WPS4685 Policy ReseaRch WoRking PaPeR 4685 Kenya's Quest for Growth Stabilization and Reforms--But Political Stability? Luca Bandiera Praveen Kumar Brian Pinto The World Bank Africa Region & Poverty Reduction and Economic Management Network August 2008 Policy ReseaRch WoRking PaPeR 4685 Abstract Kenya has long had a reputation of being politically risky, important development. The paper draws on an analysis manifested in corruption, uncertainty about policies, of debt dynamics, the evolution of domestic interest and the importance of political connections in doing rates, and the latest Investment Climate Assessment to business. Kenya began its economic liberalization in present evidence on the criticality of low political risk 1993. Reform picked up speed after a tightening of aid in facilitating good economic outcomes after 2003. The by donors on governance grounds and an attempt to December 2007 elections have highlighted other aspects re-establish credibility following the costly Goldenberg of political risk--ethnic and social tensions with roots scandal uncovered in 1992. But tangible results in the in inequality. The findings of this paper underline the shape of favorable government debt dynamics and a importance of establishing a foundation for long-term pick up in growth took a decade to materialize. The political stability and social cohesion in view of the paper argues that the peaceful presidential election and disruptions following the December 2007 elections. This transfer of power in December 2002 was central to the process is likely to be at least as difficult and lengthy as economic upswing after 2002. The subsequent decline in fundamental economic policy and institutional reform. political risk was singled out by the private sector as an This paper--a joint product of the Poverty Reduction and Economic Management Department, Africa Region, and the Economic Policy and Debt Department, Poverty Reduction and Economic Management Network--is a background paper for the Country Economic Memorandum on economic growth in Kenya. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at lbandiera@worldbank.org, bpinto2@ worldbank.org and pkumar1@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Kenya's Quest for Growth Stabilization and Reforms--But Political Stability?1 April 28 2008 Luca Bandiera, Praveen Kumar and Brian Pinto2 . 1This title is a take-off on "Mexico: Stabilization, Reform and No Growth" (Dornbusch and Werner 1994). With growth picking up in 2003, Kenya appeared to be one step ahead. But it will need to make a fresh start to overcome the disruption and distrust engendered by the December 2007 Presidential elections. 2The authors are all at the World Bank. This paper is based on background analysis for a Country Economic Memorandum being prepared on Kenya. We thank Viktoria Hnatkovska, Mohan Krishnaswami, Claudio Raddatz and Terry Ryan for helpful discussion and comments. The findings, interpretations and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. 1. Introduction For decades, Kenya has suffered from low and volatile economic growth--per capita real GDP growth averaged 1.2 percent over the period 1960-2000, overwhelmed by a standard deviation of 4.8 percent.3 More recently, notwithstanding deeply-entrenched skepticism, Kenya has done well on growth and macroeconomic stabilization. Over the four years 2003-2006, growth averaged 4.9 percent in contrast to virtual stagnation between 1991 and 2002.4 Inflation remained relatively contained at an average 9 percent, but more convincingly, the ratio of government debt-to-GDP fell by 5.8 percentage points per year over this period as interest rates fell and growth picked up. These impressive developments fueled optimism that Kenya might at last be turning the corner and embarking on a path of sustained growth and poverty reduction. However, the violent and disruptive aftermath of the Presidential elections of December 2007 have placed such optimism on hold. 1.1 Objectives and Background Kenya's reputation has been overshadowed to such an extent by its image of being corrupt and weakly governed that it might come as a surprise that serious reform began as early as 1993. We highlight the crucial role played by declining political risk after 2002 in at last bringing the reforms implemented by Kenya over the previous decade to fruition. Our analysis was completed before the December 2007 elections. It began in mid-2006, at which point our priors were formed on the basis of available reports. Typical were the following quotes from the May 2006 Economist Intelligence Unit report:5 The president....will struggle to see out the remainder of his term because of the damage to his authority and credibility caused by the "no" vote in the constitutional referendum and corruption scandals that have led to the resignation of three senior ministers. The Government of National Unity continues to confront serious divisions.... 3Hnatkovska and Loayza (2005), Appendix A pp. 95-96. 4The 2007 growth estimate is 6.8 percent, a little short of 5 percent per capita. For 8 out of 12 years from 1991 to 2002, per capita income growth was almost zero or negative. 5All three quotes from "Kenya at a glance: 2006-07" of Country Report Kenya May 2006, EIU. 2 Only a few months later, in October 2006, a substantially better assessment appeared in a Standard & Poor's report:6 Major Rating Factors Strengths: Macroeconomic and political stability becoming entrenched Improving economic growth prospects Weaknesses: Low level of economic development with severely limited infrastructure and vulnerability to exogenous shocks High debt compared with similarly rated peers Governance issues The EIU quotes convey the image of corruption and political instability, a combination normally associated with stagnation, bad public finances and a poor investment climate. The S&P quotes contain good news on stability and growth but caution that infrastructure is deficient and governance a weakness. We will not try to explain this disconnection in perceptions separated by only a few months but will instead present evidence on a perceptible change for the better in Kenya's macroeconomic outcomes starting in 2003/04. The first step in this study was to attempt an explanation for the puzzling observation that real interest rates on Kenyan treasury bills were surprisingly low, of the order of 1-3 percent. How could a corrupt and politically unstable country (recall the EIU quotes) have such low interest rates? The possibility of manipulation was rejected because Kenya has a fully convertible currency and an open capital account. If interest rates were being kept artificially low, there would be downward pressure on both the exchange rate and foreign exchange reserves; but the opposite was true. When we asked officials or the private sector why interest rates were so low, the standard reply was "Because government's borrowing requirements have come down." However, a review of the government's debt dynamics strongly suggests that borrowing requirements came down because interest rates fell substantially, not the other way round. 6S&P Credit Research Report Kenya (Republic of) Oct. 23, 2006. A later report in February 2007, by JPMorgan titled `Kenya: A return to Macroeconomic Stability' also notes that `the turning point for Kenya was the 2002 election of President Mwai Kibaki who was elected on an economic reform and anti- corruption platform.' 