Selected Issue


Selected Issue

Private Savings in Turkey—Developments and Policy Options1

A. Introduction and Analysis of Developments


1. Over the last decade and half, Turkey successfully stabilized its macro economy. In the aftermath of the 1999–2001 economic crises, the country pursued a highly successful policy of macroeconomic stabilization. Public sector consolidation increased the primary surplus to over 7 percent of GDP at its peak in 2004, reducing public debt from 90 percent of GDP in 2001 to 33 percent in 2015. Inflation decreased from close to 70 percent annually to single digits. The economy grew fast, by almost 7 percent on average between 2002–07 and close to 5 percent on average between 2002–15.

2. At the same time, however, the private sector saving rate decreased significantly, leading to a current account deficit. While the private sector saving rate averaged 18 percent over 1998–20032, it dropped to 9 percent in 2013 and has stayed below 13 percent since 2010. The decrease in the saving rate was particularly pronounced in the years since 2003. Meanwhile the public saving rate stands at around 3 percent, while the investment rate increased from around 17 percent in 2002 to 20 percent in 2014. Thus, domestic savings, private and public, no longer covered investment, opening up a large gap between savings and investments and hence a current account deficit.


Domestic Saving

(Percent of GDP)

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Source: IMF, WEO.

Saving Rates

(Percent of GDP)

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Sources: CBRT; IMF, WEO; and IMF staff calculations.

3. The current account deficit is high and financed by ample capital inflows. Between 2010–15, the deficit averaged over 6½ percent of GDP. Ample capital inflows, intermediated by the financial sector, and financial deepening, eased credit constraints and led to rapid growth of private sector credit and consumption. Hedged external wholesale foreign currency borrowing by the banking sector has become a key feature sustaining loans growth, and corporates net foreign exchange liabilities have risen to US$176 billion at end October 2015.


Saving -Investment Gap

(Percent of GDP)

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Source: IMF, WEO.

Foreign Exchange Exposure of the Non-Financial Corporate Sector

(Billions of USD)

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Source: CBRT.

4. These trends point to the external vulnerability of the Turkish economy and may ultimately prove unsustainable. Staff projects that the current account deficit will remain in the 4 to 5 percent of GDP range on current policies. Reflecting in large part the large share of short-term external debt, gross external financing requirements will continue to exceed a quarter of GDP per year. Turkey’s net foreign asset position, which has already deteriorated by about 25 percent of GDP since 2008, will deteriorate further. Recent IMF analysis (IMF, 2016) suggests the current account deficit is about 1–2 percent higher than warranted by medium-term fundamentals and desirable policy settings.

International context

5. The current Turkish saving rate is low in international comparison. Comparing the Turkish saving rate internationally shows it is considerably below the world wide average. Even when comparing to regions in the world with relatively low saving rates, such as Europe and Central Asia, sub-Saharan Africa or Latin America, Turkey stands out as having a saving rate that is lower still. Comparing Turkey to other countries by income level shows a similar picture. The domestic saving rate is well-below the average of upper middle income countries. Moreover, in other middle income countries, on average, the saving rate increased over time. Turkey hence stands out not only for the level of its saving rate, but also for the decline its saving rate shows.


Saving Rates by Region

(Percent of GDP)

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Sources: Worldbank, WDI; and IMF, WEO.Note: MENA = Middle East and North Africa; SA =South Asia; SSA =Sub-Saharan Africa; ECA =Europe and Central Asia; LAC = Latin America and Caribbea ; EAP =East Asia and Pacific.

Saving Rates by Transition, Takeoff Countries

(Percent of GDP)

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Sources: Worldbank, WDI; and IMF, WEO.Note: The transition countries in our sample are Belarus, Bulgaria, Czech Republic, Estonia, Hungary, Kazakhstan, Kyrgyz Republic, Latvia, Lithuania, Poland, Romania, Slovak Republic, Slovenia, Turkmenistan, Ukraine, and Uzbekistan. The group of takeoff countries includes China ; Chile; Hong Kong, SAR China; Indonesia; Korea; Malaysia; Mauritius; Singapore; Taiwan, China; and Thailand.

6. Maybe the most interesting international comparison is with other emerging market. Here we follow the World Bank (2011) and distinguish between take-off countries that have achieved high and sustained growth rates between 1980 and 2008 and transition countries that had generally seen a collapse in saving during their 1990s transition.3 This comparison shows the Turkish saving rate to be modestly below with that of transition countries, but far below that of take-off countries, let alone China. In addition, the Turkish saving rate is decreasing, whereas take-off countries on average show a slight increase in saving over time.


Saving Rates by Income Group

(Percent of GDP)

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Sources: Worldbank, WDI; and IMF, WEO.Notes: Low-income economies are those in which 2013 GNI per capita was $1,045 or less. Middle-income economies are those in which 2013 GNI per capita was between $1,046 and $12,745. High-income economies are those in which 2013 GNI per capita was $12,746 or more.

This paper

7. The aim of this paper is to provide policy options that would raise the private sector saving rate in Turkey. To do so, it first briefly reviews some of the literature on private savings in order to gauge different savings motives and possible causes of the drop in the saving rate. The paper does not aim to empirically test these different explanations, but reports on others’ findings. Subsequently, the paper presents policy options. The various options are discussed in terms of feasibility, contribution to increased savings, and potential positive or negative side effects (externalities). Importantly, the paper does not discuss issues related to the monetary policy framework and stance. While these issues are certainly directly relevant for private sector savings, they are extensively dealt with in IMF (2015, 2016).

8. The paper does not aim to come up with a single best policy. Rather, it is best seen as presenting a menu of policy options. These are not mutually exclusive and in some cases can be mutually reinforcing. A full impact analysis is beyond the scope of this paper. It is up to the authorities, possibly in cooperation with the World Bank, to study the impact of the various options presented and, taking account of possible side benefits or detriments, implement the best combination of policies to raise the saving rate.

B. Factors Explaining the Private Sector Savings Rate

9. Several macroeconomic factors are generally acknowledged to play a role in the savings decisions of the private sector. They include public saving, the real interest rate, and inflation. Under the permanent income hypothesis, higher public saving is linked to a decrease in private saving through Ricardian equivalence: as private agents expect lower future taxation or higher transfers, they hence need to save less to keep with the same level of future consumption. Higher real interest rates make delaying consumption more attractive by benefitting saving and punish borrowing. Through this substitution effect higher interest rates can be expected to increase savings. In an economy with many poor households, the effect may be less visible, as poor households’ consumption is near subsistence level and hence cannot be decreased to allow room for saving. In addition, an income effect may also be in play, as lenders will increase their income, while borrowers will see their income decrease. Inflation is often seen as a proxy for uncertainty. Volatile and high inflation increases uncertainty about future income, and would suggest an increase in precautionary savings. However, inflation also presents uncertainty about the future value of savings, decreasing incentives to save. Instead, agents may want to invest in long lasting assets such as land, housing, or durables.

