Report No. 81686-KZ Kazakhstan OIL RULES: Kazakhstan’s Policy Options in a Downturn Joint Economic Research Program May 21, 2013 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank CURRENCY AND EQUIVALENTS (As of June 3, 2013) Currency Unit = Tenge $1 = 151.39 WEIGHTS AND MEASURES (metric system) Vice President Philippe Le Houérou Country Director Saroj Kumar Jha Country Manager Sebnem Akkaya Sector Director Yvonne Tsikata Sector Manager Ivailo Izvorski Task Team Leader Francisco Galrao Carneiro Contents Acknowledgments v Executive Summary vii Chapter 1 Kazakhstan’s Economic Policy Stance and the Global Financial Crisis 1 Before the crisis 1 Policy responses to the crisis 3 The banking crisis in international perspective 4 The lessons from the crisis 5 Chapter 2 Fiscal Rules Should Be Simple and Transparent 9 Main features of the Kazakh fiscal rules 9 Fiscal rules in Chile and Norway 11 A bird-in-hand? 12 The Kazakh fiscal rule and the Kopits and Symansky criteria 13 “If it ain’t broke, don’t fix it” 15 Chapter 3 How a Changing World Would Affect a Changing Kazakhstan 17 The baseline scenario 17 Three extreme negative shocks 20 Scenarios for optimal resource management 23 Chapter 4 Evaluating the Different Fiscal Rules 28 Forecasting problems 28 Flexible rules perform worse than fixed rules 29 Simulating welfare gains from different fiscal rules 30 Chapter 5 Theoretical Considerations for Countercyclical Policy 35 Fiscal rules and countercyclical policy 35 The role of multipliers 36 The need for financial sector development 38 Appendix 1 Details of the National Fund of the Republic of Kazakhstan 41 iii OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN Appendix 2 Survey on International Monetary Fund Advice to a Sample of Resource-Dependent Countries 42 Appendix 3 Fiscal Rules in Different Countries 43 Appendix 4 Economic and Social Indicators, 2005–15 44 Appendix 5 A One-Time 3 Percent Shock in the Mineral Sector 46 iv Appendix 6 A One-Time –6.4 Percent Shock in the Mineral Sector 48 Notes 50 References 51 Boxes 1.1 Description of the social safety net programs 7 4.1 Current practice in other resource-exporting countries 31 Figures 1.1 Kazakhstan real GDP, 1999–2012 2 1.2 National oil fund reserves, 2002–12 2 1.3 Expansion of credit to the nonfinancial sector and external indebtedness of Kazakh banks, 2000–12 3 1.4 Composition of real GDP growth by sector, 2000–12 3 2.1 The real exchange rate was kept almost flat, despite some fluctuations in oil prices 10 2.2 Oil revenue savings exceeded 32 percent of GDP, the biggest fiscal buffer ever 10 2.3 Permanent income-based annuity and projected revenue profile for Kazakhstan 11 3.1 Real price of oil to stabilize in the long run 18 3.2 Strongest growth in oil demand to come from developing countries 18 3.3 The baseline forecast suggests continuing (but declining) reliance on the oil sector 19 3.4 Assumptions for long-term simulations 25 3.5 Macroeconomic projections under different oil price scenarios 26 4.1 Predicted output gap and final output gap for 175 countries, 1990–2011 29 5.1 Nonperforming loans are well provisioned but constrain banks’ lending 39 5.2 BTA remains the biggest outlier in nonperforming loans 39 A5.1 Oil and nonoil GDP targeting assuming a 3 percent growth shock in 2014 46 A6.1 Oil and nonoil GDP targeting assuming a –6.4 percent growth shock in 2014 48 Tables 2.1 Permanent income annuity values of oil 11 3.1 Impact of Euro Area crisis 21 3.2 Impact of U.S. fiscal paralysis 22 3.3 Impact of an abrupt fall in Chinese investment 24 4.1 Simple rules 32 4.2 Income and expenditure shares, 2007 33 5.1 Structure of expenditure of the representative household, 2005 37 5.2 Imports and domestic sales, 2002 38 A5.1 Outcomes under fixed and short-term fiscal rule scenarios assuming a 3 percent growth shock in 2014 47 A6.1 Outcomes under fixed and short-term fiscal rule scenarios assuming a –6.4 percent growth shock in 2014 49 Acknowledgments This report was written by a team led by Fran- of Kazakhstan; and representatives from the cisco Carneiro, Lead Economist and Sector Economic Research Institute. A consultation Leader for Poverty Reduction and Economic event held in Astana on June 12, 2013, with gov- Management for the countries of Central Asia ernment officials as well as written comments at the World Bank. The current volume sum- received from staff of the Minister of Economy marizes the main findings and policy rec- and Budget Planning provided additional ommendations. Background papers for this insights to the team. report were prepared by Eduardo Engel (Yale Ilyas Sarsenov and Dorsati Madani were University and National Bureau for Economic part of the core team. Sarah Nankya Babirye Research), Eduardo Ley and Harun Onder provided editorial assistance. Gulmira Akshaty- (PRMED), Andrew Burns, Theo Janse van rova and Xeniya Novozhilova from the Astana Rensburg, and Trung Thanh Bui (DECPG), country office and Oxana Shmidt from the Constantino Hevia (DECMG), and Steven Kyle Almaty country office supported the team dur- (Cornell University). ing the several missions associated with this The team is grateful to the government of task. Jorge Araujo provided peer review com- Kazakhstan for its support to this work since ments and guidance to the team. The work was its early stages. The high-level brainstorming carried out under the overall supervision of held in Astana on March 16, 2013, with senior Ivailo Izvorski, Sector Manager, Poverty Reduc- Kazakh authorities provided useful insights. tion and Economic Management Sector Unit: The team also benefited immensely from early Macro Economics 1 (ECSP1) and Yvonne Tsi- interactions and guidance received from Yer- kata, Sector Director, Poverty Reduction and bol Orynbayev, Deputy Prime Minister; Kairat Economic Management Sector Unit (ECSPE). Kelimbetov, Deputy Prime Minister; Madina Saroj Kumar Jha, Country Director for Central Abylkassymova, Vice-Minister of Economy and Asia, and Sebnem Akkaya, Country Manager Budget Planning; government officials from for Kazakhstan, provided guidance and sup- the Ministry of Finance and the National Bank ported the team. v Executive Summary This report assesses the role of fiscal rules varying annual revenues. Transitory changes for countercyclical economic management in to resource-related revenues caused by fluc- Kazakhstan and simulates the behavior of dif- tuations in commodity prices or temporary ferent types of fiscal rules. The work represents changes to production are not relevant; per- technical assistance to the Ministry of Economy manent changes are. By keeping the annual and Budget Planning and has been financed by amount transferred from the fund to the bud- the Joint Economic Research Program between get constant, Kazakhstan’s rule is more strin- the government of Kazakhstan and the World gent than pure application of the permanent Bank. It responds to government demand to rule would imply. Under the latter, permanent assess the following four topics, in line with the changes to prices, production, or other relevant terms of reference agreed between the World factors would result in changes to the trans- Bank and the authorities: ferred amount. The authorities have expressed • Assessing the cyclicality of the govern- interest in learning more about the experience ment’s economic policy stance over the last of other countries with the use of fiscal rules 10 years and identifying lessons from this and the scope for fiscal rules in countercyclical experience. policy. • Exploring how different global economic This report’s five chapters outline the scenarios could affect key variables in the advantages of Kazakhstan’s current fiscal rule Kazakh economy. and its performance against other types of • Assessing the expected impacts of alterna- rules. Chapter 1 reviews the performance of tive fiscal rules on key economic variables. the Kazakh economy and the policy stance • Reviewing the literature on counter- of the authorities over the last 10 years. The cylical fiscal policy and its relevance for analysis focuses on the policies that worked Kazakhstan. well before and after the recent crisis—and on Kazakhstan adopted a fixed fiscal rule in vulnerabilities that have yet to be addressed. A 2010 and introduced changes to it in 2012. few good-practice principles on dealing with The current rule is simple and easy to imple- severe recessions are presented as lessons for ment. Its main feature is that each year the the future. Chapter 2 argues that fiscal rules National Fund of the Republic of Kazakhstan should be simple and transparent, presenting (national oil fund) transfers the equivalent to evidence that Kazakhstan’s current fiscal rule $8  billion to the budget. This rule is similar meets these two criteria. Chapter 3 presents in scope to the permanent income approach, the results of simulations of different global where government spending out of resource economic scenarios and the expected impacts revenues is based on the net present value of its on key variables in Kazakhstan, considering estimated wealth and not on the government’s scenarios as elected in close consultation with vii OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN the Kazakh authorities. Chapter 4 uses the account moved from a surplus of $3.7  billion results of these simulations to test the perfor- in the first half of 2008 to a deficit of $3.5 bil- mance of different types of fiscal rules. Chap- lion in the first half of 2009. The fiscal posi- ter 5 responds to a government request to tion deteriorated sharply, with the nonoil fiscal review the literature on the use of fiscal and deficit widening to 10.5  percent of GDP in monetary policy for countercyclical economic 2008 before peaking at 13.8 percent in 2009. management. Together with a weak capital account, this placed the tenge under significant pressure, viii Kazakhstan’s economic policy stance leading the authorities to devalue the currency and the global financial crisis by 20  percent in February 2009. Commercial Kazakhstan grew rapidly in the first half of the banks also lost virtually all access to rollover 2000s, following a conservative fiscal policy opportunities in global financial markets. that allowed the buildup of sizable fiscal buf- In response, the authorities implemented a fers. Annual gross domestic product (GDP) fiscal stimulus program of $21 billion. About growth averaged 10 percent, and average real $11 billion of the funds were devoted to direct incomes more than doubled, transforming assistance to struggling commercial banks Kazakhstan into an upper middle-income (primarily BTA), while the rest financed anti- economy in a short time. Despite large oil- crisis stimulus measures in various sectors of related revenues, the fiscal stance remained the economy. As part of the anticrisis program, prudent. Total spending averaged 22  percent the government protected social expenditures of GDP for most of the precrisis period, with from budgetary sequestration, insured all pen- the nonoil deficit averaging 3 percent of GDP, sion payments indexed to inflation (including even as oil prices and revenues rose. Fiscal pru- those from private pension funds), extended dence allowed the authorities to accumulate unemployment benefits from four months to significant savings in the national oil fund, six, and promoted the signing of memoranda which reached $27 billion by the end of 2008, between regional authorities and larger enter- equivalent to 21 percent of GDP. prises to preserve jobs. Massive capital inflows complicated mac- Stepped-up liquidity support for com- roeconomic management, limiting scope for mercial banks and a substantive increase in a decline in inflation, boosting asset prices, deposit insurance prevented the collapse of and planting the seeds of crisis in Kazakhstan. the banking sector. By March 2009, reserve These inflows were related to soaring real requirements for commercial banks had been estate prices in Kazakhstan’s larger cities and progressively reduced from 6.0  percent to a related boom in housing construction. Inter- 1.5  percent for internal liabilities, and from national creditors supplied larger (private) 8.0 percent to 2.5 percent for external liabili- commercial banks with abundant unsecured ties. The $11  billion that the authorities pro- credit, expecting that the government would vided in direct equity and debt financing use the national oil fund to prevent failures of (10 percent of GDP) helped stabilize credit to these large, systemically important banks. This the private sector. The interventions succeeded expectation, combined with soaring commod- in maintaining depositor confidence in the ity and real estate prices and limited banking banking system. regulation and supervision, set the stage for There are five lessons from the global finan- a severe financial bubble. Most of this credit cial crisis. First, the crisis challenged the belief flowed to construction and real estate, and that stable and low inflation was the primary— the share of such lending amounted to 30 per- and many times the exclusive—mandate of cent of GDP at the end of 2007. As a result, the central banks. Second, it forced central banks finance and construction sectors became the to rethink the conventional principle that major drivers of economic growth. monetary policy should be focused on a single The onset of the global financial and eco- instrument: the policy interest rate. Third, it nomic crisis hit the Kazakh economy severely. highlighted the limitations of monetary policy Export revenues declined, and the current and underscored the role of fiscal policy as an effective tool for macroeconomic manage- the upper end of the specified range. Mean- ment in a severe downturn. Fourth, it changed while, the complementary provisions to prevent perceptions of the role of financial regulation, government borrowing from “undoing” the which was largely ignored as a macroeconomic fund’s savings remain unchanged. policy tool before 2008 in most countries. And Kazakhstan’s fiscal rule performs well fifth, given that discretionary fiscal measures against other types of rules. Several criteria can may come too late to fight a standard recession, be used to assess the robustness of a given type the crisis strengthened the need for improving of fiscal rule, and this report considers those automatic stabilizers. put forward by Kopits and Symansky (1998) to ix For Kazakhstan, automatic stabilizers can test how well the current set of rules in Kazakh- be a useful complement to monetary and fis- stan performs from a normative point of view. cal policies in a downturn. As discussed above, These criteria include several aspects that monetary policy may not be effective in a would be expected in any given fiscal rule, but severe economic downturn, and fiscal mea- the most important is that fiscal rules should sures may yield positive results only with a lag. be simple and transparent for reasons of cred- Automatic stabilizers do not suffer from these ibility and enforceability. Kazakhstan does well shortcomings. With large fiscal stabilizers, the on both criteria—and on many others. impacts in the economy can be felt in a timely A comparison with the fiscal rules used in and gradual fashion as taxes and expenditures Chile and Norway highlights differences with react in a countercyclical manner to changing the approach used in Kazakhstan. These differ- economic conditions. In addition, automatic ences do not, however, mean that the Kazakh stabilizers minimize both the interference of rule is inferior. Chile’s fiscal policy is anchored political decisions and implementation lags in on a structural balance target that depends on the response to the shock. Automaticity also the long-term price of copper—while Norway guarantees a timely reversal of a fiscal expan- follows a complex structural deficit rule that sion as the fiscal loosening during bad times is depends on the nonoil deficit. While these automatically corrected by tightening in good types of rules have desirable properties in the- times. ory, they are very hard to implement and are not entirely transparent. And if they are too Fiscal rules should be simple difficult to implement, they can undermine and transparent the government’s credibility in the markets. In the aftermath of the crisis, Kazakhstan The Kazakh rule is simple, effective, and trans- adopted a fixed fiscal rule to govern the trans- parent—and has been able to deal satisfactorily fers from the national oil fund to the budget. with transitory shocks. The fiscal rule rested on two principles. The The bird-in-hand rule would not be a first set the annual transfer from the fund to good replacement for the current Kazakh fis- the budget at $8 billion in nominal terms (with cal rule, either. It sets the amount of transfer the exchange rate for setting the amount in from the oil fund to the budget equal to the tenge fixed over a three-year horizon). The investment returns from the oil fund similarly second limited the interest cost of the public to what is currently implemented in Norway. debt to 4.5  percent of the fund balance. The Applying the bird-in-hand rule over the life national oil fund was modified by the govern- of the extraction cycle implies higher savings ment’s Presidential Decree 289 (dated March and lower transfers to the budget in the ear- 16, 2012), changing the transfer of funds to the lier years, because the driving principle is the budget from the previously fixed $8 billion to uncertainty about future revenues. Such an the flexible amount of $8 billion plus or minus approach is appropriate for mature producers 15  percent ($6.8–$9.2  billion), depending on such as Norway, with large accumulated finan- the cyclical position of the economy. These cial reserves and an oil extraction cycle close adjustments will remain valid until the end of to the end. Current projections for Kazakh- 2013, and the budget will receive a transfer of stan show that the annual transfers from the $9.2  billion, keeping the transfers pegged at oil fund to the budget are expected to surpass OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN its earned investment income only by 2025. So, and the structure of expenditure programs adopting this type of rule in Kazakhstan could (by doing such things as making poverty significantly reduce the nominal amount trans- or other assistance programs conditional ferred to the budget until at least 2025. on income or employment status) to make A fixed transfer rule with an escape clause countercyclical revenue and expenditure allows nearly as much gain in welfare as does a changes as automatic as possible. theoretically optimal rule. That is, a fixed rule with clear criteria to allow for a deviation from How a changing world would x the fixed amount that is supposed to be trans- affect a changing Kazakhstan ferred to the budget every year for a limited Living standards in the long term crucially period (the escape clause) performs as well as depend on the pace of productivity growth, an optimal (and very complex) rule. The coun- the ability of the economy to create jobs, and tercyclical policy stance adopted to cushion the absence of large swings in economic activ- the effects of the 2008–09 economic crisis was ity. Whether productivity growth averages a de facto implementation of such a rule. For 2.0 percent a year or 2.5 percent makes a huge the future, this approach could be codified. difference—under the former assumption, per But it would be important to determine the capita GDP in Kazakhstan could reach as much threshold for triggering an escape clause and as $100,000 by 2050; under the latter, $130,000. the number of years such an alternative regime Larger extraction of natural resources and would remain in effect before assuming that spending by the government could lead to a the shock is chronic rather than transitory. pace of capital accumulation that may offset The main recommendations for dealing weak productivity growth for a while. But in the with transitory shocks within the framework of absence of productivity growth the sustainabil- the current fiscal rule can be summed up as ity of that high income will be challenging, and follows: job creation can come only at a high cost. Ulti- • First, consider raising the $8 billion base amount, mately, the pace of productivity growth in an which has not been revised since the introduction economy on the cusp of high-income status will of the rule in 2010. Applying the growth of depend on the ability of companies to inno- nonoil GDP since 2010 to this amount would vate, which will require a well-educated and adjust the base amount to about $11 billion. highly skilled labor force, sound institutions This should preserve the nondistortionary that guarantee property rights, and a good characteristics of the original $8  billion investment climate. fixed amount while allowing for the greater The world economy has been through a tur- absorptive capacity of the now-larger domes- bulent period, and Kazakhstan can be affected tic economy. For the future, periodic adjust- by different exogenous shocks. While there is a ment of the base might be desirable. The multitude of such potential shocks, three were adjustments could be made based on trend chosen to illustrate the possibility for external growth in nonoil GDP, something that could events to seriously affect Kazakhstan’s eco- be calculated only over longer periods—at nomic performance: a serious Euro Area crisis least four to five years, if not longer. resulting from a financial sector breakdown, a • Second, add two limiting factors for implementing United States–centered shock stemming from 15 percent deviations from the rule. One would fiscal paralysis in the U.S. government, and a be to specify the size of the shock that would China-centered shock stemming from a col- trigger such a deviation; the other would lapse of Chinese investment and thus demand impose a time limit for how long such addi- for Kazakh exports. These scenarios, chosen tional transfers could be used (such as two by the authorities, are aimed to illustrate how years). their materialization in the short term could • Third, implement “automatic stabilizers” to the possibly affect the Kazakh economy over the extent possible. That is, use the structure of short and medium terms. the tax code (by doing such things as mak- Over the medium term, the effects of these ing taxes conditional on income or profits) three extreme scenarios on Kazakhstan would be differentiated. The detailed simulations the outcome of policy reforms to improve the presented in the background paper on differ- quality of education in the country, the busi- ent scenarios for the world economy show that ness environment, and the institutions that the worst impact would be felt in Kazakhstan regulate economic activity and the delivery of if Chinese investments were to collapse. In the public services. That would help, much along event of a 10 percentage point drop in Chinese the lines of the recommendations of the recent investments, China’s GDP would slow down World Bank Country Economic Memorandum, by 3  percentage points. In the World Bank’s Beyond Oil: Kazakhstan’s Path to Greater Prosperity global macroeconomic model, this would hurt through Diversifying. If such productivity growth xi the world economy and cause spillover effects were to materialize, Kazakhstan’s income per that could hit Kazakhstan through a reduction capita would rise to the level of the 30 richest in the demand for oil and other commodities. countries by 2050. As a result, Kazakhstan’s share of oil exports would drop from an estimated 57  percent in Evaluating the different fiscal rules the baseline estimate for 2010 to 38 percent by The simulations from different global eco- 2050, leading to significantly slower growth in nomic scenarios were used as a starting point GDP per capita. While Kazakhstan’s GDP per to compare the performance of Kazakhstan’s capita in the baseline is estimated to reach fiscal rule with other rules. This was particu- $105,000 by 2050, the combined impacts of a larly useful to evaluate how the different rules slowdown in the Chinese