This chapter contributes to the literature on the relationship between financial development and economic growth by including variables related to both the banking sector and the development of stock markets and by distinguishing among developed, emerging, and “frontier” stock markets. The results provide evidence that, in line with recent empirical literature, the banking sector has an overall negative impact, once the effect of stock markets is taken into account, although the results do not provide evidence of a threshold leading to instances of “too much finance.” However, in countries with emerging stock markets and particularly in countries with frontier markets, the impact of credit depends on the nature of the provider. Finally, the impact of stock markets depends on their own level of development, with a positive impact only for developed stock markets, but an insignificant impact for both emerging and frontier markets.
We have found a large effect of stock markets on subsequent development. We have failed to find a similar effect of bank lending. That this differential effect should exist is in itself surprising. But if it is true, then it is even more surprising that more countries are not developing their stock markets as quickly as they can as a means of speeding up their economic development. (Atje and Jovanovic, 1993, p. 663)
Table 14.1
List of Variables
Economic growth variable (dependent) | |
GDPPCGRO | Growth of real GDP per capita (%) |
Financial development variables (independent) | |
PRICRE | Private credit by deposit money banks to GDP (% of GDP) |
DOMCRE | Domestic credit to private sector (% of GDP) |
LL | Liquid liabilities to GDP (%) |
SECAP | Stock market capitalization to GDP (%) |
SEVAL | Stock market total value traded to GDP (%) |
SETURN | Stock market turnover ratio (%) (ie, total value of shares traded/average market capitalization) |
Instrumental variables | |
ENROLLSEC | Lower secondary completion rate, total (% of relevant age group) |
OPEN | (Exports + imports)/GDP (% of GDP) |
GOVCONS | Government consumption to GDP (% of GDP) |
INFLATION | Inflation rate via CPI (%) |
GCF | Gross capital formation to GDP (% of GDP) |
IIP | Initial real GDP per capita in USD |
Source: World Development Indicators, the World Bank.
Table 14.2
Descriptive Statistics
Developed (mean) | Emerging (mean) | Frontier (mean) | |
GDPPCGRO (%) | 1.87 | 2.55 | 2.61 |
Financial and banking sector development variables | |||
PRICRE (% of GDP) | 99.76 | 45.55 | 36.43 |
DOMCRE (% of GDP) | 114.14 | 52.73 | 41.12 |
LL (% of GDP) | 94.08 | 56.25 | 49.64 |
SECAP (% of GDP) | 83.98 | 52.22 | 29.88 |
SEVAL (% of GDP) | 72.28 | 24.51 | 8.58 |
SETURN (%) | 83.50 | 54.86 | 27.73 |
Control variables | |||
ENROLLSEC (%) | 107.72 | 78.16 | 79.81 |
OPEN (% of GDP) | 91.60 | 68.39 | 84.23 |
GOVCON (% of GDP) | 18.17 | 14.63 | 15.66 |
INFLATION (%) | 2.49 | 50.69 | 44.47 |
GCF (% of GDP) | 23.33 | 24.57 | 22.86 |
IIP (USD) | 17,691 | 3,826 | 2,542 |
Table 14.3a
Correlations of the Variables All Countries (Obs. = 1238)
PRICRE | LL | DOMCRE | SECAP | SEVAL | SETURN | |
PRICRE | ********* | |||||
LL | 0.7862 | ********* | ||||
DOMCRE | 0.9125 | 0.7418 | ********* | |||
SECAP | 0.4716 | 0.5879 | 0.5176 | ********* | ||
SEVAL | 0.4646 | 0.5470 | 0.5510 | 0.7535 | ********* | |
SETURN | 0.2734 | 0.2248 | 0.3452 | 0.1686 | 0.5542 | ********* |
Values in bold are not significant, even at 10% level of significance.
Table 14.3b
Correlations of the Variables Developed (Obs. = 492)
PRICRE | LL | DOMCRE | SECAP | SEVAL | SETURN | |
PRICRE | ********* | |||||
LL | 0.6579 | ********* | ||||
DOMCRE | 0.7891 | 0.6237 | ********* | |||
SECAP | 0.2898 | 0.6050 | 0.3147 | ********* | ||
SEVAL | 0.2556 | 0.4854 | 0.4101 | 0.7497 | ********* | |
SETURN | 0.0700 | 0.0426 | 0.2793 | 0.1023 | 0.5742 | ********* |
Values in bold are not significant, even at 10% level of significance.