10. In the Turkish experience with savings rate, specific factors may have played an important role. These include, the macroeconomic stabilization that took place post-crisis (increased public saving, low and stable inflation and interest rates), a stable banking system that made consumer credit widely available to the general public for the first time, growth and wealth effects, and demographic developments. This paper does not aim to empirically test the importance of these different contributing factors. However, we will discuss these factors briefly below, selectively citing empirical results of other authors.

11. The macroeconomic stabilization that took place post–2001 may have influenced private savings in several ways. First, one can expect a reduced need for precautionary savings as a consequence of reduced economic uncertainty. Partly this can be seen as a form of Ricardian equivalence, where higher public saving leads to the expectation of lower future taxation and hence a lower saving need for the same level of future consumption. However, the effects of economic stabilization can be seen more broadly. Lower inflation and real interest rates implied a greater ability to borrow externally and attract increased foreign direct investment. In addition, the provision of essential services such as health care and education improved, further reducing the need for precautionary savings.

12. At the same time financial development and deepening took off. The banking system was restructured and, from its now stronger position, started to supply credit, thus easing liquidity constraints. To the extent that financial deepening simply allowed economic agents to reach their desired level of leverage, theoretically the effects on the saving rate should be transitory. I.e., the saving rate should revert back to where it was once the new steady state leverage is achieved. However, empirically, it is not clear whether agents behave in this way, or what the desired level of debt is. Taken together, macroeconomic stabilization and financial development implied that more resources became available in the economy, some of which were employed to provide credit and hence crowd in the private sector.

13. In combination with other economic policies, the stabilization triggered strong economic and productivity growth. Productivity growth raises household income. This implies that the income for each cohort in the population will be larger than that of the cohort before it. Assuming a two generation economy, the young generation—which is in the savings phase of the life cycle—enjoys higher income than the old generation in the dissavings phase of the life cycle. Hence, assuming the generations maintain the same saving rate—in order to be able to enjoy the same share of life time income for consumption in retirement—the average private saving rate will increase in a growing economy. In the literature this is known as Modigliani’s (1970, 1986) aggregation effect. Productivity growth also implies households can look forward to higher future—and hence lifetime—income. In anticipation of this higher future income, households will consume more in the current period. This human wealth effect (Tobin, 1967, Caroll and Summers, 1991) would hence imply a lower saving rate. With these two effects in play, link between productivity growth and private saving is an empirical question. And even though cross-country empirical evidence finds a fairly robust positive link between growth and saving, the same studies also suggest the conditions under which the aggregation effect does not occur in practice (Deaton and Paxson, 1994, 1997, 2000). Hence the empirically established positive correlation between growth and saving is difficult to reconcile with standard theories of savings.

14. Turkey is undergoing a demographic transition. It features a relatively large young population (49 percent of the population is under 30 years old) and a modest number of pensioners (just 8 percent of the population is over 64 years old, and an additional 3.7 percent of the population is aged between 60 and 64). Thus the youth dependency ratio is relatively high, while the old-age dependency ratio is low. As fertility has decreased and longevity increased, the country is undergoing a demographic transition. The youth dependency ratio is decreasing gradually over time, while the old-age dependency ratio has started to increase. A reduction in the youth dependency ratio is generally associated with an increase in the household saving rate, as overall spending on the fewer children around will decline. In addition, parents can no longer rely on children supporting them in old age, prompting the need for additional retirement savings. Conversely, an increase in the old-age dependency ratio implies more people are living off their retirement savings, which implies a decrease in the saving rate. In the interim period when the youth dependency ratio is falling fast and the old-age dependency ratio is only increasing modestly, the country enjoys a demographic dividend, consisting of a larger potential labor force and higher private saving to fund investment.

15. Other demographic factors influencing the private saving rate include migration from the country side to the cities and an increase in female labor participation. Between 1980 and 2015, urbanization in Turkey increased rapidly. Urbanization is found to influence the propensity to save mainly through the precautionary savings channel. People in rural areas tend to hold larger precautionary savings than urban dwellers, which is directly related to the generally better availability of government health and other social services in urban areas, and the generally higher income volatility in the (rural) agricultural sector. Hence urbanization can be expected to lower the private saving rate. In tandem with urbanization, female labor participation increased rapidly since 2006, albeit from a low base. This would be expected to raise the saving rate, as households with both partners working generate more income and are able to accumulate more financial wealth over time.




Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Source: WDI.

Female Labor Participation


Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Source: Eurostat.

16. Non-financial wealth accumulation may also influence financial saving decisions. For Turkey, non-financial wealth in the form of home and land ownership and gold are likely relevant factors. Households owning their home or land may see this as a substitute for financial savings. Specifically, a house may be seen as an asset that can provide income in retirement, either by renting the property or selling it for a lump sum. Thus home ownership can be expected to lower a household’s propensity to save. Still, such a wealth effect should ease over time, suggesting the saving rate should increase again over time. In addition, Turkish households traditionally invest some of their savings in gold, and the number of Turks that report gold savings doubled to almost 25 percent of respondents in the three years to 2015 (Hurriyet, 2015). While the exact volume of these savings is not know, external sector statistics suggest that on average, the country imports more gold than it exports, adding to the stock. As with other non-financial wealth, gold stocks may be seen by the households as a partial substitute to financial savings, hence lowering the private saving rate.

17. Which of these factors has been dominant in Turkey in the recent era? Given the myriad different economic links discussed above, this question can only be answered empirically. Van Rijckeghem (2010), Van Rijckeghem and Ucer (2009), IMF (2007), Pirgan, Sabauncu, and Bahceci (2012), and World Bank (2014) represent the most prominent empirical studies trying to explain the post-crisis drop in the private sector saving rate in Turkey. In its recent public finance review, the World Bank (2014) examines household saving behavior in detail. Using Turkish household budget survey data, the study shows that an increase in female labor participation increases a household’s saving rate, the precautionary saving motive is strong, and increases in home ownership lead to a temporarily lower saving rate. The study IMF (2007) employs regression analysis on (inflation-adjusted) macro saving data from 1980 to 2005, and concludes that public saving, growth, and inflation are the key determinants of private saving. It argues that these results suggest that the lower private saving rate chiefly reflects improved economic stability, a notion also supported by World Bank (2014). Van Rijckeghem (2010) and Van Rijckeghem and Ucer (2009) employ macro regressions on data from 19984 to 2007, supported by analyses of household budget survey data. These studies support the conclusion that economic stabilization and a reduction in uncertainty (inflation) have been important, but in addition find a prominent role for credit availability and financial deepening. In addition, they find that demographic changes in Turkey should have led to an increase in the saving rate, and that, going forward, further demographic change will be at best neutral for the private saving rate. Ozcan, Gunay and Ertac (2003) study earlier 1968–1994 data, and conclude private savings behavior exhibits strong inertia and find a positive effect of inflation on savings ascribed to the precautionary motive. They do not find support for Ricardian equivalence or macroeconomic stabilization as factors influencing private saving, which may, however, be an artifact of the period under study. Pirgan, Sabauncu, and Bahceci (2012) empirically test for Ricardian equivalence on more recent data. While they find some evidence of partial Ricardian equivalence, they nevertheless conclude that fiscal policy does not have a prominent role to play in increasing domestic savings.

