Selected Issues


Selected Issues

National Savings Certificates and Budget Financing1

A. Introduction

1. In recent years, the issuance of National Savings Certificates (NSCs) has increased rapidly and become the dominant instrument to finance government borrowing, repeatedly exceeding the budgeted amount by a large margin. The proliferation of the NSC issuance has complicated fiscal policy, debt management, and financial market development. If left unaddressed, this could lead to significant costs and distortions.

2. NSCs were introduced as a saving scheme for small savers, when the banking system was less developed. It was intended to motivate people to save money, collect scattered small savings, and provide a safety net for certain groups of population, including senior citizens, women, and physically handicapped. NSCs pay a higher interest rate than bank deposits with similar maturity. There are presently eleven different savings instruments with maturities of 3 or 5 years, available on demand and managed by the Directorate of National Savings under the Ministry of Finance (see Annex for details). To ensure that the scheme is used mainly by small savers, the ceiling for investment in most NSCs schemes is set at Tk. 3–5 million (about $37,000 – $61,000) for an individual, and Tk. 6 million (about $72,300) for a family.

B. Recent History of NSC Issuance

3. Following relatively modest issuances in the past, sales of the NSCs accelerated in FY15 and FY16, and then jumped further by almost 60 percent in FY17, reaching a record Tk. 524 billion (2.7 percent of GDP). This sharp acceleration was not envisaged by policy makers, as evidenced by the progressively widening gap between the budgeted and actual amount of NSCs financing (Figure 1). With actual deficits and external financing broadly in line with budget targets, domestic bank borrowing had to be reduced, which turned significantly negative in FY17. The government borrowed by issuing expensive NSCs to repay less expensive T-bonds and T-bills.

Figure 1.
Figure 1.

Budgeted and Actual Budget Domestic Financing 1/

Citation: IMF Staff Country Reports 2018, 159; 10.5089/9781484360347.002.A002

Source: Bangladesh Bank1/ For FY18, actual NSCs in H1 plus projected issuance in H2 equal to H2 FY17. FY18 bank financing is derived as residual, assuming that total domestic (bank and NSC) financing is equal to 75 percent of the budgeted amount (it was 60 percent in FY17).

4. The large NSC issuance has continued in FY18 as well. During the first half of FY18, the net sales of NSCs have reached to broadly the same level as during the first half of FY17. However, in the last three months (November 2017-January 2018), the net issuance has declined by about 10 percent compared to the same period of FY17 (Figure 2). The NSC issuance in FY18 will continue to exceed the budgeted amount by a sizable margin and this will result in a further reduction of bank borrowing, depending on the overall domestic financing requirements.

Figure 2.
Figure 2.

Monthly Sales of NSCs

Citation: IMF Staff Country Reports 2018, 159; 10.5089/9781484360347.002.A002

Source: Bangladesh Bank, Staff Estimates

5. The main reason for the recent sharp increase in the NSC issuance appears to be the large difference between the NSC rates and the bank deposit or government securities rates (Figure 3). Starting in FY14, inflation began to decline gradually, followed by a similar decline in bank deposit rates and T-bill rates. Given the relatively stable administratively determined nominal NSC rates (over 11 percent), this resulted in a larger spread between the NSC rates and market-determined rates, making the former more attractive to investors. More recently, deposit rates have begun to increase somewhat and the difference with the NSC rates have become slightly smaller, reflecting tighter liquidity conditions in the banking sector. This may have led to slower growth in the NSC issuance from November 2017.

Figure 3.
Figure 3.

Inflation and Yields

Citation: IMF Staff Country Reports 2018, 159; 10.5089/9781484360347.002.A002

Source: Bangladesh Bank, Staff Estimates1/ NSC rate is assumed to be the average of 11.04–11.76 percent

6. With a rapid increase in the NSC issuance, the stock of NSCs as a percent of GDP has doubled from 5 to 10 percent between FY13 and FY17. It has also reached almost one third of the total government debt, compared to 15 percent in FY13 (Figure 4). At the same time, the share of NSCs in total budget deficit financing jumped from close to zero in FY12-FY13 to 60 percent in FY17, and is likely to remain close to that level in FY18 as well.

