Youths and their probability to enter middle-class jobs during the COVID-19 pandemic in Indonesia

ABSTRACT

While the middle class has historically served as an engine for economic growth, Indonesia’s performance in encouraging the growth of middle-class jobs is still relatively low. This is also the case for youth, which is considered to have a crucial role in economic development. This study examines the possibility for young people to enter middle-class jobs in Indonesia, as well as the socioeconomic factors which may be driving the process. Using the 2015 and 2020 National Labor Force Surveys (Sakernas), our findings show that youth groups have a lower chance of becoming middle-class workers, both before and during the COVID-19 pandemic. However, interestingly, the study finds that the chance of some youth subgroups – in particular young males – of becoming middle-class workers during the pandemic is greater than during previous periods. Other characteristics that can increase youth opportunities of becoming middle-class workers are being male, being married, having higher education, and living in an urban area. In addition, there are four dominant sectors in supporting middle-class jobs for youth, namely mining, financial services, industry, and transportation.

KEYWORDS:

• Youths
• middle-class job
• pandemic
• Sakernas

1. Introduction

Indonesia is one of the most promising economies in Asia, with an average annual economic growth of approximately 5% in the last decade (Badan Pusat Statistik 2020). In 2019, with per capita income of USD 4050 it has become a country with upper middle-income status (Pratomo, Syafitri, and Anindya 2020). Besides, it has managed to escape the lower-middle income trap in which it has been mired since 1985 by enhancing the quality of its human resources through improving education and reducing poverty. Along with a significant reduction in its poverty rates, the population of its middle and upper-middle classes have grown significantly. Based on the National SocioEconomic Survey (SUSENAS), middle-class households grew from only 5% of the population in 1993 to 20% in 2019 (World Bank 2019). The middle class has also served as an engine of growth, supporting almost half of Indonesia’s total national consumption.

The middle class is a crucial part of a country's economic growth. Banerjee and Duflo (2008) explained three reasons for the importance of the middle class. First, the middle class is the foundation of the growth of entrepreneurs who drive innovation and growth. Second, the increase in the size of the middle class encourages the accumulation of human capital and savings. Third, the middle class is the driving force behind domestic consumption, so it has the potential to encourage product diversification and increase investment and market expansion, allowing for increased economies of scale production.

The definition of middle class itself is quite diverse, depending on how welfare is defined (Pratomo 2020). Several studies in the large economics literature determine the middle class based on their income or level of consumption (Ravallion 2010; Banerjee and Duflo 2008 Kharas and Gertz 2010; and Asian Development Bank 2010a). Afif (2014) for example, characterizing the middle class as a population with daily per capita spending between USD 2–20 per day. The World Bank (2019), in their report ‘Aspiring Indonesia–Expanding the Middle-class,’ defines the middle class as those who are less than 10% likely to fall into poverty or become financially vulnerable in the following year (especially those whose household consumption per capita is approximately 3.5–17 times the poverty line).

The concept of middle class has also been defined in terms of occupational status, self-identification, or even democracy in sociology or other social sciences (López-Calva and Ortiz-Juarez 2014). Loyaza, Rigolini, and Llorente (2012) find that a growing middle class increases democratic participation, reduces corruption, and increases health and education awareness. Birdsall, Graham, and Pettinato (2010) argue that middle class constitutes the backbone of democracy ensuring social and political stability between the poor and the rich, and further suggests that economic growth is more likely to be sustained if it is driven by and in the interests of the middle class.

As reported by the African Development Bank (2011), middle-class growth is an important medium- and long-term development indicator in Africa, as it is strongly associated with faster rates of poverty reduction. In addition, according to the World Bank (2019), the growth of middle-class jobs can play a role in achieving inclusive growth, including aspects of poverty alleviation to reduce societal inequality. The World Bank (2021) also suggests that middle-class provides economic security, unlimited work contracts, and various forms of social benefits, such as health insurance, work accident insurance, death insurance, old-age insurance, pension insurance, and/or paid annual leave, sick leave, and maternity leave.

Moreover, Sen (1983) mentions that middle class will be more relative in terms of income, while in terms of functioning, middle class is absolute. Regardless of which concept is used, the measurement of middle class is dependent on a particular period and place and it is determined by several factors, such as history, culture and the development stage of a society. While the concept of the middle class is diverse, there have been considerable obstacles in applying it to developing countries, including Indonesia. The focus of this paper, however, is limited to the concept of middle class at income level, not incorporating the possible long listed concepts of economic, political and also social benefits of the middle class.

Regarding employment, Indonesia is one of the developing countries where most of the jobs created are in the low-wage sector, which is relatively far from the minimum threshold for middle-class. In 2018, out of 85 million workers only 13 million belong to the middle-class cohort (World Bank 2021). That means only 15% of workers having an income that allows them to be categorized as middle class. Nonetheless, most individuals are still vulnerable to being poor and do not have the economic security to move up to the middle class. In September 2018, 126.6 million individuals, or 47.6% of the total population, belonged to the aspiring middle class (World Bank 2019). If including the vulnerable group, this number means that more than half of Indonesia's population is not actually poor but still lacks the opportunities to move up the middle-class ladder. In addition, 77.4 million people – that is equivalent to 29.1% of the population – are still poor or vulnerable to falling back into poverty.

