The Recovery in Employment
The longest recession in New Zealand for thirty years came to an end in the second quarter of 2009. The economy is now recovering, and economic growth over 2010 is expected to be relatively robust. However, since the end of the recession, employment has continued to fall - by 29,000 over the past three quarters. Further falls in employment are possible and even when employment does start to increase, growth is expected to be sluggish. This paper aims to explore the reasons behind the weak employment outlook. This will provide policy makers with more insight into the future demand for jobs and skills, and allow government to better plan for training and assistance. This paper follows an earlier paper by the Department of Labour which examined labour market downturns since the 1960s.
The employment recovery
History suggests employment usually lags changes in economic activity. Just as firms generally won't reduce staffing levels at the first sign of a downturn, they also generally won't hire at the first signs of recovery. Employers want to be assured the recovery is permanent, and not just a temporary upturn, before committing to the costly exercise of hiring and training new people. The most efficient and low risk option is often for firms to increase the hours of their existing staff before hiring.
Figure 1 shows the recovery in employment following the end of each economic recession (ie where GDP reached its lowest point). It shows that employment often lags changes in economic activity (which can be seen in the initial dip of the lines) but then grows quite strongly. However, it is noteworthy that the two most recent recoveries - in 1991 and 1998 - have been weaker, or at least more delayed, than the three preceding recoveries.
The current recovery in employment is looking like it will be even weaker. Employment has continued to fall following the end of the recession and using the latest NZIER Consensus Forecasts (see dotted line on the below graph), employment is not expected to return to its pre-recession level until 2011.
Figure 1: Employment during recoveries
Are we seeing a new type of recovery?
While comparing recoveries over time can be troublesome in New Zealand because of a lack of data and a lack of recessions, it does appear that recoveries in employment are getting weaker over time. A similar trend has also been seen in other OECD countries, particularly the United States.
So what is behind this pattern? Is it that we are seeing a new type of recovery - a recovery where output can increase without the addition of jobs, such as the "jobless recoveries" seen in the United States over recent history, or is it that recoveries now just take longer? Are jobs being lost from the economy permanently, or are they just taking longer to come back?
This paper begins by looking briefly at how employment has tracked in previous recessions before examining some possible causes of weaker employment recoveries. These include the possibility of increased structural change, greater labour hoarding, the use of temporary workers, and increased uncertainty.
Employment and recessions
Figure 2: Number of people employed in New Zealand, 1959-2009
Following the end of a recession (as defined by a rise in real GDP), employment levels generally begin to recover quickly. In four out of the past five recessions in New Zealand, employment has recovered to its pre-recession peak within five quarters. In the 1976/78 recession, employment was actually higher at the end of the recession than at the beginning. The only recession where employment has taken a substantial time to return to its peak was during the 1987-91 recession. After that recession, it took 13 quarters, or just over three years, for employment to return to its pre-recession peak. This can be attributed to the length of the downturn but also the large structural changes that occurred over this period.
Employment rate and recessions
Due to changes in population growth, it is often more appropriate to look at the employment rate (or employment to population ratio) rather than the total number of people employed. If, for example, a recession coincides with a strong rise in net migration then employment can rise, but the employment rate may fall or remain steady. Given that it accounts for changes in the population, the employment rate can sometimes provide a more accurate picture of the health of the labour market.
History shows that on average, the employment rate falls by just under 2 percentage points in recessions. During the course of the most recent recession (ie between December 2007 and March 2009) the employment rate fell 1.1 percentage points. However, it has since fallen a further 1.6 percentage points over the three quarters to December 2009 to be 2.7 percentage points lower than its peak in late 2007.
Figure 3: Employment rate, 1959-2009
Excluding the 1987-91 recession (where the employment rate didn't fully recover before the next recession), it takes around seven quarters, on average, for the employment rate to return to its pre-downturn level. If the labour market recovers based on the average of these four previous recessions, the employment rate would need to return to its pre-recession level of 66.0% in the December 2010 quarter. Given that the employment rate is currently 63.3%, this would require employment growth to be well above that which is forecast over the next year.
Employment and GDP during the recovery
Many factors determine the recovery in employment, with the strength of the recovery in GDP a key determinant. In general, the strength of the recovery in output is correlated with the size of decline over the recession (ie the deeper the decline, the stronger the recovery). So it is possible that the weak employment recoveries were the result of recent recessions being shallower and economic growth in the recovery being slower. However, there is no evidence that recessions have been getting milder.
Both the 1991 and the 2008/09 recessions saw a larger than average fall in economic activity. The most recent recession has also lasted slightly longer than most previous downturns. In addition, the economic recovery coming out of this recession is forecast to be just as strong as that seen in 1991-93 and 1978-79, although admittedly slightly weaker than average.
