This Selected Issues paper and Statistical Appendix provides an assessment of the financial regulations as a basic framework for public expenditure management and control in Kenya. The paper discusses a number of “external” factors that have led to a deterioration of formal implementation and control mechanisms, as well as recent efforts that have been undertaken to improve them. The paper also discusses various aspects of monetary policy in Kenya, including the instruments available to the monetary authorities, current challenges in the area of monetary control, and the state of financial markets in Kenya.


This Selected Issues paper and Statistical Appendix provides an assessment of the financial regulations as a basic framework for public expenditure management and control in Kenya. The paper discusses a number of “external” factors that have led to a deterioration of formal implementation and control mechanisms, as well as recent efforts that have been undertaken to improve them. The paper also discusses various aspects of monetary policy in Kenya, including the instruments available to the monetary authorities, current challenges in the area of monetary control, and the state of financial markets in Kenya.

I Public Expenditure Management and Control1

A. Introduction and Background

1. Major fiscal adjustment was undertaken in Kenya in the first half of the 1990s, as the overall budget deficit (on a commitment basis and excluding grants) fell from the equivalent of 11.0 percent of GDP in 1992/93 (July-June) to 1.4 percent of GDP in 1995/96; nevertheless, public expenditure management and control, which is discussed in this section, remains a major problem. The fiscal slippage in 1996/97, which saw the overall budget deficit rise from an original target of 1.4 percent of GDP to a final outturn of 3.4 percent of GDP, was largely associated with the failure of the government to observe proper oversight with regard to new spending commitments.

2. The lack of major improvement in public expenditure management and control in recent years is evident in a number of areas.2 Stark discrepancies have emerged between the intended and realized level and composition of public sector expenditure, especially in areas related to nonwage recurrent spending on education, health, and other social services and on operations and maintenance. In addition, the completion rate for development projects is low—an estimated 3 percent per annum. Moreover, for both recurrent and development outlays, financial “informality,” defined as the use of financing instruments outside the legal framework, continues to persist, as evidenced by the large stock of domestic arrears.

3. The problems that face Kenya with regard to public expenditure management arise largely out of a failure of the government to observe rules governing budgetary implementation and control, as spelled out in the Financial Regulations of Kenya.3 In normal circumstances, this legally binding set of regulations would appear to provide an adequate framework for managing public expenditure. Often, however, observance of legally mandated financial rules and regulations has given way to ad hoc decisions to increase public spending in politically favored ministries.

4. The next subsection provides an assessment of the Financial Regulations as a basic framework for public expenditure management and control. The following subsection discusses a number of “external” factors that have led to a deterioration of formal implementation and control mechanisms, as well as recent efforts that have been undertaken to improve them (particularly in cash management and control). The final subsection draws several conclusions.

B. Framework for Expenditure Management and Control

5. The Financial Regulations establish a complete framework of rules for budget preparation, implementation, and control. They assign highest importance to the approved annual budget as the document that is expected to guide fiscal policy throughout the year. In doing this, they provide a comprehensive set of rules and procedures to guide the preparation and approval by parliament of the annual budget, which is expected to include highly disaggregated expenditure ceilings for each agency, institution, and program. Parliament is the only institution empowered to make changes to the annual budget, including the reallocation of resources between ministries and even between spending items within an individual ministry’s budget.

Main rules in the Financial Regulations

6. For budget implementation, the Financial Regulations establish a tightly knit set of rules aimed at reducing the scope for deviation by accounting officers—the officers responsible in each ministry for its financial transactions—from the prescribed detailed expenditure program, as set out in the annual budget. Technically, the accounting officers are to be held personally responsible for any deviation from the procedures specified in the regulations. The degree to which proper budget implementation is ultimately observed is to be set out annually in a report to be issued by the Comptroller and Auditor-General to the Minister for Finance within seven months of the end of each fiscal year. The report is expected to include a full investigation of any infractions of the regulations. The Minister for Finance, in turn, is required to forward the report to parliament, where it is initially to be considered by the Public Accounts Committee (PAC) (dealing with the appropriations accounts) and the Public Investments Committee (PIC) (dealing with the public investment projects). Each committee is expected to produce a timely report, which is to be deliberated on by parliament as a whole. From these reports, parliament is expected to discuss and decide sanctions against those accounting officers who were found to surpass their mandate, as laid out in the Financial Regulations.

7. The budget implementation process is guided by a number of clearly defined rules and procedures. For each expenditure item, the treasury issues an authorization to incur expenditure, which shows the annual budgeted amount and the accounting officer who is responsible for overseeing the execution of this expenditure within the established rules and budget ceilings. The accounting officer is expected to request periodically the transfer of funds under the authorization to incur expenditure to the line ministry’s accounts for recurrent or development spending. The Treasury evaluates the accounting officer’s request and authorizes the transfer of funds. The accounting officer then can proceed to issue payment orders, or “vouchers,” after certifying the delivery of goods or services requisitioned by the ministry.