Furthermore, access to middle-class jobs is generally limited for youth. Among young workers aged 15-24, only 15% have middle-class jobs. The proportion of youth who have middle class jobs is only 6.4% when compared to the overall middle-class workers in Indonesia (World Bank 2021). This condition illustrates that the Indonesian workforce, especially young workers, is still not set up to be part of the middle class.

In 2020, youth participation in middle-class jobs in Indonesia has been increasingly threatened due to the coronavirus (COVID-19) pandemic. The imposition of strict restrictions on community activities results in a shock to business activities for education, which ultimately means that youth – especially those who complete their education during the crisis period, or the ‘lockdown’ generation – will feel the burden of this crisis for a long time. A study conducted by Kahn (2010) emphasizes that entering the labor market during a crisis or economic downturn has long-lasting negative consequences for new workers. One of the consequences for the young individuals entering the labor market during an economic crisis is getting wages that tend to be lower than during normal times for decades (Naidoo, Truman, and Ilmiawan 2015). Such situation is most likely to be even worse during the period of pandemic, which may lead to intergenerational inequalities.

Availability of decent job for youth remains an important issue. As future leaders, their participation in labor force market is required to encourage their country’s sustainability. Should they fail to find decent jobs, it may result in lower per capita income in the long term. Hence, individuals will face a lower standard of living and can hardly accumulate the capital necessary for a country's growth and development accordingly. In addition, the lack of productive job prospects for young people is believed to have a major impact on critical global issues such as poverty and migration and is believed to increase the risk of social discontent (Escudero and Mourelo 2014).

To date, discussions on how to encourage youth to enter middle-class jobs have received relatively little attention. A large body of studies has attempted to identify the factors that most determine youth's probability of finding a job. Okicic et al. (2020) have revealed that increased opportunities for young individuals entering the labor market can be influenced by age, gender, and household characteristics. Moreover, youth participation in the work force market is also influenced by four main factors: (1) individual and social demographic characteristics; (2) family characteristics; (3) regional economic characteristics; and (4) government policies in the labor market (Ahmad and Azim 2010; Bernal-Verdugo, Furceri, and Guillaume 2012; Bruno et al. 2017; Bremer 2018; Hedvicakova 2018; Okicic et al. 2020; Pastore 2019; Kang 2021; and Scandurra, Cefalo, and Kazepov 2021). Research conducted by Ahmad and Azim (2010) and Scandurra, Cefalo, and Kazepov (2021) also explained that education is an important foundation that enables youth to get better opportunities in the labor market. Government policies in the labor market – such as those concerning minimum wage, unemployment benefits, work-related taxes – might also influence the probability of youth getting a job (Bernal-Verdugo, Furceri, and Guillaume 2012).

In the case of Indonesia, while youth are the most vulnerable group given the very low proportion of middle-class youth workers, yet no studies – i.e. those that specifically discuss the empirical and theoretical aspects of their probability of getting middle-class jobs – are available. In addition, the influence of individual and sociodemographic characteristics on the employment probabilities of middle-class young people in Indonesia has not been studied either. As a result, it is not possible to know precisely which of these factors is dominant.

Based on this background, there are two main objectives of this research. First, the study examines the relationship between youth and their probability of entering a middle-class job in Indonesia by controlling all other socioeconomic characteristics. Second, focusing on youth, the study identifies what factors can affect the probability of youth getting middle-class jobs. The definitions of youth in this study are derived from research conducted by Pratomo (2016), where young individuals are individuals aged between 18 and 24 years with the assumption that adolescents aged 15–17 have the potential to be paid less than what is regulated and are usually categorized as child workers. The determinants of middle-class youth employment will focus on the economic and sociodemographic characteristics of individuals. Furthermore, this study will also compare the conditions before and during the pandemic.

Following the introduction section, this paper will discuss the main literature used in this study. The next section presents a discussion of the research methods used, followed by a discussion of empirical results from research findings with data used before and during the pandemic. Finally, this paper ends with conclusions and some policy implications.

2. Literature review

Some studies discuss the probability of youth getting a formal job, which may be quite similar to the concept of middle-class work. Shehu and Nilsson (2014), for example, looked at the empirical evidence to confirm standardized informal work among young workers in 2012–2013 in 20 developing countries. They analyzed the employment characteristics of informally employed youth according to age, gender, area of residence, education level, and health status using survey data. The results of this study indicate that the majority of young workers aged 15–29 are still engaged in informal work, which results in lower wages, lower job satisfaction, and a high proportion of underemployed workers. In addition, other findings from the study explain several characteristics of youth who are relatively vulnerable to entering the informal sector namely, female youth, youth living in villages and single, youth experiencing health problems, and youth having low education.