economy would mean behaved in a period of economic contraction that Kazakhstan’s GDP per capita would reach and uncertainty about the future. One diffi- $98,000 by 2050. This contrasts with the milder culty associated with the use of fiscal rules as effects expected from a deterioration in the a countercyclical policy instrument is the com- Euro Area, which would imply a GDP per cap- plexity of identifying in real time where the ita of just under $105,000, and an intermedi- economy is in the business cycle. This is much ate impact from slower growth in the United more important in the context of an oil-rich States, which would bring Kazakhstan’s GDP country, given the inherent volatility of inter- per capita to $101,000 by 2050. The next para- national oil prices. Countries that try to use a graphs look at the short-term impact of these forecast of the structural deficit balance as the scenarios on the Kazakh economy. target for cyclically adjusted fiscal policy face Kazakhstan’s role in the world economy will enormous challenges. The results of the simu- depend on the policies the authorities imple- lations show that the adoption of short-term ment—and with the right policies, the country fiscal rules designed to be countercyclical can could reach the income per capita of the top 30 create even greater volatility in the presence of richest nations by 2050. The baseline scenario forecast errors. for Kazakhstan assumes total factor productiv- Theoretical considerations about an opti- ity (TFP) growth of 2.5 percent a year, the sta- mal rule show sizable gains. An optimal rule bilization of the international price of oil, and would entail accumulating savings of about a growing share of the demand for oil coming 25 percent of GDP over the next 25 years. But from developing countries. Given that Kazakh- an optimal rule simply is not practical given all stan’s TFP growth averaged 4.5 percent during the complexities involved with its implementa- 1995–2005, but before that a modest 0.5 per- tion. Even so, additional simulations show that cent, this report simulates a 0.5 percent incre- a simple fixed rule with an escape clause allows ment in TFP growth over the baseline during nearly as much gain in welfare as a theoretically 2015–19 then gradually tapering off to around optimal rule. This is very close to what Kazakh- 2.0 percent by 2025. stan is already implementing. The implica- This optimistic assumption for future pro- tion of adding an escape clause to a rule that ductivity growth rests on the premise that sig- follows the permanent income hypothesis is nificant structural changes would be needed found to deliver roughly 75 percent of the gains to reach the income per capita of the richest under the optimal rule (that is, a 12.6 percent nations. These structural changes could be improvement in national welfare). Additional OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN evidence suggests that the current fiscal rule sector. Liquidity has to be adequate when mon- in Kazakhstan already performs well compared etary policy requires boosting the money sup- with other more complex options. ply. And financial development has to be deep enough so that the links between financial Theoretical considerations for markets and instruments and the real economy countercyclical policy are developed enough. This is important since The case for strong countercyclical fiscal policy it is precisely through such channels that mon- rests on a public stimulus when private demand etary policy must act if it is to be an effective xii is weak. According to the literature, the wel- instrument of policy in affecting cycles in the fare implications of doing this are clearly bet- real economy. The fewer the options for pri- ter than a balanced budget rule, which simply vate entrepreneurs to access credit or financial transfers current cyclical ups and downs to the markets, the fewer the channels through which government budget and thus to all affected by economic authorities can expect money and it. It is also better than structural surplus rules, interest rates to provide a stimulus or a brake. which take a longer term view and target a A major impediment to a broadly effective desired long-run government surplus-to-GDP monetary policy in Kazakhstan is the fragility ratio—and respond to cyclically low (high) of the banking system. This became particu- government surpluses by increasing (reducing) larly important after the 2008–09 world finan- government debt rather than instantaneously cial crisis and the subsequent end of the Kazakh changing fiscal instruments. A key insight from real estate bubble that accompanied the world- this theoretical debate is this: if minimizing wide boom before the crash. Many banks got business-cycle volatility were instead the main into major financial trouble as a result of the objective of policy, there are significant gains crisis, and many were taken over by the state. to implementing a much more countercyclical A pervasive problem for the system as a whole rule through strong automatic stabilizers such is the high proportion of nonperforming loans as progressive taxation and unemployment carried by banks. While the Kazakh authori- insurance. ties have taken concrete steps to address this Monetary policy could also help, but its effec- problem, Kazakhstan’s nonperforming loans tiveness would require an efficient financial remain among the world’s highest. Chapter 1 Kazakhstan’s Economic Policy Stance and the Global Financial Crisis This chapter reviews the policy choices pur- GDP, transforming Kazakhstan into an upper sued by Kazakhstan during the previous middle- income economy in a short time (fig- decade and the government’s response to the ure  1.1). The share of the population living global financial crisis. The analysis starts with below the official poverty line declined from a description of the macroeconomic conditions 44.5  percent in 2002 to 12  percent by 2007. that prevailed in the first half of the decade Meanwhile, the unemployment rate, as defined and identifies the seeds of the crisis and how by the International Labor Organization, it hit Kazakhstan. The analysis next focuses declined from an estimated 13 percent in 2000 on the policy responses of the Kazakh authori- to less than 7 percent in 2007. ties to deal with the effects of the crisis.1 This In the years preceding the crisis, Kazakh- involved a fiscal stimulus that topped $21 bil- stan maintained a prudent macroeconomic lion, including targeted measures to help the policy stance and avoided procyclicality in fis- most vulnerable and the private sector and cal spending. Total spending averaged 22 per- avoid a collapse of the banking sector. The cent of GDP for most of the precrisis period, chapter then briefly discusses the lessons from with the nonoil deficit averaging 3  percent the 2008–09 crisis on how to use monetary and of GDP, even as oil prices and revenues rose. fiscal policies more effectively in combination Efforts to resist pressures for major increases in with macroprudential regulation and auto- state expenditures during 2000–07 helped to matic stabilizers to deal with a significant and accumulate significant savings in the national prolonged economic downturn. oil fund, which reached $27 billion (21 percent of GDP) by the end of 2008 (figure 1.2), and Before the crisis to maintain relatively low inflation for most of this period. Rapid growth and fiscal discipline The national oil fund helped prevent exces- could not avoid the crisis sive real appreciation of the exchange rate. Kazakhstan grew rapidly in the first half of The tenge appreciated in real terms by an esti- the 2000s and made considerable progress in mated 20 percent in 2000–08, but this was less reducing poverty. The country attracted large than that in many other resource-exporting inflows of foreign direct investment, privatized countries. This partly reflects the steriliza- small and medium enterprises and housing, tion of a large part of the oil-related inflows and adopted modern public resource man- through the national oil fund. The fact that agement institutions with significant benefits a significant share of oil profits were expatri- for the economy and its people. Annual GDP ated from Kazakhstan provided another coun- growth averaged 10  percent, and average terbalancing force to the impact of oil prices real incomes more than doubled along with on the external balance. But the pace of real 1 OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN Figure Kazakhstan real GDP, 1999–2012 1.1 300 250 Index (1999 = 100) 200 2 150 100 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: Statistical Agency of Kazakhstan; World Bank staff calculations. Figure National oil fund reserves, 2002–12 1.2 75 50 $ billions 25 0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: National Bank of Kazakhstan. appreciation did pick up considerably in 2006– capacity, set the stage for a severe financial 08 because of substantial unsterilized capital bubble. inflows. The tenge appreciated in real terms by 2 percent during 2000–05 but by 18 percent in Drowning in foreign exchange 2006–08. Easy access to foreign borrowing by banks The crisis in Kazakhstan was aided by mas- fueled credit and real estate booms. The sive capital inflows that were not sterilized in banking sector’s external debt escalated from a timely fashion, and the economy overheated. $8 billion to $46 billion (42 percent of GDP) These inflows were related to soaring real during 2005–07. The loan-to-deposit ratio estate prices in Kazakhstan’s larger cities and a reached 2.2  percent, and the nominal value related boom in housing construction. Interna- of commercial bank credit to the economy tional creditors supplied larger (private) com- almost doubled (figure 1.3), raising its share mercial banks with abundant unsecured credit. from 25  percent of GDP in 2004 to around Creditors largely expected that the government 56 percent in 2007. Most of this credit flowed would use national fund resources to prevent to construction and real estate, and the share failures of these large, systemically important of such lending was 30 percent of GDP at the banks. This expectation, combined with soar- end of 2007. The finance and construction ing commodity and real estate prices and sectors became the major drivers of economic limited banking regulation and supervision growth (figure 1.4). Expansion of credit to the nonfinancial sector and external indebtedness of Figure Kazakh banks, 2000 –12 1.3 Commercial credit to nonfinancial sector External indebtedness of banks 75 Percent of GDP 50 3 25 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: National Bank of Kazakhstan; World Bank staff calculations. Figure Composition of real GDP growth by sector, 2000 –12 1.4 Industry Construction and finance Other sectors 15 10 Percentage points 5 0 –5 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: Statistical Agency of Kazakhstan; World Bank staff calculations. The government and the regulators were capacity was inadequate. By the time pruden- late in taking action to reign in the credit tial regulations on foreign borrowing were boom. While the prudential position of banks tightened in early 2007, it was too little too late. seemed comfortable, macroprudentially there were signs of growing pressures. The surge Policy responses to the crisis in real estate prices supported bank comfort, The onset of the world economic crisis in but the heavy exposure to short-term external the second half of 2008 and associated sharp sources of financing created significant vulner- declines in commodity prices had a severe abilities. The absence of a sound legal frame- impact on K azakhstan. Monthly export work for resolving weak and troubled banks, revenues fell from $3.5  billion in the first insufficient independent monitoring of the 11  months of 2008 to less than $1.5  billion conditions of banks to generate early warning in December 2008 and in the first quarter of signs, and the perception that the government 2009. The current account moved from a sur- might be willing to bail banks out meant that plus of $3.7 billion in the first half of 2008 to a the banks had few incentives to manage their deficit of $3.5 billion in the first six months of risks. The porous regulatory framework could 2009. With a weak capital account, this placed not restrain excessive external borrowing or the tenge under significant pressure. Kazakh domestic real estate lending, and supervision banks also lost virtually all access to rollover OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN opportunities in global financial markets. those from private pension funds), extended Lower commodity prices worsened investment unemployment benefits from four months to sentiment toward Kazakhstan. six, and promoted the signing of memoranda After the country lost $6  billion trying to between regional authorities and larger enter- defend the tenge in January and early February prises to preserve jobs. In the spring of 2009, 2009, the National Bank conducted a one-off an additional public works program of $1.5 bil- 20 percent devaluation of the currency. While lion (1.5  percent of GDP), called the “Road the devaluation proved effective in relieving Map,” was set up for job creation, training, pressure on the tenge and stabilizing expecta- and temporary wage subsidies to vulnerable 4 tions, a weaker tenge put additional pressure groups at the regional level through special on the banking sector, with roughly half of the budgetary transfers to local authorities. Public domestic loan portfolio denominated in for- salaries were increased significantly. The share eign currency. of social spending in the consolidated govern- Kazakhstan’s fiscal position deteriorated ment budget increased from 52 percent in 2008 sharply, and the nonoil fiscal deficit widened to to 60 percent in 2009. 10.5 percent of GDP in 2008 and hit 13.8 per- The National Bank stepped up its liquid- cent in 2009. Consolidated state budgetary ity support for commercial banks in late 2008 revenues dropped by about 4 percent of GDP and 2009, while deposit insurance increased in 2008 due to the recession, low commod- sevenfold. By March 2009, reserve require- ity prices, a lower corporate income tax rate, ments for commercial banks had been progres- and the granting of uniform investment cred- sively reduced from 6.0 percent to 1.5 percent its. Outlays increased from 23 percent of GDP for internal liabilities, and from 8.0  percent in precrisis years to 27.9  percent in 2009 to to 2.5 percent for external liabilities. Reverse finance higher social commitments and off- REPO operations were used for liquidity sup- budget stimulus spending. port ($2.9 billion by the end of the third quar- The fiscal difficulties highlighted the need ter of 2009). The accounts of a number of for more budget flexibility and more effective state-owned organizations were also shifted targeting of countercyclical fiscal policies. The from the National Bank to commercial banks. immediate priority was to make public spend- To support depositor confidence, household ing more efficient, budgets more strategic, and deposit insurance was increased from 700 budget execution more flexible. The govern- thousand tenge to 5 million tenge. ment undertook a major revision of the 2004 Budget Code, and parliament approved its first The banking crisis in three-year national budget for 2009–11 based international perspective on the strategic plans of government ministries The basic patterns in Kazakhstan are qualita- in 2008. tively similar to those in many other countries The authorities implemented an anticrisis that have experienced financial and banking support program of $21 billion. The majority crises. The initial financial instability came of this came from the republican budget or the from an investor panic and sudden stop in large state corporate conglomerate Samruk- capital inflows that followed rapid foreign bor- Kazyna. About half the support funds ($11 bil- rowing and credit expansion to housing and lion) were devoted to direct assistance to construction in the context of soaring real struggling commercial banks (primarily BTA), estate prices. This is the most common pat- while the other half (more than $10  billion) tern for financial crises in a large number of financed anticrisis stimulus measures in vari- countries. ous sectors of the economy. Several country-specific factors in Kazakh- Several measures were introduced to allevi- stan increased the severity of the bubble that ate the social consequences of the crisis. The burst in August 2007. As in other countries, a government protected social expenditures primary cause of the initial financial instabil- from budgetary sequestration, insured all pen- ity was excessively risky behavior by commer- sion payments indexed to inflation (including cial banks and their foreign creditors. This behavior reflected the expectation that, in the • Changing prudential and other regulations event of failure, a good share of the losses might to prevent a repetition of the excessive risky be passed on to the government. Such distorted borrowing and lending that led to the 2007 incentives can be mitigated to the degree that instability. the government can either commit to limit the These and other measures prevented a col- extent of such bailouts (enforce hard budget lapse of the banking sector—and have done constraints) or enforce prudential regulations much to create conditions for the recovery that place direct restrictions on the ability of of credit and financial markets under new banks to assume risks. As a transition economy, and healthier principles. Through various 5 Kazakhstan still had quite limited capacity for programs, the government and state sector the latter, and the rapidly accumulating fiscal injected close to $10 billion into the banking oil reserves of the government made the even- sector during 2007–09, while foreign creditors tual bailout of systemically important distressed are financing a similar sum through write- banks appear probable to investors. Under downs of their debts in BTA and Alliance. The these conditions, foreign creditors supplied fact that Kazakhstan succeeded in leaving a abundant cheap credit to Kazakh banks with lit- good share of the losses in BTA and Alliance tle concern about how the money was used, and with creditors and shareholders is of primary many of the banks used this money to finance importance. First, it reduced the direct costs high-risk strategies in real estate. Foreign bor- of revitalizing (or shutting down) these banks rowing came to account for as much as 40 per- to the government and taxpayers. Second, it cent of bank liabilities, while foreign currency sent a strong signal to financial markets that lending reaching roughly half of the aggregate the government is not implicitly insuring pri- credit portfolio. vate unsecured foreign credits, a signal that now motivates foreign creditors and investors Managing the banking crisis in Kazakhstan to devote more attention to ensuring efficient The government and National Bank devel- use of their funds in Kazakhstan. oped an effective strategy to manage the cri- sis. Authorities met the initial instability with The lessons from the crisis extensive liquidity support and a sevenfold The global financial crisis challenged some increase in deposit insurance for households, macroeconomic principles that policymakers and later with a recapitalization plan for the used to live by. First, it challenged the belief four largest banks. Following a further dete- that stable and low inflation was the primary rioration of the situation, authorities seized and many times exclusive mandate of cen- control of two large problem banks in Febru- tral banks. Second, it forced central banks to ary 2009. By mid-2009, the government and rethink the conventional principle that mone- National Bank developed an effective and con- tary policy should focus on a single instrument: sistent strategy, including: the policy interest rate. Third, with low and sta- • Passing revised legislation in 2009 that ble inflation secured, the conventional wisdom allows for the effective resolution of large was that fiscal policy should have a limited role problem banks, including purchase and in macroeconomic management; but this prin- assumption powers and problem-bank divi- ciple also had to be set aside during and after sions into “good bank/bad bank.” the crisis. Fourth, the crisis urged in a change • Revising the legal framework for bank in perception of financial regulation, which restructuring. was largely ignored as a macroeconomic policy • Successfully renegotiating the unsecured tool before 2008 in most countries. Fifth, given private credits to BTA and Alliance to write that discretionary fiscal measures may only down more than $10 billion of the foreign come too late to fight a standard recession, the debts of those banks. crisis strengthens the need for improving auto- • Increasing the capacity and effectiveness of matic stabilizers.2 banking supervision, particularly for larger While stable and low inflation remain valid banks. objectives of macroeconomic policy, they are OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN no longer considered enough. When the crisis governments should promote fiscal adjust- hit, core inflation was stable in most advanced ments to reduce debt-to-GDP ratios rather than economies. With the benefit of hindsight finance spending increases or tax cuts. many have argued that core inflation was not Financial intermediation is now seen as the right measure of inflation, and that the an important aid to macroeconomic policy. increase in oil or housing prices should have Before the crisis, conventional wisdom was been taken into account. What has emerged that financial intermediation should focus on as a postcrisis lesson is that while inflation individual institutions and markets only—and and the output gap may be stable, the behav- that it had little or no role as a macroeconomic 6 ior of some asset prices and credit aggregates, policy tool. In many countries, there was great or the composition of output, may be undesir- enthusiasm for financial deregulation, and able (for example, excessively high consump- prudential regulation for cyclical purposes was tion, housing, or current account deficits) and considered improper interference in the credit can potentially trigger major macroeconomic markets. After the crisis, again with the benefit adjustments later. of hindsight, it became common sense to com- Monetary policy alone is no longer seen bine monetary policy with macroprudential as an effective way to deal with prolonged regulatory tools. While the policy interest rate recessions. When the crisis started in earnest can be used primarily in response to aggregate in 2008, many central banks started quickly demand activity and inflation, macropruden- reducing interest rates to close to zero, but tial regulatory instruments can help deal with this proved insufficient to revive aggregate financing or asset price issues. At the same demand. They had to extend the scope and time, regulatory instruments can help central scale of their traditional role as lenders of last banks handle excessive leverage or risk taking resort and provide liquidity to the economy by banks. more broadly. While it is questionable whether The crisis highlighted the need to develop the provision of liquidity should be kept in efficient automatic stabilizers to ease the more tranquil times, due to moral hazard and effects of a deep recession. As discussed, mon- risk taking by banks, the lesson from the crisis etary policy may not be effective in a severe is that monetary policy alone could not bring economic downturn, and fiscal measures may the economy back from a deep recession. Fiscal yield positive results only with a lag. Automatic policy had a huge role. stabilizers do not suffer from these shortcom- Another important lesson from the crisis is ings. With large fiscal stabilizers, the impacts that it pays to build fiscal space in good times. in the economy can be felt in a timely and Because the use of monetary policy, includ- gradual fashion as taxes and expenditures ing credit and quantitative easing, reached a react in a countercyclical manner to changing limit, policymakers had no choice but to rely economic conditions.3 In addition, automatic on fiscal policy. With the expectation of a pro- stabilizers minimize both the interference of longed recession, the fiscal stimulus became political decisions and the implementation lags the norm in the economies hit hardest by the in responding to the shock. Automaticity also crisis, like Kazakhstan. But economies that guarantees a timely reversal of a fiscal expan- entered the crisis with high levels of debt and sion as the fiscal loosening during bad times large unfunded liabilities had limited ability to automatically becomes tightening in good use fiscal policy. Similarly, the economies that times. ran highly procyclical fiscal policies driven by Automatic stabilizers can be enhanced consumption booms were forced to cut spend- through tax and expenditure policy changes ing and increase taxes despite unprecedented without increasing the size of government. recessions. Only those that entered the crisis This could be done by making tax collections with lower levels of debt were able to rely on responsive to changes in economic conditions countercyclical fiscal policies to fight the cri- and by raising the share of taxes collected sis. An associated lesson for the next decade from income-based taxes, which usually have is that, whenever cyclical conditions allow, higher income elasticities. In addition, some expenditure programs, such as unemploy- time-bound rebates in the personal tax or ment benefits, have a stabilizing impact on dis- reductions in value-added taxes or sales taxes posable household income during economic during severe recessions; allowing companies slowdowns. But these traditional mechanisms to offset corporate tax losses against past prof- are already widely used in most countries and its; scaling up unemployment benefits when tend to have only limited automatic stabilizing unemployment rates exceed a certain thresh- effects in a downturn.4 To enhance their effi- old; and automating a system of budgetary cacy, a suggestion is to change tax and expen- transfers to states or provinces in prolonged diture parameters temporarily in response recessions. 