John P. Martin Director of Employment, Labour and Social Affairs OECD The Greek Labour Market: Features, Developments and Challenges Bank of Greece Conference Athens, 22 March 2010 Tackling the jobs crisis: An OECD perspective John P. Martin Director of Employment, Labour and Social Affairs OECD

The jobs crisis An unprecedented jobs crisis OECD-area UR jumped from 25-year low of 5.8% at the end of 2007 to a post-war high of 8.7% in January 2010 While recovery is underway, the jobs crisis is far from over OECD UR expected to rise to 9.1% by end 2010, but still be at 8.6% by end 2011. Bold action needed to tackle high and persistent unemployment Individuals in jobless households 5x more likely to be poor on average Social costs go well beyond the loss of income (e.g. health, crime etc.) High risk of hysteresis effects

Part I What are the labour market impacts of the crisis? Historical patterns and recent trends

The unemployment impact so far differs greatly across countries Wide variation in UNRs across countries: from lows of 3-4% in Norway, the Netherlands, Switzerland and Korea to a high of almost 20% in Spain. Greece which had an above-average UNR before the crisis, has so far experienced a moderate jump to around 10%. Large hikes in several countries (Spain, Ireland, US, Iceland), note also quite a jump in Denmark (reverse side of the coin of the « flexicurity » model). Note small hikes in some continental countries (Italy, Belgium, France, Switzerland); small drop in GER, unchanged in POL. Has there been a new  "German miracle" during the crisis? [ Will return to this intriguing question] The variation in UNRs across countries during the current downturn – as measured by the standard deviation of the OECD harmonised UNRs – is larger than that recorded in three of the four preceding downturns.  

Different responses of employment to output declines Total percentage change, 2007Q3 to 2009Q3 >While there is a reasonable correlation (0.59) between changes in GDP and changes in total labour input (proxied by total hours worked) over the course of the current downturn, there are some notable exceptions (e.g. Spain where the drop in total hours is the largest at almost 14%; in Sweden, on the other hand, the drop in hours is below-average).

Different margins of adj. in the labour market: employment vs. hours Total percentage change, 2007Q3 to 2009Q3 Now labour input can adjust in different ways over the business cycle, with differential impacts on unemployment. One aspect of the short-run labour market adjustment story which has attracted much attention is that of the intensive (hours worked) vs extensive margin. Prime case is that of GER … often attributed to Kurzarbeit (short-time working scheme) which was made more generous in response to the crisis. I will return to this issue shortly.

The current crisis is the worst in recent decades Index base 100 = unemployment rate at the preceding business-cycle peak (based on output gap), OECD area, quarterly data The hike in OECD unemployment tracks closely the hike after the first oil shock in 1973 but is projected to be higher, bearing in mind my caveat about our projections. Source: OECD Economic Outlook, November 2009.

Recessions not only hurt lots of people, but also take a long time to fix Harmonised unemployment rates in Greece, January 1970 – January 2010 The next five slides illustrate the phenomenon of UNR persistence (so-called « hysteresis »), linked to a build up of long-term unemployment. We begin with Greece which experienced almost a decade of rising UNR following the downturn of the early 1990s. It took about 8 years to unwind most of the UNR hike – and note it never got back to the pre-1990 low of just over 6%. Now with the current crisis it is on the rise again and the specific Greek austerity packages could well push the UNR above the previous peak of over 12%. If that does occur, past experience suggests it could take up to a decade to unwind this.

Recessions not only hurt lots of people, but also take a long time to fix Harmonised unemployment rates in the United Kingdom, January 1970 - January 2010 The sole good news for the UK is that it is unlikely that the current peak in UNR will reach the 10-12% levels recorded in the mid-1980s and mid-1990s. This, together with the drop in hours worked, is an indication that the LM reforms (activation) in the 1990s/2000s are bearing fruit during the current recession?

Recessions not only hurt lots of people, but also take a long time to fix Harmonised unemployment rates in the United States, January 1970 - January 2010 Unlike the UK case, the US UNR is very close to a 30-year high, and could well pass it in the coming months. The LTU rate is now rising very fast, suggesting that it could take a long time to return to the pre-crisis UNR of around 5%.

Recessions not only hurt lots of people, but also take a long time to fix Harmonised unemployment rates in Germany, January 1970 - January 2010 The slide provides suggestive evidence that the Hartz reforms a few years ago seem to have « worked » and that, combined with flexibility in working hours, may mean that Germany will avoid any significant persistence effects on UNR this time round. >Hartz IV: (i) merged the UB and SA schemes; (ii) cut the duration of UBs; (iii) increased emphasis on activation. Estimates by OECD and IZA researchers suggest that these reforms cut the NAIRU by 0.5 percentage points or more. >One indicator in support of this hypothesis is that there has been an inward shift of the German Beveridge curve during this recession to date indicating an improved job-matching process.

Recessions not only hurt lots of people, but also take a long time to fix Harmonised unemployment rates in France, January 1970 - January 2010 For France, the jury is still out as to whether the UNR (currently 10%) will reach or exceed the previous cyclical peaks of 11-12% recorded in the mid-1990s. > Unlike Germany, few major reforms on the activation front until recently (merger of PES with benefits agency to form a new one-stop shop, Pôle Emploi, took place at the beginning of 2009 when the crisis hit the French labour market).

