Abstract
This Selected Issues paper on Kenya reviews Kenya’s external stability in a context where the exchange rate has strengthened and capital inflows are playing an increasingly important role. Kenya’s external current account deficit has widened, reflecting strong import volumes as well as rising import prices, particularly for oil, but external debt as a percent of GDP has declined steadily. Underlying these developments have been a steady increase in capital inflows and a remarkable rebound of economic growth since 2003 after two decades of stagnation.
I. Assessing Kenya’s External Stability and Competitiveness 1
A. Introduction
1. The Kenyan shilling has appreciated considerably over the past three years (through end-2007), both in nominal and real effective terms. Over the same period, Kenya’s external current account deficit has widened, reflecting strong import volumes as well as rising import prices, particularly for oil, but external debt as a percent of GDP has declined steadily. Underlying these developments have been a steady increase in capital inflows and a remarkable rebound of economic growth since 2003 after two decades of stagnation. As a result, despite a widening current account deficit, foreign exchange reserves held by the Central Bank of Kenya have risen and reserve coverage of imports has increased. A key challenge to maintain the recent growth momentum—assuming that the recent post-election violence has only been a temporary setback—will be to ensure that the real value of the shilling is consistent with maintaining both medium-term external stability and competitiveness.
This paper assesses this consistency from three different angles. We first apply three commonly used methodologies to examine whether the shilling’s real effective exchange rate (REER), Kenya’s current account balance, and its net foreign asset position are in line with the country’s economic fundamentals. We then evaluate the competitiveness of Kenya’s export sector based on estimates of unit labor cost. Finally, we examine Kenya’s export growth in recent years using a constant market share analysis.
B. External Stability Assessment
2. We begin our assessment by updating and refining an earlier study by Cheng (2004), which applies a behavioral equilibrium real exchange rate approach. Under this approach, a set of economic fundamentals that underpin REER developments is identified and regressed on the REER. The regression-predicted REER levels are then compared with actual observations. If the results indicate a higher (more appreciated) level of the actual exchange rate, this could signify a possible overvaluation.
3. The literature on determinants of real effective exchange rates is extensive, but empirical analysis differs depending on which fundamentals are appropriate for each economy and data availability. The REER is hypothesized to depend on the following five fundamental variables (see Data Appendix for further discussion of variables used in this analysis):
Productivity – The Balassa-Samuelson theory suggests that faster productivity increases in the traded goods sectors relative to major trading partners would lead to REER appreciation over time. We compared Kenya’s relative productivity to: i) major competitors for coffee and tea – Colombia, Sri Lanka, Uganda, and Tanzania (see discussion below) and ii) its major trading partners. Real GDP per capita was used as a proxy for productivity.
Terms of Trade – An improvement in the terms of trade appreciates the REER by raising a country’s income and increasing consumption, which will be divided between traded and non-traded goods. If demand and supply for non-traded goods is in equilibrium, then an improvement in the terms of trade will increase demand for non-traded goods and may lead to REER appreciation. Coffee and tea are two main commodities that account for nearly a quarter of Kenya’s total merchandise exports. Cheng (2004) finds that this variable has more power to explain REER developments than terms of trade, which was tested as an alternative to tea and coffee prices in this study.
Trade Openness – A more liberal trade regime lowers domestic prices and leads to a more depreciated REER, whereas trade restrictions may appreciate the rate. However, a more liberal trade regime is also likely to improve productivity in the tradable sector, which could counter the downward pressure of trade liberalization on the real effective exchange rate. Thus the actual effect of trade restrictions depends on individual country circumstances, although empirically, most studies have found that trade openness leads to REER depreciation. As trade restrictions and controls are difficult to quantify, openness is used as a proxy here.
International Investment Position (IIP) – Economies with relatively high net foreign assets (NFA) can afford a more appreciated REER, while debtor countries will need a more depreciated REER to generate trade surpluses given their external liabilities. In the absence of IIP data for Kenya, two IIP proxies are used: i) banking sector NFA and ii) a crude estimate of Kenya’s IIP derived from current account data.