8. With regard to the establishment of spending commitments, the Financial Regulations allow them to be entered before the Treasury transfers new funds as long as these commitments stay within the limits established through the authorizations to incur expenditure. This practice seems appropriate if time lags between the order and delivery of certain goods or services are correctly anticipated and if the treasury is informed about the level of spending commitments that will likely require the transfer of cash in the near future. However, the regulations expressly prohibit the accounting officers from establishing commitments if the intended expenditure surpass the budgeted ceilings.

9. As the Financial Regulations aim at disaggregating expenditure control, accurate accounting and a smooth flow of information between the accounting officers and the treasury are key to ensuring proper budget implementation. The regulations mandate that each accounting officer account for all spending commitments entered, vouchers issued, and checks cashed in the ministry’s “vote book.” This vote book is the basis for detailed monthly reports to the treasury on commitments and cash outlays. In addition, the accounting officers are requested to reconcile their records of account balances and funds received throughout the month with those maintained at the treasury. Lastly, in each ministry, internal auditors who report to the treasury are expected to certify that all activities of the accounting officers are carried out in full compliance with the Financial Regulations.

10. Vouchers are used to carry out the cash outlays. Prior to the issue of a voucher, an accounting officer has to verify the legality of the claim and to certify that the amount to be authorized for payment falls within the approved expenditure ceiling. The Financial Regulations prescribe that vouchers must be issued immediately after a good or service has been supplied to the public sector.

Amendments to the approved budget

11. As mentioned before, the Financial Regulations start with the assumption that changes to the approved annual budget should be kept to a minimum. Any changes to the original budget must be justified as extraordinary cases, such as spending that is unexpected (i.e., generated by natural disasters) or “cannot be postponed without detriment to the public service.” Any changes to the approved budget have to be sanctioned by parliament, and line ministries are expressly prohibited from incurring any spending commitments in anticipation of parliamentary approval. The treasury can grant exceptional permits for additional spending commitments only if compliance with the general rule would “inhibit the maintenance of public services.”

12. If an accounting officer foresees the need to exceed the approved annual spending ceiling for an item, he or she has to seek, through the treasury, parliamentary approval of what is designated as a “supplementary request.” Without the treasury’s written approval, the accounting officer cannot incur any commitments that would be covered by the expected approval of additional funds. In any event, the treasury must seek parliamentary approval for the supplementary request. These requests are normally made via the supplementary budget, which is brought before parliament for consideration every year around March.

C. External Factors Weakening Budget Implementation and Corrective Efforts

13. Against this background, recent experience has seen the annual budget become subject to amendments and modifications soon after its approval. While unforeseen events, such as natural disasters, have required that certain changes be made, political decisions also have led to the establishment of new spending commitments (such as unbudgeted wage increases or new development projects outside the public investment program) beyond those in the approved budget. At the same time, approved budget projections have tended to be based on overly optimistic foreign financing flows (grants and loans). This overoptimism has made it necessary for the treasury to compensate for shortfalls in this area either by seeking higher domestic financing or by restraining less politically favored expenditure (i.e., spending on social services and operations and maintenance) below budgeted levels. Owing to these considerations, the treasury has been forced to manage public expenditure with a crisis management approach, trying to adjust cash releases to line ministries based on variations in available cash.

14. To formalize this situation, the treasury, at the start of fiscal year 1994/95, began to issue monthly spending ceilings to line ministries on their recurrent expenditure and quarterly ceilings on their development expenditure, with a view to keep spending commitments in line with short-term cash projections. Moreover, the treasury began to focus greater attention on monitoring the cash flow, as manifested in the daily movements in the accounts held by each ministry at the central bank. In addition, the treasury has had to constrain the actual release of funds to the ministries even beyond the established monthly and quarterly ceilings to accommodate de facto spending reallocations or reductions, as determined by the available cash flow to, and additional spending undertaken by, certain ministries. Therefore, a comparison of realized with budgeted expenditure reveals a larger variation at the individual ministry level than at the aggregate level. In such a situation, it has become virtually impossible for the accounting officers to implement their annual or supplementary spending plans.

15. Similarly, it is inevitable that accounting officers, relying on their spending limits, would regularly run the risk of entering spending commitments (within these limits) even when the actual flow of funds would not permit their orderly payment. Therefore, the rise of domestic arrears, or “pending bills,” or of overdrafts in the ministries’ accounts, is not surprising. While these informal mechanisms of financing stand in clear violation of the Financial Regulations, the treasury apparently has tolerated these practices, because it has been unable to provide the funds envisaged for an orderly implementation of the approved budget.