Figure 4.
Figure 4.

NSCs, Budget Financing and Government Debt

Citation: IMF Staff Country Reports 2018, 159; 10.5089/9781484360347.002.A002

Source: Bangladesh Bank, Staff Estimates1/ NSC rate is assumed to be the average of 11.04–11.76 percent

7. The recent high level of NSC issuance has several negative implications. These include a higher cost of government borrowing, slower development of the government securities market and other financial markets, and the inability to plan and execute government borrowing and debt management strategy. Specifically: NSCs provide an expensive source of government financing. The NSC rate is well above the rate of government securities (Figure 3). As a result, the cost of government debt is increasing with higher interest payments on NSCs both in nominal terms and as a share of domestic debt interest payments (Figure 5). While increased NSC issuance has not yet had a significant impact on public debt sustainability, this could change if the current high issuance persist for an extended period.

  • NSC interest rates now significantly exceed those on bank deposits and this serves as an indirect source of subsidy to households. If the NSC holders include investors not originally targeted, this would imply unwanted and wasteful subsidies to wealthy individuals or households.

  • NSC instruments are supplied on demand and the government is unable to control the amount of financing it receives from NSCs. This makes it more difficult to plan the issuance of T-bills and T-bonds, hampering the development of the government securities market. In fact, the outstanding amount of T-bills and T-bonds peaked in 2014 and has begun to decline in nominal terms in 2015. With the NSC issuance repeatedly exceeding the budgeted amount, the government had to cancel T-bill and T-bond auctions, which has adversely impacted the liquidity of secondary markets in government securities.

  • The failure to develop a well-functioning liquid market in government securities is hampering the development of the broader financial market, including the market for corporate bonds and the equity market.

  • With longer-term savings channeled into the NSC instruments, less savings is available to finance private investments, adversely affecting the long-term growth performance.2

  • The absence of a well-functioning government securities market is also adversely impacting the efficiency of monetary policy transmission, and thus ability of Bangladesh Bank (BB) to use monetary policy to pursue its objectives.

  • The diversion of savings into NSCs also has adverse implications for the financial health of the banking sector. While there are other reasons for the deteriorating financial performance of deposit-taking banks, in particular, state-owned commercial banks, the shrinking supply of government securities is likely to have contributed to this deterioration, as it implies reduced availability of safe assets, which in turn could push banks to extend more risky loans.

Figure 5.
Figure 5.

Government Debt and Interest Payments

Citation: IMF Staff Country Reports 2018, 159; 10.5089/9781484360347.002.A002

Source: Ministry of Finance, Staff Estimates

8. Private credit growth has been strong despite with the recent slowdown in deposit growth due to higher NSC issuance (Figure 6). The difference between private credit and deposit growth began to widen around mid-2015, when the NSC issuance accelerated. Negative government borrowing and banks’ excess liquidity from a strong balance of payment position allowed banks to maintain rapid increases in private credit until recently. However, excess liquidity fell from Tk. 1.26 billion in September 2016 to Tk. 0.867 billion in December 2017 because of a weaker external position and strong private sector demand for loans. With the reduction of the maximum advances-to-deposit ratio by BB, deposit and lending rates have started to increase recently.

Figure 6.
Figure 6.

Banking Sector Deposits and Credit

Citation: IMF Staff Country Reports 2018, 159; 10.5089/9781484360347.002.A002

Source: Ministry of Finance, Staff Estimates1/ Nominal growth deflated by increase in CPI over previous twelve months.

