Equality and Efficiency, Berg: International Monetary Fund

8/3/2019 Equality and Efficiency, Berg: International Monetary Fund

1/4

12 Fnance & DevelopmentSeptember 201112 Fnance & DevelopmentSeptember 2011

equality and efficincy

IN his inuential 1975 book Equality andEciency: Te Big radeof, Arthur Okunargued that pursuing equality can reduceeciency (the total output produced with

given resources). Te late Yale University andBrookings Institution economist said that notonly can more equal distribution o incomes

reduce incentives to work and invest, but theeforts to redistributethrough such mecha-

nisms as the tax code and minimum wagescan themselves be costly. Okun likened thesemechanisms to a leaky bucket. Some o theresources transerred rom rich to poor willsimply disappear in transit, so the poor willnot receive all the money that is taken rom therichthe result o administrative costs and

disincentives to work or both those who paytaxes and those who receive transers.

I

12 Fnance & DevelopmentSeptember 2011

is there a trade-off between the two or do they go hand n hand?

Andrew G. Berg and Jonathan D. Ostry


2/4

Fnance & Development September 2011 13

INEQUALITY

Do societies inevitably face an invidious choice between effi-cient production and equitable wealth and income distribution?Are social justice and social product at war with one another?

In a word, no.In recent work (Berg, Ostry, and Zettelmeyer, 2011; and

Berg and Ostry, 2011), we discovered that when growth

is looked at over the long term, the trade-off between effi-ciency and equality may not exist. In fact equality appearsto be an important ingredient in promoting and sustaininggrowth. The difference between countries that can sustainrapid growth for many years or even decades and the manyothers that see growth spurts fade quickly may be the levelof inequality. Countries may find that improving equalitymay also improve efficiency, understood as more sustainablelong-run growth.

Inequality matters for growth and other macroeconomicoutcomes, in all corners of the globe. One need look no fur-ther than the role inequality is thought to have played increating the disaffection that underlies much of the recent

unrest in the Middle East. And, taking a historical perspec-tive, the increase in U.S. income inequality in recent decadesis strikingly similar to the increase that occurred in the1920s. In both cases there was a boom in the financial sec-tor, poor people borrowed a lot, and a huge financial crisisensued (see Leveraging Inequality, F&D, December 2010and Inequality = Indebtedness in this issue of F&D). Therecent global economic crisis, with its roots in U.S. financialmarkets, may have resulted, in part at least, from the increasein inequality. With inequality growing in the United Statesand other important economies, the relationship betweeninequality and growth takes on more significance.

How do economes grow?

Most thinking about long-run growth assumes implicitlythat development is something akin to climbing a hill, thatit entails more or less steady increases in real income, punc-tuated by business cycle fluctuations. The pattern in Chart1which shows the level of real (after-inflation) per capitaincome in two advanced economies, the United Kingdom

and the United Statesis consistent with this idea.The experiences in developing and emerging economies,

however, are far more varied (see Chart 2). In some cases,the experience is like climbing a hill. But in others, the expe-rience is more like a roller coaster. Looking at such cases,Pritchett (2000) and other authors have concluded that an

understanding of growth must involve looking more closelyat the turning pointsignoring the ups and downs of growthover the horizon of the business cycle, and concentrating onwhy some countries are able to keep growing for long periodswhereas others see growth break down after just a few years,followed by stagnation or decay.

A systematic look at this experience suggests that ignitinggrowth is much less difficult than sustaining it (Hausmann,Pritchett, and Rodrik, 2005). Even the poorest of countrieshave managed to get growth going for several years, onlyto see it peter out. Where growth laggards differ from theirmore successful peers is in the degree to which they havebeen able to sustain growth for long periods of time.

income dstrbuton and growth sustanablty

In our research we looked at the extent to which the dura-tion of a growth episode is related to differences in countrycharacteristics and policies. The quality of economic andpolitical institutions, an outward orientation of an economy,macroeconomic stability, and human capital accumulationhave long been recognized as important determinants of eco-nomic growth. And we found that they matter for the dura-tion of growth episodes too.

