Page 1 of 22

Journal for Studies in Management and Planning

Available at http://edupediapublications.org/journals/index.php/JSMaP/

e-ISSN: 2395-0463

Volume 02 Issue 2

February 2016

Available online: http://edupediapublications.org/journals/index.php/JSMaP/ P a g e | 342

The Impact of Agricultural Mechanization on

Employment. Substitution and Complementarity in Spain.

Pedro Arias1 ; Silverio Alarcón2 ; María Botey3;

Universidad Politécnica de Madrid Ciudad Universitaria s/n 28040 Madrid (Spain)

(pedro.arias@upm.es); (silverio.alarcon@upm.es) ; (maria.botey.fullat@gmail.com);

Abstract

This study is focused on analyzing the

relationship between two factors traditionally

seen as substitutable, namely mechanization

and agricultural employment. To this aim, we

have estimated mechanization levels

according to the number of registered

tractors and employment levels according to

personnel employed in agriculture. The

methodology used is based on the univariate

analysis of time series in order to assess their

individual behavior. Following that,

causality amongst these variables was

analyzed to establish their dynamic

relationship by defining the transfer function.

Key words: agricultural

mechanization, employment, time serie,

causality, transfer function.

Classification JEL: C53, E27, Q16

Abbreviations used: INE (Spanish

National Institute of Statistics); EPA (Labor

Force Survey); ARIMA (Autoregressive

Integrated Moving Average Models); AR

(autoregressive component); MA (mobile

component); SMA (seasonal mobile

component); MAGRAMA (Ministry of

Agriculture, Food And Environment); ADF

(Augmented Dickey-Fuller test).

1. Introduction

In general terms mechanization can be

defined as the regular use of machinery

equipment and installations in

productive activity. In the agricultural

context, the FAO (2006) defines under this

heading the manufacturing, distribution and

functioning of all kinds of tools, implements,

machinery and equipment used for land

preparation, agricultural production,

harvesting and primary processing.

Agricultural mechanization is – along

with the development of further agricultural

techniques – one of the productive factors

with the highest impact on the growth of the

world’s agricultural production, as declared

by the FAO (2000). In the specific case of

Spain this has entailed a process of

agricultural transformation with social and

economic effects. It has increased

productivity but at the same time, and

according to numerous authors including

Daly (1981), Garcia et al (1985), Picazo and

Reig (1990), Márquez (1999), Singh (2006),

Akinbamowo (2013) and Zhizhang and

Hanlin (2014), it has led to a decrease in the

size of the workforce. Other authors have

diverging opinions as mechanization and

other factors reduce work in certain areas but

the increase in production, productivity and

crop intensification allows for absorbing the

employment loss in other agricultural

activities, as shown by Rochin (1978),

Reimer (1984), Sindhu and Grewal (1991)

and Verma (2005).

Page 2 of 22

Journal for Studies in Management and Planning

Available at http://edupediapublications.org/journals/index.php/JSMaP/

e-ISSN: 2395-0463

Volume 02 Issue 2

February 2016

Available online: http://edupediapublications.org/journals/index.php/JSMaP/ P a g e | 343

In fact, the decrease in the size of the

workforce cannot be attributed solely to

mechanization given that there are other

variables which have also had an effect on

this loss. Escudero (1986) links it to farmers

quitting agricultural activity due to retirement

or transfer to other sectors. Other authors like

Garrido and González (1992) note the

influence on the mechanization effect of a

series of factors and administrative measures

aiming at lowering the number of farms.

Further studies, (Leguina and Naredo, 1973;

Naredo, 1974; Ortega, 1983) have shown that

mechanization was caused initially by

migratory dynamics, which since the 1950s

have been gradually diminishing workforce

availability in the agricultural sector and

causing an increase in wage costs. This

situation favored machinery acquisition and

the use of new production techniques to

lower costs and increase productivity.

Therefore, mechanization can be both the

cause and effect of the decrease in workforce.

