We selected all accidents resulting in sick leave of more than one day that occurred in Spain over the period from 2013 to 2018. We took the data from the reports that companies sent to the Ministry of Labour. There was a total of 3,420,087 reported accidents in Spain during this period, 158,166 of which were in the agricultural industry (NACE codes included are 011, 012, 0,13, 015, and 016, as we also report below) (see Table 1) (EUROSTAT, 2008). The high number of occupational accidents allows us to detect small differences that are statistically significant, therefore relevant. Although completing occupational accident forms is mandatory in Spain, it may be possible that some accidents have not been properly reported (Arana et al., 2010).

Additionally, Spain’s health authorities are required to diagnose the severity of each occupational accident. For this purpose, the severity of an accident can be classified into three different levels: minor, serious, or fatal. Of the total accidents analysed for the agricultural sector between 2013-18 (158,166), 156,065 accidents (98.67%) are classified as minor, 1,828 (1.16%) as serious, and finally, 273 accidents (0.17%) as fatal (Baraza & Cugueró-Escofet, 2021).

At the level of frequency accident, it is important to show the information in the form of total (AR_total) and fatal (AR_fatal) accident rates. The definition of these accident rates is as follows:

The most relevant information for the comparative study between Spanish and immigrant groups are the working days lost as a cause of the accident. As the objective was to analyse average values and avoid distortions due to accidents with more lost working days, the fatal accidents that are reported with a day off work have been eliminated from the sample, as well as the 99th percentile (equivalent to more than 468, therefore the study considers accidents with one to 467 lost work days).

The group variable considered was the origin of the injured worker, differentiating between native Spaniards and immigrants. The comparison variables were categorised into five groups that are explained and detailed below: (a) personal, (b) business, (c) material, (d) temporal, and (e) geographical (Camino-López et al., 2008). The variables chosen are presented grouped in the proposed categories in Table 2.

Personal variables (a) are the characteristics of the injured worker, including age and nationality. Business variables (b) include the type of contract and the seniority of the worker involved in the accident, the number of workers in the company employing him, and the agricultural activity he performs. Material variables (c) include aspects that are unique to the accident, such as how it happened, and the type of injury suffered. Temporal variables (d) refer to the moment when the accident took place (day of the week and hour of the day). The geographical variable (e) used in this study describes the severity of accidents according to the geographical area where they occurred, in this case the regions of Spain.

The statistical package used for the analysis is Stata version16/MP. The average number of working days lost per group (Spanish and immigrants) was calculated according to gender, age, size of the company, seniority of the worker, the cause of the accident, injuries resulting from the accident, the day of the week, hour of the day and the region of Spain in which the accident occurred. For each case, a mean comparison analysis is performed using Student’s t-test for independent comparing nationals and immigrants for each variable. We based our analysis on the method used by Fontaneda et al. (2019), who used the days lost as a proxy to determine the severity of the accident, so as the independent variable measuring this severity. We then compared the days lost (proxy for severity) between the two groups we were interested, immigrants vs nationals. And we compared these days lost between the two groups, by using several dependent variables (gender, age, size of the company, seniority of the worker, the cause of the accidents, injuries resulted, day of the week, hour of the day and the region of Spain the accident occurred) to examine significant differences of days lost/severity between the two groups. We examined which variables were determining when we studied the severity of the accidents comparing immigrants vs nationals.

Agriculture is a significantly male-dominated sector (Alamgir et al., 2009): during the period 2013-2018, for every 100 accidents in the agricultural sector in Spain, 20 involved women and 80 involved men. Table 5 shows that the average number of days lost due to work accidents was significantly higher in the case of Spanish workers compared to foreigners for both men and women, although the difference was greater in the female group.

In gender, the difference comes mainly from the fact that immigrant women showed fewer days lost compared to their male counterparts showing a gender effect against immigrant women. This aspect is consistent with other studies that have analysed occupational accidents in relation to gender; if immigrants are exposed to worse conditions, for immigrant females this is even worse, and one conclusion may be that women come back to work without being completely recovered from the accident (Lin et al., 2008; Fontaneda et al., 2019).