18. Hence, while many factors seem to play a role in explaining decline in Turkey’s private saving rate, not all of these explanation are consistent with the speed with which decline took place. The fast drop in the saving rate the years directly following the 2001 crisis suggest that the rapid implementation of a thorough macroeconomic stabilization program may have played a large role. The more gradual, but still fast, decline thereafter is consistent with rapid financial deepening, primarily through bank credit becoming available to a large proportion of households. Demographic trends may also have played a role on the margin, but by their nature they materialize slowly over a longer period of time.

C. Policy Options

19. Raising the low private sector saving rate is the preferred way to reduce the current account deficit, and thus Turkey’s external vulnerability. To increase the domestic saving rate, either the public or the private savingsrate (or both) would have to rise. The public saving rate is equivalent to the budget balance of the overall public sector, discussed extensively in IMF (2014). Analysis in the same report shows that increasing the private saving rate delivers better macroeconomic outcomes than focusing purely on increasing public savings. Of course the nature of private savings implies that, in contrast to public savings, it is not under the direct control of the authorities. Nonetheless, policies can be devised that influence the private sector’s saving behavior and thus work towards raising private saving. Such policies can have further positive effects besides reducing the external vulnerabilities (such as improving labor market efficiency or reducing old age poverty). For the sake of this paper, these effects are seen as positive externalities, and will be discussed only briefly. Below, we divide the policy measures into two distinct categories, based on whether they target a decrease in credit growth (dissaving) or an increase in actual gross saving.

Measures to slow down credit growth

20. Since 2001, credit growth in Turkey has been volatile. It has also been high, even when adjusted for inflation. As discussed above, this is partly a consequence of macroeconomic stabilization and the availability of consumer credit that came with a stronger banking system which focused more on the retail sector. It represents financial deepening, which comes naturally with increased wealth and should be seen positively.


Credit Growth

(Percent; y-o-y, 13 weeks MA, FX adjusted)

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

1/includes consumer, commercial and other credits Sources: CBRT; Haver Analytics; and IMF staff calculations.

21. However, the pace of credit growth raises two issues particularly relevant for Turkey. First, fast credit growth reduces net saving. This directly worsens the saving-investment imbalance, and hence increases its current account deficit. Second, fast credit growth may negatively influence credit quality. Cross country evidence (e.g., Kelly, McQuinn and Stuart, 2013, or Fernandez de Lis, Martinez Pages and Saurina, 2000) suggests that excessive credit growth is often associated with credit quality problems down the line. This is as such rapid increases of credit present quantitative and qualitative challenges to the credit scoring systems banks employ. The quantitative challenges simply amount to not having enough qualified and experience personnel to thoroughly probe credit applications, while qualitative challenges are related to data covering only benign periods of economic expansion. The result is that often credit quality deteriorates much faster than expected when the economy enters a downturn.

22. In response, the authorities employed various macroprudential tools to slow down credit growth. The choice for macroprudential measures over raising policy interest rates was driven by considerations that higher interest rates may attract capital inflows, and hence increase external vulnerabilities. The macroprudential measures were primarily geared at preserving financial stability, by targeting a slower, more sustainable pace of financial deepening. Macroprudential measures taken since 2010 are summarized in the appendix of IMF (2016).

23. The measures have successfully slowed down retail credit growth. Retail credit growth started to slow down shortly after the introduction of the main macroprudential measures targeting this market segment (marked in the credit growth figure above). Considering the structurally above-target inflation in Turkey, the current nominal growth rate of 10 percent translates into a real retail credit growth rate of just 3 percent, in an economy that is projected to grow by 3 percent this year. Hence financial deepening on the consumer side virtually came to a halt.

24. Corporate credit growth, however, has remained relatively high at 30 percent on an annual basis, but seems to have decoupled from investment. Credit to the corporate sector is generally seen as positive for the real economy, as long as it is invested and hence serves to boost growth (potential). The authorities believe corporate credit growth supports the economy and therefore are not currently considering measures to curb it. In fact, they have taken measures to support corporate credit growth to the SME segment in particular. However, while domestic corporate bank credit growth increased to over 40 percent year on year in 2013–14, the investment rate remained broadly flat at around 20 percent of GDP. In other words, the link between the increase in the absolute amount of additional corporate credit and the absolute amount of new investment that emerged from the data over the period 2004–10 seems to have changed. If not used for investment, it remains unclear towards which spending purposes the additional volume of corporate credit has been channeled since the link broke down in 2013.


Real Corporate Credit Changes and Investment

(Billions of TL at 1998 constant prices)

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Sources: CBRT, and TurkStat.

25. As with retail credit, the high volume of corporate credit growth year after year may yet result in decreasing credit quality in a downturn or slow growth environment. In such a scenario, specific economic sectors may be affected more than others. For instance, cross country experience suggests that after a number of years of good economic growth, the construction sector specifically often suffers heavily in a prolonged downturn. Given that banks’ exposure to this sector has increased rapidly over the last decade, it may therefore make sense to consider further macroprudential measures targeting this economic sector in particular.


Construction Sector Loans

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Sources: BRSA; Haver Analytics; and IMF staff calculations.

26. Therefore, it is important to identify emerging risks in the corporate credit segment. Risks in specific parts of the corporate sector can be assessed by looking at the general health of economic sectors as well as developments in cash flow, leverage, credit, and exposure to interest rate and exchange rate risks. Assessing economic sectors can be done using standard metrics such as profitability and price and cost developments. More profitable industries that face little pressure on margins present less credit risk than industries with low or volatile profitability. In addition, one can look at the prospects for specific economic sectors going forward, given the overall macroeconomic developments. Cash flow analysis on the level of individual large corporates in certain sector gives and idea of liquidity in this sector, while leverage is an indicator of loss-absorption buffers present in the sector. Credit growth itself is another risk indicator, as discussed above. Finally, the exposure to interest rate risk and exchange rate risk (of an economic sector or individual corporates) is important to get an idea of a sector’s robustness to shocks. Loans with short duration (variable rates or short maturity) or foreign exchange loans increase liquidity and solvency risk.

27. Were elevated risks from strong credit growth to be found in corporates in certain economic sectors, the authorities may want to consider prudential or macroprudential measures for corporate credits. These could be general, or could be targeted to specific parts of the corporate sector where risks are found to be elevated. Such measure could entail a mix of options. A prime prudential instrument is to increase risk weights, either for corporate credit as a whole or for specific segments of the corporate sector. Macroprudential measures for corporate loans could be modeled on the successful measures employed to slow down credit to the household sector. Such measures could include providing guidance on implicit nominal credit growth targets, increasing provisioning, or introducing debt service limits or limits on credit maturity. Consideration should be given to differentiate the measures between credit in Lira and credit in foreign exchange, as the risks differ between these types of credit.

