A Consistent Set of Time Series Data on Labour Market Flows for the Netherlands

This paper develops a consistent set of worker flow data for The Netherlands, based on the availability of actual time series and a number of additional assumptions, between all essential states in the labour market: employed, unemployed and out of the labour market. Furthermore, we also construct, based on the same principle, series of job flows, where job creation, job destruction and job reallocation are most important. We find these data to eorrespond to evidence found in surrounding countries and evidence derived from panel data sets. Sensitivity analysis applied to our main assumptions gives an indication of their importance. First version: October 1994. Applied Labour Economics Research Team (ALERT), Free University and TI, De Boelelaan 1105, 1081 HV Amsterdam, The Netherlands. 1 The empirical analysis of the flow approach of the labour market at the macro level is hampered by a lack of consistent time series data. The collation of such data now becomes feasible as recently much progress has been made in establishing proper operational defmitions for the flows of workers and the flows of jobs which play a major role in the theoretical models of the flow approach. However, up to now empirical analysis of labour market dynamics, especially in relation with the cyclical situation, has been conducted mainly on the basis of panel data sets (see e.g. Davis and Haltiwanger, 1990; and Gautier and Broersma, 1994). The flow approach distinguishes various types of worker and job flows (see Hamermesh et al., 1994, for a taxonomy at the micro level). Worker flows can be associated with demographic change and with job mobility. Workers may leave the labour force (flow from employment to non-participation) because they reach the age of retirement or because for some other reason, e.g. marriage, they prefer to stop working. Workers may also leave their job because they found another more suitable job. Finally, there is the possibility that workers are laid-off, which determines the flow of workers from employment to unemployment. Most worker flows are associated with structural change and hence with the processes of job destruction and job creation. But worker flows and job flows are by no means identical. Vacant jobs of those who have quit may be taken by others without changing the character of the job. On the other hand, some job leavers may also have left their jobs because they became obsolete. In that case leaving the job coincides with job destruction. Then there is not only a worker flow but also a job flow. In general outflow from employment to involuntary unemployment will be the result of job destruction (lay-offs = fires). But it can also be true that the person who is fired, was unfit for the job, but that the job remains the same and will be taken by someone else with adequate capabilities. Anyhow, this short outline of various types of job and worker flows shows that in general labour turnover the sum of job mobility and the flows of persons into and out of employment is larger than job turnover the sum of job creation and job destruction. Yet it may happen that within a firm, due to technological progress, someone changes his or her job and this change does not involve a quit and a hire. In that case we have simultaneous job destruction and job creation (job flows), and hence job turnover without labour turnover. The above distinction between various types of labour market flows is relevant from an analytical point of view because the adjustment costs involved in the process of reallocation are dependent upon the type of flow. For instance, for a contractor the adjustment costs may differ when 5 carpenters are to be replaced by 5 other carpenters or when, by the start of a new project, 5 carpenters are to be fired and replaced by 5 bricklayers. Therefore it is 2 This type of flow will not be considered in the macro data of this paper. 2 essential that data on labour market flows consistently describe these relationships between flows of persons and flows of jobs. Job creation takes place either when a new vacancy is created or when someone takes a new job for which no vacancy existed. When someone leaves a job either a vacancy will originate or it involves job destruction. Therefore, flows into and out of the stock of vacancies form an important part of a consistent data set on labour market flows. This paper shows how such consistent data set of annual time series for labour market flows at the macro level in The Netherlands can be constructed. For the construction of the data we use all available information to our knowledge on these flows from various sources. From published data we firstly derive a number of additional time series using the definition equations which are implied by the system. However, for The Netherlands there are not sufficiënt data available from published sources for the construction of the full data set. For that reason we need a number of additional assumptions in order to set up the remaining time series. These assumptions are based on (scanty) information at the micro level, but are also selected on the basis of restrictions on the flow data, e.g. that flows do not become negative. We note that in order to come to a consistent set of data we need time series for all variables in the system and cannot leave one series out. That is because the data set uses a closed accounting framework like in the national accounts. The contents of the paper is as follows. The next section gives an overview of all relevant flows and stocks at the macro level for which time series data are to be constructed. Section 3 discusses the construction method of the data and indicates what assumptions are needed in order to complete the data set. The data which are constructed for the reference period 1970-1991 are presented in section 4. It discusses stylized facts on labour market flows, exposed by the data. Section 5 gives an sensitivity analysis which respect to the major assumptions made for the construction of the additional data. It shows how changes in the assumptions may affect the contents of the stylized facts which are derived from the data. Finaliy section 6 concludes. 2. Stocks and flows at the macro level Figure 1 shows all stocks and flows to be included into a comprehensive data set on labour market flows at the macro level. The figure displays 18 relevant flows and 3 relevant stocks. The 3 relevant stocks are unemployment (U), employment (E) (stocks of persons) and vacancies (V, stock of jobs). The other stock of persons of the figure, the non-participants outside the labour force (N), is the residual stock; for the consistency of the system there is no need to have data on it. Yet it can be set equal to WP U E, where WP is the working age population. The flows are indicated by the general symbol F^,, which denotes the flow from x to y (x,y = e,u,n) with, when relevant, z=j in case of newly created jobs and z=v in case of jobs for which vacancies existed.