A continuous increase in the duration of sick leave can be observed both for native and migrant workers as they age (see Table 6). A further increase in this duration can be observed for the groups aged over 55 in the case of Spanish workers. The results for the days of work lost by age of the injured worker were in line with those of other studies, highlighting a longer duration of sick leave among older workers (Blanch et al., 2009; Fontaneda et al., 2019) and, specifically, in the agricultural sector (Pickett et al., 2008). The increase in the duration of sick leave with age could be due to a reduction in musculoskeletal, sensory, and motor abilities among older people, which may affect women disproportionately (Laflamme et al., 1996). In general, for older people recovery is usually slower (Margolis, 2010).

Comparison of Spanish workers with immigrants by age group shows that there was a significant difference (p<0.005) in favour of Spanish workers who have more days of sick leave. This difference increased with age: 1.30 days for workers under 24, 1.08 days for ages 25 to 34; 3.23 days for ages 35 to 44; 7.57 for ages 35 to 54; and 8.69 for those 55 and older.

Table 7 presents Morocco as the country of origin with the highest number of accidents, with 14,850 (37.9%), followed by Romania (9,210 accidents, 23.5%) and Ecuador (4,221 accidents; 10.8%). These nationalities coincide with highest immigration to Spain, due to the scarcity of native workers in the agricultural sector (Gadea et al., 2015).

There were no significant differences with the number of days lost due to an accident at work, exception for Ecuador that is well below the average (27.48 days), with 23.60. From the analysis of fatal accidents, looking at the difference between the fatal accident rate minus the total accident rate (%FAR-%TAR), there is a significant reduction for Morocco (-19.5%), which is offset by the case of Romania (+9.1%) and Bulgaria (+16.7%). This data indicates that Easter Europe countries have more serious accidents, an aspect reflected when observing the number of days lost due to an accident (29.80 days for Romania and 30.80 days for Bulgaria). The special incidence of severity of accidents for immigrants from Eastern Europe has already been studied in other countries (Guldenmund et al., 2013). The rest of the countries present a difference rate (%FAR-%TAR) around 0, which indicates a logical proportionality between the number of total accidents and fatalities.

Table 8 shows the average number of days lost due to work accidents in agriculture by nationality and size of company. Lack of experience is frequently cited as one of the causes generating the highest frequency of accidents in different sectors (Cattledge et al., 1996; Bande & López Mourelo, 2015).

Regarding the size of the company, in all cases, the average number of workdays lost due to accidents among Spanish workers was significantly higher than that of foreigners. The highest average number of workdays lost due to accidents occurs in companies with fewer than 5 workers. The number of days lost decreases as the size of the company increases, especially in companies with more than 100 workers and even more in the largest companies (more than 250 workers). This situation shows how workers in larger companies are less likely to suffer accidents, and that, if they occur, they are less serious (Salminen et al., 1993; Fabiano et al., 2004). These differences by company size reflect work risks, the efficiency of safety programmes, and many factors in the work environment that are clearly more favourable in larger companies (Jeong, 1999).

The comparison of the number of days lost due to an accident between Spanish workers and immigrants according to the size of the company shows in all cases a significant difference in most cases in favour of Spanish workers who have more lost days. This difference decreases as the size of the companies increases, going from 8.16 days in companies with less than 5 workers and reaching an equilibrium situation (there is no difference) in companies with more than 250 workers. Larger companies have better formal justice and equality conditions as they are easily forced by authorities to comply with the law. This type of formal justice does not need to depend on managerial decisions, as legal requirements are clear, and therefore obliging larger companies to fulfil these legal requirements is usually easier as this aspect is usually watched closely by unions and workers’ representatives. This means better working conditions both for Spanish workers and for immigrants, and similar efficiency of safety programmes for both groups. In general, managerial decision-making in terms of generating justice in decisions is greatly dependent on informal justice, but once this informal managerial justice combines with social pressure and translates into formal aspects of required enforcement, then these few, clear requirements are precisely the ones that guarantee equal conditions of justice for all (Cugueró-Escofet & Rosanas, 2013; Cugueró-Escofet & Rosanas, 2020). This is the case of legal enforcement of safe conditions at work. And as large companies are better-aligned with legal requirements such as workplace safety, this means that larger companies are better at providing equal condition for nationals and foreigners.

