Variables

The elaboration of the ranking of the 100 most attractive companies for professional development in Spain is carried out using six independent and one dependent variables (Table 1), whose potential weighting varies according the variable and can reach a maximum of 1,000 points. The study focuses on two of the independent variables: Training and Talent Management, due to them being directly linked to knowledge. Additionally, these two variables are the most relevant from a quantitative point of view as they reach a potential value of 46% of the total value of the dependent variable.

For ^{Gallie (2009}, pp. 6), the concept of work quality lies in the level of competencies, degree of autonomy, discretion in the tasks, opportunities for skill development, safety at work, and degree of compatibility with the work-family balance.

^{Grimshaw and Rubery (2007}); ^{Fields, Chan, and Akhtar (2002}); and ^{Jackson and Schuler (1995}) emphasize the role of the contextual factors (markets, political, institutional, social environment, culture) to explain the differences in human resource management practices and the results between companies. Recent literature underlines the relationship between the universal standardization of practices and the local context adjustment (^{Amossé et al., 2016}; ^{Gallie, 2007}; ^{Hodulak, 2017}; ^{Ibrahim and Shah, 2013}; ^{Quintanilla and Ferner, 2003}).

The scant theory construction effort on the rankings of the best companies to work for (^{Joo and McLean, 2006}) extends to the lack of a specific theory on the association between the scores obtained by the companies in those rankings and the values of other organizational, economic, and cultural factors, among others. Therefore, the explicative factors have been selected based on empirical research dedicated to human resource management practices, which are at times solely focused on the detection of the effects of one or more of the factors of interest.

Economic activity.^{Jackson and Schuler (1995}, pp. 251ff.) include, among the contextual factors influential on human resource practices, the characteristics of the activity sector. According to Conway et al. (2018), these can be categorized in different ways: services / industry. In this study, the companies are classified according to the activity sector for which the National Classification of Economic Activities (CNAE for its acronym in Spanish) is used.

Nationality. Despite the fact that Ibrahim and Shah (2003) found no effects of the country of origin on the human resource practices in Malayan companies, ^{Ferner (1997}) presented a relationship of the systematic differences in the human resource management of multinational companies based on the country of origin. ^{Liu (2004}) and ^{Guthrie et al. (2008}) document findings in agreeance with the most recent empirical researches. In this research each company is identified with the country of origin, which allows to obtain statistics of the variables under study grouped by country. In turn, they have been grouped into geo-cultural areas such as Anglo-Saxon countries, Central and Northern Europe, Mediterranean, and others; this allows a more global vision of this dimension.

Geographical location. In some jobs there are certain differences between human resource practices in different countries, detected by comparative researches that use different econometric techniques (^{Amossé et al., 2016}; ^{Conway et al., 2008}; ^{Fields, Chan, and Akhtar, 2002}; ^{Grimshaw and Rubery, 2007}; ^{Paawee and Boselie, 2007}). One of the objectives of this research is to find out whether the specific location in a geographical area influences on the Training and Talent Management variables. For this reason, all the companies were classified according to regional situation (municipality, province, and autonomous communities) of the headquarters in Spain. The analyses are done using the regional dimension of the autonomous communities, grouped also into cultural areas such as: Center, Mediterranean, North of Spain, and others. This will allow to carry out studies at different regional scale levels.

Company size. Size is possibly one of the most influential factors on human resource practices (^{Fields, Chan, and Akhtar, 2002}, pp. 265ff.), though there is no consensus on the sign-positive or negative-of its effects. On the one hand, ^{Kortekaas (2007}) finds a notable positive impact of small/medium size on employee behavior (including commitment or job satisfaction) and a negative impact of the same size on an operational performance indicator (absenteeism and sick leaves). ^{Kok and Uhlaner (2001}) find an association between the increase in size of companies and the formalization of human resource practices. ^{Morgan (2014}) suggests that small companies have several advantages such as role flexibility, close employer-employee relationship, functions, and work styles for workers, among others. On the other hand, ^{Ibrahim and Shah (2013}, pp. 7, 14ff.) argue that small companies lack the necessary resources to adapt progressive human resource management practices. In this work, the measure used is the number of workers of said organization in Spain.

Stock market listing.^{Conway et al. (2008}, pp. 638) conclude that the research on its effects did not generate clear conclusions. The empirical results found in the United Kingdom show that the listing is positively associated with teamwork and the remuneration related with performance. In the case of France, the listing is associated with the autonomy and training of the worker. This variable will be used to try and analyze whether the stock market listing or market value of the companies placing in the ranking of AE is relevant.

Sample and Analysis Techniques

The sample size corresponds to the one hundred most attractive companies of the ranking published each year by AE for the 2013-2016 period, which corresponds to 400 observations and 182 companies-as some of them repeat in all or some years-resulting in a ratio of 2.2 observations per company. Of the 182 companies, 40 appear in all four years, which represents only 21.9%, 33 appear in three years (18.13%), 32 appear in two years (17.5%), and 77 appear in only one year (42.3%), thus a continuous entry and exit of companies can be observed in the ranking of AE.

