Macroeconomic policy,labour market institutions and employment outcomes

Introduction

The Great Recession began in the USA in December 2007 and did not reach its trough until 19 months later in June 2009. Exactly two years after the recession officially ended, the Bureau of Labour Statistics reported an unemployment rate of 9.2 per cent. A little more than 14 million US workers are officially counted as unemployed, 6 million of them for 27 weeks or more. Using a broader definition that includes those who are working part-time because they cannot find full-time work and those who have given up searching for a job, 25 million workers out of a labour force of 137 million are unemployed, underemployed, or have given up looking but are ready to take a job. Although policy makers responded early with a mix of macroeconomic and labour market policies that reduced the severity of the contraction and the extent of job destruction, a comparison of harmonized unemployment rates shows that the increase in the unemployment rate in the USA was among the largest among the wealthy OECD countries. Nearly four years after the onset of the recession, the US labour market remains in a serious crisis; it is not expected to return to pre-recession levels of employment for several more years.

While policies to stimulate the economy were undertaken early in the recession, there has been only minor subsequent effort to address the jobs crisis. The reason for this failure of policy makers to act more aggressively to create jobs rests, in part, on the nature of mainstream macroeconomic models. These models have not provided a persuasive analysis of the underlying developments that led to the recession; do not include an accurate description of the proximate cause of the contraction; do not examine the role of labour market institutions in moderating or exacerbating the decline in demand; and cannot explain how such an employment contraction can persist. For example, in a recent issue of the Journal of Economic Perspectives, Robert E. Hall – president of the American Economic Association – wrote (Hall, 2010: 3):

Existing macroeconomic models account successfully for the immediate effects of a financial crisis on output and employment. I will lay out a simple macro model that captures the most important features of modern models and show that realistic increases in financial frictions will generate declines in real GDP and employment of the magnitude that occurred. But the model cannot explain why GDP and employment failed to recover once the financial crisis subsided – the model implies a recovery as soon as financial frictions return to normal.

In this article, we develop the argument that the increase in income inequality and household debt of middle- and lower-income households over several decades increased the probability of default on this debt. It also led to an increase in the size and importance of the financial sector which led to inflation in asset prices including, prominently, house prices. These developments fed on each other and made the financial system increasingly fragile. This set the stage for the Great Recession – an economic crisis whose severity was unparalleled in the previous 60 years. The crisis did not begin in financial markets; rather, the proximate cause of the recession, which began in 2007, was the bursting of the housing bubble. The collapse in house prices that began in 2006 had a two-fold effect. It reduced the ability of households to consume and, in particular, to borrow and purchase durable goods – cars, furniture, appliances – or houses. And it left residential construction companies with a large inventory of new homes that could not be sold at their offering prices. The result was a demand shock that triggered the recession. The collapse in house prices also reduced the value of assets – most directly mortgage-backed securities – on the books of banks and other financial institutions, and caused many of them to become technically insolvent. This created the financial crisis that began in the spring of 2008 and took hold in the fall of that year. The financial crisis that began in the USA quickly spilled over to much of the world economy, causing sharp declines in output (GDP) in many countries and a collapse in world trade. We note, but do not take up these world-wide ramifications.

The fiscal and monetary policy responses to the crisis were swift in the USA and, at least with respect to monetary policy, innovative. Yet the performance of the US labour market in the aftermath of the economic crisis has lagged behind that of most other wealthy nations. In part, this reflects the inadequacy of the fiscal policy response to the demand shock and the challenges facing monetary policy as the US interest rate approached the zero lower bound. In large measure it is also the result of the relative weakness of US labour market institutions. As a result of this, the decline in GDP imposed higher costs in terms of unemployment than was the case for most OECD countries (Schmitt, forthcoming). We argue that whether a demand shock that reduces the demand for goods and services leads to a fall in employment or to a fall in average hours of work (or to some combination of these two approaches) depends on the nature of labour market institutions. If employment protections are weak and the costs to employers of reducing employment are low, much of the adjustment to a lower demand for products will fall on employment rather than hours of work and much of the cost will be borne by workers. Absent a sufficiently large, positive demand shock, scarce jobs will be matched to job seekers in a process that increasingly disadvantages those experiencing longer spells of unemployment. The result is high and persistent cyclical unemployment (Diamond, 2011: 1066-9).