7 to macroeconomic conditions. This would Kazakhstan has responded to the crisis in include, for example, providing temporary a way consistent with these lessons. While the Box Description of the social safety net programs 1.1 State special allowances. State special allowances were introduced in 1999 to replace earlier cash and noncash preferences and discounts designed for various groups of people—for housing maintenance and utilities, fuel, telephones, medicines, glasses, public transportation, periodical subscriptions, and the like. Special state allowances are now paid in cash to those in the following eligible categories: World War II veterans, the disabled, individuals who participated in the liquidation of the Chernobyl catastrophe, families of military persons who died in service, individuals awarded orders or medals, victims of political repression, and mothers and families with many children. There are 17 categories overall, funded from the national budget and with the average $20 per person a month. State social allowances. This category of allowances comprises disability allowance, survivor allowance, and old-age allowance. Disability allowance is paid in cash on a monthly basis during the disability period. Disability must be certified. Once the disabled beneficiary reaches retirement age and becomes eligible for pension payments, disability allowance is terminated. But if the total of the individual’s pay-as-you-go pension and accumulation pension is less than the previously paid monthly disability allowance, the allowance will continue to be paid, its size being equal to the difference between the awarded pension and the earlier disability allowance. Survivor allowance is paid to all eligible family dependents of deceased breadwinners. Survivor allowances are awarded for the period during which the recipi- ents remain unable to work or until they reach retirement age. Old-age allowance is paid to individuals who have reached retirement age but are either ineligible for pension payments or whose pension entitlements are below the minimum. Retirement-age persons who have been making the mandatory contributions to the pension system may also receive an old-age allowance sufficient to bring their income to the minimum monthly pension. These are funded from the national budget, and the average is $50 per person a month. Targeted social assistance. Targeted social assistance was introduced in 2002. It is a cash payment provided to individuals (or families) with average per capita monthly income below 40 percent of the subsistence minimum of their oblast. The size of the payment for each eligible household member is the estimated difference between the household per capita income and the poverty line in the region. The calculation of income of beneficiaries includes cash income as well as noncash proceeds from the use of assets (such as income from poultry, livestock, and crops). Housing assistance. The housing subsidy, introduced in 2003, is compensation to low-income and vulnerable groups to cover expenses for housing maintenance, utilities, and leasing. Housing assistance is extended to poor households if their actual housing and utilities expenses exceed a certain percentage of the aggregate household income defined by the local government. Local authorities define the size of and procedure for obtaining housing assistance, as well as the eligibility for it. (The size is determined by local executive bodies and varies from 10 percent to 25 percent.) Local budgets fund this assistance. Allowances for children under 18 years. Also called “child allowances,” this program was introduced in 2006 and comprises birth allowance, child care allowances for children under age 1, and monthly allowance for children under 18 years of age. Eligibility is based on a mix of principles. Childcare allowance is available to caregivers who do not participate in the mandatory formal sector social insurance. Birth allowances and monthly child allowances are available to children of families with a per capita income below 60 percent of the subsistence minimum. The national budget funds these allowances. Local transfers. Because Kazakhstan law prohibits regional authorities from providing targeted social assistance on a means-test basis, regional budgets, for the most part, provide additional categorical benefits and allowances. Typically, wealthier regions (Atyrau is one of wealthier regions, due to oil extraction there) can afford a broader set of categorical benefits and allowances. In poorer regions, regional legislation often does not establish any additional benefits. Source: Authors. OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN roots of the crisis reflected a lack of prudence banking sector. A description of Kazakhstan’s that contributed to the onset of a deep reces- current social safety net programs is presented sion, the response that followed was measured in box 1.1. and effective. Aided by good fiscal discipline The next chapter discusses Kazakhstan’s before the crisis, the authorities built sizable postcrisis fiscal policy and the robustness of fiscal buffers in the national oil fund, which its fiscal rule. Having weathered the effects allowed them to use a fiscal stimulus package of the global financial crisis fairly well, the that helped the economy recover. Social mea- authorities are concerned whether Kazakh- 8 sures were also used to alleviate the impact stan could be affected by another global crisis of the downturn on the most vulnerable, and and whether they would be well equipped to targeted measures were implemented to help deal with the effects of a new and prolonged the private sector cope with the crisis. The recession. The next chapter discusses the response was slower and less effective in deal- effectiveness of the current fiscal rule that ing with the problems that hit the financial governs the transfers from the national oil sector hard. Although this is an area where fund to the budget against different robust- more remains to be done, the strategy adopted ness criteria and the ability to deal with transi- by the authorities avoided a collapse of the tory shocks. Chapter 2 Fiscal Rules Should Be Simple and Transparent This chapter tests the robustness of the current budget have been adequate. As a result of this set of fiscal rules in Kazakhstan and compares adequate fiscal stance, when growth stalled in its main features with those of other popular the global economic crisis of 2008 (which hit rule-based approaches. Building on an ear- Kazakhstan through falling real estate prices lier assessment prepared by the World Bank and insolvency in the banking sector), the gov- in 2012 that discussed the virtues and pitfalls ernment was able to use a discretionary fiscal of different options for the rules governing expansion, largely financed with oil funds, to the transfers from the oil fund to the budget, reduce the welfare impact of the crisis. this chapter advances the main conclusion Current fiscal policy is governed by a set of from this report to avoid introducing drastic rules originally adopted in 2010 and further changes to a mechanism that has worked well. modified in 2012. The main features of these The first section discusses the main features rules are an annual transfer from the oil fund of Kazakhstan’s existing fiscal rules. The sec- to the budget of $8  billion in nominal terms ond describes what other countries have done (the exchange rate for setting the amount in and focuses in particular on the cases of Chile tenge is fixed over a three-year horizon), and and Norway. The third considers the merits a limit to the interest cost of the public debt to of the bird-in-hand approach, and the fourth 4.5 percent of the oil fund’s balance. The main compares the main features of Kazakhstan’s features of the current rules are as follows: current fiscal rule with the permanent income • Fixed annual guaranteed transfer (made in rule, the bird-in-hand rule, and a countercycli- monthly or quarterly installments) to the cal policy that would allow expenditures to rise budget of $8 billion starting in 2011. in downturns. This is done by using the criteria • Minimum oil fund balance of 20 percent of put forward by Kopits and Symansky (1998). projected GDP at end of the respective fis- The fifth section discusses why there seems to cal year. (The guaranteed annual transfer be no case to significantly change the current will be reduced if the expected balance falls fiscal arrangements in Kazakhstan. short of this minimum.) • Annual expenditures on public debt service, Main features of the Kazakh fiscal rules however defined, not to exceed 4.5 percent In 2012, the World Bank prepared a technical of the imputed fixed investment return on note that assessed the fiscal management of the fund. (It remains unstated whether the its natural resource flows. 5 The main finding fund balance is projected at the end of the from that study was that fiscal management has respective fiscal year, the average over some been responsible and that the rules governing period, or the actual balance prior year.) the accumulation of savings in the national oil • Average cost of service and repayment fund and the transfers from the oil fund to the of public debt over 10 years not to exceed 9 OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN 15 percent of total budget receipts including will remain valid until the end of 2013, and the cash transfer from the fund. budget will receive a transfer of $9.2  billion, • Target nonoil deficit of 3 percent of GDP by keeping the transfers pegged at the upper end 2020. of the range specified. Meanwhile, the comple- • No off-budget financing—that is, no guar- mentary provisions to prevent government bor- antees or lending for domestic activity rowing from “undoing” the oil fund’s savings including Samruk-Kazyna and KazAgro. remain unchanged. There were further changes in 2012, and the The oil fund has accumulated significant 10 government is considering whether it would savings and helped build sizable fiscal buffers be wise to introduce yet additional changes while avoiding volatility in the exchange rate. in the fiscal rule governing transfers from the Despite more abrupt variations in oil prices, the oil fund to the budget. The oil fund was modi- exchange rate has been kept flat (figure 2.1). Oil fied by government’s Presidential Decree 289 fund assets rose to more than $64 billion at the (March 16, 2012). The annual transfer of funds end of 2012, helping Kazakhstan accumulate to the budget was changed from the previ- fiscal buffers in excess of 32  percent of GDP, ously fixed amount of $8 billion to the flexible the highest ever (figure 2.2). With oil prices amount of $8 billion plus or minus 15 percent remaining largely volatile and unpredictable, ($6.8–$9.2 billion), depending on the cyclical preventing the transmission of huge swings to position of the economy. These adjustments the Kazakh economy is paramount. Figure The real exchange rate was kept almost flat, despite some fluctuations in oil prices 2.1 Real exchange rate (dollar to tenge) Export oil price 110 150 Index (2008 Q1 = 100) 100 100 $ per barrel 90 50 80 0 2008 2009 2010 2011 2012 2013 Source: National Bank of Kazakhstan; World Bank staff calculations. Figure Oil revenue savings exceeded 32 percent of GDP, the biggest fiscal buffer ever 2.2 Domestic debt External debt National Fund assets 40 30 Percent of GDP 20.0 20 16.2 16.7 13.7 13.9 10 0 2008 2009 2010 2011 2012 Source: Finance Ministry of Kazakhstan; World Bank staff calculations. The $8 billion rule governing transfers from parliamentary budget approval process targets the oil fund to the budget is consistent with the a structurally adjusted measure of the budget permanent income approach. Using histori- balance. cal data for the main parameters, a range of annuity values for Kazakhstan’s oil stock was Chile’s structural balance target calculated in line with the permanent income Chile’s fiscal policy is anchored by a structural approach (table 2.1). Under the base case for balance target for the central government. This oil reserves (which uses estimates of the main feature was institutionalized in the 2006 Fiscal industry participants), and using the approxi- Responsibility Law. The framework is designed 11 mate mean Brent crude price over 40 years to ensure long-term fiscal sustainability while (1970–2010) and a 3  percent long-run finan- avoiding cyclical policies and allowing the cial rate of return, the implied annuity value is full operation of automatic stabilizers, mostly about $8.4 billion (figure 2.3). on the revenue side. Under the structural bal- ance rule, government expenditures are bud- Fiscal rules in Chile and Norway geted in line with structural revenues. Those Other countries can serve as a benchmark for are the revenues that would be achieved if the the adequacy of Kazakhstan’s fiscal rules. A economy were operating at full potential, the further feature of fiscal rules adopted in some prices of copper and molybdenum—Chile’s countries is the ability to maintain or increase major exports—were at their (assumed) long- spending in economic downturns, financed term levels, and, more recently, the return on through a tighter fiscal stance during booms. accrued financial assets were in line with the Take Chile’s fiscal rule, which rather than set a (assumed) long-term interest rate. target for the actual budget surplus targets the Two sovereign wealth funds support the structural budget balance. Similarly Norway’s Chilean fiscal framework. The Pension Reserve Table Permanent income annuity values of oil 2.1 (Under certainty about future revenues, prices in millions of dollars) Rate of return Amount Price (3%) 3.3 billion tons Approx. 40-year mean $40 $6,306 Low case (24.8 billion barrels) Mean+approx. ½ std. dev. $60 $8,927 4.0 billion tons Approx. 40-year mean $40 $8,434 Best case (35.5 billion barrels) Mean+approx. ½ std. dev $60 $12,121 Source: World Bank staff calculations. Figure Permanent income-based annuity and projected revenue profile for Kazakhstan 2.3 15 Shaded area Extraction profile at $40 a barrel constitutes 10 NFRK savings 2010 $ billions Sustainable expenditure under permanent income hypothesis at 3% real return on investment 5 0 2012 2015 2020 2025 2030 2035 2040 2045 2050 2055 NFRK is National Fund of the Republic of Kazakhstan. Source: World Bank staff calculations. OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN Fund accumulates and invests fiscal savings receipts to finance increased nonoil deficits, as to cover long-term pension liabilities (related nonoil revenues were depressed and expendi- to Chile’s aging population). The minimum tures rose as a result of higher unemployment annual amount paid into the pension fund and other recession-related spending. Con- is 0.2 percent of the previous year’s GDP, but versely, from 1996 onward, and through the if the effective fiscal surplus exceeds this 2000s commodity boom, substantial reserves amount, the contribution can rise to a maxi- were accumulated. During the recent global mum of 0.5 percent of the previous year’s GDP. crisis, the realized (not structural) deficit once The Economic and Social Stabilization Fund again exceeded the 4 percent target. 12 is financed from the remainder of the effec- Norway also operates two separate sovereign tive fiscal surplus after payments into the pen- wealth funds with different funding sources, sion fund. Withdrawals from the stabilization management arrangements, and investment fund can finance fiscal deficits during periods policies. The Government Pension Fund Global of weak growth or low copper prices; they can (formerly the Petroleum Fund) is financed with pay down public debt or finance the pension surplus oil revenue (net of the annual transfer fund. Both funds may invest only in assets to the budget, explained above). Serving sav- abroad, and under current investment policies ings and stabilization purposes, it is required to are limited to fixed income and money market invest its capital abroad, to avoid overheating the instruments. economy and to shield it from oil price fluctua- Structural balance targets have desirable tions. The diversified investment portfolio com- properties in theory, but they are complex to prises international equity, fixed-income, and implement. Their advantages may thus come real estate assets, with the aim of maximizing at a cost of reduced transparency. If they are the risk-adjusted rate of return within the invest- viewed as too difficult for a government to exe- ment policy set by the Ministry of Finance. The cute properly, they may also undermine cred- Government Pension Fund Norway (formerly ibility in the markets. Strong institutions are the National Social Insurance Fund) is a sepa- needed for successful implementation. Chile’s rate fund, financed from surpluses in the social approach involves using outside, independent insurance scheme (social security contribution expert panels to set key parameters and recom- minus payments). Since its revenue originates mend policy to government. from domestic sources, the fund is permitted to invest in equity and fixed income instruments Norway’s structural deficit rule in regulated markets in Denmark, Finland, Nor- In Norway, withdrawals from the oil revenues way, and Sweden. finance the nonoil deficit, determined each year during the annual budget process. Since A bird-in-hand? 2001, fiscal policy has been anchored by a The Kazakh authorities are also interested in structural deficit rule requiring the structural the bird-in-hand rule, transferring the invest- nonoil deficit to equal the imputed real return ment returns from the national oil fund to the on the assets in the pension fund (currently budget annually. Norway currently pursues a 4 percent). The rule ensures that fiscal policy version of this approach, setting its structural does not deviate over time from a sustainable deficit spending to a “smoothed” estimate path—by constraining government expendi- (using a 4 percent rate of return) of the returns tures to long-term government revenues minus to its global fund. Norway, however, is toward the structural balance target. the end of its extraction cycle: it has already The Norwegian rule allows for deficits to allowed its fund to accumulate more than exceed the 4 percent target during economic 100  percent of GDP. Current projections for downturns. It also enforces lower deficits dur- Kazakhstan envisage the annual transfer from ing boom periods, reducing pressures on aggre- the national oil fund to reach or surpass its gate demand and stabilizing economic activity. investment earned income only by 2025. For example, during a severe recession in the One objection to this approach is that it early 1990s, Norway used its entire oil revenue would severely constrain current spending. Applying the bird-in-hand rule over the life of the country. That is, which organ of the gov- of the extraction cycle implies higher savings ernment (central bank, ministry of finance, and lower transfers to the budget in the earlier ministry of economy) should be responsible for years because the driving principle is uncer- these functions? Who will ensure compliance tainty about future revenues (only today’s fund and impose enforcement mechanisms when balances can be counted on for returns). This needed? Perhaps the most sensitive area is to policy is appropriate for mature producers, determine under what circumstances “escape such as Norway, with large accumulated finan- clauses” would be invoked and by whom. Insula- cial reserves. In Kazakhstan, at an earlier stage tion from ordinary political processes is espe- 13 in the extraction cycle, disregarding future oil cially important, given the normal tendency of revenues would imply a contraction in current politicians to take a short-term view of matters consumption. This may be viewed as incompat- involving expenditures as well as the history of ible with the principles of intergenerational politically motivated “escapes” from fiscal rules fairness underlying the permanent income in numerous other countries. approach. A further objection to bird-in-hand is that Well defined: The target variable should be the shift is only in the quite distant future. In clearly defined, as should the institutions one scenario, the switch from an annuity to for and the coverage of the rule bird-in-hand would occur in 2025. Setting fis- The Kazakh rule is well defined and transpar- cal rules so far in the future is not credible. ently defined. If the size of fiscal transfers to More advisable would be rules governing the the national budget is the target variable, the present (and near future) that are consistent government has clearly done a good job defin- with longer term principles of fairness and pru- ing it under any of the fiscal rules under con- dent economic management. sideration. All mineral revenues accrue to the fund, which is the sole conduit for transfers to The Kazakh fiscal rule and the the national budget. This allows for no ambigu- Kopits and Symansky criteria ity or lack of definition. Under any variant of a An analysis of how the current fiscal rule in fixed amount transfer rule, the target is simi- Kazakhstan compares with the two leading larly well defined—the amount of the transfer rule-based approaches under different criteria is clear, and there is no attempt to influence shows a robust performance. A useful frame- any other outcome. work for evaluating fiscal rules is in Kopits and Under the bird-in-hand rule, there is simi- Symansky (1998). The following discussion larly little scope for ambiguity—the limit on takes each of their criteria in turn. It compares transfers relates to “money in the bank.” As the current fixed rule with the permanent long as this is a publicly known