What does it mean to have restrictions in Indonesia, and how can a systematic procedure for determining the presence and relative level of restrictions be applied to all companies in the data set. The next explanatory variable, 𝑑!(!!!), is a proxy variable for debt, which in this case is the size of the firm's annual interest payments. The β coefficients will be an indication of the effect that the explanatory variables have on firm performance.
All variables here are the same as before, except for the presence of two interaction terms. Furthermore, adding the tariff variable as an instrument will address the problem of the interaction between productivity and export. This new p-hat probability will then be used in both initial benchmark regressions to instrument each instance of the export decision variable, including the interaction terms between the export decision and high equity/debt levels.
The linear probability model is a relatively simple application of the standard multiple regression model to a binary dependent variable. One of the main disadvantages of the LPM is that the adjusted probabilities it generates can be less than zero or greater than one. Furthermore, for the purposes of the IV regression, the p-hat value must replace the binary export decision {1, 0} in the two main regressions, so again a value greater than one will not work.
Another disadvantage of the LPM is that the partial effects of any explanatory variable in an LPM are constant.
Data
The sign of the coefficient can be used, but the magnitude must be determined by partial effect calculations that use the common scale factor derived from the regression. Different industries inherently have different capital requirements and other industry-specific considerations that affect their debt and capital levels, so industry dummy variables were one of the necessary dummy variables. Each of these individual ISIC codes was then converted into a binary variable “yes, belongs to this industry, no, does not belong to this industry” for use in the regression analysis.
Lending rates exhibit location-specific variation, so the province variable in the survey captures these local differences in economies, interest rates, and other factors such as personal relationships. First, to improve precision and aid the interpretation of the results, all data points that were regressed except for the binary variables (the export decision and dummies) were converted to log files. Labor productivity was measured as a fraction of the company's total sales divided by their number of employees.
Interest payments and capital levels needed to be associated with the export decision, a binary variable, so they were converted into their own binary variables. For each industry, the median level of capital and interest payments was determined, and then each company was compared to this value. If their capital/interest payments were above the median for their industry, that company was assigned a {1} for {high capital/debt}.
Substantial measures have been taken to reconcile debt, as this is one of the most important explanatory variables. For whatever reason, debt figures were not taken from the Indonesian companies, and the only suitable indicator found in the data was the annual interest payments on loans. Although interest payments performed admirably during the first regressions and behaved largely as would be expected: increasing as exports increased, decreasing as exports decreased, it was still critical to generate a new reasonable measure of debt burden.
The study misleadingly reported that liabilities were equal to assets at all times, so initial equity and added equity at year end were subtracted from this to arrive at a more realistic number that could be used to allocate debts. Ultimately, however, this empirical measure was not used to replace interest payments as a measure of debt because there was no compelling evidence that it was a more accurate measure. So in the end, imputed debt did not replace interest payments as a measure of debt, but it did serve as confirmation of the accuracy of interest payments because the two correlated so well.
1st Stage Probit Regression - Export Decision
Conclusion
In the thinking proposed by LBE, the decision to start exporting should result in higher turnover and productivity. On the other hand, these results were biased because there were endogeneity issues in the sample and the regression methods to account for. This was added to an initial probit regression that generated a predicted probability of export, which was used to replace the export decision in the final regression.
"Globalization" has turned into a hot buzzword, and developed and undeveloped economies alike are figuring out how best to compete in the new global marketplace. Competition in this market is critical as exports have steadily increased over the past century as a percentage of the overall economy among developed countries. Several broad aspects of the crisis were critical in each country, including the dumping of Asian assets, free-falling exchange rates and poor access to new credit.
When this speculation imploded, banks were left with long-term assets in the form of real estate that were incredibly difficult to liquidate due to weak insolvency laws. With recalibration, the government focused on three general areas of reform: greater transparency and supervision of the financial sector, more careful handling of "hot money" through measures such as the ban on certain short-term loans, and a recapitalization of balance sheets. In addition, the Banking Law Amendment of 1998 added additional long-term structural reforms, for example removing restrictions on foreign parties owning more than 49% of the shares of a domestic bank.
The net result was a 37% decline in the number of commercial banks from 1996 to 2000, and the loan-to-deposit ratio more than halved from a peak of 104% over the same time period. With the recapitalization and sale of state shares, foreign-owned banks emerged as leading players instead of banks linked to business groups. The Central Bank, which until then had only partial authority under the authority of the Minister of Finance, now has all the powers over the banks.
Coming out of the crisis and the reform it brought about, it becomes important but difficult to understand what impact the crisis had on Indonesian firms and their lending environment, especially since the survey data used spans the crisis with the response years of 1996 and 2006. There was considerable uncertainty about the future of the exit from the crisis, both political and legal, creating a lack of investment demand. High rates also prompted some to seek alternative financing, as seen in the expansion of private sector bond issuance that hit Rp 25.6 trillion in 2003, compared to Rp 6.5 and Rp 2.3 trillion in 2002 and 2001, respectively.
Microlevel Evidence from the Republic of Korea and Taiwan. The World Bank Economic Review. Banking Deregulation in Indonesia: An Updated Perspective in the Light of the Asian Financial Crisis”, 20 J.INT'L L.
