Borrador

Original Text:

The estimation results presented in this section are related to our hypothesis seeking to prove that over time, the efficiency of For-Profit MFIs tends to be higher that that of Non For-profit MFIs. The analysis started with the estimation of a unitary model (pooled model) and two separate models (For-Profit and Non For-Profit). With these estimations, we aimed to test whether the two separate models were nested in the unitary model and were therefore using the same technology, or whether they were two different groups which used different technology. The results of the likelihood ratio test yielded statistically significant differences between groups as the mean Chi2 (14) value for ten specifications was of 145.84. Consequently, the two separate groups (For-profit and Non For-profit MFIs) were non-nested, and therefore the they used different technologies.


The procedure we used to generate these results was carried out by the parametric Stochastic Frontier Analysis (SFA) assuming a truncated distribution (trunc) of the error term where we included a vector of explanatory variables (Zi) that we think may have an impact on the efficiency scores of the production function. In addition, whilst estimating a unique production function with only one output and three basic inputs can lead to misleading conclusions, we discard this possibility by estimating ten different models by varying the main outputs and one of the inputs to captured the sensitivity of our results to changes in the outputs (yi) or in one of the basic inputs (xi) that enter the production process as pervious studies of this kind do.

Sections 8.1 to 8.3 below provides the estimation results of three different efficiency measurements, the intra-firm efficiency (IA), the Catch-up component (CU) and the Inter-firm efficiency (IE), and Section 8.4 summarize the results for the overall efficiency measurements we estimated and presents a ranking of MFIs.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Revised Text:

The estimated results presented in this section are related to our hypothesis seeking to prove that, over time, the efficiency of For-Profit MFIs tends to be higher than that of Not For-profit MFIs.

The analysis started with the estimation of a unitary model (pooled model) and two separate models (For-Profit and Not For-Profit). With these estimations, we aimed to test whether the two separate models were nested in the unitary model, and were therefore using the same technology, or whether they were two different groups which used different technologies.

The results of the likelihood ratio test yielded statistically significant differences between groups, as the mean Chi2 (14) value for ten specifications was 145.84. Consequently, the two separate groups (For-profit and Not For-profit MFIs) were non-nested, and therefore they used different technologies.

The procedure used to generate these results was carried out using the parametric Stochastic Frontier Analysis (SFA), assuming a truncated distribution (trunc) of the error term where we included a vector of explanatory variables (Zi) which we think may have an impact on the efficiency scores of the production function.

In addition, while estimating a unique production function with only one output and three basic inputs can lead to misleading conclusions, we discarded this possibility by estimating ten different models, varying the main outputs and one of the inputs to demonstrate the sensitivity of our results to changes in the outputs (yi) or in one of the basic inputs (xi) that enter the production process, as previous studies of this kind have done.

Sections 8.1 to 8.3 below provide the estimated results of three different efficiency measurements; the intra-firm efficiency (IA), the Catch-up component (CU) and the Inter-firm efficiency (IE). Section 8.4 summarize the results for the overall efficiency measurements we estimated and presents a ranking of MFIs.

Editor:

1. Please review the text carefully to ensure your intended meaning was not mistakenly changed.

2. Before you leave this page, please take a moment to click on the "Click here to post comments." link at the bottom of the page to rate the proofreading and leave a comment if you wish.

Proofreading and editing pages that receive ratings and comments are moved to the top of the list.