We argue that income distribution may alsoand inde-pendentlybelong in this pantheon of critical determinants

of growth duration. At the level of simple correlation, moreinequality seems associated with less sustained growth.

Chart 2

Roller coaster

In developing and emerging markets long-run growth paths

can be steadyor not so steady.

(real GDP per capita, log)

Source: Penn World Tables Version 6.2.

Note: Real GDP per capita is measured in logarithms, which means that the straighter the

line, the more constant the growth rate. The vertical dashed lines represent periods when the

growth rate makes a signicant and persistent change up or down.

1950 60 70 80 90 2000 1960 70 80 90 20007.5

8.0

8.5

9.0

7.4

7.6

7.8

8.0

8.2

1951 61 71 81 91 2001 1954 64 74 84 94 20048.0

8.5

9.0

9.5

10.0

7.6

7.8

8.0

8.2

8.4

8.6

Brazil

Chile Jordan

Cameroon

Chart 1

Climbing the hill

For advanced economies like the United Kingdom and the

United States, income grows at a more or less steady paceover the long run.(real GDP per capita, log)

Source: Penn World Tables Version 6.2.

Note: Real GDP per capita is measured in logarithms, which means that the straighter

the line, the more constant the growth rate.

1950 60 70 80 90 2000

United Kingdom United States

1950 60 70 80 90 20009.0

9.5

10.0

10.5

9.0

9.5

10.0

10.5

11.0


3/4

14 Finance & DevelopmentSeptember 2011

Chart 3 shows the length o growth spells and the averageincome distribution during the spell or a sample o coun-tries. We deine a growth spell as a period o at least iveyears that begins with an unusual increase in the growthrate and ends with an unusual drop in growth. he measureo inequality is the Gini coeicient, which varies rom zero

(all households having the same income) to 100 (all incomereceived by one household).

It may seem counterintuitive that inequality is stronglyassociated with less sustained growth. Ater all, some inequal-ity is essential to the eective unctioning o a market econ-omy and the incentives needed or investment and growth(Chaudhuri and Ravallion, 2007). But too much inequalitymight be destructive to growth. Beyond the risk that inequal-ity may ampliy the potential or inancial crisis, it may alsobring political instability, which can discourage investment.Inequality may make it harder or governments to make di-icult but necessary choices in the ace o shocks, such as rais-ing taxes or cutting public spending to avoid a debt crisis. Or

inequality may relect poor peoples lack o access to inancialservices, which gives them ewer opportunities to invest ineducation and entrepreneurial activity.

Against this background, the question is whether a system-atic look at the data supports the notion that societies withmore equal income distributions have more durable growth.

We study growth spells as medical researchers mightexamine lie expectancy. hey study the eects o age,weight, gender, and smoking habits on lie expectancy; welook at whether actors such as political institutions, healthand education, macroeconomic instability, debt, and tradeopenness might inluence the likelihood that a growth spell

will end. he result is a statistical model o growth dura-tion that relates the expected length o a growth episode(or, equivalently, the risk that it will end in a given year) toseveral o these variables. We compare the risk that the spell

will end in a given year with the values o these variables inprevious yearsat the beginning o the spell or the previousyearto minimize the risk o reverse causality. In the aceo the usual diiculties involved in disentangling cause andeect, and the risk that we have been unable to ind goodmeasures o important variables, the results we report below

should nonetheless be interpreted only as empirical regulari-ties (stylized acts).

he analysis suggests that a number o variables oundto be important in other contexts also tend to be associatedwith longer growth spells (see Chart 4). o show the impor-tance o each variable, the chart (which covers 1950 to 2006)reports the increase in the expected duration o a growthspell or a given increase in the variable in question, keepingother actors constant. o compare the eects o the dierent

variables on growth duration, we calculate expected durationwhen all the variables are at their median values (the valuegreater than that observed in 50 percent o the observationsin the sample). hen we increase each variable, one variable

at a time, and look at what happens to expected duration. Wewant the size o each o these increases to be readily compa-rable. o achieve this, we increase each variable by an amountsuch that it moves rom the median value to a value greaterthan that observed in 60 percent o the sample (a 10 percen-tile increase).