C. Alternative Interest Cost Scenario with Continued High NSC Issuance

9. Using more expensive NSC financing to repay less costly market borrowing will result in higher interest payments and higher fiscal deficit and debt. While these additional borrowing costs do not appear to be significant yet, they will accumulate over time, and limit the government’s capacity to finance other spending without running a higher fiscal deficit and weakening fiscal sustainability. The FY17 data illustrates the magnitude of increased borrowing costs from higher-than-budgeted NSC issuance.3

10. Additional borrowing costs from higher NSC issuance can be estimated by using the difference between actual FY17 NSC issuance and its hypothetical issuance with the budgeted share of NSC in domestic deficit financing. This enables to separate the effect from different level of domestic finance needs, and focus solely on the effect of changes in the NSC/bank financing share. The average cost of NSC borrowing is assumed at 11.4 percent, a mid-rate for the NSC rates (11.1–11.7 percent). Further, it is assumed that the reduction in bank financing (purchases of government securities) is split 50:50 between T-bills and T-bonds, and that the T-bill rate and T-bond rate would be 4 and 8 percent, respectively.

Table 1.

Bangladesh: FY17 Additional Borrowing Costs from NSC Issuance

(In Billions of Taka)

article image

11. Under these assumptions, additional NSC issuances (Tk. 341.6 billion) implies additional interest costs for the financing of the FY17 deficit of Tk. 18.4 billion (around 0.1 percent of GDP). This is derived as the difference between additional NSC interest costs of Tk. 38.9 billion, and lower interest costs on T-bills (Tk. 6.8 billion) and T-bonds (Tk. 13.7 billion), reflecting lower issuance of these government securities compared to the issuance that would have been achieved with the actual deficit and the budgeted share in the total domestic financing. Using the same approach, additional net interest payments resulting from higher share of NSCs during FY14 and FY17 amount to Tk. 43 billion, or about ¼ percent of FY17 GDP.

Table 2.

Bangladesh: Interest Costs from Additional Issuance

article image

12. Continued NSC issuance at the present level would imply a notable increase in interest payments over the medium term. The medium-term implications for interest payments on government domestic debt can be estimated by using the same assumptions for the interest costs of NSCs and market securities as described in the previous paragraph. Figure 7 shows two paths of interest payments on domestic government debt as a percent of GDP. The baseline scenario assumes a gradual decline in the NSC issuance as a percent of GDP from 2.7 percent of GDP in FY17 to close to 1 percent of GDP. The alternative scenario assumes that the issuance will remain twice as high as the baseline (as a percentage of GDP) during the projection period, close to the recent peak. Over the medium-term, the difference between the interest payments in these two scenarios becomes more pronounced: while interest payments remain broadly stable as a percent of GDP in the baseline scenario, under the alternative scenario, interest payments continue to increase gradually to about 2.2 percent of GDP by 2036, 0.5 percentage points higher than in the baseline scenario. 4

Figure 7.
Figure 7.

Interest Payments on Domestic Government Debt

(In percent of GDP)

Citation: IMF Staff Country Reports 2018, 159; 10.5089/9781484360347.002.A002

Source: Staff estimates

D. Possible Measures to Address the NSC Problem

13. Addressing the NSC problem in a relatively speedy and comprehensive manner should remain an important policy priority. The above projection of interest payments illustrates that the high NSC issuance could, in the medium term, have tangible impacts on debt sustainability. The actual impact on debt dynamics could be worse if the possible adverse implications of the higher NSC issuance on investments and growth are considered. As discussed above, the current NSCs scheme could – if continued – carry significant costs to the economy that would undermine the authorities’ efforts to sustain high growth and reach middle-income status.