28. Recently introduced policies to provide incentives for equity financing for the non-financial corporate sector will also work in the direction of containing debt finance. A new notional interest deduction for cash capital increases came into effect in mid 2015. Accordingly, the balance between tax-deductible interest on debt finance and taxable profits on equity finance has shifted in favor of equity finance. While the exact magnitude of the shift will remain unclear for some time, certainly less debt finance will imply lower corporate credit growth. Whether this measure will increase the overall corporate saving rate is unclear though.

29. (Macro) prudential measures would be relatively easy to introduce and could quickly have a significant impact. A reduction in corporate credit growth of 10 percentage point per year (i.e., from the current level of 30 percent to 20 percent) would, ceteris paribus, contribute some 0.7 percentage point to the private saving rate. To the extent that corporate borrowing seems to have decoupled from investment over the last few years, a moderate reduction in corporate credit growth may not have large real effects. However, over the medium-term, it seems likely that the historically strong link between corporate borrowing and investment would still apply. Hence restricting corporate credit growth in general may have negative effects on GDP growth as well as potential future growth. For these reasons, measures targeted at specific segments of the corporate sector where risks seem elevated may be more suitable. In addition, a gradual introduction of such measures would allow for calibration along the way, depending on the actual observed impact on credit and growth.

Measures to increase savings

30. Increasing the gross private saving rate presents another way to increase private savings. This can be done in many ways. Here, we focus on three related possible policy areas: pensions, severance pay and subsidized savings.


31. The pension system in Turkey consists of a large first pillar, a small second pillar and a modest but growing third pillar. The first pillar encompasses pay-as-you-go (i.e., unfunded) social security benefits. Participation is mandatory and collection and administration are done by the government through the social security institution. The second pillar consists of pension foundations in some private companies. Employees in these companies automatically participate, but participation could be noncontributory for employees, and funding from the employer is often in the form of (supplementary) dividend payments to the foundation. Such foundations cover only a small fraction of Turkish employees. In addition, some companies feature unfunded book reserves for pensions. The third pillar consists of voluntary pension accounts, introduced in 2003. Employer contributions to the third pillar are tax deductible up to 15 percent of gross salary, capped at the minimum gross wage. Initially employee contributions benefitted from a tax advantage, but in mid 2012 the tax advantage was replaced by a (capped) government matching contribution of 25 percent of investment5. Capital gains in the third pillar are taxed at 5 percent at retirement or 10 (more than 10 years contribution) or 15 percent (less than 10 year contribution) in case of early withdrawal.

32. The funded third pillar currently contributes some 2 percent of GDP to savings and may have boosted the private saving rate by as much as 0.3 percentage points. Third pillar pension funds have grown fast, especially since the introduction of the government matching contribution in 2012. Between mid 2012 and October 2015, the amount of funds invested in third pillar funds has more than doubled to a total of TL42 billion (or 2 percent of GDP). Eren and Genç-İleri (2015) find that the introduction of the third pillar pension system may have boosted the net saving rate by almost 0.3 percent of GDP, and increased the capital stock by some 15 percent. Over the same period, the number of participants also steadily increased to now close to 6 million.


Third Pillar Pensions

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Source: Pension Monitoring Center.

33. The current pension system is too generous. In international comparison, the replacement rate is high and the retirement age is low (OECD, 2006). While reforms to both have been put in motion, the transition periods are lengthy, and the end goals are modest. The reduced replacement ratio, in effect for people entering the pension system after 2008, remains well-above OECD averages for both women and men (Figure 1). Meanwhile, the minimum age at which employees can become eligible for a pension increases only slowly, to reach 65 by 2043. Overall this implies that the current system is very generous, and it will take a long time for this to change. These generous incentives to retire early have negatively impacted the employment rate for older workers.

Figure 1.
Figure 1.

Replacement Rates in Internatitonal Comparison

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001


Employment Rate for People Aged 55-64


Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Source: OECD.

34. Revamping the second pillar would raise savings while simultaneously contributing to pension system sustainability and capital market development. A funded system implies that individuals save for their own retirement. This can be done through individual accounts or collectively, and through defined contribution of defined benefits systems. A defined contributions system using individual accounts is per definition sustainable as pension benefits are paid from accrued pension savings. In contrast, collective defined benefit systems feature some elements of risk pooling. As such systems are not per definition sustainable, the challenge remains to keep (projected) benefits in line with (projected) assets, taking account of developments in, e.g., longevity. A benefit formula based on conservative return projections and taking account of longevity makes for a sustainable fund, while retaining risk sharing features. Saving in a funded system would boost domestic savings. Over time, it would result in large pools of institutionally-managed money available for investment through the domestic market. In fact, in order for a pension fund to yield economic benefit, the investment role is crucial.6 This provides incentives for the development of the local capital market, which should serve to improve the allocation of capital more broadly in the economy.

35. Given Turkey’s relatively young population, now is the time to start funding the pension system. With 32½ percent of people under the age of 20, Turkey has a young population, but fertility rates have decreased sharply over time. The country is hence enjoying a demographic dividend, with many people of working age (68 percent of the population), while the young dependency ratio is falling and the old age dependency ratio is not yet increasing much, as only 8 percent of the people are aged 65 and older. This dividend present an opportunity to save in anticipation of population aging that will inevitably follow if these demographic trends persist. Given their role in providing income into old age, pension savings are generally seen as an appropriate vehicle.

36. Peer country experience suggests that the introduction of a second pillar pension system could yield several percent of GDP in additional private savings. It is beyond the scope of this paper to fully quantify an actuarial pension model for Turkey. Instead, we look at some peer countries that have successfully reformed their pension systems, and the boost to private savings these countries subsequently experienced. While many countries have reformed their pension systems in various ways, the examples of Chile and Mexico stand out as useful peer comparisons with Turkey. Chile, as it pension reform in the early 1980s is generally seen as the first in the big wave of countries reforming their systems. As such, it is now arguably the most mature system among the countries that have reformed, and is often held up as an example. Mexico presents a good peer as its economy is in a similar stage of development as Turkey. It is a solidly middle income country, with a diversified economy, and a large advanced economy in close proximity as its main trading partner, which started a broadly successful transition to a funded pensions system in 1995. Some of the features of both systems are described in Box 1, while the figure in the box illustrates the development of the private saving rate in both economies.

37. The authorities are working towards gradually introducing changes to the pension system. Expanding a pilot program of auto-enrollment of new employees into a funded pension system is on the agenda. Other changes would increase the focus on long-term investment, and include a reduction in overall costs of administration and management. Simplification of the regulatory and supervisory frameworks and the creation of well-designed default options for employees and employers are also among the goals. Such reforms would all go in the right direction. In the end, however, the intricacies of the actual design of both the accumulation and the payout phases are what will determine the long-term effects of the reforms on the private saving rate. Over the medium-term, the design of the transition phase is crucial.