Hazad to sustained gowth

Somewhat surprisingly, income inequality stood out or thestrength and robustness o its relationship with the dura-tion o growth spells: a 10 percentile decrease in inequality(represented by a change in the Gini coeicient rom 40 to

37) increases the expected length o a growth spell by 50 per-cent. he eect is large, but is the sort o improvement that a

Chart 3

Lasting effects

More inequality seems to spell less sustained growth.

(years in growth spell)

Sources: Penn World Tables; and Wide World Inequality Database.

Note: Inequality is measured by the Gini coefcient, which ranges from zero, where all

households have the same income, to 100, where one household has all the income. All

spells lasted a minimum of ve years. No incomplete spells are included. The data cover the

period from 1950 to 2006. Countries in the sample include Belgium, Brazil, Cameroon,

Colombia, Ecuador, El Salvador, Greece, Guatemala, Jamaica, Jordan, Pakistan, Panama,

Singapore, Thailand, and Zambia.

Inequality30 35 40 45 50 55 60 650

5

10

15

20

25

30

35

40

45

Chart 4

Growth spells

Factors have differing impacts on how long growth periods last.

Income distribution appears quite important, whereas other

factors are less so.

(change in expected growth duration, percent)

Sources: Berg, Ostry, and Zettelmeyer (2008); and authors calculations.

Note: The height of each factor represents the percentage change in a growth spell

between 1950 and 2006 when the factor moves from the 50th percentile to the 60th

percentile and all other factors are held constant. Income distribution uses the Gini

coefcient. The political institutions factor is based on an index from the Polity IV Project

database that ranges from +10 for the most open and democratic societies to 10 for the

most closed and autocratic. Trade openness measures the effect of changes in trade

liberalization on year-to-year growth. Exchange rate competitiveness is calculated as the

deviation of an exchange rate from purchasing power parity, adjusted for per capita income.

Income Political Trade Exchange rate External Foreign directdistribution institution openness competitiveness debt investment

0

10

20

30

40

50


4/4

Finance & Development September 2011 15

number o countries have experienced during growth spells.We estimate that closing, say, hal the inequality gap betweenLatin America and emerging Asia would more than doublethe expected duration o a growth spell in Latin America.

Remarkably, inequality retains its statistical and economicsigniicance even when we include many potential determi-

nants at the same time, a claim that we cannot make or manyo the conventional determinants o good growth peror-mance, such as the quality o institutions and trade openness.Inequality still matters when we allow or regional dierencesin expected growth duration (such as between emerging Asiaand Arica). his all suggests that inequality seems to matterin itsel and is not just proxying or other actors. Inequalityalso preserves its signiicance more systematically across di-erent samples and deinitions o growth spells than the other

variables do. O course, inequality is not the only thing thatmatters but, rom our analysis, it clearly belongs on the list owell-established growth actors such as the quality o politicalinstitutions or trade openness.

Do these statistical results ind a voice in the political andeconomic narratives o the actual country growth episodes?It appears to be the case in, or example, Cameroon. Growthaveraged 7 percent rom 1978 through 1985. hen the econ-omy ell apart and declined by 6 percent a year over the sub-sequent decade. Oil wealth in the 1970s initially inancedlarge increases in the public sector, particularly in publicemployee wages, which proved very diicult to cut whenoil prices ell. Although these measures [to cut governmentspending] were necessary to rescue the country rom urthereconomic crisis, they were very unpopular because they leastaected the political elite and those in the upper echelon o

government, whose privileges remained intact (Mbaku andakougang, 2003). Our statistical model o growth durationsuggests that the risk that the growth spell would end in 1985was very highmore than 100 times higher than would betypical or a country enjoying a growth spell. he modelattributes this high risk mostly to Cameroons unusually highinequality as well as its low inlow o oreign direct invest-ment and high degree o autocracy.