So is it possible that the relationship between GDP and employment has changed?
Figure 4 below suggests that employment growth relative to growth in GDP has in fact been getting weaker in the two years following the end of a recession. In the three recoveries beginning in 1967, 1978 and 1983, employment grew by more than half the rate of GDP growth in the two years following the end of the recession. The next two recoveries in 1992 and 1998 however, saw employment grow at about one-quarter of the growth rate in GDP.
Figure 4: GDP and employment growth in the two years following the end of the recession
During the current recovery, employment growth is expected to be even weaker compared to growth in GDP. Using the latest NZIER Consensus Forecasts, the economy is expected to grow by over 5% between March 2009 and 2011, while employment is expected to be only just above its March 2009 level in March 2011.
However, this could just be due to employment taking longer to recover rather than employment growth being weaker. To test this, we can examine how employment grows following the trough in employment (rather than from the trough in GDP as presented above). This would allow for the fact that the lag between employment and GDP may have increased. Indeed Figure 5 shows that when employment does begin to grow following a recession, growth has been fairly typical across recessions.
Figure 5: GDP and employment growth in the two years following their respective troughs
This suggests that employment recoveries are not so much becoming 'weaker' but are instead just becoming more delayed. The lag between employment and GDP appears to have become more pronounced in recent recoveries. Some possible causes of this change are briefly discussed below.
Why is employment taking longer to recover?
As outlined above, it appears that employment has taken longer to recover following recent recessions. The following section offers some speculative reasons for why the pattern may have changed and why employment growth may be weak during the current recovery. Note that these reasons aren't mutually exclusive. It is likely that a number of these factors have contributed to more delayed employment recoveries.
Increased structural change
During a downturn, especially a relatively long one, many firms use the time to re-organise production. Low productivity employees are laid off, some work is reassigned and various production processes are streamlined. A number of firms will cease to exist altogether.
As a result of more efficient processes, employers will often find they can create a similar amount of output with fewer workers. While this is a positive for productivity and the long-term well being of the economy, the result can be the permanent loss of some jobs. For example, it is unlikely that all of the 34,100 manufacturing jobs lost during the recession will return given the long-term decline in industries such as textile and apparel manufacturing. While some manufacturing firms would have closed for good, others will become more productive and require less labour. However, this re-organisation of production processes happens during every recession. What makes this one different?
Layoffs may have become more permanent...
In earlier recessions, it was common for firms or industries to layoff workers with the intention of rehiring them when demand picked up again. However, in more recent recessions (especially the late 1980s/early 1990s downturn), the emphasis has shifted to permanent layoffs with relationships severed between workers and firms/industries. Figura (2003) found that in the United States, structural changes have become more important in recent decades. It was shown that restructuring has become more common during downturns with a larger share of job losses in recessions being permanent. It was also shown that restructuring has mainly affected people with strong attachment to their industry and workers are at first not certain that restructuring related employment changes are permanent. This has led to longer periods of unemployment as often workers need to re-train to develop the skills to find work in another industry. While New Zealand is going through less structural change than the United States, the key challenge for policy makers is still to ensure that the people most affected by the recession are equipped with new and relevant skills that meet employers' needs.
The labour market has become increasingly global...
Technological changes, such as the expansion of the internet, have meant that the labour market has progressed from a national market to a global one where labour on the other side of the world can easily be accessed. The offshoring of many jobs, particularly manufacturing and service jobs, from developed countries like New Zealand to developing countries such as China has accelerated in recent decades. Some jobs lost during the downturn will not return in the same area where they were eliminated and affected workers will take time to find new employment. A number of these changes may have been inevitable in the long run, but the recession has simply brought these changes forward.
The recent period of expansion was both robust and long...
Since the early 1990s, employment and output grew strongly for fifteen years, interrupted only by the relatively mild 1997/98 recession. During boom times, it is usually more profitable for firms to devote resources to maximising sales rather than reorganising production processes. After a long period of growth, there can often be an extended phase of reducing staff numbers and low levels of hiring due to the accumulation of surplus jobs and postponed restructurings. It is also possible firms over-hired during the expansion and that the rise in employment was unsustainable in some areas, for example, such as in property services. Employment in property services grew at over four times the pace of total employment in the five years preceding the recession.
This recession may have also allowed firms to reorganise their production process to take full advantage of the heavy investment in plant and machinery seen over the most recent long expansion. Investment in plant and machinery by businesses grew by an average of 12% per annum over the six years to June 2008. Increased productivity during the recovery could see firms initially increase output without needing to take on new workers.