The article proceeds as follows. The next section examines the literature on income inequality and financial fragility to identify the source of the underlying developments that led to the economic crisis. The third section examines the chronology of events to establish the proximate cause of the recession and the macroeconomic policy response in the USA. We briefly discuss the challenge to the effectiveness of monetary policy as short-term interest rates approach zero. The fourth section examines the role played by weak labour market institutions in the failure of macroeconomic policies to create more robust job growth. The article concludes with a discussion of the implications for policy.

Income inequality and financial fragility

A compelling case can be made that rising income inequality in the USA since the late 1970s and rising household debt led to increasingly fragile financial institutions and set the stage for both the recession that began in December 2007 and the financial crisis that followed soon after (Blair, 2010–11; Palley, 2009, 2011; Rajan, 2010; Reich, 2010). A great deal of work has been done documenting the long run changes in the distribution of income and wealth in the USA (see, for example, Autor et al., 2008; Card and DiNardo, 2002; Di, 2007; Gottschalk and Moffitt, 1994; Mishel et al., 2009; Wolff, 1998). Wage inequality for men increased steadily from the 1970s, for women from the early 1980s. In the 1970s, the increase in the dispersion of hourly wages occurred mainly at the bottom of the wage distribution; in the 1980s, throughout the distribution; and since 1990, mainly at the top (Heathcote et al., 2010). With respect to household income inequality, the most authoritative and widely cited study was carried out by Piketty and Saez (2003, updated by Saez, 2010) and finds that income disparities increased dramatically in the USA beginning in the 1980s, with incomes at the very top growing much more rapidly than other incomes and top income earners capturing an increasing share of total income. The authors show that for four decades, from the late 1940s until 1988, the top 1 per cent of income earners received about 10 per cent of all income. After 1988, however, the share of the top 1 per cent grew steadily, and stood at 23.5 per cent of total income in 2007. The top 5 per cent of income earners captured 34 per cent of all income in 2007.

Several writers have documented the rise in household debt over this period, arguing that households in the middle and the bottom of the income distribution responded to the prolonged period of falling real wages and stagnant household income by borrowing in order to maintain consumption and living standards (Palley, 2009; Rajan, 2010; Reich, 2010). Strong evidence that households do indeed borrow in order to smooth and maintain consumption can be found in Heathcote et al. (2010). Using time series data from the Consumer Expenditure Survey (CEX) for both income (adjusted for taxes and transfers) and for consumption (excluding the purchase of durable goods such as automobiles or appliances) over the period 1980 to 2005, the authors find that

Household income at the 90th percentile of the income distribution was nearly 4.5 times household income at the 10th percentile in 1980; by 2005, it had increased to more than 6.5 times. However, the ratio of household consumption at the 90th percentile, which was 3.5 times that at the 10th percentile in 1980, increased to just under 4.5 times in 2005. Thus, dramatic increases in income disparities were accompanied by more modest increases in consumption disparities. This suggests that poor and middle-class households were able to borrow to maintain rising living standards, and that much of the run-up in household debt preceding the economic crisis occurred among these households (Heathcote et al., 2010).

Household borrowing increased dramatically after 1983. The ratio of debt to personal disposable income, which had increased only modestly from 55 per cent of disposable income in 1960 to 65 per cent in the mid-1980s, more than doubled over the next two decades. It skyrocketed to an all-time high of 133 per cent in 2007 (Federal Reserve Bank of San Francisco, 2009). Viewed in relation to GDP, household debt nearly doubled from 75 to 140 per cent of GDP between 1983 and 2007. Mortgage debt accounted for most of the increase in debt-to-income ratios preceding the economic crisis (Kumhof and Rancière, 2010). The increase in debt enabled personal consumption expenditures to grow faster than disposable income and this provided a significant boost to economic growth (Federal Reserve Bank of San Francisco, 2009).

This increase of debt in relation to income (i.e. increased leverage) greatly increased the financial fragility of the economy. Declines in asset prices (e.g. house prices) have a large effect on the wealth of highly leveraged households and increase the risk of default. This risk was exacerbated by the dramatic increase in the share of mortgage loans to households that did not qualify for conventional mortgages. The share of so-called subprime mortgages increased from 4 per cent of mortgages in 2000 to 15 per cent in 2006 (Kumhof and Rancière, 2010).