Cameroon is typical. We have examined six historicalcases, including Colombia, Guatemala, and Nigeria. hesecases, and our broader statistical analysis o a large numbero growth episodes, suggest that inequality is an underlyingeature that makes it more likely that a number o actors

external shocks, external debt, ethnic ractionalizationcome together to bring a growth spell to an end.

raising the tide

One reasonably irm conclusion is that it would be a big mis-take to separate analyses o growth and income distribution.o borrow a marine analogy: a rising tide lits all boats, andour analysis indicates that helping raise the smallest boatsmay help keep the tide rising or all crat, big and small.

he immediate role or policy, however, is less clear. Moreinequality may shorten the duration o growth, but poorlydesigned eorts to reduce inequality could be counterpro-

ductive. I these eorts distort incentives and undermine

growth, they can do more harm than good or the poor. Forexample, the initial reorms that ignited growth in Chinainvolved giving stronger incentives to armers. his increasedthe income o the poor and reduced overall inequality as itgave a tremendous spur to growth. However, it probably ledto some increased inequality among armers, and eorts to

resist this component o inequality would likely have beencounterproductive (Chaudhuri and Ravallion, 2007).

Still, there may be some win-win policies, such as better-targeted subsidies, better access to education or the poorthat improves equality o economic opportunity, and activelabor market measures that promote employment.

When there are short-run trade-os between the eects opolicies on growth and income distribution, the evidence wehave does not in itsel say what to do. But our analysis shouldtilt the balance toward the long-run beneitsincludingor growtho reducing inequality. Over longer horizons,reduced inequality and sustained growth may be two sides othe same coin.

he analysis calls to mind the developing country debt cri-ses o the 1980s and the resulting lost decade o slow growthand painul adjustment. hat experience brought home theact that sustainable economic reorm is possible only whenits beneits are widely shared. In the ace o the current globaleconomic turmoil and the need or diicult economic adjust-ment and reorm in many countries, it would be better ithese lessons were remembered rather than relearned. Andrew Berg is an Assistant Director and Jonathan D. Ostry isDeputy Director in the IMFs Research Department.

Reerences:Barro, Robert J., 2000, Inequality and Growth in a Panel o

Countries,Journal o Economic Growth, Vol. 5, No. 1, pp. 532.

Berg, Andrew, and Jonathan D. Ostry, 2011, Inequality and

Unsustainable Growth: Two Sides o the Same Coin? IMF Staf

Discussion Note 11/08 (Washington: International Monetary Fund).

, and Jeromin Zettelmeyer, 2011, What Makes Growth

Sustained? orthcoming in Journal o Development Economics.

Chaudhuri, Shubham, and Martin Ravallion, 2007, Partially

Awakened Giants: Uneven Growth in China and India, in Dancing with

Giants: China, India and the Global Economy, ed. by L. Alan Winters

and Shahid Yusu (Washington: World Bank).

Hausmann, Ricardo, Lant Pritchett, and Dani Rodrik, 2005, Growth

Accelerations,Journal o Economic Growth, Vol. 10, No. 4, pp. 30329.

Mbaku, John M., and Joseph Takougang, eds., 2003, Te Leadership

Challenge in Arica: Cameroon under Paul Biya (Trenton, New Jersey:

Arica World Press).

Okun, Arthur, 1975, Equality and Eciency: Te Big radeof

(Washington: Brookings Institution Press).

Polity IV Project, www.systemicpeace.org/polity/polity4.htm

Pritchett, Lant, 2000, Understanding Patterns o Economic Growth:

Searching or Hills among Plateaus, Mountains, and Plains,World Bank

Economic Review, Vol. 14, No. 2, pp. 22150.

Wacziarg, Romain, and Karen Horn Welch, 2008, Trade

Liberalization and Growth: New Evidence,World Bank EconomicReview, Vol. 22, No. 2, pp. 187231.

Documents

Equality and Efficiency, Berg: International Monetary Fund