勞動供給三維與政策分析

Three Margins of Labor Supply and Policy Analysis

賴志芳(Chih-Fang Lai)

指導教授 ： 陳明郎

國立臺灣大學/社會科學院/經濟學研究所/博士(2014年)

https://doi.org/10.6342/NTU.2014.00821

摘要

歐洲的勞動供給在過去的30年間─70年代初期到2000年代初期─相對於美國下降了30%，其中部分來自於每位工人勞動工時的下降，部分來自於就業人數的減少。目前已有許多文獻探討其成因，Prescott (2002, 2004)認為此差異全為歐洲的高勞動所得稅所造成，Ljungqvist and Sargent(2007, 2008)則認為歐洲優厚的失業救濟金導致居高不下的失業率。然而絕大多數文獻的討論都專注在政策對工時的影響或政策對就業的影響，卻鮮少有文獻探討政策對兩者的相對影響。又就業的變化同時受失業率和勞動參與率變動的影響，因此本博士論文試圖建構一套完整的理論模型涵蓋勞動供給的三個維度：每工人工時、失業率和勞動參與率，以供政策分析之用。
既有文獻中，目前已知涵蓋勞動供給的工時和就業兩個面向的只有Fang and Rogerson (2009)所提出的理論模型。Fang and Rogerson (2009)將工時植入標準的Pissarides配對模型中，得到勞動所得稅會同時降低每工人工時和就業的結果，但他們並未討論救濟金的影響，也未作定量分析。本論文第一篇延伸Fang and Rogerson (2009)的架構，將勞動搜尋的特性加入新古典成長模型的設定中，分析勞動稅和失業救濟金對勞動供給兩個面向的相對影響，並作定量分析。我們發現勞動稅提高的確會同時降低工時和就業；而較高的失業救濟金則會降低就業，但可能提高工時；兩者同時增加約能解釋75%歐美勞動供給的差異，解釋力則受勞動供給彈性和勞動尋職強度左右。
第一篇採用Fang and Rogerson (2009)的設定，每工人工時和工資皆由勞資雙方談判決定，然而文獻上工時有不同的決定機制，即使是在談判工時時，相對談判能力也會有所不同。第二篇延續Fang and Rogerson (2009)的架構，變動勞資雙方的相對工時談判能力。我們發現工人的工時談判能力越大，提高勞動稅對工時的負向影響越大，而對就業的負向影響越輕微。當工人的工時完全由家計單位決定，也就是談判能力為100%時，勞動稅對工時的影響達到最大，搭配效用函數中若是線性於工時，則對就業完全無影響。而工時若是由官方管制，勞資無法隨意更動時，勞動稅提高，想當然耳只會降低就業。
談完工時，第三篇分析就業。工人要就業必須先想工作而後找工作，前者決定於工人主觀的勞動參與意願，後者受限於勞動市場客觀的結構限制，主觀意願與客觀限制的不同連帶影響著政策的效力。因此第三篇在第一章的架構下內生化勞動參與，將就業分離為勞動參與率和失業率兩個維度，同樣以提高勞動稅和失業救濟金為例，分析在勞動參與內生與外生下，政策對每工人工時和就業的影響。本章發現勞動稅提高雖會升高保留工資而降低配對，但當勞動參與內生時，勞動參與的意願也降低了，勞動市場上尋職者減少，配對成功機率相對提高，就業反而較勞動參與外生時降得少。失業救濟金在勞動參與外生時為尋職者的考量，但當勞動參與內生時，工人在決定是否想工作時就已列入考慮，尋職時反而已不是保留工資的一部分。因為失業救濟金的提高鼓勵了勞動參與，就業也跟著增加，得到與傳統勞動搜尋模型截然相反的結果。而兩者同時提高的定量分析在勞動參與內生時對工時和就業的解釋度較佳。
綜合三篇的研究結果顯示，勞動供給可分為三個維度，各個維度的決定機制各有不同，連帶影響政策對各維度的運作效果，又各維度之間互有牽引，如若未將勞動供給的三個維度放在一個完整的架構裡討論，則政策分析的結果可能失真。

