Parametric and Non-Parametric Methods of Efficiency Analysis of Company

Parametric Methods of Efficiency Vs Non-Parametric Methods of Efficiency

Parametric and non-parametric methods of efficiency measurement have gained significant momentum in the current decade. However, the choice of the method for efficiency management has been a matter of lengthy debate. The parametric method has the advantage of being free from the random noise of inefficiency. However, separating true noise from the process may be a restrictive process.

The non-parametric method has the advantage of justifying the axiomatic properties. It also does not have as many restrictive properties as that in the parametric method. However, non-parametric methods do not differentiate between true inefficiency and statistical noise. Therefore, just like the parametric method, it also has both advantages and disadvantages.

In contrast to parametric methods, non-parametric methods of efficiency analysis are based on the hypothesis that efficiency is a result of the assimilation of empirical results of the most efficient DMUs (Decision-making units) and/or from the benchmarks.

Some of the advantages of Non-Parametric Methods of Efficiency Analysis are as follows −

  • The calculations are easier to estimate

  • The data can be qualitative in format, and does not need to be quantitative

  • The data can be in ordinal ranking

  • The method does not have as many restrictions as in the case of parametric efficiency analysis.

However, the non-parametric method of efficiency analysis is deficient because it does not rely upon all the data obtained as a reference, they do not use all information from the sample.

Nature and Types of Parametric Methods of Efficiency Analysis

The parametric efficiency technique can be divided into three major frontiers, all of which need a selective functional form to gain the cost of profit frontier.

These are as follows −

  • Distribution-Free Approach (or DFA)

  • Stochastic Frontier Approach (or SFA), and

  • Thick Frontier Approach (or TFA).

The production function in parametric efficiency analysis is defined by a set of explanatory variables. Two components of the composite error terms of the regression and the inefficiency term are also used in the analysis.

Distribution Free Approach (or DFA)

Distribution Free Approach is used in panel data and also used in the relaxation of composite error terms of assumptions of distributions.

Here, the core inefficiency is distinguished from any random error by assuming that core inefficiency is being persistent over time, whereas random errors tend to be run-of-the-mill over time.

Stochastic Frontier Approach (or SFA)

Stochastic Frontier Approach treats the deviation arising from production function that comprises both random error and inefficiency. This is why Stochastic Frontier Approach has a two-sided distribution of errors and one-sided distribution of inefficiency.

Thick Frontier Approach (or TFA)

Thick Frontier Approach does not impose restrictions on composite error terms, but it considers inefficiency at the highest and lowest efficiency quartiles to be different. Thick Frontier Approach also assumes that the random error is in these two quartiles.

All of these approaches have potential specification errors as the specified cost of profit is approximate to the true counterpart.

Nature and Types of Non-Parametric Methods of Efficiency Analysis

There are two approaches to the non-parametric efficiency analysis. They are as follows −

  • Data Envelopment Analysis (DEA) and

  • Free Disposal Hull (FDH) method.

Data Envelopment Analysis (DEA)

Data Envelopment Analysis was developed to measure the performance of various non-profit organizations. These organizations were highly resistant to traditional performance calculation techniques. This was so because the organizations had complex and often unknown inputs and outputs, and non-comparable items that had to be considered to get a complete picture of the condition of efficiency of operations.

In the case of DEA, price details are not often available or reliable. However, as there are differences between the objectives of private and public organizations, the common ground to test their performance is technical efficiency.

The DEA model is non-stochastic, so there is enough noise present in the samples of the non-parametric efficiency model.

Free Disposal Hull (FDH) Method

The Free Disposal Hull (FDH) method for measuring efficiency is a nonparametric method that relaxes the convexity of the Data Envelopment Analyses (DEA) model to a large extent. It is the most used nonparametric method of efficiency measurement analysis.

Free disposal hull method is of two types − the input-oriented model and the output-oriented model. The version that tends to minimize inputs while maintaining a given set of outputs is known as the input-oriented model. The other version that aims to maximize outputs without needing inputs is known as the output-oriented model.

Free disposal hull method is mostly used to obtain estimates of farm efficiencies. It is particularly useful tool for analyzing public sector efficiency questions. Moreover, FDH method of efficiency analysis requires minimal assumptions with respect to the production technology.


Both parametric and non-parametric methods of efficiency analysis are important. However, their use depends upon their attributes and qualities in order to fit the best with the samples and to utilize the data most prudently. The idea is to have the least noise and the best outcome with the least number of errors.