Public Consumption – Higher public consumption is expected to appreciate the REER by increasing the demand for goods. To the extent government expenditure is more heavily weighted towards non-tradable than tradable goods, the real effective exchange rate may appreciate. Cheng (2004) finds that public consumption is not an explanatory fundamental variable for Kenya’s REER.
4. In light of the recently discovered overstatement of inflation by Kenya’s official CPI, we constructed a new CPI series to recalculate Kenya’s REER. The new series removes the upward bias in the food component of the official CPI after 1997.2 The results show a much slower pace of appreciation of the shilling over the past few years (Figure 1).
Kenya: Real Effective Exchange Rates (2000 = 100)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Sources: Kenyan authorities and IMF staff estimates.Kenya: Real Effective Exchange Rates (2000 = 100)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Sources: Kenyan authorities and IMF staff estimates.Kenya: Real Effective Exchange Rates (2000 = 100)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Sources: Kenyan authorities and IMF staff estimates.5. The results of the estimated equilibrium long-run relationship between the real exchange rate and the aforementioned variables which have explanatory power for Kenya’s real exchange rate are presented in Table 1. All the presented coefficients have the expected signs and are statistically significant. All the models were estimated with the dynamic ordinary least square method (Stock and Watson, 1993).
Equilibrium Real Exchange Rates: Regression Results
Equilibrium Real Exchange Rates: Regression Results
| Explanatory variable | Estimated coefficient | t-statistic |
|---|---|---|
| Net foreign assets (percent of GDP) | 0.014 | 3.86 |
| Trade Openness (in logs, percent of GDP) | -0.275 | -3.25 |
| Real tea and coffee prices (in logs, index) | 0.192 | 7.51 |
| Productivity differential (in logs, index) | 0.300 | 2.83 |
| Number of observations | 110 | |
| Adjusted R2 | 0.654 | |
Equilibrium Real Exchange Rates: Regression Results
| Explanatory variable | Estimated coefficient | t-statistic |
|---|---|---|
| Net foreign assets (percent of GDP) | 0.014 | 3.86 |
| Trade Openness (in logs, percent of GDP) | -0.275 | -3.25 |
| Real tea and coffee prices (in logs, index) | 0.192 | 7.51 |
| Productivity differential (in logs, index) | 0.300 | 2.83 |
| Number of observations | 110 | |
| Adjusted R2 | 0.654 | |
6. The regression results show that the real appreciation of the shilling over the three years through end-2007 has essentially been driven by improvements in economic fundamentals.3 In fact, the Kenyan shilling appears to have been undervalued over the period 2002-04—after three years (1998-2001) of overvaluation driven by high inflation. The 2002-04 undervaluation seems to reflect a slow market reaction to improved economic conditions.
Actual and Estimated Equilibrium REER: 1980-2007 (2000 = 100)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Actual and Estimated Equilibrium REER: 1980-2007 (2000 = 100)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Actual and Estimated Equilibrium REER: 1980-2007 (2000 = 100)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
7. Decomposing the change of the equilibrium REER over the period 2005-2007 shows that the bulk of the appreciation has been driven by the upsurge of tea and coffee prices in the world market and a strengthening net foreign asset position.4 Recent improvements in productivity also contributed to the shilling’s real appreciation, while changes in openness during this period had only a marginal effect on the exchange rate.