amount (there income hypothesis, which depends on the pres- are no hidden or secret accounts and no use ent value of the wealth represented by the coun- of contingencies such as futures or options), a try’s resource base and national savings. The rule such as one limiting transfers to a fixed bird-in-hand approach allows consumption to percentage of the wealth on hand is difficult to be based only on already-extracted resource game or misunderstand. wealth (on the value of the national savings Under a permanent income rule, there is fund alone and not on the value of oil still in some scope for differing opinions on such mat- the ground). And with a countercyclical policy, ters as the discount rate and the value of oil in which would allow expenditures to increase in the ground. But it should be possible to gen- downturns, as with the Chilean copper fund erate a range of plausible values using differ- rule based on a structural deficit estimate. ent assumptions for these values, allowing the Some of these rules might make sense in authorities to choose between, say, the most particular circumstances, but the real question conservative and least conservative values that here is whether one of them is clearly superior to can be derived. the Kazakh rule. Their characteristics must also But if the variable to be targeted is some be clearly mapped to the institutional structure measure of budgetary surplus or deficit, as OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN with a countercyclical policy rule, the problem bird-in-hand, which depend directly on the becomes more complicated. While a variety of value of the oil wealth as calculated on the fiscal balance concepts could be targeted, even basis of market prices, can generate procy- if there is a consensus that a well-known bal- clical outcomes and are not preferred poli- ance such as the nonoil primary deficit should cies if the goal is to smooth shocks to current be used, it is difficult to know in real time income and output rather than to smooth exactly where that balance lies. Similarly, there transfers themselves. Countercyclical policies are enough judgment calls involved in comput- can indeed accomplish some degree of output ing the target variable that it is quite possible or income smoothing, but the extent to which 14 for substantial disagreements or simple mea- this can be achieved is not likely to be perfect. surement errors to persist for long periods. Not only is it virtually impossible to know in That makes achieving a consensus difficult or real time what degree of countercyclical policy allows political interests to dominate debate. is truly necessary, but the tools to achieve this Indeed, these issues pose most of the problems goal are less than perfect even in the best of in countries often pointed to as “best cases” cases. So, while the goals of such a policy may using independent boards of experts to deter- be laudable, the best that can be expected is mine these variables such as the Chilean Cop- that government policy would “lean against per Fund. the wind” of economic cycles and at worst could even exacerbate them where measure- Transparency: The operation of the rule and ments of variables or expectations of future the actions to ensure compliance should developments are mistaken. be clear and visible to all observers Many of the issues under this criterion are Consistency: The rule should be both related to those under the first criterion. While internally consistent and consistent with any of the proposed arrangements for fiscal other policies and goals of economic policy rules can be made transparent, the simpler Evaluation of this criterion is quite similar to rules are much easier to make clear to the that for the “adequacy” criterion above. If fixed public and, for those not versed in economic transfers are chosen, there is little chance for concepts, to understand and report on. For inconsistency with other policies as there is no example, while fixed amount rules are quite variation to generate any lack of synchronic- obvious to all, as are rules involving fixed per- ity. In contrast, policies such as bird-in-hand centages of accounts owned by the govern- and permanent income, capable of generat- ment, rules involving concepts like present ing procyclical results, can easily generate discounted value (as in permanent income) outcomes at odds with the government’s legiti- will not be readily understood by the general mate desire to smooth fluctuations in output public. Countercyclical policies may be under- and income. While countercyclical policies are stood in the abstract (“we will increase trans- in theory a remedy to this problem, the diffi- fers when oil prices are low”) but may be more culties in implementation and measurement difficult to explain. mean that these rules can at times be consis- tent with other policies and goals and at times Adequacy: The mechanisms envisioned in not. Worse, this may not be clearly known until the rule should be capable of achieving well after the fact. the desired level of the target variable If the goal is to prevent transmission of oil Simplicity: The rule should be easily price f luctuations to the national budget, stated and easily understood any variant of a fixed transfer rule is a good The simplicity of fixed transfer rules is obvious. way to achieve this. In fact, the performance The bird-in-hand and permanent income rules of the government to date in achieving this are less simple but still easily stated and under- goal could well be regarded as a model of stood. Countercyclical rules are much less clear how to do this in an effective and sustained in practice though they can work quite well in way. Rules such as the permanent income or theory. Flexibility: The rule should be flexible Thomas Bertram Lance, the Director of the enough to deal with changing economic Office of Management and Budget in Jimmy circumstances by containing adequate Carter’s 1977 administration, was quoted in the escape or emergency clauses newsletter of the U.S. Chamber of Commerce, Fixed transfer rules are by their very nature not Nation’s Business: flexible. It is certainly possible to alter the fixed amount to be transferred in light of changing Bert Lance believes he can save Uncle economic circumstances but only at the cost of Sam billions if he can get the government losing the positive attributes of a fixed trans- to adopt a simple motto: “If it ain’t broke, 15 fer rule. Periodic reviews of the amount to be don’t fix it.” He explains: “That’s the transferred in light of economic growth and trouble with government: Fixing things changing oil markets and production would that aren’t broken and not fixing things seem reasonable. But the reputational benefits that are broken.” of adhering to a fixed rule would be lost in direct proportion to the extent that it is subject The various alternatives for fiscal rules to frequent revision. under consideration in Kazakhstan all have Bird-in-hand and permanent income rules merit in theory but generate quite differ- are certainly flexible in that they would gener- ent results when subject to actual numerical ate different transfer amounts in response to simulation. In particular, the proposal to use different oil market conditions. But their flex- transfers from the sovereign wealth fund in a ibility is the opposite of what the government countercyclical manner seems attractive but is seeking in its investigation of countercyclical suffers in practice from the near impossibility policies. of identifying cycles in the national economy, Countercyclical rules are flexible and in the national budget, or international oil prices the “right” direction, but this flexibility car- in a time frame that would allow implementing ries the risks noted above. Only in theory can a such a policy. While long-run values for trends countercyclical rule be counted on to perform in GDP or for international equilibrium prices as desired. In practice it will be less than per- may be estimated, the confidence in such esti- fect (and at times far less than perfect) in its mates, given the history of past projections performance. and the performance of the oil price, should not be overestimated. Accordingly, rules that Enforceability: The mechanisms to depend on such forecasts or estimates are likely enforce compliance must be clear to run into trouble in very short order. In all All of the fixed transfer, permanent income, cases analyzed in this report, rules that allow and bird-in-hand rules can be enforced by for short-term flexibility in budgetary transfers the managers of the sovereign wealth fund, generate more volatility rather than less due to as in past years. There is no reason to try to the inevitable forecast errors in even the best alter these arrangements given the past per- of cases. formance. A countercyclical rule, by contrast, Some of the proposals have the potential to would require a mechanism to specify the state seriously change the way the oil fund is man- of the business cycle and the structural budget aged and how it relates both to containing deficit, difficult in the best of circumstances, demand pressures and smoothing economic as shown by the numerical exercises comparing cycles. In particular, proposals to allow invest- output forecasts with actual outcomes in chap- ment in domestic assets, though perhaps rea- ter 3. sonable in some cases, have the potential to undermine one of the most important suc- “If it ain’t broke, don’t fix it” cesses of macro policy management over the There is a famous saying in the United States past decade: insulating the domestic economy about the usually well-intended action by from the overheating caused by a too-rapid policymakers to introduce reforms that in expansion fueled by oil revenue. In fact, avoid- some cases may not be necessary. In May 1977, ing doing exactly this is why sovereign wealth OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN funds are so important in the first place. This should preserve the nondistortionary Undermining the rationale for the national characteristics of the original $8  billion oil fund cannot be defended on economic fixed amount while allowing for the greater grounds and should be avoided if at all possi- absorptive capacity of the now-larger domes- ble. In addition, the political pitfalls of govern- tic economy. For the future, periodic adjust- ment asset purchases in the domestic market ment of the base might be desirable. The could well be difficult to avoid as the govern- adjustments could be made based on trend ment assumes the role of choosing where and growth in nonoil GDP, something that could where not to invest. be calculated only over longer periods—at 16 Theoretical considerations discussed in the least four to five years, if not longer. next chapters show that a simple fixed trans- • Second, add two limiting factors for implementing fer rule with an escape clause allows nearly as 15 percent deviations from the rule. One would much gain in welfare as a theoretically optimal be to specify the size of the shock that would rule. The use of the oil fund to cushion the trigger such a deviation; the other would effects of the 2008–09 economic crisis was a de impose a time limit for how long such addi- facto implementation of a fixed rule with an tional transfers could be used (such as two escape clause, something that could be read- years). ily codified in the future. Important values to • Third, implement “automatic stabilizers” to the be fixed in any such codification would be the extent possible. That is, use the structure of threshold for triggering an escape clause and the tax code (by doing such things as mak- the number of years such an alternate regime ing taxes conditional on income or profits) could be continued before assuming that the and the structure of expenditure programs shock is chronic rather than transitory (that (by doing such things as making poverty the “cycle” is in reality the new “normal”). or other assistance programs conditional The oil fund as currently constituted has on income or employment status) to make been able to deal with transitory shocks. As countercyclical revenue and expenditure the next chapter will show, even a “meltdown” changes as automatic as possible. of the European Union does not produce a Such a rule would retain the base amount short-run shock larger than can be smoothed at a level consistent with current GDP, oil rev- today. Longer run simulations show that the enue, and fund total holdings—at $9.2 billion most important factor for the government is a year—but would allow an increased amount to ensure the efficiency of investment expen- of, say, 15 percent if GDP growth falls below a ditures at the present time since even modest specified level. The two-year limit is in effect an success has the potential to create very large “escape from the escape clause.” Other mecha- increases in welfare. But the danger that a pro- nisms could be envisioned, but the idea is to longed shock might be mistakenly perceived as avoid triggering an escape clause in a repeated transitory is a problem inherent in any attempt sequential fashion. to use the oil fund to smooth economic Bear in mind that simplicity is a great vir- fluctuations. tue in rules of this type. The discussion of the The main recommendations to deal with Kopits–Symanski criteria for evaluating fiscal transitory shocks within the framework of rules highlights the desirable qualities of the the current fiscal rule can be summed up as current fixed rule and gives reason to question follows: the practical ability of implementing a more • First, consider raising the $8 billion base amount, complicated structural surplus rule. In fact, which has not been revised since the introduction the excellent performance of the current rule of the rule in 2010. Applying the growth of is itself an argument for making only modest nonoil GDP since 2010 to this amount would modifications. That is the main recommenda- adjust the base amount to about $11 billion. tion of this report. Chapter 3 How a Changing World Would Affect a Changing Kazakhstan This chapter simulates different global eco- in an economy on the cusp of high-income sta- nomic scenarios and how they might affect tus will depend on the ability of companies to key variables in the Kazakh economy. Section innovate, which requires a well-educated and one describes the baseline scenario, with total highly skilled labor force, sound institutions factor productivity (TFP) expected to grow that guarantee property rights, and a good steadily at 2.5  percent a year, the oil price to investment climate. stabilize, and developing countries to consti- To construct the baseline, assumptions tute a growing share of the oil market. Section about the rest of the world can strongly influ- two presents three extreme shocks: continu- ence Kazakhstan’s export earnings and domes- ing deterioration in the Euro Area, prolonged tic economic conditions. Overall, we assume fiscal uncertainty in the United States, and that the real price of oil will stabilize in the declines in China’s investment rates and com- long run as the current massive increases in modity prices. Section three presents long-term demand from major consumers, like China and scenarios for the Kazakh economy, considering India, stabilize as they too attain higher per how key strategic macroeconomic variables capita incomes. Even so, developing countries would behave under an optimal resource man- will constitute an increasing share of the world agement strategy.6 oil market. Developing country oil demand is expected to exceed that of high-income coun- The baseline scenario tries by 2023 and account for 60  percent of Living standards in the long term depend on global oil demand by 2050 (figures  3.1 and the pace of productivity growth, the ability of 3.2). China and India together are expected to the economy to create jobs, and the absence account for nearly half the world oil demand of large swings in economic activity. Whether in 2050. productivity growth averages 2.0  percent a In the baseline, TFP is assumed to grow year or 2.5 percent makes a huge difference— steadily at 2.5 percent a year over the next six under the first assumption, per capita GDP in years and then decline gradually to 2.0 percent Kazakhstan could reach as much as $100,000 from 2025 onward. Prior to 1995, TFP growth by 2050; under the second, $130,000.7 Larger in Kazakhstan averaged 0.5 percent a year. It extraction of natural resources and higher then increased to 4.5 percent over 1995–2005. government spending could lead to a pace of The baseline used in the simulations assumes capital accumulation that for a while may offset an average for these two distinct periods. This weak productivity growth. But in the absence is an optimistic assumption compared with of productivity growth, sustaining high income the past, but not out of line with what might will be a challenge, and job creation would be reasonably expected in a post-Soviet envi- come only at a high cost. Productivity growth ronment where investment can be expected 17 OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN Figure Real price of oil to stabilize in the long run 3.1 Nominal oil price growth Real oil price growth Real oil price 60 125 40 100 2005 $ per barrel 20 75 Percent 18 0 50 –20 25 –40 0 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 Source: World Bank staff calculations. Figure Strongest growth in oil demand to come from developing countries 3.2 Developing countries Developing countries’ High-income countries World share of world demand 10 60 5 40 Percent Percent 0 20 –5 0 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 Source: World Bank staff calculations. to be more productive than it was during the significantly since the 1990s and remains close immediate pre- and post-independence years. to 0.5 percent a year.10 At 2.0–2.5 percent, TFP growth is on par with The optimistic assumption for the baseline the average for developing countries (includ- can be realized only if investment remains high ing China), but is higher than the average in Kazakhstan and is complemented by struc- 1  percent for the six oil-exporting countries tural reforms likely to increase TFP growth from the Gulf Cooperation Council (Bah- in the medium to long term.11 Evidence on rain, Kuwait, Oman, Qatar, Saudi Arabia, and the main determinants of TFP growth sug- United Arab Emirates). 8 For Malaysia, TFP gests that the factors that matter the most in grew on average 4  percent a year between oil- producing countries include a high initial 1995 and 2004.9 Research on the G7 countries income per capita, the quality of education, over 1960–95 shows that their TFP growth the size of the government (the smaller the was higher in the 1960s (at close to 3 percent better), macroeconomic stability, trade open- a year) when they invested more in new tech- ness, positive terms of trade, and good institu- nologies. For the United States, TFP responded tions.12 These factors are closely related to the positively not only to investments in new tech- recommendations of the Kazakhstan Country nologies but also to the entrance of baby boom- Economic Memorandum, which postulates ers in the labor market in the 1960s and 1970s. that Kazakhstan’s path to greater prosperity TFP growth in the G7 countries has slowed and diversification will be paved by reforms that reduce volatility in the economy, improve Kazakhstan’s real GDP per capita is forecast the quality of education (and thus contribute to increase from $12,500 in 2013 to $23,000 to productivity growth), and promote improve- by 2023, reaching $105,000 by 2050.14 Kazakh- ments in the quality of institutions and in the stan’s GDP per capita is now about 30 percent provision of public services.13 of the per capita income in high-income coun- With rising and then stabilizing world oil tries, but is forecast to rise to 41.3 percent in prices, the oil sector will remain a reliable 2025 and to 70 percent by 2050. As a share of source of revenue. Kazakhstan is projected the per capita income in high-income countries to see some moderation of growth but with in 2012, it rises to 46.5 percent by 2023 and to 19 continuing heavy reliance on oil (figure 3.3). almost 83.8 percent by 2030. Even though most Kazakh oil production now contributes about of the growth will be from the nonoil sector, oil 21  percent of GDP, and nonoil production will continue to play a major role in economic 79  percent. Through 2020, oil production is activity and exports and as a source of tax reve- expected to rise 6.8  percent a year and then nues. By 2050, the share of oil in nominal GDP 3.0 percent a year. The remainder of the econ- would decline by only 4.4  percentage points omy is projected to grow by 4.4 percent a year. (from 25.4 percent of GDP in 2013 to 21.0 per- With real oil prices broadly stable, oil’s share in cent). The tax ratio would be off the highs of the real economy is forecast to decline gradu- 58 percent in 2020, but at 49.5 percent in 2050, ally from 21.3 percent in 2013 to 19.5 percent still higher than it is today (45.9 percent). in 2050. Continuing reliance on oil exposes the Achieving this fairly optimistic baseline country to vulnerabilities that cannot be forecast is not guaranteed. Its success is predi- forecast with precision. Maintaining a bud- cated on several assumptions: get surplus of about 1.5  percent of GDP over • The United Nations population and labor the long term would allow total government force growth rates suggest that the 15–64 expenditure to rise 13.5–14.5  percent a year age cohort will grow by 0.9 percent a year. (or between 4.5  percent and 6.0  percent in • The capital stock expands by 4.2 percent a real terms). The direct contribution (ignoring year, requiring real growth of some 4.7 per- second round multipliers) of government out- cent in real fixed investment, substantially lays to GDP growth is estimated at more than below the 12.5 percent a year in the 10 years 20  percent (or almost 1  percentage point of ending in 2011. the 4.2 percent average growth expected over • Strong productivity growth and oil sector 2013–50). This continuing heavy reliance on performance are the main factors likely to oil not only makes the country vulnerable to improve Kazakhstan’s economic prospects. adverse developments in the oil market but also Figure The baseline forecast suggests continuing (but declining) reliance on the oil sector 3.3 Oil exports’ share of total exports Oil tax revenue’s share of total tax revenue Oil sector’s share of nominal GDP growth 75 50 Percent 25 0 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 Source: World Bank staff calculations. OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN increases the local economy’s dependence on Impact of a deterioration of government—something unsustainable in the conditions in the Euro Area long run. A major Euro Area crisis could hurt Kazakh- stan in different ways. Measures over the past Three extreme negative shocks several years, including those by the European The world economy has been through a turbu- Central Bank, have restored confidence in lent period, and Kazakhstan is exposed to exog- the Euro Area and reduce the likelihood of a enous shocks. Three were chosen to illustrate serious deterioration. But the recent Cypriot 20 the possibility for external events to seriously bank bailout shows a significant deterioration affect Kazakhstan’s economic performance: remains possible. a serious Euro Area crisis resulting from a The Euro Area scenario illustrates the financial breakdown, a United States–centered impacts on global growth of a deterioration shock stemming from fiscal paralysis in the that freezes two smaller Euro Area economies U.S. government, and a China-centered shock out of international capital markets. This in stemming from a collapse of Chinese invest- turn forces a sharp decline in government ment and thus demand for Kazakh exports. and business investment spending (equal to These scenarios, chosen by the authorities, are around 9  percent of the GDP of each coun- aimed to illustrate how their materialization in