Similarly, Angel-Urdinola and Tanabe (2012) examined the main micro-determinants of informal employment consisting of age, gender, education level, employment sector, marital status, employment status, and geographic areas within a selected group of Middle Eastern and North African regions. The results show that the level of informality is highest among young people between 15 and 24 years old, which accounts for 24–35% of the total employment, but after the age of 25 years, the level of informality will decrease. In particular for women, education level and marital status are still dominant factors in determining worker informality. The most important finding is that workers in the tertiary sector – such as financial services, transportation, tourism, and communication – have a greater chance of leaving the informal sector than those working in the secondary sector including jobs in industry and construction.

Doğrul (2012) examined the determinants of formal and informal employment in urban areas of Turkey using variables of gender, marital status, head of household, and education by applying the multinomial logit model. Moreover, Başlevent and Acar (2015) examined trends in informal employment using microdata from the 2000, 2006, and 2012 Turkey’s Household Labor Force Survey (HLFS). They studied occupations and age groups based on gender differences, following an econometric approach to the main determinants of informality status. Both studies provide the same view regarding the drivers of informality among workers, namely, that there is still a gender gap in efforts to reduce informality among workers. Women are more likely to be employed in subsectors where informal work is more prevalent.

Apart from that, the discussion on the determinants of middle-class employment still received relatively little attention. Previous studies on the dynamics of the middle class can generally be divided into four main dimensions: (1) discussion of the definition and measurement of the middle class (Ravallion 2010; Banerjee and Duflo 2008; Kharas and Gertz 2010; Asian Development Bank 2010a; Ncube and Shimeles 2013; Tschirley et al. 2015); (2) the relationship between the middle class and economic growth (Handley 2015; Dabla-Norris et al. 2015; Tschirley et al. 2015); (3) its relationship to government (Cheeseman 2015; Mattes 2015; Resnick 2015); and (4) the determinants of middle-class society in general (Albert, Santos, and Vizmanos 2018; Kodila-Tedika, Asongu, and Kayembe 2016; Asian Development Bank 2010b).

Albert, Santos, and Vizmanos (2018) examined the determinants of households joining the middle class in the Philippines using a multinomial logistic model. The results of this study indicate that unmarried male workers living in urban areas are more likely to enter the middle-class groups, depending on certain household characteristics, family size, and age. Kodila-Tedika, Asongu, and Kayembe (2016) examined the determinants and consequences of the middle class on a continent with relatively high economic growth in 33 African countries in 2010. They found that GDP per capita, education, size of the informal sector, openness and democracy had a positive effect on middle-class growth. However, on the other hand, there are negative relationships with ethnic fragmentation, political stability, and economic vulnerability.

Ncube and Shimeles (2013) estimate the determinants of the middle class in Africa by utilizing microdata consisting of more than seven hundred thousand household histories in 1990–2011. The results show that education, quality of institutions, and ethnic fractionalization have played an important role in determining the growth of the middle class in recent years. In addition, they show that the size of the middle class is higher in countries where mutual trust among citizens is stronger. Research conducted by Castellani, Parent, and Zentero (2014) in Latin America adds to the literature on the determinants of the middle class. That study showed that gender, age, and education were associated with the likelihood of being middle class. In this case, female-headed households tend to be less affluent and more likely to be in lower-income classes, whereas referring to age, young individuals tend to be poorer while the older workers are better off – the older they are, the more likely they become rich. Meanwhile, regarding education, the higher their education level is, the higher their possibility of entering the middle class.

3. Research method

To determine the probability of youth getting a middle-class job, a logit model will be estimated using the 2015 and 2020 Indonesian Labor Force Survey (Sakernas) data. Sakernas data comprise several indicators, such as sociodemographics and labor market characteristics at individual level. The years 2015 and 2020 were chosen to compare the conditions before (normal condition) and during the pandemic. This survey is conducted annually by the Central Statistics Agency (BPS) and is often used as the main medium through which the dynamics of the workforce in Indonesia could be identified.

In general, the equations used in this study can be written as follows: (1) $\ln \left[{\displaystyle \frac{\pi (x)}{1-\pi (x)}}\right]={\alpha }_0+{\alpha }_{1}Yout{h}_i+\psi X_i+{\epsilon }_i$(1) where $\pi (x)$ is probability for youth to become middle-class workers. This study uses a definition of middle-class employment that can be measured using data from the Sakernas based on income levels. An individual’s employment status is categorized as a middle-class job (Y = 1) if their current income is equal to or above the minimum threshold and not a middle-class job (Y = 0) if they are below it. The middle-class job definition used in this study refers to research conducted by the World Bank (2021), where the minimum income threshold for class work is obtained through the following formula: (2) $\mathrm{Middle}\mathrm{Class}\mathrm{Income}\mathrm{Minimum}\mathrm{Limit}=(3.5\ast \mathrm{Poverty}\mathrm{Line}\ast 4)/1.5$(2)

By referring to Equation (2), we will find the minimum income limit for middle-class workers for each province in Indonesia as follows (Table 1):

Table 1. Minimum Limit for Middle-Class Job Income.