The elaboration process of the ranking consists in delivering a survey with a hundred questions on behalf of AE to the Human Resources departments of the companies that have more than five years in Spain and that have had more than 100 employees in the studied period. Once the companies respond to said surveys, expert consultants and independent Human Resources professionals of AE proceeded with the measurement and validity of the ranking analysis, to then publish the evaluation of the variables for each company: training, talent management, corporate social responsibility (CSR), remuneration and compensation, work environment, employee perception of the company, and total rating.

Thus, to reach the objective, new variables have been added to the ranking used such as nationality, location, activity of the company, size, and whether it is listed in the stock market or not. The objective is to analyze whether these added variables influence the value obtained by the companies for training and talent management and to better know their profile.

The analysis techniques used will be: multiple linear regressions through ordinary least squares for each of the four years, Pearson correlations, unifactorial variance with Levene’s test, and statistics that allow measuring the Training and Talent Management of companies in number, percentages, and average values.

Training and Talent Management according to the Economic Activity of the Companies

Table 2 shows the value of Training in the companies, classified by economic sector. First place is taken by sanitary activities and social services, which achieve an average value of 183.3 points. In second place are professional, scientific, and technical activities with 182.2 points, this being where international consultancy companies such as Deloitte, PWC, or KPMG place. It should be noted that a sector as related with knowledge as education places ninth, with little representativeness in the number of companies.

Focusing on the Talent Management variable, the differences regarding the above are notable. First of all, the Pearson correlation with Training based on the average values obtained by sector is of only 12.2%. It is here that it can be observed how the sector that occupies first place is that of professional, scientific, and technical activities. The education sector stands out once more, placing in last place with 127.5 points.

Training and Talent Management according to the Nationality of the Companies

Table 3 shows the values of the Training variable, grouped by countries, where the companies from South Korea take first place in the ranking with a value of 188.8 points obtained with the electronics company LG. Second place is taken by Irish companies, which obtain a score of 187.5, with the consultancy Accenture standing out. Both South Korea and Ireland have an insufficient sample size. In third place are the Swiss companies such as Zurich insurance or the pharmaceutical Roche Farma, among others. Therefore, focusing on countries with a significant presence of companies, Switzerland takes third place with 183.4 points presented by 14 observations; France is in fifth place with a mean value of 178.4 points with 34 references; followed by the USA in sixth place with 177.2 points and 65 companies; and Spain, the country with the most companies (155), takes seventh place with 176.4 points. The United Kingdom is in third place based on the number of companies; however, it is in tenth place concerning mean value in Training, with Germany in twelfth place.

Regarding the Talent variable, the companies from Finland and South Korea standout, although both countries are lacking in representativeness in the number of companies. In third place, with a score of 183.2, very similar to that obtained in the Training analysis, are Swiss companies. The companies from the United Kingdom, with a low average value for Training, obtain fourth place in the case of Talent Management. In the case of Spain, the fact that it goes from seventh place in Training to twelfth place in Training Management stands out, which reveals that both factors do not always go in the same direction or that sometimes the treatment of human capital is not adequate. It is also worth mentioning that Portuguese companies have some of the best values with regard to Training (fourth place) and yet, when analyzing Talent and the total values they are considered some of the worst, taking sixteenth and thirteenth places, respectively. This explains why the correlation between Training and Talent grouped by country is of 43.4%.

In the analysis by international areas, shown in Table 4, the companies from Mediterranean Europe take first place in Training and last place in Talent Management, which once more shows the low correlation between both variables. Another aspect that stands out is the low values of the companies from North-Central Europe, with countries such as Denmark, Luxemburg, Finland, or Sweden. The companies from Anglo-Saxon countries are those that maintain more constant spots.

Training and Talent Management according to the Geographical Location of the Companies

Table 5 shows that, in the case of the Training variable, Cantabria is the Autonomous Community that occupies first place in the ranking with 184.2 points, this being where companies such as the Santander Bank, one of the largest banks in Europe, are located. Second place corresponds to organizations located in the Balearic Islands, where companies related to tourism stand out, as it is one of the areas with the highest touristic index in the world, achieving a total of 181.7 points. Almost two points below, in third place is the Basque Country, which has in its territory the financial institution BBVA (another large European bank) and large energetic companies. In the case of the communities with the greatest presence there is Catalonia with 51 companies, taking fifth place; with Madrid being the area with the greatest number of companies, having 76.8% of the total.

In the case of the Talent variable, Galicia is the community with the highest average value, placing first with 188.3 points, with the textile company Inditext standing out as one of the companies with the highest value in the world; it is followed by the Canary Islands and the Balearic Islands. As can be observed, the relative positions of the communities change when classifying them according to Training or Talent Management, which explains why the Pearson correlation between both variables is of 23.8%, much lower than when classified by countries, although higher than it is when sorted by activities.