The increase in household debt was facilitated by looser regulation of financial institutions and by the rapid development of innovative financial products and vehicles. Regulators allowed increases in leverage for investment banks and other financial institutions that did not carry the explicit government guarantees that applied to commercial banks. In particular, the Securities and Exchange Commission reduced capital requirements for investment banks, leading to an increase in leverage in these institutions (see Blair, 2010–11: 308–11 for a discussion of the regulatory changes that allowed financial institutions to take on excessive leverage). Financial innovation – asset backed securities (including residential mortgage-backed securities or RMBS), derivatives, and special purpose entities that moved assets off the books of regulated banks – made it possible for unregulated financial institutions to provide credit and liquidity to the economy. The historically low rate of default on residential mortgages allowed them to be pooled and securitized, with some tranches of these securities considered safe enough for regulated financial institutions to invest in them. The result of this deregulation and innovation was a rapid increase in the availability of credit to finance the purchase of homes (Blair, 2010–11; Brunnermeier, 2009). These financial innovations fuelled an expansion in credit that outpaced the ability of the real economy to create new assets, leading investors to bid up the price of existing assets – and families as well as speculators to bid up house prices. In the process, asset bubbles were created (Blair, 2010–11: 268; Brunnermeier, 2009: 78–9). Moreover, high and rising house prices fuelled a boom in residential construction and led to substantial overbuilding as contractors were able to secure financing for projects (commercial as well as residential) based on the assumed appreciation in the value of real estate. Construction peaked in 2005 (US Census Bureau, 2010).

This increase in the use of leverage by both homeowners and financial institutions increased the financial fragility of the economy. While the housing market was rising, individual homeowners and financial institutions could make themselves better off by using as much leverage as possible. But the financial system as a whole was made more fragile, that is, more likely to suffer a major financial crisis. Even a small fall in house prices can cause serious problems for highly leveraged individuals who bought houses at inflated prices with little or no down payment. The sharp decline in house prices that began in 2006 led to unprecedented rates of default by homeowners and created severe problems of solvency for the highly leveraged institutions that held the mortgage-backed securities that had financed home purchases, effects that were quickly transmitted to other financial institutions.

Taking the argument full circle back to the increase in income inequality over the last three decades, Blair argues that asset bubbles are behind the growth in compensation in the financial sector that fuelled the rise in incomes at the top of the income distribution.

[B]y generating inflation in the asset classes they were financing, participants in the financial sector were able, for an extended period, to show gains on the portfolios they were managing that appeared to more than offset the costs of their own compensation. Investors are more than happy to pay high fees, salaries, commissions and bonuses to financial market actors [in these circumstances]. As long as the bubble had not yet burst, the illusion of value creation therefore caused investors to accept higher leverage and to justify extraordinary compensation packages for the participants in the financial sector. (Blair, 2010–11: 232)

Over the last three decades, average compensation in the financial sector has grown much faster than in other parts of the economy. This timing parallels the rise in the share of income going to the top income earners and this rise in compensation contributed to the dramatic increase in income inequality in the USA (Blair, 2010–11: 279-90).

Economic policy in the great recession

To anticipate the potential for a serious crisis and to devise policies capable of restoring growth and creating jobs after the demand shock would seem to require a clear understanding of the proximate cause of the economic crisis. The key point is that it was the collapse of the housing bubble that caused both the recession that began at the end of 2007 and the financial crisis that erupted in 2008 with the failure of Bear Stearns in the spring of that year and the demise of Lehman Brothers in the fall. To see this, consider the following.

Inflation-adjusted house prices in the USA had been flat for a century, from 1895 to 1995 (Schiller, 2006). Beginning in 1996, however, house prices began to rise faster than the rate of inflation. Low interest rates, lax underwriting and a frenzy of innovative financial products, including residential mortgage-backed securities and collateralized debt obligations backed by subprime mortgages expanded the availability of credit and fuelled the rise in house prices. House prices, adjusted for inflation, increased by 70 per cent between July 1996 and July 2006. The increase in house prices added $8 trillion to household wealth – wealth that households would not have had if inflation-adjusted house prices had followed their historical path and remained flat (Baker, 2008). The additional housing wealth had two effects – it added between $400 billion and $500 billion a year to household consumption spending (see Gramlich, 2002 for an estimate of the wealth effect on consumption) and it led to a precipitous drop in personal saving out of disposable income to 1 per cent (St. Louis Federal Reserve Bank, 2010). The run-up in house prices encouraged builders to ramp up construction of new homes: housing starts peaked at 2.1 million in 2005, 50 per cent above their pre-bubble level (US Census Bureau, 2010). House prices fell substantially and housing wealth declined by $6.4 trillion between 2006 and 2008, dragging down consumption spending. Uncertainty about house prices led to a sharp decline in the demand for housing and residential construction collapsed (US Census Bureau, 2010).