38. Of course there are many design and transition-related issues that would need to be dealt with. Any pension claims should be transferable, in order not to decrease labor market flexibility. Beyond that, some of the most important question with respect to the design of the system would include: What will the eligibility criteria and criteria for withdrawal be? What is the contribution rate? What incentives, if any, through the tax system or otherwise, will the government provide for participation? What will the benefit formula look like? How will it take account of the length of contributions and expected future changes in longevity? Who will manage the funds, how will the mandate to manage the funds be allotted (e.g., competition in the market vs. competition for the market), and what will the fee structure be? Transitional issues that need to be decided include how to treat workers that will not be in a position to build up a full pension under the new system (and have built up right under the existing system), as well as the financing of such transitional arrangements. In the case of Turkey, some of the subsidies currently paid under the third pillar voluntary savings scheme—which disproportionately accrues to higher income earners who participate most in the scheme—could be redirected to finance part of the transition to a funded second pillar. Many of the key design considerations were discussed with authorities and are contained in Impavido et al. 2010.

39. More broadly, one would need to integrate the second pillar with the first and third pillars. The introduction of a funded second pillar would serve to alleviate pressure on the unfunded first pillar pension system. Hence it would create further room for parametric reform of the first pillar, gradually transferring this pillar toward a basic pension income or solidarity payment. Over the longer term, this would serve to shield public finances from some of the pension expenditure-related consequences of population aging. At the same time, the need for a voluntary third pillar diminishes. While the system could be maintained, government contributions incentivizing third pillar savings could hence be scrapped or reduced, and (some of) the current outlays could be employed towards incentivizing second pillar savings.

Pension Reform In Selected Countries

Mexico and Chile present two examples of countries that have successfully transitioned from pay as you go to funded pension systems. Both countries feature individual accounts and private management of pension assets.


In the 1990s, Mexico faced demographic trends similar to those Turkey is facing today. It saw a decrease in the fertility rate and an increase in life expectancy, leading to slower population growth. At the same time, it led to a decrease in the youth dependency ratio, and initially the old age dependency ratio. However, in the 1990, Mexico’s old age dependency ratio was already increasing rapidly.

Mexico introduced a three pillar pension system in 1997 for private sector employees. The first pillar consists of a minimum guaranteed pension equivalent to the indexed minimum wage for low-income workers. This pillar also included a severance pay component. The second pillar is a fully-funded mandatory system of (defined contribution) individual savings accounts with competitive mutual fund management, while the third pillar consists of a voluntary savings account. Public sector employees only started to transition into the new system in 2007.


Mexico: Pension Assets under Management

(Millionsof Mexican Peso)

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Source: CONSAR, Mexican pension regulator.

Contributions for the system were set at 8½ percent of earnings, covering both pensions and severance pay. These contributions consisted of 6½ percent of earnings for the joint first pillar pension and severance system, split between the employer (3.15 ppt), the employee (1.13 ppt) and the government (2.22 ppt), and a separate 2 percent of earning contribution paid for by the employer to fund the employee’s individual savings account under the second pillar.1

The impact of these reforms has been a pronounced increase in the private saving rate. The average saving rate post-reform rose to almost 19 percent of GDP (1998–2015 average) from a little under 15 percent of GDP on average (1976–1996) prior to the reform. Meanwhile, assets under management rose steadily, at a pace of over 16 percent annually since 2005, and have now reached almost 14 percent of GDP.

The overall impact on the national savings rate has been more modest. Taking account of the decrease in public savings to partly finance the transition deficit, as well as the interaction between public and private saving, Villagomez and Anton (2013) estimate that the pension reform raised the national saving rate by about 1 percentage point.


By 1979 the Chilean pension system featured 32 pension funds, all operating under a pay-as-you-go system. The number of active participants per pensioner had steadily decreased to about 2.5, and evasion of social security contributions was widespread. As a result, the system faced a financing crisis, and contributions and investment returns no longer covered even current pension payments.

The 1981 reform replaced the pay-as-you-go system with a fully-funded pension system based on individual accounts. These accounts were privately managed, with employees given a choice of management company. Employees were given a choice to join. However, an incentive for joining was provided, as contribution rates were set at a low enough level to significantly increase take-home pay for participants. The transition was financed by the issuance of “recognition” government bonds, which were placed in employees’ individual accounts. Amounts coincided with the rights employees had built up under the old system.

Chilean workers contribute 13 percent of pensionable salary to their pension fund, disability and term life insurance. Of this total, 10 percentage points concerns pension savings, while the remaining 3 percentage points pays for disability and term life insurance and administrative fees and commissions (Soto, 2007).

The Chilean private saving rate increased significantly over the years since the pension reform. While the private saving rate amounted to a paltry 6.7 percent of GDP in the five years leading up to the reform, it steadily increased to an average of 18.3 percent of GDP between 1987 and 2015. Of course, starting from the mid 1970s, the entire Chilean economy was transformed through a program of deep market-oriented structural reforms. Hence the increase in the private saving rate cannot easily be ascribed to the pension reform. In fact, the actual contribution of the pension reform to the increase in the saving rate remains an issue of lively debate even 30 years after the system’s inception. The most thorough analysis of the reforms in Corbo and Schmidt-Hebbel (2006) focuses on the effects on total national saving. To do so it separately analyzes all different channels through which the reform influences aspects of the saving rate: the transition deficit, Ricardian equivalence, new mandatory household saving, and the response of households to this new mandatory saving. It concludes that, based on plausible assumptions, the overall increase in national saving that can be ascribed to the pension reform are between 0.7 and 4.6 percentage points of GDP.


Private Savings Rate

(percent of GDP)

Citation: IMF Staff Country Reports 2016, 105; 10.5089/9781484338209.002.A001

Source: IMF World Economic Outlook.
1 In addition, the system features a special funded housing account into which employers pay 5 percent of payroll.

Severance pay

40. The Turkish severance pay system is generous but inflexible (see, e.g., Gursel and Imamoglu, 2012). In the system, employees in the formal sector become eligible for severance pay after a year with the same employer. They build up the right to a month of severance pay per year worked, based on the last-earned gross wage.7 Severance pay is payable upon firing or death of the employee, or when the employee retires.8 Crucially, an employee who quits voluntarily is not eligible for severance pay, and built up rights to severance pay are not transferable. In other words, quitting for another job implies giving up any existing rights to severance pay, and restarting the buildup of rights from scratch. The state (social security institute) is not involved in the severance pay scheme.

41. And benefits are often not paid. Full compliance with the severance pay according to the labor law is high in the public sector, but limited in the private sector. Many private sector employers avoid paying severance pay. Informal workers are not eligible. In addition, many workers in the formal sector do not receive severance pay, either because they are kept on repeated temporary contracts or a made to formally resign while in fact they were fired. Of course workers can also not be paid because they do not meet the eligibility criteria. Furthermore, for many employees the severance pay is less than a month per year of employment, as the wage base is underreported. Overall this implies that employers’ cost of the severance scheme is much below the 8.3 percent headline rate.