16. Because of the increase in overdrafts and pending bills, the treasury eliminated the possibility of ministerial overdrafts as of the beginning of the fiscal year 1997/98. Since July 1997, only the treasury has been allowed to make use of the overdraft facility at the central bank, whose ceiling was established in the recently amended Central Bank of Kenya Act. At the same time, the approved budget for 1997/98 included provisions for eliminating the stock of pending bills as of end-June 1997.

17. At this stage, it is too early to assess whether the tightening of cash controls in the current fiscal year will result in the intended control over spending commitments. The current system still focuses exclusively on the monitoring of cash movements, while formal controls regarding spending commitments are not firmly in place. In early 1998, the authorities initiated a process to strengthen the monitoring of spending commitments by improving the timeliness and accuracy of the monthly expenditure reports that are being submitted by line ministries (as discussed above).

18. Parliament’s recurring delays in discussing the Auditor-General’s reports are a major area of concern. The deliberations on these reports are seen as a primary means of enforcing accountability in the management and use of public funds. The Auditor-General’s reports covering the fiscal years 1992/93 to 1996/97 have not yet been discussed by parliament. Also, parliament only received its last set of reports from the Controller and Auditor-General for the fiscal year 1996/97 in April 1998—a clear violation of the established deadlines. The delays in this area partly arise from continued interference in the operations of the Auditor-General’s office and in part from a decision to delay discussion of these reports in parliament. The Auditor-General’s office has been hampered in its functions by an underprovision of budgetary funds for needed equipment, such as computers, and by the frequent removal of specially trained staff to other positions within the public sector.

D. Conclusion

19. The above analysis has shown that deficiencies in budget implementation are largely related to changes to the approved budget that stem from ad hoc policy decisions and overly optimistic projections of foreign disbursements. While the Financial Regulations establish appropriate rules for budget implementation and control, their functioning depends critically on the reliability of the approved budget’s spending ceilings as a guide to the execution of spending at the ministerial level. However, the annual budget has become subject to frequent and unforeseen changes throughout the fiscal year that introduce a high degree of uncertainty into the orderly execution of annual spending programs at the ministerial level. The treasury’s crisis management of the cash flow has, at least implicitly, acknowledged this situation by tolerating informal means of financing, in particular the accumulation of domestic arrears, which stand in clear violation of the Financial Regulations. This situation, together with the prevailing notion that infringements against the regulations are normally not penalized, has led to a further spread of fiscal indiscipline that is in urgent need of correction.

II. Monetary Policy and Financial Sector Issues4

A. Monetary Policy Framework

20. This section discusses various aspects of monetary policy in Kenya, including the instruments available to the monetary authorities, current challenges in the area of monetary control, the state of financial markets, and the financial aggregate target that could be more appropriate to use for monetary programming in Kenya.

21. The central bank can affect monetary conditions in Kenya in the following ways:

  • Primary auctions of government paper. American-style auctions are conducted once a week (on Thursdays), with the volume of paper sold in principle determined by both budgetary financing needs and monetary policy considerations. The process begins on the previous Friday, when the auction is advertised in all major daily newspapers in Kenya and the volume of paper to be auctioned is announced. Participants in the auction must physically drop off at the Central Bank of Kenya (CBK) their bids, which must include the face value of the paper to be bought,5 the price offered, and the maturity (91 or 182 days) desired. The Monetary Policy Committee of the central bank (which includes Ministry of Finance representatives) meets on Thursday afternoon to determine the amount to be actually sold, taking into consideration not only the financing need of the government and the amount of liquidity needed to be withdrawn as implied by the reserve money target of the central bank, but also the cutoff rate of interest that various volumes sold would imply.6

  • Open market operations (“tap sales”). Between primary auctions, the central bank engages in sales of government paper from its portfolio at rates that it negotiates individually with each bank. The central bank initiates these sales, often in the aftermath of primary auctions when only a low volume of paper has been sold—due either to undersubscription or to a small volume of bids accepted by the CBK, which might have perceived the rates (in the auction bids) to be excessively high.

  • Repurchase agreements (repos). The central bank engages in repurchase and reverse repurchase agreements with a view to mopping up liquidity in and injecting liquidity into the banking system, respectively. Decisions to offer repos are made on a daily basis on account of the perception of the tightness of monetary conditions and deviation from targets. The CBK will post electronically both a buying and a selling rate,7 including the amount and the maturity. Offers by the market are taken on a first-come, first-serve basis, but the CBK reserves the right to ration the liquidity injected.

  • The required reserve ratio. The required reserve ratio has been used on occasion by the central bank to tighten or loosen monetary conditions. The last time the ratio was changed was on October 1, 1997, when it was decreased from 18 percent to 15 percent. Banks have to fulfill the ratio twice a month: in the middle and at the end of every month. The ratio is determined as the average amount of deposits of a bank (or a nonbank financial institution) with the central bank over a 15-day period ending in the middle and at the end of every month to the amount of domestic currency deposits held with the bank by residents and nonresidents on the last working day of the penultimate month through the middle of each month, and on the last day of the previous month thereafter.