Table 9 shows the effect of the worker’s experience on the severity of accidents, differentiating between Spanish and immigrant workers.

In reference to the effect of the worker’s experience on the severity of accidents, in both groups, Spanish workers and immigrants, a greater number of accidents was observed in workers with less experience (up to three months). This aspect is related to the job insecurity of the sector and the short-term contracts associated with harvesting campaigns (Gadea et al., 2015). The number of accidents fell substantially in workers with 1 to 4 years of experience, which may be related to the departure from the agricultural sector of a significant part of the workers who have been involved in the sector only in a circumstantial way. Finally, an increase in the number of accidents from 5 years and up to 30 years of experience is once again noted, a group that corresponds to those who have consolidated their employment in the agricultural sector.

The comparison of the number of days lost due to an accident between Spanish workers and immigrants based on the experience of the injured worker shows in all cases a significant difference (p<0.005) in favour of Spanish workers, who have more days of sick leave. Due to the lack of significance, the case of workers with more than 30 years of experience is not considered in this case. More experience should imply more training and experience in the work being done and, consequently, a lower probability of accidents (Llacuna et al., 2013).

Table S1 [suppl] (ordered by the total number of accidents) shows the aggregate data of the deviations causing accidents in the agricultural sector. The classification of the deviations follows the methodology of European Statistics on Accidents at Work, or ESAW (EUROSTAT, 2013).

The type of accident has been frequently discussed in the literature (Melchior & Ruviaro, 2019). In the case of agriculture, the main types of accident that have been analysed are those that occur due to overturning during the handling of heavy machinery (tractors), since these accidents are more serious (Arana et al., 2010; Valero & Abril, 2016) and are related to older and more experienced workers, who are usually the workers who use tractors (Goldcamp et al., 2004; Arana et al., 2010).

Accidents caused by the “loss (total or partial) of control of the means of transport or cargo (with or without motor)” (code 042) are those that present the greatest severity in terms of fatal accidents (a total number of 107 fatal, which 50 were not in itinere) that are not included in this study. In the case of non-fatal accidents, the number of days lost due to an accident in the case of Spanish workers is 46.00 days on average and that of foreign workers 42.33 days. These results may be conditioned by the underreporting of accidents in the agriculture sector; according to the study by Arana et al. (2010), not all the fatal accidents associated with the use of the tractor have been reported, and it establishes that in the period 2004 -2008 only 61.88% of the total number of the accidents that occurred in Spain were officially registered.

Other deviations that should be highlighted in terms of the number of days lost due to accidents correspond to falls at different levels and at the same level. “Fall of a person from a height” (code 051) with 52.70 days lost on average by Spanish workers and 41.53 by immigrant workers; “Fall of a person at the same level” (code 052) with 43.28 days in Spanish workers and 35.44 in the case of immigrant workers. A significant difference (p<0.005) could be observed in accidents due to falls at the same level or at a different level in favour of Spanish workers, who had more days of sick leave. In the case of accidents due to the “loss (total or partial) of control of the means of transport or cargo (with or without motor)” the conclusion is that there was no difference between the days lost by Spanish and foreign workers. For this last type of accident, there are several studies that work to improve the intrinsic safety conditions of this machinery (Di Nocera et al., 2018; Rondelli et al., 2018; Kim et al., 2019).

Only the case of “Caught between objects” (code 063) shows a greater number of days lost due to accidents in favour of foreign workers (43.46 days lost) compared to Spanish counterpart (42.66 days lost), although the difference was not significant.

Table S2 [suppl] shows the aggregate data of injuries derived from the various accidents that occurred in the agriculture sector. Injuries are categorised according to ESAW (EUROSTAT, 2013).

In reference to the type of injuries, in accidents with major damages, the difference in days lost due to sick leave is lower than in less significant accidents. The situations of “bone crushing” (code 020) and “amputations” (code 040) of greater severity presented minor differences between Spanish and foreign workers, 81.77 days lost compared to 75.43 days respectively in the case of “bone crushing” and 87.42 days lost compared to 84.06 days lost in the case of “amputations”.