Kenya: Contributions to REER Appreciation, 2005-07
Kenya: Contributions to REER Appreciation, 2005-07
| Variable | Percent of total |
|---|---|
| Productivity | 5.8 |
| NFA to GDP | 9.7 |
| Coffee/Tea Price Index | 81.0 |
| Openness to Trade | 1.6 |
Kenya: Contributions to REER Appreciation, 2005-07
| Variable | Percent of total |
|---|---|
| Productivity | 5.8 |
| NFA to GDP | 9.7 |
| Coffee/Tea Price Index | 81.0 |
| Openness to Trade | 1.6 |
8. As mentioned earlier, several other configurations of explanatory variables were tested. The crude estimate of Kenya’s IIP described above was not statistically significant and did not perform as well as banking sector NFA. We found that the terms of trade (in place of tea and coffee prices) is marginally significant (at 10 percent) but substantially reduces the goodness of fit of the regressions, and splitting the terms of trade into export and import prices did not produce significant coefficients. The oil price was also insignificant as an explanatory variable. Similarly, we found that relative productivity with Kenya’s competitive partners, Colombia, Sri Lanka, Tanzania and Uganda – which are major tea and coffee producers, produced a model with a better fit than productivity relative to trading partners. Similar to Cheng (2004), public consumption did not seem to increase the overall goodness of fit of the model and was also insignificant.
9. Under the second approach we employed to assess the REER, the regression results from an IMF study (Lee et al., 2008) were used to derive the levels of the equilibrium current account balance for Kenya. This approach is similar to the first approach in that the economic fundamentals and demographic characteristics of a particular country are used to derive current account “norms”—levels considered to be sustainable. If the current account balance is found to deviate from the norms, estimates of REER adjustments to bring the current account back to equilibrium can be calculated based on trade elasticities with respect to the REER. One important caveat to note is that the results (Lee et al., 2008) are based on advanced and emerging market data, and as a result, the regression coefficients and the current account norms derived from them may not be fully applicable to Kenya, or for that matter, to other low-income countries.5 Indeed, it seems plausible that Kenya’s current account norm could be lower (i.e., larger current account deficits) than even emerging market economies as the country receives foreign aid on a continuous, albeit modest, basis.
10. Our calculations show that Kenya’s current account deficit has been generally smaller than its estimated norms except for a short period during the second half of the 1990s (Figure 3). However, it is important to note that the margin has been narrowing rapidly since 2002, when Kenya ran a current account surplus of 2 percent of GDP. With the current account deficit forecast to widen further over the medium term and perhaps breach the norms in 2008, future developments need to be monitored closely.
Kenya: Actual and Estimated Current Account Balance, 1990-2007
(in percent of GDP)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Sources: Kenya authorities and IMF staff estimates.Note: The three lines showing estimated current account norms are based on, respectively, pooled estimation, hybrid pooled estimation, and fixed effects estimation in Lee et al. (2008).Kenya: Actual and Estimated Current Account Balance, 1990-2007
(in percent of GDP)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Sources: Kenya authorities and IMF staff estimates.Note: The three lines showing estimated current account norms are based on, respectively, pooled estimation, hybrid pooled estimation, and fixed effects estimation in Lee et al. (2008).Kenya: Actual and Estimated Current Account Balance, 1990-2007
(in percent of GDP)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Sources: Kenya authorities and IMF staff estimates.Note: The three lines showing estimated current account norms are based on, respectively, pooled estimation, hybrid pooled estimation, and fixed effects estimation in Lee et al. (2008).11. A third approach examines the relationship between the level of net foreign assets—as measured by the IIP—and Kenya’s current account balance. The so-called external stability approach described in Lee et al. (2008) links a country’s net foreign asset position to its net financial flows and allows estimation of the corresponding levels of the current account balance that will stabilize certain levels of net foreign assets for a given set of macroeconomic assumptions, particularly on GDP growth and inflation. A simplified version of the resulting formulas states:
where cas is the current account balance in percent of GDP, g is the growth rate of GDP, π is the inflation rate, and bs is the net foreign asset position in percent of GDP stabilized by cas.
