and ISIC rev. 2

The map** between the I–O classification and 4-digit SITC rev. 2, described in Appendix A.1.1, is further linked to the International Standard Industrial Classification (ISIC) using the concordance between the 4-digit SITC and 3-digit ISIC, constructed by Muendler (2009). Then, each HS10 code has a matching I–O and ISIC code.

To organize the dependence on external finance in terms of the I–O classification, for each I–O code, I choose one ISIC code that has the highest number of HS10. For the I–O codes with two or more ISIC codes, which contain the same number of HS10, I use the average value of the external financial dependences. For the data on external dependence in the 4-digit ISIC levels, I choose the I–O codes that are matched to the industry description of the 4-digit levels. Since the ISIC is in a higher level than the I–O classification, the same values of the external dependence can be matched to different I–O codes. Note that for a robustness check, the data on the external finance dependence listed by the ISIC are directly matched to the SITC by Muendler (2009)’s concordance.

1.4 A.4 Map** between the 1987 SIC and SITC rev. 2

The SIC codes are matched to the HS10 and SITC codes by Feenstra (1997)’s concordance between the HS10 and 1987 SIC and Feenstra (1996)’s concordance between the HS10 and 4-digit SITC. Then, each HS10 has a corresponding SITC and SIC code. Since one or more SIC codes are matched to a SITC level, I use the same methodology described in Appendix A.1.2. That is, by calculating the portions of SIC codes according to their matching HS10 numbers within a SITC level and by adding up the factor intensities weighted by these shares, factor intensities, including $s_z$ and $k_z$, are recalculated and reorganized by the SITC. These factor intensities are used for a robustness check.

Appendix B Additional robustness checks

See Tables 10 and 11.

Table 10 Robustness check with the SITC and Nunn’s measure

Table 11 Robustness check with the SITC and the Herfindahl index