try). The shock is spread over two years, with the short term could possibly affect the Kazakh three- quarters of it felt in 2013 and a quarter economy over the short and medium terms. in 2014. That reduces average growth in devel- Over the medium term, the effects of these oping countries by 1.1  percentage points in three extreme scenarios on Kazakhstan would 2013. As economies gradually recover, the over- be different. The detailed simulations in the all impact declines—but developing country background paper on different scenarios for GDP would still be 0.3  percent lower than in the world economy show that the worst impact the baseline even two years after the simulated would be felt in Kazakhstan if Chinese invest- crisis begins. Important transmission mecha- ments were to collapse. In the event of a slow- nisms and some of the vulnerabilities of devel- down in the Chinese economy accompanied oping countries include: by a 10  percent drop in investments, China’s • Remittances to developing countries could GDP growth would slow by 3 percentage points. decline by 1.7 percent or more, representing In the World Bank’s global macroeconomic as much as 1.4 percent of GDP for countries model, this would hurt the world economy and heavily dependent on remittances. cause spillover effects that could hit Kazakh- • Tourism, especially from high-income stan through a reduction in the demand Europe, would be reduced, with significant for oil and other commodities. As a result, implications for countries in the Caribbean Kazakhstan’s share of oil exports would drop and North Africa. from an estimated 57 percent in the baseline • Short-term debt for many developing countries estimate for 2010 to 38 percent by 2050 lead- has come down, in part because of Euro ing to a significantly slower growth in GDP Area deleveraging. But countries with high per capita. While in the baseline Kazakhstan’s levels of debt could be forced to cut into gov- GDP per capita is estimated to reach $105,000 ernment and private spending if financial by 2050, the combined impacts of a slowdown flows to riskier borrowers become scarcer. in the Chinese economy would mean that by • Commodity prices. The weakening of global 2050 Kazakhstan’s GDP per capita would reach growth in the Euro Area causes a 7.5  per- $98,000. This contrasts with the milder effects cent decline in oil prices and a 7.4 percent expected from a deterioration in the Euro decline in metal prices. Such declines are Area, which would imply in GDP per capita likely to cut into government revenues and of just under $105,000—and an intermedi- incomes for oil and metal exporters, but ate impact from slower growth in the United cushion the blow for oil importers. States, which would bring Kazakhstan’s GDP • Banking sector deleveraging. A crisis could per capita to $101,000 by 2050.15 accelerate bank deleveraging in Europe, with economies in Europe and Central Asia commodity prices and the domestic loss of con- most likely (among developing countries) to fidence (table 3.1). be affected. Kazakhstan’s GDP would decline 2.5  per- Impact of continuing fiscal cent in 2013 and 1.6 percent in 2014—relative uncertainty in the United States to baseline—with its current account deterio- Prolonged uncertainty about the size and length rating by 1.5 percent in 2013 (and its fiscal bal- of the U.S. fiscal adjustment could dampen the ances by 0  percent) and 1.4  percent in 2014. world economy. Although the January 2013 fis- Real activity in Kazakhstan is principally hit by cal deal resolved most of the tax side (the larger 21 a lack of confidence, as domestic households portion) of the fiscal cliff, the spending portion cut back on spending, while business delays of the fiscal cliff was not resolved. The seques- investment. Nearly 75  percent of the 2.5  per- ter was adopted at the end of February 2013 as cent decline in real GDP in 2013 (1.9 percent) Congress was unable to reach agreement on comes from the decline in confidence. With expenditure cuts. The fiscal compression (about global activity and import demand falling, 1.8 percent of GDP) forms part of the baseline Kazakhstan’s exports also come under pres- and helps explain why, despite a strengthen- sure. In this context, trade contributes about ing private sector, U.S. growth is projected to 0.3 percentage point to the overall decline in remain just below 2.0  percent in 2013. While GDP. Lower global growth also reduces the this is the most likely outcome, the uncertainties demand for commodities, exerting downward generated by failure to agree on a medium-term pressure on commodity prices, which also plan and the kicking in of debt-ceiling legisla- reduces Kazakhstan growth by 0.3  percent- tion could have significant additional downside age point in 2013 and 0.2  in 2014. For the effects. In a worst case scenario, where debt is current account, the commodity price impact downgraded again, such uncertainties could dominates, and for the fiscal balance, the dete- weigh heavily on investment and consumer rioration is caused almost equally by falling durable spending in the United States. Table Impact of Euro Area crisis 3.1 Real GDP Current account balance Fiscal balance (percent change in level) (change, in percent of GDP) (change, in percent of GDP) 2013 2014 2013 2014 2013 2014 Euro Area –2.3 –1.4 1.3 0.4 –1.0 –0.8 World –1.3 –0.8 0.0 0.0 –0.7 –0.5 High-income countries –1.4 –0.9 0.1 0.0 –0.8 –0.6 United States –1.0 –0.6 0.0 0.0 –0.6 –0.4 Developing countries –1.1 –0.7 –0.2 0.0 –0.5 –0.3 Low-income countries –0.6 –0.4 –0.2 –0.1 –0.1 –0.1 Middle-income countries –1.1 –0.7 –0.2 0.0 –0.5 –0.3 Developing oil exporters –1.3 –0.9 –0.9 –0.2 –1.0 –0.7 Developing oil importers –0.9 –0.6 0.2 0.1 –0.2 –0.2 East Asia and Pacific –1.0 –0.7 0.2 0.1 –0.1 –0.1 Europe and Central Asia –1.3 –0.9 –1.0 –0.3 –1.1 –0.8 Latin America and Caribbean –1.2 –0.8 –0.2 0.0 –0.7 –0.4 Middle East and North Africa –1.0 –0.7 –0.9 –0.3 –1.4 –0.9 South Asia –0.5 –0.3 0.5 0.1 0.0 0.0 Sub-Saharan Africa –1.0 –0.7 –1.2 –0.4 –0.8 –0.6 Kazakhstan –2.5 –1.6 –1.5 0.0 –1.4 –1.1 Commodity effect –0.3 –0.3 –1.6 –0.3 –0.7 –0.3 Confidence effect –1.9 –1.1 0.1 0.3 –0.6 –0.7 Slowdown in Portugal and Spain –0.3 –0.2 –0.1 0.0 –0.1 –0.1 Source: World Bank staff calculations. OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN In an unlikely scenario where repeated fail- contributes 1.6  percent in 2013 and 1.2  per- ures to set U.S. fiscal policy on a sustainable cent in 2014 to the decline in real GDP. Lower path result in a downgrading of debt, growth global growth also reduces the demand for in the United States could slow by some 2.3 per- commodities, which puts downward pressure centage points—pushing it into recession. The on commodity prices, reducing Kazakhstan’s Euro Area would be pushed into a deep reces- growth by 0.3 percent in 2013 and 0.4 percent sion, potentially increasing the risk of a second in 2014. For the current account, the commod- crisis there, and developing country GDP would ity price impact again dominates, contributing 22 decline by 1 percent, relative to baseline. The nearly 1 percent to the overall deterioration in bulk of these effects come from an assumed both 2013 and 2014. increase in precautionary savings of U.S. busi- ness and consumers of 1.0 and 2.0 percentage Impact of an abrupt fall in China’s points, of firms and households in other high- investment rates and commodity prices income countries of 0.5 and 1.0  percentage An ever-growing fear is that the Chinese econ- point, and of 0.3 and 0.7 percentage point in omy might lose steam and bring commodity developing countries (table 3.2). prices down. China has, on average, recorded Under this scenario, Kazakhstan’s GDP will close to 10.0 percent annual growth for more decline by 1.9 percent in 2013 and 1.7 percent than 30 years and 10.3  percent growth dur- in 2014, while its current account balance will ing the first decade of this millennium, with deteriorate by 1.5 percent of GDP in 2013 and growth as high as 14.2 percent in 2007. During 0.5  percent in 2014, and its fiscal balance by most of this high-growth period, investments 1.2 percent of GDP in both 2013 and 2014. The (and savings) were at a high 30–35 percent of major transmission mechanism to Kazakh- Chinese GDP. In the 2000s, investment rates stan’s output is the “confidence” channel, as jumped initially to 40 percent of GDP (partly in domestic households cut back on expendi- reaction to the low cost of international capital) ture and businesses postpone investment. It and then again to 45 percent of GDP, because Table Impact of U.S. fiscal paralysis 3.2 Real GDP Current account balance Fiscal balance (percent change in level) (change, in percent of GDP) (change, in percent of GDP) 2013 2014 2013 2014 2013 2014 United States –2.0 –2.0 0.5 0.5 –0.4 –0.7 World –1.3 –1.2 0.0 0.0 –0.5 –0.6 High-income countries –1.4 –1.4 0.1 0.1 –0.6 –0.7 Euro Area –1.1 –1.0 0.2 0.0 –0.4 –0.5 Developing countries –1.0 –0.9 –0.1 0.0 –0.4 –0.4 Low-income countries –0.7 –0.5 0.1 0.0 –0.2 –0.2 Middle-income countries –1.0 –0.9 –0.1 0.0 –0.4 –0.4 Developing oil exporters –1.1 –1.1 –1.0 –0.6 –0.9 –0.9 Developing oil importers –0.9 –0.8 0.3 0.2 –0.2 –0.2 East Asia and Pacific –1.1 –1.0 0.4 0.3 –0.2 –0.2 Europe and Central Asia –0.9 –0.8 –0.9 –0.6 –0.8 –0.8 Latin America and Caribbean –1.1 –1.0 –0.3 –0.2 –0.6 –0.5 Middle East and North Africa –0.7 –0.7 –0.7 –0.4 –1.3 –1.2 South Asia –0.4 –0.4 0.6 0.2 0.0 0.0 Sub-Saharan Africa –0.8 –0.8 –1.1 –0.8 –0.7 –0.7 Kazakhstan –1.9 –1.7 –1.5 –0.5 –1.2 –1.2 Commodity effect –0.3 –0.4 –1.5 –0.8 –0.7 –0.5 Confidence effect –1.6 –1.2 0.0 0.2 –0.5 –0.6 Fiscal cliff –0.1 –0.1 0.0 0.0 0.0 0.0 Source: World Bank staff calculations. of China’s fiscal and monetary stimulus dur- would likely have serious consequences world- ing the global financial crisis. As a result, the wide. Simulations suggest that a 10 percentage contribution of investment to Chinese growth point drop in Chinese investment would slow rose from 2.3  percentage points during the Chinese GDP growth by about 3  percentage 1980s and 1990s to around 5.0  percentage points. The high import content of investment points in the 2000s. And China’s capital-output implies that a significant share of the slowdown ratio, which for an economy in a steady-state leaks out as reduced imports—reducing the growth equilibrium will be broadly stable, has impact on China but extending it to the rest of increased since 2000 by 20 percent and is still the world. 23 rising rapidly. Such a strong decline in investment rates High investment rates are required to sus- is, however, unlikely. An abrupt deteriora- tain the capital stock in a fast growing economy tion in the investment climate would likely like China’s. But such high investment-to-GDP elicit a strong policy response. A 5 percentage ratios are unprecedented. Neither Japan nor point decline in Chinese investment growth Korea—two countries that also enjoyed long would result in a 6.0  percent decline in Chi- periods of high growth—ever saw investment nese imports and a 1.4 percent decline in GDP rates exceed 40 percent. A level of 35 percent of relative to baseline (table 3.3). Lower Chinese GDP is seen to be more sustainable and consis- imports in turn reduce global exports, and tent with underlying productivity growth and world GDP declines 0.5  percent from 2013 population growth. baseline and 0.3 percent for developing coun- China’s authorities have identified the need tries excluding China. Reflecting the compo- for a more balanced pattern of investment. sition of Chinese import demand, high- and This would require not just a lower investment middle-income countries are hit harder than rate, but also higher investments and expendi- low-income countries. tures in services and in such intangible assets Under this scenario, Kazakhstan’s GDP as human capital. This can be achieved, in would decline by 0.9  percent in 2013 and part, by reducing implicit subsidies that favor 0.7 percent in 2014. Its current account would capital investments over investment in labor. deteriorate by 0.8  percent of GDP in 2013 The gradual rebalancing and reduction in and 0.5 percent in 2014, and its fiscal balance physical capital investment rates is expected to by 0.5  percent of GDP in 2013 and in 2014. be compensated by faster consumption growth Although China is a larger direct trade part- over an extended period. China’s economic his- ner of Kazakhstan than the United States and tory suggests that it has the instruments, per- European Union are, the impact of a Chinese haps more than any other country, to achieve slowdown on Kazakhstan is smaller due to the such a transformation. But orchestrating such smaller role of China in the global economy a transition should not be underestimated. than of the United States and European Union. Many other countries have failed to smoothly The transmission channel is the demand for adjust their investment profiles. and the price of oil. As the United States and While a smooth transition is the most likely European Union are major global consum- outcome and the one retained in the baseline ers of oil, when their economies come under scenario, the risks are considerable. For exam- pressure, this has a larger impact on the price ple, the transition to a lower investment rate of oil. The lower oil prices, in turn, have a big could be abrupt, perhaps provoked by a fail- impact for Kazakhstan through lower domestic ure of a significant share of new investments to income and private consumption spending. realize hoped-for profits, resulting in a spike in unpaid loans and a rapid tightening of credit. Scenarios for optimal In such a scenario, investment growth would resource management likely come under significant pressure. A general equilibrium model analyzes the Given China’s much increased weight in dynamic consequences of the anticipation and the global economy and its role as an engine realization of oil wealth in Kazakhstan. Differ- of global growth, a sharp decline in investment ent scenarios take into account the uncertainty OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN Table Impact of an abrupt fall in Chinese investment 3.3 Real GDP Current account balance Fiscal balance (percent change in level) (change, in percent of GDP) (change, in percent of GDP) 2013 2014 2013 2014 2013 2014 China –1.4 –0.1 1.7 0.9 –0.3 –0.2 East Asia and Pacific –1.3 –0.9 1.3 0.7 –0.3 –0.2 East Asia and Pacific (excluding China) –0.6 –0.5 0.0 0.0 0.0 0.0 World –0.5 –0.4 0.0 0.0 –0.2 –0.2 24 High-income countries –0.4 –0.3 –0.2 –0.1 –0.2 –0.2 Euro Area –0.4 –0.3 –0.1 –0.1 –0.1 –0.2 Developing countries –0.7 –0.5 0.4 0.2 –0.3 –0.2 Developing countries (excluding China) –0.3 –0.3 –0.3 –0.1 –0.2 –0.2 Low-income countries –0.3 –0.2 –0.2 –0.2 –0.1 –0.1 Middle-income countries –0.7 –0.5 0.4 0.2 –0.3 –0.2 Developing oil exporters –0.4 –0.4 –0.5 –0.2 –0.3 –0.3 Developing oil importers –0.8 –0.6 0.8 0.4 –0.2 –0.2 Europe and Central Asia –0.3 –0.3 –0.5 –0.2 –0.3 –0.3 Latin America and Caribbean –0.3 –0.3 –0.2 –0.1 –0.2 –0.1 Middle East and North Africa –0.3 –0.2 –0.4 –0.2 –0.5 –0.4 South Asia –0.2 –0.2 0.0 0.0 0.0 0.0 Sub-Saharan Africa –0.3 –0.3 –0.6 –0.3 –0.3 –0.3 Kazakhstan –0.9 –0.7 –0.8 –0.2 –0.5 –0.5 Source: World Bank staff calculations. about Kazakhstan’s oil extraction capacity and the oil price and the country’s oil extraction the international price of oil.16 Oil enters the (figure 3.4). The simulations allow for assessing model as both an input of production and as how macroeconomic variables would evolve in an internationally tradable commodity. It is the long run under various scenarios related owned by both the government and house- to the size and long-term value of future oil holds. The role of government is essential in discoveries. This is particularly important for the structure of the model that assumes that fiscal policy and public debt, highlighting the the government is benevolent, in the sense that risks related to the uncertainty of the size of it is interested in maximizing social welfare these oil discoveries (figure 3.5). subject to resource constraints. The govern- ment affects social welfare in three interrelated Simulation results ways. The first is the provision of public con- The growth of GDP per capita fluctuates sub- sumption goods (such as security and justice stantially under the three price scenarios. services, as well as targeted assistance for edu- While it is lower under the baseline price rela- cation, nutrition, and health). The second is tive to the low and high price scenarios until the supply of public capital (such as transport 2015, it is highest under the baseline scenario and communication infrastructure) to comple- during the 2020s. Under the reference oil price ment private capital production. The third is scenario, it is projected to peak at 4.1 percent managing and servicing the public debt, which by 2017 and again at 4.3 percent by 2020 (see can influence directly the cost of capital and figure 3.5). GDP is highly sensitive to oil pro- resource constraints for both the public and jections because the oil sector is extremely the private sectors. important in Kazakhstan’s economy, especially The model is calibrated to fit key charac- if measured in per capita terms. This becomes teristics of the Kazakh economy. This includes clear when comparing the evolution of GDP demographic projections to simulate the path per capita with that of nonoil GDP per capita of various macroeconomic variables for alter- (see figure 3.5), which grows faster in the low native projections of the future evolution of oil price scenario because the economy needs Figure Assumptions for long-term simulations 3.4 Projected demographic structure Projected oil price Population growth Labor force Reference High Low 1.25 70 250 1.00 68 200 $ per barrel 0.75 66 150 Percent Percent 0.50 64 100 0.25 62 50 25 0.00 60 0 2010 2020 2030 2040 2050 2010 2020 2030 2040 2050 Projected oil extraction Reference High Low 3 Millions of barrels a day 2 1 0 2010 2020 2030 2040 2050 Source: Hevia (2012). to compensate for the lower expected oil of GDP. Public consumption, under the refer- wealth by investing more and importing less, ence and low oil price scenarios, is projected to relative to the other scenarios. be about 18 percentage points of GDP (though Private and public investments (as ratios to gradually declining), while under the high oil GDP) are also quite sensitive to different oil price scenario, public consumption is about price scenarios. Along with the behavior of net 23 percent of GDP. These extremely high con- oil exports, the pattern of private and public sumption rates reflect the extraordinary wealth investments helps explain the substantial dif- from oil exports. Indeed, the income effect due ference in nonoil production across different to oil wealth is so large that total consumption oil price scenarios. Wealth and substitution (private plus public) is more than 100 percent effects imply that public investment remains of GDP. larger than private investment when oil prices The fiscal and public debt dynamics under are higher or lower than the reference price. the three scenarios are particularly interesting The investment effort is thus weaker under the from a policy viewpoint. If the price of oil is reference oil price. Higher oil prices lead to a very high, it is optimal for the public sector to decline in the use of oil for domestic produc- run fiscal deficits in the early years, thus also tion in favor of oil exports (substitution), low- raising public debt as a share of GDP. This is ering demand for capital as a complementary due to the public sector’s realization that lower factor of production. In addition, higher pro- taxes are optimal under more optimistic oil jected oil prices imply larger expected wealth windfall scenarios, because tax revenues are for domestic residents, expanding consump- no longer needed to fund current consump- tion relative to investment. tion and investment because the government Both private and public consumption (as can borrow against its newfound oil wealth. ratios to GDP) are projected to remain roughly Fiscal outcomes tend to improve as the oil constant. Under the reference and low oil price windfall is realized over time. Debt dynamics scenarios, private consumption is expected to are driven by the fiscal outcomes. Under the be around 80 percent of GDP. Under the high most optimistic oil price scenario, public debt oil price scenario, it is more than 100 percent increases as a share of GDP in the initial years, OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN Figure Macroeconomic projections under different oil price scenarios 3.5 Reference High Low Growth of GDP per capita Growth of nonoil GDP per capita Growth of GDI per capita 5 5 15 4 4 10 3 3 5 Percent Percent Percent 2 2 1 1 0 26 0 0 –5 –1 –1 –2 –2 –10 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 Net oil exports to GDP Private investment to GDP Public investment to GDP 15 25 5 10 20 4 5 15 3 Percent Percent Percent 0 10 2 –5 5 1 –10 0 0 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 Private consumption to GDP Public consumption to GDP Primary surplus to GDP 120 25 15 100 20 10 80 5 15 Percent Percent 60 Percent 0 10 40 –5 20 5 –10 0 0 –15 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 Public debt to GDP Net exports to GDI Net foreign liabilities to GDI 60 20 90 80 40 10 70 20 0 60 Percent Percent Percent 50 0 –10 40 –20 –20 30 20 –40 –30 10 –60 –40 0 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 2010 2015 2020 2025 2030 GDI is gross domestic income. Source: Hevia (2012). but stabilizes in the later years. In contrast, if about the long-run price of oil as long as the the price of oil is low, the debt burden declines reserves last. because the public sector has an incentive to Given uncertainty over the future trajec- maintain higher fiscal surpluses. It is notewor- tory of both oil prices and oil reserves, smart thy that the high price scenario corresponds to savings and public investment policies should a price that reaches about $200 per barrel by be advocated. Consider the risks of pursu- 2030, but remains between $80 and $150 over ing apparently optimal fiscal and debt poli- 2010–15. That is, what matters for the optimal cies based on specific expectations about fiscal and debt dynamics are the expectations these two variables. If the expectations are inflated, the government might be tempted to due to seemingly optimal public and private pursue overly expansionary public consump- sector increases in consumption. Thus, in the tion in the initial years, before realizing of face of uncertainty, it might be prudent to act the oil windfalls. This would lead to potential under the most conservative expectations for increases in public debt. But if the windfalls the future price of oil and the size of Kazakh- turn out to be smaller than expected, the stan’s oil reserves. And to prevent excessive necessary fiscal adjustments down the road retrenchment of the private investment, a pru- might be severe. Further, this adjustment dent public sector could take measures to save would be needed when the economy’s pro- and invest more than it would under optimis- 27 ductive capacity would have been diminished tic scenarios. Chapter 4 Evaluating the Different Fiscal Rules Using fiscal policies to mitigate the cyclical fluc- business cycle conditions in real time (and tuations of the economy is desirable in theory. therefore bad at estimating structural deficit But in practice there are limits on successful gaps) and equally bad at predicting oil price implementation of these policies because there behavior (the main underlying causal factor in are limits on establishing the cyclical position Kazakh economic cycles). of the economy in real time. Section one briefly Onder and Ley (2013) provide an excellent discussed problems associated with forecasting example of the difficulty in estimating busi- the business cycle. Section two then presents ness cycle conditions. They demonstrate that the results of simulations of the performance of assessing the cyclical position of the nonmin- different fiscal rules in the presence of forecast eral economy is very challenging by contrasting errors and concludes that flexible rules perform best-practice predictions of output growth in worse than fixed rules. Section three explores a cross-section of countries with the final data the results of original work commissioned for for the output gap and growth projections. this report that looks into the expected welfare They use International Monetary Fund (IMF) gains to the economy as a whole of adopting dif- data, which covers more than two decades. For ferent types of rules derived from the perma- each year, the predictions correspond to the nent income approach. The main conclusion: a projections in the previous year’s fall World Eco- fixed rule with an escape clause for countercy- nomic Outlook. These get subsequently revised, clical response, in the same fashion as the one and the final numbers correspond to the most that Kazakhstan is already implementing, gen- recent vintage in the dataset.17 erates welfare gains close to those generated by A comparison of predicted output gaps an optimal fiscal rule. with the final estimates of the same output gaps reveals the difficulties in making accu- Forecasting problems rate forecasts. Figure 4.1 shows that assessing One major unknown in any countercyclical the cyclical position of a nonmineral economy policy is the extent to which cycles may indeed is very challenging. This difficulty is evident be cycles. Volatility in world oil markets has when the real-time predictions of the output become more marked in recent years. But how gap are contrasted with the final data. If pre- long would a cycle have to be before Kazakh dictions were reasonably good, the scatter plot authorities should start to regard the changes should lie along the diagonal, however, as fig- as permanent rather than merely temporary? ure 4.1 shows, the dispersion is very large. The And could such a cycle be accurately identified horizontal axis shows the prediction of the out- in time for appropriate policies to be imple- put gap in real time. This prediction is from mented in a timely fashion? Unfortunately, the the fall World Economic Outlook for the follow- economics profession is quite bad at estimating ing year (close to budget preparation time). 