42. Severance pay is an unfunded obligation, and therefore does not contribute to savings. There is no obligation to fund the rights the employees build up, and most employers therefore choose not to fund these obligations. Such unfunded severance obligations do not contribute to domestic savings. While the actual obligation is often less than a month per year (due to underreporting of wage), it is still very substantial, increasing the incentives for employers to come up with ways to avoid paying.

43. Funding the severance pay scheme would contribute to savings and could at the same time fix some of the shortcomings highlighted above. While funding the scheme would directly result in savings, crucial design questions include whether the savings would accrue to funds on corporate balance sheets or to individual employees’ severance accounts, and what the contribution rate should be. 9 Pre-funded individual severance accounts would ensure that eligible employees receive the benefits. Individual accounts would also remove penalties on flexibility in the labor market, as benefits would no longer be tied to staying at a specific company, and increasing coverage to include employees on temporary contracts and self-employed would be straightforward.10 At the same time, it would reduce employers’ incentives to force resignations or use temporary contracts. In order not to increase the labor wedge, contribution rates should be roughly in line with the actual cost of the current scheme for fully-compliant businesses.

44. The authorities are considering reforms in this direction, but there is no timeline or clear action plan. First of all, the authorities plan to extend severance pay to all employees. They plan to do so by establishing pre-funded individual severance accounts (Ministry of Labor and Social Security, 2014), under the responsibility of the Ministry of Labor. The strategy to establish the scheme and the timeframe within which this should be achieved remain under negotiation between the authorities and the social partners.

45. The introduction of individual severance accounts could boost the private saving rate by as much as 0.8 percent of GDP. The actual savings that would result depend mainly on the contribution rate and the coverage, as well as the extent to which severance pay savings would substitute other savings. Assuming a contribution rate of 4 percent—or about half of the current unfunded obligation—unchanged coverage, and a substitution effect of 50 percent, would imply additional annual savings of more than 0.8 percent of GDP. In this scenario, the contributed amounts roughly correspond to the amount reported to TURKSTAT as paid out as severance and termination pay. Hence during the transition phase, when current workers’ right would also need to be paid out in addition to the funding of the new system, total severance expenses would roughly double. If the premium were to increase gradually over time to 8 percent, and under the same assumptions, private saving could be boosted by as much as 1.7 percent of GDP. All of this is under the assumption that the labor market behavior of employees will remain unchanged.11 Importantly, and in contrast to the current unfunded obligation, if the total wage bill in the economy increases, the amount of savings under this scheme also increases.

46. The transition phase of such a reform needs to be carefully thought through. For employers that were used to avoid paying, or in case of increased coverage to include employees that were not previously covered, the new system will increases costs through a higher labor wedge. In addition, vested interests from employees in the public sector—who by and large receive the current generous benefits—may oppose reform. A gradual transition, e.g., by first introducing the new system only for new employees and/or gradually moving existing employees over to the new system, could help overcome these problems. However, given that one needs to grandfather existing rights in some form and at the same time strives to build up a funded system, a period of overall higher outlays is unavoidable. Another important issue is the integration of individual severance accounts with the general system of unemployment insurance. In the short run, unemployment insurance could be adjusted to the severance pay reform through parametric changes. In the medium term, an approach integrating both systems (as done, e.g., in Chile, Austria and Mauritius, see Box 2) would likely be most beneficial. Such integrated systems combine, first, severance payment from an individual account with, second, a solidarity component in the form of a stripped down unemployment insurance for workers who have exhausted their severance account balance.

47. Such reforms would also have positive impact beyond increasing savings. Most prominently, making the benefits portable would directly increase labor market flexibility by no longer penalizing switching jobs. A more flexible labor market improves productivity, and may in turn encourage job creation. Individual accounts would also increase the likelihood of the individual receiving a payout and would hence provide additional security to employees. In addition, like with second pillar pension savings, a steadily growing pool of severance savings could help boost capital market development, and thus enhance capital allocation in the economy, boosting long-term growth.

Severance Pay Reform In Selected Countries

The most successful experiences with severance pay reform have been in countries that fully integrated the unemployment and severance pay systems. Austria and Chile come to mind as good examples.


In 2003, Austria replaced its existing severance pay system with a system based on individual savings accounts. Until then, employers were obliged to pay employees that had been employed for more than three year severance pay upon dismissal, starting with 2 months wages after three years and increasing to a year’s wages after 25 year of employment. Employers had to make provisions in their accounts for part of this obligation, but did not need to fund the obligation.

The scheme was criticized for inhibiting labor mobility and creating an insider-outsider problem, with only about a third of employees eligible for severance pay. The reform dealt with these problems in two ways. First, the individual accounts are by their nature portable. In addition, contributions in the new system started from the first day of employment, taking away the three year eligibility “cliff”. The introduction of the system was done in one go at a rate of 1.53 percent of gross wages, but existing severance pay entitlements under the old system were grandfathered.1 Over the first ten years the reform yielded cumulative savings of 2 percent of GDP. In other words, in the build up phase, where the number of covered employees was slowly increasing, annual saving increased by some 0.2 percent of GDP on average.


Chile combined its severance pay and unemployment subsidy systems into one in late 2002. The system was reformed in 2009. The prior separate systems suffered from low coverage, with severance pay only applicable to open ended contracts and unemployment subsidy consisting of a low amount targeted at low income households (just 2 percent of the unemployed population benefitted from the unemployment subsidy in 2001). It came under pressure when unemployment spiked after the Asian crisis and again during the global financial crisis.

The new system is based on individual accounts, but also features a solidarity component, and covers all private employees in the formal sector, i.e., including employees on fixed term contracts. The individual accounts are funded by contributions from both employees and employers. The solidarity fund, which contributes to payouts for those fired because of economic hardship to the firm, is funded by contributions from employers and an annual lump sum from the state. In addition, beneficiaries have access to health benefits, training and job search assistance while drawing benefits under this scheme.

The new system also has focused on keeping down administration costs and management fees. Over time such costs and fees are a significant factor in funds’ total returns. To keep them down, the Chilean scheme is operated by a single private party, which is selected through a competitive bid for a period of 10 years. Such “for the market” rather than” on the market” competition combines both competition to keep costs and fees in check and a single operator able to realize efficiencies of scale in operations.

1 Austria introduced the new system for all new labor contracts concluded after January 1, 2003. For older contract, transitioning to the new system was voluntary. Therefore, while covering the vast majority of employees, the new system still does not yet encompass all employees.