3 Indeed, with the primary fiscal balance having shrunk and concerns about the composition of public spending, the only factor which could explain the big fall in interest rates was a decline in country or political risk.7 Informal interviews with businessmen supported this idea. The encouraging refrain was: "We no longer need political connections to do business in Kenya." In addition, it was noted that investments and business would continue regardless of the outcome of the then-upcoming Presidential elections. In order to obtain systematic evidence on perceptions of politics and business, questions were added to the Investment Climate Assessment (ICA) conducted in June 2007. We report these results in section 5. 1.2 Main Findings and Organization of the Paper Serious data deficiencies--in particular, the availability of a consistent GDP series only after 1995/96--are an impediment to an analysis requiring long time series. Our strategy therefore is to persuade the reader by "connecting the dots" among the extent of past reform; a review of the evolution of government indebtedness (measured by the debt-to-GDP ratio) and its underlying determinants; a decomposition of interest rates to gauge devaluation and default risk; and the survey results from the ICA. We conclude that the positive trends after 2003 were being driven by three factors: · lagged benefits of price, trade, exchange rate and interest rate liberalization forced by reduced aid after the Goldenberg scandal uncovered in 1992; · solid foundation for solvency based on significant revenue collection as a payoff to the reform of tax policy and administration which started in the mid-1990s; and critically, · declining political risk after the successful 2002 elections fueling an improvement in sovereign creditworthiness and the private investment climate. 7To paraphrase Sherlock Holmes: When you have eliminated the impossible, whatever remains, however improbable, must be the truth. 4 Two messages are embedded in the above: first, the back-to-the-wall effect of a drastic reduction in aid starting in 1992 stimulated substantial reform. Second, it took the successful December 2002 elections and smooth transfer of power from President Arap Moi to President Kibaki for visible results to materialize. This suggests two interacting effects: long lags before payoffs to reform appear; and threshold events (such as a successful election) which trigger these payoffs. Not surprisingly, we concluded (prior to the December 2007 elections) that the maintenance and acceleration of these positive trends depended crucially on continued success in entrenching political stability, improving governance and strengthening fiscal and financial institutions; and relaxing the infrastructure constraint on private investment. The unfortunate December 2007 election fiasco has reinforced the key insight of this paper: that sustained growth and poverty alleviation will eventually depend upon Kenya's ability to engender political stability and social harmony for a prolonged period; experience from countries as disparate as China, India and Vietnam suggests that such stability is needed for a minimum 10-15 years for positive trends to take hold.8 The reasoning is that irrespective of the political system, good governance and stability in the rules of the game for a long period are needed for having well-managed public finances, investor confidence and hence long-run growth. Section 2 sketches out an analytical framework which underpins the subsequent presentation of empirical evidence. Section 3 contains a snapshot of Kenya's economic liberalization dating to 1993. Section 4 analyses government debt dynamics, discussing key facets such as revenue mobilization and expenditure composition. While impressive strides have been made towards an efficient revenue system based on broad-based taxes, the composition of expenditure remains a serious constraint on growth. Specifically, spending on infrastructure is inadequate. Section 5 presents evidence on declining country and political risk based on a 8Rodrik (1999) contains a seminal analysis of the importance of social cohesion in responding efficiently to exogenous shocks and nurturing long-run growth. 5 decomposition of interest rates and perceptions from Investment Climate Assessments. Section 6 discusses challenges--economic and political--looking ahead. 2. Analytical Framework A reduction in political risk could have a growth dividend through two channels: at the macro level, a reduction in real interest rates which improves government debt dynamics and contributes to stabilization; and at the micro level, a lengthening of business horizons with its corollaries of lower hurdle rates of return for investment projects and greater private investment. We start with the micro channel using a slight adaptation of the model in Razin and Sadka (2004). The representative competitive firm invests in line with the following Bellman equation: V[(1- )K0] = Max{1 1- t (1- ) V[(1- )K]} , K + r + F(K) -[K - (1- )K0]+ 1 + r + where F(.) is the production function, V(.) a value function, (1- )K0 the initial net capital stock ( is the rate of depreciation, K0 the inherited capital stock), K the chosen capital stock for the current period, t the current tax rate, r the world interest rate, a political (country) risk premium external to the firm and the future tax rate. The parameter captures the investment horizon or equivalently the importance given to the future. We assume 0 1 with a higher value connoting a longer horizon. The first order condition is:9 (1) (1- t)F(K) = r + + + (1- )(1- + ). We now mimic a short-horizon situation by setting = 0. Equation (1) gets modified as follows: (2) (1- t)F(K) = r +1+ . Note that the RHS of (2) is the maximum value of the RHS of (1) for a given world interest rate and country risk premium. In other words, when the horizon is short the hurdle rate for private investment is set at its maximum so that the investment can be recouped in short order. This is likely to coincide with a high country risk premium, . Now consider the polar opposite case of 9For a derivation, see Razin and Sadka (2004)--equation (1) is the same as their equation (2) with = = 0. Butnotethattheinterpretationof isdifferent. 6 low political risk and a long horizon, which we mimic by setting =0 and = 1. Equation (1) becomes: (3) (1- t)F(K) = r + (1- ) + . Comparing the RHS of equation (3) with that of (2), we see a substantial decline in the hurdle rate of return for investment.10 Figure 1 summarizes these possibilities. Figure 1 Length of Horizon and Optimal Investment A (1-t) F (K) Short horizon r + 1 + Long Horizon r + (1-) + Long horizon + Tax cuts B K1 K2 K3 K The figure shows three levels of the optimal capital stock decisions, K1 K1 and stagnation would result. On the other hand, investment would resume if political risk falls and 10Note that (1- ) + << 1 for reasonable values of and . For example for a 10 percent depreciation rate and a marginal tax rate of 30 percent, (1- ) + = 0.37 << 1. In addition, we have the country risk premium on the RHS of (2). 