Historically, disadvantaged groups bear the brunt of falling labour demand EULFS data, index of relative business-cycle volatility (national average=100) Youth, low-skilled, temporary workers historically most vulnerable to rising UNR.

A similar pattern holds in the current downturn Percentage change of employment over 2008 Q2 to 2009 Q2 This pattern has been generally borne out during the current downturn, especially for youth and temporary workers.

In addition to job losses, hours reductions and LF withdrawals also account for falling LD Unweighted average across countries, 1983-2007 In terms of the cyclical adpictment of LD, it is interesting to note how it has differed in terms of Average hours worked; Employment rate; LF participation rate. The chart shows the pre-crisis patterns over the 25-year period to 2007. Note: Declines in average hours worked have been more important for prime-age workers; Employment rate variations more important for men than for women; LF part rate changes more important for youth and older workers.

Part II What are the policy challenges? Strategic choices and policy response

LM policy challenges Long-term challenges Short-term challenges Should labour demand policies play a major role? Is the social safety net adequate? Is the work-first approach recession-proof? Long-term challenges How to avoid high unemployment from persisting? How to avoid undermining long-run labour supply? How to avoid undermining long-term labour market efficiency?

Resources available for LM policies differ across OECD countries On average, 1.3% of GDP of which: 0.8% passive and 0.6% active But large differences across countries: e.g. from 0.4% in US to 2.8% in DEN Spending on UBs exceeds spending on ALMPs in almost all countries For Greece prior to the crisis, the LMP spending effort amounted to 0.5% of GDP in 2007, far below the OECD average of 1.5% in that year. It is worth noting that Ireland, Spain and Portugal – the other members of the PIGS group – spent an average of 1.8% of GDP on LMPs. The bulk of Greek spending on LMPs in 2007 went on passive spending (0.34%) and only 0.15% on ALMPs.

Governments have taken many types of measures in response to the jobs crisis Source for this information is a joint OECD – EU Commission questionnaire – see OECD EmO 2009. We are in the throes of updating this questionnaire for the 2010 EmO and the G20 Labour ministerial.

Supporting labour demand Vigorous macro-economic policy response, including large fiscal packages, to boost AD Estimated to save 3.2 to 5.5 jobs in 2010 in the 19 OECD countries included in the analysis These employment estimates are based on OECD simulations of the impact of the fiscal packages – for more details, see OECD EmO (2009).

Ensuring that S-T measures to support labour demand are well targeted and temporary Most OECD countries have introduced measures to support labour demand: Subsidies for the reduction in working time (22 countries) Reductions in non-wage labour costs, in some cases targeted at SMEs, disadvantaged groups (16 countries) Hiring subsidies and work experience, generally targeted at disadvantaged groups (16 countries) S-T labour demand should be unwound when recovery becomes firmly established By protecting job matches these measures have reduced socially and economically inefficient job losses … …but risk increasing dualism (U highly concentrated on already disadvantaged groups and new entrants)... ….and slow-down efficiency-enhancing labour reallocation Hiring subsidies may help promoting employment for disadvantaged groups, but have to be well targeted and with strict conditions for employers. S-T working subsidies (22 countries have such schemes; 15 had them before the crisis and 7 put in place such a scheme during the crisis): In order to minimise the dynamic inefficiency associated with short-time working, we recommend building in some incentives, e.g. NLD has introduced a scheme in which the duration of the subsidy depends on the share of the workforce involved – the highest the share the shortest the duration; it also requires employers to pay back 50% of the subsidy if they dismiss a worker in the three months after the end of the subsidy . Reduction in social security contributions (estimated cost per additional job: 1.7 times average job costs in SR; 7 times in LR ). 16 countries Targeting: Youth: ITA; KOR; UK; US; Older workers BEL; POL; Low paid CZ; Hiring subsidies Targeting: Youth: AUS; CA; FRA; ITA; KOR; UK; US; Older workers BEL; FRA; JPN LTUs PRT; SWE SMEs JPN; US Administration first announced a targeted tax credit for firms which create net additional employment in 2010 and/or pay higher wages. The tax credit was to be worth up to $5000 per hire and the total subsidy per firm was capped at $500K, to ensure that most of the tax credit will go to small firms. The total cost of the Tax Credit was put at $33 bill. But, faced with opposition in Congress, the Administration dropped this proposal quickly and opted to support a much smaller jobs bill ($15 bill) which exempts hires of those unemployed for over 3 months from payroll taxes, and pays a $1000 tax credit if the hires are maintained for a year.

