Structural Reforms and Labour‐market Outcomes: International Panel‐data Evidence

• Author: Claire H. Hollweg,Daniel Lederman,Devashish Mitra
• Date: 01 July 2016
• Published date: 01 July 2016
• DOI: http://doi.org/10.1111/twec.12279

Structural Reforms and Labour-market

Outcomes: International Panel-data

Evidence

Claire H. Hollweg

1

, Daniel Lederman

2

and Devashish Mitra

3

1

Trade and Competitiveness Global Practice, World Bank, Washington, DC, USA,

2

Office of the Chief

Economist, Latin America and the Caribbean, World Bank, Washington, DC, USA, and

3

Department of

Economics, The Maxwell School of Citizenship & Public Affairs, Syracuse University, Syracuse, NY,

USA

1. INTRODUCTION

THE debate over the effects of structural reforms has been revived in recent years, in part

as a result of the ongoing debt crises in European economies. As countries engage in fis-

cal stabilisation through austerity policies, pressure has been mounting for debt-ridden econo-

mies to undertake structural reforms in conjunction with macroeconomic stabilisation. The

main argument in favour of structural reforms is that they tend to improve productivity and

thus may be one of the few potential sources of growth in the context of fiscal austerity. From

the viewpoint of developing countries, we have seen this move before, when numerous econo-

mies undertook both fiscal stabilisation and structural reforms throughout the 1980s and 1990s

(see Figure 1).

This paper contributes to the literature by focusing on the impact of structural reforms on

labour-market outcomes at the country level. Research on the impact of structural reforms on a

comprehensive set of labour-market outcomes is needed to inform current debates about the

potential impacts of comprehensive structural reforms on workers. The outcomes that we study

are the unemployment rate, employment level, average wage index, labour force participation

rates (overall and female) and female employment level. The data on labour-market outcomes as

well as on the dates of countries’ reforms to assess how labour-market trends were influenced by

the advent of structural reforms within countries are publicly available. We adopt a structural

reform index from the literature that proxies for the date when countries reached a threshold of

broad reforms including macroeconomic stabilisation, privatisation, trade opening, as well as the

end of interventionist states such as communism. In principle, the liberalisation date is the date

after which all of the Sachs–Warner reform criteria are continuously met.

These labour-market outcomes jointly capture the aggregate welfare of workers and the

overall health of the labour market. In a labour market with full employment, wages and

employment are the only two outcome variables. For workers who are employed, their wage

will determine their well-being, and for a given wage, greater employment leads to grea ter

welfare of workers as a whole. When frictions in job search and matching are introduced, the

set of all outcome variables describing the labour market also includes the unemployment

rate. When the possibility of opting out of the labour force (exercising the option of not look-

ing for a job) is further introduced, the labour force participation rate becomes a fourth out-

The authors thank Markus Br€

uckner, Jakob B. Madsen, Peter Robertson, Diego Rojas and an anonymous

referee for detailed comments.

©2015 The World Bank The World Economy

©2015 John Wiley & Sons Ltd 925

The World Economy (2016)

doi: 10.1111/twec.12279

The World Economy

come variable. Changes in employment can now take place through changes in the unemploy-

ment rate for a given labour force participation rate or through changes in the labour force

participation rate for a given unemployment rate (or a combination of changes in unemploy-

ment and labour force participation). An increase in the labour force participation rate also

indicates a smaller proportion of discouraged workers and thus an improvement in their over-

all well-being. Yet in the case of the decision regarding whether or not to participate in the

labour market (whether to look for a job or not), the factors affecting men and women can be

different (maternity, raising children, discrimination, etc.). Therefore, we separately look at

the female employment and labour force participation rates.

To our knowledge, there seems to be no other comprehensive empirical investigation in the

existing literature of the impact of structural reforms at the cross-country macro-level on labour-

market outcomes other than the unemployment rate. On the theoretical side, however, there is a

long list of papers that analyse the impact of trade policy on labour-market outcomes. The mod-

els of trade and search unemployment are all extensions of the basic search unemployment mod-

els of Mortensen and Pissarides (Pissarides, 2000). Most take the labour force participation rate

as 100 per cent and focus on the unemployment rate and the wage, which decrease and increase

in labour-market tightness, respectively. However, the lack of comparable wage data across

countries thus far has prevented researchers from also empirically looking at the impact of trade

on the overall wage at the cross-country macro-level. Pissarides (2000) also extends the basic

model to include an endogenous labour force participation rate, and the possible impact of struc-

tural reforms on participation rates is also discussed in this paper. In this context, another litera-

ture shows the discrimination-reducing impact of product market competition (Becker, 1957).