  • The rediscount window. Banks can approach the CBK and rediscount their holdings of government paper without limit at a price determined by the central bank on a daily basis. The rediscount rate in recent times has been maintained at 5 percentage points above the last treasury bill auction rate.

  • The Lombard window. This facility permits a bank to borrow from the central bank up to 5 percent of its paid-up capital at a cost equal to the interest rate in the last treasury bill auction plus 3 percentage points. Collateral has to be in the form of government paper maturing in no more than 91 days or paper quoted on the Nairobi Stock Exchange.

  • The lender-of-last-resort window. This facility permits a bank to borrow funds from the CBK in excess of the 5 percent of its paid-up capital available under the Lombard facility but at a higher cost—at 5 percentage points above the last treasury bill auction rate. The required collateral is the same as for the Lombard facility.

  • Foreign exchange market operations. The CBK effectively injects or withdraws liquidity by engaging in foreign exchange market intervention. Intervention in the foreign exchange market in recent times has been primarily determined by a) the need to acquire foreign exchange for the government’s debt-servicing needs and b) the desire to moderate sharp movements in the exchange rate in a market that the central bank considers excessively thin. Thus, intervention has not been geared to affect domestic liquidity, although it can do so. In this light, the central bank often makes an effort to sterilize its intervention in the foreign exchange market through corresponding repos in the local currency market.

Issues of monetary control

22. Monetary policy in Kenya has been carried out not only via the first four operations mentioned above (primary auctions of government paper, open market operations, repos, and the required reserve ratio), but also through the last four operations mentioned, which are either not under the direct control of the monetary authorities (as in the rediscount, Lombard, and lender-of-last-resort windows, where the central bank is obliged to provide the amount of liquidity demanded and can only change the price) or are aimed at influencing the exchange rate or fulfilling the debt-service obligations of the government. These last four kinds of transactions have complicated the control of monetary aggregates, as they have had to be counterbalanced by the first four kinds of transactions, something which has not always been easily done. These difficulties have been exacerbated by the fact that there have been constraints in using the first four operations. For example, the primary auctions are dominated by the need to roll over maturing government debt, which, given its short maturity, has been very large every week.8 The large gross financing need implies that interest rates on government paper are kept high, which effectively precludes efforts to increase the auctioned amount further in order to absorb liquidity for monetary policy purposes, for fear of raising interest rates higher. Another example of a constraint on the conduct of monetary policy is the relative inflexibility afforded by the required reserve ratio as an instrument for monetary management over the short run. Finally, the remaining two instruments in the first category of operations—tap sales and repos—are hampered by the asymmetry between operations to mop up liquidity and operations to inject liquidity. Specifically, tap sales and sterilization repos are truly under the control of the banks and not of the monetary authorities, as the latter have no way to force the mopping up of liquidity but can only try to induce it by offering attractive rates to the banks. Often, the banks will not respond to seemingly attractive rates; this can be especially true when the need to tighten is most urgent, as during a time of speculative attack against the currency. In these circumstances, the monetary authorities in Kenya can find themselves unable to control the quantity of liquidity despite the apparent sophistication of the instruments available.

23. The solution to this problem appears to be in reducing the automaticity with which liquidity is provided under the second category of operations. This could be done by putting maximum limits on the amounts that were made available under these facilities so that they are consistent with monetary/liquidity targets of the CBK. Regarding foreign exchange market operations, an approach consistent with the exchange rate regime to which the authorities aspire is to avoid intervention in the foreign exchange market within a broad band, thus eliminating the influences of such intervention on domestic monetary conditions in most circumstances. In this approach, the government would typically acquire foreign exchange needed to service its foreign debt as any other debtor in the country by entering the market. If a “pure float” would appear to be unfeasible or undesirable even within a broad band, some intervention in the foreign exchange market could be manageable if the CBK had the means to control the other sources of domestic liquidity9—specifically, if it had the option of withdrawing liquidity offered to banks under its various facilities (as the problem is not to inject liquidity but to mop it up). Following this approach would imply that liquidity would have to be offered to banks mostly in the form of repurchase agreements and very short-term loans under the Lombard and lender-of-last-resort windows; the CBK could let these agreements and loans to mature and not renew them, thus automatically mopping up liquidity without directly relying on interest rate increases to do so.

