Seppo Laaksonen

Who bears the burden of wage cuts? Evidence from Finland during the 1990s

This paper focuses on the share and incidence of nominal and real wages cuts in the Finnish private sector. It complements other analyses of downward wage rigidities especially by looking for individual and employer characteristics that might explain the likelihood of observing an individual’s wage cut. The examinations are based on Probit models that include individual characteristics, employer characteristics, and the form of remuneration as explanatory variables. We find relatively few individual or employer characteristics that have a strong and common influence on the likelihood of wage decline across the different segments of labour markets. However, the full-time workers have had a lower likelihood of nominal and real wage declines during the 1990s compared with part-time workers. Declines in wages have also been more common in small plants/firms. In addition, nominal wage declines have been more transitory by their nature within the segments of the Finnish labour markets in which they are more common. Overall, the frequency of nominal wage declines has been fairly low for manufacturing non-manuals and service sector workers but somewhat higher for manual workers in manufacturing. However, nominal wage moderation together with a positive inflation rate produced real wage cuts for a large proportion of employees during the worst recession years of the early 1990s.

Technology, job creation and job destruction in Finnish manufacturing

Job creation, destruction and reallocation rates is examined using data on Finnish manufacturing establishments for the period 1987-93, focusing on differences in these rates across the technology levels, and on the effects of the 1990s recession on the rates. The results indicate that the high technology sector has higher job creation and destruction rates, and it has responded to the recession somewhat differently than other sectors. It is also found that the high and low technology sectors are contributing differently to job reallocation: high technology is more important (compared to its employment share) in job creation, entry, and gross reallocation, while low technology is more important in job destruction, exit, and net job decrease.

Retrospective Two-stage Cluster Sampling for Mortality in Iraq

The corresponding total was also published, claiming that about 601,000 (95% confidence interval 427,000 to 794,000) violent excess deaths had occurred in Iraq by July 2006. This is more than ten times the sort of official truth obtained from the register of deaths maintained by the Iraq Body Count (between 43,450 and 48,164 deaths for the same period). It is quite conceivable that the registration of deaths cannot be done correctly in such circumstances, but this difference is much more than anyone expected. This was just one reason for the debate.

Alternative imputation techniques for complex metric variables

This paper deals with imputation techniques and strategies. Usually, imputation truly commences after the first data editing, but many preceding operations are needed before that. In this editing step, the missing or deficient items are to be recognized and coded, and then it is decided which of these, if any, should be substituted by imputing. There are a number of imputation methods and their specifications. Consequently, it is not clear what method finally should be chosen, especially when an imputation method may be best in one respect, and another method in the other. In this paper, we consider these questions through the following four imputation methods: (i) random hot decking, (ii) logistic regression imputation, (iii) linear regression imputation, and (iv) regression-based nearest neighbour hot decking. The last two methods are applied with the two different specifications. The two metric variables have been used in empirical tests. The first is very complex, but the second is more ordinary, and thus easier to handle. The empirical examples are based on simulations, which clearly show the biases of the various methods and their specifications. In general, it seems that method (iv) is recommendable although the results from it are not perfect either.

Designing a Questionnaire and Survey Modes

We start from the cornerstones of survey research that give a general understanding of this chapter (Scheme 3.1) based on Salant and Dillman (1994) and De Leeuw, Hox, and Dillman (2008b). This chapter concentrates on measurement in surveys, the target being specifically to avoid measurement error or to evaluate its impact on estimates.

Missingness, Its Reasons and Treatment

We have described many things relating to missing values in the previous chapters but have not described them precisely. The focus of this chapter is how to deal with missing values. The outcomes of this are then used in the remaining chapters, particularly as they concern reweighting, imputation, and survey analysis. It therefore would be good to come back and look at this chapter if something is not clear when one is reading the later chapters. The examples here are mainly taken from the ESS, which includes missingness information. They thus have been calculated from the fieldwork data available. Other considerations and examples can be found in many sources and often in conference papers. We do not give many references, nonetheless it is good, for example, to compare the two ESSs (the European Statistical System, abbreviated to ESS). Stoop (2017) makes a useful comparison between the two. Gideon (2012) is also a good book to read, particularly the chapter by Stoop and Harrison (2012). Koch, Halbherr, Stoop, and Kappelhof (2004) focus on quality comparisons.

Concept of Survey and Key Survey Terms

We determine the survey in its relatively short form, as follows, but it can be defined in many other forms as well (Laaksonen, 2012):

Summary and Key Survey Data-Collection and Cleaning Tasks

This is a summary of survey actions and consists of a long list of steps and tasks in the order that is roughly the one followed in practice.

Basic Survey Data Analysis

This chapter includes basic survey data analysis, such as estimating frequencies, means, and statistical models, using ‘survey instruments’ but not going into most complex cases. The purpose is to give instructions for survey analysis using general statistical software, such as SAS, SPSS, STATA, or R, but without details about them. Examples using PISA and ESS are the main part here, some being derived from ESS-related test data as well (see Sect. 6.1). The chapter is primarily based on these examples. They do not cover complex samples, but such can be rather straightforwardly calculated using one of the software packages. The SAS, SPSS, STATA, or R software work well with the following sampling designs in cross-sectional surveys: Simple or stratified random sampling Equidistance sampling, assuming that it corresponds to simple random sampling Unstratified or stratified two-stage cluster sampling

Imputation Methods for Single Variables

This chapter considers imputation methods for single variables. Naturally, it may be necessary to impute the values of several variables in each dataset and to carry out several imputations for each dataset. It is essential to understand the basics of Chap. 11, which presents the starting point for imputation methods. It is helpful to look at that chapter for the core terms, but an important question is also why one should, or should not, use imputation. Before answering this question, it is necessary to analyse the missingness and the reasons for it thoroughly. Then again, it is good to remember that the imputation methodology always depends on the case; thus, each variable should be separately imputed even though the principles of the method used can be similar. Successful imputation therefore is ‘tailored’ to the specific case, and the best results are obtained if the ‘imputation team’ has sound knowledge of the basis of the data and its quality.