Providing an adequate safety net while minimising benefit dependency The jobs crisis leads to longer average unemployment spells Where unemployment benefit durations are short, temporary extension of benefits during the crisis helps cut the poverty risk among LTU (CAN, FIN, JAP, PORT, US) Extensions should be temporary and targeted to the most vulnerable with enforcement of job-search requirements The crisis can also be an opportunity to reform unemployment benefits to provide a more adequate safety nets to vulnerable groups, conditional on job search, and participation in ALMPs For temporary and other non-standard workers, benefit eligibility has been loosened (FIN, FRA, JAP)

Helping job-seekers find a job Maintain core jobs-search assistance to help jobseekers Even in recessions firms continue to create many new jobs (see EmO 2009, Ch. 2) Cost of job loss increases due to longer expected unemployment duration and loss of human capital Many countries made good progress prior to the crisis in implementing effective back-to-work policies Don’t relax activation strategies; adapt them For those at risk of LTU, re-employment services need to be adapted to specific conditions of slack LM Some shift in emphasis from “work-first” approach to “train-first approach” through training and work-experience programmes: Negative effects of programme participation on job-search less of an issue in recessions Helps provide jobseekers with “the new skills for the new jobs” in the recovery Is there a need for a public sector job creation scheme as a backstop for an “adapted” activation regime in a steep downturn? Requires more resources for ALMPs One thing is clear: maintaining an effective activation regime during a steep cyclical downturn cannot be done on the cheap. As my colleague David Grubb has emphasised, it is vital to maintain interventions and monitoring of those at-risk of LTU (the size of this group will inevitably expand during a downturn) and to have an arsenal of effective ALMPs at the disposal of the PES. This will inevitably require some additional resources and some shift in the mix of ALMPs towards more reliance on training programmes. Question: I confess to some mild schizophrenia on this issue. As David Grubb and I have argued in our review of the evaluation literature, (published in the Swedish Journal of Economic Policy, 2001), most such programmes did not work in the past. It is hard to be optimistic that they would work better today. But it is at least conceivable that with better design (incl. incentives to reward providers who get participants into regular jobs eventually), they might work a bit better. Still a gamble, especially at a time of very strong pressures on public spending.

Discretionary funds for ALMPs limited with some notable exceptions Average annual planned additional expenditure in response to the economic downturn Only three OECD countries have an automatic link between spending on ALMPs and unemployment: Australia, Denmark, Switzerland. In all other cases, the correlation is usually much less than unity, unlike the case with spending on UBs. Maybe useful to extend this automaticity to other countries?

Inappropriate government responses can undermine LS in LR Allowing the unemployed to drift into LTU and inactivity Essential to maintain mutual-obligations approach (“bend but not break”) Early retirement schemes and exemptions from job search of older UB recipients Did not free up jobs for youth and took a long time to unwind So far so good? More recently, large inflows of WA persons into sickness and disability programmes Some countries have made reforms aimed at promoting employment and employability of people with remaining work capacity Will they stay the course?

Helping youth “ride out the storm” Youth unemployment has increased disproportionately in many countries Youth E twice as sensitive to cycle as that of prime-age workers Youth UR much higher than other groups (e.g. 1/3 Spain; 1/4 Ireland) Need for decisive actions targeted on at-risk youth Efforts to prevent youth entering the LM without qualifications should be redoubled (e.g. second-chance school, subsidies for apprenticeships for un-skilled youth) Out-of-school youth should have access to appropriate ALMPs even if they do not qualify for UB

The youth unemployment crisis in Greece: what helps? More than 1 in 4 Greek youth unemployed in Q3 2009 Part of the increase due to the crisis (3 pp since Q3 2008) Youth unemployment likely to have risen further in Q4 2009 + Q1 2010 But youth LM outcomes already dismal prior to the crisis (in 2008, UR 7pp higher than OECD average ; ER 20pp below OECD average) Ensure that Greek youth can access effective ALMPs: Require youth to participate in job-search training early in the UN spell, reserve more costly interventions for those who fail to find work; Target job subsidies on: long-term unemployed youth; unemployed early-school leavers; and youth who have been NEET for over 6m (outreach services required, see Connexions in UK); Set up rigorous evaluations of cost-effectiveness.

The youth unemployment crisis in Greece: what helps? (cont.) To ensure youth enter LM with valued skills: Create a single vocational route in secondary school combining class-based and work-based learning; Expand apprenticeship training to include more professions and encourage SMEs to join forces to provide apprenticeship places. In the longer run, an overhaul of the education system? Expand ECEC; Raise compulsory schooling requirement to at least 16 (OECD median); Change entry system to tertiary education (preparatory year; national exam at end of HS; universities free to manage admissions).

Concluding remarks Governments are intervening actively to minimise the scale of the jobs crisis Some have been more successful to date than others With the recovery in sight, governments must not reduce their efforts to tackle high and persistent U and some countries may have to do more Governments must tackle the jobs crisis without undermining LM inclusion in the long-run Discretionary LM policy measures should be timely, temporary and targeted A severe recession such as currently underway also a requires a sufficiently strong response Need to pursue reforms in labour and product markets. Spain is a classic example: it needs to reform its EPL so as to lessen duality in the LM; A much stronger benefit activation stance (little control of ALMPs in regions, a benefit system which allows temporary workers to qualify relatively easily for up to 2 years of wage-related benefit with no activation, etc.). But is there sufficient political will?

Thank you! www.oecd.org/els/employment/youth