Conversely, accidents with less serious consequences (number of days lost due to accidents) show a greater difference between Spanish and foreign workers in favour of the former group. This is the case of “extreme temperatures effects” (code 100) with 42.00 days lost and 18.63 days lost and the case of “effects of noise, vibration and pressure” (code 090) with 28.81 days lost and 12.85 days lost, and the case of “drowning and asphyxiation” (code 080) with 17.36 days lost and 7.63 days lost, respectively.

These results indicate that, for the same injury derived from the accident, in general terms, Spanish workers take more days off work with less intrinsic severity compared to immigrant workers. This situation is related to a greater labour presenteeism (Mateo, 2017) associated with ignorance of their rights and protections at work (Smith & Mustard, 2010) and to more unfavourable benefits during the period of sick leave (Solé & Rodríguez, 2010) for the case of immigrants. This situation cannot occur in more severe accidents, associated with an inability to work, which places foreigners nearer to the same level as nationals.

In Spain, accidents classified as “Heart attacks, strokes, and other non-traumatic diseases” (code 130) are considered work accidents, except when there is evidence to consider otherwise. In this case, no difference was observed between the days lost due to an accident between the two groups under study, an aspect that is related to the less clear working nature of the accident. The average number of days lost by Spanish workers was 180.47 days and by foreigners 182.50.

Table 10 presents data on the average number of days of work lost by nationality by the day of the week the accidents occur.

In reference to the day of accident occur (see Table S3 [suppl]), the average duration of sick leave steadily increased as the week progressed from Monday to Sunday, and this was the case for both groups (nationals and immigrants). We also found statistically significant differences for both groups when comparing successive days of the week (p<0.005) and confirming an increase in the number of days lost in both cases. This situation of greater number of days of absence in accidents that occurred during the weekend can be considered in relation to injuries that are difficult to diagnose (Moral de Blas et al., 2012; Fontaneda et al., 2019).

Table 10 also shows the evolution of the number of accidents, highlighting the following aspects: first, the reduction of these during the weekend; second, the highest number of accidents registered on Mondays; and third, from Tuesday to Friday the number of accidents remains stable. This situation can be considered normal as established by various studies carried out in various industrial sectors such as construction (Camino-López et al., 2008). The higher accident rate on Monday can be attributed to the so-called “Monday effect”, associated with the circumstance that some of the accidents that occur on weekends are not reported until Monday, the first working day of the week, due to the worker’s social benefits in relation to the insurance company (Campolieti & Hyatt, 2006; Butler et al., 2013).

Table S3 [suppl] shows the duration of sick leave due to accidents comparing the time of the day the accident occurred. The differences between Spanish and foreign workers were statistically significant (p<0.005), except in accidents that occurred between 2 a.m. and 7 a.m. and between 10 p.m. and midnight. In all time slots, the average number of lost workdays for foreign workers was less than for Spanish workers.

Table S4 [suppl] shows that the distribution of accidents is not the same in all the regions of Spain. There was a greater number of accidents concentrated in the southern (Andalusia) and eastern (Valencia and Murcia) areas of Spain, coinciding with the main agricultural areas. In most regions the number of days lost due to accidents was higher for Spanish workers, except in some regions, such as Madrid or Valencia, but without statistically significant results regarding the difference presented. Some regions stand out with greater differences between the number of days of sick leave between nationals and foreigners, such as Aragón (with a difference of 9.09 days), even if in this case more accidents have been reported for foreign workers than for Spanish ones. Another example is Castilla y León (with a difference of 16.29 days); however, in this case the total number of accidents for Spanish workers was greater compared to foreigners. The different behaviours observed depending on the region should be subjected to more detailed analysis.

This study analyses accidents in the agriculture sector in Spain in the period 2013-2018, but its conclusions may be different in other countries of the European Union or the rest of the world. The MTMSS Workplace Accident Registry system collects the total number of reported accidents, but it may be that some have not been reported, and therefore have not been considered in this study.

Only accidents that have led to at least one day’s absence from work have been considered, which means that accidents without a personal injury are not analysed in this paper. Likewise, we analysed the severity of accidents once they have occurred, but not the probability of their occurrence. The nature of the data used allowed us to study and compare groups of accidents in agriculture sector that have already occurred, hence, to show the probability that once the accident occurs in that specific category, it may be minor, serious, very serious or fatal.