7 horizons lengthen. Applied to Kenya, this suggests that the initial binding constraint to overcome would be political risk; once this is achieved and the investment climate improves, attention might have to turn to other constraints such as infrastructure. Turning to the macro, few economists would quarrel with the proposition that macroeconomic stability is necessary for sustained economic growth and a good investment climate. However, the focus has shifted (especially after the most recent emerging market public debt crises which began in 1997-98 and ended with the Argentine default of 2001) from short-run fiscal deficits and inflation to balance in the government's intertemporal budget constraint.11 Satisfying this constraint rests on the capacity to eventually generate adequate primary fiscal surpluses, which in turn depends upon future growth and tax collection.12 Two key points emerging from recent empirical work are that balancing the government's intertemporal budget calls for (i) a comprehensive approach to managing the public finances. Paying attention to primary deficits, real interest rates and growth rates is not enough--equally important are bailouts and contingent liabilities, the efficiency and predictability of taxation, expenditure composition and the rate of return on public investments; and (ii) sound macro-micro linkages.13 We briefly sketch the latter and make the connection with political risk. Define government net worth as the present value of future primary surpluses (expressed as a ratio of GDP) minus the initial government debt-to-GDP ratio.14 Since the discount rate is the real interest rate minus the real growth rate (assumed to be positive for convergence), a reduction in interest rates as a result of reduced political risk will unambiguously increase net worth and strengthen government solvency. What about an accompanying reduction in tax rates? This will have two opposing effects: it will tend to lower revenues and hence the present value of primary 11The framework which dominated IFI policy thinking on macroeconomics and growth until 1997 might well have been Fischer (1993). 12Solvency requires that the present value of future primary surpluses (discounted at a rate equal to the real interest rate minus the real growth rate) equal or exceed the initial debt-to-GDP ratio. 13See IMF (2003), Budina and Fiess (2005), Gill and Pinto (2005). 14See Serven (2007). 8 surpluses; but it will tend to raise private investment and hence growth especially in conjunction with a fall in political risk (increasing the capital stock to K3 in Figure 1). The combination of rising growth and lower interest rates will in turn exert an upward effect on net worth which could in principle offset the effect of a fall in revenues. Besides, the fall in revenue could be curtailed by improving tax administration which increases the tax base and thus partially or completely offsets the impact of lower marginal tax rates.15 Now consider government capital expenditure. If it is cut, the immediate effect will be to raise primary surpluses and improve debt dynamics; but net worth could decrease as a result of a fall in long-run growth (via complementary reductions in private investment). This in turn could trigger a rise in interest rates if initial indebtedness is high and tax revenues fall with growth. Of course, if capital expenditure is of low quality, a cut would be unambiguously beneficial. In Kenya's case as we shall see, the reduction in interest rates and cut in marginal tax rates have contributed to an accelerated decline in indebtedness (government debt-to-GDP ratio) after 2003; but growth diagnostics indicate that more public spending is needed to alleviate the infrastructure constraint on private investment. The challenge therefore is to choose infrastructure projects with high rates of return which will ensure that the government's net worth is protected.16 However, in line with the main point of this paper, this is secondary to managing political risk. 3. Kenya's Economic Liberalization: A Snapshot 17 The economic liberalization started with Sessional Paper Number 1 of January 1986, when the government decided to shift from dirigisme to freeing up the economy. By the late 1980s, Kenya had adopted a managed float. Then a series of incidents intervened after the Berlin Wall collapsed in November 1989 and Kenya began to lose its geopolitical bargaining power (as 15Many developing countries offer the prospect of raising average effective tax rates (and hence total revenues) even when marginal tax rates are cut by enlarging the tax base through better administration. 16A detailed analytical discussion is contained in Serven (2007). 17Sources: Meetings with academicians, private sector firms, IMF files. 9 it was the most prominent market economy in the region and the US fleet was able to use Mombasa port). Here is a brief timeline: · Nov 1991: Emphasis in Consultative Group shifts abruptly from economics to governance. Donors freeze aid. · Jan 1992: IMF refuses to go to its Board with an unfinanced gap. The World Bank pulls back loans in response. · 1992: Kenya starts building up external debt arrears; foreign exchange becomes the critical constraint. · End 1994: IMF shadow program enables Paris Club rescheduling. Prior to the 1994 IMF program, turbulence hit in the shape of the notorious Goldenberg scandal (1991-93), interest rates on treasury bills went to 80 percent and banks allied to Goldenberg's Exchange Bank collapsed. This scandal arose out of a fraudulent effort to exploit schemes set up by the central bank to encourage exports and help alleviate the shortage of foreign exchange that occurred as external debt payments fell due and new loans and aid were cut back.18 Liberalization continued and by the end of 1994, prices, interest rates and the exchange rate were liberalized. In the meanwhile, aid became small and unpredictable and since 1994/95, net foreign financing to the government of Kenya has been negative except for 2001/02 and 2005/06, when there was a debt rescheduling. As a result, domestic borrowing served not only to finance the deficit but also pay off maturing external debt. Two points are worth noting. First, the hardening of the donors' stance on governance grounds and the need to re-establish credibility in the wake of Goldenberg were major spurs to 18 Goldenberg International was set up in 1990 to export gold and diamond jewelry from Kenya in return for a 35 percent premium on the exchange rate in contrast to the normal premium of 20 percent for exporters. Little or no gold was actually exported; a scheme was established to round-trip the sale of export dollars to the Central Bank of Kenya (essentially, buy dollars at the official exchange rate, sell them back at a 35 percent premium as "gold exports" and then repeat the process). It was finally uncovered in 1992. Warutere (2005). 10 reform in Kenya. Second, private sector firms interviewed in mid-2006 invariably pointed to the liberalization which picked up steam in 1993/94 as a positive turning point. 4. Government Debt Dynamics A key requirement for macro stabilization noted in section 2 is balance in the government's intertemporal budget constraint, for which a sufficient condition is that public debt be on a sustainable trajectory, i.e., there is no need for a drastic change in fiscal policies to rein in indebtedness. If intertemporal balance is not assured, there will be a tendency for macroeconomic uncertainty to increase, engendering expectations of rising inflation and interest rates, which will slow down private investment and growth. Governments borrow when they run fiscal deficits, which can be decomposed into the primary fiscal deficit and interest payments. The path of the debt-to-GDP ratio is then determined by the primary deficit and a term proportional to the difference between the real interest rate and the real growth rate. If the primary deficit is positive and large, and real interest rates exceed growth rates, then the debt-to- GDP ratio will grow until corrective action is taken or a crisis--typically involving a burst of inflation and/or a disruptive debt default--forces such action. Debt can also increase when the government issues recapitalization bonds to bail out banks or takes over the guaranteed loans of loss-making parastatals, while the opposite happens when government assets are privatized. Box 1 summarizes difficulties encountered in assembling the debt data as well as data on macro variables more generally, while Figure 2 shows the path of Kenya's debt-to-GDP ratio and its currency composition over the past 11 years. The government debt-to-GDP ratio has fallen by some 27 percentage points over the past 11 years to a level of 46 percent by the end of 2006/07, while the share of foreign-currency denominated debt has gone down from about 80 percent of total debt in 1996/97 to 50 percent at the end of 2006/07. This shift towards domestic debt was occasioned by the need to pay off maturing external debt and compensate for the reduction in external aid after 1993; but at the same time, there has been an impressive reduction in indebtedness as shown in Figure 2. 