Kenya: Illustrative scenarios of stabilizing the IIP
Kenya: Illustrative scenarios of stabilizing the IIP
| Assumptions | Net Foreign Assets (percent of GDP) | Current Account deficit (percent of GDP) | ||
|---|---|---|---|---|
| GDP Growth | Inflation | |||
| NFA (2007) stabilizing Current Account | 6.5 | 5.0 | -33.5 | 3.5 |
| Current Account (2007) stabilizing NFA | 6.5 | 5.0 | -43.5 | 4.6 |
Kenya: Illustrative scenarios of stabilizing the IIP
| Assumptions | Net Foreign Assets (percent of GDP) | Current Account deficit (percent of GDP) | ||
|---|---|---|---|---|
| GDP Growth | Inflation | |||
| NFA (2007) stabilizing Current Account | 6.5 | 5.0 | -33.5 | 3.5 |
| Current Account (2007) stabilizing NFA | 6.5 | 5.0 | -43.5 | 4.6 |
12. Assuming a medium-term GDP growth rate of 6.5 percent per year, and an inflation rate of 5 percent (Kenya’s current target), the current level of Kenya’s current account deficit (about 4.6 percent of GDP in 2007/08) would stabilize NFA at about -43.5 percent of GDP. To compare this level of NFA with Kenya’s actual IIP position at the end of 2007, we made a rough estimation of the latter using annual balance of payments data.6 Our results suggest that Kenya’s IIP at the end of 2007 was around -33.5 percent of GDP. Thus, should Kenya continue to run a current account deficit at its current level, its NFA would decline over time and eventually stabilizes at -43.5 percent of GDP. If Kenya wants to keep its NFA at the current level in the future, it would have to reduce the current account deficit to about 3.5 percent of GDP.
13. Two caveats are in order. First, the estimated current account norms and IIP for 2007 are both very preliminary and any conclusions drawn from them would need to be interpreted with great caution. Second, the sustainability of Kenya’s foreign debt situation should be considered also in the context of its total public debt. Although Kenya’s public foreign debt has declined in recent years, international experience suggests that Kenya should aim for some further reduction in the total public debt relative to GDP (see Chapter II). More broadly, the macroeconomic balance and external stability approaches applied here should be complemented by other approaches, including debt sustainability analysis.
C. Competitiveness Assessment
14. One indicator of a country’s external competitiveness is unit labor cost, which is obtained by dividing labor cost by output (value added) per worker. An improvement in labor productivity without a fully offsetting increase in labor cost would, other things being equal, improve a country’s cost-competitiveness. Figure 4 suggests that Kenya made impressive strides in reducing unit labor cost during 2002-06, especially when compared with its two major East African Community (EAC) neighbors. However, it still lags behind China, which experienced a drastic decline in unit labor cost during 2002-05.
Change in Unit Labor Cost 1/ (percent)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Source: IMF calcualtion based on World Bank Investment Climate Surverys.1/ Changes are from 2002 to 2006 for Kenya and Tanzania, and 2002 to 2005 for all other countries.Change in Unit Labor Cost 1/ (percent)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Source: IMF calcualtion based on World Bank Investment Climate Surverys.1/ Changes are from 2002 to 2006 for Kenya and Tanzania, and 2002 to 2005 for all other countries.Change in Unit Labor Cost 1/ (percent)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Source: IMF calcualtion based on World Bank Investment Climate Surverys.1/ Changes are from 2002 to 2006 for Kenya and Tanzania, and 2002 to 2005 for all other countries.15. The question is whether Kenya’s reduced unit labor cost has translated into improved export competitiveness. This question can in part be answered by a constant market share analysis.7 However, lack of up-to-date and consistent data makes this a difficult task. For example, Kenya’s exports appear to be under-reported in the United Nations COMTRADE database when compared to data from the IMF’s Direction of Trade Statistics (DOTS) and official balance of payments statistics. Under-reporting in the 2002 COMTRADE figures would be consistent with the finding of a World Bank study (Ng and Yeats, 2005), and appears to be confirmed by import data from Kenya’s trading partners. For this analysis, we used data from the DOTS for values of total exports and adjusted Kenya’s individual exports (by commodity and by destination, based on COMTRADE data) across the board by an equal proportion so that they sum to the total DOT export values, both for 2002 and 2006.