In short, the 1950s effectively

inaugurated a process of intensification of

machinery use parallel to a loss of size of the

agricultural workforce. This has led to more

agricultural activity with fewer farmers and

to the specialization and simplification of

alternatives.

This process of technological

transformation can be observed when we

compare the active agricultural population

and the number of tractors registered up to

the early 1980s; these variables evolved in

opposite directions, which could be

interpreted as factor substitution i.e.

mechanization replacing workforce (Figure

1).

Figure 1

Tractors (registered number)

1950 1955 1960 1965 1970 1975 1980

period

0

10000

20000

30000

40000

tractors

Active agricultural population (thousands of people)

1960 1964 1968 1972 1976 1980

period

2100

2600

3100

3600

4100

4600

5100

employment

However, there is no general consensus

regarding the cause-effect relationship

between these variables. In view of all

this and in the face of the observed

behavior of both variables, the object of

this study is to analyze their evolution

and define their causality. This paper is

then focused on the following questions:

 Estimating a time model for

mechanization and employment in

agriculture throughout the period 1985-

2014.

Page 3 of 22

Journal for Studies in Management and Planning

Available at http://edupediapublications.org/journals/index.php/JSMaP/

e-ISSN: 2395-0463

Volume 02 Issue 2

February 2016

Available online: http://edupediapublications.org/journals/index.php/JSMaP/ P a g e | 344

 Establishing the causal relationship

and defining the transfer function

between these variables.

The study is structured as follows:

section 2 presents materials and

methods, section 3 describes the

mechanization process and the

evolution of agricultural employment,

section 4 presents the estimations of the

time models for mechanization and

workforce. Finally, section 5 deals with

causal relationships and defines the

transfer model.

2. Materials and methods

Achieving the proposed objectives

requires the previous characterization of

the two variables. In the case of

workforce we have studied the

personnel employed in agriculture; in

terms of mechanization level, a series of

indicators can be found in the literature

to estimate it: (Ferrer (1978), Spoor

(1987), NCAER (1981), Binswanger

(1982), Andrés and Espejo (1989), Vico

(1996), Binimelis et al (2005), Singh

(2006), Sharabiani and Ranjbar (2008),

Ghadiryanfar et al (2009), Fadavi (2010)

and Zangeneh and Banaeian (2014)).

These indicators are basically structured

in two groups according to their nature,

which can be either technical or

economic. Within the latter group, some

authors lean towards using tractor

demand, given that it is one of the

pieces of machinery that has contributed

most to agriculture modernization

(Cooper (1994) and Pérez (1984) and is

thus a robust indicator of the

mechanization process. In view of this,

we have chosen the latter approach and

we estimate mechanization by means of

the demand for tractors and specifically

through the number or registered

tractors.

As to workforce, we have used

information provided by INE through

quarterly EPA survey. As for

mechanization, we use the tractor

registration data collected in the Boletín

Mensual de Estadística [Monthly

Bulletin of Statistics], issued by the

Spanish Ministry of Agriculture, Food

and Environment.

The methodology used to achieve the

first objective is based on the univariate

analysis of the time series. This

technique was developed by Box and

Jenkins (Box-Jenkins, 1976; Peña,

2005) in the 1960s and has since been

applied extensively in different fields,

mechanization analysis amongst them

(Kuchler and Vroomen (1987), Biondi

et al (1998), Arias (2001), Unakitana

and Akdemirb (2007) and Kim et al.

(2013)). This methodology focuses on

the univariate analysis of time series

through ARIMA models. The first stage

involves defining and estimating the

model1

and making the necessary

transformations so that the series is a

stationary process (in median and

variance). The second stage is the

1 Estimating the ARIMA model requires

the defining of the different parameters

constituting it. Thus, an ARIMA (p, d, q)x(P, D,

Q)s requires the defining of the autoregressive

component for the regular stationary part s

(orders p and P respectively), the mobile

average (orders q and Q) and the regular and

seasonal differentiations (d y D) to be used in

the series.