11 Box 1: Debt and Macro Data In Figure 2, domestic debt is net of government deposits in the banking system and on-lending. External debt is on a gross basis and includes borrowings from the IMF. In compiling statistics on external debt, data from 3 different sources were compared: World Economic Outlook and staff reports (IMF); Global Development Finance (World Bank); and the statistical bulletins of the Ministry of Finance (MoF) and the Central Bank (CBK). All institutions report data on gross external debt in dollars, by creditors or groups of creditors (See table). The difference reached as high as $1.2 billion in 2003/04. External Debt Comparison (In millions of US dollars) 1996/97 1997/98 1998/99 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 IMF 5,666 5,755 5,454 5,286 4,713 4,907 5,088 5,119 5,068 5,000 5,370 World Bank 1/ 6,814 6,465 6,824 6,475 6,145 5,521 6,128 6,869 6,919 6,428 6,534 Kenya 5,915 5,366 5,476 5,426 4,988 4,795 5,488 5,573 5,695 5,837 5,955 Difference max-min 572 389 22 253 275 538 861 1197 627 837 585 Source: IMF WEO and Staff Reports; GDF; and Statistical bulletins of the Ministry of Finance of Kenya 1/ Calendar year. It includes use of IMF credit Eventually, the choice fell on the data published by the MoF as better documented, consistent with monthly data published by CBK and consistent with past external borrowing requirements from the fiscal accounts. Also, the collection of macro aggregates proved difficult. In 2005, the Central Bureau of Statistics updated the methodology for estimating GDP so as to better capture the contribution of services in the economy. The published series now goes back only to 1996, preventing a longer time series on the debt-to-GDP ratio. In addition, data on key macroeconomic variables such as the fiscal accounts, balance of payments, interest rates, prices and exchange rates are difficult to find from the same source on a consistent basis (i.e. IMF staff reports or GDF), for years before 1996. Figure 2 Kenya: Public Debt (In percent of GDP) 80 Domestic debt 70 External debt 60 50 40 30 20 10 0 1995/96 1996/97 1997/98 1998/99 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 12 4.1 What Has Lowered the Debt Ratio? In Table 1, the annual average change in indebtedness is algebraically apportioned to the primary fiscal balance (including grants, which averaged approximately 1 percent of GDP per year over both the sub-periods in the table), growth, real interest and exchange rates and other factors like bank bailouts, debt relief, etc. The starting point is the debt-to-GDP ratio at the end of the fiscal year 1995/96. The subsequent 11 years are divided into two sub-periods: 1996/97- 2002/03 and 2003/04-2006/07. For each sub-period, the average annual change in the debt-to- GDP ratio is given, as well as the portion attributable to the factors mentioned above. Table 1. Factors Explaining Falling Indebtedness 1996/97-2006/07 (% points of GDP, annual average) 1996/97 2003/04 1996/97 -2002/3 -2006/07 -2006/07 Change in public sector debt -1.8 -5.8 -3.2 Contribution from 1. Primary Deficit (- surplus) -1.7 -0.1 -1.1 2. Real GDP growth -1.4 -2.8 -1.9 3. Real interest rate 2.1 0.3 1.5 4. Real exchange rate (- appreciation) 0.6 -2.5 -0.5 5. Other Factors -1.4 -0.7 -1.1 The first compelling observation is that the rate at which indebtedness has been falling jumped from 1.8 percentage points of GDP per year over the first 7 years to 5.8 percentage points per year over the last four years. The biggest factor explaining this sharp acceleration is the joint effect of the large decline in real interest rates and the real appreciation of the Kenyan shilling after 2003.19 During the first 7 years, as a result of the combined effect of the real interest rate and real exchange rate (lines 3. and 4. in the table), debt rose by 2.7 percentage points of GDP per year, while the combined effect has been to lower debt by 2.2 percentage points of GDP per year after 2003, a swing of close to 5 percentage points of GDP. The acceleration in growth rates has also helped significantly in the reduction of debt ratio. Second, the accelerated decline in 19Strictly speaking, one has to look at the effective real interest rate, which would capture the effects of both inflation and exchange rates and related capital gains/losses, by summing up rows 3. and 4. in Table 1. 13 indebtedness occurred even though the average primary fiscal surplus (including grants) fell by 1.6 percentage points of GDP; we will return to this point later when political risk is discussed. Third, big collapses of the real exchange rate (RER) and bank/other bailouts, which were salient features of most of the emerging market debt crises of the past decade, have not been major factors in Kenya.20 4.2 Primary Fiscal Balance and Interest Payments Figure 3 plots the primary balance and interest payments over the period 1995/96- 2006/07. Two observations stand out: (i) the primary fiscal surplus including grants went up in the wake of the Goldenberg scandal and reached close to 6 percent of GDP by the mid-1990s but has been on a declining trend ever since, i.e., the primary deficit has been rising; and (ii) interest payments have fallen from 6 percent of GDP in 1995/96 to 2.8 percent in 2006/07.21 Figure 3 Primary Balance and Interest Payment (% of GDP) 7 Primary balance Interest payment 6 5 4 3 2 1 0 96/ 97/ 98/ 99/ 00/ 01/ 02/ 03/ 04/ 05/ 06/ 07/ 95 96 97 98 99 00 01 02 03 04 05 06 19 19 19 19 19 20 20 20 20 20 20 20 The fall in interest payments (and interest rates, as we shall see later) even as the primary deficit was rising (primary surplus shrinking) is puzzling. A plausible explanation is that the 20See the country case studies in Budina and Fiess (2005). It also contains a derivation of the flow budget constraint of the government forming the basis for the decomposition shown in Table 1. 21Unfortunately, we do not have a consistent GDP series for pre-1995. Earlier numbers showed government interest payments to be as high as 10-12 percent of GDP in the aftermath of Goldenberg. 