The collapse of house prices undermined the financial calculations of highly indebted households and the default rate on residential mortgages spiked, causing the value of mortgage-backed securities – and complex financial products based on them – to decline precipitously. The result was the financial crisis that erupted in 2008 and a decline in the stock market as equity prices fell by half between late 2007 and early 2009. Thus, the collapse of the housing bubble had two effects – it led to the recession that began in December 2007 and it sparked the financial crisis that took hold in the fall of 2008.

By the time Barack Obama took office in January 2009, financial markets were in turmoil and the combination of a decline in housing and stock market wealth plus a decline in residential and non-residential construction opened up a hole in private sector demand of about $1.2 trillion. Most economists had missed the housing bubble, did not expect real estate valuations to come down significantly, and did not consider that a decline in house prices could set off a major financial crisis (Hall, 2010: 17–18). When the economy began to contract, they focused on the financial crisis; most continue to view the crisis in financial markets as the real culprit behind the steep contraction in output and employment. ‘The dominant view among macroeconomists today is that a financial crisis causes real economic activity to collapse by raising frictions’ (Hall, 2010: 6; see also Bernanke et al., 1999; Hall, 2011). In this view, reducing financial friction is the key to ending the financial crisis and restoring economic activity. Monetary policy, including bold and innovative policies to bail out the financial sector, was undertaken for just this purpose and was extremely successful in returning this sector to profitability. Capital is available and corporate borrowing costs as well as mortgage interest rates are low. The real economy, however, continues to struggle.

In contrast to the dramatic steps taken by government to rescue the financial sector, fiscal policy – while undertaken quickly – was timid and conventional. In February 2008, just 10 weeks after the start of the recession, President Bush signed a fiscal stimulus package that provided a total of $170 billion in tax rebates for lower and middle-income taxpayers and tax incentives for business investment. This was followed a year later, after Obama became president, by a second fiscal stimulus that passed with almost no support from Republicans. The legislation provided $784 billion in stimulus funds over two years or about $400 billion a year. In December 2010, a small effort was made to provide further fiscal stimulus, mainly through tax breaks. The main provisions were an extension of the Bush era tax cuts, including those for the wealthiest Americans, for two years; an extension of emergency unemployment benefits for 13 months; and a reduction in payroll taxes paid by workers for one year. These measures stopped the free fall in output and employment (Council of Economic Advisors, 2010), but were too small in relation to the $1.2 trillion hole in private demand to ignite robust job growth. The impact of the collapse in house prices was too great.

Mainstream macroeconomics is sceptical that a large fiscal stimulus can be effective in closing output gaps and prefers the use of monetary policy for promoting economic stabilization (Woodford, 2011). In mainstream models, a big increase in borrowing by the government will increase the demand for savings, cause interest rates to rise, discourage consumers and private investors from borrowing and offset most of the effects of government spending. Yet despite the 2009 fiscal stimulus, interest rates remain low and monetary policy has not been effective in stimulating private borrowing in the USA. The reason for this is that the hole in demand opened up by the collapse of the housing bubble is so large that nominal interest rates would have to be negative in order stimulate sufficient private spending to fill the output gap. This, of course, is not possible. Indeed, the short-term interest rate that the Federal Reserve targets is currently hovering near the zero lower bound so that it is not possible for the central bank to reduce interest rates further.

In these circumstances, even mainstream macroeconomic models find that fiscal policy is both effective and urgently needed since it is no longer possible to stimulate the economy through interest rate cuts.

[W]hen monetary policy is constrained by the zero lower bound on the short-term nominal interest rate […] it is plausible to assume not merely that the real interest rate does not rise in response to fiscal stimulus, but that the nominal rate does not rise […] Hence, government purchases should have an especially strong effect on aggregate output when the central bank’s policy rate is at the zero lower bound. (Woodford, 2011: 15–16)

While the model finds that a large fiscal stimulus is effective in increasing output ‘when a disturbance to the financial sector results in insufficient aggregate demand even with the central bank’s policy rate at the lower bound of zero’, this policy is to be undertaken only ‘under Depression-like circumstances’ (Woodford, 2011: 23).