Table 14.3C
Correlations of the Variables Emerging (Obs. = 383)
PRICRE | LL | DOMCRE | SECAP | SEVAL | SETURN | |
PRICRE | ********* | |||||
LL | 0.8542 | ********* | ||||
DOMCRE | 0.9149 | 0.6937 | ********* | |||
SECAP | 0.4572 | 0.3137 | 0.6146 | ********* | ||
SEVAL | 0.5033 | 0.4595 | 0.5694 | 0.6989 | ********* | |
SETURN | 0.1800 | 0.2408 | 0.1418 | −0.0674 | 0.4381 | ********* |
Values in bold are not significant, even at 10% level of significance.
Table 14.3D
Correlations of the Variables Frontier (Obs. = 363)
PRICRE | LL | DOMCRE | SECAP | SEVAL | SETURN | |
PRICRE | ********* | |||||
LL | 0.7691 | ********* | ||||
DOMCRE | 0.9793 | 0.7566 | ********* | |||
SECAP | 0.5028 | 0.5185 | 0.5022 | ********* | ||
SEVAL | 0.2499 | 0.2828 | 0.2550 | 0.7123 | ********* | |
SETURN | −0.0428 | 0.0182 | −0.0396 | 0.0642 | 0.5498 | ********* |
Values in bold are not significant, even at 10% level of significance.
Table 14.4
Growth, Stock Market Development, and Credit Allocation (Dependent Variable GDP per Capita Growth)
All | Developed | Emerging | Frontier | |
Banking development variables | ||||
PRICRE | −0.089 (−2.38) [0.017]** | −0.014 (−0.80) [0.422] | −0.20 (−3.13) [0.002]*** | −0.33 (−3.33) [0.001]*** |
DOMCRE | 0.038 (1.44) [0.151] | −0.015 (−0.96) [0.336] | 0.10 (1.62) [0.105] | 0.16 (2.41) [0.016]** |
LL | −0.094 (−3.39) [0.001]*** | −0.014 (−1.19) [0.236] | −0.025 (−0.62) [0.537] | −0.006 (−0.14) [0.887] |
Stock market development variables | ||||
SECAP | 0.021 (4.14) [0.000]*** | 0.031 (5.31) [0.000]*** | 0.025 (1.72) [0.085]* | 0.012 (0.85) [0.394] |
SEVAL | −0.001 (−0.23) [0.818] | −0.008 (−5.27) [0.0000]*** | 0.005 (0.32) [0.745] | 0.024 (1.19) [0.233] |
SETURN | 0.004 (0.52) [0.607] | 0.013 (3.15) [0.002]*** | −0.009 (−0.71) [0.480] | 0.004 (0.44) [0.661] |
Number of observations | 930 | 401 | 288 | 241 |
Wald chi-test (7) | 51.06*** | 156.93*** | 170.83*** | 64.02*** |
Arellano–Bond test for order 1 | Z = −4.45 [0.0000]*** | Z = −3.68 [0.0002]*** | Z = −3.00 [0.0027]*** | Z = −2.85 [0.0044]*** |
Arellano–Bond test for order 2 | Z = 1.76 [0.0779] | Z = −1.40 [0.1622] | Z = 1.38 [0.1669] | Z = 1.63 [0.1031] |
1. For the Wald chi-test, the null hypothesis is that none of the variables are worth including, and the alternative is that some variables are needed.
2. For the Arellano–Bond test of order 1, the null hypothesis is that the errors in the first-difference regression exhibit no first-order serial correlation.
3. For the Arellano–Bond test of order 2, the null hypothesis is that the errors in the first-difference regression exhibit no second-order serial correlation.
In parentheses are Z-statistics values, while in square brackets are the corresponding p-values.
*** Denotes statistically significant at 1%, ** statistically significant at 5%, * statistically significant at 10%.
3.143.205.27