improving india’s saving performance · for annual gdp growth over the next 10 years. ... 1950/51...

38 Finance & Development / June 1997

One of the prerequisites forputting India on a high growthpath is a substantial rise indomestic saving. This willrequire tighter fiscal policiesand strong structural reforms,including liberalization offinancial markets.

HE DOUBLE task of alleviatingpoverty and keeping up with fast-growing Asian neighborsprompted the Indian government

to announce a target of 7 percent or morefor annual GDP growth over the next 10years. A key question is whether India willbe able to finance the investment necessaryto reach this target through increaseddomestic saving and avoid a much greaterrecourse to foreign saving with its associ-ated risks on the external front.

A strategy to improve India’s saving per-formance needs to take account of recentinsights in the saving literature. Over thepast few years, several studies of saving in developing countries have found that tax and interest rate incentives have

been largely ineffective. Moreover, empiri-cal studies suggest that higher growth gen-erally tends to precede higher saving. Inlight of this evidence, it may be more effective to increase domestic saving byraising public saving and implementing astrong structural reform program, includ-ing financial liberalization.

Measuring savingIndia’s saving rate is relatively high,

compared with that of other countries. Ithas shown an uneven upward trend overthe past four decades (Chart 1), and therehave been considerable changes in its com-position. Historically, domestic saving has

been dominated by household saving in physical assets. However, the recentincrease in saving has been driven mainlyby financial household saving, partlyreflecting a continuing expansion of finan-cial institutions’ branch networks into ruralareas and, more recently, the increasingavailability of alternative investmentopportunities. Private corporate saving hasalso shown a steady increase over the lasttwenty years, although it remains below 5percent of GDP. Public saving weakened inthe early 1990s to reach a low of 0.5 percentof GDP in 1993/94, a significant reductioncompared with the levels of 4–5 percent ofGDP seen in the early 1980s.

Martin Mühleisen,a national of Germany, is an Economist in theIMF’s Asia and Pacific Department.

TChart 1

Components of domestic saving in India(percent of GDP)

Source: India's National Accounts.

01950/51 55/56 60/61 65/66 70/71 75/76 80/81 85/86 90/91

5

10

15

20

25

30

Physical

Domestic

Financial

Public Corporate

Improving India’s SavingPerformance

M A R T I N M Ü H L E I S E N

Measurement problems. The inter-pretation of Indian saving trends is com-plicated by a number of weaknesses in the Central Statistical Office’s (CSO)methodology for measuring both invest-ment and saving. The most importantshortcomings are:

• The estimate for physical householdsaving is set equal to household invest-ment, which itself is calculated only indi-rectly as a residual. Not surprisingly,measured physical household saving hasbeen highly volatile.

• There are errors and omissions in theestimates of both savings and investment,but adjustments are made only to invest-ment. The CSO thinks the saving estimateis more reliable (based on the greater accu-racy of public, financial, and corporate sav-ing data) and therefore adjusts investmentto equal the sum of domestic and foreignsaving.

• The commodity flow method used toestimate total investment—based on fixedproduction coefficients—has remained un-changed for decades. While it might still beuseful for comparing investment in adja-cent years, new technologies and the grow-ing amount of investment in the informalsector are not adequately reflected in theestimates.

• The estimates of corporate saving andinvestment are based on a small, unrepre-sentative sample, and rely largely on volun-tary responses from enterprises.

• Finally, the CSO estimates do notcover some assets preferred by households,namely jewelry and gold. Household savingin gold probably increased after importrestrictions were liberalized in 1992, imply-ing an increase in the underestimation ofsaving.

What can be said about underlying sav-ing trends in the face of these problems? Wehave tried to generate alternative estimatesof saving in two ways:

• First, reversing the present CSO prac-tice, we adjusted domestic saving to includeerrors and omissions, so that the sum ofadjusted domestic saving and foreign sav-ing equals the original investment estimate.This yields a much smoother, more plausi-ble path for domestic saving (Chart 2).

• The second approach is based on analternative estimate of physical saving, theweakest component in the CSO’s methodol-ogy. Assuming that physical saving is neg-atively related to financial saving, weestimated physical saving using an econo-metric regression on financial saving and atime trend to obtain a second path fordomestic saving.

Although these two alternative savingmeasures differ in some respects, they bothsuggest that the recent fluctuations indomestic saving may have been exagger-ated. Both measures show that the savingrate was fairly constant for most of the1980s, before it picked up in the early-to-mid-1990s.