14 1991-93 Goldenberg scandal engendered a great deal of macroeconomic volatility and left little option but to raise the primary surplus sharply; and both the primary surplus and interest payments then gradually returned to `normal' levels. The Goldenberg scandal is estimated to have cost $600 million to $1 billion, some 9 to 16 percent of 1994 GDP (Warutere 2005). Efforts to mop-up excessive monetary infusion arising from the scandal pushed interest rates on T-bills to over 80 percent (Figure 8 below). Kenya was also forced to shift to domestic debt to pay off maturing external debt as a result of the accompanying aid squeeze. A large increase in the primary fiscal surplus was engineered in an attempt to bolster credibility and a significant liberalization was set in train. Import licenses were abolished in May 1993, a unified, market- based exchange rate adopted in November 1993 and exchange controls lifted in May 1994. The era of price controls ended in 1994. As a result of this freeing up, foreign exchange flowed in, helping drive interest rates down.22 Thus, positive macroeconomic benefits started appearing before 2002 but growth stagnated. A second possible explanation for falling interest rates in spite of the fall in the primary surplus is that the latter was driven by adjustments in fiscal policy designed to support private investment and long-run growth. This could happen for example if marginal tax rates were cut and/or public spending to alleviate infrastructure bottlenecks were raised. In the short run, this would reduce the primary surplus but with the promise of better prospects for long-run growth and tax collection, as discussed in section 2. Countering this is the argument that markets tend to be myopic and that the quality of Kenya's past public investments has been low--which would strengthen the argument that interest rates fell because of reduced political risk rather than market perceptions of higher government net worth. Revenue mobilization is now discussed, followed by expenditure composition. 22Source: Interview with Kenyan academics. 15 4.3 Revenue Mobilization We shall show that Kenya has managed a successful transition from a high tax rate, low tax base system with considerable reliance on import taxes and seigniorage to a more efficient and equitable revenue system based on broad-based taxes like income tax and VAT. This accomplishment is considered difficult to achieve--countries marked by political instability, social division and weak governance, which is the stereotypical image of Kenya, are more likely to rely on seigniorage and import tariffs, taxes which are relatively easy to collect but highly distortionary, and probably regressive in the case of seigniorage.23 The rise in the primary surplus in the wake of Goldenberg was achieved in part by raising tax rates. This plus high inflation boosted the tax take and pushed the revenue-to-GDP ratio to about 25 percent in the early-to-mid 1990s.24 It was felt that maintaining such a high rate of revenue mobilization would deter growth, so tax rates were cut and the revenue-to-GDP target lowered to 21-22 percent of GDP. Considerable reform in tax policy and administration has been implemented since the mid-1990s. After 2003, governance improvements have boosted direct tax collections notwithstanding large reductions in the CIT rate and the top PIT rate to 30 percent (from peak levels 45 percent each earlier in the decade). VAT rates and import tariffs have also been cut significantly. A `large taxpayers office' was set up in 2006 whereby 800 companies pay some 70 percent of total direct taxes. Figure 4 plots four important revenue components: import and excise duties, income tax and VAT as a percentage of GDP. Revenues fell as tax rates were cut but then picked up after 2003 as a result of stronger tax administration. The steady decline in import duty collection is by design, given the trade liberalization policy. Income tax displays a u-shaped pattern, falling in the late 1990s, but then, unlike import duty, recovering over the last three years. This is encouraging to the extent that it captures lower tax rates combined with a broadening of the tax base and 23Cukierman, Edwards and Tabellini (1992). 24Source: Meeting with Kenya Revenue Authority officials, February 2007. 16 improving compliance--which is the case in Kenya. VAT was introduced in 1990; the number of rates has been reduced from 15 (with the highest rate once 210 percent) to a unified rate 16 percent (except some zero-rated products), with obvious advantages for reducing misclassification and tax evasion and easing administration. Excise duties have been largely replaced by the VAT, but remain on oil imports and other specific products. Figure 4 Revenue Mobilization (% of GDP) 30.5 25.5 20.5 15.5 10.5 5.5 0.5 69 79 89 99 00 10 20 30 40 50 60 70 95/ 96/ 97/ 98/ 99/ 00/ 01/ 02/ 03/ 04/ 05/ 06/ 19 19 19 19 19 20 20 20 20 20 20 20 Revenue, excluding grants Income Tax VAT Import Duty Excise Duty What about seigniorage? Figure 5 plots both the growth rate of reserve (base) money and the change in reserve money divided by GDP (an indicator of seigniorage). 17 Figure 5 Reserve Money 60 4.5 4.0 50 3.5 40 3.0 30 2.5 2.0 20 1.5 10 1.0 0.5 0 0.0 -10 -0.5 1981 1983 1985 1987 1989 1991 1993 5 7 1995 1997 1999 2001 2003 200 200 Growth in Reserve Money (%) Change in Reserve Money (% of GDP, right scale) Both the level and volatility of these indicators have declined sharply after 1999. This in turn could have contributed to lower macroeconomic uncertainty and inflation, lowering interest rates. Likewise, Figure 6 shows that the effective import tariff rate and the share of import duties in total revenues, which were at peak levels after Goldenberg, have both come down sharply. The immediate impact of cutting import tariffs would be to increase the primary deficit (as we have seen) and lower the cost of imported machinery and raw materials as well as the tax on exports by Lerner symmetry. In combination with cuts in income tax rates, which raise the private returns to capital, the cut in import tariffs would increase competition from imports, spurring both more investment and higher productivity in firms, improving growth prospects. Figure 6 Import Duty and Effective Tariff Rate 17.5 15.5 13.5 11.5 9.5 7.5 5.5 3.5 Effective Tariff Rate (%) Import Duty (% of Revenue) 1.5 -0.5 92 93 94 95 96 97 98 99 1800 01 02 03 04 05 06 07 1991/ 1992/ 1993/ 1994/ 1995/ 1996/ 1997/ 1998/ 1999/ 2000/ 2001/ 2002/ 2003/ 2004/ 2005/ 2006/ Private sector firms interviewed in mid-2006 unanimously concurred that the economic liberalization starting in 1993 has been a major stimulus to higher efficiency and productivity (Box 2). Microeconomic evidence from Investment Climate Assessment 2007 corroborates this anecdotal evidence. It confirms significant growth in total factor productivity during the period 2002-2006 after stagnation during the 1990s.25 Box 2: Vignettes from Interviews with Private Sector Firms, July 2006 · Firms noted that reductions in import tariffs had spurred exports because it had lowered their costs, exactly as Lerner symmetry would predict. One firm described this as decisive in doubling its capacity in 2003, even more important than the fall in interest rates. · All firms mentioned pressures to increase efficiency and productivity by reducing waste and technological upgrading because of greater competition and falling profit margins as a consequence of lower import barriers. Many mentioned adopting Gemba Kaizen techniques to reduce waste, get workers more involved in preventive maintenance and doing things better on the shop floor.a/ One multinational said it had reduced manufacturing unit costs by 30-40 percent in spite of rising labor and energy costs. · Kenyan labor was typically described as well-educated and the best in the region. · While the increased competition has not so far led to bankruptcies, the expectation was voiced that loss-making parastatals will eventually close. In short: competition and hard budget constraints are forcing firms to become more efficient and innovative, while the increased incentive for exports is spurring capacity expansion. a/Gemba Kaizen in Japanese means change for the better (kaizen) in the workplace (gemba). It was originated by Masaaki Imai, a Japanese management consultant, who established the Kaizen Institute in 1986. A point worth making is that in spite of the cut in VAT, income tax and import tariff rates, which contributed to the decline in revenues since the mid-1990s till 2001/02, the revenue- to-GDP ratio at around 21 percent is still high in Kenya compared to other countries of a similar income level. Kenya collects at least 4-7 percentage points of GDP more in revenue than its EAC partners, Tanzania and Uganda, and is a positive outlier among developing countries. 