The state of markets

24. The secondary market for treasury bills continues to be relatively underdeveloped in Kenya. Its development has been partially constrained by the extensive involvement of the CBK in the market through the extensive use of its tap sales of treasury bills and its rediscount window.10 The secondary market for bonds is somewhat more developed as certain banks have assumed the role of marketmakers.11

25. The interbank market for deposits is considered to be thin, as only overnight deposits are quoted and the volumes tend to be small, in the range of 5 percent of required reserves. Interest rates charged on overnight funds can vary significantly from bank to bank. The dispersion of rates charged for overnight funds has often been 30 percent of the weighted average interest rate on daily interbank transactions and can reach much higher. This situation reflects, inter alia, the perceived differences in the creditworthiness and vulnerability of certain banks. Banks also place other deposits with each other, but there is limited information about the terms at which these deposits are accepted, which are not quoted as part of the interbank market. Altogether, deposits between banks are relatively small in total volume, reaching only about 25 percent of required reserves, or 3 percent of broad money; however, the bulk of these deposits is concentrated in certain banks, which appear to be excessively dependent on them.

26. The foreign exchange market is also thin, and small volumes have been known to move the exchange rate significantly. The actual rates at which trades are made are not disclosed to the public, but the CBK publishes the average of the opening (buying and selling) rates of nine major banks on a daily basis.

Monetary aggregates and financial programming considerations

27. In recent years, the conduct of monetary policy in Kenya has focused on the behavior of the broad monetary aggregate M3. This aggregate has been defined to include currency in circulation and term and nonterm domestic currency deposits with banks as well as with nonbank financial institutions, but not to include foreign currency deposits (FCDs) held by residents. This omission appears to have caused the growth of liquidity in the hands of the public to be understated—at least in the course of 1997, when there was both an increase in the foreign currency value of FCDs and in their domestic value after the depreciation of the Kenya shilling in July-August 1997. Specifically, while annual M3 growth was 9.8 percent at end-December 1997, the growth of M3X, which includes FCDs held by residents, reached 11.9 percent. Understating the growth in liquidity could lead to looser-than-warranted monetary conditions, with negative effects on inflation and balance of payments performance.

28. Another potentially important source of liquidity and of other services offered by money has been the holding of government paper by the nonbank public. This paper is by and large of short maturity (up to one year, and in its majority no longer than 91 days). Also, it is quite liquid because the CBK has stood ready to rediscount all treasury bills presented by nonbanks, as well as banks; in addition, treasury bonds can be readily sold by nonbanks to commercial banks, designated as marketmakers, which subsequently can quickly rediscount them with the CBK. In this light, treasury bills and bonds have characteristics close to those of term deposits with commercial banks, as the latter also have penalties associated with early withdrawal (which is the equivalent of rediscounting costs). Finally, government paper has been offering very attractive rates of return and has the backing of the state, which may imply (to some lenders) lower risk than that on commercial bank deposits. The rates on 91-day government paper have averaged about 27 percent since August 1997, while the rates on wholesale term deposits with commercial banks for three to six months have been in the range of 17-20 percent over the same period.

29. As government paper is a close substitute for bank liabilities, a measure of liquidity in the economy that includes bank term deposits but does not include government paper could understate the amount of liquidity held by the nonbank public, as in the case of excluding FCDs, and thus lead to the setting up of monetary targets that are inconsistent with the inflation and balance of payments objectives of the authorities. Alternatively, this point can be formulated in terms of projecting the demand for M3. Specifically, projecting the demand for M3 without considering the demand for government paper12 could lead to an excessively loose monetary policy if portfolio shares of government paper and deposits changed in favor of a greater share of government paper. By the same token, a portfolio shift in the direction of relatively larger holdings of bank liabilities could leave the monetary policy stance tighter than originally anticipated. In contrast, if the monetary programming exercise includes the nonbank holdings of government paper in the aggregate whose demand is being projected (or targeted), or the exercise projects M3 by taking explicitly into consideration the projections of the behavior of holdings of government paper, its potential for accuracy increases.13

30. The above considerations become increasingly relevant the more the nonbank public undertakes large shifts in its portfolio between government paper and bank liabilities. In cases where the portfolio shares remain stable for long periods of time or have only small effects on the growth of bank liabilities, monetary programming can proceed with small cost at the level of M3 and without regard to movements in potentially close substitutes, like government paper. However, the increasing sophistication of cash managers in large enterprises and (usually later on) of the population can rapidly change the monetary environment; in this case, persisting in the use of M3 under the assumption of broadly constant velocity14 could lead to misjudgements in monetary policy. The same point applies also in cases where monetary programs incorporate expected changes in velocity that, however, are not systematically revised in light of unanticipated shifts between M3 and its substitutes.