14. The authorities should start with establishing a database that records the existing holdings of NSCs.5 The monitoring of the investment ceilings could aim: investors with excessive holdings would have to bring them below the ceilings before becoming eligible for new purchases; or they may become ineligible to collect the interest on the NSC holdings above the applicable ceilings. Consideration should also be given to whether the current eligibility criteria, the investment ceilings, and the number of instruments still serve the original purpose. Presently, almost all citizens of Bangladesh are eligible to buy some NSCs (see Annex for details). Thus, even rich people can invest in the scheme and receive an implicit subsidy. While excluding persons above a certain income threshold would reduce this untargeted subsidy, it could come with increased monitoring costs and administrative challenges.

15. Currently, little information is available to assess the extent of misuse in the NSC scheme. In 2013, BB released a survey report about the NSC scheme. The findings indicated that most of the NSC investors were of modest income, and that their average invested amount was fairly small.6 This suggests that the scheme was successful in targeting the intended group. However, since 2013, NSC issuance has increased significantly, raising the question whether the conclusion about accurate targeting and the absence of widespread misuse remains valid. Unless the number of NSC investors has increased dramatically since the time of the survey, the size of the NSC issuance increases (from Tk. 7.7 billion in FY13 to Tk. 516 billion in FY17) would suggest possibly a significant breach of the investment limit of Tk. 3 million for an individual (or Tk. 6 million for a family). Thus, it can be expected that effective monitoring of the NSC scheme and enforcing corrective measures could result in the reducing the issuance from the recent elevated levels.

16. However, a more resolute solution may be necessary to reduce distortion and costs from the current NSC scheme. The demand for NSCs will be strong as long as their relative attractiveness over other investment facilities remains and their issuance will continue to be outside of the control of the government. Therefore, more far-reaching reform of the NSC scheme would entail:

  • Changing the pricing policy by linking the NSC rates closer to the market rates (bank deposit rates, T-bill and T-bond rates), which will reduce the current high spread. To the extent that the original objective of the NSC scheme – to provide a means of savings to small savers – remains valid, the scheme with a reduced number of instruments could still be maintained. However, this does not require maintaining the large subsidy element in the current high NSC rates.

  • Delinking budget financing from providing income support to the poor and vulnerable population. Instead, such support should be provided as part of the budget spending in the form of direct transfers to the targeted recipients. For example, expanded pension coverage would reduce the need to use NSCs to provide support to the elderly. That would allow the authorities to regain control over budget financing.

17. A radical solution would be to completely abolish the NSC scheme, or to stop the issuance of the NSCs once the annually budgeted amount has been reached.7 However, completely abolishing NSCs is not necessary to contain the problems of the current scheme. And limiting the issuance to the budgeted amount could create new problems. First, the existence of the limit on issuance could significantly accelerate the issuance as everyone seeks to secure their purchases. Second, once the issuance stops, the government would have to switch to other means of borrowing. As government borrowing usually increases in the final months of the fiscal year, that could force a volatile schedule in issuance of T-bills and t-bonds, possibly resulting in a large volatility in yields which, in turn, would limit the potential benefits of this approach in terms of lower interest costs.

E. Conclusion

18. In recent years, the NSC issuance has accelerated rapidly, creating various side-effects. The increasing role of NSCs in government borrowing has led to higher interest costs, while the unpredictability of the “on tap” NSC issuance has complicated budget financing and debt management strategy, which has hampered the development of capital market. The question remains on whether the original objectives of the NSCs are still pertinent and whether there are more effective ways to achieve them.

19. Reforming the NSC scheme has become a matter of urgency. The near-term steps could include: (i) creating the database of the NSC subscribers to increase the monitoring capacity and reduce the scope for misuse by breaching the access limits to NSCs; (ii) reducing the number of savings instruments to simplify the system and further increase the capacity to monitor NSCs; and (iii) limit the access to NSCs to retail investors. In the medium to long term, the NSC scheme should be separated from budget financing, and the NSC rates should be linked to market interest rates, to reflect the new role of NSCs as a savings instrument for targeted population segments rather than a tool for subsidy provision.