The employment response to macroeconomic policy

The drop in GDP in the USA between the onset of the recession at the end of 2007 and the trough reached in 2009, while steep, was not more severe than in most of the other OECD countries that also experienced large, negative demand shocks. At the same time, the macroeconomic policy response in the USA, while it fell short of what was needed to stimulate vigorous job growth, was, according to most accounts, better than in most other wealthy countries (see, for example, ILO, 2009; OECD, 2009). The USA implemented the largest explicit fiscal stimulus package over the period 2008–10 as a share of 2008 GDP among the major OECD countries – almost twice the OECD average – and did so more quickly. It is true that transfer payments including unemployment insurance benefits are more generous in many other countries, making automatic stabilizers a more important component of the fiscal response elsewhere. Even so, it is likely that the fiscal response in the USA was larger and faster than in most other countries. As for monetary policy, the actions of the Federal Reserve Board led interest rates to fall farther and more quickly in the USA than was true of the European Central Bank and interest rates in the EU (ILO, 2009: Figure 2.2). Yet, a comparison of harmonized unemployment rates among wealthy nations finds that the USA experienced the third largest increase in the unemployment rate among these countries between 2007 (the year before the downturn in most economies) and 2009 (the year that GDP reached its trough in most countries), with only Spain and Ireland doing worse (IMF, 2010; OECD, 2010: Figure 1.10). An examination of the ‘jobs gap’ – defined as the increase in the number of jobs required to restore the employment-to-population ratio to its value in the fourth quarter of 2007 and that takes the growth of the working age population into account – shows that the US jobs gap is 5.5 per cent of projected employment in the fourth quarter of 2011 compared to a jobs gap of 3.1 per cent of projected employment in the Euro area and 3.3 per cent in the UK (OECD, 2010: Table 1.1). National labour markets performed very differently in translating declines in GDP into job loss and a rise in unemployment. The rise in unemployment and the decline in the employment-to-population ratio in the USA suggest that labour market institutions in the USA did not perform very well.

A large literature has developed that examines the effect of labour market institutions – wage setting, unemployment insurance, active labour market policies and employment restrictions governing hiring and dismissal – on labour market outcomes, with more protected, presumably more rigid, labour markets generally expected to perform more poorly. The standard view was captured in the OECD’s 1994 Jobs Study. By the mid-2000s, this view was called into question as some countries with apparently rigid labour markets performed well in terms of labour market outcomes (Blanchard, 2006; Layard et al., 2005 [1991]; OECD, 2006: chapter 3). Some studies have found mixed results for strict labour market regulations and only weak effects of employment protections on employment and unemployment, due partly to lower labour force participation where employment protection is high (Nickell, 1997; Nickell et al., 2005), while Howell et al. ( 2007) found no systematic effects at all.

Employment protection regulations – i.e. the rules governing the involuntary separation of workers from employment, both individual and collective – restrict employers’ ability to dismiss workers. They are usually justified on the grounds that they protect workers from arbitrary actions by employers and that dismissal imposes costs on both employees and society. Costs are imposed on workers in terms of job search, deterioration of skills and retraining, while society experiences costs such as unemployment insurance payments and spending on training and job placement programs. Employment protection statutes require companies to internalize some of these costs and raise the cost to firms of laying off or firing workers. The statutes may define the grounds for terminating employment, set out the procedures to be followed when dismissing an individual employee, require that employees be given a specified amount of advance notice, require the involvement of unions or works councils, set minimum amounts of severance pay for workers according to length of job tenure and specify payments to workers found to have been unfairly dismissed. Employers generally face additional requirements in the case of mass layoffs (Venn, 2009). Thus, employers face dismissal costs that may be substantial, depending on the generosity of these provisions and the strictness of the regulations. The emphasis in analyses of the effects of employment protections has been on the effects of higher costs of firing workers on employers’ willingness to hire. Severance payments, for example, may make firms less likely to lay off workers when demand for their products or services declines, but anticipation of this future cost may discourage them from adding workers when demand increases. Where the costs of laying off workers are substantial, however, other institutional arrangements may be developed that offset these costs, e.g. institutions may evolve that support the adjustment of hours of employees and provide employers with greater internal flexibility with respect to hours of work (Addison and Teixera, 2003). In a comprehensive review of the theoretical and empirical literature on the effects of employment protection, these authors note that a surprising finding of the literature on employment dynamics is the apparent ability of some nations with high employment protections to rather quickly adjust labour input to variations in demand. They argue that this is partly due to their enhanced ability to substitute hours for workers, although doing this imposes some cost.