31. If the hypothesis is true that nonbank holdings of government paper are an essential part of liquidity, and thus the demand for M3 is contingent on the demand for such paper by nonbanks, one should be able to observe, inter alia, the following: (a) the variance of the velocity of the broader aggregate that includes M3 and nonbank holdings of government paper (called here M4) is smaller than the variance of the velocity of M3; (b) the rate of change in the velocity of M4 is nearly always lower in absolute terms than the rate of change of the velocity of M3; and (c) the correlation coefficient of the inverses of the velocities of M3 and of nonbank holding of government paper is a negative number significantly greater than 0 (but less than 1). The observation of these phenomena could serve as supportive evidence of the hypothesis that M4 rather than M3 is the more appropriate aggregate to project in monetary programming if one uses a constant-velocity assumption (which is usually the case in Fund- supported programs).15

32. For Kenya, monthly data were compiled for the period December 1992-March 1998 to calculate the velocity of M3 and of M4 (Figures 1 and 2).16 It was found that M3 velocity had a mean of 1.19, while M4 velocity had a mean of 0.99. Moreover, the variance of M3 velocity was 0.0099, while that of M4 velocity was 0.0028, indicated that velocity was much more stable for the broader aggregate.17 The same finding holds when the variances of M3 and M4 velocities are calculated over various subperiods in the sample. When the rate of change of the velocities of M3 and M4 were calculated, it was found that the velocity of M4 had a lower rate of change (in absolute terms) than that of the velocity of M3 in 48 out of the 52 months in the period December 1993-March 1998 (Figure 3). Finally, the correlation coefficient of the inverses of the velocities of M3 and of nonbank holdings of government paper was equal to minus 0.6, which is also consistent with the portfolio shift hypothesis that changes in M3 are strongly associated with opposite changes in nonbank holdings of government paper. Thus, the data appear to support the view that the velocity of M4 has been more stable than that of M3 since at least 1993, and that financial programming could have benefitted from using M4 as the aggregate—the projections of which would form the basis of the monetary programming exercise.18

Figure 1.
Figure 1.

Kenya: Inflation and Change in M3 and M4, December 1993-March 19981

(In percent)

Citation: IMF Staff Country Reports 1998, 066; 10.5089/9781451821055.002.A001

Sources: Kenyan authorities; and Fund staff estimates.1M3 is defined as currency in circulation plus term and non-term deposits with deposit money banks and nonbank financial institutions; M4 is defined as M3 plus nonbank holdings of government debt.
Figure 2.
Figure 2.

Kenya: Velocity of M3 and M4, January 1993-March 19981

Citation: IMF Staff Country Reports 1998, 066; 10.5089/9781451821055.002.A001

Sources: Kenyan authorities; and Fund staff estimates.1M3 is defined as currency in circulation plus term and nonterm deposits with deposit money banks and nonbank financial institutions; M4 is defined as M3 plus nonbank holdings of government debt.
Figure 3.
Figure 3.

Kenya: Change in Velocity, December 1993-March 19981

(In percent)

Citation: IMF Staff Country Reports 1998, 066; 10.5089/9781451821055.002.A001

Sources: Kenyan authorities; and Fund staff estimates.1M3 is defined as currency in circulation plus term and nonterm deposits with deposit money banks and nonbank financial institutions; M4 is defined as M3 plus nonbank holdings of government debt.

33. In an effort to get a sense of the implications of considering M3 as the monetary aggregate most stably related to prices and output, instead of the more appropriate M4, an alternative inflation rate for January 1994-March 1998 was simulated and compared with the actual inflation rate. The simulated inflation rate was constructed by constraining increases in M4 to equal the observed (actual) increases in M3 and by using the observed path of M4 velocity in this period. The simulation also assumes that output changes would be invariable to monetary conditions. Finally, it should be noted that the simulated inflation rate is not the optimal inflation rate, as potentially it could still be too high or too low. In that sense, no conclusions are to be drawn about the desirability of policies in a given period;19 the exercise is aimed only at indicating which direction inflation would have taken had M4 (as the more appropriate measure of liquidity in the economy) increased by the amounts finally observed for M3.

34. The results of the simulation are presented in Figure 4, where simulated inflation is shown to be lower than actual inflation in two periods—in most of 1994 and, more recently, in 1997 and the first quarter of 1998. Conversely, simulated inflation is higher in 1995 and, to some extent, in 1996. These results illustrate the potential problem of ignoring alternative forms of liquidity to M3. In 1994, as in 1997 and early 1998, there were significant portfolio shifts by the nonbank public away from money and into government paper holdings. As a result, the 12-month inflation rate at certain times in the first part of 1994 was probably up to 14 percentage points higher than it could have been.20 Similarly, in the last three quarters of 1997 and the first quarter of 1998, inflation could potentially have been up to 7 percentage points lower than actually observed, reaching as low as the 0-2 percent range. In this sense, one can conclude that the actual monetary conditions in these periods were noticeably looser than they appeared if one concentrated on M3 alone. Alternatively, one can conclude that, had the authorities adjusted their monetary policy so as to contain increases in liquidity (as measured by M4) to the M3 increase that they finally allowed, inflation would have been significantly lower than actually observed (Figure 4).

