Time preferences and career investments

TL;DR: In this article, the authors examined the role of time preferences in career investments and found that on-the-job search and work effort increase with patience, whereas the relation between job mobility is more ambiguous.

About: This article is published in Labour Economics.The article was published on 2015-08-01 and is currently open access. It has received 10 citations till now. The article focuses on the topics: Promotion (rank).

Summary

1. Introduction

• Climbing up thewage ladder—like any other ladder—takes time and effort.
• There are two distinct career paths one can follow to reach a higher position: through promotions or jobmobility.
• Since on-the-job search andwork effort involve immediate costs and delayed rewards, they can be considered as investment activities.
• Recent literature in economics demonstrates that time preferences predict important social and economic outcomes.
• Other papers (e.g., Fouarge et al., 2014) assess to what extent economic preferences of recent graduates predict their occupational choice.

2.1. A career model

• Internal and external mobility are generally examined in isolation.
• If search and work effort are considered jointly in a theoretical model, this may lead to different predictions.
• Drago (2006) shows that more patient workers invest more in effort, but less in onthe-job search: Drago’s model therefore predicts a negative relation between patience and job mobility, for which he finds empirical support.
• The authors propose an alternative career model where the total level of on-the-job search intensity and work effort is endogenous and both activities are modelled as investments (i.e., generating immediate costs and delayed rewards).

2.2. The optimization problem

• The structure of ourmodel is in the spirit of Moen and Rosen (2013), who developed a model where on-the-job search and work effort are substitutes.
• In period 1, the worker receives wage w and makes career investments c1(s) and c2(e).
• As the model does not rule out the possibility of a negative relation between patience and on-the-job search, there may exist a hump-shaped relation between δ and this career investment.
• Ourmodel predicts that both work and search increase with patience (at least for sufficiently low δ).

2.3. Discussion

• In the model presented above, the authors made several assumptions.
• Onemay argue that there are plausible alternative assumptions, for instance, concerning the costs functions.
• On-the-job search may in that way lead to wage growth (‘promotions’) within the current firm.
• Moen and Rosen (2013) assume a convex cost function of search effort, but a linear cost function of work effort.
• Drago (2006) proposes a theoretical model of search and work effort that allows for hyperbolic discounting.

3.1. Sample

• To examine the relations between time preferences and career investments, the authors make use of the DNB Household Survey (DHS).
• All household members aged 16 or older complete the questionnaire online.
• As the questions about time preferences were not asked in 1993–1995, the authors exclude the first waves of the DHS.
• During this period, the entrants could have been unemployed and in that case their answers may not reflect on-the-job search effort.
• Due to panel attrition and refreshment, the authors make use of an unbalanced panel, consisting of almost 7000 observations (over 2000 individuals).

3.2. Time preferences

• 6 Van Huizen and Plantenga (2014) demonstrate that the items are associated to job search behaviour of the unemployed.
• The authors would expect that CFC01, CFC02 and CFC06–08 are positively related to patience, whereas the other six items can be expected to be negatively correlated with patience.
• Furthermore, various statistics point out internal consistency: the average interitem covariance is equal to 0.38, the value of the Cronbach’s alpha is 0.73 and the overall KMO value is 0.77 (varying between 0.72 and 0.83).

3.3. Work effort

• Wemeasure the effort exerted by employees on the job using two different indicators: a statement about shirking behaviour andworking overtime.
• It may be argued that respondents who agree with this statement have a tendency to shirk at work.
• Several empirical studies examined the investment character of working overtime hours and found a positive relationship between overtime hours and the incidence of promotion (Francesconi, 2001; Booth et al., 2003; Pannenberg, 2005).
• Less than 5% of the workers state that they work less than their contractual working hours.

3.4. On-the-job search intensity

• In the literature on (on-the-) job search behaviour, search intensity has beenmeasured in a variety of ways.
• The latter may be interesting other job”; “Yes, I am considering searching for another job”; “No, I just found another job”; “No, I am not looking”.
• The authors make use of the answer to this question to construct the job search attitude variable.

3.5. Mobility

• Themodel presented in Section 2 leads to ambiguous predictions on the relation between time preferences on the one hand and jobmobility and promotions on the other hand.
• These relations therefore remain an empirical question.
• While respondents are not asked directly whether or not they moved to another job, by using data on the length of tenure and exploiting the panel structure of the data, the authors can infer whether a workers has accepted an outside job between two consecutive waves.
• It appears that whether or not the worker uses a written statement to answer the wage questions is a major determinant of receiving a substantial wage raise at the current employer: this points out that measurement error in the wage data is a problem.
• Estimations based on this restricted sample show generally positive relations betweenpatience and theprobability of receiving a promotion.

4.1. Work effort

• In order to examine the relation between timepreferences andwork effort, two equations are estimated using different dependent variables: a self-assessed measure of shirking and a variable indicating whether the employee works less, equal or more than his contractual hours (‘overtime’).
• Table 7 shows the average marginal effects of patience on the two work effort variables for estimations with and without a squared patience term .
• This is confirmedby themarginal effects estimated at different patience levels.
• This alternative specification (estimated with OLS) leads to the same qualitative result.
• Overall, the findings are in line with the prediction that more patient workers invest more in work effort.

