The Fiscal System and Income Distribution: The Case of Puerto Rico (1977) by La Colección Puertorriqueña

EXCESO DE COPIA CPR UPR RP

TEMAS SOBRE ECONOMIA

DE PUERTO RICO

SEPARATAS

Número 4: Arthur J. Mann PUBLIC EXPENDITURE PATTERNS IN THE DOMINICAN REPUBLIC AND PUERTO RICO, 1930-1970

Número 5:

Arthur J. Mann

THE FISCAL SYSTEM AND INCOME DISTRIBUTION: THE CASE OF PUERTO RICO

Número 6:

Arthur J. Mann y Donald Perlis

UTILITY FUNCTIONS, PUBLIC GOODS AND INCOME DISTRIBUTION

UNIVERSIDAD DE PUERTO RICO FACULTAD DE CIENCIAS SOCIALES DEPARTAMENTO DE ECONOMIA RIO PIEDRAS, PUERTO RICO

EXCE~a c~ pARd4 D( C9f,t4 CpP UpR RP~

UNIVERSIDAD DE PUERTO RICO Facultad de Ciencias Sociales DEPARTAMENTO DE ECONOMIA Río Piedras, Puerto Rico

SEPARATAS NUMERO 4, 5 Y 6

Arthur J. Mann

PUBLIC EXPENDITURE PATTERNS IN THE DOMINICAN REPUBLIC AND PUERTO RICO, 1930-1970

Arthur J. Mann

THE FISCAL SYSTEM AND INCOME DISTRIBUTION: THE CASE OF PUERTO RICO

Arthur J. Mann y Donald Perlis

UTILITY FUNCTIONS, PUBLIC GOODS AND INCOME DISTRIBUTION

Unidad de Investigaciones Económicas

Angel L. Ruíz Director

Miguel A. Ramirez Pérez Coordinador de Publicaciones Ad Honorem

Septiembre 1977

NOTA Las tres publicaciones que aquí se reproducen siguen el objetivo básico expuesto antes de que material de calidad relevante a la economía de Puerto Rico y temas relacionados debe recibir la más amplia divulgación. Para ello nuestro compañero claustral y colaborador, el doctor Arthur J. Mann ha adelantado su autorización para publicar por este medio. Cada uno de los ensayos está precedido de una página a color señalando los datos básicos del número de la separata, autor, título y fuente. Esto facilitará la lectura selectiva del material. Aprovechamos para dar la bienvenida al compañero Donald Perlis, del Recinto Universitario de Mayagüez, como colaborador.

S EPA RA TA Número 4

Arthur J. Mann

PUBLIC EXPENDITURE PATTERNS IN THE DOMINICAN REPUBLIC AND PUERTO RICO, 1930-1970

/Social and Economic Studies, 24 (March 1975:

47-82/

PUBLIC EXPEND1:T1i1,.E P.17TEI;i,NS IN THE 1~()141INfCAN REPUBLIC IND PUERTi) RICO, i93(?-1970 r

.4.tz rl l ~ ; u 1. M ►tári iNCRODUCTION Over the past several decades there have appeared numerous empirical studies which attempt to measure and explain the absolute and relativo growth of public sector activities in the economies of many nations? Almost mva iabty such analyses have demonstrated that in both the developed and the developing countries the government's share of national output has risen over the varying spans of years coveredby each investigation. The purpose of this paper is dual: (a) t-~ fill the data gap with respect to public expenditure growth in the Dominican Republics and Puerto Rico, and (b) to present a tentative hypothesis to better identify life principal factors which have led to an increasingly active role for the public sector in both countries. The hypothesis to be tested is not really novel, for It involves combining the Peacock-Wiseman displacement effect concept (23 pp. 24-281 with two cases of small, open economies. The displacement effect, which postulates an unsated demand for public disbursements, relates to the stepwise manner in which government spending as a proportion of some gross output measure increases (or decreases) over time. Public expenditure, an obvious (unction of public revenue, is constrained by the citizenry's ideas as to tolerable tax levels; under the influence of a 'social disturbance' such tolerability restraints are lifted, thereby providing the mechanism by which a "new plateau of expenditure may be reached, with p+.►blic expenditures again taking a broadly constant share of gross national product, though a different share from the former one" [23 p. xxivj . Use of such a concept thus enables one to measure the time pattern of public expenditure growth relative to that of natfona► output growth.

As generally conceived, the. social disturbance' phenomenon deals with events such as wars and economic depressions which directly influence people's ideas about

the tolerability of a given tax level. The two countries under analysis in this paper were not oveuiy involved, during the 1QÚ -l97(: : period., in world .var, but undoubtedly the repercussions of the lefts depression rcverbt°rated to their detriment. Thus, the second,,; effects of economic cycles in their larger, more developed trading partners certainly were causal ficrors leading to reduced levels of econc mic activity in the more economically dependent, smaller countries. If such a causal relation is valid, then it may be postulated that the social disturbances generated by :Kars, their a terrnath, and other event in iar2er, developed uainna uSditioruaily influence lax and public expenditure patterns (and citizenry attitude) in dependent economies. Such influences are transmitted to the tatter national entities with increased strength due to the extreme openness of their ecurfomic strucl'u-es;4 over the 1931)1970 intervaf the sum of

SOCIAL AND ECONOMIC STUDIES

merchandise imports and experts by value normally comprised some 25 to 40 per cent of GNP in the Dominican case, and a much higher 60 to 90 per cent in the Puerto Rican case. And in each instance the principal trading partner was the United States, which experienced the social upheavals of a great depression and a world war. LIMITATIONS OF T E ANALYSIS The restrictions imposed b} the lack of a reliable gross output measure for Puerto Rico prior to 1930s and for the Dominican Republic prior to 3940 limit the secular relevance of the hypothesis, especially if it is desired to carry out analysis against the backdrop of Wagner's 'law of increasing state expenditures'.6 The displacement effect itself, although su ected to ratter rigorous statistical testing by Gupta 141, has fallen under serious criticism fro;n, among others, Pryor [25 pp. 443.4461, M.usgr:<ve [18 pp. 87-901 and bird 41. There are numerous statistical and conceptual difficulties inhcrent in this type of collection of time-series data for one country alone poses a variety of non-comparable and oft-times non-reconcilable figures. Such difficulties are evident in the large number of sources employed to ferret out ,ihe required primary data (see Appendix B). The quality of the data, especially in the Dominican case during the Trujillo era, may be questionable, and many of the assumptions which necessarily had to be made by tits researcher to approximate essential time-series can only be justified by reasoning that no alternative assumption would represent an improvement over the one actually utilized; e.g., the deriv'alion of Dominican gross national product data for the period 19401949. Although it, is tritely observed that expenditure is a function of revenue, and especially, of tax revenue, little_ heed is taken of the income side of the public budget, concentratioa is ro be. almost solely upon the historical movements of public expenditures. The entire conceptual question of what constitutes government expenditure on cofecti":iy dermrnded goads and services as opposed to public produclion f,i ~., tlt. activities c:r he public corprrrat ores in Puerto Rico or the `autonomous institutions' in the I►orninicar: Republic) is reduced ro the leveT of presenting data for the 'consolidated' public sector in an 4ppeniiix at. the cud of the paper. Titus, emphasis is ;driest entirely on t'ie non-corruuercial expenditure activities of the central government in each rrat'iori, regardless of revenue sol: i 'e1.

study. Mere

This latter point „terits brief elat)oration. The question of which expenditures are most appropriate to measure public activity in any country at any time is conceptu, liy uniesciwd 4n the PescOi;k-WISe ;earl vire'point [.23 p. tj the rctivities of the i:uhlic c:orporalions should bt exci,.:ded from the definition of public sector activity since their production is `;ffectud by market criteria'. However, it might well be gi.esti ar.ed vvhethcr or not ,tram' other public sector activttie.s aomuronly provided are not also of a commercial nature. .k ref tae organiLationai form of a public corporation is not neceasarih' a. reasonable u;clicator of the degree of market criteria affectation. This is especially true of i:he Puerto Rican case, where institutions such as the Univeráity of Puerto Rico area the lirstitute of Puerto Rican Culture are formally classified as public corporations. In the opriion of Ifird ty p. 19íj, which is similar to that of Peacock arel Wiseman, those l uhilc:ogerrries that are price-chargers or profit-motivated

E"CITEY,Tr

-ri`1'tERNS

are to hP :..torte ~...:;. any definition of government. `yet, does the fact that a r .scat tuition fee is levied against the uni ~rsity student by the public authority lead to the conclusion that disbursements on oobiie university education be excluded from public expenditure? To .lo so would surely- introduce distortion. 1t appear' then, that a good case can be made for adopting a broad view of what constitutes public activity. Unfortunately, however, serious and complf x statistical difficulties regarding the expenditures of the public enterprise sector cause one to fall hack upon ₹he narrower definition of governmental activity which focuses upon the outlays of he central government alone, ir.. she process excluding meet of the spending of the public corpor tions. The sole reliable data with respect to the exYenditu-c°s cf the Puerto.. Rican public corporations cover the years 1962-1966, and were coripiled in a special and unpublished study carried out by the Puerto Rican Budget Bureau in 1966. To gather similar data for the years 1940-1961 would require the coa~eration of both the Puerto Rican central government and all of the public corporations, and even if the requisite cooperation were forthcoming the task would be of gargantuan proportions. The figures that do appear in the Puerto Rican `consolidated' public sector table for the years 1930-1960 (Table C-3! are this author's own rather crude estimates gathered from a variety of independent published sources. As such, they are reluctantly employed to any great extent to expir:sÉ apfaaent changes in he public sector's relative economic role. The Dominican public enterprise data (Table C-1) are even more deficient, being available e to this researcher) for the public corporations only since 1967. Thus, although it would be conceptually more appropriau to consi cntiy use one definition of the public sector (that r,fhich =nc►:ides the activities of all public enterprises), dare deficiencies render impossible the use of the preferred methodology. The result is that the text discussion concentrates solely upon the activities of the central government f sc, a procedure which in reality understates the relative role of the pu1s1ic sector. SUMMARY OF 1^li Dll?ti: S Over the period of years encom a.ssed by this !tudy sg regnte public expenditures in current price;. multiplied many times over in both countries. From 10.6 n illion pests irs 1936 the disbursements of the Dontir.tean central goverru tent rose to 259.6 million pesos in 1910 (see Table 1);7 these of he central governs.:at of Puerto Rico increased from 14.7 mthon dollars in 192?i to 975.7 million dollars in 1970 (see Table 2).a As is revealed in Table 3 the growth rate of public citpenditures as measured in current prices over he 1940-191'0 period outstril ped that of asgregate expenditure on gross output, thereby leading to an upward displace tent (Table 1 and 2, column 7). The public share of gross national product in the Renzbiic rose from around 11 per cent in 1940 to lg per cent in 1970, and actually averaged over 22 per cent for the 15 year period subsequent to 1948. The time pattern of displacement is found to be quite irre y lar, and further contemplation of the data reveals that dis~tiacentent in fact occurred over a much srortc. time span in the 1940s, when he public portion of national output tripled before falling back to a relatively steady plateau fuetuatiig slightly between 20 and 26 per cent until

SOCIAL AND ECONOMIC STUDIES TABLE 1 GROSS NATIONAL PRODUCT AND CENTRAL GOVERNMENT EXPENDITURE, CURRENT r1 ND CONSTANT PRICES AND PER CAPITA CONSTANT PRICES, DOMINICAN REPUBLIC, 19364970 (millions of RD$) Gross Rational Product

Centrist_..._ Ggvernment E _.

nditure

t3xpenditure a5 Pratportion of

(:Np (1't Current Prices Year

(2) Constant Prices (1962 100)

(3) Per Capita Constant Prices (in pesos)

(4) Current Prices

1936

10.6

1937

10.5

1938

(1962 = 100)

(6) Per Capita Constant Prices (in pesos)

(5) Constant Prices

(1)

(8)

(4)1(1)

(6)((3)

11.6

1939

11.5

1940

110

1941

120

324

189

12.0

32.4

18.9

10.9 10.0

10.0

1942 1943

149

331

188

14.2

18.0

9.5

.9.6

163

267

148

17.5

11.8

1944

194

259

140

19.2 26.4

31.6 31.5 35.2

19.1

13.6

40.3 49.9

21.3

13?

13.7

25.8

19.1

70.4

35.5

2,8.4 29.4

12.0

1945

218

295

156

29.8

1946 1941

225

135

42.9

257

262 247

1948

293

282

125 139

73.2 86.4

40.9

1949

335

376

181

40.1

22.2

1950

383

459

215

74.3 90.7

83.1 83.5

28.5 29.5

105$

49.4

23.7

23.0

1951

455

527

238

Itl2.2

106.5

48.1

22.5

20.2

19.52

50)

542

236

129.1

131.7

25.8

24.3

t953

510

539

221

129.9

132.6

57.3 55.7

25.5

24.5

1954

537

107.9

116.0

47.1

20.1

19.9

5' S

583 599

231

~55

235

127:2

136.8

53.6

22.9

22.8

?956i

630

670

253

144.6

157.2

59.5

23.0

23 5

1957 1954

705

739

270

153.0

151$

55.3

2l.7

20.5

110

753

265

164.5

16?_"

1959

730

248

154.9

163.1

75.:

1960

689 715

781

250

147.3

150.3

493

au.b

19.3

1961

688

139

236

151.1

48.3

20.2

20.5

57.4

23.2

2l.7

22.5

22.4

1962

870

270

139.0 184.7

184:?

57.4

21.2

213

1963

994

282

189.5

175.5

529

I9.1

18.8

1964

1,082

•935 989

290

212.5

193.2

56.6

19.6

I9.5

196$

949

f,'48

241

145.7a

123.5

35.2

14.6

1966

1,043

yS 3

264

188.7

16 .5

46.6

15.4 18.1

1967

1,096

Y86

265

198.0

172,2

46.3

1968 1969

1,145

1,003

262

207.7

168.9

1,298 1,453

1,111

282 290

235.4 259.6

199.5

1970

1,163

Saun•es ana Noxes: See Appendix R.

220.0

44.1

18.1 18.1

17.7 17.5 16.8

50.6 549

18,! 17.9

18.9

17.1)

PUBLIC EXPENDITURE PATTERNS

GROSS NATIONAL PRODUCT AND CENTRAL GOVERNMENT 2 EXPENDITURE, CURRENT AND CONSTANT PRICES AND PER CAPITA CONSTANT PRICES, PUERTO RICO, 1928-197O (millions of USSi

TABLE

Gro

(1) Current Prices Yest 19:8 1129 1130 193) 1932 3933 (934 1935 1936 19'37 1938 1939 I940 (941 1942 1943 1944 I945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 i965 1966 1967 1968 1969 1970

241 241 215 199 I79 228 226 254 268 271 264 287 330 .380 437 503 553 564 612 65) 719 755 8) 5 968 1,048 1,1Ü4 1,142 1,198 1,265 1,3'70 1,518 1,681 1,839 2,045 2,270 2,491 2,766 3,063 3,368 3,735 4,126 4,602

N~on~l P~Sx1uct

(2) Constant Prices (1954 100)

359 366 392 372 439 395 427 436 458 463 499 554 593 587 605 641 635 705 732 819 879 925 1,Olti 1,081 1,104 1,139 1,186 1,'222 1,258 1.364 1,475 1,559 1,674 1,810 1,923 2,070 2,234 2,325 2,462 2,628 2,806

Sources and Notes: Sec Appendix B.