Subsidized savings

48. Subsidized saving outside of the pension system is currently a limited part of the policy mix in Turkey. The new government recently introduced a system of dowry accounts. Young unmarried people that deposit contributions into the account for at least three years will be eligible for a 20 percent state subsidy not exceeding TL5,000. In many other countries though such subsidized saving scheme have been, or still are, more omnipresent. In a seminal study, the OECD (2007) compares tax-preferred savings accounts across its membership and evaluates the effectiveness of the schemes in raising private and national savings. A main distinction it employs is between savings accounts purely intended for the purpose of funding future education needs and general savings accounts that can be used for a broader set of outlays or without any restrictions on usage.

49. The OECD concludes that tax-preferred accounts other than educational plans create new savings only when moderate-income households participate in them. While participation among high-income households is generally high, and they deposit more funds on average than moderate or low-income households, these household are in general not liquidity constrained, and their tax-preferred savings almost entirely substitute for other savings. In addition, as percentage of income their total contributions are smaller than those of households in lower income brackets. Low-income households often face binding liquidity constraints, and hence their ability to divert funds to savings accounts in response to tax incentives is limited. Moderate-income households also face a liquidity constraint, but it is less binding than for low-income families. Hence this the income segment where tax-preferred savings accounts are found to boost the saving rate most. In addition, as in most countries moderate-income households face a lower marginal tax rate than high-income households, the loss of tax revenue is more modest, and hence the boost to national savings is larger. However, even in this study, estimates of the impact of tax-preferred savings on private savings are scant and in general do not suggest large significant effects on the macro level (Box 3).

50. For tax-preferred educational plans the conclusions differ slightly. These plans attract wealthier households in even greater numbers; as such households have more resources and may be more sensitive to educational matters. An interesting phenomenon is that educational accounts rarely attract the maximum tax-exempt contribution. This seems to signal that households may not want to lock substantial wealth into accounts whose spending goals are narrowly targeted. As such, these kinds of plans—while they may be useful for other purposes such as boosting the take up of (higher) education—are not thought to boost private saving much.

51. Another question the OECD study touches upon is the efficiency of tax-preferred savings schemes, i.e., whether the schemes increase national savings at the lowest possible costs. In general, incentives to save through tax credits are found to be quite expensive, precisely because many richer households will participate in lieu of their other savings. Encouragement to save through tax exemptions of the part of earning that is saved is found to be less expensive. However, if such schemes are coupled with a savings bonus from the government, plans can become expensive when the government matching rate is generous.

52. Overall, subsidized savings scheme do not seem to provide an efficient way for Turkey to boost the private saving rate. The effects of tax-preferred savings schemes on the private saving rate would likely be limited, and such scheme would possibly be regressive. More broadly applied subsidies for savings (i.e., savings bonuses) could quickly become expensive to the budget.

Tax-Preferred Savings Accounts In Selected Countries

Several OECD countries have set up a myriad of tax-preferred savings schemes, both for educational outlays and for more general purposes. Some examples include:

United Kingdom

The United Kingdom tax code features tax incentives for individual savings accounts (from 1999), personal equity plans, tax-exempt special savings accounts (until 1999), and child trust fund accounts. In addition, its savings gateway scheme features tax incentives for savings targeted specifically at low and moderate-income households. The child trust fund and savings gateway scheme feature savings bonuses from the government, while the other plans are limited to tax exemptions.

The pace of increase under most schemes was high in the first few years, but slowed rapidly thereafter. This may indicate that when new plans became available, people shifted assets from taxable savings to tax-preferred accounts. However, some studies suggest that up to 15 percent of ISA holders would not have saved had ISAs not existed. Another indication that these accounts may have boosted the private saving rate lies in the fact that the take up among moderate-income households was substantial.


Between 1982 and 2000, Norway’s tax code features a tax-favored savings scheme targeted at equity investments. It allowed for a 15 percent tax credit for share purchases, limited to NOK 10,000 for tax payers in tax class 2 (mainly couples and single parents) and NOK 5,000 for most other households. In 1999, of the tax payers in tax class 2, about 60 percent received the maximum deduction, while of all other tax payers, about 80 percent received the maximum deduction. This less-than-full participation suggests that for a significant part of the participants, the marginal return to savings has changed which could have led to an overall increase in private savings. In addition, this effect is stronger for class 2 taxpayers, which may on average be in lower income brackets than other tax payers, thus further suggesting a positive effect on private savings.

D. Summary and Conclusion

53. Boosting the private savings rate is important to reduce the external imbalance of the Turkish economy. The large and persistent current account deficit has been financed by ample capital inflows. This situation presents a vulnerability of the Turkish economy and may ultimately prove unsustainable. The vulnerability can be reduced by decreasing investment—which would lower the potential growth rate of the economy—or increasing domestic saving—which would lower consumption but increase the economy’s growth potential. As government finances are already in decent shape, most of the savings boost should come from the private sector.

54. The authorities have recognized this and put raising the private saving rate forward as an important policy goal. To that effect they have introduced a subsidized third pillar pension scheme, and, more recently, a savings subsidy for dowry accounts. They have also piloted an auto-enrollment funded pension scheme, and are considering scaling up this pilot. Lastly, proposals to reform the severance pay scheme by making it a funded and transferable benefit have been put forward. The authorities have also used macroprudential tools to limit credit growth, motivated by the (systemic) prudential risks in the banking system very high credit growth may cause.

55. Going forward, full and swift implementation of the pension and severance pay reform plans is key. Given the urgency of reducing vulnerabilities and the time lag with which new policies will affect the savings rate, the time to start is now. In addition, Turkey’s relatively young population implies the country is enjoying a demographic dividend, which provides a window of time to increase savings in anticipation of almost inevitable population ageing in the future. Macroprudential policies limiting credit growth should also remain part of the policy mix.

56. Careful attention needs to be paid to the details of the systems’ designs, including transition issues. For the pension system in particular, the design of the fee structure to a large extent determines the ultimate welfare benefits of the system. Attention should not only be paid to the accumulation and investment phase of the system, but also to the payout phase and the balance between the associated risks for the participating individuals and the system. In addition, it is crucial to fund the transition to the new pension and severance pay systems fairly and efficiently, possibly by diverting some of the subsidies currently being allocated to third pillar pensions and other subsidized savings schemes. Studies performed by the authorities together with the World Bank provide much detail and options on these crucially important issues.

57. The reforms will likely have other benefits besides increasing domestic savings and reducing external vulnerabilities. These include more labor market flexibility through severance pay portability; lowering of old age poverty rates through higher pension savings; and boosting domestic capital market development. They may hence provide an important contribution towards the authorities stated goal of boosting the long term growth potential of the economy.


  • Carroll, Christopher D. and Summers, Lawrence H. “Consumption Growth Parallels Income Growth: Some New Evidence,” in B. Douglas Bernheim and John B. Shoven, eds., National Saving and Economic Performance. Chicago: Chicago University Press, 1991, pp. 30543.