關鍵字

搜尋與配對 ； 工時 ； 失業 ； 勞動參與 ； 勞動稅和失業救濟金

並列摘要

This dissertation decomposes labor supply into three margins step by step and studies the relative effects of two adverse labor market institutes on labor supply. Labor supply in Europe declined about 30% relative to the US over the past 3 decades. The decline in labor supply comes from both hours worked per worker and employment. Some studies attributed the declining hours worked to higher labor taxes, while other studies accredited high unemployment rates in Europe to generous non-employment benefits. Fang and Rogerson (2009) is the only exception which incorporates two margins of labor supply.
Fang and Rogerson (2009) embedded working hours into Pissarides matching model and found that higher labor taxes decrease both hours per worker and employment. The first essay of this dissertation starts from Fang and Rogerson (2009) to compares the relative effects of increases in labor taxes and non-employment benefits on hours per worker and employment and quantifies them. We find that increases in labor taxes decrease hours per worker and employment, with an overstated adverse effect on hours per worker if extensive margins are not taken into account. Moreover, increases in non-employment benefits decrease employment and increase hours per worker, with an understated adverse effect on employment if intensive margins are not considered. In the baseline parameterization, we find that increases in labor taxes and non-employment benefits together explain about 75% of declining labor supply in Europe, with the fraction accounted for being increasing in the labor supply elasticity and decreasing in the labor’s contribution in matching.
The second essay adopts the same setup of Fang and Rogerson (2009) but varies the relative bargaining power of workers on working hours. We find that the mechanisms shaping the supply of hours per worker play an important role. In the mechanism when the working hour is bargained by matched job-worker pairs, a higher labor income tax reduces both employment and hours per worker. When the laborer’s hour bargaining power is larger, the negative effect on employment is smaller while the negative effect on hours is larger. In the mechanism when labor hours are decided exclusively by the household, i.e., the laborer’s hour bargaining power is 100%, the negative effects on hours per worker approach to the maximum. In extremis, when the utility of leisure is linear in hours, there is no any effect on employment. In the mechanism when the working hour is effectively regulated by an authority, a higher labor tax only reduces employment without any effect on hours.
The third essay further splits employment into unemployment rates and labor force participation which is endogenous, and compares with the model with exogenous LFP. Because of discouraging LFP, labor taxes decrease employment in our model less than the model with exogenous LFP, have ambiguous effects on hours, and decrease less labor supply in our model. Due to boosting LFP, unemployment compensation increases employment in our model and decreases in the model with exogenous LFP, but with opposite effects on hours, labor supply is ambiguous in both models. With endogenous LFP, the quantitative result explains the difference in labor supply better than the model with exogenous LFP.

並列關鍵字

search and matching ； hours worked ； unemployment ； labor force participation ； labor taxes and unemployment benefits

參考文獻

Alesina, A., E. Glaeser and B. Sacerdote, 2006, Work and leisure in the U.S. and Europe: why so different? in M. Gertler and K. Rogoff eds., NBER Macroeconomics Annual 2005.

Azariadis, C., B.-L. Chen, C. Lu, and Y. Wang (2013) “A two-sector model of endogenous growth with leisure externalities,” Journal of Economic Theory 148, 843–857.

Andolfatto, D. and P. Gomme, 1996, Unemployment insurance and labor market activity in Canada, Carnegie Rochester Series on Public Policy 44, 47-82.

Arseneau, D. M. and S. K. Chugh, 2012, Tax smoothing in frictional labor markets, Journal of Political Economy 120, 926–985.

Azariadis, C., B.-L. Chen, C.-H. Lu, and Y.-C. Wang, 2013, A two-sector model of endogenous growth with leisure externalities, Journal of Economic Theory 48, 843–857.