25Based on a firm-level production function, Soderbom (2004) reported that average firm TFP increased by 7percent (but was not statistically significantly different from zero) over the entire 1999-2002 period. Using a similar methodology, ICA (World Bank 2007) provides preliminary evidence that TFP increased in manufacturing firms by a statistically significant 15 percent over 2002-2006. 19 4.4 Expenditure Composition The major complaint voiced by the private sector pertains to the quantity and quality of infrastructure. Roads in particular are singled out--the fuel levy has not translated into better roads.26 This brings us to the Achilles' heel of the Kenyan public finances: expenditure policy. Figure 7 plots total primary expenditure and two of its key components: development expenditure and recurrent expenditure in the past ten years. Development expenditure consists largely of capital expenditure, and most of it is on public infrastructure. During the fiscal consolidation of the 1990s, development expenditure was cut from a high of 6.4 percent of GDP in 1994/95 to 1.9 percent in 1999/2000. These cuts were accompanied by deterioration in the quality of public investment portfolio. A number of construction projects with questionable economic rationale were taken up (Eldoret airport was a prime example) and a large stock of incomplete projects accumulated.27 Figure 7 Primary Expenditure % of GDP 30.0 25.0 20.0 15.0 10.0 5.0 0.0 79 89 99 0 1 2 3 4 5 6 7 6/ 7/ 8/ 200 200 200 200 200 200 200 200 199 199 199 99/ 00/ 01/ 02/ 03/ 04/ 05/ 06/ 19 20 20 20 20 20 20 20 Primary excl drought o/w Recurrent o/w Development excl net lending 26The Road Maintenance Fuel Levy was established in 1993. It consists of an automotive fuel levy and transit toll collections. The objective of the levy was to provide funds for maintenance of roads; its proceeds are also used for rehabilitation and reconstruction. In 2006/07, approximately KSh 18.6 billion (1% of GDP) was collected as fuel levy. 27Kenya, Public Expenditure Review 1997. World Bank (1997). 20 The neglect of infrastructure in the 1990s resulted in a huge backlog of road maintenance and the network is now not maintainable at the current levels of spending. It is estimated that clearing the backlog of maintenance alone would cost KSh 150 billion (about 9 percent of 2006 GDP); periodic and routine maintenance require an additional 15 billion KSh per year. If urgently-needed network expansion and capacity enhancement are included, it is estimated that the annual spending requirement over a 7-year period is about 4 percent of GDP against the 2006/07 allocation of about 2.7 percent of GDP.28 The challenge therefore is to scale-up spending in order to maintain and rehabilitate infrastructure while ensuring the quality of investment portfolio. This challenge needs to be seen in conjunction with the competing demand for social spending, which has been increasing at a fast pace following the adoption of free primary education in 2003 and various health goals by the government. 5. Political Stability and the Country Risk Premium Interviews with private sector and participants in financial markets show that 2003 was considered a watershed year. The change in political leadership was seen as the culmination of a successful transition to multi-party democracy, a process that began in 1992. It was widely believed that Kenya's political development was irreversible and economic policies unlikely to change dramatically with changes in political leadership.29 In other words, there was a decline in perceived political and economic risk by investors. 28Sessional Paper: The Development and Management of the Roads Sector for Sustainable Economic Growth:, Ministry of Roads and Public Works, Nairobi, 2006. 29A statement by Titus Naikuni, chief executive of Kenya Airways, reported in Financial Times, June 13, 2007, characterizes the changed political scene and attitude of businesses. He said, "I think we've sort of separated the politicians, who are making noise, from those of us who want to make money. Because the politicians realize that without this money we won't have development." Of the election, he says, "It's time people realized Kenya has become very mature, in the sense that the business community couldn't care less about what happens. I don't put our business plans based on whether we're going to elect Kibaki or someone else." 21 5.1 Inferences from Domestic Interest Rates What can be inferred about country risk from the Kenyan government's domestic borrowing costs? Figure 8 shows the annualized yield on the 91-day T-bill rate. There is a clear declining trend after the spike associated with the Goldenberg scandal, with interest rates continuing to fall and going below 10 percent starting in 2002, with lows in 2003 and 2004. Since Kenya has no exchange controls, the following formal construct can help shed light on risks: iK = i$ + x^ + DRP + SRP , where iK is the nominal interest rate on Kenyan government T-bills (denominated in Kenyan shillings), i$ is that on US government T-bills (denominated in USD) and x^ is the target devaluation rate for the KSh/USD exchange rate. This is standard interest parity. Risk could arise on two counts: suppose the market believes the eventual depreciation rate will turn out higher than that (explicitly or implicitly) targeted by the central bank, CBK. It will then demand a devaluation risk premium, DRP, to compensate for this. Suppose it also believes, given Kenya's image of political instability and weak governance that the government could default on its debt. It would then demand a sovereign or default risk premium, SRP, to compensate for this. The SRP could be regarded as a function of Kenya's credit and inflation history, its political risk and the quality of its fiscal and financial institutions. 22 Figure 8 Kenya 91-Day T-bill rate (1991-2006 annualized yield, %) 90 91-Day Tbill (%) 80 70 60 50 40 30 20 10 0 7 4 1 1 7 4 1 1 7 4 1 1 7 4 1 1 7 4 1 1 7 4 91M 92M 93M 93M 94M 95M 96M 96M 97M 98M 99M 99M 00M 01M 02M 02M 03M 04M 05M 05M 06M 07M 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 Table 2 presents data organized around the preceding equation. It shows that Kenyan t- bill rates have fallen along with US t-bill rates, suggesting that the favorable global interest rate climate has filtered through to Kenya; this is particularly evident in 2003 and 2004. In addition, changes in monetary policy may have contributed to the very low interest rates in 2003 and 2004: the cash reserve ratio was lowered from 10 percent to 8 percent and then 6 percent. This coincided with the maturing of a large tranche of repo operations, boosting short-run liquidity and lowering domestic interest rates.30 Of course, since Kenya has an open capital account, this money could leave and be invested in US or UK T-bills; this would put upward pressure on Kenyan T-bill interest rates. Alternatively, banks and other investors may be hampered from investing overseas because of prudential restrictions; however, there is little evidence that such restrictions are binding.31 Besides, investors have the option of investing in capital market or real estate if returns on T-bills are considered too low. One additional factor may be contributing to 30Source: Meeting with CBK officials. Interviews with current and former senior officials indicate a deliberate strategy to stimulate the economy after 2003 by lowering interest rates and targeting moderate inflation rather than inflation in the 0-5 percent. 