Sufficient savings? Econometric re-gression analysis suggests that private sav-ing is likely to continue to increase—albeitgradually—over the coming years, drivenby rising per capita income and continuedfinancial deepening. In addition, a lowershare of agriculture in the economy and anincrease in the age dependency ratio wouldtend to increase private saving. Taking intoaccount likely developments in public sav-ing, this would result in a saving rate ofabout 28 percent of GDP after 2000. But thisis not likely to be enough to finance the in-vestment needed to reach the government’s

growth objective. Even assuming someimprovement in investment efficiency (inthe absence of reliable employment and cap-ital stock data, there are no estimates oftotal factor productivity growth), thegrowth target implies that the investmentrate would need to increase to well above 30 percent, which—even with higherrecourse to foreign saving—would requirea domestic saving rate of around 30 percentby the turn of the century. If the growth tar-get is to be achieved, stronger action onboth the public and private saving fronts iscalled for.

Strategy for higher savingWhile higher domestic saving is needed

to finance faster growth, policies aimeddirectly at mobilizing saving are not neces-sarily the best instrument to achieve thistarget. In the case of India, it has beenargued that growth has suffered less from alow saving rate than from inefficient invest-ment, in part because of the dominant roleof the public sector in the economy.

There is also a growing literature that,based on cross-country studies, has foundlittle evidence that policy efforts to boostsaving have been very effective. Thisresearch suggests that the main policyfocus should be on initiating a virtuousgrowth-saving circle by fostering growththrough fiscal consolidation and strongstructural reforms, including privatizationand financial liberalization. Under such astrategy, initially, growth would need to befinanced mainly through higher public sav-ing. Private saving, which eventually wouldhave to provide the bulk of additionalinvestment financing, would follow with alag, responding to higher growth. Financialliberalization—in particular, reform of

39Finance & Development / June 1997

Chart 2

Alternative estimates of saving in India

Adjusted for errors and omissions

CSO estimate CSO estimate

Source: India, Central Statistical Office (CSO), and IMF staff calculations.

(percent of GDP)

5

1950/5154/55

58/5962/63

66/6770/71

74/7578/79

82/8386/87

90/9194/95

1950/5154/55

58/5962/63

66/6770/71

74/7578/79

82/8386/87

90/9194/95

10

15

20

25

30

10

15

20

25

30Using an estimate for household investment

(physical saving)

AdjustedIMF staff estimate

long-term saving instruments—would helpto ensure that private saving was efficientlyallocated.

The case for an indirect approach tohigher private saving is supported byrecent findings that traditional saving pol-icy instruments—like higher interest ratesor special tax incentives—fail to raise theprivate saving rate in the long run.Although these results were establishedmainly for industrial countries, they arelikely to apply just as forcefully in develop-ing countries. For example, Ogaki, Ostry,and Reinhart (1996) have found that theresponsiveness of private saving tochanges in real interest rates is less atlower levels of per capita income, as ahigher share of income must bedevoted to subsistence consump-tion. They estimated that theresponse of saving to changes ininterest rates in India was amongthe lowest in the developingworld. Moreover, Chelliah (1996)and others have pointed out thatmost Indian households do notpay income tax, either because their incomeis too low or because they fail to report tothe tax authorities. Changes in the taxregime would therefore affect only a smallpart of the population, and would beunlikely to significantly alter overall savingbehavior.

Public saving. Studies suggest thatthe most direct way to raise domestic sav-ing is by generating higher public saving.However, India has seen a steady decline inpublic saving over the past two decades,both at the central and state governmentlevels. This trend has been partly reversedsince 1993/94, but further strong effortswould be needed to restore public saving tothe level of the early 1980s. Such effortswould need to involve a series of actions inthe areas of tax policy, expenditure man-agement, center-state relations, and publicenterprise reform.

To some extent, higher public savingmay be offset by lower private saving.However, judging from an estimated long-run relationship between private and pub-lic saving, the offset factor for India couldbe as low as 25 to 30 percent. Nevertheless,in the short run, the trade-off could besomewhat larger, as fiscal consolidationwould have to be achieved partly throughhigher taxation.

Incentives for private saving.While instruments to directly raise privatesaving have proven largely ineffective,structural reform measures could have alarge impact on growth and, indirectly, on

private saving, mainly by improving theefficiency of resource allocation and raisingtotal factor productivity growth. Broadly,the reform agenda for India would includepublic enterprise restructuring and privati-zation; increased private involvement ininfrastructure; agricultural reform; labormarket reforms and exit policies; lifting ofprivileges for smaller enterprises; and,especially, financial reform.