35. In the period 1995-96, the simulation results illustrate that ignoring portfolio shifts could also had led to an unexpected tightening of monetary conditions. The simulation actually hints at an additional explanation for the particularly positive inflation performance in 1995, besides other factors not discussed here (like external sector developments). Inflation in that year would have been much more in line with performance in 1996 and 1997 if the monetary authorities had recognized and adjusted (by loosening policies, in this case, so that growth in M4 would have been equal to actually observed growth in M3) for the shift away from government paper and into bank liabilities. This conclusion, however, does not imply that such a response would have been appropriate, as (a) the low inflation rate actually observed would probably have been preferred to a higher inflation rate, and (b) growth in liquidity equal to realized M3 growth is not assumed to have been an optimal target.

Figure 4.
Figure 4.

Kenya: Actual and Simulated Inflation, January 1994-March 1998

(In percent)

Citation: IMF Staff Country Reports 1998, 066; 10.5089/9781451821055.002.A001

Sources: Kenyan authorities; and Fund staff estimates.

B. Banking Supervision Issues21

Structure of the financial sector

36. As of December 31, 1997, Kenya’s financial sector comprised the Central Bank of Kenya, 53 domestic- and foreign-owned commercial banks, 15 nonbank financial institutions, 2 mortgage finance companies, 4 building societies, and numerous insurance companies and other specialized financial institutions. The Nairobi Stock Exchange was established in 1954 and is overseen by the Capital Markets Authority (CMA)—the supervisory body in charge of capital markets development in Kenya. Foreign participation in the Nairobi Stock Exchange was allowed in January 1995. As of December 1997, 58 companies, with a total market capitalization of US$ 1.8 billion, were listed on the exchange.

37. Despite the existence of numerous banks, the banking sector in Kenya is dominated by four large banks, which together control 50 percent of bank assets and 52 percent of bank deposits.22 The largest bank in the system is the state-owned Kenya Commercial Bank, with 17 percent of bank assets and 18 percent of bank deposits. It is followed closely by the foreign-owned Barclays Bank, with 16 percent of bank assets and 15 percent of bank deposits. The group of large banks also includes the state-owned National Bank of Kenya and the foreign-owned Standard Chartered Bank, each having 8 percent of bank assets and 9 percent of bank deposits.

Bank supervision practices

38. Bank supervision in Kenya is conducted by the Bank Supervision Department of the CBK, which operates under the laws of Kenya, The Banking Act, and the Central Bank of Kenya Act. The Banking Supervision Department has the power to grant and revoke licences of institutions, to set minimum capital requirements, prudential ratios and regulations, and the frequency of financial statements, to audit and review institutions, and to approve external auditors. Overall, the department can be deemed to have the legal and regulatory authority to effectively supervise the operations of the banking system in Kenya.23 It lacks, however, the legal authority to impose penalties for violation of regulations that do not relate to monetary policy; such authority is reserved for the Ministry of Finance.

39. To ensure prudent management of banking institutions, the Banking Supervision Department conducts both on-site inspections and off-site surveillance of banking institutions. Regarding on-site inspections, they are intended to be conducted regularly; the goal is to visit each institution annually. The off-site surveillance is a continuous process whereby information is obtained from banks and other financial institutions in the form of a set of returns. In this context, banks are required to periodically furnish the Banking Supervision Department with statistical and other information covering key areas of their operations to enable the department to form an opinion of their financial soundness. For example, liquidity positions are submitted every ten days, balance sheets and foreign exchange operations statistics on a monthly basis, profit and loss statements on a quarterly basis, and externally audited accounts on an annual basis. On the basis of this information, the Bank Supervision Department rates the reporting institutions, and, upon identification of weaknesses, the institutions are asked to undertake any necessary remedial measures.

40. While on-site examinations of each institution are to be performed as often as possible (with annual inspections the goal), on-site examinations in recent years have been conducted relatively infrequently. For example, in 1997, full on-site examinations were performed at only 19 out of 74 institutions (26 percent), all of which were banks; no inspections were conducted on any of the 15 nonbank financial institutions, the 2 mortgage companies, or the 4 building societies. Only 23 percent of the total assets in the system were scrutinized during the on-site inspections in 1997. The picture that emerges from a review of the records of the Banking Supervision Department in recent years is that of only sporadic full on-site examinations of financial institutions; some institutions have not been subjected to an on-site inspection in several years. Specifically, 19 institutions (30 percent of total) have not had an on-site examination since 1994, including 3 of the 7 largest institutions (Peer Group 1, as defined by the CBK).

41. With full on-site inspections lagging, the Bank Supervision Department has carried out instead targeted inspections (i.e,. inspections of a particular area of a bank’s operations) and special investigations. In 1997, the on-site division conducted 31 targeted inspections and 4 special investigations. Targeted inspections focused on areas such as asset quality, capital adequacy, and earnings. Special inspections included areas such as imprudent banking and lending practices and insider lending.