(3) Per Capita Constant Prices (in dollars)

234 234 245 228 264 233 247 248 .255 253 268 2'92 307 299 303 316 307 336 343 378 399 415 455 488 500 510 527 542 553 590 630 655 690 733 764 808 860 886 929 979 1,035

CentraLGcwernment Expenditure_

(4) Current Pricey

18.7 (8.8 19.7 20.3 20.6 17.6 22.2 20.6 21.3 22.8 25.5 29.5 29.0 31.5 44.2 52.8 64.6 72.8 105.1 102.3 103.4 115.5 116.5 124.7 141.2 161,6 171..; 189.0 189.0 225.0 233.3 242.1 274.1 307.7 319.6 375.5 425.2 479.8 566.0 686.0 743.6 834.2 975.7

(5) Constant Prices (1954 = 100)

(6) Per Capita Constant Prices (in dollars)

Expenditure as Proportion of GNP (7) (8)

('4)1(1) (6)1(3)

: 24.9 29.2 34.2 30.8 36.1 30.5 30.4 31.3 35.5 43.4 42.7 51.7 56.4 67.5 81.2 95.6 115.4 119.(i 1.17.5 129,2 130.3 138.1 146.2 166.1 171.3 188.9 182.5 213.7 215,7 222.4 236.7 258.0 259.8 296.0 328.2 363.2 418.7 493.3 526.9 579.2 640.0

16.2 18.6 21.4 18.9 21.7 18.0 17,6 , 17.8 19.8 23.7 22..9 27.3 30.3 34.4 40.7 47.1 55.9 57.0 55.1 59.6 59.2 62.0 65.5 75.0 77.5 84.6 81.1 94.8 94,9 96.2 101.2 108.4 107.1 119.9 130.4 141.8 161.2 188_t 198.9 215.7 236.1

7.8 8.2 9.4 I0.4 9.8 9.1 9.1 8,4 8.5 9.2 11.2 10.1 9.5 11.6 12.1 12.8 13.2 18.6 16.7 15.1 16.1 15.4 15,3 14,6 15.4 15.5 16.5 15.8 17.8 17.0 15.9 16.3 16.7 15.6 16.5 17. I 17.3 I8.5 20.4 19.9 20.2 21.2

6.9 7.9 8.7 8.3 8:2 7.7 7.i ' 7.2 7.8 9.4 8.5 9.3 9.9 11.5 13.4 14.9 18.2 11.0 16.1 15.8 14.8 14.9 14.4 15.4 15.5 16.6 I5.4 )7.5 17.2 16.3 16.1 163 15.5 16.4 11.) 17.5 18.7 21.2 21.4 22.0 22.8

SOCIAL AND ECONOMIC STUDIES

late

the 1Q6 s. it is this six year period in the 1940s which deserves deeper analysis. A similar but less drarnatu upward displacement is also noted in the Puerto Rican case during the i94os, at which time central government expenditures as a percentage of GNP rose from a low of 9.5 per cent in 1941 to a high of 18.6 per cent in 1946. That this new plateau remained at a higher level than the previous plateau of the 1930s is evident, i.e., displacement was not a temporary henomenon. Thus, there appears tc` be strong evidence, in terns of current prices, of a perinanent upward displacement having occurred in both countries during the mid to 194Qs. Before pursuing this apparent phenomenon further, the so-called permanent influences of population and price level changes on the relative expenditure pattern must be accounted for and removed, for undoubtedly a large part of the absolute growth observed in both the CNt' and expenditure series is due to population growth and currency depreciatio:i. Allowance for price level changes generally means elimination of the influence or inflation, for over the period analysed the overall price level, as measured by various price 1n fires, rose in both cases (although it remained quite stable in the t)omiai.cah Republic of the t 950x). The indices employed to deflate the data are eviderr ly the best availat''t., and are most appropriate in the Puerto Rican rase after i947. Use of a wholesale price index to deflate Dominican public expenditure data is dictated by the abs* rir:e of a more appropriate rrkflator, and the results are severely limited since it applies to the city of Santo Domingo alone. Expenditure data expressed in real terms are Sound in columns 2 and 5 of Tables 1 and 2. Although absolute experntiture growth is consequently diminished, it nevertheless, exhibits a multiple h.crease when comparison of the initial year data is made with the tetminat year rratir,tics. Since the population of each country increased steadmiy over the years (with the exception of several years in the 1950s when Puerto i2ircan emrgt-atino ro the United States was so great a to counteract naturi growth), growth rtes of real per capita data are further lowered; ref eretice should be rte;, ,o columns 3 and b of Tsbes 1 and 2. Despite removal t:£ the i;armanenr t ences the time pattern of relative public expenditure grr.::wh ierrsairts esaentiall) unchanged (cotnpare columns 7 and 8 of Tables 1 and recou _.re rmrr:r h,r had to other methods and thta breakdowns to explain the observed propottioraal fluctuations.

data

FUNCTIONAL i3XTENDI1'UkE Ch ECORIES gi rder to Letter i(entify the reasons 1'ur the upward displacement the are dz54gregated into functional and economic categories.9 The functional categories selected are basically self tiescripUve, end an effort war made to insure comparability between the two country cases. One difference does stand out, however. The bulk of Dominican outlays on `t'rote, ton of Persons and Property is composed of spending [.n the national armed forces, wherc Puerto Rico is spared such exhaustive expendaures simile they are assumed by the United States federal government. Although it is not customary to prer.ens debt. service (both principal and interest payments) as a

P1It3 IC EXPENDITURE PATTERNS

year

separaa.a category, nor indeed to include ,Iebt principal payments as part of public expenditure, there are several compelling reasons for thus proceeding. The public debt in both nations is/was largely external, and since both are/were highly dependent on foreign capital the government must manage its debt in such a way as to maintain good external credit standing. Such management draws on funds that might have been employed in other functions, and certainly causes a drain on the national treasury in the particular year the amortization payment is effectuated. It is true, of course; that this procedure involves .double-counting; i.e., in the year loan revenues are spent and again in the year the principal is repaid. However, double-counting arises in terms of the entire time period covered by loan contraction and liquidation, a time span usually much longer than one year. Therefore, in terms of expenditure data relating to one alone little double-counting occurs. Moreover, debt service payments have been directly linked with political policy, especially in the Republic's case, and in the final analysis this study deals with the politics of public expenditure patterns. As will be shown below, these expenditures represent a vital factor in idéntifying the causes of Dominican upward displacement in the mid-i940s.10 In the Puerto Rican case debt service payments remained at a proportionally high level throughout the 1930s, and since 1950 a large portion of the increase in the Commonwealth's public capital has been financed through bond issues floated in New York.

public

Only the disbursements of the central government are accounted fqr in the functional breakdowns presented below, and grants to municipalities comprise only those monies emanating from 'owned' central government revenues, thereby excluding those revenues collected for the municipalities to augment administrative efficiency. Since both nations are unitary states, such a procedure involves ignoring the disb'.trsentenes of the municipal governments. lrxclusion of Dominican municipal data hardly distorts at all the growth of Dominican public expenditure, for if grants to murúcipali:ies are subtracted from municipal spending the total expended by the local government never rises above S per cent of central fist outlays for the entire period. However, some distortion is introduced in the Puerto Rican case. Clark (9 p. i49) found that in 19'28 municipal spending comprised approximately 50 per cent of central government spending after allowance is made for grant-financed disbursements; this proportion steadily dropped over the years, teaching r bout LS per cent in 1945 as the Puerto Rican displacement began to make itself manifest. Since that time it has fluctuated between 12 and 17 per cent. Given the difficulty of distributing municipal non— budgetary expenditures among the tunctional categories these disbursements (both budgetary and non-budgetary) are wholly excluded, hereby understating the magnitude of the 'true' ron-commercial/trading public sector. However, the municipal.totals are to be found in Table C-1 and C-3, which contain data concerning the `consolidated' sector. Finally, central government grants to public corporations, public service enterprises, and trust funds have been distributed among the categories aecordirig to purpose; e.g., contributions to a public corporation engaged in the promotion of industrial development are allocated to economic development.

TdlRL.T 3

CU%tL`1..ATP/E AN?áUA7.. (.;. "•ow~á Éi RAT£S v'`b~ C~F`S3 NATt4'.`13. Lfs~1JCT, ($ i :tA.1.. G'L~ F.RNMiNA `4ts1TURE, .AND N1C~~,,.~i PPUEL1C MIL) PUESTO RICO, 1530-1970 t.;'LA.i.~.ON,

Ef';

Peuoá

Gross Ntiomi Pro uct (curzc ~t puces) (ccsnstat p: ces) Dam. Rep. P.R. . Dom. Rep.

Pcpt tfe,r

Goverment Expenditure (currcnt prices)

Dana, step.

P.R.

(constant prices) Dom. Rep.

F.R.

Go, Rep.

P.R. ~

1940-f4

9.G

9.7

194ú>50 195~yC

I3.3 6.4

1969-7G

r.M

144á-48

11.3

1530`7O

4.S"

5.9

1<).8

12.4

6.6

9.4

3.0

10.2

3.9s

5.$

22.4

11.8

2.5

1,7

3.5

5.3

5.0

14.9 8.9

14.Oa

8.3

3.6

6.2

3.6

0.6

10.6 9.4

4.1

6.6

5.8

13.5

3.9

10,5

2.8

1.5

-2.6

3.6 5.3

35.1

33.2

17.5

12.4

2.4

1.?

8.5

2.9

1.4

7.7

~~egs~ng yeac of pe4iod ss 1941.

10..2

1.3

PUBLIC EXFENDITURE PATTERNS

TAJ3Llr 4

CENT[LAL ('iOYERNnENT ExPENnrrUftE Bx FUNCTIONAL CA':'LGDitY, 1)fJMINICAN REPUBLIC, 1936-1970

(millions of RD$)

Year

1936 t937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 ¡953 ¡952 1953 ¡954 1955 1956 1957 ¡958 1959 1960 1951 1942 ¡963 ¡964 3965 ¡966 1967 1968 I959 1970

General fhvtectiºn o' Administra• Persons and Dehl tion 4erviee Property

Education

Glints to 5oc~s! Economic Municipal. Total Development Development ities

t.6 1.2 ¡.5 1.2 1.5 1.5 2.3 NA NA NA NA NA 10.2 11.9 12.8 12.0 ¡5.8 12. ) 15.5 20.0 l9.6 16.1 19.2 ¡G.3 16.8 18.4 2110 20.4 '_0.9 I1.7 21.5 15.4c 18.5 2<.Y, 26.0

1.0 t.0 1.1 13 1.4 1.3 13 ¡.6 1.8 19 NA 3.0 34 3.7 4.0 4.6 8.3 7.9 7.9 9.5 10.4 11.6 11.8 1e.3 9.1 9.1 ¡3.2 17.11 24.6 ¡8.5 26.2 28.8 3J.4 36.3 42.0

0.9 0.8 0.9 0.9 10 0.9 ¡.4 1.5 2.2 2.4 NA NA 10.3 7.3 7.8 9.2 10.8 9.3 ¡0.9 ¡4.¡ 14.7 13.7 17.8 15.9 15.7 16.7 18.7 23.1 23.0 2!.! 25.9 31.8 35.3 36.8 45.3

2.7 3.0 3.0 33 3.5 3.6 4.0 2.2a

_.9a

3.2a NA NA tS.O 15.1 23.5 24.9 35.1 31.0 29.6 26.4 31.6 37.7 42.5 52.3 45.5 38.8 52.2 52.5 56.7 43.3 52.8 45.5 46•3 47.5 47.9

t.0 1.7 1.4 1.E 1.6 1.3 3.6 1.5 ¡.2 1.9 5.7 18.0 23.0 17.1 24.8 26.6 249 39.3 0 0.5 1.6 3.1 2.4 3.3 3.5 2.7 17.2 2.2 7.5 12.1. 6.5 6.3 6.5 f0.8 ¡1.8

3.1 2.5 3.4 2.9 2.i 3.0 3.1 4.1 6.9 5.8 NA NA 23.3 18.1 11.5 23.7 32.3 27.7 40.8 54.1 62.9 66.7 66.7 49.7 52.0 46.5 49.6 57.4 63.6e 22.! 45.0 56.7 g9.7 70.2 7..9

0.3 0_t 0.3 0.3 23 0.4 0.3 0.4 0.5 08 0.9 09 1.2 1.1 3.3 1.2 1.9 2.4 3.2 2.6 3.8 4.1 4.1 6.1 4.7 6.8 13.t¡ ¡6.1 ¡6.2 ¡0.9 10.8 ¡3.3 I1.0 11.0 9.7

10.6 30.5 11.6 31.5 12.0 t2.0 14.2 t9.2b 26<b I9 fib 42.9b 73.2b 86.4 74.3 90.7 102.2 129.t ¡29.9 .107.9 ¡27.2 i4t.6 153.0 164.5 154.9 147.3 139.0 184.7 189.5 145.7 1€8.7 ¡98.0 207.7 235.4 259.6

alne!tN:d defcase expenditures or~iy, as other fncljsi vst (e.g., po¡ice) are not rvaei~l)1e. blkrad+ do ri ;t add to total Sue to speeif~c funet;urul category exduaion9; rite w tat itself ú, bowaver, retu b¡a. CSirwY L intnic3n ofíikcJ s4stists Wrre presented unlit 1967 rt, the basis of 5p>zsx ing by ministries, the data ¡a ¡Mr category are pr0b bi' orerst.ated 1urIny ahE ps'sinus years, foe to tt eeedisbur+euK' . which atre not ctxufabl ¡a any otbet category were :::cludad kere. ¡'JA .. data not avaii Wc. Soorcee Append t B,PuWi, Ex+~rer.4inrmJlare.

SOCIAL AND ECONOMIC STUDIES

56 TABLE 5

Year

General Administj. titer

1928 2.2 1929 2.3 2.8 2930 1931 3.1 1932 3.3 1933 3.0 1934 28 29 2935 1936 3.2 1937 3.5 3.8 1938 3.8 1939 3.4 1910 4.5 1941 1942 S.l 1943 7.0 2944 8.4 194$ t0.7 1946 12.0 1947 128 194$ 13,4 1949 13.5 13.8 1950 1951 14.1 1952 15.9 1953 18.0 1954 184 29.? 2915 2956 . 17.9 1957 20.1 2956 20.9 1459 22.7 1960 ?5.9 1961 1:.4 24.5 1902 1963 35.1 1964 40.0 1965 46.2 1966 4d.5 53. ' 1967 IsbB 59.r1 1969. f,3.0 671 1970

BY FUNCTIONAL CENTRAL GOVERNMI SIT EXPENDITUI CATEGORY, PUERTO RICO. 1928-1970 (millions of currant USS) Protection of Persuns and Debt Property Servi~t

kdut soon

Gantt to Social &0n0elic Munleipnb Tolel Development Development (ties

2.8 2.6 2.i 2.1 2.3 1.9 2.1 2.1 2.2 2.6 2.7 2.9 2.8 2.9 3.3 4.0 4.6 4.6 . 5.1 5.2 7.9 7.8 8.3 10.8 14.2 15.0 22.4 17.1 18.7 22.0 22.5 21.7 25.6 25.8 27.4 33.6 34,1 419 434 56.5 70.1 73.1 79.R

4.8 49 4.$ 5.0 54 4.5 4.6 5.2 53 6.2 ', .4 7.9 8.0 8.1 9.7 1.2 12.0 26.1 16.7 27.2 32.6 3S:5 33.1 322 34.0 37.8 40.3 4333 52.1 61.6 62.9 63.1 74.4 92.5 98.6b 106.8 124.3 134.8 250.9 194.9 209.1) 2.31.6 276.5

2.0 2.! 3.1 3.3 3.3 2.0 5.4 3.8 2.3 2.5 4.2 3.9 4.2 43 5.3 63 13.8 17.6 32.7 30.3 21.1 32.6 35.8 38.9 44.5 50.8 56.8 56.6 55.5 72.7 69.1 63.0 79.0 79.3 84.2 201.4 l374 134,8 159,7 174.9 189.0 223.8 264.3

2.2 26 2-d

3.3 3.4 3.7 3.0 2.7 5.6 3.6 2.6 5.0 2.1 3.0 4.4 8.6 4.8 2.5 3.0 1.0 0.9 1.1 1.2 2.0 2.4 2.6 2.6 3.7 5.7 4.3 5.1 8.1 12.6 13.3 15.5 17.9 20.7 23.6 26.4 28.9 33.4 36.6 43.0

4.7 4.2 4.0 3.4 2,8 2.3 4.2 3.7 2.2 4.2 4.6 5.7 7.7 8.t . 15.7 24.5 206 18.5 34.6 . 253 26.7 23.8 23.1 23,9 29.0 34.2 32.0 45.7 35.0 41.7 50.3 52.9 52.1 04.8 60.0 74.2 8?.4 66.4 12fi.3d 164.9 168.5 290.5 2163

0.1 0.1 0.1 0.1 0.1 0.2 0.1 0.2 0.3 • 0.2 0.2 0.3 0.4 0.4 0.7 I.0 1.4 3.8 1.0 0.5 0.8 1.1 I.R. 2.8 1.2 3.2 5.6 32 4.2 2.6 2.5 S.l 4s 4.6 4.4 6.3 6.3 6.6 7.4 12.2 13.8 15.6 18.5

18.7 T8.8 19.7 20.3 20.6 17.6 22.2 10.6 21.3 22.8 25.5 29.3 29.0 31.5 44.2 524 64.6 728 105.1 102.3' 103.4 115.5 116.5 124.7 141.2 162.6 1713 189.0 189.0 225.0 2333 242.1. 274.1 307.7 319.6. 3753 425.1 479.8 566.0 686.0 ',43.6 8342 975.7

aCurtenl aad capital grants to p.iUu euepnt46una ate distributed truing the categn fee. bt;adodea those expeoahrurrs of the UnMnsiry nl P,acato R ca, financed by 0te antral guvernnteet sldtough for badgstary parpnses the enivereary became a pe.blie .Vtpnr4tiltn an t 162_ clnciades can tral gowtnmt at faaaunced capea4 tires al' Iba Medical ('en Les' in Rlu Piedras. dlncludes current and apitaf espcndittues .7f dw ttay,hways Au:taurity Itepnmag in 1966. 5a)umt'3: Appcndi' B. Petb6e txpend0aate Data

P13BI.IC EXPEN1311'17RF PAT3'EIiNS

C`,17VTTtAL GOVERNMENT EXPE.1dD1T1JR1: BY FUNCTIONAL CATEGORY, DUMINICAN REPUBLIC AND !kTERTO Rico, 1928-1970

TABLE 6

(percentagc. distribution)

Year

Cenerat Adn'dnist. C1.R- P.R.

1928 1929 ¡930

11.2 12.2 14.2

I932 ¡933 ¡934 1935 1~36 1937 1938 939 1940 1941 !942 1943 1 rt4 1945 ¡446 ¡947 1948 1949 1950 19:01 1<¡52 1953 1454 1955 1956 1957 1958 1959 1960 1961 1962 1983 1•H+4 1 965 !96f. 1967 1968 1969 1970

•

16.0 17-13 12.6 14.1 15.0 15.4 35.0 ¡2.9 I 3.4 ¡4.3 11.5 23.? 13.0 14.1 ¡1.4 12.5 13.J 11.7 ¡1.8 11.3 11.3 ¡ 1. i 10.9 10.4 9-5 8.9 9,0 9.4 v.4 8.9 9.2 9.3 9.4 96 3.7 7.8 8.0 7.ü 6.9

15.1 11.4 ¡2.9 ¡0.4 12.5 12.5 16.2 -¡1.8 ¡a.0 14,1 11.1 L.2 9.5 ¡4.4 t 5.1 13.6 ¡0.5 ¡1.1 ¡05 ¡1.4 13.2' 10.8 1(L8 98 12.1 11.4 7.8 8.9 9.7 ¡0.0

ProtectionPers. and ?rap. DR. P.R.