20. Income support to the vulnerable population should be done directly from the budget. Such a mechanism would be more efficient, better targeted, and more transparent, with appropriate political control and approval. Under the present system, this support is not well targeted, and the size of indirect subsidies extended from the budget (in the form of higher interest payments) is outside of political control and approval.

Annex I. National Savings Certificates – Overview

article image

National Savings Schemes and the Outstanding Stock

(Billion BDT, outstanding stock as on end the period) FY16

article image

Annex II. Pakistan’s Reform of Savings Certificates

1. In the past, Pakistan faced similar problems with the national savings scheme that currently exists in Bangladesh. There were many saving instruments, with pricing delinked from the benchmark instruments, and it was difficult to control and project the issuance. During the 2000s, Pakistan undertook a reform of the scheme, successfully addressing many of the issues. While the implementation of the reform has proceeded gradually, it has been done comprehensively, covering simultaneously different fronts.

  • Pricing mechanism and benchmarking. Under the previous system, the pricing of the savings instruments was delinked from the benchmark government securities, creating room for arbitrage and fragmenting the financial market. To correct this, the return on the savings certificates were linked to the yield of the government security (T-bills and T-bonds) of the corresponding maturity (the adjustment of return does not have to be made in one jump, and can be done in several steps). For example, the 6-month saving certificate rate was based on the 6-month T-bill rate from the latest auction. This reform brought several benefits. It implied reduced borrowing costs to the government, and removed the source of arbitrage. All savings instruments and other securities were now linked via the new pricing mechanism, which improved the liquidity of the government securities market, broadened the investor base, and strengthened the monetary policy transmission mechanism. Initially, the rates on savings certificates were set equal to the auction yields on government securities, but later, a 50-basis points margin was added.

  • Eligibility. Originally, there was no limit to access: all types of investors could invest in saving certificates, including individuals, groups, companies, pension funds etc. The stated objective was to provide secure savings instruments to any investor. Following the reforms, the access was limited to individuals only, with the savvier institutional investors assumed to invest in the wholesale market. At the same time, individual investors were allowed access to T-bond and T-bill auctions via noncompetitive bidding (with limits of 5–10 percent of the issuance).

  • Reduction of the number of savings instruments. Originally, there were over 30 different instruments, often several instruments with the same maturity, resulting in fragmentation and low liquidity of individual issuance. Thus, the number of instruments was reduced to eight, which reduced the need for monitoring and simplified the scheme’s implementation. In addition, to increase liquidity, some less liquid maturities were abolished. The savings certificates currently come in two forms, prize bonds and coupon-bearing instruments.

  • Limits on purchase. The effective removal of room for arbitrage between the saving certificates and government securities obviated the need for monitoring and limiting access. Thus, savings certificates are treated as any other security. The only exception is a special certificate for retirees (person over 60 years old) and widows, which can purchase up to 5 million rupees (around US$45,000) of special certificates that provide an additional return of 150 basis points over the benchmark return. The use of national identification cards and death certificates made controlling the access to these instruments straightforward.

  • Issuance. Commercial banks can issue the savings certificates on behalf of the government, and are paid a commission for this service. This also allows the banks to keep the investors as their customers, because the investors are using the bank accounts, and thus could advertise their services. The National Savings Office, attached to the Department of Finance in the Ministry of Finance, is in charge of managing the issuance of savings certificates. The ad hoc (“on tap”) issuance of savings certificates can take place any time, regardless of the timing of the T-bill and T-bond auctions, and the minimum investment requirement could be calibrated to meet the specific needs of the retail investors.

  • Tax treatment. The tax treatment of saving certificates and other securities has been aligned. For registered taxpayers (holders of tax identification number), a 10 percent withholding tax is applied on interest income, while non-registered taxpayers are subject to a 20 percent tax rate.

