The empirical analysis of this chapter begins with simple, one variable regressions of the legal measures against the ultimate financial variables of concern. Given the very large numbers of variables in the analysis and enor- mous numbers of permutations of independent variables, we start by looking for some basic associations. While this study includes no intersect- ing relationships or control variables, that mere fact has some value. The law is constitutive of much in society and is likewise somewhat endogenous to societal preferences, and legal variables surely have a cybernetic associ- ation with one another and with potential independent control variables.
To the extent that additional variables are endogenous of the law (or vice versa), they will not be truly independent checks on the effects of the law.
Hence a simple regression is of some value.
We begin the simple regression with a consideration of the basic foun- dational legal variables. Because the basic law and its enforcement governs and theoretically facilitates all business transactions, one might expect that the recognition and enforcement of property rights and contracts would associate with richer financial markets. Similarly, one would expect that the actual operation of the rule of law, with consideration for judicial efficiency and judicial independence, might also yield stronger financial markets, because they presumably mean that the law on the books is better trans- lated into practice. And regardless of the substantive content of the law on the books, the rule of law should provide economic benefits by creating some stability and avoiding costs associated with corruption.
The first analysis involves the correlation among the basic foundational legal variables themselves. Different studies have used different combina- tions of these variables, but collinearity between the scaling of the variables may obscure the truly significant determinants. Intercorrelations among the variables should be identified as a threshold matter. Bivariate Pearson correlations were taken for the paired foundational legal variables, and Table 5.1 reports their statistical significance. Those positive correlations significant at 0.01 are marked with , those significant at 0.05 were marked with , and insignificant associations were marked with . In this analysis, none of the correlations were significant in a negative direction.
Plainly there is a high intercorrelation among the foundational legal vari- ables, with most significant at the 0.01 level. The procedural legal variables (such as judicial independence, judicial efficiency, and rule of law) not only significantly correlate with one another, they also significantly correlate with the more substantive legal variables (such as property rights and enforcement of contracts). The one foundational legal variable that is not significantly associated with the others is English origin, which, ironically, is the variable Table 5.1 Associations among legal foundational variables
origin riskprop riskcon prop cim rol wbrol icrgrol judeff judindep
origin —
riskprop —
riskcon —
prop —
cim —
rol —
wbrol —
icrgrol —
judeff —
judindep —
that has been most widely used in the cross-country studies of the effect of the law. This gives some reason to question whether legal origin best captures the role of the foundational law but also gives some confidence that its significance can indeed be isolated from other measures of the law, so that the true practical importance of legal origin can be assessed.
The next analysis continues the simple regressions of the foundational basic legal variables but introduces financial measures as independent vari- ables. Each of the foundational legal variables is regressed as an indepen- dent variable against the World Bank’s measures for financial markets as dependent variables. These variables are all traditional quantitative mea- sures offinancial markets. Again, statistically significant bivariate correla- tions are reported with or . For each result, a better performance (that is, lower overhead costs) merits a .
The table demonstrates a consistent and strong statistical association between most of the basic foundational legal variables and sundry measures of financial markets. The associations were highly significant and consis- tently positive in direction. The only independent legal variable lacking in consistent positive significance was English legal origin. None of the vari- ables explained the relative degree of stock market turnover, though. Both the substantive legal measures and the purely procedural assessments of the legal system showed a significant association with larger and more efficient financial markets, as measured by the World Bank variables. This finding is especially significant, because the foundational law would also be expected to affect the overall GDP, which is the denominator in many of these ratios.
It appears the law has a positive effect on financial markets that surpasses its general economic value to GDP. The next analysis considers the effect of the same basic foundational variables on other non-World Bank financial variables from a variety of sources, reported in Table 5.3. These sources Table 5.2 Legal foundational variables and World Bank financial variables
ll/gdp bvic/gdp ohc nim mcap/gdp smv/gdp smt/gdp
origin
riskprop
riskcon
prop
cim
rol
wbrol
icrgrol
judeff
judindep
include both the La Porta, et al. measures and international business survey responses on the cost of capital and adequacy of national capital markets.
The effect of the basic foundational legal variables on these measures is not as strong as was the case with the World Bank variables, but the results are typically positive and relatively consistent in their pattern. The vast majority of the variables had a strong positive association with perceptions of lower cost of capital and adequate supply of capital, with lessened market volatility and with the increased proportion of Worldscope firms to domestic firms. None of the variables were significantly associated with the measures of market capitalization to sales or cash flow or cash flow to price.
The latter absence of an association is rendered much less troubling, though, by the consistent significant positive associations of these variables with the quantitative World Bank measures of markets in Table 5.2. While the reason for the lack of association in this table is not perfectly clear, it might be that the foundational variables are affecting both the numerator and the denominator in the same positive direction and similar amount, so that they do not show up as a significant determinant of the ratio.
Because these are simple regressions, and because the independent legal variables are highly intercorrelated, one should be cautious about drawing any significant conclusions at this point. A significant association with one legal variable might be attributable to another collinear legal variable or to some yet unmeasured third factor. In addition, the absence of an associa- tion of the basic foundational law with some measures offinancial markets leaves some doubt about the nature of the positive effect from this law.
On balance, though, the evidence on the financial value of the basic foun- dational law is strong. None show any significant negative association, and all but legal origin and contract-intensive money have a pattern of highly significant positive associations for a majority of the financial variables of Table 5.3 Legal foundational variables and other financial variables
mcap/s mcap/c wldomes coc adeqcap cash/price mvol
origin
riskprop
riskcon
prop
cim
rol
wbrol
icrgrol
judeff
judindep
interest. For these variables, we reach some preliminary tentative conclu- sions. First, we find that legal origin is not a fruitful test for legal systems’
effect on financial markets, at least given the available alternative measures.
Second, we find that both the substantive and procedural basic founda- tional legal rules strengthen national financial markets, though the high association between procedural protection and substantive protection leaves the independent importance of each in some doubt.