Province	Minimum Income Limit (Indonesian Rupiah)
	2015		2020
	Urban	Rural	Urban	Rural
Aceh	3,923,024	3,681,244	5,074,393	4,803,521
Sumatera Utara	3,545,715	3,291,279	4,858,271	4,541,992
Sumatera Barat	3,951,164	3,650,995	5,278,289	4,943,615
Riau	3,899,168	3,889,947	5,346,656	4,934,860
Jambi	3,955,980	3,079,020	5,156,648	4,244,371
Sumatera Selatan	3,534,897	2,986,611	4,394,133	3,960,777
Bengkulu	3,972,659	3,772,337	5,384,596	4,755,212
Lampung	3,609,461	3,230,155	4,707,080	4,079,665
Kep. Bangka Belitung	4,823,793	5,065,499	6,686,960	6,877,267
Kep. Riau	4,531,296	4,264,708	5,781,636	5,580,391
DKI Jakarta	4,695,021	–	6,377,831	–
Jawa Barat	2,970,772	2,979,461	3,889,191	3,839,192
Jawa Tengah	2,876,188	2,896,087	3,774,876	3,660,683
DI Yogyakarta	3,355,053	3,027,603	4,558,969	3,770,993
Jawa Timur	2,933,653	2,972,135	4,000,201	3,791,256
Banten	3,412,939	3,141,525	4,966,229	4,428,545
Bali	3,187,837	2,932,701	4,215,251	3,801,616
Nusa Tenggara Barat	3,129,317	2,925,683	3,919,291	3,700,667
Nusa Tenggara Timur	3,493,980	2,710,055	4,631,116	3,520,963
Kalimantan Barat	3,243,483	3,148,021	4,708,153	4,279,557
Kalimantan Tengah	3,166,231	3,499,421	4,469,481	4,656,951
Kalimantan Selatan	3,470,068	3,294,405	4,749,285	4,626,673
Kalimantan Timur	4,709,143	4,448,397	6,303,724	6,123,311
Kalimantan Utara	4,715,779	4,458,020	6,752,461	6064436
Sulawesi Utara	2,822,195	2,903,301	3,756,685	3,666,301
Sulawesi Tengah	3,513,963	3,295,413	4,599,075	4,467,745
Sulawesi Selatan	2,558,640	2,375,557	3,504,321	3,274,049
Sulawesi Tenggara	2,634,147	2,467,463	3,582,497	3,361,148
Gorontalo	2,562,756	2,568,188	3,506,907	3,489,836
Sulawesi Barat	2,511,413	2,609,544	3,331,692	3,287,844
Maluku	3,779,337	3,784,685	5,451,236	5,287,305
Maluku Utara	3,533,021	3,325,700	4,617,228	4,289,413
Papua Barat	4,467,857	4,267,405	5,931,697	5,621,373
Papua	4,153,865	3,662,829	5,808,563	5,249,179

Source: Sakernas.

The main objective of Equation (1) is to determine the factors that affect the probability of youth getting middle-class jobs before and after the pandemic. Youth are predicted to have a lower probability of getting into middle-class jobs during the pandemic due to the worsening economic conditions. The coefficient vector (ψ) on the variable X acts as a control variable that includes sociodemographic characteristics, including education, participation in pre-employment training, gender, education level, place of residence, employment sector and internet use (digitalization).

The probability of youth becoming middle-class workers also depends on the labor market, community, and households’ characteristics. However, not all of these aspects are available in the Sakernas, since Sakernas are focused on individual level. Meanwhile, ${\epsilon }_i$ is the error term, explaining variables outside the model that also have the potential to affect the likelihood of youth becoming middle-class workers, whereas much of them – such as skills, personalities, or cultural aspects – is unpredictable.

Equation (1) is then re-estimated into the following equation using only a sample of young workers (15-24 years): (3) $\ln \left[{\displaystyle \frac{\pi (x)}{1-\pi (x)}}\right]={\alpha }_0+\psi X_i+{\epsilon }_i$(3) Equation (3) is used to answer the second objective, namely, to determine what factors have the potential to increase youth's chances of getting middle-class jobs. The main variables used in this equation follow what is in the first equation. The summary statistics for the main variables across middle-class workers (and lower-class workers as comparison) can be seen in Table 2.

Table 2. Summary statistics.

Variables	Middle-Class Workers	Lower-Class Workers
	Mean	Mean
Youth	0.04	0.22
Male Workers	0.70	0.49
Training	0.48	0.11
Urban	0.63	0.43
University	0.53	0.28
Senior High School	0.21	0.20
Vocational High School	0.11	0.09
Digitalization	0.29	0.03
Sector of Employment
Agriculture, forestry, livestock, and fishing	0.07	0.25
Manufacturing	0.10	0.07
Mining and quarrying	0.04	0.09
Wholesale and retail trade, restaurants, and hotels	0.13	0.35
Transportation, storage, and communications	0.05	0.16
Finance, insurance, real estate, and business Services	0.06	0.10
Community, social and personal services	0.49	0.28

Source: Sakernas.

4. Results

Table 3 reports the results of the logit regression regarding the probability of youth entering middle-class jobs, separated for all employees (including self-employment) and for paid employment only. The results show that youth are less likely to enter middle-class jobs both before and during the pandemic and apply for any level of employment. International Labor Organization (2012) noted a similar result in Brazil, the Philippines, and South Africa, where young workers are disproportionately represented in low-paying jobs, i.e. jobs that pay less than two-thirds of the average wage. This phenomenon is actually seen as normal, considering that early-career work is a stepping stone to middle-class jobs, which are considered to be more stable and pay higher wages. On the other hand, this condition shows the mismatch of supply and demand for labor, where limited work experience and a lack of professional skills often cause young people to face serious difficulties when trying to find work, especially to get quality work in the case of middle-class jobs.