-25.5 284 25.9 28.7 29:2 300 28.2 11.5a 11.0i 10.7° -17.4 20.3 25.9 24.4 27.2 23.9 27.4 20.8 21.9 24.6 25.8 :13.8 319 27.9 28.3 27.7 26.7 29.7 28.0 23.0 22.3 20.2 ¡9.5

15.0 13.8 10.7 l0.3 U.? ¡0.8 9.5 10.2 10.3 11.4 10.6 9.8 0" 9.2 7.5 7.6 7.1 6.3 49 5,1 7.6 o.b 1.1 8.7 10.1 9.3 9,0 9J) 9.9 9.8 9.6 9.0 9.3 8.4 8.t> 8.9 8.0 8.7 7.7 8.2 9-4 8.8 8.2

Debt Service

r1.Ii.

P.R.

••--

11.8 ¡3.R 14.2 16.3 36.5 21.0 13.5 13.1 28.3 ¡56 102 17.0 7. 2 95 10.0 irr.1 7.4 2.1 :A . 1.0 0.9 LQ 1.0 -t.ti 1.1 1.8 1.5 2.0 3.0 1.9 2.2 3.3 4.6 4; 4.8 4_8 4-9 4.9 4.7 4.2 4.5 4.4 4.4

-9.4 16.2 ¡2..1 ¡3.9 13.3 ¡c1.A ¡1.3 7.8 4.5 6.4 73,3 24.6 266 22"==:0 26.2 26.0 19.3 30.3 0.0 0.4 1.1 2.0 ¡.5 2.1 2.4 1.9 9.3 1.2 3.5 8.1 3.4 3.3 3.) 4.6 i.5

IiJucatia.t D.R. P.R.

---

9.4 9.5 9.5 11.3 1!.-r :08 ¡0.6 8.3 6.8 6.4 - . 41 34 5.0 44 4.5 6.4 6.1 73 73 7.2 7.6 7.2 7.3 6.2 6.5 7.1 9.4 I1.6 12? 13.9 745 24.6 ¡5,4 16.2

25.7 26.1 24.4 24.6 26.2 25.6 20.7 25.2 25.8 27.2 29.0 26.8 • 27.6 25.7 25.7 220 i 7.0 22.1 (Sc, 26.6 31.5 30.8 284 25.8 24.1 24.4 23.5 22.8 27.6 27.4 27.0 28.7 ?7.1 3ü.1 •30.9 28.4 29.2 29.1 26.7 28 4 28,1 27.8 28.3

aDefence enpeo'36uos only. N;te' DelMl .n y nol add to 100.0 p9 cent dne to rounding. Suur."r: Tables 4 ad 5.

Soci:71 I)evelop¡nent D.R. P.R.

-•

•8_5 7,6 7.8 7.8 9. 3 7.5 9..9 7.8 8.3 8.1 -,11.9 9.8 8.6 9.0 64 7.<^ 10,1 71.2 10.2 9.0 10.8 10.3 10.'1 12.0 10.1 ¡2.2 ¡0.8 14.5 ¡2..7 76.7 ¡7.0 15.0 11.4

¡0.? 11.2 1S,7 18.3 16.0 li-4 24.3 18.4 10.8 i lA 16.5 13-2 14,1 14.3 12.0 12.3 21.4 24.2 31.1 29.6 20.4 28.2 30.7 31.1 31.5 31.4 33.2 29.9 29.4 32.3 29.6 26'2 28.8 25.8 26.3 21.0 27.6 28.1 28.2 25.5 25.4 26.8 21.1

Grantr ta MuniciptiEcornunic Lievelopnncnl Sties 1).R. P.R. t?.R P.R.

v 29.2 29.8 29.1 23:I 22.5 25.0 21.8 21..4 26.1 19.5 . .. ' -•• ' 27.0 24.5 19.3 23.2 25.0 21.3 37.8 42.5 43.5 43.6 40.5 32.1 35.3 33.5 26.9 303 29.9 15.2 23.8 28.6 28.7 29.8 29.6

25.1 22.3 20.3 16.7 11.6 13.1 18.9 18.0 l0.3 18.-7 18.0 ¡9.3 26.6 25-7 35.5 27.5 31.9 25.4 32.9 24.7 25.8 20.6 19.8 19:2 20.5 21.2 18.7 24.2 ¡65 18.5 2t,6 21.9 19.0 21-1 18.8 19.8 19.4 18.1 22.7 24.0 22.7 22.8 23.2

•• •2.8 2.9 ¡.6 2.6 2.5 3.3 2,1 2.1 1.9 17 1.2 1.4 ¡ .5 1.4 1.2 1.5 1.8 3.0 2.0 2.6 2.7 2.5 3.9 3.2 4.9 7.5 8.5 7.6 7.5 5-7 6.7 5.3 4.7 3.7

0.5 0.5 0.5 O.S 03 ¡.1 0.5 1.0 14 0.9 0.8 1.0 1.4 ¡.3 14 1.9 2.2 5.2 1.0 0.5 0.8 1.0 1.0 2.2 0.9 2.0 3,3 ¡.7 2:2 1.2 i.l 2.1 1.6 1.5 1.4 1.7 1.5 1,4 1.3 1.8 1.9 1.9 1.9

SOCIAL AND ECONOMIC STUDIES

Reference is made to Tablas 4-6 to discuss the functional expenditure cattepuiy data. The Dominican data in Table 4 covering the years 19364947 (1942 excepted) are budget figures, not actual expenditures. However, due to the conservative nature of Dominican budgetary practices in those years there is most likely a close approximation of actual disbursements to projected outlays [8, p. 253). Unfortunately, this was not the case for the four vital years (1944-1947) which witnessed the elusive upward displacement. Expenditures in these years greatly exceeded budgeted outlays because of extensive revenue rises due primarily to high export prices; the latter subsequently increased the capacity to import, and a major portion of public revenue was derived from import duties and taxes An additional problem was created in interpreting the functional data for these four wears because the rnodc: of presentation was completely revamped. As a result it is necessary to omit enurely,anny figures for certain functional categories in the years 1943-1947, and to omit, for all categories excepting debt service, the year 1946. However, since the data for the initial year and the terminal year (1942 and 1948) of the displacement period are actual and reliable expenditure figures, comparison of the changes undergone within the functional distributions is feasible, although it is not passible to chart the exact year to year variaticxns. Although the reliability of the Dominican data may be fairly well accepted in the just-mentioned sense. a knowledge of the historical background of the years 1930-1961 reduces one's confidence in the validity of the official figures. Under the Trujillo dictaroahip the Republic was c+»Everted into a tightly controlled, totalitarian state, in which unrestrained economic tyranny was the role. The country became the private fiefdom of une family and an ever-changing coterie of friends, thereby blurnnf the distinction between government expenditure on public and on private account; t in such an environment it is certainty legitimate to question the validity of the data presented in the official publications, and there is no alternative but to do so. Table 4 reveals the fairly steady growth of Dominicans public outlays by functional category over the 1926`!970 period, the one significant exception being debt service spending (to be discussed below). The percentage distributions presented in Table b reveal that spending on he protection of personas and property (military included) normally claimed one-quarter of public resources, in stark contrast lo the Puerto Rican percentage distributions. Expenditures on education only in recent years form a significant portion of public ditbur'emens, whereas outlays on economic development were consistently high, especially in juxtaposition to hose on social development. Over the 1942-1948 interval the principal determinant of the upward displacement was the debt service category. in 1942 debt service comprised 113 per cent of total expenditures; this proportion had risen to 26.6 get cent by 194$. The reason Dominican debt payments rose so abruptly as to partially cause a tripling of the relative public expenditure level has its base in history and in he psyche of one man. The Dominican external debt, accumulated sine 1869, had led to various foreign interventions in Dominican affairs. The Dominican-American Convention of 1907 established a U.S. Customs Receivership, under which 55 per cent of

PUBLIC EXPENDITRE PATTERNS

wield

for

their

collections were to he applied to discharging debt obligationL t2 A period of increasing internal debt was followed by a U.S. inilitury occupation from 2916 to 1924. By the T mujillo•Huil Treaty of 1940 the Receivership was terminated, and ftonm 1940 to 1946 amortization was made at an accelerated pace. Establishment in 1947 of a central bank, the substitution as legal tende.r of a new Dom inican peso for the previous medium, the dollar, and the accumulation of large gold and foreign exchange balances during the war and early post-war years enabled the government to totally high relative and liquidate ids external debt in 1947. Debt payments maintained absolute levels in are endeavour to also liquidate the internal debt, which was accumpiished in 1953'3 The drop in total expenditures in 1954 should be noted, it is due almost entirely ro changes in the debt service category. Thus, the how" of he 1942-t94í upward displacenrentis partially explained. To answer the question of why' it happened is more diftócult, for recourse cannot be made to the `crisis' explanation of the Peacock-Wisernart hypothesis. No read internal social disturbance took place. Rather, and in obvious speculation, Trujillo may be regarded as an economic nationalist, at least in this case. His desire to liquidate he external debt may well have been rooted in the sordid history of the Republic, so prone to foreign interference. It is not difficult to posit that a man of so great an ego would chafe under the bards::; of ouiside control.''' In addition to the psychological' theory of Dominican displacement, events in the extem:d sector allowea the requisite revenues to be collected. The Republic is a classic export economy. During most of the period under study exports accounted for approximately 25 per cent of GNP., while, as previously statue, exports and imports combined averaged up to 40 per cent of grass output. Import-export related tax revenue fluctuated between 40 and 50 Per cent of tax collections. Actually, two schedules of import duties were extant, e customs duty proper and internally collected taxes on the sale, use, and consumption of the same goods. !~'ram 1945 to 1948 imports grew munch faster than eYp'irts.. although a favourable trade balance was maintained. World War U had left the nation with shortages of both capital and consumer goods; pent-up demand ton imparts and large foreign exchange reserves accumulated as marl, t prices for liornnucan primary exports a result of exceptionally high permitted imports as a percentage of GN1' to .rise front S per rent in 1945 to 23 per cent in 394'i; K irnpo t export related tax revenue increased 1mm 5 to Ill per, cent of GNI' between 1945 and 1947., Moreover, r;ee effective tax rate ml. vent to imports (calculated by dividing imrxut tax revenue by total import value) did not, at this time, rise significantly. r6

revenue

sax

import

consumption

sates

;ax

the

raising

effective

proportion

Aitl;oug1t upward displacement attained. its peak in 1948 the subsequent decline placed he economy at a h.igem pntean than prior to 1942 (see Table 1, column 7). Esports and expo,? prices seld not continue to increase at a fasier rate than national product, so that the continued relatively high 1evci of public spending was maintained by of imports forre from 33 per cent in 1948 to 57 pem cent. in 1950. Throughout time entire decade of time. 1.9`t1c to effeeuve import tat rate varied from 35 to 57 per cent, averaging 44 per cent. Furtlterrnors, the effective tax rates not related to the external

StN.'1.á1L AND EcOrtoMtc ;yT0DiES

sector (principally indirect taxes) averaged 12 per cent of GNP after 1948, having risen front 6 to 12 percent of GNP in the 1942-.948 period, and reached a maximum of 16 per cent in 1953 when the interord debt was liquidated. Since it is in the nature of such eonsumptien taxes to be regressire, the tax burden on the i)omínican was heavy, due in part to an extremely tineyual income distribution. Thus, Dominican displacement was partaaily financed by a rather high tax burdtn on the Dominican populace. it may well have exceeded the bounds of tolerability, and certainly cit±tans lacked the privilege co articulate their views. This, of course, is in direct contrast to what hsa been conohrded in other studies pertinent to more developed and/cr more ocratic countries. Peacock and Wiseman (21 p. 261 state, for example, rh. t no govertrncat, dictatorship or democracy, "is likely to act without any considerstion at all cr tite iCWS of its cit.tzens." A tp ism indeed, but even a superficial knowledge of the Dominican Republic under Trujillo leads one to conclude that little, if any, heed was paid to citizens' desires. The Puerto Rican functional data olTable 5 fit well into their historical context. The decade of the 1930s was an era of economic stagnation, high unemployment, and a basically passive insular government. Real per capita gross output remained, for all practical purposes, almost constant throughout the epode its low point being reached in the heart of the depression in .993. Public measures to stimulate the economy. were, of course, contrary to prevailing phiiesophy, atthuugh the u.s. government did fund emergency hurricane relief after the hurricanes of 1928 and 1932. As Table 2 demonstrates, until 1939 central government expenditure never comprised more than t0 per cent of gross output spending (1932 a slight exccption). r ~ The late 19_í0s did witness the rapid growth of public disbursements on social and economic development, the latter primarily involving the conrtmction of roads and hydroelectric plants. At this time well-documented political changes were occurring, changes which would in the coming decade(s) significantly alter the structure of the economy and the pattern of public expen(litures. The formation of the reform-minded Popular Democratk Party iii t'~38 under the leadership of Luis Munoz Marin, its election victories in 1940, and i')44 anti 1948. ;snd the 1941 Roosevelt appointbent n: liberal Rexford G. Tugwe;l as gti,emor combined to forge a pacific socio-economic revc,lutkni. Charge was retarded in the early 194-Os due to the constraints imposed by wartime, but accelerated rbn)ptly as Cite War Care to an end. it is precisely ;his historic perod whit witnessed the Puerto Rican upward displacement effect (see Table 2, úoiv,imti 7j. 7 steles ; and 6 reveal the emphasis placed upon publicly-protn.oted ec ,nocriir deveiopriaent in th'* 1940s. While at the same time social development and education were not toeing slighted. In the latter years of the 1940s and into the 195Os a swite'x in direct emphai was erfectuated in public expenditure policy, as she promotion o social development took precedence over that of economic development. The pr'vace seCt.or, th rough tax exemption offered to newly esrabiished firms regardless of the origin (resident or non-resident) of the capital, was given primary responsibility for economic development, although the central fsc provided gra:rt atitl to public. corporations (amú to its own dependencies) to develop the economy's infrastruetc_e. Of interest is the relative increase in the late

PUALIC t~XPI:NBI7URE PATTERNS

1960s in o:u(ays on economic development, when road and highway conxtruction rece"ved renewed priority. Returning in the 1940s, promotion of social attcl economic development was fashioned along twin channek, separate but closely linked. The central government itself established the c.verall refertn-oriented policies with its expenditure patterns; furthermore, it legislated the est3bfrshmernt of an eventual myriad of semi-autonomous public corporations which also served t carry out'policy objectives. The data in Table S do not reflect the aggregate impor of the public corporations, the most important of watch became largely sell=tinancing. Only central government grants to these bodies are incprporaFed in the data, and as such understate the relative growth of the public sector. It is postulated that many of the functions undertaken by the public corporations might just as well have been carried out by the central government itself; however, operating under the politi.cal•philosophicai constraint of a `free enterprise' stem, the explicit presence of the public sector was reduced in magnitude.18 Table . -3, the 'consolidated' public sector. permits one to gauge the 'real' proportional growth or the Puerto Rican public sector between 1944 and 1950, and it becomes q.,parent that the Puerto Rican upward displacement effect was still not quite as significant aS was that of the Dotnimean Republic. What :: e today The most important public corporations twith the exception of the Housing and Hignwrtys Authorities) all began operations in the fiscal years 1942 ! 945. ~9 Through 30 June 1945, actual insular goverurnei,t grants (as opposed to appropriations) to them semi-public entities amounted to 22.ti million; through 30 June 1947. giants totaled .S6t.7 millíon.. whereas general fund authorizations summed +o $l2' million as of 31 December 1947; through 30 Jurte 1950, insular grant expenditure an public corporations had aggregated $12& million. Of this latter total the Lard Aohrrity, the Industrial Development Company, and the Development Bank had received almost exactly 50 per clot, Thus, part of the cause of the Puerto Rican ttísplace.rnent i. identified and quantified, as a large portion of the economic and socísl development expenditure otercase is cute to these grants. Further accentuating the relative growth of the public sector during the mid-1940a arc the expenditures of the War t'tnergency= Programme en social nd economic development. Spending on this temporary pcograrnrne was a mere S1.F million on fiscal 1943, but had risen to $16.2 million by 1945. A good part of the capital expenditure increase in 1945-1946 was due to the construction and mainten,noe of roads fmanced through this programme ,sec Table . ). The goventrnent was able to partially finance the displacement through a forKuitous sct of circumstances. Attitottgh the domestic tax burden did rise, such effectwe rate increases cannot entirely explain the relativa expansion of public sector dishurnencuts. World War 11 had cut off European eAports of alcoholic beverages to the United States, and American domestic pry:duction derived from cereals had also deereased. ittc Puerto Rican rum export industry was thereby stimulated, and by 1944 rum distrllation and bottling had become the island's single most important manuiactsrn:g industry. Federal (U.S.) excise tax collections on the sale of Puerto Ricat runts within the U.S. r::vertcd to the alar treasurr/, so that the years 1941 to

SOCIAL AND ECONOMIC STUDIES

62 TABLE 7

CENTRAL GOVERNMENT CURRENT AND CAPITAL EXPENDITURE, DOMINICAN REPVI1L1C AND PUERTO RICO, 1928.1970

(millions of current pesos/ dollars)

Current' lIP.. P.R.