16. Kenya’s exports expanded slightly faster than world exports during 2002-06 (Table 4). This has stabilized Kenya’s share of world exports after a long-term decline since 1977, when Kenya accounted for about 0.12 percent of total world exports (Figure 5). During 2002-06, however, there was no improvement in competitiveness; most of the expansion above the world pace was attributable to Kenya’s increasing reliance on the SSA and the “rest of the world” markets, whose imports grew more rapidly than the world average (Table 5).8 Interestingly, the share of Kenyan exports destined for the fast-growing Asian market declined during the period.
Decomposition of Cumulative Changes in Kenyan Exports, 2002-06
Decomposition of Cumulative Changes in Kenyan Exports, 2002-06
| Contribution from: | In Millions of US dollars | In Percent of Total |
|---|---|---|
| Global expansion | 1,790 | 91 |
| Export composition | -143 | -7 |
| Market distribution | 330 | 17 |
| Competitiveness | -12 | -1 |
| Total | 1964 | 100 |
Decomposition of Cumulative Changes in Kenyan Exports, 2002-06
| Contribution from: | In Millions of US dollars | In Percent of Total |
|---|---|---|
| Global expansion | 1,790 | 91 |
| Export composition | -143 | -7 |
| Market distribution | 330 | 17 |
| Competitiveness | -12 | -1 |
| Total | 1964 | 100 |
Kenya: Exports as Percent of World Exports (1950-2007)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Kenya: Exports as Percent of World Exports (1950-2007)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Kenya: Exports as Percent of World Exports (1950-2007)
Citation: IMF Staff Country Reports 2008, 337; 10.5089/9781451821192.002.A001
Effects of Market Distribution on Kenyan Exports, 2002-06
Effects of Market Distribution on Kenyan Exports, 2002-06
| Kenyan exports (percent of total) | Annual import growth (percent) |
||
|---|---|---|---|
| 2002 | 2006 | ||
| Asia | 13 | 10 | 19.4 |
| EU 25 | 37 | 26 | 16.0 |
| NAFTA | 10 | 8 | 11.7 |
| SSA | 29 | 39 | 20.6 |
| Rest of the world | 10 | 16 | 19.3 |
| World | 100 | 100 | 16.4 |
Effects of Market Distribution on Kenyan Exports, 2002-06
| Kenyan exports (percent of total) | Annual import growth (percent) |
||
|---|---|---|---|
| 2002 | 2006 | ||
| Asia | 13 | 10 | 19.4 |
| EU 25 | 37 | 26 | 16.0 |
| NAFTA | 10 | 8 | 11.7 |
| SSA | 29 | 39 | 20.6 |
| Rest of the world | 10 | 16 | 19.3 |
| World | 100 | 100 | 16.4 |
17. Despite the booming global commodity markets, Kenya’s export composition in general remained unfavorable for growth during the period. Among Kenya’s top five exports at the two-digit level of the Standard International Trade Classification (SITC, Revision 1), which accounted for about 50 percent of Kenya total exports in 2006, only one (petroleum and petroleum products) had a world growth rate that was higher than the average for all exports during 2002-06 (Table 6). Similarly, only three of Kenya’s top 10 exports faced world markets that were growing faster than the average. There are, however, some encouraging signs about Kenyan exports. Several of the top 10 exports, mostly manufactured goods, such as fishery and fish preparations, chemical elements and compounds, and non-metallic mineral manufactures, were growing strongly—and the latter two products did so despite relatively slow world market expansion.