Table 3. Logit Results of Middle-Class Worker Status (All Employees).

Variables	2015								2020
	All Employees		Paid Employment		All Employees		Paid Employment		All Employees (With Digitalization)		Paid Employment (With Digitalization)
	Coef	Odd Ratio	Coef	Odd Ratio	Coef	Odd Ratio	Coef	Odd Ratio	Coef	Odd Ratio	Coef	Odd Ratio
Youth	−1.148***	0.3174	−1.211***	0.2980	−0.788***	0.4549	−0.758***	0.4687	−0.780***	0.4583	−0.746***	0.4743
Male Works	0.574***	1.7758	0.510***	1.6659	0.708***	2.0292	0.655***	1.9248	0.731***	2.0779	0.674***	1.9612
Training	0.736***	2.0875	0.744***	2.1039	0.560***	1.7499	0.590***	1.8042	0.540***	1.7162	0.573***	1.7729
Urban	0.235***	1.2646	0.277***	1.3191	0.297***	1.3456	0.362***	1.4363	0.283***	1.3266	0.344***	1.4103
University	2.409***	11.1261	2.772***	15.9946	2.188***	8.9217	2.627***	13.8273	2.073***	7.9501	2.516***	12.3749
Senior High School	1.115***	3.0486	1.533***	4.6336	1.009***	2.7442	1.475***	4.3703	0.995***	2.7045	1.456***	4.2883
Vocational High School	1.165***	3.2066	1.616***	5.0349	1.001***	2.7211	1.465***	4.3291	0.984***	2.6744	1.443***	4.2334
Married Workers	0.773***	2.1668	0.821***	2.2726	0.657***	1.9298	0.755***	2.1284	0.650***	1.9148	0.749***	2.1149
Digitalization	–		–		–		–		0.507***	1.6596	0.434***	1.5440
Sector of Employment
Agriculture,	−0.437***	0.6462	−0.379***	0.6842	−0.275***	0.7597	−0.474***	0.6224	−0.263***	0.7686	−0.467***	0.6268
Manufacturing	0.237***	1.2680	0.038	1.0387	0.553***	1.7385	0.392***	1.4800	0.552***	1.7365	0.399***	1.4907
Mining and quarrying	1.339***	3.8161	1.428***	4.1707	1.372***	3.9427	1.337***	3.8069	1.371***	3.9412	1.340***	3.8178
Wholesale/retail trade, restaurants, and hotels	0.336***	1.3989	−0.472***	0.6235	0.152***	1.1642	−0.656***	0.5188	0.162***	1.1755	−0.644***	0.5254
Transportation and communications	0.269***	1.3082	0.246***	1.2783	0.065*	1.0674	0.110**	1.1167	0.062	1.0637	0.099**	1.1037
Finance	0.655***	1.9256	0.252***	1.2865	0.684***	1.9817	0.207***	1.2298	0.658***	1.9313	0.192***	1.2119
Community, social and personal services	0.640***	1.8974	0.325***	1.3841	0.388***	1.4740	−0.066*	0.9358	0.303***	1.3541	−0.134***	0.8749
Constant	−4.745***	0.0087	−4.801***	0.0082	−4.876***	0.0076	−4.925***	0.0073	−4.886***	0.0076	−4.927***	0.0072
Observations	202,158		111,233		314,364		161,997		314,364		161,997

Note: *** p < 0.01, ** p < 0.05, * p < 0.1

According to O’Reilly, Grotti, and Russell (2019), youth aged 15–24 in Europe have historically been more likely to enter low-paying and low-skilled jobs, which allow less competition from their older counterparts. This is why the employment distribution of youth tends to skew in certain sectors that have relatively low barriers to entry, low wages or in sectors offering part-time or temporary contract work. Furthermore, Graham, Williams, and Chisoro (2019) have mentioned three structural problems faced by youth in the labor market. First, the mismatch between education output and labor market skills that leads to high unemployment among university graduates. On the other hand, labor market participation among high-skilled youth is very low, whereas young graduates are likely to face a labor market transition of three years on average. Second, the inability of the private and public sectors to absorb new labor markets. Third, lack of access to quality national programs that allow a smooth transition from school to work life in the form of vocational training or career guidance. Hence, all those factors have limited youth opportunities to enter directly into middle-class jobs.

However, there is an interesting finding that youth’s chances of getting middle-class jobs are even higher during the pandemic compared to normal periods, although their chances are still lower than those of older workers. In 2015, the coefficient for the variable of youth was −1.148, and by 2020, it had fallen to −0788. This means that the log of odds in favor of becoming middle-class worker is higher in 2020. This phenomenon seems to have arisen due to the acceleration of the use of digital technology due to restrictions on social activities, which then encourage the birth of new digitally-based professions or jobs. This argument is also evidenced by the increased opportunities for youths to become middle-class workers when entering digital variables into the model. Widespread use of technology seems to benefit youths who are relatively more adaptive to technology compared to older generations. Should young workers have better technology-maximizing skills than their older counterparts, demands for their participation will get higher; and therefore, they may enter the digitalized work market more easily than their older counterpart (Kim et al. 2019).