The impact of financial sector reformcould be large. In India, the link betweenlow growth and the inefficient allocation ofsaving has become increasingly relevant,particularly in infrastructure. Althoughoverall investment has increased in recentyears, investment in infrastructure has

declined, and worsening infrastructure con-ditions have become a major obstacle togrowth. The Mohan Committee on infra-structure development recently concludedthat the lack of long-term financing was asubstantial hindrance to such investment,and listed the development of domesticdebt markets and the effective use of long-term saving among the highest reform priorities.

Consequently, efforts to raise private sav-ing should focus on financial liberalization,particularly on the development of long-term saving instruments, such as pensions,life insurance, and mutual funds. Whileproviding an attractive investment vehiclefor individual savers, their main role wouldbe to improve the allocation of savings,ensuring that funds would flow to the mostproductive investment projects, thus gener-ating the highest rate of growth for a givenamount of investment. As a result, the vir-tuous growth-saving circle would becomemore dynamic, and savings could accumu-late faster.

Long-term instrumentsIn India, unlike other countries, the share

of major instruments for long-term house-hold saving—pension and life insurance—in gross financial saving has stagnatedover the past 30 years. By contrast, mutualfunds had growing success through theearly 1990s, particularly after the sectorwas opened to competition from the privatesector. However, they have also experienced

considerable problems in recent years.These developments reflect two fundamen-tal weaknesses:

• The markets are dominated by thepublic sector. The three largest institutionalinvestors in India—the Life InsuranceCorporation of India (LIC), the Unit Trust ofIndia (UTI), and the Employees’ ProvidentFund (EPF)—account for about a third oftotal financial saving. These public sectorinstitutions face little competition. In thepension and life insurance sectors, the EPFand the LIC hold a near monopoly, while theUTI still accounts for more than 80 percentof the mutual fund business.

• Portfolio allocation is heavily regu-lated, resulting in comparatively low re-

turns and little flexibility to reactto market developments.

Owing to these weaknesses,long-term saving markets havefailed to attract investors at atime when more lucrative alter-natives have emerged in othermarkets, particularly as interestrates on bank deposits and cor-

porate paper have increased. A sustainedincrease in long-term saving would requiregiving market participants greater flexibil-ity in portfolio allocation, while greater pri-vate sector involvement would help toboost competition and more innovativeproduct development. The government hastaken preliminary steps in this direction,but a stronger reform impetus is stillrequired in some areas.

Provident funds. The Indian provi-dent fund system—consisting of the EPFand a number of smaller provident funds—provides fully funded defined-contribu-tion retirement schemes for about 8 percentof the labor force. Those not covered underthese schemes—over 90 percent of the pop-ulation—rely mainly on extended familynetworks and informal saving arrange-ments for old-age security. The entire port-folios of provident funds are invested ingovernment or quasi-government securitiesand special deposit schemes.

The funds have been a sizable factor inthe financing of the public sector deficit;however, their investment yields have beenrelatively low. The average real rate ofreturn was below 1 percent in the 1980sand only slightly higher—1.5 percent—in1994/95. These returns are too low to gener-ate a sizable accumulation of pension assetsduring a lifetime, and they also encouragewithdrawal of funds when allowed undercertain defined circumstances.

The government has begun to respond tothese problems. Mutual funds have been

Finance & Development / June 199740

“Efforts to raise private savingshould focus on financial

liberalization, particularly on the development of long-term

saving instruments.”

allowed to offer pension plans—althoughthey are still subject to provident fund port-folio allocation rules—and the recent in-troduction of an LIC pension plan wasdirected mainly at workers in the informalsector who had no access to the systembefore. Moreover, the government hasrecently changed the investment schedulefor private pension funds, increasing theceiling for investment in debt instrumentsissued by the public financial institutionsto 40 percent, and lowering the ceiling onspecial deposit schemes that earn a lowerrate of return. As a result, the return oninvestment could increase by up to 2–3 per-centage points.

The Indian provident fund system couldbe reformed to follow the examples of coun-tries like Chile that have successfully raisedtheir saving rates through pension andfinancial market reforms. Key aspects ofreforms would involve providing an in-creased role for private pension fund man-agement and substantially liberalizingportfolio allocation rules, with an appropri-ate regulatory structure to ensure prudentinvestment allocation.

Life insurance. The life insurancesector was nationalized and consolidatedinto the LIC in 1956, jointly with the gen-eral insurance sector. Since then, the LIChas been a monopoly operator, chargedwith the tasks of making life insuranceavailable throughout the country, particu-larly in rural areas, and mobilizing savingsby providing attractive insurance products.On the first count, LIC has been fairly suc-cessful. Having built up a large regionaldistribution network comprising some2,000 branches, rural areas now account forover 40 percent of new policies. However,the Indian insurance market, with an esti-mated $5 in annual premiums paid percapita, has not made a significant contribu-tion to savings mobilization.