Annual growth of Kenya’s ten top exports, 2002-06 (percent per year)
Annual growth of Kenya’s ten top exports, 2002-06 (percent per year)
| 2006 value | Kenya growth | World growth | |
|---|---|---|---|
| SITC item | (US$ million) | (percent) | (percent) |
| Coffee, tea, cocoa, spices | 957 | 12.8 | 13.9 |
| Fruit and vegetables | 448 | 14.8 | 10.3 |
| Crude animal and vegetable material | 381 | 9.7 | 13.1 |
| Petroleum and petroleum products | 291 | 4.1 | 28.7 |
| Clothing | 275 | 17.9 | 9.7 |
| Fish and fish preparations | 151 | 40.0 | 25.9 |
| Chemical elements and compounds | 150 | 22.4 | 13.0 |
| Non-metallic mineral manufactures | 131 | 20.3 | 5.4 |
| Iron and steel | 104 | 27.4 | 12.3 |
| Machinery, other than electric | 96 | 23.3 | 17.7 |
| All exports | 4108 | 17.7 | 16.4 |
Annual growth of Kenya’s ten top exports, 2002-06 (percent per year)
| 2006 value | Kenya growth | World growth | |
|---|---|---|---|
| SITC item | (US$ million) | (percent) | (percent) |
| Coffee, tea, cocoa, spices | 957 | 12.8 | 13.9 |
| Fruit and vegetables | 448 | 14.8 | 10.3 |
| Crude animal and vegetable material | 381 | 9.7 | 13.1 |
| Petroleum and petroleum products | 291 | 4.1 | 28.7 |
| Clothing | 275 | 17.9 | 9.7 |
| Fish and fish preparations | 151 | 40.0 | 25.9 |
| Chemical elements and compounds | 150 | 22.4 | 13.0 |
| Non-metallic mineral manufactures | 131 | 20.3 | 5.4 |
| Iron and steel | 104 | 27.4 | 12.3 |
| Machinery, other than electric | 96 | 23.3 | 17.7 |
| All exports | 4108 | 17.7 | 16.4 |
18. The finding that Kenya’s external competitiveness has not improved seems to contradict the prediction of falling unit labor cost. However, as Eifert et al. (2005) have shown, Kenya’s labor productivity at factory floors has been relatively high among low-income countries; what seems to be hampering improvements in Kenya’s competitiveness are nonfactor costs. The World Bank’s latest investment climate surveys (2008) suggest that costs related to infrastructure (roads, energy, and telecommunications), taxation, security, and bribes in Kenya are considerably higher than those in most of its competitors.
D. Concluding Remarks
19. The real appreciation of the shilling over the past three years through end 2007 is found to be driven largely by strong tea and coffee prices, a strengthening net foreign asset position, and improving productivity. Consistent with this finding, Kenya’s external current account deficit also seems to have been at a sustainable level, but room for further widening without compromising external sustainability has shrunk over the past few years as the current account deficit continues to increase. Indeed, future developments need to be closely monitored given the projected further widening of the current account deficit over the medium term. Kenya’s current account deficit in 2007 is near the level required to stabilize its 2007 NFA position although a somewhat lower level of NFA (i.e., larger negative value as a percent of GDP) may be sustainable. It is worth emphasizing, however, that these findings are preliminary and should be interpreted with caution given some of the data weaknesses. External sustainability should also be analyzed in conjunction with the assessment of public debt sustainability.
20. The shilling’s recent appreciation does not appear to have undermined Kenya’s export competitiveness so far. However, the recent productivity gains, which have been reflected in falling unit labor cost, do not appear to have translated into improved competitiveness in terms of an expanding market share, although it has contributed to the recent real appreciation of the shilling and may have helped to arrest the secular decline of Kenya’s weight in the world export market. Strong import demand in Africa has helped Kenyan exports to expand, but this increasing reliance on regional markets makes Kenya vulnerable to an economic slowdown in the region. Diversification into emerging markets would help build a globally competitive export sector as well as reduce this vulnerability. While there are some encouraging signs that some of Kenya’s key manufactured exports are growing strongly, its overall export composition remains unfavorable. Further improvements in Kenya’s export competitiveness would require reducing nonfactor costs as well as continuing to raise factor productivity.