With regard to several sociodemographic variables that affect the likelihood of individuals getting middle-class jobs such as gender, in general men are more likely to find middle-class jobs than women. This condition provides an explanation why the gender gap still occurs in the workforce in Indonesia. It is more difficult for women to find middle-class jobs compared to their male counterparts because their access to the labor market is generally more limited. Women often have greater household responsibilities such as childbirth and childcare, which causes them to choose a more flexible type of work (Shehu and Nilsson 2014). This may result in female youths’ reduced formal working hours and also their work experience, which in turn becomes an obstacle for them to get middle-class jobs.

Women's labor force participation remains lower than that of men because women are still responsible for domestic work, and even if women are employed in paid work, their distribution is skewed in the informal sector, among the poor and underpaid. They also face a sizeable wage gap, a larger career opportunity gap, and an entrepreneurial ability gap compared to their male counterparts (O’Reilly, Grotti, and Russell 2019). This segregation of opportunities can create variations in the quality and risk or vulnerability of work because women are too often in sectors characterized by poorer working conditions and lower status and wages.

Moreover, marital status also affects the likelihood of individuals getting middle-class jobs. A married person's chance of entering middle-class work market is 1.9–2.3 times higher than those who are not married. In line with previous findings regarding the possibilities for men and women to get middle-class jobs, married men were more likely to get middle-class jobs. In contrast, married women are less likely to be middle-class workers; and they are most likely to be unpaid family workers or self-employed. By working in the self-employment sector, married women generally have more flexible working hours to accommodate their household responsibilities, while paid employment usually requires more commitment and fixed working hours (Pratomo 2014).

Furthermore, education level is also a crucial factor in helping individuals entering middle-class work market. The estimation results show that the probability of being employed in the middle-class jobs increases as the level of education increases. Looking at the results presented in Table 3, labor force in Indonesia with a higher education/university degree (as the highest level of formal education obtained) are approximately 8–15 times more likely to be employed in middle-class jobs compared to those who have lowest levels of formal education. Individuals with a high school education are 3–5 times as likely to enter middle-class jobs, and their chances are not much different from those with vocational education.

Education is indeed a very important piece of human capital in expanding one's job opportunities. The higher the educational level reached by individual, the higher the competence possessed. Individuals with higher skills or competencies have greater competitiveness and are able to compete for access to the labor market, including middle-class jobs. Low educational attainment and early entry into the labor market are strongly correlated with poor job quality, lower wages and a higher likelihood of becoming unemployed at some stage later in life (O’Higgins 2017, 2020). In this regard, Pratomo (2016) also argues that people with higher levels of education have better job prospects, and the difference is very striking between those who have attained higher education and those who have not.

Nonetheless, it should be noted that the unemployment rate among youths with secondary vocational education is much lower than that of graduates with higher education qualification. This is because young people with university qualifications in most cases have higher job expectations, especially in terms of wages, social conditions, and career growth. Therefore, it is more preferable for employers to hire young graduates with vocational competencies, rather than inexperienced university graduates with higher qualification but have not proven their practical experience.

Like formal education, informal education – such as training – is an important provision for individuals seeking middle-class jobs. Participation in training certainly provides more skills so that they can be more competitive in fighting for positions in the middle-class employment sector. Furthermore, innovation, technology, and market developments have turned the world of work into a rapidly changing environment, and training must be ongoing to keep up with this reality. Education and training systems are challenged to equip a growing young workforce with the skills needed for their future jobs, particularly soft skills.

In addition, the results in Table 3 show that the probability of getting a middle-class job is higher in urban areas than in rural areas. The odds of getting middle-class job in urban area are 1.2–1.4 times more than rural area. This is because urban areas tend to have high levels of commercial and government activities as well as several industrial centers with high added value. Therefore, workers in urban areas are more likely to be involved in better paying jobs than in rural areas. Besides, young individuals in rural areas are generally involved in agriculture-based jobs, which tend to have low added value and are vulnerable to bad weather, harvest failure, or other climatic shocks. This may explain why young people living in rural areas are less likely to find middle-class jobs compared to those living in urban areas.

Meanwhile, if we look at the characteristics of middle-class jobs by sector of work, individuals who work in the agricultural sector have the lowest probability of becoming middle-class workers. Individuals working in the agricultural sector are less likely to find a middle-class job (the parameter estimate is −0.28 to −0.77). On the other hand, the mining sector is more than twice as likely to be categorized as middle-class jobs. Other sectors that play a relatively important role in supporting the growth of the middle class are the services, industry and trade sectors.