Like other long-term saving instruments,life insurance has experienced a relativedecline recently, mainly owing to the com-paratively low interest rate paid on lifeinsurance funds. The LIC is subject to simi-lar, although somewhat less restrictiveportfolio allocation constraints as pensionfunds. Some 75 percent of annual portfolioinvestments must be allocated to govern-ment securities or socially oriented pur-poses, while the remaining 25 percent canbe invested in private sector debt. The aver-age yield has remained low, reaching only1–2 percent in the early 1990s. In addition,high administrative costs, related to highstaffing levels and insufficient computeriza-tion, have dampened profitability.

Based on far-reaching recommendationsby the Malhotra Committee, the govern-ment has been considering plans to openthe insurance sector to private competitors,including those from abroad, within thenext four years. So far, an InsuranceRegulation Authority has been set up toestablish rules for the broader marketstructure. In order to prevent private com-petitors from focusing exclusively on prof-itable, specialized (urban) markets, theMalhotra Committee recommended thatnew entrants be obliged to cover to someextent rural sectors and to contribute to thefinancing of socially oriented projects. Italso recommended strengthening the LIC’scompetitiveness by lowering the currentmandatory investment norm to a level thatallowed portfolio allocation more in linewith international levels.

Mutual funds. When the mutual fundsindustry was liberalized in 1992, the UTIhad held a monopoly in the market foralmost 30 years. Indian retail investors(some 24 million shareholders) had beenaccustomed to guaranteed high returns ontheir UTI investments. This good record,combined with aggressive marketing bynew entrants, led to expectations of highprofits by investors who began to investstrongly in the new private mutual funds.However, they were generally unpreparedfor the risks they were taking after liberal-ization.

The net asset value of mutual fundsdeclined when stock prices began to fall in1992. The situation was exacerbated be-cause existing market regulations did notallow portfolio shifts into alternativeinvestments, leaving funds with no choicebut to hold cash or continue investing inshares. Moreover, since only closed-endfunds had been introduced in the market,investors who wanted to disinvest had tosell their holdings at a loss in the secondarymarket. These losses, the after-shocks ofthe 1992 stock market scandal, and the lackof transparent rules rocked confidence inshares and mutual funds. Partly owing to arelatively weak stock market performance,mutual funds have not yet recovered, withfunds trading at an average discount of10–20 percent of their net asset value.

The stock market supervisory authorityhas recently adopted a set of measures cre-ating a transparent and competitive envi-ronment for mutual funds. These includerelaxing investment restrictions into moneymarket and debt instruments, listing open-ended funds, and permitting mutual fundsto launch pension schemes. In response tothese changes, the UTI is to be reorganized

internally into a number of separate, com-peting units, and foreign banks have againbegun to launch new funds. The intentionis that mutual funds could become the keyinstrument for long-term saving, offering avariety of investments ranging from pureequity funds to pension plans.

These measures should help to increasepublic confidence in the stock market.Nevertheless, the key to a revival of in-vestor interest would be a solid recovery ofIndian stock markets—something thatdepends to a large extent on governmentpolicies. As long as public financing needscontinue to keep real interest rates high,both lower enterprise profitability and thehigher attractiveness of competing invest-ment alternatives will have a negativeimpact on Indian stocks.

Looking to the futureHow should India raise its domestic sav-

ing rate? Traditional tax and interest rateincentives are unlikely to lead to a strongresponse of the private saving rate. Instead,the most promising way to boost domesticsaving is through increased public savingand a strong structural reform program,including financial liberalization, whichwould initiate a virtuous circle in whichhigher growth would prompt further in-creases in private saving. With a view toincreasing the efficiency of savings alloca-tion and financing the heavy infrastructureneeds of the Indian economy, particularattention should be paid to long-term sav-ing instruments.

A sustained increase in long-term savingwould require two major policy changes.First, the government would need tosharply reduce its recourse to captivefinancing from pension funds and the LIC,thus giving market participants greaterflexibility in their portfolio allocation. Atthe same time, greater private sectorinvolvement would be required to helpboost competition and more innovativeproduct development, which would makesaving instruments more remunerative andthus attractive to individual investors.

41Finance & Development / June 1997

References:R. Chelliah, 1996, “How to Hike Savings,”

Economic Times, July 12.Martin Mühleisen, 1997, “Improving India’s

Saving Performance,” IMF Working Paper No. 97/4 (IMF: Washington).

M. Ogaki, J.D. Ostry, and C.M. Reinhart,1996, “Saving Behavior in Low- and MiddleIncome Developing Countries,” IMF StaffPapers, Vol.. 43 (March), pp. 38–71.

F&D