42. The Banking Supervision Department has operated in recent years under certain handicaps that help explain the small number of full on-site investigations. In 1994-95, the department was restructured, and many of its experienced staff members were transferred to other departments. Moreover, in 1997, in the context of an early retirement program, a number of experienced examiners opted to retire. As a consequence, the number of fully trained and experienced examiners declined from 37 in 1994 to 16 in 1997. Thus, the department is currently lacking the number of staff needed to examine the large number of financial institutions in the system as frequently as envisaged.

Portfolio quality and related issues

43. As of December 31, 1997, nonperforming loans in the 74 institutions supervised by the Bank Supervision Department amounted to K Sh 52 billion (20.5 percent of total loans),24 this total represents an average increase of 61 percent in each of the past two years.25 There are 19 banks (35 percent of the total) with nonperforming loans comprising more than 20 percent of their loan portfolio, and, for some, nonperforming loans reach as high as 87 percent of their total loans.

44. On an aggregate basis, financial institutions appear to possess adequate capital and reserves. The provision of doubtful loans amounted to K Sh 28 billion on December 31, 1997, (48 percent of nonperforming loans).26 If, however, the estimate of losses on the disposal of nonperforming loans is understated, additional losses would have to be recognized, negatively affecting the institutions’ capital and reserve accounts. The currently very small return garnered by the depositor protection fund on nonperforming advances it has received for disposition indicates that the provision for nonperforming advances may need to be adjusted upward. Such an adjustment would reduce, but not eliminate, the total equity in the system.27 Individual banks and nonbank financial institutions, however, could see their total equity eliminated if provisioning needs were adjusted upward. An additional risk arises from the lack of on-site examinations of many institutions, as the true state of their portfolios may be worse than reported and, thus, the current level of provisioning undertaken by these institutions may be insufficient to prevent an important decrease in their capital.

45. An important issue is the diversity of experiences among banks in Kenya in the areas of both portfolio performance and capital adequacy. A preliminary review indicates that even among Peer Group 1 banks there are significant differences. While certain banks have (report) only a small share of nonperforming loans, large parts of other banks’ portfolios are impaired. The potential additional exposure of these latter banks could significantly affect their capital. Additional effects on capital adequacy may be understated by inadequate reporting of the true state of the portfolios of certain banks that have not been subjected to on-site examinations for long periods of time.

Table 1.

Kenya: Gross Domestic Product by Origin at Constant Prices, 1991-96

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues; and Fund staff estimates.

Includes general government.

Table 2.

Kenya: Gross Domestic Product by Origin at Current Prices, 1991-96

(In millions of Kenya shillings)

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues; and Fund staff estimates.

Includes general government.

Table 3.

Kenya: Expenditure on Gross Domestic Product at Constant Prices, 1991-96

article image
Sources: Government of Kenya, StatisticalAbstract and EconomicSurvey, various issues; and Fund staff estimates.
Table 4.

Kenya: Expenditure on Gross Domestic Product at Current Prices, 1991-96

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues; and Fund staff estimates.
Table 5.

Kenya: Gross Domestic Product, GDP Deflator, Population, and Real per Capita Income, 1987-96

article image
Sources: Government of Kenya, Economic Survey, various issues; World Bank, World Development Indicators, various issues; and Fund staff estimates.
Table 6.

Kenya: Gross Fixed Capital Formation at Current Prices, 1991-96

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues; and Fund staff estimates.
Table 7.

Kenya: Sales of Agricultural Production to the Marketing Boards, 1991-96

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues; and Fund staff estimates.

Except pyrethrum, which is expressed in metric tons.

No purchases of rice paddy by the National Cereals and Produce Board in 1990.

Table 8.

Kenya: Value of Agricultural Production Sold to the Marketing Boards, 1991-96

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues; and Fund staff estimates.
Table 9.

Kenya: Average Prices to Producers For Selected Commodities, 1991-961

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues; and Fund staff estimates.

These prices are for calendar-year deliveries and reflect actual payouts, although average prices for two seasons that overlap during a calendar year may have differed. For coffee and tea, the prices are processed coffee and made tea, respectively.

Table 10.

Kenya: Quantity Index of Manufacturing Output, 1991-96

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues; and Fund staff estimates.
Table 11.

Kenya: Selected Statistics on Construction Activity, 1991-96

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues; and Fund staff estimates.
Table 12.

Kenya: Energy Supply-and-Demand Balances, 1991-96

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues; and Fund staff estimates.
Table 13.

Kenya: Employment by Industry and Sector, 1991-96

article image
Source: Government of Kenya, Statistical Abstract, various issues.