Year 1928 1929 1930 ¡931 1932 1933 1934 1935 1936 393'1 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 ¡952 1953 ¡954 1955 lySti 1957 1958 1939 tr!60 1961 ¡962 1963 ¡964 1965 1906 1967 1;168 ¡969 1970

pet

9.1 8.8 9.7 9.7 13.6 10.2 12.1. IS.S 20.9 22.7 31.8 523 56.3

14.7 15.5 17.3 18.2 18.8 16.5 20.0 18.4 19.3 20.4 21.9 25.3 24.:i 25.1 36.: 48,9 60,0 59.7 89.7 1i6.4 84.4

49.3

,16.a

65.0 72.3 78.3 89.9 55.6 66.2 80.9 94.7

97.7. 105.1 120.2 1.35.7 144.1 163.0 164.5 193.? 204.2 210.5 241.0 272.7 285.7 331.1 373.3 412.3 4'14.'1 562.4 ti22.: 7001 8.32.:;

---.. --

¡e2.a 107.1 108.7 106.6 160.1 166.3 183.8 13$.7 167.3 1630 i63.' 171.& 183. >

Current s % of Total D.R. 1'.I1

85.t. 83.8 83.6 84.3 88.3 115.0

852

80.7 79.:; 76.2 74.1 11,4 65.2 66.1 11.7 70.7 60.7 69.2 51.5 52.0 559 61.9 62.0 69.1 73.6

76.2 11E•.7 e7.x 66.5 93.8 88.7 h2.1 78.t+ 73.0 7ri.0

Capítula D.R. P.R.

78.6 E1,4 117.8 89.'7 91.3 93,7 90.1 89.3 'S9.2 89.5 85.9 85.6 9;1,8 '19.7

---

1.5 1.7 1.9 1.8 1.4 1.8

4.0 3.3 2.4 2.1 1.8 1.1 2.2 7.2 2.3 2.4 .3.6 4.:: 4.7 6.4

al.•r

2.1

St2.o 92.9 112.0 85 3 R4.S 8,.n

3.7 5.5 7.1 11.1 20.9 3 .l

3".9 4.6 1'3. 15.4 1".9 ¡JO 19.5 19.:+ lv.ti 21.0 25.9 77.2 26.0 243 31.6 29,1 3: e 33.1 3`.•.0 33. 44.4 31.5 67.5 918 123.0 1'w;.1 134,1 14'34

:;.1.1

250

83.4 84.3 11.5.1 940 841 86._ 87.^ 86.3 87.3 8.9 a7.9 fAs.6 $34 56.2 87' 115.9 528 R:.G 837 112.9 RS ':

25.7 29.9 54.` 40.0 52.3 61.0 fi.3.? 5ii._t ú2.5 •47.6 38.6 32.4 24.6 23.1 28.7 >.0 21.4 i5.3 43.6 631, 1'4

Capital as % of TOtsl D.R. P.R.

a 14.'_ lti.'2 16.4 15.7 ¡1.') 15.r1 13.8 19.5 25.8 23.6 259 2k,ú 34.8

33.6 26,3 29.3 39.'3 30.8 48.5 48.0 44.1 381 380 3119 2d2 2::+.? ii 3 ¡2.2 ¡:4.5 62 ¡3.? 17.-, .t.: 27.0 29.4

Total D.R. P.R.

21.4

17.6 12.2 10.3 8.7 6.3 9.9 10.7 10.8

104

18.7 I8.8 19.7 20.3 20.6 17.6 22.2 10:6 21.3

103 14.1 14.2'

205 11.6 11.5

22.8 25.5 2).3

¡62 20.3 18.3

12.0 12,0 14.7

29.0 31.5

!9.2

44.2 52.8

26,4

64.b

7.4

7.¡ 1'8.0 14.1 15.5 18.4 169 16.6 15.7 14.9 ¡6.0 15.9 13.b 13.0 14.0 12.5 13.! 12.1 11.4

29.8

72.8

42.9 7.3.2 86.4 74.3 90.7 ¡02;2 119.1 1.*?.9 107.9 127_2 144.6 153.0 164.5 154.9 147.3

105.1 302.3 103.4 115.5 116.5 124.7 141.2 161.6 171.3 189.0 ¡89.0

10.6

139,0 184.7

11.8 12.1 14.1 16.2 18.1, 16.3 16.1 14.7

189.5 212.5 ¡45.7 188.7 198.0 24Y).7 235,4 254.6

225.0 233.3 242.2 274.1

3"07.7 314.6 375.5 425_2 479.8 566.0

686.0 743.6 834.3 '1"S 7

alneludes rattt,Vs cat geti,Js s.nd wvi~es anil ,ranafnri an<1 euhsld:cs,,s n tnalt4 _„rsih,.t1 fq;ure, sincr all ex• us' not ar nved u, ir• ,•a}tira! di:bur.sar•',an'. wete presvn,rd M i,as,+ bren nade ..n euue,n acetwnt. ylnd„úes ,r.:,iay•. ioc du:abw asxte; r-y., raºda, lat:ildinh, m.^•.:J,exp. S~tar:-,

Appen4itc 11, f>r5lrr. Exprmtir:+n Ll;rm.

'UI3I,1C EXPENDITURE PATTERNS

1948 testified to the fovowir ' -fir .tore excise tax revenues flowing into island public coffers (in millions of dollars, fiscal ycat 1941-1948): 4, 14, 14, 64, 37, 33, 1.8 and 3. The government was able to conserve part of these exported tax balances from the abnormally high yield years of 1944-1946 to equip the economy with social and economic development facilities once wartime shortages had been overcome. In fact, Puerto Rico's current account on balance of payments shifted from a surplus in 1946 (and during the war years) to a deficit thereafter. Central government expenditures as a-percentage of GNP peaked in 1946, and leveled off at 15.16 per cent in the late 1940s and during the 1950x.. After the sudden drop in off-shore excise receipts subsequent to 1947 the government returned to more steadily recurrent revenue sources to maintain its share of gross domestic expenditure. Economic expansion permitted income tax revenues to íncfease at a faster rate than national income, so that only a very slight increase in domestic effective tax rates was required to preserve expenditures close to previously attained levels. Moreover, grants from the federal government continued to provide around 10 per cent of total revenue, and bond íssues were increasingly emitted (particularly after 1960) to :finance capital spending. Thus, the essentially separate discussions of the Dominican and Puerto Rican functional expenditure categories do offer support for the postulate that there is a probable connection between displacement in small, open economies and social disturbances in their larger commercial partners, in each country case events exogenous to these two dependent economies reverberated to the benefit of their public treasuries, consequently permitting increased (relatively and absolutely) public expenditure Ievels? t Here the similarity ends, however, for Table 6 dramatically reveals the different uses to which these revenues were put. The percentages in 'fable 6 are further significant, for a much-sought goal in both countries is that of economic growth (defined here as a sustained rise in real per capita income).zr The main contrasts noted in this table are those between expenditures on the protection of persons and property, education, and social development; whei as the percentage distribution of e;.:tlays on the fo:mer is consistently higher in the Dominican Republic due to the maaJntenance of a military establishment, the proportions allocated to he latter two areas are consistently higher in Puerto Rico (where absolute expenditures are also greater). In terms of these fac rs atone it is not surprising that Puerto Rico's real per capita GNP Ievcl rosc nearly 300 per cent over the 1940-1970 period, whi`G that of the Republic increased only a bit over 50 per cent. Krueger 1161 estimated that educatior)al levels accounted for one-third to one-half of the differences in per capita income levels between the United States and a sample of 24 less d : L lr>oed. cc':n)tries. Of course, ii is nserly simplistic to suggest that public expenditure allocation is even the principal determinant of the observed increasing relative gap between per capita gross incom e levels in the two countries under study, for a host ad factors corr:e into play (e.g., political sability ).23 Nevertheless, Table 4-6 are at icast partially indicative of the policyunplications regarding public expenditure aliocatiun between functions.

SOCIAL AND ECONOMIC STUDIES

CURRENT AND CAPITAL EXPENDITURE CATEGORIES Table 7 breaks down public expenditures into current and capital (asset-creating) account, and from first glance it is evident that relative expenditure on capital account in the Republic was amazingly high in comparison to that in Puerto Rico, comprising well over one-third of total spending from 1948 to 1960. Since capital grants to the Puerto Rican public corporations are included in this table under current expenditure due to data deficiencies prior to 1946, the figures are really understatements of the central government's contribution to capital formation after 1941. But even if this were accounted for the Commonwealth proportions would generally be lower than those given by the Dominican data over the 1943-1961 span. This is not surprising, however, given the greater reliance placed by the former's public policy on the role of private capital in economic development. Furthennore, Puerto Rico's relatively greater access to and attraction for private capital diminishes its need to seek capital growth through public sector mechanisms. It is observed that Dominican expenditure on capital account grew at a faster rate than that on current account and on gross output during the 194?-1948 interval, thereby accounting for another portion of the upward displacement (debt amortization payments, which conceputally might be classified as capital account expenditure, are not here defined as such). Again, it is necessary to speculate on the grey boundary separating public expenditure for public welfare from public spending for private welfare. Curiously, and perhaps not coincidentally, many of the major Trujillo family investments coincided with capital expenditure peaks. The 1946-1948 period, when capital expenditure as a proportion of GNP doubted, witnessed the construction of a cement plant which later became the hub of the Trujillo-controlled Ozama Construction Company. As public capital disbursements remained at high levels in the early 195Ús work was completed on the giant sugar mill at Rio Haina, which became the centre of the dictator's twelve-mill complex. On the other hand, the years 1954-1956 saw public capital outlays reach almost 50 per cent of total disbursements, and two major investments which ultimately benefitted the country can be linked to this peak. The first was the construction of a superhighway connecting Santo Domingo with the interior city of Santiago, located in the heart of the Republic's richest agricultural area; the second was the International Trade Fair of 1955-1956. Also of note is the drastic fall in both absolute and relative capital outlays during the tumultuous years of 1961-1965 and their recovery under t%a relatively stable Balaguer regir , after I %6. CONCLUSION In the final analysis there are probably more similarities than dissimilarities to be found in Dominion and Puerto Rican public expenditure patternss over the past four decades. The dissimilarities revolve around the political decision-making mechanism and the overall socio-political framework, the pace of economic growth and structural change, and the public resource-consuming xolc of the armed forces. Undoubtedly the average Puerto Rican had (and still has) more voice in orchestrating expanded public intervention in the national economy than his Dominican counterpart. Between 1940

PUt3T.iC FXPENDtTGRI: t'ATTERNS

public

and 1970 real per capita in.otne tev:k in the Commonwealth quadrupled, while in the Republic they rose by approximately 50 per cent. Assuming income elasticity of goods aid services to be greater than one, demand coefficients far most such real income rises served to maintain (and augment) relative public sector influence in domestic economic of€airs_ An important factor which continues to affcct Dominican spending patterns are the expenditures of the military, an interest group whose desires must L satisfied by any post-Trujillo politician who values his political life. The rather preeminent position of the armed forces is not principally due, of course, to external threats to he nation's welfare, but to the need to maintain internal equilibrium.

Rico

In addition to the fact that upward dispiace.neht occured in both countries during the 1940s, other analogies are apparent. Abnormal events wl'ich enveloped the developed nations iu war in oie 194us retierherated to the ultimate benefit of the and the I)omiñican Republic, generating open, depent 'nt economies of Puerto the revenues required to iinanee increased mending, and initiating changes in public attitudes toward the public sector. It is this latter phenomenoni the demonstration effect, which is most likely responsible for the longer nun perni ent variations evidenced in the data presented above. Not cs~%y does the demonstration effect operate to transform private sector consumntion habits, t• ' also alters public sector consumption forms. The drive toward economic development, an oft-expressed and fervently sought goal of each government, stimulates public expenditure growth from both the demand and supply sides; i.e., economic growth requires social infrastructure creation, while at the same time growth nay serve to further enhance demand for publicly provided goods and services. Certainly it may be demonstrated from the data of Tables 4 and 5 that expenditures on social and economic development amid on education have increased at a more rapid rate than those on the more traditional areas (administration, protection of persons and property), in the Puerto Rican case i °ginning in the early 1940s and in the Domi:,ican case since the early l 9€;0s.

Although Puerto Rico's disl;a.ein nt effect oeeurred during a period of socio— economic disturbance, and was accelerated by exogenous factors, it may be labeled, in quasi•Marxian fation, quite inevitable. Value judgement as to J e duties of the state had changed, and, judging by the electoral mandates given ti.e Popular Democratic Party, it is nigh improbable to assume that the stpulace desired a continuation of pre-]ái40 policies. The marriage of good governine.at v.ith rational economic policy inexorably led to a larger relative roL nor the St«te in the island's affairs, a necessity dictated, so to speak, b the dyaarnics of the system. Adopted were essentially socialist measures, but always within the framework of mixed, modern capitalism. Although there are obvious historical contrasts, the sar e phenomenon was work in the Dominican Republic, the data of he post-Trujillo years support this statement. It is interesting to note thst even wher the relative poverty of both countries is taken into account, their displacement effects are similar to those of richer nations. In other words, displacement is not a function of real income levels but of prevailing attitudes toward the public sector, attitudes which are influenced by improved inter and intra— country communication, th e "demonstration effect of a F':ence and welfare measures

SOCIAL AND ECONOMIC STUDIES

Naturally, a host of other social abroad, and the conflict of political ideologies" and institutional factors interact to mould the citizenry's ideas at any one point in time. And finally, another possible key factor affecting the rising demand for public spending is the internal migration from rural to urban areas, the pace of which has picked up in the past decade in both countries (see Table A-2). Not only does the urban resident appear to require (and expect) more public services per capita than does his rural brother, but hand in hand with urbanization occur other social changes; e.g., less reliance on the extended family for social needs, thereby transferring burdens to society as a whole. In a sense an ideological evolution/revolution regarding the state's functions has converted what was formerly individual responsibility into collective accountability. Whether or not past trends can be successfully extrapolated to the future is mere conjecture.

FOOTNOTES tme author wishes to acknowledge a debt to Fuat and Suphan Andic, Richard Bird, David Morgan, and an anonymous reader. 2See, for example, Musgrave and Culbertson [191, Peacock and Wiseman (23), Arr}leand. Veverka [31, Goflman and Mahar [111, and Bird [51. 3The Goffman-Mahar article [11J does contain aggregate data alone for the Dominican Republic covering the period 1940-1963; i.e., no disc grcgations are presented. Moreover, their income elasticity coefficients are based on GNP data which were substantially underestimated during 1950.1963. For further comment on this point see Mann 1171. 41t is recognized that the special nature of the political relationship between Puerto Rico and the United States further influenced transmittals. 5F. Andic (2) does present a `rough' estimate of Puerto Rican national income for the years 1920.1929. To use his data here would require calculation of a GNP value for each year on the basis of a known relationship between GNP and national income. Moreover, it is doubtful, one priori grounds, that significant relative changes in public expenditures occured during this decade. 0See Wagner [291. For further expositions of Wagner's ideas see Bird [61, Peacock and Wiseman [23, pp. 16-21 J, and Musgrave and Peacock [20, pp. 1-8(. Throughout the entúe period under study the Dominican peso remained at a one to one par with the U.S. dollar, which circulates in Puerto Rico. 1n fact, before 1947 the dollar was used as the Dominican national currency. Therefore, no exchange rate computations are required in order to make direct comparisons between amounts stated in either Dominican pesos (RD$) or dollars (US$). Additionally, the Dominican data refer to the natural year (January 1-December 31), whereas the Puerto Rican data refer to the fiscal year (July 1-June 30). These data do not include the direct expenditures of U.S. government agencies in Puerto Rico. They do include, however, those expenditures made by the Puerto Rican central government rmanced with federal grants. 9An economic category disaggregation normally involves two distinct expenditure breakdowns, that between outlays on goods and services as opposed to transfers, and that on current as

PUBLIC EXPENDITURE PATTEt'NS

opposed to capital account. Given numerous data source problems reliable figures for the former classification could not be produced for Puerto Rico prior to 1946 nor for the Dominican Republic prior to 1967. 10Table A-3 presents the Dominican displacement over the years 1940 19S5 eliminating debt principal outlays. It is observed that this exercise does not change the pattern of displacement, although it does reduce its magnitude. 11

For further amplification of these points gee Rodman [27J, Bosch [71, and Ornes [221.

Total Dominican external indebtedness at this time amounted to over 40 million dollars, perhaps surpassing the value of national output. Refer to [28 p. 3531. 13 "Government finances were good enough to permit President Trujillo to retire $31,000;000 worth of internal bonds. As a result, the country 's now one of the few in the world to be entirely free of debt." See [211. 14Of course, what really is needed lure is a theory of Trujillo himself. 15 Ovea this period the four principal Dominican exports, sugar, cacao, coffee, and tobaecn, averaged about 80 per cent of total exports. The price of sugar between 1942 and 1948 rose from $53 to S l l 0 per metric ton, and expos t volume doubled; the price of cacao increased from $146 to $661, that of coffee from $182 to $485, and that of tobacco from $95 to $292. Data taken from [ 10, pp. 146-158 J . 16 The effective rate concept applies to total imports; i.e., it reveals nothing as to import mix. During this period duty-free capital imports rose from 20 to 34 per cent of total imports, so that the rates on eonsunption goods may well have risen. Tax rates on a number of prime necessities (salt, shoes, vegetable lard) were quite high_ 17 To reiterate, direct federal agency spending is excluded. Commencing in 1934 the federally-funded Puerto Rico Emergency Relief Administration (later renamed the Puerto Rico Reconstruction. Administration) channeled funds toward rehabilitation and relief programmes, and is estimated to have spent $72 million by the time funds were terminated in 1941. Perloff [24,p. 32J estimates that total federal agency contributie ns in relief and loans between 1933 and 1941 amounted to $230 million. These disbursements certainly served to prop up income levels, and, as Hanson [l5J points out, "pioneered in the development of economic and social programs for the PUP as well as in training and condition.n$ men and women to carry them out".