Trade and other structural reforms therefore would encourage women to look for work and

should have a positive impact on female labour force participation. This effect is also empiri-

cally tested herein using female employment and labour force participation rates. This paper thus

fills important gaps in the literature.

The analysis first documents average trends (across countries) in our labour-market out-

comes up to 10 years on either side of each country’s reform y ear. This is done by controlling

for country fixed effects and another specification also controls for real GDP, the labour force

participation rate and the working age population. On average, we find that unemployment

0 5 10 15

Frequency

1960 1970 1980 1990 2000

Year

FIGURE 1

Histogram of Year of Reform

©2015 The World Bank The World Economy

©2015 John Wiley & Sons Ltd

926 C. H. HOLLWEG, D. LEDERMAN AND D. MITRA

rates are higher after reforms than before reforms. In addition, we find that employment is

higher on average and its trend is more positive after as compared to before reforms. In the

case of wages and labour force participation rates (overall and female), it is the trend that

takes an upward turn after reforms.

In turn, we estimate fixed-effects ordinary least squares and instrumental variables regres-

sions of our labour-market outcomes on a reform dummy variable, a time trend, square of the

time trend and the same set of controls mentioned above. Focusing on wages, the Stolper-

Samuelson theorem implies that in developing countries, which generally are abundant in

unskilled labour, liberalisation will increase the wage rate of unskilled workers. However, the

assumptions under which this result holds, including free mobility of labour across sectors,

can be questioned. When labour does not move across sectors (is sector specific) and earns a

higher wage in the protected manufacturing sectors, the fear is that removing protection will

instead lower these workers’ wages. Also, the presence of adjustment costs could lead to

higher unemployment after reforms at least in the short run, if not in the long run. While

trade reforms are a very important component of structural reforms, other aspects of structural

reforms include macroeconomic stabilisation, privatisation and deregulation. Such reforms

provide a better macroeconomic environment along with better economic incentives for every-

one, leading to higher productivity and therefore higher wages and employment. However,

moving to a better environment also involves short-run adjustments with temporary adverse

consequences for labour-market outcomes.

The effects shown by our regressions for the employment level and the wage index are

quite strong and conclusive: even in the presence of a strongly significant and positive time

trend that we control for, reforms have a positive and significant impact on employment and

wages. The interpretation of our instrumental variables estimates of the effect of reforms on

unemployment rates instead range from inconclusive to some weak support for the presence

of an unemployment-reducing effect of reforms. Because the labour force participation rate of

the formal sector is included in the regressions and due to the weaker results on unemploy-

ment after reforms, the results suggest that some of the increase in employment may be com-

ing from informal workers joining the labour force.

The evidence on the effect of structural reforms on labour force participation rates (overall

and female) is somewhat inconclusive even though we find that there is a positive time trend

and positive effect of real GDP. Finally, there is some indication that structural reforms may

have increased female employment. Because we control for the time trend and real GDP, our

results show the impact of structural reforms on labour-market outcomes beyond what hap-

pens through the impact on growth. Redistributive effects in favour of workers, along the

lines of the Stolper-Samuelson effect, may be at work.

While the literature that finds evidence of a positive effect of structural reforms and

greater openness on growth is well established, to our knowledge, there are only two major

cross-country empirical studies that look at the impact of trade policy on unemployment

rates.

1

One is the paper by Dutt et al. (2009) and the other is by Felbermayr et al. (2011a).

2

1

Included in the literature that finds evidence of structural reforms leading to growth is the seminal

work of Sachs and Warner (1995). While the paper’s measurement and estimation framework came

under criticism, Wacziarg and Welch (2008) show that liberalisation has, on average, robust positive

effects on growth. See Baldwin (2003) for a survey of the literature.

2

See also Hasan et al. (2012) for an intra-country study on cross-state and cross-industry variations in

unemployment in the context of trade reforms.

©2015 The World Bank The World Economy

©2015 John Wiley & Sons Ltd

STRUCTURAL REFORMS AND LABOUR-MARKET OUTCOMES 927