Data Appendix for Real Effective Exchange Rate Estimation
Quarterly time series data were used from 1980Q1 -2007Q4. Where quarterly data were not available, linear interpolation was used to convert annual data to quarterly data.
Real Effective Exchange Rate (REER) – The actual REER comes from the IMF Institute through 1997. Beginning in 1997, a modified REER is calculated using IMF staff estimates of the CPI due to an overstatement of the official Kenyan CPI. The modified REER is scaled to the previous series for continuity around the 1997 period.
Productivity Differential – Real GDP per capita levels (using constant 2000 US Dollars), from the World Bank’s World Development Indicators (WDI), are used as a proxy. The differential is calculated as the average index of GDP per capita levels for Colombia, Sri Lanka, Tanzania and Uganda over the index of GDP per capita level for Kenya.
Tea and Coffee Prices – An average of coffee and tea price indexes from the IMF’s International Financial Statistics (IFS) is taken and deflated by the price of world merchandise exports.
Trade Openness – The sum of exports and imports of goods and services as a percent of GDP is taken from the IMF’s World Economic Outlook (WEO) database.
Net Foreign Assets (NFA) – Banking sector NFA data as a percent of GDP are taken from IFS.
Other variables were considered but were not used because they either were not statistically significant or did not improve the fit of the model:
International Investment Position (IIP) – Because Kenya’s IIP data are not available, a crude estimate of the IIP was constructed by using current account data from the IMF’s WEO database.
Terms of Trade – Terms of trade data are taken from the IMF’s WEO database.
Crude Oil Price – The crude oil price index is taken from the IMF’s WEO database.
Public Expenditure – Public expenditure is measured as central government expenditure and net lending as a percent of GDP and is taken from the IMF’s WEO database.
References
Cady, J. and J. Liu, 2003. “External Competitiveness” in “Madagascar: Selected Issues and Statistical Appendix”, IMF Country Report, No.03/7, International Monetary Fund, Washington, January.
Cheng, K. C., 2004. “Estimation of the Equilibrium Real Exchange Rate for Kenya”, in Kenya—Selected Issues and Statistical Appendix, International Monetary Fund, Washington.
Eifert, B. A. Gelb, and V. Ramachandran, 2005. “Business Environment and Comparative Advantage in Africa: Evidence from the Investment Climate Data”, World Bank, Washington, August.
Kiptui, M. and L. Kipyegon, 2008. “External Shocks and Real Exchange Rate Movements in Kenya”, paper prepared for presentation at the 13th Annual Conference on Econometric Modeling in Africa, University of Pretoria, 9-11 July 2008.
Leamer, E. and R. Stern, 1970. Quantitative International Economics, Allyn and Bacon, Boston.
Lee, J., G.M. Milesi-Ferreti, J. Ostry, A. Prati, and L.A. Ricci, 2008. “Exchange Rate Assessments: CGER Methodologies”, Occasional Paper, No.261, International Monetary Fund, Washington.
Ng, F. and A. Yeats, 2005. “Kenya: Export Prospects and Problems”, Africa Region Working Paper Series No.90, World Bank, Washington.
Stock, James H., and Mark W. Watson, 1993. “A Simple Estimator of Cointegrating Vectors in Higher Order Integrated Systems,” Econometrica, Vol. 61, Issue 4 ((July))pp.783–820.
World Bank, 2008. “Kenya: Accelerating and Sustaining Growth”, Report No. 42844-KE, World Bank, Washington.
Yang, Y., 2003. “Trade Strategy for Sustained Growth”, in “The Republic of Kazakhstan: Selected Issues and Statistical Appendix”, IMF Country Report 03/211, International Monetary Fund, Washington, July.
Prepared by Yongzheng Yang.
This paper was prepared by Yongzheng Yang, Olessia Korbut, and James John.