18Úf course, the partial divorce of economics from politics may favour economic efficiency in the semi-autonornuus agencies; i.e., there may very wed exist a solid economic raison d'érre favouring establishment of bodies ]biked to but apart from he central government apparatus. Further speculation as to reasons for such extensive use of the public corporation might lead one to look at the public der: limit. hi the 1940s he ec ial government's debt-incurring capacity amounted to 10 per cep: of .he acrer,-ed va.l,ie. of taxable proper; ty. Perhal .t to avoid this restriction recourse was made to the etiabisshmen't of public corporations which did not encounter this legal barrier to tebt emissions. 19A partial listing of carne -r'id purpose t the more important .: es: Land Authority— acquisition and redistribution of land: lndustr d Devefopmen'. Company (Fomento) and Development Bank-Industrial development and cre lit extension; Water Resources Authority-electric power production; Aqueduct and Sewer Ai_thotity-sewage disposal and water supply; Agricultural Company-agricultural development.

This series of data tc.lien from [26, p. 91, [12, Exhibit #33J. [24, p. 3801, and [13, Table #331. 2t Admíttedly, what has been establis:.ed is more of a pcnnissh'e relationship than a functional one.

SOCIAL AND ECONOMIC STUDIES

22 T his statement may not be totally valid as applied to irusilio's Dominican Republic. 23 For a thorough analysis of the relationship between per capita income levels and numerous social, economic, and political indicators, refer to Adelman and Morris [1]. Musgrave (18, p. 72] goes on to state: It follows that the relationships between per capita income and fiscal structure that are obtained from the historical view may well differ from those revealed by the cross-section appro ich. Low income countries today do not operate under the same technical, political, and value conditions as prevailed in the past when now developed countries were at similar low levels of income.

REFERENCES [1]

ADELMAN, Irma and Cynthia T. MORRIS, Society, Politics, and Economic Development: A Quantitative Approach. Baltimore: The Johns Hopkins Press, 1967.

[2]

ANDIC, Fuat, "Research Note: National Income of Puerto Rico, 1920-1929," Caribbean Studies V.2 (1962), pp. 34-39.

[3]

ANDIC, Suphan and Jindr .ch VEVERKA, "The Growth of Government Expenditure in Germany since the Unification," Finanzarchiv, V.23, 1 (1963), pp. 169-277.

[4]

BIRD, Richard M., "The `Displacement Effect': A Critical Note," Finanzarchfv, V. 30, 3 (1972), pp. 454-463.

[5j

, The Growth of Goverment Spending in Canada Toronto: Canadian Tax Foundation, 1970.

[6]

__ ._ , "Wagner's Law' of Expanding State Activity," Public Finance/Finances Publiques, V. 26, 1(1971), pp. 1-26.

[7]

BOSCH, Juan, Trujillo: Causas de una ttrania sin ejemplo. Caracas: Grabados Nacionales, 1959.

[8]

BROOKINGS INSTITUTION, Refugee Settlement in the Dominican Republic. Washington: The Brookings Institution, 1942.

[9]

CLARK, Victor, et al, Porto Rico and Its Problems. Washington: The Brookings Institution, 1330.

[10]

fJIRECCION GENERAL DE ESI'ADISTICA 21 Años de Estadkticas Domirticanas. Ciudad Trujillo: 1957.

[ 11 ]

GOFFMAN, Irving J. and Dennis J. MANAR, "The Growth of Public Expenditures in Selected Developing Nations: Six Caribbean Countries, 1940.65," Public Finance/Finances Publiques, V. 26, 1 (1971), pp. 57-74.

[12]

GOVERNMENT OF PUERTO RICO, Annual Report of the Auditor, 1947. San Juan.

PUBLIC EXPENDITURE PATTERNS [13]

_ ,Annual Report of the A. d tor, 1950. San Juan. GUPTA, S.P., "Public Expenditures and Economic Growth", Public Financ% Finances Publiques, V. 22,4, (1967), pp. 433-443.

[15]

HANSON . Earl Parker, "James Bourne and Puerto Rico", San Juan Star, March 22, [973, p. 32.

[16]

KRUEG:ER, Anne O., "Frc.torEndowments and Per Capita Income Differences Among Countries " TonomicJournal, V. 73 (1968), pp. 641-659.

[17]

MANN, Arthur J., "The Growth of Public Expenditures in Selected Developing Nations: Six Caribbean Countries, 1940.65--A Comment," Public Finance/Finances V. 27, 3 (1972), pp. 368-369.

[18]

MUSGRAVE, Rich:mod A., Fisc-J Systems. New Havan: Yale University Press, 1969.

(191

, and J.M. CULB VI SON, "The Gro'!th of " ublic Expenditures in the United States, 1890-1948." National 7'm Jc<irnal, V. 6, 2 (1953), pp. 97.115.

[20)

and Alan T. PEACOCK. Ckssics in the The,':-,' of Public Finance. London: Macmillan and Co., 1958.

Puvhque:,,

[14]

[21 J

NEW YORK TIMES, January 6, 1954, p. 47.

(22]

ORNES. German .E., Trujillo: Paqueño César del Caribe. Caracas: Editorial Las Novedacie, 1958.

[23 J

PEACOCK, Alan T. and Jack WISEMAN, The Growth of Public Expenditure in the United Kingdom. Princeton: Princeton University Press, 1961.

[24]

PERLOFF, Harvey S., Puerto Rico .r Economic Future. Chicago: University of Chicago Press, 1950.

[25]

PRYOR, Frederic L., PubtiE: L'zpenditures in communist and Capitalist Nations. Homewood: Richard D. Irwin, 1968.

(26)

PUBLIC ADMINJSTPATIGN SERVICE, Summary Report on the Accounting and Reporting Practices of the Public Service Enterprises of the Govern-

ment of Puerto Rico. Chicago: 1946. [27J

RODMAN, Selden, Quisqueya.Seattle: Univérsity of Washington Press, 1964.

[28]

SCHOENRICII, Gro, 590010 Domingo: A Country with a Future. New York The Macmillan Co., 1918.

[29J

WAGNER, Adolph, Finanawissensc;taft. Leipzig: 1890 (3rd edition).

SOCIAL AND ECONOMIC STUDIES APPENDIX A

TABLE A-i

TOTAL POPULATION OF THE DOMINICAN REPUBLIC AND PUERTO RICO, 1930-I970 (in thousands)

Year

Dominican Republica

1930 1935 1940 1945 1950 1955 1960 1965 1970

Puerto Ricob

1,256 1,484 1,674 1,889 2,136 2,554 3.047 3,513 4,006

1,535 1,695 1,862 2,031 2,202 2,232 2,340 2,562 2,711

a1 July of each year, excluding census years of 1935, 1950 and 1970. b1 January of each year. Sources: Dominican Republic — Dirección General de Estadistica, Anuario Estaailatieo de la República Dominicana, 1951 (Ciudad Trujillo: Imprenta San Francisco, 1953), p. 445, Oficina Nacional de Estadística, República Dominicana en Cifras, 1971 (Santo Domingo: 1971), p. 21; Puerto Rico — Puerto Rico Planning Board, Bureau of Economics and Statistics, Statistical Yearbook, historical Statistics; U.S. Dept. of Commerce, Census of Population, 1930-70.

TABLE A-2 PERCENTAGE DISTRIBUTION OF RURAL AND URBAN POPULATION, DOMINICAN REPUBLIC; AND PUERTO RICO, SELECILD YEARS, 19204970

Year

Dominican Republic Rural Urban

Rural

1920 1935 1950 1960 1970

83.4 82.0 76.2 69.7 60.2

78.2 70.3 59.5 55.8 41.9

16.6 18.0 23.8 30.3 39.8

Sources. S^e sources to Table A-?.

Puerto Rico Urban 21.8 29.7 40.5 44.2 58.1

PUBLIC EXPENDITURE PATTERNS TABLE A-3

Year 1940 1941 1942 1943 1944 1945 1946 1947

DOMINICAN DISPLACEMENT EFFECT EXCLUSIVE OF DEBT PRINCIPAL PriYMENTSa Expenditure as `So of GNP 10.3 9.5 8.1 10.4 12.7 12.4 16.2 20.8

elncludes public debt interest payments. Sources: See Appendix B.

Year

Expenditure as % of GNP

1948 1949 1950 1951 1952 1953 1954 1955

20.5 16.1 16.1 15.6 19.4 16.2 20.1 22.7

SOCIAL AND ECONOMIC STUDIES TABLE A•4

(2)

(1)

Yeas 1930 1931 3 932 1933 1934 ¡935 1936 1937 I'J38 4939 3940 3941 1942 1943 1944 1945 (946 1947 1948 1949 1950 1951 1952 1953 3954 1955 i95tí 1957 1958 1959 1960 1963 1962 1963 1964 1965 (966 1967 1968 t 969 1970

Wholesale Price Index, Dom. Rep. (1962 = 100)

37 45 61 75 74 416 104 104 89 86 96 98 98 93 93 92 I0 101 95 96 't2 100 t 0>3 110 118 ( (: 115 123 i IS i 3x

PRICE INDEXES, 1930-1970 PRICE INDEXES, 1930.1970 (3)

Wholesale Price Index, U.S.

(1954=100)

Consumer Price Index, P.R. (1954 = 100)

53.4 46.9 40.6 38.6 41.5 45.6 47.3 49.1 48.5 45.8 45.8 48.8 54.7 59.5 63.0 61.7 66.7 79.1 90.9 92.2 91.6

54.4 68.7 70.9 72.6 73.9 76.1 88.7 92.0 87.6 81.3

(4)

(5)

Implicit Price Deflator. Govt. Consumption, P.R. (1954 = 100)

Implicit Price Deflator, Govt. Capital Exp.,

(it?

64.7

S5.0 87.6 89.2 89.8' 89.9 96.2 96.9 100.0 99.8 103.2 1046 107.5 108.1 115.8 119.0 142.7 126.6 129.3 131.6 135.0 139.3 141.0 143. 2 151.4

88.3 89.8 90.3 87.3 92.5 98.4 99.4 100.0 ]01.0 106.0 109.7 113.2 114.5 115.9 121.5 125.5 128.7 131.4 135.2 136.2 138.0 141.8 148.5 358.8

P.R. (1954 = 100)

Sources: Column (I) - work sheet provided by the Oficina National de Estadistica de la Krpúh4ica Dominicana. Santo Domingo, January (97^_; Column (2) - Economic Report of the President, 1966 (Wathingtun: U.S Government Printing Office, I966). p. 257. A two calerniar year average was taken and then converted to a 1954 base; (Column (3) '- Junta de Ptanificaciún de Puerto Rico, .4nusjio Estartisstko, EsrmJúiicas Nisorieas. 1959, pp. (47.3. Index Is for wage•earner3 femilies only, and base is converted to 1954; C'oluntns (4) and (5) ' Years 1940 and 1947.54 Junta de Planifieaci6n, /ngrrto p Prrxfucto, Piaerrn Rio's. 1940 a 1947-60, pp. A-242 to A-243; years (955-70: Junta do I9anificacion, Ingreso y Aattui n+. Put:m Rico, 1970. o6-7. Column (5) is a weighted average of expenditure on gross fixed domestic Investment and or rtecltmery and equiprncnt.

PUBLIC EXPEND1TUR1? PATTERNS

APPENDIX B NOTES AND SOURCES TO TABLES Dominican Republic GNP Data Current GNP, 1940.1946: Pan American Union, Fiscal Receipts, Expenditures, Budgets, and Public Debt of the Latin American Republics (Washington), p. 57. The data were given as national income; since during the 1950s national income averaged 80 per cent of GNP the latter was calculated dividing the national income data by 0.F. Current GNP, 1947.1949: Given the apparent absence of either GNP or national income data for these three years interpolation assuming a constant growth rate of 14.2 per cent annually was utilired. Current GNP, 1950.1959: Banco Central de la República Dominicana, Cuentas Nacionales de is Republica Dominican. Producto Bruto Nacional, 1950-1964 (Santo Domingo: 1966), Cuadro 1. Current GNP, 1960.1970: Banco Central de la República Dominicana, work sheets, 3 January 1973. Constant GNP, 1941-1949: The current GNPdata were deflated by the wholesale price index for Santo Domingo (base changed to 1962), the only alternative price indexes being a cost-of-living index and a retail price index, both for the capital city alone. Constant GNP, 1950-1959: Same source as current GNP for the years 1950-1959, Cuadro 19. Dominican constant price GNP is not estimated using implicit GNP price deflators. Rather, it is estimated extrapolating base year values by physical volume indexes for each sector. Thus, this explains the absence of an implicit gross product price deflator. Constant GNP, 1960.1970: Banco Central de la Republica Dominicana, work sheets, 3 January 1973. Constant Per Capita GNP, 1941-1970: The constant GNP figure for each year is divided by the total population figure for the corresponding year (see Table A-1). Dominican Republic Public Expenditure Data Current Government Expenditure, 1936-1970: Secretaria de Estado de Finanzas, Memoria, varying years, 1954-1970; the Brookings Institution, Refugee Settlement in the Dominican Republic (Washington: The Brookings Institution, 1942); Dirección General de Estadistica, Anuario Estadístico, years 19361954; Secretaria de Estado dei Tesoro y Crédito Publico, Memoria, years 1947-1953; Dirección General de Estadistica, Republica Domínicana: Album Estac(wico Crafico, 1944; Banco Central de la Republica Dominicana, Boletín Mensual, years 1955-1970; Secretaria de Estado del Tesoro y Credito Publico, Boletín

SOCIAL AND ECONOMIC STUDIES

Especial: Evolución Fiscal d:wrante elPeriodo 1930.1947; Dirección General de Estadistica, 21 Años de Estadísticas Dorninicanas (1956); Dirección General de Estadística y Censos, Finanzas Municipales, several years; Dirección General de Rentas Internas, Leyes y Reglamentos de Rentas Internas, (1944); Cesar A. Herrera, Las Finanzas de la Republica Dominicana (Ciudad Trujillo: 1955); International Monetary Find, international Financial Statistics, years 194fá-1970; Nicolas H. Rizik, Trujillo y la Estadística (Ciudad Trujillo: 1945); Secretaría de Estado de Finanzas, Esradisticas Fiscales de la República Dominicana, 1955-1961; Oficina Nacional de Planificación, work sheets; Oficina Nacional de Planificación, Evolución de la Econornia Dominicana en el Periodo 19541970; Oficina Nacional del Pn supuesto. Resumen de! Presupuesto, years 1971 and 1972; Oficina Nacional de Planificación, Plataforma pura el Desarrollo Económico y Social de !a República Dominicana (1968); varied annual reports of the Ministries of Education, Public Works, Health, Police, and the Armed Forces. Constant Government Expenditure, 1941-1970: Deflated by the wholesale price index for Santo Domingo (base changed to 1962). Constant Per Capita Government Expenditure, 1941-1970: The constant figure for each year is divided by the total population figure for the corresponding year. As far as possible all infra-governmental transactions have been eliminated. Sorne double-counting may have occurred, but was unavoidable given the 'complicated nature of the accounts and the many different sources employed. Most purely financial expenditure items (e.g., loans) were eliminated, with the exception of public debt amortization payments. These data deal only with the disbursements of the central government. All Dominican data in this and subsequent tables refer to the calendar year (January 1-December 31). aAdditioraai spending of 57 million, excluded here, was financed through irregular external channels (mos+ly USAID) due to the civii war. Puerto Rican CNP Data Current GNP, 1929: Dudley Smith, Puerto Rico's Income (Washington: 1943), p. 10. Current GNP, 1930-1939: Daniel Creamer, The Net Income of the Puerto Rican Economy, 19441944 (Rio Piedras: Social Science Research Center, 1948), p. 22. For each year a net nco ne figure is given. From the official Puerto Rican national accounts, which present national income estimates for 1940 and 1947 to the•present, it was calculated fated that in 1940 GNP was 1.275 times net income. Therefore, all the Creamer net income d2ta (and he Smith 1929 figure) were situp:y multipl;^d by 1.'75 to arrive at an admittedly crude GNP estimate for each over the 1929-1939 period. Current GNP, 1940: Junta de Planificación de Puerto Rico, Area de Planificación