Table 6 shows the average values for Training and Talent Management by regional areas in Spain, where there are barely differences between areas, without considering the “Remaining” grouping , which occupies the last place and repents the least number of cases. As for the number of companies, it can be observed that companies located in the Central area of Spain predominate, more specifically in Madrid, which is the capital of the country.

Training and Talent Management according to the Size of the Company

It can be observed in Table 7 that 87% of companies have mean workforces with more than 250 workers. The Training variable shows some results where the mean value is greater in the larger companies (176.06) in comparison to the others (169.47), and in this case it is statistically relevant.

On the other hand, the behavior of the Talent Management variable is different, as smaller companies have a higher mean value (180.88) than large ones (176.14), without said differences being significant. Therefore, it can be concluded that company size is not relevant in Talent Management.

Training and Talent Management according to the Stock Market Listing of the Company

It can be observed in Table 8 that 76.25% of the most attractive companies for professional development are listed in the stock market. The data referring to Training do show that companies listed in the stock market have higher values (177.00) than the others, given that their bilateral significance is < 0.05; therefore, it can be considered that these means are relevant. Regarding Talent Management, the bilateral significance is of 0.085, values that are very close to statistical relevance in favor of companies listed in the stock market with a value of 178.03 against an average value of 174.09 for the others.

Global Analysis of Training and Talent Management

At this point, it will be estimated the degree to which the differences in the values of the independent variables modify the value of the Training of the companies, so that the rest of the variables remain constant. For this, a multiple linear regression model is estimated through ordinary least squares for each of the four years (year = model). The dependent variable in each of the models is the value of the training level. The independent variables are:

In all the models, the number of observations is identical: the 100 companies in the ranking by AE for the year in question. Annex I shows the results of the four regression models. The model for the year 2016 is statistically significant (F (24, 75) = 1.7; p(F) < .05); for all the others, p(F) > .1).

Stock market listing achieves statistical significance (at a level of .01) in at least two models (2014 and 2016). In both cases, the companies listed in the stock market have a higher value and the intercept increases by 10.17 and 10.61 points, respectively. Other variables are not statistically significant except in one of the models:

Apart from the statistical significance of the variables, there are certain patterns or trends detected that should be unraveled in a more systematic research. For example, in the case of the sign of the coefficients, in three of the models the belonging of the companies to both the Mediterranean geo-cultural area (2014, 2015, and 2016) and the area of the rest of the countries (2013, 2014, and 2015) seems to increase its score when compared to companies of the Anglo-Saxon area, while the belonging to the Central and Northern Europe area seems to contribute to reducing the value of the companies (2013, 2015, and 2016).

Considering the sign, large companies have greater scores in the four models, and stock market listing shows a growing positive relationship.

Focusing on Talent, the measure in which the value of said variable modifies the differences in the values of the independent factors if all other variables are kept constant would be estimated. For this, a multiple linear regression model is estimated through minimum least squares for each of the four years; the value for talent management is the dependent variable in this case. Annex II shows the results of the four regression models. The model for the year 2015 is statistically significant (p(F) < .05); for all others, p(F) > .1).

Two independent variables achieve statistical significance (at a level of .05) in at least two models. The first is company size, although the sign of its effect in one of the models is opposite to the other; thus, in 2013, large companies are more valued than small companies (the value of the intercept increases by 16.17 points), while in 2016 they are less valued (the value of the intercept is reduced by 17.11 points). The second of these variables is stock market listing. In both models (years 2013 and 2015), companies listed in the stock market have a higher value than those not listed, increasing the intercept by 11.35 and 12.21 points, respectively.

Despite the fact that in the 2013 model the category of Mediterranean Countries has statistical significance, no excessive value can be granted to the result as the geo-cultural area of the country of origin of the company has no statistical significance.

The sector of economic activity has statistical significance in the 2013 and 2015 models. Certain patterns or trends should be noted in Annex II. For example, if the sign of the coefficients is considered, in three of the models the belonging of the companies to both the Central and Northern Europe geo-cultural area (2014, 2015, and 2016) and the Mediterranean area (2013, 2014, and 2015) seems to reduce their score in comparison to companies from the Anglo-Saxon area, while their presence in other countries appears to contribute to increasing the attractiveness of the companies.

The Spanish nationality of the company (statistically significant at a level of 0.5 in the 2013 model) has a positive sign in all of the models, except for the 2014 model. The stock market listing variable has a positive sign in the 2013, 2015, and 2016 models; stock market listing increases the attractiveness of the companies.

Limitations

This article reveals a series of limitations related to the established objective. First of all, it is a quantitative work limited to two independent variables (Training and Talent Management), which assesses the attractiveness of a company for professional performance. Another limitation is due to focusing on adding other objective variables, such as sector, nationality, size, stock market listing, or regional location, while avoiding factors such as leadership, management capacity, variables of a qualitative character and whose assessment is more complex. Another limitation is the fact that the data are for a concrete period between the years of 2013 and 2016. This is additional to the national limitation, as the focus is on companies that operate in a European country.