Cross-Country Comparative Research and Quantitative Data Analysis
In my field of research, which combines Economics and Japanese Studies (trained at the University of Bonn), it soon became clear that eventually I would have to deal with cross-country comparisons. Even for quantitative analysis, which deals with statistical data, the challenges of cross-country comparisons – not to mention the time intensity – should not be underestimated. First, one needs “comparable” data. To assess whether data sets are really comparable, one has to understand how the data were collected. To interpret and explain differences in the data and in the statistical analysis, one has to control for differences in the respective national contexts. Here, difficult decisions have to be made as to what differences should be considered and how they should be operationalized and controlled for.
My most recent endeavour with cross-country comparisons was a journal article which I wrote together with Yuji Genda (University of Tokyo) and Ryo Kambayashi (Hitotsubashi University) (Genda et al. 2019). For this research, I visited the Institute for Economic Research (IER), Hitotsubashi University several times between 2015 and 2017. We were interested in what kind of employees did not know about their labour contracts, especially about the duration of their employment.
In Japan, employment contracts tend to be opaque. This contradicts the national labour law, which provides employees with the right of a labour contract specifying among other things the length of employment. Genda (2016, 2018) had already found that 8% of Japanese employees did not know their contract length. We termed this group of employees DNK. We used the data set of the Employment Status Survey (ESS), which provides basic data on Japanese employment structure including the variable that interested us: the knowledge of employees on their contract length.

A typical job advertisement in the restaurant sector, a sector with a high share of opaque labour contracts | Picture by Markus Heckel.
Our assumption was that DNK is a Japan-specific problem. We wanted to confirm this by comparing the situation in Japan with that in another OECD country. It proved extremely difficult to find a country with a comparable set of data, but in the end, we were successful. In Spain, the Economically Active Population Survey (EAPS) provides similar data. For comparative reasons, it was also advantageous that both countries shared a similar labour market structure in terms of a high ratio of non-standard employment.
At first glance, our assumption seemed correct. Spanish data suggested that only 2% of the employees were DNK. However, after carefully checking and controlling the data, it turned out that the comparable figure was actually 11% – a number even higher than for Japan. This was a big surprise and meant that our assumption proved wrong. DNK seemed not to be a Japan-specific problem, but to be of more general relevance.
The next step was to analyse in more detail the economic and social factors characterizing the DNK, possibly explaining why this group tended to be ignorant about an important aspect of their employment contract. To do so meant deciding what factors would be relevant, how they could be operationalized, whether comparable data sets would be available and, if not, how to work around such incompatibilities. In the end, it turned out that DNK in both countries could be characterized by the same socio-economic variables: Women, singles, and less-educated workers were in both countries more likely not to know their labour contract duration. Different data sets revealed, that in both countries this group tended to be dissatisfied with their current job, more likely to search for other jobs and less likely to seek more work in their current employment. It was striking how similar the results for both countries were despite their geographic distance and institutional and cultural differences. Regardless of the consistency of most of the results, there were also non-negligible differences in some of the variables and results. We constructed two additional data sets to combine the advantages of both data sets. We took an important lesson from the results and concluded that despite the different institutional design of their labour markets, Japan and Spain shared similar structural problems. Over time, the ratio of employees not knowing the length of their labour contract stayed the same or even slightly increased, which underlines the severity of the issue.
Overall, we concluded that a better education – starting in school and continuing in the work place – possibly supported by work councils or unions – is necessary to raise awareness among workers about their legal rights as employees. So, although cross-country comparisons are extremely challenging, the results can be highly rewarding. On a more basic level, we also learned that one should never be content with attributing seemingly non-rational results to country-specific cultural factors.
References
Genda, Y. (2016): Shifting the Employment Debate: “Nonregular” Focus Distracts and Misleads, nippon.com, 2016.03.22.
Genda, Y. (2018): Koyō wa keiyaku: funiki ni makenai hatarakikata (Employment is a contract: a working style that does not give in to atmosphere), Tokyo: Chikuma Shobo.
Genda, Y., Heckel, M., & Kambayashi, R. (2019). Employees who do not know their labour contract term and the implications for working conditions: Evidence from Japanese and Spanish microdata. Japan and the World Economy, 49, 95-104.
Markus Heckel is a Senior Research Fellow at the German Institute for Japanese Studies (DIJ) Tokyo. His main research interests include macroeconomics, political economy of central banks, the discourse of monetary policy and labour economics.
Citation: Markus Heckel, “Cross-Country Comparative Research and Quantitative Data Analysis” in: TRAFO – Blog for Transregional Research, 28.10.2020, https://trafo.hypotheses.org/25275.
OpenEdition suggests that you cite this post as follows:
Max Weber Stiftung (October 28, 2020). Cross-Country Comparative Research and Quantitative Data Analysis. TRAFO – Blog for Transregional Research. Retrieved March 25, 2025 from https://doi.org/10.58079/usye