اخراج های بخش عمومی، حق سنوات، و تورم در اقتصاد باز کوچک

Because severance pay is worth 2–5 years of wages in many LDCs, public sector layoffs increase the fiscal deficit in the short run. Nevertheless, generous severance pay is not as serious a macroeconomic problem as generally thought. In the case where the fiscal deficit is financed by printing money, inflation is continuously lower under plausible conditions. When the government can borrow in world capital markets and layoffs reduce the present-value wage bill, there exists a sequence of bond sales and subsequent redemptions that guarantees continuously lower inflation. This result does not hold, however, if the reform lacks credibility.

Public sector overstaffing in developing countries has lowered productivity, increased the fiscal deficit and inflation, worsened the distribution of income, and contributed to balance of payments problems. Consequently, structural adjustment programs often call for deep employment cuts in the civil service and at state-owned enterprises. This is especially true of adjustment programs in Sub-Saharan Africa. Ghana, Sierre Leone, Uganda, Kenya, Tanzania, and Zambia reduced civil service employment by 20–50% in the nineties (Harvey, 1996, Sahn et al., 1997 and African Development Bank, 2005); Guinea and Gambia went even further, eliminating one-fourth of all public sector jobs ( Jabara, 1994 and Mills and Sahn, 1996). The long-run benefits of lower inflation, higher productivity, and a better distribution of income add up to a powerful case for reform. There is a catch, however: under existing contracts, retrenched workers in many LDCs are entitled to severance pay worth 2–5 years of wages (see Table 1). Because severance pay is so generous, the wage bill increases sharply before it starts to decline. Policy makers fear therefore that employment cuts will lead to higher fiscal deficits and higher inflation in the short run. This has slowed the pace of reform to a crawl in many countries. Surveying the scene in the Middle East and North Africa, Page (2001) observes that “financial incentives for voluntary redundancy such as severance pay… have not been widely used due to fiscal constraints” (p. 72); hence governments are unlikely to “undertake more active labor redundancy policies [until] increased tax revenues from rising income will permit greater fiscal space” (p. 73). In a similar vein, Helbling (1999) argues that civil service reform may threaten macroeconomic stability if its short-term costs are not paid for by a temporary tax surcharge. Recently the World Bank amended its operational rules to allow lending for severance pay. The new rules were adopted for the express purpose of loosening the fiscal constraint in public sector restructuring programs (Rama, 1999). Nevertheless, fiscal concerns still inhibit reform (African Development Bank, 2005) and “cost-benefit” assessments in the literature still tend to be highly ambiguous, e.g.: “The voluntary approach to reductions—especially paying workers to leave—… is now broadly supported by IFIs (International Financial Institutions). But the costs of needed severance pay and other safety-net enhancements—early retirements, counseling, retraining—may be greater than the financial and economic benefits” (Cook and Murphy, 2002, pp. 7–8). Table 1. Ratio of severance pay to the annual wage in various LDCs. Senegal, 1990–2000 3–5 Ghana, 1986–1989 4.3 1989–1995 >4.3a Kenya, 1990s 3.3 Mali, 1990s 4.0 Zambia, 2000–2007 10 Sierra Leone, 1990s 2.0 Egypt, 1996–1997 3.0 Guinea-Bissau, 1990s 9.6 Africa, 1990s 5.2b Ecuador, 1990s 2.3 Bangladesh, 1990s 3.1 India, 1993–1994 14.3c Argentina, 1990–1994 2.0d Peru, 1990s 2.4 Average in the 1990s for 37 LDCse 2.05/3.0 Sources: Tait Davis, 1991, Rouis, 1994, Kikeri, 1997, Rama, 1999, Rama and MacIsaac, 1999, Gupta et al., 1999, Haltiwanger and Singh, 1999 and Thurlow and Wobst, 2004, and African Development Bank (2005). a 52 months of wages plus early payment of retirement benefits. b Ratio is inflated by inclusion of programs that used part of the fiscal saving from layoffs to increase real public sector wages. c For state-owned textile firms. d For railways, telecommunications, and steel companies. e 2.05 is the figure implied by Haltiwanger and Singh's (1999) calculation that, at a discount rate of 10%, it takes 2.27 years on average before the financial saving from lowering the wage bill equals total outlays on retrenchment. But Haltiwanger and Singh's figure (i) includes some programs that paid little or nothing to retrenched workers and (ii) excludes programs that generated a present-value loss (i.e., programs that did not have a finite breakeven period). The higher value 3 reflects an arbitrary guess that adjustment for these factors would increase the average payback period by 50%. Table options I argue in this paper that generous severance pay is not as great a macroeconomic problem as generally thought provided layoffs are credible (i.e., expected to be permanent). Severance pay three times the annual wage is certainly a severe fiscal shock. But the shock is strictly temporary. Forward-looking agents realize that the fiscal deficit and inflation will be lower in the future. Due to the deflationary pull of the fundamentals, the outcome is uncertain even when severance pay is financed entirely by printing money. Much depends on the trajectory of the wage bill and the sensitivity of money demand to its future return, but there is no general presumption that inflation increases in the short run. I demonstrate that inflation is continuously lower when the interest-elasticity of money demand is slightly above .50 and the ratio of severance pay to the annual wage is less than the reciprocal of the nominal interest rate, a number on the order of 2–6 in most LDCs. And if the interest-elasticity is significantly higher, severance pay can be obscenely generous without endangering price stability. In the case where the interest-elasticity equals unity, for example, layoffs are unambiguously deflationary even when the ratio of severance pay to the annual wage equals the reciprocal of the real interest rate. It is natural to question the relevance of these results. After all, nominal money growth increases very rapidly when severance pay is 2+ years worth of wages; for inflation to decrease right away, nominal money demand has to grow even faster. This might seem far fetched, but at least one real world episode supports the analytical results. In 1989 monetary financing of the fiscal deficit soared to 6% of GDP in Vietnam to cover the costs of massive public sector layoffs. 1 Surprisingly, the sharp increase in money growth was not inflationary. Thanks to a huge drop in the velocity of money, inflation abruptly decreased from 300% to less than 100% ( Dodsworth et al., 1996). The analytical results and the experience of Vietnam are encouraging but not decisive; given uncertainty about the true value of the interest-elasticity of money demand, they cannot fully allay the fear that large severance payments might prove highly inflationary in the short run. The situation is quite different, however, for governments that can borrow in the world capital market or from the World Bank. Access to the external securities market allows the government to effectively amortize the cost of severance pay by borrowing against future fiscal adjustment. When the reform is fiscally sound and credible, this strategy—issuing bonds to cover the cost of severance pay in the short term and retiring the extra debt in later periods as the fiscal deficit declines—smooths the path of nominal money growth enough to ensure continuously lower inflation. The qualifier “fiscally sound” means only that layoffs reduce the present-value wage bill, or, in the language used earlier, that the ratio of severance pay to the annual wage is smaller than the reciprocal of the real interest rate. No restrictions are needed on the interest-elasticity of money demand. The efficacy of temporary bond sales is an important albeit obvious application of Barro's (1979) principle that debt should be used to smooth the path of distortionary taxes (here the inflation tax). Unfortunately, the result is not completely general. It requires the reform to be fully credible in the eyes of the public. This qualification matters in practice. Haltiwanger and Singh (1999) report that significant rehiring occurs in 20% of public sector retrenchment programs. Rehiring is especially common in voluntary separation programs for civil servants. Typically adverse selection results in the departure of the most talented professionals who have the best opportunities in the private sector. After efficiency declines and the flow of public services slows to a trickle, ex-civil servants get recalled to their old jobs in large numbers (African Development Bank, 2005). Rama (1999) refers to this as the “revolving door syndrome.” Lack of credibility is very damaging. If jobs cut during the austerity phase are restored at a later date, then the wage bill jumps back to its previous level while the government is still saddled with debt issued to finance severance pay. The debt overhang implies faster money growth and higher inflation in the period that immediately follows the policy reversal. Private sector anticipation of this scenario creates inflationary pressure regardless of whether the scenario actually materializes. Numerical simulations suggest that bond sales can secure lower inflation at the start of the reform as long as severance payments are not too extreme and the credibility problem not too severe (i.e., employment cuts are expected to last at least three years). But if the government borrows on commercial terms there is almost no hope of avoiding higher inflation as the date of the expected policy reversal draws near. For average levels of severance pay and a high rehiring rate, the debt-financing scheme works only when the World Bank is willing to lend long-term at a very low real interest rate (0–4%). The rest of the paper is organized into six sections. Section 2 lays out the model and Section 3 derives conditions for inflation to decrease monotonically when the fiscal deficit is financed every period by printing money. Section 3 also presents numerical results to make the point that the price dynamics are quantitatively sensitive to the interest-elasticity of money demand; more specifically, that generous severance pay is highly inflationary when the interest-elasticity lies in the 0–.40 range. Following this, I introduce debt-financing schemes in Section 4 and prove that layoffs combined with the right sequence of open market operations guarantee lower inflation over all time horizons. Section 5 generalizes the results to two alternative scenarios and Section 6 examines their sensitivity to weak credibility. The final section contains concluding remarks.