During a downturn firms often hold on to labour despite declining demand. It can be less costly to hold on to staff than to lay a worker off and have to rehire and train a new worker when the economy recovers. This is especially the case when workers are highly skilled or have a large degree of firm specific-skills. This practise is commonly referred to as 'labour hoarding'.
During the current downturn, there has been evidence of many employers reducing hours and limiting wage increases rather than laying workers off. Hours worked has fallen by 4.2% since December 2007 compared to a fall in employment of only 1.5%. Hours worked per worker are also at a record low and around one quarter of part-timers want to work more hours. As a result, there is a significant amount of spare capacity in the current workforce. While we believe there is somewhat of a structural element in the decline in hours worked, we still expect a cyclical rebound in hours as demand picks up. During the recovery, businesses should be able to increase production by getting more hours out of existing workers rather than needing to hire new staff. However, the theory behind labour hoarding was discussed as early as 1962. So why has there been an increase in labour hoarding in more recent recessions?
The workforce has become more specialised...
The decline in manual jobs and advances in technology has meant more jobs require specialised expertise, training and education. As a result, the cost of letting workers go is higher and the degree of labour hoarding may have increased. A greater degree of labour hoarding will see an increase in the time taken before employers begin hiring again.
The current downturn has followed a period of severe skill shortages...
The unemployment rate was at 4% or lower for the four years from mid-2004 and at times up to one-quarter of firms were identifying a shortage of labour as their biggest constraint on expansion (see Figure 6). This experience may have made employers more willing to hold on to staff, especially highly productive workers, for when the economy recovers.
Figure 6: Percentage of firms identifying a shortage of labour as their main constraint on expansion
More temporary workers
There is a growing trend in many countries of hiring temporary workers or contractors instead of hiring permanent-full time staff. This is a sign of an increasingly demand-led labour market. It is sometimes referred to as 'just in time' employment and has been a common explanation of the jobless recoveries seen in the United States.
Temporary workers don't directly contribute to weak employment recoveries as they are still counted in the Household Labour Force Survey - the official measure of employment. However, indirectly temporary workers can allow firms to wait and see if a recovery is sustained before hiring.
Coming out of a recession, businesses are usually uncertain about the robustness of the recovery and whether there will be demand for their products. The increased use of temporary workers, who are able to be hired on short-notice and dismissed more easily, allows firms to wait and see if the recovery is robust and to only hire workers when they are required. Schreft and Singh (2003) show this is exactly what happened in the jobless recoveries in the United States. Schreft, Singh and Hodgson (2005) argue this did not happen in earlier recessions because it was more costly to delay hiring or continue reducing staff because of a lack of flexibility in the labour market. This lack of flexibility meant that hiring too late could cause firms to forgo revenue while it hired and trained new workers.
In New Zealand, little data exists on temporary employment trends and so it is difficult to assess its impact on the labour market. However, many OECD countries have experienced growth in temporary workers over the past two decades and anecdotally New Zealand has also seen growth in this area following the deregulation of the labour market in the early 1990s. In 2008, Statistics New Zealand reported that 9.4% of employees in New Zealand were working in temporary jobs.
A possible factor that has been mentioned as attributing to jobless recoveries is greater uncertainty. For example, it has been argued that in the United States during the post-2001 recovery the September 11 attacks, the wars in Afghanistan and Iraq, and the accounting and corporate governance scandals, increased uncertainty for businesses and made them less willing to expand their workforces. Hiring decisions are usually based on future demand and greater uncertainty makes that demand harder for firms to predict, thereby inhibiting hiring.
Uncertainty around the current recovery is significant given the global nature of the downturn...
While it is difficult to assess the impact of uncertainty on recessions over time, it is clear that uncertainty surrounding the current recovery is significant given the downturn has been unprecedented in many ways. The turmoil in financial markets was the worst seen in decades and the global downturn was felt across most parts of the economy and in most countries in the world. While uncertainty has eased over the past year, the economic environment remains unpredictable, and considerable questions remain about the strength and sustainability of both the domestic and global recovery.
The global outlook has improved strongly over the past six months although the recovery is expected to be unsteady. Sustained growth in many developed countries is not assured and there continues to be significant risks around the outlook. Recently there have been concerns over the fiscal position of many European countries and it is unclear how countries will deal with the removal of fiscal and monetary stimulus.
There may be structural changes...
The degree to which there will be structural changes following this recession will also affect the recovery. For example, both business investment and consumer spending may be much lower than in previous recoveries as businesses pay off debt and households increase savings following a period of excessive borrowing. Data suggests this may already be happening. Both the cost and availability of credit may also have permanently changed. This large degree of uncertainty will affect businesses hiring and investment decisions. Many firms may wait longer than previously seen to see concrete evidence of a recovery before hiring.