Given the large weight of food in Kenya’s CPI basket and the nature of the upward bias—it increases with the degree of price volatility, removing the bias in the food component of the CPI should correct for most of the overall bias.
This finding is consistent with that of Kiptui and Kipyegon (2008), which used some different explanatory variables and a different estimation method from this paper.
To the extent that there has been a common boom in most commodities, the strong effect of rising tea and coffee prices on the shilling may have also captured the effect of Kenya’s booming horticultural exports, which account for another quarter of Kenya’s total exports.
The IMF’s Research Department is in the process of estimating such coefficients for low-income countries.
Kenya does not yet compile statistics on IIP. The estimation involves cumulating flows (including reserves) in the capital and the financial accounts of the balance of payments over time.
For an exposition of the constant market share analysis, see Leamer and Stern (1970). For applications of this approach in the context of IMF surveillance work, see Cady and Liu (2003) and Yang (2003). The analysis decomposes changes in exports over a period into four components: (i) global demand as indicated by global export growth; (ii) effects of export composition, i.e., whether the country happens to export commodities that are growing faster than the average of all commodities; (iii) effects of market distribution, i.e., whether the country’s exports are concentrated in markets that are expanding faster than the global average; and (iv) a residual that indicates the supply side effect, often considered to reflect changes in competitiveness.
Using the same approach, it was found in the 2006 Staff Report on Article IV Consultation that Kenya’s export competitiveness improved during 2002-2005. This finding was influenced by the preliminary nature of the 2005 data on Kenya’s exports, which were subsequently revised downward in the UN COMTRADE database.
References
Reinhart, Carmen, Kenneth S. Rogoff, and Miguel A. Savastano, 2003, Brookings Papers in Economic Activity.
International Monetary Fund, 2002, Assessing Sustainability, Washington D.C., available via the internet: www.imf.org/external/np/pdr/sus/2002/eng/052802.htm.
International Monetary Fund, 2003, Public Debt and Fiscal Policy in Emerging Market Economies in World Economic Outlook, Fall, September 2003.
International Monetary Fund, 2005, Fiscal Responsibility Laws, Washington D.C. (unpublished).
Di Bella, Gabriel, 2008, A Stochastic Framework for Public Debt Sustainability Analysis, International Monetary Fund Working Paper 08/58.
Prepared by G. Everaert (FAD). The author wishes to thank Gabriel Di Bella for sharing data on Institutional Investor ratings, and Tom Krueger, Yongzheng Yang, Jan Gottschalk, Kadima Kalonji, and participants at the June 28 seminar in Nairobi for valuable comments.
Unless spending from privatization receipts is spread over time.
Central government domestic debt and public and publicly guaranteed external debt.
However, it does not include specific thresholds for assessing total public debt.
Their work focused on external debt, as opposed to total public debt.
A higher rating reflects a country’s higher attractiveness as an investment destination.
To avoid potential endogeneity between the variable periods of high inflation and sovereign risk, regressions were also performed using the time spent under high inflation since 1958, such that the inflation variable is mostly pre-determined to the sample period used. Results are generally robust to using this variable.
Correlation coefficient is 0.34 percent.
In principle, an infinite number of scenarios are possible. However, a reasonable set of assumptions was chosen to derive a relevant public debt threshold, or to illustrate sensitivity to assumptions used.
This is much in contrast with developed countries, whose primary surpluses react progressively stronger to rising debt up to public debt levels of 80 percent of GDP, and where the sustainable level of debt is estimated to average 75 percent of GDP (IMF, 2003).
Countries are classified as ‘weak’, ‘medium’, and ‘strong’ according to certain country-specific and institutional characteristics, summarized in the CPIA rating.
E.g. In 2000/01 when Kenya was hit by a severe drought, external debt service peaked to 3.6 percent of GDP. Eventually, a Paris Club rescheduling was agreed following the approval of a new IMF arrangement.