4.2. On-the-job search intensity

• The authors estimate the relation between time preferences and search effort using different estimation methods: the equations using search attitude and a dummy indicating whether the worker applied for another job as the dependent variable are estimated by ordered probit and binary probit, respectively.
• The main findings are presented in Table 8 .
• In all models, the average marginal effects of patience on search intensity are positive and significant.
• This is especially the case in the models estimating the number of applications: the marginal effects seem to be concentrated at lower patience levels.
• Given that patient workers search more intensively for other jobs, the findings suggest that search is an investment activity.

4.3. Job mobility

• To assess the relation between time preferences and job mobility, the authors estimated a probit model with a dummy as dependent variable, indicating whether the worker has made a job-to-job transition between the current and the consecutive wave.
• The average marginal effect of patience and marginal effects estimated at different patience levels are negative and insignificant.
• The theoretical model leads to ambiguous predictions on the effect of patience on jobmobility: patience has a positive job arrival effect and a negative job acceptance effect (at least for lower levels of patience).
• Standard errors in parentheses (clustered at the individual level).

4.4. Robustness tests

• The authors performed several tests to examine the sensitivity of the empirical results.
• Alternatively, the authors used an aggregate patience measure using the factor scores of the first factor of all CFC items (or 8 items, excluding CFC01, CFC04 and CFC05).
• Estimations using such alternative measures lead to similar results as presented above.
• Second, the authors examined whether involuntary job search drives the results.
• Again the estimations lead to the same qualitative results: patience is positively associated with work effort and search intensity, but is not significantly related to job mobility.

4.5. Measuring time preferences by behavioural proxies

• The authors patience measure is fundamentally different from those used in several other studies.
• 14 The DHS includes several questions indicating an individual’s risk preference, such as “I amprepared to take the risk of losingmoneywhen there is a chance that Iwill gainmoney aswell”, which is answered using a 7-point scale.
• The most interesting results concern the estimations of job mobility: the average marginal effects as well as the marginal effects evaluated at different levels of patience are positive (Table 10).
• A potential explanation for the discrepancy in findings is that the proxy measure may capture other unobserved factors.

5. Conclusions

• Workers can pursue different career paths by investing in their current job and by searching on the job.
• Given that the two activities are substitutes,the authors show that at higher 15 The Cronbach alpha is 0.225 and the interitem correlation of 0.068.
• The theoretical model does not lead to unambiguous predictions on the association between the discount rate and job mobility.
• Moreover, Drago (2006) predicts that patience is negatively rather than positively related to search intensity and job mobility.
• First, relying on behavioural proxies to measure patience may generate misleading outcomes: these proxies are rather noisy measures and are likely to capture other characteristics, such as risk aversion.

Figures

Table 2 Patience measure: summary statistics.

Table 6 Correlations between search effort variables.

Table 3 Work effort.

Table 4 Job search effort.

Table 5 Number of channels and applications.

Table 10 Job mobility and patience proxies.

Table 7 Marginal effects of patience: work effort.

Table 8 Marginal effects of patience: search intensity.

Table 9 Marginal effects of patience: job mobility.

Frequently asked questions

Q1. What contributions have the authors mentioned in the paper "Time preferences and career investments" ?

In this paper, the authors examined how time preferences are related to career investments and thereby shape the individual 's career path. 

Q2. What have the authors stated for future works in "Time preferences and career investments" ?

A potential explanation for this inconsistency is that the authors use a different patience measure: whereas previous studies rely on behavioural proxies ( e. g., smoking and alcohol consumption ), they exploit a battery of items indicating the individual ’ s orientation towards the future. The authors test this explanation by re-estimating their models using a patience measure based on behavioural proxies. Future research could exploit more general ( self-assessed ) psychological constructs such as the CFC scale. The results complement recent findings in economics showing that time preferences predict the individual ’ s income level ( Golsteyn et al., 2014 ; Cadena and Keys, forthcoming ), suggesting observed income inequality can to some extent be explained by heterogeneity in the discount rate. 

Q3. What are the common methods of searching for a job?

Answering advertisements, directly contacting employers and asking friends and relatives are frequentlyused job search methods as well. 

Q4. What is the effect of a higher level of work effort on the job acceptance rate?

A higher level of search effort results in a positive effect on the job arrival rate,whereas an increase in the level of work effort generates a negative effect on thejob acceptance rate (1 − F(ŵ)). 

Q5. How does Cadena and Keys (forthcoming) show that impatience negatively?

Cadena and Keys (forthcoming) also demonstrate that impatience is negatively related to school performance and thereby depresses lifetime income: the earnings gap between ‘impatient’ and ‘patient’ individuals is over $75.000 by the time they reach middle age. 

Q6. Why do workers who just (re)entered the labour market have rather distinctive job search behaviour?

The rationale is that workers who just (re)entered the labour market may have rather distinctive job search behaviour, as they may for instance accept a job that they perceive as temporary. 

Q7. What is the role of job search in the wage formation of unemployed workers?

Although promotions and job mobility are typically studied in isolation, recent literature stresses that on-the-job search may play an important role in the wage formation of workers staying in the firm (e.g., Cahuc et al., 2006; Postel-Vinay and Turon, 2010; Moen and Rosen, 2013). 

Q8. What does the model show about the relationship between patience and work effort?

Themodel shows that patience increases work and search effort (at least within a certain range of the discount rate), but that the relation with mobility is ambiguous. 

Q9. What methods are used to estimate the relation between time preferences and search effort?

The authors estimate the relation between time preferences and search effort using different estimation methods: the equations using search attitude and a dummy indicating whether the worker applied for another job as the dependent variable are estimated by ordered probit and binary probit, respectively.