  • Debt management strategy. One of the problems caused by financial instruments not linked via a pricing mechanism to other securities is inability to predict their demand and issuance with reasonable accuracy. This inability seriously complicates the debt management strategy and the annual budget financing planning. Linking all financing instruments via a pricing mechanism and benchmarking makes it easier for the authorities to predict demands with a reasonable accuracy. Given the absence of differences in pricing and tax treatment, government securities and savings certificates become close substitutes which improves the liquidity of both markets, and thus the attractiveness of both instruments to investors. In turn, this improves the ability of the authorities to formulate and execute budget financing and debt management strategies.

2. Figure A1 highlights several contrasting developments between the national savings scheme in Pakistan and Bangladesh.

  • In Pakistan, returns on the savings certificates have been broadly declining in real terms (calculated as nominal return per annum adjusted for the CPI). CPI inflation fell from over 10 percent in 2011 to below 6 percent in 2017, and the nominal returns on savings certificates have been broadly following the decline in inflation. In Bangladesh, the rigidity in nominal returns, in combination with a similar decline in inflation, has resulted in a sharp increase in the real returns of the NSCs, which in turn has triggered the significant increases in the NSC issuance. In Pakistan, linking the returns of the savings certificates to the benchmark securities did not open the difference in returns, and the corresponding scope for arbitrage.

  • In Bangladesh, the stock of savings instruments has been growing rapidly in the last five years, both as a percent of total government debt and as a share of GDP. In Pakistan, the outstanding stock of savings instruments has been relatively stable (as a percent of GDP) during the last decade, and it has fallen from about one half to about one fourth of the domestic public debt stock. While the share of NSCs in financing the budget has increased sharply in recent years in Bangladesh, it has fallen markedly in Pakistan, and it is now less than one third of the share in Bangladesh (Figure A2).

Figure A1.
Figure A1.

Pakistan: National Savings Scheme

Citation: IMF Staff Country Reports 2018, 159; 10.5089/9781484360347.002.A002

Source: State Bank of Pakistan, National Savings Pakistan, CEIC, Staff Estimates1/ Nominal return adjusted for annual CPI
Figure A2.
Figure A2.

Budget Financing by Savings Certificates

(In percent of total budget financing)

Citation: IMF Staff Country Reports 2018, 159; 10.5089/9781484360347.002.A002

Sources: Bangladesh Bank, National Savings Pakistan, staff estimates


  • Bangladesh Bank, 2013. “Sample Survey on Socio-Economic Characteristics of Buyers of Saving Instrument in Bangladesh”, Dhaka.

  • Mansur, A., 2017.“Reform of NSD Instruments – How long Will the Government Wait?”, The Financial Express, November 30, 2017, Dhaka.

    • Search Google Scholar
    • Export Citation

Prepared by Jiri Jonas.


Strictly speaking, this problem only arises with NSC purchases by those who are not originally targeted recipients. Channeling the same amount of funds as a means of income support to the targeted groups via NSCs or direct budgetary transfers should not have significantly different impact on government savings.


Another way to look at the costs of NSCs would be to compare the evolution of interest payments on NSCs over time – both in nominal terms and as a percent of GDP and of total interest payments – see Figure 5.


This illustration does not capture the effect of higher interest payments in the alternative scenario on the deficit and the overall borrowing needs. Accounting for this would further add to interest payments under the alternative scenario.


As discussed in the Staff Report, the authorities are planning to launch such database, using the National Identification Number Database.


Bangladesh Bank (2013). According to the survey, the average monthly family income of the investors in the NSC scheme was Tk. 28,700 (about US$360). About 55 percent of investors in NSCs invested less than Tk. 0.5 million (less than $6,250).


Mansur (2017) suggested that one option to address the NSC problem would be to completely stop selling new NSC instruments. The argument is that with modern banking system in place accessible to most people, there is little justification for the NSC scheme originating at the time when banking system was underdeveloped. Instead, the savings could be directed from NSCs to the banking system and capital market.

Bangladesh: Selected Issues
Author: International Monetary Fund. Asia and Pacific Dept