31For example, the Retirement Benefits Authority requires pension funds to invest no more than 15 percent of their funds overseas; but the actual volume is 5-10 percent. 23 low interest rates: Kenya as a regional safe haven for financial assets.32 Such investors are not likely to be fussy about the rate of return. Table 2. Interest Rate Decomposition (In percent) 1999 2000 2001 2002 2003 2004 2005 2006 2007 91-Day Tbill 13.3 12.1 12.7 8.9 3.7 3.0 8.4 6.8 6.8 U.S. Treasury bill 4.7 5.8 3.5 1.6 1.0 1.4 3.2 4.7 4.4 Diference 8.6 6.2 9.3 7.3 2.7 1.6 5.3 2.1 2.4 Actual KSh./US$ depreciation 17.8 7.0 0.7 -1.9 -1.2 1.6 -6.4 -4.1 -9.7 EMBI Africa (in basis points) .. .. .. 529 232 177 145 74 166 Nigeria 1338 2037 1426 2276 732 667 523 481 .. South Africa .. .. .. 238 141 95 85 85 164 EMBI+ Spread (in basis points) 824 756 731 765 418 356 245 169 239 Source: Central Bank of Kenya; International Financial Statistics; JPMorgan; and Staff estimates. In Table 2, the difference between US T-bill and Kenyan T-bill rates has fallen sharply starting in 2003, indicating declining devaluation and default risk; the average difference during 1999-2002 was 7.9 percentage points but went down to 2.8 percentage points during the 2003- 2007 period. Unfortunately, there is not a way of disentangling the DRP and the SRP. A standard way of doing so is to assign the same SRP to Kenyan T-bills as that on an instrument of similar maturity issued in the international capital markets (with SRP given by the spread on such an instrument relative to an identical currency and maturity borrowing by a benchmark country, such as the US) and then obtain the DRP as a residual; but Kenya's external debt is mainly on concessional terms and it has not had access to the international capital markets. Ex post, the Kenyan shilling has been on an appreciating trend since 2002, as can be seen from the table. In October 2006, S&P rated Kenya's capacity to repay long term sovereign obligations denominated in foreign currency as B+, which indicates that the sovereign has the capacity to meet its financial commitment, but adverse business, financial, or economic conditions could impair its capacity. Countries that received the same S&P rating of B+ in 2006 and are tracked by the EMBI+ index exhibited a spread of over 200 basis points (bps) at end-2006 over the U.S. 32For example, for oil money from southern Sudan. Anecdotal evidence suggests that the Kenyan shilling is widely tendered in the region, including in troubled areas like Somalia. 24 Treasuries, about 30 bps above the overall EMBI+ spread at end-2006.33 Table 2 also presents the EMBI spread for Africa and the EMBI+ spread, which have been on a declining trend since 2002, attesting to the improved global interest rate climate, from which Kenya has also benefited. What is exceptional when Kenya's experience is stacked against that of even emerging market countries is its ability to issue nominal (i.e., un-indexed) local currency debt at single-digit interest rates when countries such as Brazil and Turkey have had great difficulty in doing so in spite of running significant primary surpluses (far greater than Kenya's) for several years in a row (Brazil since 1999). This indicates a high degree of macro policy credibility and the absence of "debt intolerance" (Reinhart, Rogoff and Savastano 2003). Further, Kenya's debt structure is less vulnerable to external shocks as a result of the falling share of external debt over the past decade. Finally, Kenya's debt dynamics are favorable and it has been growing impressively over the past three years in contrast to the debt sustainability problems and slow growth in many of the emerging market countries. In January 2008, following post-election violence, S&P downgraded Kenya's long-term local currency rating to B+ from its earlier level of BB- and long-term foreign currency to B from B+, five levels below investment grade. 5.2 Survey Evidence from the Investment Climate Assessment What would be the implications of greater confidence and lower country risk? One would expect to see firms more willing to reinvest profits instead of depositing these in an overseas bank account. One would also expect investment horizons to lengthen, so that the hurdle rate for investments goes down.34 Indeed, the rising political and macroeconomic stability 33The Emerging Markets Bond Index Plus (EMBI+) tracks total returns for traded external debt instruments in the emerging markets. In addition to serving as a benchmark, the EMBI+ provides investors with a definition of the market for emerging markets external-currency debt, a list of the instruments traded, and a compilation of their terms. The EMBI+ tracks instruments mainly from the three major Latin American countries (Argentina, Brazil, and Mexico), reflecting the size and liquidity of these external debt markets. The non-Latin countries are represented in the index by Bulgaria, Morocco, Nigeria, Philippines, Poland, Russia, and South Africa. 34See section 2. 25 has led to a significantly reduced perception of risk on these two counts by the surveyed firms. In the Investment Climate Assessment (ICA) conducted by the World Bank in June 2007, the percentage of manufacturing firms which perceived political instability or uncertainty as a "major-to-severe" constraint dropped to 18 percent in June 2007 from 47 percent in 2003. Similarly, the percentage of manufacturing firms that consider macroeconomic stability to be a major constraint dropped to 28 percent in 2007 from 50 percent in 2003. Table 3 presents the changes in perceptions of manufacturing firms about the investment climate between 2003 and 2007.35 ICA 2007 also reported that a larger percentage of surveyed manufacturing firms were preparing multiyear business plans in 2007 (58 percent) than five years before (49 percent). A last remarkable observation is in order from Table 3. A higher percentage of firms considered electricity and transportation to be constraints in 2007 than in 2003. One possibility is that these sectors deteriorated between 2003 and 2007; but that does not appear plausible. The operational performance of Kenya Power and Lighting Company ­ the power operator - improved considerably after its management was contracted out to an international company in 2006. The number of power outages went down by about 40 percent between 2004 and June 2007, and new connections have increased by more than 50 percent. Table 3 Percent of Manufacturing Firms Perceiving Issue as Being a Major or Severe Constraint to Business Issue 2003 2007 Crime, Theft and Disorder 69 59 Tax Rates 69 56 Electricity 47 55 Corruption 73 54 Transportation 36 53 Practices of Competitors in Informal Sector 64 50 Tax Administration 52 50 Customs and Trade Regulations 40 42 Telecommunications 45 28 Business licensing and Permits 13 28 Macroeconomic Instability 50 28 Access to Finance 71 26 Political Instability 47 18 Access to Land 23 16 35These data are for a panel of 169 manufacturing firms that participated in both surveys. 