PUBLIC EXPENDITURE PATTERNS

Económica y Social, ,ingreso y Producto, Puerto Rico, 1970 (San Juan: 1971), p. S. Current GNP, 1941 1946: Due to the unusual and unstable conditions prevailing in the Puerto Rican economy during these years, GNP data are not officially estimated. Reference was made to Puerto Rico Planning Board, Economic Division, Economic Development oaf Puerto Rico, 1940.1950, 1951-1960 (San Juan: 1951), p. 162. By following the rates of increase of gross insular product for 1940-1946 contained in this publication (the absolute data are unusable since they represent large overestimates) a 15 per cent constant growth rate for 1941-1944 was assumed; the subsequent growth rates for the years 1944-1945, 1.945.1946, and 1946-1947 were assumed to be 10 per rent, 2 per cent, and $.5 per cent respectively. Current GNP, 1947-1960: Junta de Planificación de Puerto Rico, Area de Planificación Económica y Social, Ingreso y Producto, Puerto Rico, 1970 (San Juan: 1971), pp. 3-9• Current GNP, 1961-1970: Junta de Planificación de Puerto Rico, División de Producción F conámica y Valor Añadido, revised and unpublished (to date) work sheet data. Constant GNP, 1930.1939: An initial attempt at deflating the current data was made utilizing separately a Puerto Rican wholesale price index for foodstuffs and the United States Bureau of Labour Statistics' overall index of wholesale prices for the continental U.S. Both indexes seriously overstated the real GNP figures (in terms of 1954 prices after their bases were changed) using 1940 as a point of reference. Therefore, an alternative procedure was followed, a methodólogy conspicuous for its lack of reliability. Employing the U.S. wholesale price index applicable to calendar years, a two-calendar year average covered by each fiscal year was computed and changed to a 1954 base year. The current GNP data were then deflated by this index, and the rate of annual change of these deflated figures was calculated. Each annual rate was assumed to reflect the rate of real GNP change using real GNP in 1930 as the reference point. 1930 real GNP was estimated by taking 1930 real GNP as a proportion of 1940 real GNP (both computed by the wholesale price index) and multiplying the resulting percentage by the more reliable published 1940 real GNP figure of $499 million; i.e., 451/627=0.719; 499 (0.719)-359. Constant GNP, 1940: See source to current GM, 1940. Constant GNP, 1941-1946: A methodology similar to that for the years 1930-1939 was followed; it is, however, probably more reliable. An index for deflating gross product for the years 1940.1946 is presented in Daniel and Henrietta Creamer, Gross Product of Puerto Rico, 1944-l94ó (Rio Piedras: Social Science Research Centre, 1949), p. 11; and in Daniel and Henrietta Creamer, Gross Product of Puerto Rico, 1940-1944 (Rio Piedras: Social Science Research Center, 1948), p. 71. Subtracting the annual rate of increase in the index from the annual rate of increase in current GNP yields an estimate of the

SOCIAL AND ECONOMIC STUDIES

annual rate of increase in real GNP. Applying these percentages to the real GNP of years subsequent to 1940, the 1941-1946 data are calculated. Constant GNP, 1947-1960: See source to current GNP, 1947-1960. Constant GNP, 1961-1970: See source to current GNP, 1961-1970. Constant Per Capita GNP, 1930-1970: Col=.unn 5 divided by the total population data contained in Table A-1. Puerto Rican Public Expenditure Data Current Government Expenditure, 1923-1950: ,4nnua1 Report of the Auditor, each fiscal year; Annual Report of the Governor, each fiscal year; Harvey S. Perloff, Puerto Rico's Economic Future (Chicano: University of Chicago Press, 1950), p. 379; Puerto Rico Planning Board. Economic Division, Economic Development of Puerto Rico, 1940-1950, ¡951-1960 (San Juan: 1951), p. 71 and pp. 169.173; S.L. Descartes, Basic Statistics on Puerto Rico (Washington: Office of Puerto 12ico. 1946); Departamento de Hacienda de Puerto Rico, Informe Anual, 1972 (San Juan: 1972), p. 26; Annual Report of the Treasurer, each fiscal year; Economic Development Administration and the Department of Agriculture and Commerce, Annual Book of Statistics of Puerto Rico Fiscal Year 1949.50, pp. 445-453; Junta de Planificación, División de Económica, Informe Eeondmico al Gobernador, 1950-1951 (San Juan: 1952); various annual reports of the Departments of Education, Agriculture and Commerce, and Interior. Current Government Expenditure, 1951-1970: Departmento de Hacienda, Servicio de Contaduria, Informe Anual, each fiscal year; work sheets and various reports, Servicio de Contaduría; work sheets, Junta de Planificación, Negociado de Planificación Económica, División de Producción Económica Y Valor A i ndido; annual reports, Highways Authority; Commonwealth of Puerto Rico, Department of the Treasury, Expenditures and Receipts for Fiscal Years Ending June 30, 1953 and 1954. Constant Government Expenditure, 1930-1939: Due to the absence of any usable insular price index for this period the current expenditure figures were deflated by the U.S. wholesale price index (base changed to 1.954). The rate of change of the deflated data from 1940 bark to 1930 was computed on a year to year basis; these rates were thén applied o the reliable 1940 real expenditure figure and to each subsequent year figure, thereby deriving the data appearing in column S. `Constant Government Expenditure, 1940 and 1947-1970: Expenditure measured in current prices was deflated by several implicit price deflators. Current account expenditure was deflated by the implicit deflator pertaining to the consumption expenditures of the Commonwealth central government; expenditure on capital account was deflated by my own weighted average of central government expenditure on gross fixed dome ttic investment and on machinery and

PUBLIC EXPENDITURE PATTERNS

Económica y Social, Ingreso y Producto, Puerto Rico, 1970 (San Juan: 1971), p. 8. Current GNP, 1941-1.946: Due to the unusual and unstable conditions prevailing in the Puerto Rican economy during these years, GNP data are not officially estimated. Reference was made to Puerto Rico Planning Board, Economic Division, Economic Development o, Puerto Rico, 1940-1950, ¡951-1960 (San Juan: 1951), p. 162. By following the rates of increase of gross insular product for 1940-1946 contained in this publication (the absolute data are unusable since they represent large overestimates) a 15 per cent constant growth rate for 1941-1944 was assumed; the subsequent growth rates for the years 19441945, 1945-1946, and 19464947 were assumed to be 10 per cent, 2 per cent, and 8.5 per cent respectively. Current GNP, 1947-1960: Junta de Planificación de Puerto Rico, Area de Planificación Económica y Social, Ingreso y Producto, Puerto Rico, 1970 (San Juan: 1971), pp. ,4-9. Current GNP, 1961-1970: Junta de Planificación de Puerto Rico, Division de Producción Económica y Valor Añadide, revised and unpublished (to date) work sheet data. Constant GNP, 1930-1939: An initial attempt at deflating the current data was made utilizing separately a Puerto Rican wholesale price index for foodstuffs and the United States Bureau of Labour Statistics' overall index of wholesale prices for the continental U.S. Both indexes seriously overstnted the real GNP figures (in terms of 1954 prices after their bases were changed) using 1940 as a point of reference. Therefore, an alternative procedure was followed, a methodology conspicuous for its lack of reliability. lrmploying the U.S. wholesale price index applicable to calendar years, a two-calendar year average covered by each fiscal year was computed and changed to a 1954 base year. The current GNP data were then deflated by this index, and the rate of annual change of these deflated figures was calculated. Each annual rate was assumed to reflect the rate of real GNP change using real GNP in 1930 as the reference point. 1930 real GNP was estimated by taking 1930 real GNP as a proportion of 1940 real GNP (both computed by the wholesale price index) and multiplying the resulting percentage by the more reliable published 1940 real. GNP figure of $499 million; i.e., 451/6270.719; 499(0 719)359. Constant GNP, 1940: See source to current GNP, 1940. Constant GNP, 194.1-1946: A methodology similar to that for the years 1930-1939 was followed; it is, however, probably more reliable. An index for deflating gross product for the years 1940.1946 is presented in Daniel and Henrietta Creamer, Gross Product of Puerto Rico, 1944-1946 (Rio Piedras: Social Science Research Centre, 1949), p. ii; and. in Daniel and Hcirrietta Creamer, Gross Product of Puerto Rico, 1940-1944 (Rio Piedras: Social Science Research Center, 1948), p. 71. Subtracting the annual rate of increase in the index from the annual rate of increase in current GNP yields an estimate of the

SOCIAL AND ECONOMIC STUDIES

annual rate of increase in real GNP. Applying these percentages to the real GNP of years subsequent to 1940, the 1941.1946 data are calculated. Constant GNP, 1947-1960: See source to current GNP, 1947-1960. Constant GNP, 1961-1970: See source to current GNP, 1961-1970. Constant Per Capita GNP, 1930-1970: Colma 5 divided by the total population data contained in Table A-1. Puerto Rican Public Expenditure Data Current Government Expenditure, 1928-1950: Annual Report of the Auditor, each fiscal year; Annual Report of the Governor, each fiscal year; Harvey S. Perloff, Puerto Rico's Economic Future (Chicago: Jniversity of Chicago Press, 1950), p. 379; .Puerto Rico Planning Board. Economic Division, Economic Development of Puerto Rico, 1940.1950 1951-1960 (San Juan: 1951), p. 71 and pp. 169-173; S.L. Descartes, Basic Statistics on Puerto Rico (Washington: Office of Puerto Rico, 1946); Departamento de Hacienda de Puerto Rico, Informe Anual, !972 (San Juan: 1972), p. 26; Annual Report of the Treasurer, each fiscal year; Economic Development Administration and the Department of Agriculture and Commerce, Annual Book of Statistics of Puerto Rico Fiscal Year 1949-50, pp. 445-453; Junta de Planificación, División de Económica, Informe Económico al Gobernador, 1950-1951 (San Juan: 1952); various annual reports of the Departments of Education, Agriculture and Conunerce, and Interior. Current Government Expenditure, 1951-1970: Departmento de Hacienda, Servn io de Contaduria, Informe Anual, each fiscal year; work sheets and various reports, Servicio de Contaduría; work sheets, Junta de Planificación, Negociado de Planificación Económica, División de Producción Económica Y Valor Añandido; annual reports, Highways Authority; Commonwealth of Puerto Rico, Department of the Treasury, Expenditures and Receipts for Fiscal Yews Ending June 30, 1953 and 1954. Constant Government Expenditure, 1930-1939: Due to the absence of any usable insular price index for this period the current expenditure figures were deflated by the U.S. wholesale price index (base changed to 1954). The rate of change of the deflated data from 1940 buck to 1930 was computed on a year to year basis; these rates were thén applied o the reliable 1940 real expenditure figure and to each subsequent year figure, thereby deriving the data appearing in column 5. Constant Government Expenditure, 1940 and 1947-1970: Expenditure measured in current prices was deflated by several implicit price deflators. Current account expenditure was deflated by the implicit deflator pertaining to the consumption expenditures of the Commonwealth central government; expenditure on capital account was deflated by my own weighted average of central government <'xpenditure on gross fixed domestic investment and on machinery and

PUBLIC EXPENDITURE PATTERNS

equipment. For the primary index data see Junta de Planificación, Ingreso y Producto, 1940 y 1947-1960, pp. A-242 to A-243; the indexes for the years 1955-1970 were derived from primary data found in Junta de Planificacidn, Ingreso y Produucto, 1970, pp. 6-67. Constant Government lixpenditure, 1941.1946: The annual percentages change in Creamer's government expenditures at constant prices was taken and applied to the 1940 and subsequent annual figure. For sources, see the already cited Creamer studies under the constant GNP data for 1941.1946. In addition, the current data were also deflated by the Puerto Rican consumer price index changed to a 1954 base; see Junta de Planificación, Anuario E'staHstico, Estadísticas Históricas; 1959 pp. 147-148. A simple arithmetic mean of these two series was computed, and it is the mean figure which appears in column 5. Constant Per Capita Expenditure, 1930-1970. Column 5 divíded by population. As in to case of the Dominican data an effort was made to eliminate intra-government transactions and purely financial expenditures (again, with the exception of public debt amortization payments). The data for the decade of the 1960s include the outlays of the Highways Authority, the University of Puerto Rico, and the Medical Center, which, although officially public corporations, must be included here for the sake of consistency; i.e., their functions had been undertaken directly by the central government prior to their creation in to 1960s, and therefore consistency demands their inclusion after such establishment. All Puerto Rican data in this and subsequent tables refer to the fiscal year 30-July 1).

APPENDIX C THE CONSOLIDATED' PUBLIC SECTOR The data presented in the text tables fail to demonstrate the total impact of the public sector in either country, for they 'basically include only the non-conunercial activities of the central fisc. Firstly the noncommercial functions of the municipal fiscs do not appear; secondly, the trading activities of the gamut of public corporations have been largely ignored, since they usually cover their service costs by levying axplicit charges. Although they are analogous to the private corporations in this sense, they nevertheless are subject to a greater degree of public control and do not pursue profits as their ultimate end. And in Puerto Rico after 1941 and in the Dominican Republic after 1961 they undertook a host of activities that might have been carried

SOCIAL AND ECONOMIC STUDIES

out by the private sector in an earlier historical period. Thirdly, the public service enterprises, the autonomous institutions, and the social security funds, although providing service without directly selling a good, nevertheless reflect public policy directives as they disburse semi-public monies, in the process affecting the economy's performance. Tables C—I and C--3 present data to crudely estimate the growth of the `consolidated' public sectors in each country. Transfers between areas have been counted as expenditures only, in those areas where they have actually been spent. In the Dominican case the relative growth of the `consolidated' sector closely parallels that of the central government, and reveals that the apparent lower plateau of central guvernrnent spending in the 1960s (Table 1, column 7) disappears, as the public corporations and autonomous institutions took up former functions of the central government as well as moving into new areas; e.g., the State Sugar Corporation—Corporacion Estatal Azucarera—administers the former Trujillo sugar properties. Most of the autonomous institutions and public corporations were created after 1961, but some have existed for years; e.g., the Agricultural Bank (I945), the Dominican Institute of Social Security (1948). However, data pertinent to their expenditures exclusive of central fist transfers could not be located, and are not therefore included in Table C —i prior to 1962. Table C-2 provides a functional breakdown of the Dominican `consolidated' public sector fo.r the year 1970 and may be compared with data for 1970 contained in Table 4. The Puerto Rican data in Table C--3 for the yeats I930 and 1940 do not include the capital outlays of the municipally-owned and operated water and light companies, although they were consolidated under public corporations wíth island-wide responsibilities in the 1940s; they do include, however, the capital disbursements of those entities which might be considered forerunners of the present public corporations and public service enterprises, e.g., the ls,ibela Irrigation Service, established in 1928. As can be readily observed, the faster rate of increase of municipal expenditures in comparison to that of central government spending led to a slight displacement effect between 1930 and 1940. However, in general the `consolidated' figures reflect a displacement very similar to that demonstrated by the central government data between 1940 and 1950.

for

It should be pointed out that these `consolidated' data do, in a sense, exaggerate the significance of the public sector i relation to national output, they comprehend certain disbursements on intermediate goods and au pure transfers; the GNP figure comprises spending on final output only; i.e., on payments to factors for current year contributions to national production. On the other hand, the exclusion of current public corporation outlays, many of which represent current factor income, understates the proportional public sector role. The above discussion is suggestive of an area which merits further and more penetrating analysis; i.e., the role the public corporations have played in the economic development of Puerto Rico anti the Dominican Republic (and other countries).

TABU C-1

EXPENDITURES OF CONSOLIDATED PUBLIC SECTOR, DOMINICAN REPUrILIC, SELECTED YEARS,

1936-197O

(millions of current F D$) Central Government

Municipal Governments

1936 1940: 1950 1954 ~►558 "1962 1954 166 1968 1970

10.3 11.7 ó>.4 104,7 160.4 171.1 183.3 166.1 174.3 2e3.9

0.9 0.9 5.0 6.6 16.2 21.9 24.7 19.6 18.9 20.6

Autonomous Institutions

16.4 33.8 37.8 47.1 55.1

Public Corporationsa

NA NA NA 76.3 70.0

Total

Totaj as Proportion of GNP

11.2 12.6 94.4 111.3 176.6 209.4 241.8 223.5 316.6 369.6

11.5 24.6 20.7 24.9 24.1 <":2.3 21.4 27.7 25.4

'( riy capital exile: diture included. NA — Data not available. Most autonorswus institutions and public corporations (organismos financieros) were established after 1961. From the standpoint of total expenditures the most important autonomous institutions (provide services without directly selling a good) ase the Autonomous University of Santo Domingo (UASD), the Dominican Social Security Institute (IDSS). the Dominican Agrarian Institute (DAD), the Dominican Corporation of State Enterprises (CORDS), the National Institute of Drinking Water and Sewerage (INAPA), and the National Institute of Hydraulic Resources (INRFl). The most important public corporations (cover most of own expenses and!or sell a good or service) are the State Susrar Corporation, the Agricultural Bank, the Dominican Electric Corporation, the National Housing Rank, and the Price Stabilization Institute. Source:

PUBLIC EXPENU[TURE PATTERNS

Ycat

Oficina Nacional de Planificación de la República Dominican, Departamento de Programación del Sector Público Consolidado, work sheets, January 4, 1973. —a

«o

CONSOLIDATED PUBLIC SECTOR EXPENDITURES BY FUNCTIONAL CATEGORY, DOMINICAN REPUBLIC, 1970 (millions of current RD$) Protection of Persons and Property

Debt Service

Education

Social Development

Economic Development

Total

Item

General Administration

Cwrent

41.0

47.0

8.4

34.5

45.1

3' 4

210.4

Capital

0.7

38.0

6.9

21.4

97.2

164.9

Total

41.7

47.7

46.4

41.4

66.5

131.6

375.3

Percentage Distribution of Total

11.1

12.7

12.4

11.0

17.7

35.1

100.G

Totai as Percent of GNP

2.9

3.3

3.2

2.8

4.6

9.1

25.8

Source:

See Table C-1

SUQlLLS Ji31IONOD3 QNV ZV¡OOS

TABLE C-2

TABLE C-3

ExPr1;DJTUFtES OF CONSOLIDATED PtiBLIC SECTOR, PUERTO Rico, SELECTED Y£ARS, 1930-196Ú (millions of current USS)

Total as Proportion of GNP

U_7

2.0

c' .4

11..8

1.& 33,1

4.3 7.7

38.0 94.9

Municipal Governments

Public Corporation sa

1940

18,9 2.'-.4

6.8 11.4

19':f

10&?