I have shown that credible public sector layoffs often reduce inflation immediately despite the handicap of high severance payments to retrenched workers. Does this mean that, distributional concerns aside, generous severance pay is not a serious problem? Not necessarily. Although the results are generally positive, the degree of macroeconomic risk varies from negligible in some cases to fairly high in others. The prospects for success are least favorable when the inflation rate is initially high and institutional constraints or a poor credit rating forces the government to finance the fiscal deficit every period by printing money. In these circumstances, the hope of avoiding a temporary sharp increase in inflation rests with the hope that the interest-elasticity of money demand is sufficiently large. For the levels of severance pay common in LDCs (2–4 years of wages), the interest-elasticity has to be .5–.8 or larger when the pre-reform inflation rate is 25–50%. The critical range drops to .20–.50, however, if inflation is initially 10%. Absent access to the debt market, policy makers have no choice but to let the path of the fiscal deficit dictate the path of nominal money growth. By contrast, a creditworthy government operates under the much looser constraint that the path of nominal money growth be consistent with long-term fiscal adjustment; it can therefore reconcile high severance payments with continuous smooth reductions in nominal money growth by issuing bonds in the short run and redeeming the debt later after the wage bill declines. This strategy succeeds nicely: subject to the trivial qualification that reform lowers the present-value fiscal deficit, it delivers unambiguously lower inflation at every point on the transition path. The result is independent of the sensitivity of the risk premium to the level of debt because variations in private borrowing exactly offset variations in public sector borrowing on the path to the new steady state. Weak credibility is the most serious threat to macroeconomic stability. Credible reforms benefit from the perception that inflation will be permanently lower in the future. This is lost when downsizing is expected to be temporary. Consequently, if the reform is 100% non-credible (i.e., 100% rehiring is expected) inflation increases in the short run regardless of how the fiscal deficit is financed. Generous assistance from the World Bank is the only antidote to private sector pessimism. If the Bank supports the reform by lending at a low real interest rate, then inflation can be reduced even during the period of low credibility. The subsidized loan enables the government to keep inflation at bay until the economy locks onto the right transition path. It is arguably the difference between success and failure in countries where policy makers can handle the political fallout from layoffs but not the fallout from both layoffs and higher inflation.