Table 4 shows the results regarding sociodemographic indicators of the probability of youth getting middle-class jobs. The reference categories in this section are gender, participation in training, place of residence, education level, marital status, and employment sector. First, our results suggest a possible gender gap in middle-class employment with men being more likely to be employed than women. In this case, O’Reilly, Grotti, and Russell (2019) have even stated that despite national differences in youth employment, many countries have striking similarities in the unequal sectoral distribution of employment opportunities for young women and men. Furthermore, Pratomo (2016) also explains that young women in developing countries have much less participation and employment than young men even after considering the effects of childbearing. This result is also in line with several other studies, such as those conducted by Adeniran, Ishaku, and Yusuf (2020), Dibeh, Fakih, and Marrouch (2019), Nunez and Livanos (2010), Garrouste, Kozovska, and Perez (2010), Ahmad and Azim (2010), Hussain, Anwar, and Huang (2016), Rodokanakis and Vlachos (2013), and Msigwa and Kipesha (2013)

Table 4. Logit results of middle-class worker status (youth only).

Variable	2015				2020
	All Employees		Paid Employment		All Employees		Paid Employment
	Coef	Odd Ratio	Coef	Odd Ratio	Coef	Odd Ratio	Coef	Odd Ratio
Male Workers	0.296***	1.3447	0.238**	1.268401	0.606***	1.8331	0.635***	1.8870
Training	0.639***	1.8950	0.530***	1.698472	0.391***	1.4780	0.431***	1.5381
Urban	0.240***	1.2708	0.293***	1.340883	0.501***	1.6503	0.534***	1.7050
University	1.854***	6.3836	2.091***	8.089527	2.003***	7.4126	2.127***	8.3868
Senior High School	0.771***	2.1613	0.923***	2.517005	1.077***	2.9353	1.232***	3.4293
Vocational High School	0.660***	1.9346	0.920***	2.510168	0.800***	2.2264	0.923***	2.5168
Married Workers	0.518***	1.6792	0.438***	1.548898	0.284***	1.3278	0.324***	1.3830
Sector of Employment
Agriculture	0.353	1.4236	0.378	1.460065	−0.237	0.7892	−0.460**	0.6312
Manufacturing	0.793***	2.2092	0.618**	1.85541	0.891***	2.4377	0.581***	1.7873
Mining and quarrying	1.761***	5.8199	2.039***	7.684508	1.259***	3.5224	1.106***	3.0237
Wholesale/retail trade, restaurants, and hotels	0.243	1.2747	−0.343	0.709746	−0.479***	0.6193	−1.101***	0.3324
Transportation and communications	0.687***	1.9871	0.380	1.462659	0.414***	1.5126	0.235	1.2651
Finance and business services	0.883***	2.4176	0.596**	1.815195	0.547***	1.7284	0.174	1.1903
Community, social and personal services	−0.008	0.9920	−0.343	0.70937	−0.206	0.8140	−0.590***	0.5543
Constant	−5.282***	0.0051	−5.191***	0.005566	−5.336***	0.0048	−5.099***	0.0061
Observations	25,524		18,642		37,199		25,433

Note: *** p < 0.01, **p < 0.05, * p < 0.1.

Second, the level of education has a significant effect on the likelihood of youths getting middle-class jobs. Youths who have completed tertiary education are approximately 6–9 times more likely to get middle-class jobs than those with lower levels of education. Meanwhile, the probability of youths with a high school education becoming middle-class workers is 2–3 times higher than those with lower level of education, and the results are not much different from those who take vocational education. As explained earlier, higher education reflects better individual abilities, thus opening up opportunities for youths to obtain better jobs. International Labor Organization (2012) also emphasizes that education and training for youths can increase their employability and encourage higher productivity, which then leads to better quality jobs with higher incomes. The absence of formal education and the inability of the formal sector to adequately absorb the youth workforce have made the informal sector the only alternative employment opportunity for young people (Ismail 2016).

Third, nonformal education, in this case job training, is an important component in helping youth obtain middle-class jobs. The results show that young women and men in Indonesia who have participated in training programs are 1.4-1.8 times more likely to be employed in the middle-class employment sector compared to those who have not attended the training. This is based on the assumption that participation in job training can improve the skills of youth so that they have better prospects in formal jobs or better paid jobs. This result is also indicated in the research of Abdullai, Tresi, and Ramadani (2012), Cairó and Cajner (2018), and Otero (2016), which also show a positive relationship between nonformal education and youth employment.

Fourth, the likelihood of youth becoming middle-class workers is influenced by location factors. The estimation results in Table 4 show that the probability of getting a middle-class job is 1.2-1.7 times higher for young workers living in urban areas compared to those living in rural areas. As explained in the previous discussion, this phenomenon occurs because urban areas tend to have many choices of formal jobs with high added value, in contrast to conditions in rural areas that are identical to agricultural-based informal jobs, which tend to have low added value. This phenomenon explains why youth in rural areas are less likely to find middle-class jobs. In addition, youth in rural areas – who usually have lower social capital than urban youth – also have lower levels of education, so this group tends to end up in informal jobs with low-quality work, low income, risky working conditions and no benefits other than wages paid for their actual hours worked (Tran 2017).