    • Search Google Scholar
    • Export Citation
  • Corbo, Vittorio, and Klaus Schmidt-Hebbel, 2006, “Macroeconomic Effects of Pension Reform in Chile,” published at

    • Search Google Scholar
    • Export Citation
  • Deaton, Angus S. and Paxson, Christina H. “Saving, Growth, and Aging in Taiwan,” in David A. Wise, ed., Studies in the Economics of aging. Chicago: University of Chicago Press, 1994, pp. 33164.

    • Search Google Scholar
    • Export Citation
  • Deaton, Angus S. and Paxson, Christina H. “The Effects of Economic and Population Growth on National Saving and Inequality.” Demography, February 1997, 34(1), pp. 97114.

    • Search Google Scholar
    • Export Citation
  • Deaton, Angus S. and Paxson, Christina H. “Saving and Growth: Another Look at the Cohort Evidence.” Review of Economics and Statistics, May 2000, 82(1).

    • Search Google Scholar
    • Export Citation
  • Eren, Okan, and Serife Genç-İleri, 2015, “Government Subsidized Individual Retirement System,” Working Paper 15/20, Central Bank of the Republic of Turkey, Ankara.

    • Search Google Scholar
    • Export Citation
  • Fernandez de Lis, Santiago, Jorge Martinez Pages, and Jesus Saurina, 2000, “Credit Growth, Problem Loans and Credit Risk Provisioning in Spain,” Working Paper No. 0018, Banco de Espana: Madrid.

    • Search Google Scholar
    • Export Citation
  • Gursel, Seyfettin and Zumrut Imamoglu, 2012, “Severance Pay Reform Problems and Solutions,” Betam Working Paper, Bahcesehir University, Istanbul

    • Search Google Scholar
    • Export Citation
  • Hurriyet Daily News, 2015, “Number of Gold Savers Doubles in Turkey,”

    • Search Google Scholar
    • Export Citation
  • Impavido, G., E. Lasagabaster, and M. G. Huitron, 2010, New Policies for Mandatory Defined Contribution Pensions—Industrial Organization Models and Investment Products. The IBRD Press.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2007, “Safe to save less? Assessing the recent decline in Turkey’s Private Saving Rate,” in IMF Country Report No. 07/364, International Monetary Fund, Washington, DC, USA.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2014, “Turkey Staff Report for the 2014 Article IV Consultation,” Country Report 14/329, International Monetary Fund, Washington, DC, USA.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2016, “Turkey Staff Report for the 2016 Article IV Consultation,” Country Report (forthcoming), International Monetary Fund, Washington, DC, USA.

    • Search Google Scholar
    • Export Citation
  • Kelly, Robert, Kieran McQuinn, and Rebecca Stuart, 2013, “Exploring the Steady-State Relationship between Credit and GDP for a Small Open Economy. The Case of Ireland,” Working Paper No. 1531, European Central Bank: Frankfurt.

    • Search Google Scholar
    • Export Citation
  • Ministry of Labor and Social Security, 2014, “National Employment Strategy 2014–23,” Ministry of Labor and Social Security, Ankara, Turkey.

    • Search Google Scholar
    • Export Citation
  • Modigliani, Franco. “The Life Cycle Hypothesis of Saving and Inter-Country Differences in the Saving Ratio,” in W. A. Eltis et al., eds., Induction, Growth, and Trade: Essays in Honor of Sir Roy Harrod. London: Clarendon Press, 1970.

    • Search Google Scholar
    • Export Citation
  • Modigliani, Franco., “Life Cycle, Individual Thrift, and the Wealth of Nations.” American Economic Review, June 1986, 3(76), pp. 297313.

    • Search Google Scholar
    • Export Citation
  • OECD, 2006, “Pensions at a Glance,” OECD: Paris.

  • OECD, 2007, “Encouraging Savings through Tax-Preferred Accounts,” OECD Tax Policy Studies No. 15, OECD: Paris.

  • Ozcan, Kivilcim Metin; Asli Gunay; and Seda Ertac, 2003, “Determinants of Private Savings Behavious in Turkey,” Applied Economics, 2003, Vol. 35, pp. 14051416.

    • Search Google Scholar
    • Export Citation
  • Pirgan Matur, Eser, Ali Sabuncu, and Sema Bahceci, 2012, “Determinants of Private Savings and Interaction Between Public and Private Savings in Turkey,” Topics in Middle Eastern and African Economies, vol. 14, September 2012.

    • Search Google Scholar
    • Export Citation
  • Poterba, James M., ed. International Comparisons of Household Saving. Chicago: University of Chicago Press, 1994.

  • Tobin, James. “Life Cycle Saving and Balanced Growth,” in William Fellner et at., eds., Ten Economic Studies in the Tradition of Irving Fisher. New York: Wiley, 1967

    • Search Google Scholar
    • Export Citation
  • Van Rijckeghem, Caroline, 2010, “Determinants of Private Saving in Turkey: An Update,” Working Paper 2010/04, Bogazici University: Instanbul.

    • Search Google Scholar
    • Export Citation
  • Van Rijckeghem, Caroline; and Murat Ucer, 2009, “The Evolution and Determinants of the Turkish Private saving rate: What Lessons for Policy?,” ERF Research Report Series 0901, Economic Research Forum and TUSIAD, Istanbul, Turkey.

    • Search Google Scholar
    • Export Citation
  • Villagomez, Alejandro, and Arturo Anton, “Impacto Macroeconomico de la Reforma de Pensiones en Mexcio,” Working Paper 560, CIDE, Mexico, DF.

    • Search Google Scholar
    • Export Citation
  • World Bank, 2011, “Sustaining High Growth: The Role of Domestic Savings”, Country Economic Memorandum 66301–TR. World Bank: Washington, D.C.

    • Search Google Scholar
    • Export Citation
  • World Bank, 2014, “Turkey in Transition: Time for a Fiscal Policy Pivot?”, Public Finance Review. World Bank: Washington, D.C.


Prepared by Alexander Tieman.


These were years that included significant economic crises. In response, the savings rate fluctuated between 12.4 and 28.5 percent in this period.


China qualifies as a take-off country but as its extremely high saving rate would distort the average, it is treated separately here.


The authors chose to start in 1998, as there is a break in the series due to national accounts revisions in 2007, in which only the series up to 1998 were retroactively revised.


While participants can save unlimited amounts in their pension account, the government contribution is capped at 25 percent of the minimum wage.


While there may be other benefits of having such funds, a second pillar fund that invests solely in government bonds yield negative economic benefits after fees compared to straightforward lowering of government debt through a first pillar system.


Severance pay is capped at a rate linked to the gratuity of a civil servant in the highest pay scale.


There are several other eligible events such as compulsory military service, or, for women only, marriage.


In addition, the scope of eligible events could be revisited.


Including employees on temporary contracts could be done by abolishing the one year minimum employment requirement.


If the functioning of the labor market would improve, higher employment and thus a higher saving rate, would be the likely outcome.

Turkey: Selected Issue
Author: International Monetary Fund. European Dept.