28 Figure Predicted output gap and final output gap for 175 countries, 1990 –2011 4.1 25 20 15 10 29 5 Final gap 0 –5 –10 –15 –20 –25 –15 –10 –5 0 5 10 15 20 Predicted gap Note: The blue plots to the right of the figure show the distribution of the actual output gaps. It is useful to see that they are symmetrically distributed around zero (which means that when one considers all countries and all years, there were both positive and negative output gaps, more or less in the same amount). The red plot on the top shows the distribution of the predicted gaps. This is also reasonably symmetric, showing that government’s measurement errors are not too biased negatively (pessimist) or positively (optimist). Data for this scatter plot come from the International Monetary Fund’s World Economic Outlook and consist of output data from 1966 to 2011 from 175 countries. The dataset contains output figures released biannually, in spring and fall, from 1990 to 2011. Thus, the dataset comprises real-time output figures from 22 different vintages where we consider the one from fall 2011 as the one that contains final figures (with some exceptions). Source: Onder and Ley (2013). For example, 5 percent on the horizontal axis Flexible rules perform worse means that the government expects a 5 percent than fixed rules output gap next year. The vertical axis shows Simulations show that short-term flexible fis- the output gap actually realized for the corre- cal rules designed to be countercyclical can sponding year. If there is a 2 percent value on create greater volatility in the presence of fore- the vertical axis, this means that the actual out- cast errors. Onder and Ley (2013) explicitly put gap was 2 percent. If a specific country pre- simulate how alternative types of fiscal rules in dicted the output gap at 5 percent (horizontal Kazakhstan would perform if the country suf- axis) but the actual output gap was 2 percent fered shocks of the size of those discussed in (vertical axis), this means that the government the previous section. Specifically, they look at made a measurement error of 3  percentage both fixed rules and rules allowing some flex- points. That point would lay off the 45-degree ibility. They analyze each of the alternatives for line because the horizontal and vertical axes its performance smoothing economic volatility have different values. and in maintaining the value of the national oil These results indicate that trying to use a fund over the long run. Their results are quite forecast structural deficit balance as the target striking and very important for the manage- for a cyclically adjusted fiscal policy provides a ment of the fund. It is worth quoting directly very uncertain anchor for the public finances. from this paper: While it has conceptual appeal, in practice it cannot be done with the precision needed. The In conclusion, short-term flexible fiscal observation of Martin Wolf in the Financial rules which are designed to be counter- Times (March 6, 2012) seems appropriate: cyclical are found to be destabilizing in almost all cases under consideration. It does make economic sense to target The main reason for this is the inability cyclically adjusted rather than actual def- of the fiscal authority to determine with icits. But the improvement in economics enough precision as to where in the busi- is at the cost of a reduction in precision. ness cycle the economy finds itself. Given Nobody knows what a structural deficit is. the inevitability of the contemporaneous OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN forecast errors, fiscal policies intended associated with oil revenues. They wrote, “to to smooth the business cycle give rise to be useful, [fiscal] rules need to be easily rep- more volatility on average than delivered licable in terms of their calculation.” Their by adherence to a fixed long-term fiscal optimal rule is indeed complicated, as it would rule. Another consequence of the fore- depend on a variety of factors, including the cast errors that lead to destabilizing fis- volatility of oil revenues, the volatility of pri- cal policy is that the present discounted vate incomes, the persistence of oil revenues, value of the GDP under the flexible rules the public sector’s targeting efficiency, and sev- 30 is less than the one generated by the fixed eral household characteristics such as risk aver- rules. A final shortcoming of short-term sion, discount factors (measuring the extent to fiscal rules is that they cost too much which households value consumption today rel- relative to the fixed ones: they withdraw ative to consumption in the future), and their more from the oil fund on average and minimum subsistence level of consumption. leave behind a smaller stock of assets in The authors compared the welfare gains under the [National Fund of the Republic of the optimal fiscal rule with those under more Kazakhstan] at the terminal planning tractable alternatives. period. The optimal fiscal rule would be complex, as the savings and spending decisions would While short-run flexibility is theoretically depend on many economic and social factors. beneficial, in reality it cannot deliver the For example, it would depend on household results promised. Indeed, active countercyclical “discount factors” (Engel and Montecinos use policy is likely to make volatility worse rather 6 percent) that reflect households preference than better. In addition, they would be very for consuming today instead of in the future; likely to deplete the oil fund relative to what risk aversion; the income of the poorest house- can be expected from a fixed transfer rule. holds relative to subsistence; the volatility of These findings are illustrated in the simula- oil revenues and private income (Engel and tions presented in appendixes 5 and 6. Among Montecinos use 0.41 and 0.13, respectively); other results analyzed by Onder and Ley is the the persistence of oil revenues (Engel and importance of the size of fiscal multipliers in Montecinos use 0.57, based on statistical the performance of flexible fiscal rules. Some- analysis of Kazakhstani data); the probabil- what paradoxically, they find that smaller fiscal ity of facing a deep depression in consump- multipliers can result in higher volatility when tion (based on international evidence, the countercyclical policies are attempted. The authors use a 5 percent probability of facing reason for this stems both from the counter- a 30 percent decline in private incomes); and cyclical nature of the spending and the prob- the targeting efficiency of social transfers, lem of inaccurate forecasting of business cycle which measures the share of public social conditions. When multipliers are smaller it is expenditures across the distribution of house- necessary to use larger amounts of spending to hold incomes. Clearly, such a rule would be get a given countercyclical effect. But if those unlikely to be implemented in practice, partly countercyclical effects are themselves inaccu- because of its complexity. Still, the welfare rately targeted due to inaccurate business cycle gains of implementing such a complex rule forecasting, greater instability can be the result could be substantial. (box 4.1). The rules considered for the simulations were variations of the permanent income Simulating welfare gains from approach. They were the Friedman permanent different fiscal rules income rule (Friedman I) and two variations of The main gains from fiscal rules for oil rev- the latter allowing respectively for optimal mar- enues would come from a countercyclical ginal propensities to spend out of permanent component. Engel and Montecinos (2013) government income and the interest on gov- report the welfare gains for Kazakhstan from ernment assets (Friedman II) and for an “exit following four different types of fiscal rules clause” through which those parameters switch Box Current practice in other resource-exporting countries 4.1 In discussing the possibilities for modifying Kazakhstan’s current rule for transfers from the oil fund to the national budget, it is useful to look at the experience of other resource-rich developing countries and draw some comparisons. Baunsgaard and others (2012) contains a summary of the current state of the debate along with a description of current practice in a selection of resource-rich countries. The paper also surveys the advice given to each of these countries and highlights the somewhat surprising result that permanent income–based rules have played a lesser role in practice than might have been expected given the importance of preservation of wealth for future generations in much of the discussion in academic and policy circles. Rules based on this goal have accordingly been much analyzed but were implemented in only 6 of 17 countries (developed and developing) surveyed, though they were discussed in two others. Dissatisfaction with permanent income–based rules has arisen in some quarters based on the desire of governments to increase 31 capital formation in capital poor countries. Accordingly, efforts to define alternatives that would allow increased investment in current periods have been investigated, though in most countries there remain limits (at least in theory) on the amounts that can be trans- ferred, primarily to prevent economic management problems in terms of overheating the economy, real exchange rate appreciation, and the like. The table in appendix 3 surveys 11 resource-rich countries and specifies the type of rule followed and the rules for transfers from the sovereign wealth fund to the national budget. Of these 11, only 4 have permanent income–based rules, and of these only Norway has adhered strictly to its rule as written. Azerbaijan, Russian Federation, and Timor-Leste have either suspended their rules or simply exceeded the amounts indicated in practice. Of the countries with a resource fund, only three (Costa Rica, Norway, and Timor-Leste) have a flexible rule based on the structural fiscal balance. The remainder has some form of revenue sharing or threshold value rule that does not depend on estimates of the fiscal situation or the business cycle. Even these have not been honored in all cases. While Baunsgaard and others (2012) do not survey the macroeconomic outcomes generated by this fact, the long literature on Dutch disease and related issues points to the potential problems. To the extent that volatility in transfers as well as real exchange rate appreciation can be avoided, there is a case for bringing investment expenditures forward in time to allow capital formation to benefit current generations as well as future ones. Kopits (2004) contains a collection of papers with a variety of “lessons learned” from the experience of other countries: • Fiscal rules need the support of the electorate. In the Kazakh case, it is hard to judge the extent to which the public understands the nuances of fiscal rules or resource management, but support of the legislature is important in legitimizing the decisions made. • It is best to have rules written into the constitution, but even informal rules can work if they are generally accepted. Here the Kazakh case presents a “best case” example but also some retreat in the last year. While the original fixed rule was written into law and well understood, the subsequent loosening of the rule to allow 15 percent deviations was less formal. Regularizing whatever is adopted next will be important. • A fiscal rule works only if there are good procedural rules (accountability and transparency). Here the Kazakh case is exemplary in that there is one sole recipient of oil funds—the National Fund of the Republic of Kazakhstan—and the amounts to be transferred are clearly seen by all. • Emerging markets have lower tolerance for high debt-to-GDP ratios than do developed countries. This is an important consider- ation for the future and the Kazakh authorities have clearly focused on this in their rules regarding the limits on debt issuance by the government. • There is more volatility in emerging markets, and this needs to be taken into account. Indeed, this is the rationale for the cur- rent exercise. • Commodity stabilization funds are good for resource economies where the resource is nonrenewable. The Kazakh use of the oil fund is exemplary in this regard. • Fiscal decentralization needs careful attention to rules. The Kazakh case is not highly decentralized at the present time, making this consideration somewhat less directly relevant. Overall it seems clear that the Kazakh performance to date has been reasonable among resource exporters. So, the need for careful consideration of any proposed changes is obvious. Source: Authors. when private income is below a certain thresh- The comparative results show the gains old (Friedman III). By definition, Friedman I with each of the different types of rules. Table and Friedman II are not countercyclical rules 4.1 shows the gains under a “balanced-budget because savings and expenditures are deter- rule” whereby all oil (annual) revenues would mined entirely with respect to fluctuations in be spent. The table also reports the gains asso- public oil revenues (for example, due to price ciated with the introduction of countercyclical- fluctuations). ity, defined as a measure of the extent to which OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN Table Simple rules 4.1 (Percent) Rule Welfare gain Countercyclicality Optimal 16.4 13.8 Friedman I 5.2 0.0 Friedman II 7.9 0.0 Friedman III 12.6 11.0 32 Source: Engel and Montecinos (2013). oil-financed transfers are countercyclical with fund. In theory, the best way to use such rev- respect to private incomes. All rules of Milton enues is to use them so that the marginal pro- Friedman are labeled in honor of the Nobel- pensities of revenue losses are minimized. For prize winning economist who invented the per- example, if the rate of return from the oil fund manent income hypothesis, in wide use today. in the short term is relatively high, one would Friedman  I is the classic permanent income want to spend more out of current revenues rule, whereby savings and transfers are a lin- than remove savings from a high-yielding fund. ear function of both oil revenues and returns But the gains from this approach appear to be on previously saved assets. Friedman II deter- small, and most sovereign wealth funds might mines savings and transfers as a function of a do this instinctively. linear function of both oil revenues and asset A simple rule with an exit clause could returns but with different coefficients on each. be close to the optimal. This is the scenario Friedman III provides some discretion for the depicted in the last row of table 4.1. It would government to fund transfers out of current have a simple built-in countercyclical compo- oil revenues and returns to assets when private nent embodied in the exit clause. This clause households face a large negative income shock, would provide discretion to the fiscal author- thus introducing some countercyclicality. The ity to tap the oil fund in situations of severe results suggest that the gains under a balanced- private sector recessions that are transitory. budget rule that would basically entail consum- For instance, this scenario would occur only if ing whatever oil windfalls are realized each private income were to drop by 30  percent.18 year would be equal to just over 16 percent of And the spending would end as soon as private national consumption. incomes stabilize. Furthermore, this could be The gains from the simpler rules based done with Kazakhstan’s notably progressive purely on the permanent income approach social targeting, whereby the poorest 20  per- would yield much lower gains than the complex cent of households in per capita income tend rule that takes into account the probability of to receive approximately close to 40 percent of facing deep recessions in the private sector. The social expenditures (table 4.2). gains under the no-rule (or balanced- budget The optimal rule would entail accumulat- rule) scenario are still positive at just over ing savings of close to 25 percent of GDP over 5 percent of the present value of long-term con- the next 25 years. Any such rule would accu- sumption. But, as mentioned, such a rule does mulate savings in an oil-related fund. Such a not have an explicit countercyclical element, fund, operating under the optimal and very because it revolves only around the behavior complex rule, would have accumulated sub- of oil revenues (prices) and not around the stantial resources over the next 25 years.19 private sector business cycle. The third rule in These savings would in principle accumulate table 4.1 corresponds to a slightly more compli- over the current stock of oil savings, but it is cated version of the Friedman rule. It is com- unclear whether the current levels of savings, plicated by an effort to use either current oil taking into account both reserves as well as the revenues emanating from current oil sales or assets of the fund, are optimal from society’s using the revenues from the accumulated oil viewpoint. Table Income and expenditure shares, 2007 4.2 (Percent) Quintile Income share Expenditure share Poorest 9.3 39.6 Second 13.4 19.5 Third 17.1 15.6 Fourth 22.4 11.8 Wealthiest 37.8 13.5 33 Source: Engel and Montecinos (2013). Using fiscal rules to enhance national wel- The basic intuition guiding government fare could be an option. When it comes to fis- expenditures is straightforward. What the cal rules aiming to use revenue windfalls to complex model developed by Engel and Monte- enhance national welfare, Friedman’s perma- cinos suggests is that it would be beneficial for nent income hypothesis plays a central role. the economy to help the private sector smooth Roughly speaking, it states that national wel- consumption by combining a precautionary fare is maximized over time if public expendi- motive with a smoothing of transitory income tures are stabilized around the level that the shocks motive. In that scenario, the govern- public sector’s permanent income can sustain. ment does not only consider its own revenue That is, welfare is maximized when real con- and assets when deciding how much to spend, sumption is stable and unaffected by transitory but also looks at how the private sector is shocks. It is common in the literature to com- doing, spending more when the private sector’s pare the welfare gains from some sort of fiscal income is low. This can work well only when rule that aims to stabilize private consump- automatic stabilizers are well developed and tion at the level justified by the government’s can play their part. According to Engel, Neil- permanent income to a balanced budget rule, son, and Valdes (2011, 14): which implies procyclical fiscal policies. The use of fiscal rules in that fashion could [B]ecause there is income heterogene- have desirable effects in a downturn. Fiscal ity across households, and the govern- rules based on Friedman’s permanent income ment has only a limited ability to transfer hypothesis do not, however, contemplate the income to the poor, the government faces possibility that households within countries a nontrivial tradeoff when implementing differ in terms of their level of income, so they its spending rule: imperfect targeting also differ in their marginal utilities of income. increases the level of expenditure needed In other words, household welfare is affected to achieve a given level of consumption differently if the household is rich or poor. A for the poorest households, which in turn transitory drop in household income might makes the optimal policy more counter- force a poor household to reduce food con- cyclical than if targeting were perfect. It sumption, for example, whereas a rich house- follows that better targeting leads to less hold’s consumption habits might be materially countercyclical government spending, unaffected by a similar proportional (transi- implying that countries that have less tory) drop in income. The background paper capacity to target transfers should run a prepared for this report by Engel and Mon- more countercyclical effort. tecinos (2013) takes into account these dif- ferences in marginal utilities. In their model, In conclusion, the results show that the pro- the national welfare implications of temporary file of public spending can be quite relevant. income shocks across households are deter- The optimal rule leads to the same welfare mined by the public sector’s ability to trans- gain as a 16  percent increase in the govern- fer temporary aid to poor households during ment’s revenues under a balanced budget downturns. rule. A Friedman-type rule attains a third of OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN the gains obtained under the optimal rule an escape clause allows nearly as much gain (5.2 percent). A linear rule in assets and per- in welfare as does a theoretically optimal rule. manent income with optimally chosen mar- That is additional evidence suggesting that the ginal propensities achieves a little under half of current fiscal rule being adopted in Kazakh- optimal gains (7.9 percent). Adding an escape stan already performs well in comparison to clause to the latter rule obtains a little over other more complex options. The recommen- 75 percent of the gains under the optimal rule dations presented in chapter 1—to rebase the (12.6 percent). The degrees of countercyclical- nominal amount in line with the growth of 34 ity of the optimal rule and the linear rule with nonoil GDP over four to five years, add limits an escape clause are similar—and much larger for the implementation of the 15 percent devia- than those of the simpler linear rule without an tion introduced in May 2012, and implement escape clause. automatic stabilizers to the extent possible— These results endorse the conclusions would bring the current fiscal rule even closer advanced in