For US employers, the costs of employment protection requirements are minimal. The USA has the weakest employment protection requirements of any OECD country, far below the OECD average. On an index of employment protections that ranges from 0 to 6, where 0 is essentially no statutory employment protections, the USA has a score of 0.2. The UK with a score of 1.1, has the next lowest score while Germany scores a 3.0 (Venn, 2009: 8). The ‘employment-at-will’ doctrine means that US employers can usually terminate a worker’s employment without having to justify or explain their action, except in cases where the dismissal is discriminatory or the worker has a contract that restricts the employer’s actions (Venn, 2009). Moreover, the nature of employer- provided benefits in the USA – especially health insurance – means that cutting hours is less effective in reducing a firm’s total compensation costs than laying off workers. Thus employers in the USA are likely to respond to a demand shock by adjusting employment and laying off workers.

Newer work on the effects of employment protection legislation on labour market outcomes examines the extent of internal flexibility with respect to hours of work in conjunction with this and other labour market institutions. In a study of 25 countries that includes measures of ‘internal numerical flexibility’, in which the enterprise responds to lower demand for goods and services by internally adjusting employee hours rather than reducing staff, Eichhorst et al. (2010) find that variations in labour market institutions generally play a weak role in explaining the response of labour markets to macroeconomic shocks. The exceptions are the USA and Denmark, where employment protections are exceptionally weak and employers are much more likely to immediately lay off workers in response to a demand shock than are employers elsewhere. They find an important role for work-time flexibility in explaining superior employment performance following a macroeconomic shock and a more limited role for traditional labour market institutions. Möller (2010) makes a similar argument in a study of Germany, stressing the importance of measures such as working time accounts and short-time work and contrasting it with employment protection legislation, which he finds to be less important. Short-time unemployment insurance programs were in effect in 17 US states at the start of the recession, and the number has increased since then. But take-up rates have been too small to have an impact. This is largely due to design problems with the programs – in most states participants are not eligible for extended benefits beyond the initial 26 weeks of unemployment compensation, and they are not eligible for emergency benefits if they are unemployed for more than 52 weeks. For employers, the failure of the federal government to subsidize employers’ increased costs for these programs during the crisis, as was done in several European countries, limited the take-up of these programs.

The recent studies suggest that institutions that support adjustments in working time – working time accounts and employer agreements with unions and works councils as well as short-time work arrangements – have reduced the significance of strict employment protections. The ‘flexibility’ cluster identified by Eichhorst et al. (2010: 6) includes Austria, France, Belgium, Germany, the Netherlands and Finland – countries that also tend to have some of the strictest employment protection regulations. Experience with strict employment protection requirements likely provided incentives to employers in some countries to implement enterprise practices that facilitate short-time work, and some governments were willing, in the crisis, to subsidize part of the added costs of these practices.

The older literature on labour market performance viewed labour market institutions as the major determinant of labour market outcomes. If it dealt at all with the relationship of employment and unemployment to macroeconomic shocks, it was to argue that employment protections and other institutions that make the labour market more rigid would limit the ability of macroeconomic policy to improve employment outcomes by restricting job creation and translating efforts to stimulate the economy into inflation (for a critical discussion of this debate, see Baccaro and Rei, 2007; Schmitt and Wadsworth, 2005). The development of new labour market institutions that facilitate employers’ internal adjustments of employee hours in response to a demand shock suggests that stricter employment protections in combination with internal adjustments can help preserve employment in the face of macroeconomic shocks.

Discussion and conclusion

As we have argued, labour market institutions play an important role in determining the extent to which a decline in demand for goods and services is transmitted into a decline in jobs and high unemployment; they cannot restore job creation by firms facing a shortfall in demand. That requires a new, positive demand shock – a large fiscal stimulus. In the US, however – and in much of the rest of the world – the emphasis is on deficit reduction and the hope is for ‘expansionary austerity’. However these policies may play out in countries that run surpluses in their current account balance and are successful in exporting more than they import, they cannot succeed in the USA.