1954

154.7

15.5 28,4

1962

263.6

57.1

19b6

436.4

78.5

179.2

Year 1930

44.9

I .6 21.5

15.3

162.0 236,4

38.8

454.4

22.3

55.8

749.9

24.5

21.4

alncluúes capital sper•<ding uniy of bath public corporations and the so-called public service enterrsire . The most important public corpora^ions

frorc: a total expenditure viewpoint are the Wav r Resources Authority, the Aqueduct and Sewers Authority, and the Urban Renewal and Houaíng

PUBLIC EXPENDITURE PATTERN~

Trust Eundb

Total

Central Goren ~pot

Corporation (CRUV); pubic service enterpthes include 20 minor funds, among whi.h are the national lottery, a coffee insurance programm., and the I .pia and Gusyama irsicration services. bExcIudes f derai t:as= fur:mss. Sow:e: See Appendix B; aI'o utilized wort: .' eet data from Office of the Governor of Pue Eo Rico, Bureau of the Budget. Although sc;ne of the data gathered from iluelget Bureau records they represent actual disbursements, not more budgeted figures.

00 ~

TABLE C•4

CoNSDLIDATED PUBLIC SECTOR EXPENDITURES BY ECQOM1C CATEGORY, PUERTO Rico,

t966

Public Sector Area

Current Expenditures

Capital Expenditures

Debt Service

Transfers to Other Areas

Total Less Transfers

Central Governments

311.6

98.4

26.4

100.5

436.4

56.8

387.3

Public Corporations

157.3

173.2

Public Service Enterprises

53.8

0.5

14.8

54.3

Trust & Agency Funds

54.7

1.1

0.2

55.8

Municipalities

53.7

16.4

8.4

78.5

Totals

631.1

288.0

93.2

115.5

1,012.3

aThese totals pertinent to the disbursemei►ts of the central government differ from those found in other tables due to conceptual differences. Since the principal focus of this table ir on the total outlays of the entire public sector rather than on those of each public sector area no reconciliation is attempted.

NA - Data not available. Source. See Table G3•

S~1Qn.LS DiwoxOD3 ario 1VDos

(millions of current USS )

SEPARATA Número S

Arthur J. Mann

THE FISCAL SYSTEM AND INCOME DISTRIBUTION: THE CASE OF PUERTO RICO

/Public Finance Quarterly, 4 (July 1976):

339-366/

This study provides statistical estimates of the incidence of Puerto Rico's fiscal system. /n developing the estimates the openness of the Gbmmon• wealth's economy and its particular internal economic structure are taken into account. The tax system is found to be slightly progressive across income classes, whereas the public expenditure structure is found to be regressive (favoring the lower income classes). Thus, the overall fiscal systemn redistributes real incomes from higher to lower income classes.

THE FISCAL SYSTEM AND INCOME DISTRIBUTION: THE CASE OF PUERTO RICO ARTHUR J. MANN University of Puerto Rico in Mayaguez

This paper presents estimates by family income class of the tax burden and of the incidence of publicly expended monies in Puerto Rico for the year 1970. Since all fiscal systems contain distinct income redistributive properties, the analysis additionally generates, as a byproduct, both preand postfisc income distributions. The incidence concept employed is that normally used in studies of this nature, i.e., net fiscal incidence is defined as the final net change in the relative current income position of the family brought about by public budget activities. Both tax and expenditure effective rate patterns have been measured against a "broad" income base, AUTUOR'S NUT}': this articic is based upon the author's study for the Puerto Rican Tax Reform Commission, entitled "La Carea de tas Contribuciones y los lkneticios de ios Gastos Publieos." Puc.rio Rk:u, t97Ú (suhmission made in October, 1973). The author is indebted to twat Andie, Irving J. Coffman, Richard A. Musgrave, I calve Viseasillas, and two anonymous referees for their comments. PUBLIC FINANCE QUARTERLY, Vol. •4 No. 3, July 1976 ©1176 Saqe Publications, Inc.

13391

[340] PUBLIC FINANCE QUARTERLY

has

which includes monetary and imputed income items (see Table A-2 for the income base components). Although the statistical results are presented with apparent awareness of the many limitations of mathematical precision, this type of analysis is essential, for surely what is produced is been noted by no more than a reasonable estimate. As others,' a principal defect of static incidence studies is that they are based upon the heroic premise that the procedure of adding in or subtracting out income changes does not alter the original income distribution. Here is a general equilibrium problem par excellence, and to resolve it would entail the complete elimination of government and the reallocation of public resources to the private sector. There is assumed to be no TABLE 1

Tax Payments, Puerto Ríco, 1970 (millions of dollars) Line 1. 2. 3. 4. 5. 6. 7. 8, 9. 10. 11.

Tax Personal Income& Corporate Income& Excises Prepertyc Customs Dutiese Estate and Gift Licenses lottery t'unicipalf ICiscellaneOua :,octal Insurance (u} Pederal (b) Commonwealth

Cross Total

Total list of Exported Portlop

149.3 85.9 239.7 57.7 42.9 6.2 17.3 23.5 10.6 15.1 342.0 213.6 (28.4

149.3 42.pó 207.8 37.0d 40.0 6.2 15.7 23.5 ' 10.6 15.1 342.0 213.6 12.8.4

Percent; 16.6. 4.7 23.4. 4.2. 4.3 0.7 1.8 2.6 1.2 1.7 38.5 24.0 14.4

1).

Total

990.2

889.2

100.0

13. 14.

Total, Excluding line 11 Total, Excluding line 11(a)

6482 776.6

72 £75.6

lí .5 76.0

SOURCES: Departamento de Hacienda de Puerto Rico, Informe Anual del Secretario, 1970 (San Juan: 1970), 1718; Department of the Treasury of Puerto Rico, Economy and Finances. Puerto Rico, 1970 (San Juan: 1971), 25; work sheets of the Office of Economic and Financial Studies of the Treasury Department; Departamento de Hacienda, Negociado de Asuntos Municipales, Informe Anual del Negoeiado de Asuntos Municipales (San Juan: 1971), tabla 6; work sheets of the Bureau of Economic Piannin9 of the Puerto Rican Planning Board. NOTES: a. One-naif of the partne' nip income tax is allocated ro personal income and the other half to corporate income since partnerships are taxed under the rates applicable to corporations, an partners are +,,xed as individuals on profits distributed to them. I,. Standard case assumptions described in text. C. Includes the portions corresponding to the central and municipal governments. d. Compromise case astumoticn6 described in text. e. Includes administrative costs. f. Excludes property tax already included in line 4.

Mann I FISCAL SYSTEM AND INCOME DISTRIBUTION 13411

excess burden cawed by taxation and no consumer surplus effectuated by public expenditures; no explicit distinction is made between current and capital spending, although their temporal effects probably differ; little definitive empirical evidence supports the shifting assumptions utilized; there may be serious but incontrollable (for the researcher) defects in the basic data series, and the extrapulation procedures described in the notes to Table A-I may inject an unknown error guantity.2 THE DISTRIBUTION OF THE TAX 3UfDEN

Table 1 presents a list of those taxes to which 4uerto Rican residents are subject. Most of the levies are imposed by the central government, although customs duties and federal (United States) social security taxes (lines 5 and 11 a) do not fall under Commonwealth jurisdiction. Since Puerto Rico is a classic example of an open economy, tax (and expenditure) exportation is accounted for by using three different exportation coefficients:3

Coefficient I—computed by taking the ratio of nonresident investtotal investment; the denominator value is, estimated ment through use of a series of incremental capital-output ratios, and the final result is a figure of 0.549. Hence, it is assumed that 54.9% of those taxes falling on capital are '.ported. Coefficient II—since goods produced in Puerto Rico but exported comprise 38% of GNP, it is assumed that resident consumers bear the burden of only 62% of certain cunswnption taxes. Coefficient III--since nonresident consumers accounted for 6.8% of total personal consumption outlays, it is postulated that only 93.2% of excise taxes fall upon resident taxpayers. Naturally, tax importation is also a distinct probability, but since such imports would not be Puerto: Rican variables they are disregarded. The results of the tax side of the analysis will obviously be a function of the tax shifting assumptions used. In the case of most Puerto Rican taxes there is no reason to make shifting assumptions that vary from those normally employed in this

13421 PUBLIC FINANCE QUARTERLY

type of analysis, and these particular assumptions will not be defended. Hence, it is assumed that the personal income tax is not shilftable, and .falls entirely on the statutory taxpayer; the excise taxes,° customs duties, and municipal gross receipts taxes are assumed to be completely shifted forward; the estate and gift taxes are assumed to burden only those families in the highest income classes (a premix based on a sample of estate TABLE 2

Tax Shifting Assumptions

Pa z rersonal Income

Burden Palle on

Exportatiol<i Aaaumption

Taxpeyer

Corporato income 40% consua,izs 50% dividends or capital income 10% wages and salaries ?xc5 see

ConetUlers

Property

Orthodox Approach: Real commercial land. Real commercial improvements Residential owners Residential renters Personal corporate Personal non-corporate Revised Orthodox Approach: All commercial All residential Nieaskowski Approach

Ci nt.o^,s Duties

Consumers

Bet-ate and Cirt

Upper income families

l.i.een≥e

Families Businesses

Distrfhutivr Series Line of Table A-t

none

50% 62,2% none

22 12 or 23 t

coefficient III

coefficient I coefficient II none none see corporate income prenises none

12 22 25 26

coefficient I none 25%

23 25,26 23

coefficient III

none none coefficient II

D a 22

ottcry

°Consumers" of lottery

none

iunicipal

Consumers

none

~~isrel :~nenun

Consumers

none

none none

1,17,21 1,17,21

none none none

1,17,21 1,17,21 22

'nri ll tn:'ur'YnCe Cane A Employee Portion--employe,_ Smployer Portion--employee Case S r^mployee Portion--employee Taployer ?ortion--e.aployee(5O%) eon-'ncrai;50%)

e. Each individual excise amount is distrlbutcd by its own speclfic distributive series which is not found in Table A-i. b. The percentages, applicable to only the three highest income classes, are 13.6. i9.i, and 73.3.

Mann / FISCAL SYSTEM AND INCOME DISTRIBUTION [3431

tax records); license taxes are postulated as burdening those who directly pay them or those who consume goods produced by economic units if the unit initially absorbs the tax; the ``lottery" tax distributes net state lottery income among families according to their propensities to gamble (i.e., it is not shifted). The cases of the corporation income tax, the property tax, and the social insurance taxes do merit further elaboration. The general lack of consensus regarding the incideiice of the corporate profits tax is well known, and the Puerto Rican case, which has never been subje~ied to econometric analysis a la Krzyzaniak-Musgrave, may be even more complicated than others. The Industrial Tax Exemption Law has created a taxed and an exempt corporate sector, and the former operates in the context of an extremely open economy. Certainly tax exportation is a possibility, and the vast series of market relations between. the two corporate sectors and between the taxed corporate sector and the rest of the domestic economy must be considered. Given this large number of. imponderables it is hypothesized that 40% of the tax is shifted forward, 50% falls on corporate profits, and 10% is shifted backward.5 With respect to the 40% presumed to be shifted forward, it might be supposed that Puerto Rico's rather highly imperfect domestic market structures would suggest a larger degree of shifting. But it must be recognized that the taxed firm might export either to United States or world markets, where it may have to take the price as given.6 Thus, 40 of the gross corporate tax, or 534.4 million, is postulated as initially reflected in higher product prices. However, since not all those goods produced by the taxed sector are consumed domestically, tax exportation by means of the export mechanism is presumed. Based on data provided by several government agencies it was estimated that 50'x- of taxed corporate sector production is exported. and in this manner only 517.'2 million burden residents. Regarding the (543.0 million) assemcd to lay; :7n profits, it must logically be hypothesized that both resident and

(3441 PUBLIC FINANCE QUARTERLY

nonresident stockholder incomes are reduced in identical proportions. Using data provided by the Puerto Rican Treasury Department and the Planning Board's Economic Planning Division, a figure of 37.8 : was computed as the ratio of the sum of dividends and retained earnings accruing to residents to after-tax corporate profits in the taxable corporate sector. Hence, only S16.2 million are allocated to the resident tax burden. The question arises as to which distributive series, dividend income or total capital income, is the most appropriate to distribute this amount among income classes. If distributed by the former the implicit assumption appears to be that the taxed corporate capital is totally immobile and will not shift to the exempt, unincorporated, and/or external sectors. This is difficult to accept, for what in effect is then maintained is that investors in taxed companies are satisfied with a lower return than they might receive in other areas, or that they are "locked-in" their low-return investments. Of course, the gross rate of return in the taxed sector might be high enough to compensate for a lower after-tax net return. If the $16.2 million are distributed by a capital income series, one is relying on a modified Harberger formulation, in that as the tax reduces the rate of return on capital invested in the taxed corporate sector, capital shifts to the unincorporated or exempt sector. As this occurs the return on capital in the corporate sector rises, while that on capital in the other sectors falls. After reaching equilibrium where the net after-tax rates are equal in all sectors, it is concluded that the corporate tax burden is shifted to total capital by means of a reduced return; tax exportation coefficient I would then be used in this instance. To resolve the quandary of which series to use a compromise measure has been adopted, and the SI6.2 million total is distributed by both series, after which a simple arithmetic average is taken. Finally, the t0 (S8.6 million) postulated to be shifted backward is based on the existence in Puerto Rico of less than perfectly competitive tabor markets. Certainly it is possible that a firm which possesses forward-shifting market power also has the influence to shift the corporate tax backward onto wages.

Mann / FISCAL SYSTEM AND INCOME DISTRIBUTION [3451

The lesser is labor mobility between the taxed sector on one hand and the exempt and unincorporated sectors on the other hand, the greater is this potential. A high unemployment rate (officially at 12-15%) would seem to strengthen the argument for backward-shifting, at least in those areas where job skill requirements are minimal. Moreover, rather weak unionization and/or collective bargaining at the plant level (as opposed to the industry level), generally the case in Puerto Rico, offer further support for a backward-shifting hypothesis. The problem of the incidence of the property tax may be even more complex than that of the corporation income tax, for there are four groups that are affected by this levy: consumers, renters, property owners, and capital income recipients. Such a classification leads to at least a three-part division regarding the various points of view on property tax incidence. The first point of view, here labeled the orthodóx approach, involves a disaggregation of total property tax collections into revenue from each type of property (i.e., real and personal property and their subdivisions—business and residential land and improvements, owner-occupied and renter-occupied dwellings). Table 2 summarizes the shifting assumptions, applied to each property type, but due to space limitations the methodology employed in the breakdown and the reasons for adopting each specific shifting assumption are not discussed in this paper. A second point of view, here labeled the revised orthodox approach, reasons that the property tax falls essentially on land and capital. In order for a firm to be able to shift its property tax forward it must possess substantial market power, but in Puerto Rico it seems reasonable to assume that the property tax base is much less concentrated in firms with oligopolistic market power than is the corporate tax base. Furthermore, many firms which do have such power are tax exempt in the first place. This suggests that the property tax on commercial real and personal property might be distributed according to capital income, after deducting the exported portion using coefficient l: that part of the tax burdening residential property remains similar to the orthodox assumptions.

13461 PUBLIC FINANCE QJARTERLY

A third viewpoint, here labeird the A ieszkowski approach, argues that the entire tax falls on capital income.' Although the Mieszkowski general equilibriu.a approach is subject to a number of limitations, it is applicable to the Puerto Rican situation despite the fact that it does not consider the property tax on land. Although each municipality imposes its own tax rte (in addition to the central government's uniform rate), this does not seem to significantly alter the fundamental conclusions. Certainly it is possible that differential rates may induce long-run excise-like effects, thereby causing capital to move toward lower-rate municipalities, with the final result that net return rates are equalized. Thus, tha entire portion of the property tax that burdens residents is distributed according to income derived from asset holdings. However, only 25`áb (instead of 54.9%) is asstr:ned exported to take into account the fact that practically the entire burden on residential property falls on residents. No single approach to property tax incidence appears in the statistical tables. For u :e in the text tables a simple unwéighted average of each atsolure total in each income class under the three approaches is taken. Thus, a "compromise" approach instead cf a "standard" :: ç p. oach (as in the corporate profits tax case) is employed. The incidence of social insurance taxes is especially important, given that these imposts comprise almost two-fifths of the overall tax burden. This l= h proportio may in itself be rather surprising, and comes aL ut because the Puerto Rican resident is subject to two levels of soc;ai security taxes, federal and Commonwealth. Moreo',er, the federal portion is so proportionally high due to the appiicaticn o; equal rates in an area where per capita income levels are less than half of those on the continent. Regarding the quo lion of sccia insurance tax incidence, it seems generally acce,tauie to posit nonshiftability of the employee-paid portion of rl,e tax. Ccntrov~rsy arises, however, as to the incidencc of thr em ioyer s portion. Brittain's empirical works on the incidence of s actin:: factor taxes in the

TABLE 3

Effective Tax Rates, Puerto Rico, 1970 (broad income base) ` .

1. o. 7. U. '?, G. .