The results in Table 4 also show that marital status can determine the likelihood of youths becoming middle-class workers. Married young individuals are more likely to find middle-class jobs. This is explained by the fact that the standard definition of youth lies between the ages of 15 and 24 and that most youths of that age are usually still being students, and thus being single and unemployed. Meanwhile, Msigwa and Kipesha (2013) further explained that married youths have a greater responsibility in taking care of their families, which requires them to work, while most single youths still depend on their parents, so they are less motivated to work. This explains why youths who are single are less likely to enter the middle-class job market.

Meanwhile, if we look at the characteristics of middle-class employment for youth by sector of work, there are four sectors with the highest probability of being a middle-class employment sector, namely, the mining sector, the financial services, the manufacturing sector, and the transportation sector. The mining sector is the largest sector that contributes to the provision of middle-class jobs. The estimation results in Table 4 show that the odds of being middle class workers in the mining sector is 5–7 times. However, in a pandemic year, the chances are even lower. The probability of youth becoming middle-class workers in the other three sectors is relatively the same.

On the other hand, the sectors that allow the lowest opportunities for youths to become middle-class workers are agriculture, trade, and social services. First, youths in the agricultural sector have the lowest probability of becoming middle-class workers. In 2015, the coefficient for variable agricultural sector was not statistically significant and became negative during the pandemic. Second, the trade sector has no effect on the likelihood of young people entering the middle-class work market in 2015. However, in 2020, the odds of young people becoming middle-class workers in this sector have decreased. In addition, the social service sector affects youth opportunities to enter middle-class work market only for those who work as employees in 2020 (the parameter estimates are −0.590).

5. Conclusions

This study aims to examine the likelihood of young people entering middle-class jobs and to analyze the effect of the socioeconomic characteristics of youths on their likelihood of entering middle-class jobs. Using data from the 2015 and 2020 National Labor Force Survey (Survei Angkatan Kerja Nasional/Sakernas), this study found that young people have less chance of becoming middle-class workers both during normal times and during the pandemic. Interestingly, however, the study also found that the chances of young people, particularly young men, being middle-class workers during the pandemic are greater than during normal periods. This phenomenon seems to have arisen due to the acceleration of the use of digital technology at times of restrictions on social activities during the 2020–2022 pandemic, which then encourage the growing of new digitally-based sectors or jobs. It is therefore necessary to continue to increase the use of digital technology in workplace. Optimizing the digitalization will enable young workers to survive during the pandemic, and even increase productivity in the post-pandemic economic recovery.

Therefore, young people represent a promise to change society for the better. However, there are not enough decent jobs for them. The youth employment crisis, exacerbated by the global economic and financial crisis, is now requiring governments, employers, and workers to work even harder to promote, create and maintain decent jobs, and productive work. Failure to generate enough decent work is dangerous because it has the potential to have long-term effects on the quality of youth competency in the middle-class work market.

On the other hand, this study provides some important findings regarding the important aspects that can increase youth opportunities to become middle-class workers. Increasing levels of education, both formal and informal, are key factors in encouraging youth participation in middle-class employment. A high level of education reflects greater competence and competitiveness so that they are able to compete for better labor market access, especially middle-class jobs. Living in urban areas increases the chances of young people entering middle-class works, since urban areas tend to have high levels of commercial and government activities as well as several industrial centers with high added value. Therefore, workers in urban areas are more likely to be involved in high-paying jobs than in rural areas.

Another variable that increases youth participation in the middle-class employment sector is marital status in which married youths have a greater chance of getting middle-class jobs. However, compared to male workers, female youths – especially those who are married – tend to have difficulties to get into middle-class job market and tend to enter the low-paid job sector because their limited access to enter middle-class jobs. Besides, women often have greater household responsibilities, so they tend to choose a more flexible type of work causing a reduction in formal working hours and work experience. Child-bearing and child-rearing years have such a negative impact on female youths, suggesting policies that support them to enter middle-class jobs, such as providing public preschool and childcare facilities in each village and sub-district that can be managed by local communities.

In addition, this study also finds that there are four dominant sectors in providing middle-class employment for young people, namely, mining, financial services, industry, and transportation. These sectors become middle-class job providers in Indonesia because they tend to have high added value, high contribution to the economy and are also highly compatible with developments in digital technology. Relating to pandemic, all of the dominant sectors show a stable positive effect, although most of the coefficients are getting smaller compared to 2015 results due to decreasing incomes affected by mobility restrictions (lockdowns) and decreasing business activities. In contrast, although agriculture and wholesale/retail trade sectors tend to become a safety valve during the pandemic and were less affected by lockdowns, they are less likely to offer middle-class jobs due to their low added value compared to other sectors (Muljaningsih et al. 2022).

One of the limitations of this paper has been to estimate the determinants of youth employment in the middle class only based on sociodemographic characteristics at individual levels. However, it will allow further studies to consider other factors – i.e. those affecting the possibility of workers getting a more decent middle-class job – such as structural factors, communities, and household factors. In addition, further studies can also consider the qualitative aspects obtained through in-depth interviews of youths who are engaged in middle-class job to find out their responses on the ground.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Direktorat Jenderal Pendidikan Tinggi, Riset, dan Teknologi through Pendidikan Magister Menuju Doktor Untuk Sarjana Unggul [grant number NKB-0267/E5/AK.04/2022].