chapter 1: a simple fixed rule with to an optimal rule. Chapter 5 Theoretical Considerations for Countercyclical Policy This chapter discusses theoretical consid- popularity following its successful application erations related to the implementation of in Chile is the structural surplus rule, which countercyclical policy with a special focus targets a long-run government surplus (or defi- on Kazakhstan. Section one presents a brief cit), derived from some version of the long-run review of the literature on the role of fiscal sustainable payout from a sovereign wealth rules on countercyclical policymaking. The fund based on an exportable mineral resource. main message is that there could be significant Such a rule allows countercyclical payouts in a gains associated with the adoption of a coun- downward phase of the cycle to be balanced by tercyclical rule, especially if it is implemented added savings in a boom phase. through strong automatic stabilizers. Section The main pitfall associated with such a rule two discusses the role of multipliers and con- is arriving at a nonpolitical estimate of long- cludes that they will work in connection with run forecasts of export prices, physical produc- automatic stabilizers, but that their final effect tion, and GDP. While smoothing deviations on the economy will depend on the composi- from trend is always the goal of the welfare tion of government spending. Section three maximization, history shows us that politicians concludes this literature review by looking can be counted on to bias estimates in favor of at the role of monetary policy in a downturn, current spending, particularly when they are arguing that for monetary policy to be effective soon to engage in an election. it needs to rely on a well-developed domestic The structural surplus rule can be counter- financial sector. cyclical and can be demonstrated to have wel- fare properties superior to a rule that results Fiscal rules and countercyclical policy in cyclical neutrality or procyclicality. Kumhof There is a vast literature on the role of fiscal and Laxton (2010) extend this by relaxing the rules for countercyclical policy in mineral econ- liquidity constraint imposed by the permanent omies. Frankel (2011) and Arezki (2011) show income approach. After all, the structural sur- how different fiscal rules can make mineral plus rule is countercyclical but is constrained income vary from procyclical to countercycli- by its grounding in some form of the perma- cal to anywhere in between. Clearly, procycli- nent income rule. A key element of this analysis cality is something to be avoided if possible, is that the ability of monetary policy to ade- but it is a natural result of some of the simpler quately smooth welfare for households in the budget rules that can be applied. Simple bal- economy is limited when many households are anced budget rules result in economic insta- liquidity-constrained (their incomes are low bility that mirrors that of the mineral export enough that they live paycheck to paycheck). sector as money is spent as it comes in rather In effect, the government can improve wel- than being saved. Another rule gaining in fare by saving on their behalf out of mineral 35 OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN income—and spending or transferring the that infrastructure or other capital projects are money in cyclical downturns. In theory, the less divisible than are other types of expendi- best way to accomplish this is through direct tures and so are less amenable to countercycli- transfers and taxes on labor and consumption. cal instruments of policy. The case for strong countercyclical fiscal To definitively quantify expenditure and tax policy rests on the targeting of the incomes of multipliers, it would be necessary to estimate a lower income groups as the appropriate goal full-fledged macroeconomic model—which is for policy. The welfare implications of doing beyond the scope of this report. Lacking that, 36 this are clearly better than a balanced budget a review of Kazakh public expenditures and rule, which simply transfers current cyclical ups taxes together with some idea of the import and downs to the government budget and thus component of spending in different areas can to all affected by it. It is also better than struc- go a long way toward classifying expenditures tural surplus rules, which take a longer term as having high or low multipliers. The most view and seek to target a desired long-run gov- recent quantitative analysis of some of the ernment surplus-to-GDP ratio and responds to relevant parameters can be found in Naumov cyclically low (high) government surpluses by (2009), which constructs a social accounting increasing (reducing) government debt rather matrix and compatible general equilibrium than instantaneously changing fiscal instru- model for Kazakhstan to evaluate the impact ments. But this tends toward minimizing fiscal of oil revenue—as well as Hare and Naumov spending volatility. A key insight of Kumhof (2008), which reports some of the findings and Laxton’s paper is that, if minimizing busi- of the more extensive 2009 study. Based on ness cycle volatility should instead be the main Kazakh household and budget surveys from objective, there are significant gains to adopt- 2002 to 2005, this study reports both expen- ing a much more countercyclical rule imple- diture patterns of households and the import mented through strong automatic stabilizers, content of overall sales for different sectors in such as progressive taxation and unemploy- the economy. ment insurance. The sectoral composition of government spending is the most important in the impact The role of multipliers on the domestic economy. But the import con- What makes a fiscal multiplier higher or lower tent of domestic sales in those sectors deter- depends on whether there is enough targeting mines the extent to which subsequent rounds of public spending and excess capacity in the of effects are contained within the country economy. Generally speaking, there are two or are leaked through import expenditures. characteristics for which to look. First, when Household expenditure structure is also a key looking for a large multiplier, a spending tar- element of the extent to which expenditure get is desired that will elicit output and employ- effects are contained or leaked, since payments ment response on the part of the real economy to labor (and therefore household income) rather than simply bidding up prices. This play a large role in determining where second- implies spending money in areas where there ary effects are felt. For example, if the govern- is either excess capacity or additional capacity ment were to spend heavily on construction can be added with relative ease. Second, the projects to increase future productivity, a large subsequent rounds of effects should be concen- share of the payments to the construction trated in the domestic economy (again in areas sector would end up in the hands of workers, with relatively elastic supply response) rather given the relatively high labor content of pro- than being dissipated through leaks out of the duction in this area. The expenditure patterns domestic economy into increased imports. of these workers determine, in turn, whether It is also important to favor expenditures the domestic economy will get a boost (if they that can increase the supply response elasticity. spend in areas dominated by domestic produc- But the likelihood that such expenditures can tion) or whether the effects will be felt only in be easily increased or decreased through the increased imports (if they spend on imported business cycle is lower. This is due to the fact items like cars). Household consumption is heavily weighted content in their own input structure with more toward basic necessities and goods with a high detailed analysis. Transport is a good example, import content. Food and drink are impor- where virtually all transport services are sup- tant, with large shares also going to apparel plied by domestic providers, which purchase and home furnishings (table 5.1). Utilities most of their vehicles and inputs to run them including electricity, gas, and water account for from outside of the country. 6.8  percent, while transport, education, and There is an important caveat to this entire agricultural goods are also important—at 3.7, discussion. Typically, discussions of multipliers 3.5, and 3.4  percent. The roughly a third of revolve around finding spending targets that 37 expenditures remaining is split among various will maximize the extent to which domestic categories. Food and fuels have a high-import demand will be boosted. But in an oil econ- content in terms of domestic market share, as omy susceptible to overheating from excessive do manufactures of all kinds (table 5.2). Other spending of oil revenue, there is a definite sectors have a far lower import share in domes- limit to how much the government will want tic sales, indicating that purchases directed to see domestic spending increase in any given toward these sectors will boost domestic period. It is entirely possible that spending on, demand more in direct effects than will other for example, construction, which shows 25 per- purchases. Some sectors appear to have a fairly cent imported content in final sales, would be high domestic content in direct expenditures an important area for spending given the long- but would almost certainly reveal a high import run priorities of the government. Table Structure of expenditure of the representative household, 2005 5.1 Expenditure Percent Agriculture and related services 3.4 Coal, other solid fuels 1.7 Food, drink, and tobacco 39.8 Clothes and shoes 9.7 Furniture, textiles, cleaning, and home products 5.4 Personal goods, televisions, and computers 1.9 Books, newspapers, and magazines 0.7 Cars and other transport equipment 1.2 Gasoline and fuels 1.3 Other personal use goods 3.5 Electricity, gas, heat and water, and central heating 6.0 Construction and housing repair 1.3 Car repair and maintenance 0.3 Repair of personal goods services 0.3 Hotels and restaurants 2.2 Transport 3.7 Post, Internet, and telecommunications 2.7 Financial and legal services, rent, and insurance 0.4 Personal services 1.5 Education 3.5 Health and medical services 2.4 Public utilities (sewage, water disposal) 0.8 Amusement and recreational services 0.7 Other (pets, plants, and related services) 0.5 Interhousehold transfers 5.2 Tax on land and real estate 0.1 Source: Hare and Naumov (2008). OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN Table Imports and domestic sales, 2002 5.2 (Millions of tenge, unless otherwise stated) Exports/output Imports/sales Total output Exports Imports (percent) (percent) Agriculture 3,800 1,110 100 29 4 Forestry 50 2 5 4 9 Fishery 170 0 1 0 0 Coal, lignite, and peat 1,030 350 9 34 1 38 Crude oil extraction 10,560 7,580 600 72 20 Other mining 2,350 160 100 7 5 Food, close, and tobacco 7,860 260 6,520 3 86 Fuels and chemicals 2,510 750 1,840 30 105 Metals and metal products 7,100 4,410 1,190 62 44 Other manufacturing 1,080 290 3,700 27 468 Electricity, gas, and water 3,490 7 50 0 1 Construction 4,840 1 1,230 0 25 Trade 8,200 2,310 1 28 0 Hotels and restaurants 690 0 0 0 0 Transport 6,740 30 750 0 11 Post and communication 1,080 90 80 8 8 Financial services 7,680 150 1,260 2 17 Public and other services 6,200 340 50 5 1 Total 75,430 17,840 17,486 24 30 Source: Naumov (2009). The need for financial sector development major financial trouble as a result of the crisis, One of the key requirements for effective coun- and many were taken over by the state. A per- tercyclical monetary policy is deepening the vasive problem for the system is the high pro- financial sector. Having a sufficient supply on portion of nonperforming loans still carried by the market so that liquidity can be ensured banks. when monetary policy requires money’s use as Although the authorities have been support- an instrument is important, since it is precisely ing the banking sector through interventions, through such channels that monetary policy restructurings, and extensive liquidity, the sec- must act if it is to be an effective instrument of tor is yet to recover. As discussed in the latest policy in affecting cycles in the real economy. World Bank economic update for Kazakhstan, As is having a deep enough financial develop- reported capital adequacy appears healthy ment so that the links between financial mar- and bank liquidity is ample, but the industry kets and instruments and the real economy are still has depressed profitability due to poor well enough developed that monetary policy asset quality that leads to risk aversion.20 Fur- can hope to affect outcomes on the real side. thermore, nonperforming loans remain high The fewer the options for private sector entre- at 37 percent of total loans at the end of 2012, preneurs to access credit or financial markets, compared with 35 percent at the end of 2011 the fewer the channels for economic authori- (figure 5.1). BTA Bank was the major contribu- ties to expect money and interest rates to pro- tor to this shift, with its nonperforming loans vide a stimulus or a brake. going up from 61.5  percent of total loans in A major impediment to broadly effective April 2012 to 87 percent in December 2012 (fig- monetary policy in Kazakhstan is the fragility ure 5.2). Although nonperforming loans in the of the banking system. This became particu- banking sector are well provisioned (93  per- larly important after the world financial crisis cent of total), they keep a third of banks’ loan of 2008–09 and the end of the Kazakh real portfolio idle, not working for the economy. estate bubble that accompanied the worldwide To address nonperforming loans, the boom before the crash. Many banks got into National Bank established national and Figure Nonperforming loans are well provisioned but constrain banks’ lending 5.1 Nonperforming loans Provisions 40 30 Percent of loans 20 39 10 0 January January January January January December 2008 2009 2010 2011 2012 2012 Source: National Bank of Kazakhstan. Figure BTA remains the biggest outlier in nonperforming loans 5.2 100 Nonperforming loans (percent of total) BTA (December 2012) 75 BTA (April 2012) Nurbank 50 Temirbank KKB Alliance ATF 25 Halyk BCC 0 0 1,000 2,000 3,000 Loans outstanding (billions of tenge) Source: National Bank of Kazakhstan. private asset management companies and The practical and theoretical case for a provided additional incentives for nonper- stable monetary policy using price stability as forming loan writeoffs. The National Bank of a target is well established. This is even clearer Kazakhstan is following its initial strategy of when the country is a large mineral exporter, setting up a national asset management com- in which case the strict inflation targeting pany. The Problem Loans Fund, established approach can be usefully modified to target in April 2012, started buying problem loans the real exchange rate rather than using some (with a discount) from banks on a pilot basis. form of a consumer price index target, as is typ- Four banks have already set up special purpose ical. Another possibility raised in the literature vehicles and are expected to start transferring is to use the primary export commodity price bad loans to their respective vehicles. As part (in this case, oil) as part of a target index of of encouraging the writeoffs, the authorities prices since this is the main source of volatility extended tax exemptions for writeoffs until affecting the economy.22 the end of 2013 and imposed a ceiling for The inability to deal with the effects of large nonperforming loans (over 90 days overdue) foreign exchange inflows can be disastrous. at 20 percent of a loan portfolio in 2013 and This older, “Dutch disease” literature under- 15 percent in 2014.21 scores the role of foreign exchange inflows in OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN bidding up the price of nontraded goods, caus- centralized monetary policies (such as interest ing an appreciation of the nominal exchange rate management) to contain the aggregate rate and thereby causing deterioration in non- demand pressures at the root of the problem mineral traded sectors due to adverse moves are limited by “diversifying” policy to other in the real exchange rate.23 This stagnation instruments that can bear part of the burden. is exacerbated by resource pull effects, which In this situation, a premium remains on cause reallocations away from nonmineral the ability of the government to avoid aggre- traded sectors (agriculture) both to mineral gate demand from overheating in the first 40 (oil) and to nontraded (construction) parts of place. The more oil proceeds that can be the economy. saved offshore, avoiding their transmission to More recently, this analysis has been the domestic economy through fiscal policy, extended to include adverse political and the easier the task of the authorities to main- institutional ramifications of resource depen- tain control over both inflation and the real dence—and to other economic problems that exchange rate. And to the extent that mone- can arise. One of the most important is real tary policy can be made more effective through exchange rate volatility, which can stem from developing a deep and liquid domestic finan- volatility of resource income, but which could cial market, the more real effect will be gained in theory result from anything that causes a from a given change in interest rates, limiting stop-go pattern in absorption or inflation.24 the extent of incipient appreciation from mon- For the Kazakh case, Hausmann (2005b) etary tightening. details some of the pitfalls of a strict infla- In conclusion, this literature review empha- tion targeting approach to monetary policy. sizes general good-practice principles that He walks through the standard response could be useful in a severe recession. First, the when confronted with fiscal expansion, capi- literature on the use of countercyclical policies tal inflows, or oil price shocks. In each case, a highlights that they can generate sizable gains, pure inflation-targeting approach would tend especially if these policies are implemented in to exacerbate short-run fluctuations in the connection with automatic stabilizers. The dis- real exchange rate, thus in effect making the cussion in chapter 1 illustrated how automatic nonoil traded sector the “shock absorber” for stabilizers could be used in severe downturns the economy. This would happen as monetary without increasing the size of the government policy was tightened in response to pressures in the economy. Second, the literature on the on price levels from the increased absorp- use of multipliers emphasizes that they are tion stemming from any of the sources of the most effective when coupled with automatic shocks. Tighter monetary policy implies higher stabilizers. The discussion on multipliers also interest rates and therefore an appreciation of covered the enhanced effects that they could the nominal exchange rate. This is not what a have when used to boost activity in sectors that policy of economic diversification should aim rely more on domestic providers rather than for in the long run. on goods and activities that rely mostly on The proposed alternative is to tighten pol- imported inputs. This chapter reiterated the icy through other means. That is, by applying message advanced earlier that the effective- limits on credit expansion through stricter ness of monetary policy will be limited in the reserve requirements, capital adequacy, or absence of a well-developed financial sector. portfolio regulations and fiscal policy rather These three major principles reinforce the les- than through traditional monetary tighten- sons from the global financial crisis discussed ing. In other words, the need to rely solely on in chapter 1. Appendix 1 Details of the National Fund of the Republic of Kazakhstan Type of fund Stabilization fund. The main goals of the national oil fund are to accumulate part of the revenues from natural resource depletion for future generations (a savings function) and to reduce the dependence on oil to meet budgetary needs (a stabilization function). Furthermore, the overall objective of the fund is to ensure efficient oil revenue management and transparency to increase the welfare of current and future generations. Date of inception January 2001. Rules for contributions Receipts for the national oil fund are generated from tax receipts from the oil and gas sector. (What event or price triggers contributions?) Rules for distributions A guaranteed amount determined in the annual budgeting process, not to exceed a third of the fund’s assets. (What event or price triggers distributions?) The guaranteed transfer should be used only to finance development budget programs designed to invest in projects that will also be used by future generations. Investment rules and restrictions Consists of a stabilization portfolio (minimum 20 percent of the fund) invested in short-term debt securities and an investment portfolio invested in longer term debt and equity securities. How distributions are used Statutory limitations on the uses of revenues are defined by the president and include the following uses: the execution of its stabilization function; earmarked transfers of the fund to republican and local budgets, for purposes to be defined by the president; operational expenditure of the fund and the conduct of the annual external audit; and reliable and liquid investments abroad. The national oil fund may neither be used to provide credits to private or state organizations nor may it provide security for other obligations. Reporting rules and disclosures The fund makes daily, monthly, and yearly reports that it submits to the council. The council prepares an annual report to submit to the president on February 1 of each year. An independent external auditor (selected by the president) also makes an annual report. The report and audit are released on national (state- owned) media. Investment management Management is provided by the National Bank of the Republic of Kazakhstan. A management council oversees the operation of the national bank and vets all reports. The president of Kazakhstan decides the composition of the management council. The overall functioning of the council is at the president’s discretion. The president chairs the management council, which includes the prime minister, chairman of the senate of parliament, chairman of the Majlis of parliament, head of the administration, chairman of the national bank, deputy prime minister, minister of finance, and chairman of the accounting committee for controlling execution of the republican budget. No representatives from trade unions, professional associations, or civil society sit on the council. Other The 2009–10 transfer of $10 billion to finance a crisis response package, including financial support to the top four banks through loans and capital injections, support to the construction and housing sector, financial assistance to the small and medium-size enterprise and agricultural sectors, and public investment in the industrial sector. Source: World Bank (2010). 41 Appendix 2 Survey on International Monetary Fund Advice to a Sample of Resource-Dependent Countries Use of permanent income hypothesis model Included in the Indicator of long-term Use of price-based Use of medium- Main fiscal indicator is Country fiscal framework sustainability fiscal rule term framework the nonresource balance Angola ✔ ✔ ✔ Azerbaijan ✔ ✔ ✔ Cameroon ✔ ✔ Chad ✔ ✔ ✔ Chile ✔ ✔ Congo, Rep. ✔ ✔ ✔ ✔ Equatorial Guinea ✔ ✔ ✔ Gabon ✔ ✔ ✔ Ghana ✔ Nigeria ✔ ✔ ✔ ✔ Norway ✔ ✔ ✔ Papua New Guinea ✔ ✔ Peru ✔ Russian Federation ✔ ✔ ✔ Timor-Leste ✔ ✔ ✔ ✔ Trinidad and Tobago ✔ ✔ ✔ ✔ Venezuela, RB ✔ Note: In some countries, the use of the permanent income hypothesis changed during the reviewed 2004–11 period, so its current use may differ. Source: Baunsgaard and others (2012). 