During the past twenty years, growth in employment during the recovery stage of the economic cycle has become weaker. This appears to be due to an increase in the lag between employment and GDP. Employment during the past two recessions has taken longer to turn following the end of the recession than seen in the three preceding recessions. Looking forward, employment is expected to take even longer to return to its pre-recession level. Similar patterns of 'jobless recoveries' have also been seen in other OECD countries, particularly the United States. This appears to be due to a major shift in the relationship between the labour market and GDP growth rather than a 'one off' break from historic patterns.
This paper has offered some possible causes for weaker recoveries in employment. These include the possibility of increased structural change, greater labour hoarding, the use of temporary workers and increased uncertainty.
In regards to the current recovery, we believe that increased labour hoarding (as a result of the difficulty firms faced in finding staff in recent years) and the large degree of uncertainty around the economic outlook (as a result of the worst financial crisis seen in decades and the associated downturns experienced in major economies) are the key explanations of employment taking longer to recover. However, whatever the reasons for its occurrence, a subdued employment recovery poses a number of challenges for policymakers and analysts.
A fragile recovery in employment suggests we will see unemployment remain elevated for some time. This has implications for vulnerable groups in the labour market such as youth, the low skilled and those with little or no qualifications. It also has implications for Maori and Pacific labour market outcomes given they are generally less qualified. With low demand for labour and a high unemployment rate, there will be a large group of people competing for a limited number of jobs. Getting these vulnerable workers into some sort of meaningful employment or training will be a significant challenge given an elevated unemployment rate. Ensuring youth have the skills and ability to enter the labour market is also of paramount importance. Long spells of unemployment when young can have long-term negative effects on subsequent labour market outcomes.
Also of particular concern will be the long-term unemployed. Around one in four people have been unemployed for longer than 26 weeks and this proportion is expected to increase over the next year. With some jobs not expected to return (eg some manufacturing jobs) there will be a pool of unemployed who will need to retrain to develop the skills needed to find work in another industry. It is important that the people most affected by the recession are able to access education and training opportunities to develop new and relevant skills and move into employment.
The Department is currently undertaking more in-depth work in this area to try to better understand the explanations for weaker employment recoveries. Results of this work will be published later in 2010.
 See Department of Labour (2009), “How bad is the current recession: labour market downturns since the 1960s”
 Care should be taken when examining data pre-1986. Official estimates of employment from the Household Labour Force Survey only began in 1986. All data used pre-1986 are estimates sourced from NZIER – see Chapple, S. (1994), “HLFS-consistent labour market data”, NZIER working paper 94/16.
 March 2010 Consensus Forecasts, New Zealand Institute of Economic Research. NZIER Consensus Forecasts are an average of New Zealand economic forecasts compiled from a survey of ten financial and economic agencies.
 Technically, there were two or even three recessions over this period, however, in the broader sense this can be treated as one long downturn given the recovery from each of the recessions lasted no more than two quarters.
 See Friedman, M. (1964), Monetary Studies of the National Bureau, the National Bureau Enters Its 45th Year, 44th Annual Report (NBER, New York), and Friedman, M. (1993) "The 'plucking model' of business fluctuations revisited", Economic Inquiry 31.
 See Box E, June 2009 Monetary Policy Statement, Reserve Bank of New Zealand.
 Using March 2010 Consensus Forecasts, New Zealand Institute of Economic Research.
 Figura, Andrew (2003), “The Effect of Restructuring on Unemployment”, Board of Governors of the Federal Reserve System.
 See also Knotek II, Edward and Steven Terry (2009), "How will unemployment fare following the recession?", Federal Reserve Bank of Kansas City.
 Oi, Walter (1962), "Labor as a Quasi-Fixed Factor", The Journal of Political Economy, Vol. 70, No. 6.
 For example, see Congressional Budget Office (2005), “Employment during the 2001-2003 recovery”, The Congress of the United States and Ranchhod, Satish (2009), “Lessons from Previous US Recessions and Recoveries”, Reserve Bank Bulletin, Vol. 73, No. 1.
 Schreft and Singh (2003), “A Closer Look at Jobless Recoveries”, Federal Reserve Bank of Kansas City, Economic Review, Second Quarter 2003.
 Schreft, Singh, and Hodgson (2005), “Jobless Recoveries and the Wait-and-See Hypothesis”, Federal Reserve Bank of Kansas City, Economic Review, Fourth Quarter 2005.
 Survey of Working Life, Statistics New Zealand.
 See Groshen, Erica L. and Potter, Simon (2003), "Has structural change contributed to a jobless recovery", Current issues in economics and finance, Vol. 9, No. 8, Federal Reserve Bank of New York or Congressional Budget Office (2005), "Employment during the 2001-2003 recovery", The Congress of the United States.