Attempts to revive the labour market and create jobs in the USA via measures to reduce the deficit are bound to fail; the reasons for this are straightforward. The USA runs very large trade and current account deficits. The current account deficit peaked in the third quarter of 2006 at 6.3 per cent of GDP, fell to its lowest point in the second quarter of 2009 at 2.8 per cent of GDP, and recovered by the first quarter of 2011 to 3.2 per cent of GDP (BEA, 2011a). It is likely to remain at or above 3 per cent of GDP for the foreseeable future. This large trade deficit, in which the country as a whole consumes more than it produces, means that either the public sector or the private sector needs to spend more than its income to maintain employment and move towards full employment. Households, however, are currently spending less than their incomes. They are saving and/or paying down debt in order to repair their personal balance sheets. Personal saving out of disposable income, which fell almost to 1 per cent during the housing bubble, stood at 5.1 per cent of disposable income in the first quarter of 2011 (BEA, 2011b), or 4.4 per cent of GDP (BEA, 2011c: Tables 1.7.5 and 2.1). Businesses are also reluctant to spend to expand operations and hiring because of doubts about the recovery. Non-financial companies had $1.84 trillion on hand in cash and other liquid assets as of the end of March 2011. This is an increase of 26 per cent compared to a year earlier and equals about 7 per cent of company assets, the highest level since 1963 (Lahart, 2011). This leaves only the public sector to take up the slack. At the current level of employment, the federal government must run a budget deficit of about 9 per cent of GDP to compensate for high private savings of households and businesses (about 5.5% of GDP) and the large trade and current account deficits (about 3.2% of GDP), and indeed, that is what has occurred. The federal budget deficit in 2010 was 8.92 per cent of GDP (Office of Management and Budget, 2010: Table 1.3). Unless private saving decreases or exports increase dramatically, neither of which seems likely in the short to medium term, actions by the federal government to reduce spending will mean a decline in employment; they will not increase jobs. In the current circumstances it will take an increase in the federal budget deficit – a positive demand shock that can offset the lingering effects of the bursting of the housing bubble – to create jobs.

Nearly four years after the onset of the recession, however, stimulus alone may not be sufficient to absorb all of the unemployed in a timely manner and bring the unemployment rate down to its pre-recession level. While such policies, if they could be enacted, are key to restoring job growth, weak labour market institutions tend to dampen the immediate impact of such policies on employment; the long-term unemployed can be expected to have a particularly difficult time finding jobs. Policies that discourage layoffs and encourage the use of short-time work programs in the public as well as the private sector can play an important role in improving labour market outcomes. The USA could do much more along these lines at minimal cost to the public purse over and above current expenditures on extended unemployment insurance benefits. The challenges lie in the low firing costs and weak unions that provide incentives to employers to dismiss workers rather than reduce hours. Direct job creation by the federal government as employer of last resort may be necessary to assist young people in the transition into employment and to reduce the ranks of the long-term unemployed. In the absence of further fiscal stimulus, policies that encourage the adjustment of employee hours rather than reductions in staff may offer the only path forward at present.

The recession and economic crisis of 2007–9 is sometimes referred to as a ‘black swan’ event – a rare and extraordinary occurrence that could not have been foreseen. Others see this as only the latest – and biggest – contraction and financial crisis in a cycle that has recurred with increasing intensity since the 1970s (Madrick, 2011; Rajan, 2006). Yet this cycle of ever more severe financial crises has its roots in the reliance on debt to fuel economic growth in the USA and in the deregulation of the financial services industry to provide the easy credit this growth model requires. Following the excesses of the dot-com boom of the 1990s, fuelled by a run-up in private debt as the Clinton administration moved the federal budget from deficit to surplus, the dot-com bust threw the US economy into recession in 2001. That recession ended quickly, however, as demand – fuelled by the housing bubble – recovered. More such boom and bust cycles, each more dangerous than the last, can be expected if the USA fails to put in place policies that support the growth of incomes in the middle and bottom of the income distribution so that rising wages, rather than rising debt levels, can again fuel growth (Palley, 2009, 2011). This is not a new insight, especially among development economists. For example, Berg and Sachs (1988) found that unequal societies are more likely to experience serious debt crises. Berg and Ostry (2011) found that inequality is a significant hazard to sustained economic growth. Their results suggest that countries with more equal income distributions experience significantly longer periods of growth and that ‘growth and inequality-reducing policies are likely to reinforce one another’ (2011: 16).