42 Appendix 3 Fiscal Rules in Different Countries Rule permanent Non–permanent income income hypothesis Resource Country hypothesis framework fund Description Azerbaijan ✔ R A nonoil balance guideline (2004) consistent with constant real consumption out of oil wealth. Never observed. More recently, reliance on ad hoc balance budget oil price. Complemented by state oil fund. Chile ✔ F Structural balance guideline (institutionalized in 2006 fiscal responsibility law). Adjustment by long-term rice of copper and molybdenum (10-year forecast) as determined by an independent committee. Targets have been changed over time. Supported by two funds (stabilization and savings). Ecuador ✔ R Various rules (nonoil balance, expenditure growth) that were mostly not observed. More recent rule states that current spending cannot exceed permanent revenue (a sort of “golden rule”). Oil funds abolished in 2008. Equatorial Guinea ✔ R Guidelines establishing that current expenditures should be limited to nonoil revenue have led to very high capital expenditure levels. Economic and Monetary Community of Central Africa convergence criteria include various fiscal targets (such as a nonoil balance target). It has a fund for future generations. Ghana R A recent petroleum revenue management framework built around a stabilization fund and a heritage fund. Benchmark oil revenue is calculated at a seven-year moving average, with 70 percent used to finance the budget. Remaining revenue allocated in fixed proportions to the funds. No fiscal anchor limiting budget deficit. Mongolia ✔ R A ceiling on the structural deficit with structural mineral revenues estimated using a 16-year moving average of mineral prices. Combined with a ceiling on expenditure growth defined by the nonmineral GDP growth rate (useful when structural revenue is growing fast). The structural balance target can be changed every four years. Flows to a stability fund linked to difference between actual and structural revenues. This framework will start in 2013. Nigeria ✔ R Three percent of GDP deficit ceiling for federal government computed at budget oil price (not strictly followed). Budget oil price set every year in political negotiations, including with subnational governments. Excess crude account receives “windfall” revenues: ad hoc withdrawals. Norway ✔ R “Bird-in-hand” fiscal guideline: the cyclically adjusted nonoil central government deficit as 4 percent (the expected long-run real rate of return) of the sovereign welfare fund assets. Guidelines are flexible: temporary deviations permitted over business cycle or if large changes in sovereign welfare fund value. Very strong political consensus. Papua New Guinea ✔ R Five-year medium-term fiscal strategy that sets a ceiling to the nonmineral deficit in line with “normal” mineral revenue. A portion of “windfall” mineral revenue (70 percent) can be spent up to a nonmineral deficit ceiling of 8 percent of GDP. It was largely followed, but there was volatile real expenditure growth due to swings in total GDP. Russian Federation ✔ R The budget code includes a long-term nonoil deficit target of 4.7 percent of GDP that was suspended in 2009. Annual budgets underpinned by rolling three-year medium-term fiscal frameworks. Two oil funds (stabilization and savings). Timor-Leste ✔ F Fiscal guidelines based on permanent income hypothesis framework (constant in real terms). Nonoil balance set in line with estimated sustainable income, which is calculated annually as 3 percent of the sum of the petroleum fund balance and the present value of expected future petroleum receipts. Deficits can exceed the estimated sustainable income if properly justified and approved by parliament. More recently, government has scaled up public investment so that total spending amounts to more than twice the level of the estimated sustainable income. Note: Resource funds can be an account or a statutory legal entry. R is contingent (linked to threshold values) or revenue share (flows in proportion to total revenue) funds; F is flexible (financing, linked to the overall fiscal position) funds. 43 Appendix 4 Economic and Social Indicators, 2005–15 2013 2014 2015 Selected indicators 2005 2006 2007 2008 2009 2010 2011 2012 (projected) (projected) (projected) Income and economic growth GDP growth (annual % change) 9.7 10.7 8.9 3.3 1.2 7.3 7.5 5.0 5.0 5.4 5.5 GDP per capita growth (annual % change) 8.7 9.5 7.7 2.0 –1.4 5.8 6.0 3.5 3.9 4.3 4.4 GDP per capita ($) 3,771 5,292 6,771 8,514 7,165 9,070 11,357 12,009 12,775 13,725 14,749 Private consumption growth (annual % change) 9.7 11.9 10.5 4.4 1.6 10.7 9.7 9.3 6.5 4.4 4.4 Gross fixed investment (% of GDP) 31.0 33.9 35.5 27.5 29.4 25.4 22.2 21.6 21.3 21.1 21.0 Public 4.6 4.5 5.6 5.8 5.5 5.1 4.6 4.6 4.6 4.6 4.5 Private 26.4 29.4 29.9 21.7 23.9 20.3 17.6 17.0 16.7 16.5 16.4 Savings (% of GDP) 29.7 31.4 27.8 32.4 25.9 26.7 29.1 25.9 24.5 24.2 24.0 Public 9.7 11.8 12.4 10.3 7.6 10.4 12.9 12.0 11.5 9.5 9.2 Private 19.9 19.6 15.4 22.1 18.2 16.4 16.2 13.9 13.0 14.7 14.8 Money and prices Inflation, consumer prices (annual % change, end of year) 7.4 8.4 18.8 9.5 6.2 7.8 7.4 6.0 6.5 6.0 6.0 Inflation, consumer prices (annual % change, period average) 7.6 8.6 10.8 17.2 7.3 7.1 8.3 5.1 5.8 5.8 5.8 Central bank refinancing rate (%, weighted average) 7.8 8.6 9.2 10.8 8.4 7.0 7.4 6.2 n.a. n.a. n.a. Nominal exchange rate (end of period) 134 127 121 121 148 147 148 150 149 148 148 Real exchange rate index (2005 = 100) 100 90 81 71 81 76 72 71 69 66 64 Fiscal accounts and external debt/assets Revenues (% of GDP) 28.0 28.3 29.2 29.7 22.7 25.0 27.7 26.8 25.5 25.5 25.1 Oil revenues 9.4 10.0 8.1 12.3 8.1 10.4 14.0 13.4 12.0 12.1 11.7 Nonoil revenues 18.6 18.3 21.1 17.4 14.7 14.6 13.8 13.4 13.5 13.5 13.4 Expenditures (% of GDP) 22.3 20.3 24.2 27.2 23.5 22.1 21.5 22.3 22.5 22.6 22.6 Current 15.8 14.2 14.8 14.2 16.1 15.0 14.9 16.2 16.4 16.5 16.6 Capital and net lending 6.5 6.1 9.4 13.1 7.4 7.1 6.7 6.1 6.1 6.1 6.1 Overall fiscal balance (% of GDP) 5.6 8.0 5.0 2.5 –0.8 2.9 6.2 4.5 3.0 2.9 2.5 Nonoil fiscal balance (% of GDP) –3.7 –2.1 –3.0 –9.8 –8.8 –7.6 –7.8 –8.9 –9.0 –9.2 –9.2 Nonoil primary fiscal balance (% of GDP) –3.3 –1.8 –2.8 –9.4 –8.4 –7.1 –7.3 –8.5 –8.5 –8.6 –8.6 External debt, total (current $ billion) 43.4 74.0 96.9 107.9 112.9 118.2 123.8 139.8 151.7 164.0 175.9 External public debt (% of GDP) 3.8 3.9 2.0 1.6 3.2 3.5 2.9 2.8 3.0 2.9 2.7 Total public debt (% of GDP) 9.8 12.7 7.8 8.7 13.6 15.3 12.4 13.0 16.1 16.9 17.8 Total official international reserves (% of GDP) 26.5 41.0 36.8 35.5 41.2 40.0 38.8 42.7 43.7 44.3 44.5 External accounts Export real growth (%, year over year) 1.1 6.5 9.0 0.9 –11.6 1.9 3.5 –0.4 1.3 4.9 4.9 44 2013 2014 2015 Selected indicators 2005 2006 2007 2008 2009 2010 2011 2012 (projected) (projected) (projected) Import real growth (%, year over year) 12.5 12.2 25.8 –11.3 –16.0 0.9 6.9 19.5 4.0 3.5 4.0 Merchandise exports (current $ billions) 28.3 38.8 48.4 72.0 43.9 61.6 88.5 92.4 90.8 95.4 100.0 Of which oil and gas 19.5 26.3 31.5 48.9 30.0 42.5 65.3 64.5 63.7 67.8 71.8 Merchandise imports (current $ billions) –18.0 –24.1 –33.3 –38.5 –29.0 –32.9 –40.5 –47.2 –48.2 –50.0 –52.0 Services, net (current $ billions) –5.3 –5.9 –8.2 –6.7 –5.8 –7.1 –6.4 –7.7 –6.3 –6.7 –7.1 Workers’ remittances, net (balance of payments, current $ billions) –1.1 –1.9 –2.9 –1.9 –1.4 –1.4 –1.5 –1.5 –1.5 –1.5 –1.5 Current account balance (current $ billions) –1.1 –2.0 –8.3 6.3 –4.1 1.8 13.6 8.8 6.9 7.3 7.6 45 Current account balance (% of GDP) –1.8 –2.5 –7.9 4.7 –3.6 1.2 7.2 4.3 3.2 3.1 3.0 Foreign direct investment (current $ billions) 2.1 6.7 8.0 13.1 10.1 3.7 9.1 12.7 12.2 11.6 9.8 Population, employment, and poverty Population, total (millions) 15.1 15.3 15.5 15.7 16.1 16.3 16.6 16.8 17.0 17.1 17.3 Population growth (annual % change) 0.9 1.1 1.2 1.2 2.7 1.4 1.4 1.4 1.0 1.0 1.0 Unemployment rate (% of labor force) 8.1 7.8 7.3 6.6 6.6 5.8 5.4 5.3 5.3 5.3 5.3 Poverty headcount ratio at national poverty line (% of population) 31.6 18.2 12.7 12.1 8.2 6.5 5.3 4.0 n.a. n.a. n.a. At $1.25 a day (PPP) n.a. 0.4 0.2 0.1 0.1 n.a. n.a. n.a. n.a. n.a. n.a. At $2.00 a day (PPP) n.a. 3.3 1.5 0.9 1.1 n.a. n.a. n.a. n.a. n.a. n.a. Inequality (income Gini coefficient) 0.304 0.312 0.309 0.288 0.267 0.278 0.289 n.a. n.a. n.a. n.a. Life expectancy (years) 66 66 67 67 68 68 69 n.a. n.a. n.a. n.a. Other GDP (current tenge billions) 7,591 10,214 12,850 16,053 17,008 21,816 27,572 30,073 32,259 34,948 37,872 GDP (current $ billions) 57.1 81.0 104.8 133.4 115.3 148.1 188.0 201.7 216.8 235.3 255.5 Doing Business ranka n.a. n.a. n.a. n.a. 74 58  56 49 n.a. n.a. n.a. Human Development Index rankingb n.a. n.a. n.a. n.a. n.a. 69 68 n.a. n.a. n.a. n.a. CPIA overall rating 3.7 3.7 3.7 3.6 3.7 3.7 3.8 3.8 n.a. n.a. n.a. Economic management 4.2 4.2 4.2 4.0 4.2 4.3 4.3 4.3 n.a. n.a. n.a. Structural policies 3.7 3.7 3.7 3.5 3.5 3.5 3.7 3.7 n.a. n.a. n.a. Social inclusion and equity policies 3.5 3.5 3.5 3.6 3.6 3.6 3.7 3.7 n.a. n.a. n.a. Public sector management and institutions 3.4 3.4 3.4 3.4 3.4 3.4 3.5 3.5 n.a. n.a. n.a. a. The Doing Business indicator is ranked out of 181 countries in 2009; 183 in 2010; and 185 in 2010–11. b. The Human Development Index is ranked out of 169 countries in 2010 and 187 in 2011. n.a. is not available; CPIA is Country Policy and Institutional Assessment; PPP is purchasing power parity. Source: World Bank (n.d.). Appendix 5 A One-Time 3 Percent Shock in the Mineral Sector The analysis below describes a single realization the loose rule. In addition, by using the 4 per- of Onder and Ley’s Monte Carlo simulations to cent discount rate to calculate the discounted provide a concrete example of what alternative sum of nonmineral GDP’s in the estimation short-term adjustment rules imply in one partic- period, Onder and Ley find that the moder- ular scenario. In this particular case, the min- ate adjustment rule brings about $0.25 billion eral sector receives a 3 percent growth shock in less than the fixed adjustment rule in terms 2014. This shock is then transmitted to the non- of 2012 values. Similarly, the loose adjust- mineral sector. In the remaining years, both ment rule brings about $0.38 billion less than sectors gradually converge back to their long- the fixed adjustment rule. In comparison, the term trajectories as characterized by the sector- short-term flexible fiscal adjustment rules draw specific propagation mechanisms (figure A5.1). fewer funds from the national oil fund than Three growth rates are traced over time: the the fixed rule: at the terminal year, the latter rate that the government has projected ex-ante leaves $267.7  billion in the oil fund whereas prior to the realization of the shock, the actual the moderate and loose short-term rules leave rate, and the rate the government projects con- $271.7 billion and $272.1 billion. This is mainly temporaneously. As the last rate is reassessed because of positive contemporaneous estima- annually, the forecast error associated with it tion errors. Based on these results, Onder and changes from year to year. Table A5.1 shows the Ley assessed that the flexible short-term rules outcomes under the fixed and each of the two generate higher volatility and lead to low GDP short-term fiscal rule scenarios. benefits. On the positive side, the fund savings The standard deviation of the variation increase under the flexible rules, but these around the potential output is 0.42 for the fixed increases are small. rule, 0.53 for the moderate rule, and 0.61 for Source: Onder and Ley (2013). Figure Oil and nonoil GDP targeting assuming a 3 percent growth shock in 2014 A5.1 Actual Ex-ante projected Contemporaneous estimation 0.12 0.09 0.10 Nonoil GDP targeting Oil GDP targeting 0.08 0.06 0.06 0.04 0.03 0.02 0.00 0.00 2012 2014 2016 2018 2020 2022 2012 2014 2016 2018 2020 2022 Source: Onder and Ley (2013). 46 Outcomes under fixed and short-term fiscal rule scenarios assuming a 3 percent Table growth shock in 2014 A5.1 ($ billions) 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 Nonmineral GDP Fixed 140.0 148.4 156.1 167.6 177.0 186.8 194.1 201.8 209.0 219.7 230.5 Limited adjustment rule 140.0 148.4 155.8 167.7 177.2 186.9 193.8 201.7 208.9 220.0 230.1 Loose adjustment rule 140.0 148.4 155.8 167.7 177.5 187.0 193.5 201.5 208.9 220.3 229.7 Potential nonmineral GDP (ex-ante) 140.0 148.4 157.5 167.9 177.4 187.0 194.3 201.8 209.1 219.7 230.5 47 Expenditures Fixed 40.5 46.6 48.9 51.6 54.1 56.6 58.5 60.5 62.4 65.1 67.9 Moderate 40.5 46.7 47.7 51.3 55.3 57.8 57.3 59.3 61.5 66.3 66.7 Loose 40.5 46.7 47.6 51.2 56.5 59.0 56.8 58.1 61.0 67.1 65.5 Ex-ante planned expenditures 40.5 46.6 48.9 51.6 54.1 56.6 58.5 60.5 62.4 65.1 67.9 National oil fund assets Fixed 43.7 56.3 71.0 87.2 105.0 124.7 147.5 173.6 203.1 234.4 267.7 Moderate 43.7 57.4 72.9 90.2 109.3 130.3 152.2 177.2 206.9 237.2 271.7 Loose 43.7 57.7 73.4 91.9 111.9 133.8 155.1 178.3 207.0 236.6 272.1 Ex-ante national oil fund assets 43.7 56.3 70.1 85.2 101.7 120.0 141.2 165.0 193.0 222.0 253.2 Source: Onder and Ley (2013). Appendix 6 A One-Time –6.4 Percent Shock in the Mineral Sector This appendix describes another realization of fixed rule, 1.01 for the moderate rule, and 1.17 Onder and Ley’s Monte Carlo simulations. In for the loose rule. In comparison, by using a this case, the mineral sector receives a –6.4 per- 4  percent discount rate, Onder and Ley find cent growth shock in 2014. This shock is then that the moderate adjustment rule brings transmitted to the nonmineral sector. In the about $0.38 billion more than the fixed adjust- remaining years, both sectors gradually con- ment rule in terms of the 2012 values. Simi- verge back to their long-term trajectories as larly, the loose adjustment rule brings about characterized by the sector-specific propaga- $1.02 billion more than the fixed adjustment tion mechanisms (figure A6.1). rule. But the short-term flexible fiscal adjust- The left panel in figure  A6.1 depicts the ment rules draw more funds from the national evolution of mineral sector GDP growth in oil fund than the fixed rule: at the terminal response to the shock. As the shock is transmit- year, the latter leaves $223.6 billion in the oil ted to the nonmineral sector, the right panel fund whereas the moderate and loose short- shows that contemporaneous estimations term rules leave $216.3 billion and $206.5 bil- magnify the actual cyclicalilty (overpessimism lion. These results show that the f lexible during economic downturn and overopti- short-term rules increase the volatility, and mism during recovery). Table  A6.1 shows the although they contribute to the GDP more outcomes under each short-term fiscal rule than the fixed adjustment rule case, this comes scenarios. at a substantial cost in the form of reductions The standard deviation of the variation in savings in the fund. around the potential output is 0.89 for the Source: Onder and Ley (2013). Figure Oil and nonoil GDP targeting assuming a – 6.4 percent growth shock in 2014 A6.1 Actual Ex-ante projected Contemporaneous estimation 0.12 0.08 Nonoil GDP targeting Oil GDP targeting 0.08 0.06 0.04 0.04 0.00 0.02 –0.04 0.00 2012 2014 2016 2018 2020 2022 2012 2014 2016 2018 2020 2022 Source: Onder and Ley (2013). 48 Outcomes under fixed and short-term fiscal rule scenarios assuming a – 6.4 percent Table growth shock in 2014 A6.1 ($ billions) 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 Nonmineral GDP Fixed 140.0 148.4 154.5 167.3 176.5 186.7 194.0 201.7 209.0 219.7 230.4 Limited adjustment rule 140.0 148.6 154.1 167.7 176.5 186.8 194.0 201.6 209.2 219.7 230.2 Loose adjustment rule 140.0 148.6 153.7 168.3 176.5 187.2 194.2 201.6 209.2 219.8 230.2 Potential nonmineral GDP (ex-ante) 140.0 148.4 157.5 167.9 177.4 187.0 194.3 201.8 209.1 219.7 230.5 49 Expenditures Fixed 40.5 46.6 48.9 51.6 54.1 56.6 58.5 60.5 62.4 65.1 67.9 Moderate 40.5 47.5 47.7 52.8 55.3 57.8 59.7 60.6 63.6 66.3 67.3 Loose 40.5 47.5 45.9 54.3 56.5 59.6 62.0 61.9 64.4 67.5 68.2 Ex-ante planned expenditures 40.5 46.6 48.9 51.6 54.1 56.6 58.5 60.5 62.4 65.1 67.9 National oil fund assets Fixed 43.7 56.3 68.1 81.1 94.9 110.4 128.2 148.8 172.2 197.1 223.6 Moderate 43.7 55.4 68.3 80.2 92.8 107.0 123.6 143.9 166.0 189.5 216.3 Loose 43.7 55.4 70.0 80.7 92.1 104.6 118.9 137.7 158.8 180.9 206.5 Ex-ante national oil fund assets 43.7 56.3 70.1 85.2 101.7 120.0 141.2 165.5 193.0 222.2 253.2 Source: Onder and Ley (2013). Notes 1. The discussion in this chapter draws on that output is obtained by combining several World Bank reports produced dur- two inputs: capital and labor. But he also ing the crisis to analyze the soundness of showed that for the United States, these two Kazakhstan’s policy stance and macroeco- factors of production did not explain out- nomic performance. These include the put well, and he interpreted the remaining Kazakhstan Economic Reports prepared unexplained part as technical change or in March 2009, June 2009, and July 2010, total factor productivity, a measure of effi- as well as the Program Document of the ciency in the use of factors of production. 2010 Development Policy Loan and its 12. Espinoza (2012). respective Implementation Completion 13. World Bank (2013a). and Results Report of February 2012. 14. Unless otherwise indicated, per capita 2. For a detailed discussion of these general income is based on nominal GDP per capita macroeconomic principles before and (weighted according to the Atlas method). after the crisis, see Blanchard, Dell’Ariccia, 15. The results of these simulations and their and Mauro (2010). impacts on the world economy and in 3. Baunsgaard and Symansky (2009). Kazakhstan can be found on the tables in 4. Baunsgaard and Symansky (2009). the annex to the background paper with 5. World Bank (2012). the global scenarios in Volume II. 6. The details underlying the global eco- 16. Hevia (2012). nomic scenarios and simulations and the 17. IMF (2011). full model for optimal resource manage- 18. Engel and Montecinos (2013). ment developed by Hevia (2012) are pre- 19. Engel and Montecinos (2013). sented in the second volume of this report. 20. World Bank (2013b). 7. See the background paper by the Develop- 21. Six banks do not meet this criterion so far: ment Economics Prospects Group in Vol- Kazkommertsbank (25  percent of NPLs ume II for the details of these estimates. over 90 days overdue), BTA Bank (78 per- 8. See Espinoza (2012) for a study of the cent), ATF Bank (43  percent), Alliance determinants of TFP growth in the Gulf Bank (46 percent), Nurbank (28 percent), Cooperation Council countries. and Temirbank (44 percent). 9. Jajri (2007). 22. Hausmann (2005a), Frankel (2005, 2011), 10. See Amador and Coimbra (2007) for and Velasco (2005). an assessment of TFP growth in the G7 23. See, for example, van der Ploeg (2011) for countries. a recent survey, or Corden (1984) and Gelb 11. The modern analysis of long-term growth (1988) for numerous case studies and a and productivity is based on the seminal survey of older literature. paper by Solow (1957). His model assumes 24. Hausmann and Rigobon (2003). 50 References Amador, Joao, and Carlos Coimbra. 2007. “Total as Social Insurance.” Central Bank of Chile Factor Productivity Growth in the G7 Coun- Working Papers 627, Santiago. tries: Different or Alike?” Bank of Portugal, Espinoza, Raphael. 2012. “Factor Accumula- Working Paper 9, Lisbon. tion and the Determinants of TFP in the Arezki, Rabah. 2011. “Fiscal Policy in Com- GCC.” University of Oxford, Department of modity Exporting Countries—Stability and Economics, Oxford Centre for the Analy- Growth.” In Beyond the Curse—Policies to Har- sis of Resource Rich Economies, OxCarre ness the Power of Natural Resources, ed. Rabah Research Paper 94, UK. Arezki, Thorvaldur Gylfason and Amadou Frankel, Jeffrey A. 2005. “On the Tenge: Mon- Sy. Washington, DC: International Monetary etary and Exchange Rate Policy.” Harvard Fund. University, Cambridge, MA. Baunsgaard, Thomas, and Steven Symansky. ———. 2011. “How Can Commodity Exporters 2009. “Automatic Fiscal Stabilizers.” IMF Make Fiscal and Monetary Policy Less Pro- Staff Position Note SPN/09/23, Interna- cyclical?” In Beyond the Curse—Policies to Har- tional Monetary Fund, Washington, DC. ness the Power of Natural Resources, ed. Rabah Baunsgaard, Thomas, Mauricio Villafuerte, Arezki, Thorvaldur Gylfason, and Amadou Marcos Poplawski-Ribeiro, and Christine Sy. Washington, DC: International Monetary Richmond. 2012. “Fiscal Frameworks for Fund. Resource Rich Developing Countries.” IMF Gelb, Alan. 1988. Oil Windfalls, Blessing or Curse? Staff Discussion Note 12/04, International Oxford, UK: Oxford University Press. Monetary Fund, Washington, DC. Hare, Paul, and Alexander Naumov. 2008. “A Blanchard, Olivier, Giovanni Dell’Ariccia, and Study of Changing Income Distribution in Paolo Mauro. 2010. “Rethinking Macro- Kazakhstan Using a New Social Account- economic Policy.” IMF Staff Position Note ing Matrix and Household Survey Data.” SPN/10/03, International Monetary Fund, Heriot-Watt University, Center for Economic Washington, DC. Reform and Transformation Discussion Corden, W. M. 1984. “Booming Sector and Paper 2008/02, Edinburgh, UK. Dutch Disease Economics: A Survey.” Oxford Hausmann, Ricardo. 2005a. “Ensuring a Com- Economic Papers 36 (3): 359–80. petitive and Stable Real Exchange Rate: A Engel, Eduardo, and A. Montecinos. 2013. “Fis- Macroeconomic Policy Strategy.” Harvard cal Rules as Social Insurance: The Case of University, Cambridge, MA. Kazakhstan.” Unpublished document. ———. 2005b. “Kazakhstan’s Macro Challenges Engel, Eduardo, Christopher Neilsen, and Ahead: A Summary of the Views.” Harvard Rodrigo Valdes. 2011. “Chile’s Fiscal Rule University, Cambridge, MA. 51 OIL RULES—KAZAKHSTAN’S POLICY OPTIONS IN A DOWNTURN Hausmann, Ricardo, and Roberto Rigobon. Onder, Harun, and Eduardo Ley. 2013. “The 2003. “An Alternative Interpretation of the National Fund of the Republic of Kazakh- Resource Curse.” In Fiscal Policy Formulation stan and Countercyclical Fiscal Policy.” and Implementation in Oil Producing Countries, World Bank, Poverty Reduction and Eco- ed. Jeffrey Davis, Rolando Ossowski, and nomic Management, Washington, DC, April Annalisa Fedelino. Washington, DC: Interna- 2013. tional Monetary Fund. Solow, Robert M. 1957. “Technical Change and Hevia, Constantino. 2012. “Optimal Resource the Aggregate Production Function.” Review Administration under Oil Windfalls: The of Economics and Statistics 39 (3): 312–20. 52 Case of Kazakhstan.” World Bank, Washing- van der Ploeg, Frederick. 2011. “Natural ton, DC, June 2012. Resources: Curse or Blessing?” Journal of Eco- IMF (International Monetary Fund). 2011. World nomic Literature 49 (2): 366–420. Economic Outlook, April 2011: Tensions from the Velasco, Andres. 2005. “Monetary and Exchange Two-Speed Recovery—Unemployment, Commodi- Rate Policy: An Alternative View.” Harvard ties, and Capital Flows. Washington, DC. University, Cambridge, MA. Jajri, Idris. 2007. “Determinants of Total Factor World Bank. 2010. “Republic of Kazakhstan— Productivity Growth in Malaysia.” Journal of Improvement of the System of Accumula- Economic Cooperation 28 (3): 41–58. tion, Use and Management of the National Kopits, George, and Steven Symansky. 1998. Fund of the Republic of Kazakhstan.” Wash- “Fiscal Policy Rules.” IMF Occasional Paper ington, DC. 162, International Monetary Fund, Washing- _____. 2012. “The Fiscal Management of Nat- ton, DC. ural Resource Revenues in a Developing Kopits, George, ed. 2004. Rules-Based Fiscal Policy Country Setting.” PREM Notes No. 164. in Emerging Markets: Background, Analysis and ———. 2013a. Beyond Oil: Kazakhstan’s Path to Prospects. London: Palgrave Macmillan. Greater Prosperity through Diversifying. Kazakh- Kumhof, Michael, and Douglas Laxton. 2010. stan Country Economic Memorandum. “Simple Fiscal Policy Rules for Small Open Washington, DC. Economies.” International Monetary Fund, ———. 2013b. Kazakhstan: On the Crest of the Oil Washington, DC, June 16. Wage. Kazakhstan Economic Update, Spring Naumov, Alexander. 2009. “An Analysis of 2013. Washington, DC. Kazakhstan and Its Energy Sector Using SAM ———. n.d. World Development Indica- and CGE Modeling.” PhD Dissertation, Her- tors database. Washington, DC. http:// iot-Watt University, Department of Econom- data.worldbank.org/data-catalog/world ics, Edinburgh, UK. -development-indicators.