26 Labor Regulations 22 16 Inadequately Educated Labor Force 31 11 Source: Kenya ICA, 2003 and 2007. Similar hard data for transport are not available but the government has been allocating increasing amounts for road rehabilitation since 2005/06. A more likely possibility is that the increased dissatisfaction with infrastructure reflects a resurgence in investment. Between 2003 and 2006, expenditure on gross fixed capital formation expanded at an average of 16 percent per year (Economic Survey 2007, Kenya National Bureau of Statistics). This evidence is consistent with the idea that after 2003, infrastructure services replaced political and macroeconomic risk as the binding constraint to private investment.36 A last piece of persuasive evidence supporting a sharp improvement in the investment climate is provided by import trends. Figure 9 shows a marked increase in the imports of machinery and industrial transport equipment after 2003. These imports averaged 5.6 percent of gross domestic expenditure (GDE, equal to consumption plus investment plus government spending) over 1996-2002. This went up to 9.3 percent of GDE over 2003-2006. Figure 9 Imports of Machinery and Transport Equipment (% of Gross Domestic Expenditure) 16.0 citse e 12.0 machinery and dom urt industrial transport machinery s ndi 8.0 os gr industrial of pexe 4.0 transport % equipment 0.0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 year 36For the binding constraint approach to growth diagnostics, see Hausmann, Rodrik and Velasco (2005). 27 6. Challenges Looking Ahead: Concluding Remarks The economic policy challenge for Kenya is to consolidate macroeconomic stabilization while continuing to accelerate growth. There are two ways of describing this challenge: 1. Increasing spending on the social sectors and infrastructure without compromising government solvency 2. Raising the rate of return to private capital while lowering the cost of capital. The good news in challenge 1 is that Kenya does not appear to have a debt sustainability problem and therefore has more latitude in generating `fiscal space' for infrastructure and social spending, than many MICs with such a problem, which has severely constrained their fiscal space options.37 Its debt dynamics have been especially favorable for the past 4 years as shown above and real interest rates have been low. With indebtedness on a steady decline over the past 10 years, the key question that policymakers need to ask is whether this trend should be preserved or whether an alternative goal may be socially more desirable, such as either keeping the government debt-to-GDP ratio constant or even letting it rise in the short-run while the infrastructure constraint is alleviated. The key point is to select infrastructure projects with high economic rates of return and ensure that infrastructure services are priced at a cost-recovery level. This together with any acceleration in growth as a result of relaxing the infrastructure constraint on private investment should ensure that over the long-term, the additional public investments pay for themselves. Resources can also be released for higher infrastructure investment through more efficient spending. For example, an ongoing study shows that the quality of roads portfolio can be improved by selecting projects with higher economic returns. There is also scope for curtailing waste through more efficient procurement. Similarly, in the education sector, better allocation of the time of existing teachers will help in financing the ongoing expansion on a fiscally sustainable basis while improving the quality of education outcomes. 37See for example Gill and Pinto (2005). 28 Challenge 2 is related to challenge 1 in that relaxing the infrastructure constraint will raise the return to private capital. In addition to macroeconomic stabilization, a crucial factor for keeping interest rates and the cost of capital low is political stability, which will also lengthen private investment horizons and thus the threshold rate of return for investment projects. We conclude by discussing this political stability challenge, which has superseded the economic policy challenge. Notwithstanding significant reform spurred by a reduction in aid by donors on governance grounds and as an attempt to bolster credibility after the costly and disruptive Goldenberg scandal broke in 1992, tangible results took a decade to emerge in the form of sharply improved government debt dynamics, faster growth and lower interest rates. We argue that the peaceful presidential elections and smooth transfer of power in 2003 was a threshold event enabling positive economic outcomes. While the latter coincided with falling global interest rates and a pick-up in growth throughout the developing world, we base our evidence on the vastly improved debt dynamics of the government and surveys of private investors, who pointed to rising political stability as a major factor lengthening their business planning horizons. Besides, interest rates fell in spite of a shrinking primary surplus. Apart from lowered country risk, this has been facilitated by a rise in the revenue-to-GDP ratio after 2003 even though marginal tax rates were cut, attesting to better governance and tax administration. Against this background, the new round of political instability unleashed by the flawed December 2007 presidential elections presents a formidable challenge. Not only is there need to achieve a speedy resolution, there is a fundamental need to establish political stability for an extended period of time--considerably longer than that achieved over the 5 years 2003-2007. This challenge promises to be at least as difficult as that of policy and institutional reform. 29 References Budina, Nina, and Norbert Fiess. 2004. "Public Debt and Its Determinants in Market Access Countries: Results from 15 Country Case Studies." PRMED, The World Bank, Washington DC. Cukierman, Alex, Sebastian Edwards and Guido Tabellini. 1992. "Seigniorage and Political Instability." American Economic Review, June. Dornbusch, Rudiger and Alejandro Werner. 1994. "Mexico: Stabilization, Reform and No Growth." Brookings Papers on Economic Activity 1;1994, 253-315. Fischer, Stanley. 1993. "The Role of Macroeconomic Factors in Growth." Journal of Monetary Economics 32, 485-512. Gill, Indermit and Brian Pinto. 2005. "Sovereign Debt in Developing Countries with Market Access: Help or Hindrance?" Chapter 4, 87-118, in: Gerard Caprio, James A. Hanson and Robert E. Litan, eds., Financial Crises: Lessons from the Past, Preparation for the Future. Washington: Brookings Institution Press. Hausmann, Ricardo, Dani Rodrik and Andres Velasco. 2005. "Growth Diagnostics." http://ksghome.harvard.edu/~drodrik/barcelonafinalmarch2005.pdf Reinhart, Carmen, Kenneth Rogoff, and Miguel Savastano. 2003. "Debt Intolerance." Brookings Papers on Economic Activity 1.2003. Rodrik, Dani. 1999. "Where Did All the Growth Go? External Shocks, Social Conflict and Growth Collapses." Journal of Economic Growth December. Razin, Assaf and Efraim Sadka. 2004. "A Brazilian-Type Debt Crisis: Simple Analytics." IMF Staff Papers Vol. 51, No. 1, 148-153. Serven, Luis. 2007. "Fiscal Rules, Public Investment and Growth." Policy Research Working Paper 4382. The World Bank. Soderbom, M. 2004. "Productivity, Exports and Firm Dynamics in Kenya 1999-2002". CSAE paper. Oxford: Centre for the Study of African Economies, Department of Economics, University of Oxford. Warutere, Peter. 2005. "The Goldenberg Conspiracy: The Game of Paper Gold, Money and Power." The Institute for Security Studies Occasional Paper 177, September. http://www.issafrica.org/static/templates/tmpl_html.php?node_id=318&slink_id=381&sli nk_type=12&link_id=28 World Bank. 1997. Kenya: Public Expenditure Review. Washington, DC. World Bank. 2003, 2007. Kenya: Investment Climate Assessment. Washington, DC. 30