12. 1 . 14.

Tex ;:or?aril lnco~+e ;nr,n::.rrtto 2ncone =.xr.icer• Pr+.ne:Ty mn n•,tiea ...,•..,te an4 Gift .'icen:es ?,ntterp ur3ciprl aneou.; ...=c Soclai 1ne+arance (a2 ?r'.eral (': ^cnacnwcalth Tata1 Total, Excluding íneu;ance r t i Erclu4 ng .dwr'I Social i.ns.:zva: e

Tamily Income Clasp le~t than £1,000- 22.000$1.097 $?,9?9 0.2 0.5 0.8 0.7 0.5 3.6 3.8 4.: 0.6 0.5 0.5 1.1 1.1 1.n 0.2

0.2 0.3 0.3

0.2 O.e

33,000$3.999 1.0 0.8

15,000•$7teg9 .

4.6 0.6

84,`00$4.999 1.4 0.8 5.1 0.6

1.9 0.8 4.7 0.7

$7.500$9,999 3.6 0.9 5.8 0.8

1.0

0.9

0.7

0.6

0.9

0.2

0.4 0.5 0.2

0.5 0.6 0.2

0.3 0.5 0.9 0.2

0,1

0.4 9.0 5.4 3.6 22.5

0.3 8.7 5.2 ~~ 23.1

0.3 7.7 4.3 j~ 24.6

0.8 0.4 O. 0.2 0,2

9.4 6.1 3,'.• 19.9

0.6 0.5 0.9 0.2 0.2 7,1 3.4 3.7 25.0

3.3 1.3 1.8 22.0

0.3 7.a 4.9 2.9 20.2

$/0,000- Tt2,500. 115,000$1 ,999 $12,499 $19.99º 4.7 6.6 7.9 1.0 0.9 1.3 5.6 4.2 5.3 0.7 0.9 1.4

$20,000 az+d over 8,3 1.8 3.7 1.8

rot4 3.4 1.0 4.8 0.8

0.5 0.2 C+.5 4.9 3.7 1.2 12.2

0.4 6.2 4.4 1.8 13.6

7.3 5.0 2.3 15.6

17.8

^.3 0.5 0.3 0.4 9.3 6.1 3.2 19,7

7.3

7.4

8.3

9.4

10.4

10.5

13.5

14.4

16.9

17,9

18.7

12.4

8.5

9.2

10.6

12.0

13.6

13.8

17.1

17.9

20.5

21.6

20.5

15.3

0.3 0.4

0.5 0.3 0.4 8.4 5.8 2.6

0.4 !'.3 0.9 0.4

0.4 0.5 0.2

SOURCES: Each absolute amount of tax paid by each income class (these quantities are not given in this paper) is expressed as a percentage of broad incorm (Table A•2, llne 18) in the corresponding class.

13481 PUBLIC FINANCE QUARTERLY

United States suggests that labor absorbs the entire burden. While his analysis cannot be totally transferred to Puerto Rico's case it is, without doubt, suggestive. Consequently, two different shifting assumptions have been used in this paper, and each assumes that the employee share falls directly upon the employee. In the first case the employer's share is hypothesized to be completely shifted backward, in line with Brittain's results. In the second case half the employer portion is assumed shifted backward, with the remainder shifted forward. It does not really seem reasonable to propose complete forward shifting of the employer's portion in Puerto Rico, for if labor is employed in export industries that lack external market power such transfer is not easily made; it is additionally difficult to conceive of those instances in which the domestic producer meets import competition. On the other hand, the absence of strong collective bargaining at the industry level might serve to intensify backward shifting. The effective rates that do appear in Table 3 are computed in a fashion similar to those for the • roperty tax; i.e., they are unweighted averages of the two postulated shifting cases. From line 12 of Table 3 it is readily perceived that the overall tax structure is progressive across the entire gamut of income classes until the final open-ended class. Whether or not the ' effective rate decline in this latter class is significant cannot be • determined, for the effective rates within it are unla~own. Therefore, the policy implications of the estimates are limited to those family income groups below $20,000. The experiments contained in lines 13 and 14, first, excluding all social insurance contributions and second, excluding only federal social insurance collections, leave unvaried the general progressivity conclusion. Of the individual taxes the personal income tax is certainly (and unsurprisingly) the most progressive (discounting estate 4axes, which do not affect all income categories); the fact that it is na as progressive as might be expected from a glance at the tax stat..t~g is due to tax base erosion caused by evasion and avoi nre. under the "standard" corporate tax shifting asump-

Mann / FISCAL SYSTEM AND INCOME DISTRIBUTION [3491

tions a basically proportional pattern is found until relatively high income levels, where the rates turn decidedly progressive. Excise taxes demonstrate an essentially progressive pattern until income levels of around 510,000, from where they turn regressive; to the outside observer such progressivity may seem quite surprising, but many of the items subject to excises have been selected precisely because lower income classes do not abundantly consume them (e.g., private motor vehicles). In other instances the statutory rate structure is progressive and/or exemptions are granted which permit lower income families to purchase the items without paying the excise.9 The property tax, under the "average" rates presented here, reveals a proportional structure until high income levels, where it converts into a progressive levy.' ° Social insurance contributions illustrate a progressive-regressive pattern, which is principally due to the federal OASDHI impositions; this is due, in turn, to the taxation of only the first 57,800 (in 1970) in wages and salaries. The remaining levies, of little relative importance, are found to be essentially proportional in character (with the obvious exception of estate taxes). THE DISTRIBUTION OF PUBLIC EXPENDITURE BENEFITS

Taxes reduce family income levels, but these same tax revenues, when spent, serve to augment real family incomes. There are various alternative methods that might be used to distribute public expenditure benefits among family income classes, the ideal being the "benefits-received" approach. Under this approach each family income group would estimate the "value" of public services it receives, and benefit determination would then be proportional to such a distribution. The difficulty in employing this approach is; of course, that of revealed preferences, for in order to estimate the benefits received by each family the family demand curve for the publicly provided good must be known. The limitations imposed by the nonoperational nature of this methodology lead

(350) PUBLIC FINANCE QUARTERLY

one to fall back upon a "cost. of service" approach, under which it is assumed that the cost incurred by the public sector in benefiting each income group is exactly equal to the value of the benefits received by each group. Thus, .the point of departure becomes the identification of those family income classes for which given public disbursements are made --the determination of public expendittt re incidence. The public expenditures to be dealt with are presented in Table 4. This functional category classitication covers the entire gamut of public disbursements, from those which are allocable TABLE 4

Public Expenditures' by Functional Category, Puerto Rico, 1970 ;millions of dollars' Line 1. 2. 3. 4. 5. 6.

Functional Cstegorv b

1O.

Education Health Nt Sanitation Sor.iai Welfare Agricultural Development Lahor 2elatlors Trannportation Q Cosenunl cation Housing Interest Social Insurance (a) Federal (b) Commonwealth 'C~neral'd

11.

ñ3a1

12.

Total, Excluding Federal Social lnsurance

7. 8. 9.

Grose ^otal`'

Total Net of F:xDorted Portion

248.1 146.1 59.2

248.1 141.1

34.6

59.2 34.6

8.8

68.2 20.2. 26.3 312.9 204.6 108.3

60.7 20.2 6.6 3129 204.6 108.5 232.2

Per 22.1

12.5 5.3 3.1 0.8 5.4 1.8 0.6 27.8 18.2

9.6 20.7

232.2 1,156.6

9,3?4.d

100.0

952.0

919.0

81.8

SOURCES: Work sheets, reports, and record: of the Accaunting Division of the Treasury Dept.; w~'rk shear of the Ecoriornu Pianning Bureau of the Planning Ftoard; work sheets and the Informe Anuat of tie Bureau of Municipal Affairs of the Treasury Dept. NOTES: a. Includes Ile expenditures of both the •:entr~i and municipal governments, with double-counting eliminated. Expenditures tinanced ny grants from the United States are excluded, consistent with the purpose of measuring the rodistributive effects of the huertn Rican fiscal system in isolation; the exception is federal outlays on social insurance, which most bA Included since tax contributions to these same programs are inceded as part of the ter. burden. b. Although most of ine dato were obtain^d from the Accounting Division of the Treasury Departrncn;, the categories co'respond more to Chase developed by the Bud at Bureau. C. Includes Central government grants to ;tirbiic c;cr•porations. d. Includes expenditures riot ahocab.a to ?ecific groups: eo+omerciat, industrial, cooperative, and lecreatlon.+l devºlopnrr;'.; protection of persons and property and consumer protection; general rubric administration; lane conservation; non -classifies services•.

Mann j FISCAL SYST„ AND INCCME DISTRIBUTION 13511

to easily defined family income groups (specific,type public outlays) to those not dearly allocable to any specific group (pure public goods). An example of the former are outlays on social welfare or on social insurance transfers (which are conceptually nothing more than negative taxes); an illustration of the latter are disbursements on the protection of persons and pr'perty. Between the poles fail public expenditures on education, which combine characteristics df both. '' In the interest of brevity a detailed discussion of the incidence of public expenditures is entirely omitted.' 2 Table 5 summarizes the fauúly groups the benefits (costs) are assumed TAE4.E 5

Public Expenditure Sh~fting Assumptions Punctlanal 5 p n:litore 77enafit

.ecetved b3

Exportation sA s.msption

Distributive Series Line of 29b1e A-1

none 25 p;rcent by a3 society tiemainiag 75 percept to fa¢ilie+ of atudents: Prey-university etudecte one Oniversity students none Nee 1 th All eoclety none Roepital P. clinic patters+.n none Children in school lunch proarar, none I+ueinesa firms coefficient II :wily residence. rove 4oc1a1 ■e'ifa:e 2ocial welf:.xe recipients none £3rieulturo Farm farm ies nore Faro iamil^ inc r._ none Lao;r Relatior.s Wage á ealacy earners nor2 roperty ewre z Transpor'++tion none hid;hwav Tamil; :.era none Consume^s of t_~tnaportPrl gccd:: coefficient .Air A marine paesenEeru note ^ubiic cue p ºngers mire Communication Telegraph h. telep;lane uaern itesiderts of ' .ninR, probe-tn Roa.~ing none Sondholdern 8O 7_ Centrce Public Debt govt. A 5C of ~uric.~l Social Insureccer • manes OASDHI rir~ct recipients uiireci recipi.e-te none Pension, none Onempi••yment Direct recipl•-e'^ IU'ita Disab.lity Direct :eeipie:.•.s kIucation

'Ge +m l ' bxpenditcres: Aseuaption I Familiec in proporti brood icccnee Ae.uvptior, 2 Fc.caliec In pre.;•ortL.n to rurbe,Pamii.iee in 1'oportior.. t .ssucption 3 disc^able money `. ^'ene

35 29 30 35 31 29 22 39 7 33 34 1

32 23

5 &

none

5? 35

none

ci`c.e

TABLE B

Effective Public Expenditure Rates, Puerto Rica, 1970 (broad income base)

I ,r•r 1'•mgti~~n1: C13esiti atinn +, Fbincatlon 2. !+e~1t? • Szritatian S. ?ncinl ,.eltºre 4. Afrleºltutal Dev!lop,eent 5. 6. 7. 1. ~. 10. 11. 12. 13.

Family Incnm! CIn^A ters than $1,000- $2,000$1,000 $1.999 $2,494 11 .5 74.9 1 5.3 60.7 9,7 12.3 26.9 3.6 9.9 1,7 10.8 2.6

:..,hnr ,r1:ltiona ?ronepartt+ti,cn-Coer•iunicution Nnuinc f~t.nr.'-t '-ri~11 insurance l,t ;`) CoamnnwealtA s,?tot.1 'nner.2."Exnendituzen :••,nr'1nn i :-anp•lnn 2 ..rr+i~,ptinn 3

n, ';11.11 )1tVrle 14. ánr•a',r;ion 1 • +5. 'xcled:ne 9(a) 16. ,ir^umntion 2 7. ==.clndir.r 9(a) 111. Ar:ur!tr•i"n 3 19. •- »dinfr 9(5)

$3,000- $4,070• 44 $4,99' 8.6 6.2 6,1 2.4 1.6 1.8 1.1 0.7

s5,o0o$7.499 4.9 1,4

s7,50019.999 3.4 1.0

$10,o0dj12,499 2.4 0.7

0.5 0.2 1.8 0.1 3.8 2.0

0.3 0.2 1.9

10.7

9.5

$12,500- 41Tj,00O- 120,000 jt4.949 119.948 and over 0.6 1.9 2,4 0.6 0.4 0.5 — — 0.3 0.2 0.6 0,2 0.2 0.2 1.8 1.42,4 0.2 0.3 0.4 3.8 2.1 3,4 1,0 1,1 0.7 2,J3 1.4 2,3 8.8 5.1 9.6

To$el 5,7 3.2 1.4 0,1 0.2 1.4 0.5 0.2 7.2 4.7 2.S 20.4

0,1 0.2 0.6 0,8 3.1 1.2 0.1 0,1 16.3 18.1 23.2 14.2 14,2 13.8 1,4 5.0 4.3 199.1 67.2 46.2

0.2 0.9 0,9 _0.1 12.4 e.7 4,7 32.2

0,2 1, 0.4 0.1 9.4 5.8 3.6 '2.1

0.3 0.6 0.2 1.4 0.1 0.1 5,1 3.6 i. 14.1

5.3 41,3 6.3

5.3 14.9 6.5

5.3 9.8 6.2

5.3 ".7 5.9

5.3 b.fl 5.6

5.3 5.1 5,5

5.3 3.5 5.2

5.3 2.8 5.1

5.3 2.3 4.9

5.3 1.5 3.9

5.3 1.2 4.3

5.3 5.3 5.3

204.3 190.1 240,4 276.2 205.4 191.2

77.5 54.2 82.1 63.9 75.7 .5.4

51.5 37.7 56.0 42.2 57.3 58.5

37.5 29,3 39.9 31.7 38.1 %9.8

27.4 21.7 74.1 22.3 77.7 72.0

19.4 15.9 19.2 15.7 a19.7 16,1

16.0 14,1 14.3 12,5 15,9 14.0

14.8 12.8 12,2 10.3 14.5 12.6

14,1 13,1 11.1 10.1 13.7 12.7

15.0 13.9 11,1 10.0 13.5 12.5

10.4 9,7 6.3 5.6 9.4 8.7

25.7 21.1 25.7 21.1 25.7 21.1

0.7 8.8

0,1 3.8 1,9

,•9

SOURCES: Each absolute amount of public expenditure benefit received by each income class (these quantities are not given in this paper) Is expressed as a percentage of broad income (Table A-2, line 18) In the corresponding class. Since this procedure was followed the details may not sum to the totals. Line 10 Is the sum of lines 1-9; lines 14, 16, and 18 are the sums of line 16 and lines 11, 12, and 13 respectively; lines 15, 17, and 19 are computed by subtracting line 9(a) from lines 14, 16, and 18 respectively.

TABLE 7

ffeetive Net iscat Incidence Rates, Puerto Rico, 1970 (broad income base)

.. 3.

e, 5. 5.

T1.:n ..:-u1ptiva 1a ' -'-fl SpotrR ?:xcl :-'.i•.T 5~.::al in:~v n nrr F.cr. 8?ng FP48Tri1 ;toci 'i 1o?u.tnCR ar, imr•tiun 2h w'rttl :, ... . Exr J^.r:nr..tce ':xr'.., -,, , ,r,l Ax~.rmption 3' •v -rí' J<ci.,l :.; r.:: T.x•:. . F~ . :i .. ., . : . . •r:7+lGCQ

Parr.;.lY Incnme 11 5 eoa ~tAan 1(1, '.)- d'T,fiUO11,000 ft,9gq ~2.0`)9

13,000l3,gF.~~

14,OC 24.gn?

15b3Of 7. 4^~9

17,50019.99Q

10,000~„1,c.494

112,y00~,i,~y,Q~

115.000. ;19.9

120,000 ead ayer

Toca±

193.0

5?.5

3R.2

19.A '

7.'+

-0.5

-6.5

-8.5

-10,4

-9.9

-11.3

5.4

i20.c

4t.9

25.2

15.6

7.4

4,0

.1.2

-3.5

.6.4

-6.3

-10,4

6.0

182.2

45.4

-7.4

17.4

R.?

2.2

-3.0

-5.2

..1

-7.7

-10.8

3.6

.•29.t

69.1

10.7

72.3

R.4

-0,1

-8.2

-11.1

-13,4

-13.8

-15.4

5.4

216.9

51,6

23.1

18:2

R,4

3.8

-3.0

••6.0

-9.5

-10.2

.14.4

6.0

219.3

55.1

3s.e

s3.9

p.t'

2.0

-4.8

-7.8

-10.2

.1T.6

-14.9

5.6

r:"c . t

60.7

37.1

2J.4

R,n

-0.3

-ó.6

-@.r1.

-10.8

.11.4

-12.4

5.4

•.fR' ,9

43.1

76.0

16.4

R,1

4.2

-1.3

-5.1

-5.8

-7.7

-r1.4

6.0

1a+.}

4E..6

28.2

18.0

2.4

-3.1

-5.5

-7.?

-9.2

-11.8

5.6

SOUR( H Each aLr;n4!' e' ama ,t of tax paid by eaU income bass l subtracted from the absolute arro5-4 01 public expenditure benefit -eraive,1 b• e. r c;aas the rest a. which are absolute net tiscd dlstrlbutior ?g ores, are then expressed as )ercentages of broad income (Table t, a orra5 and:-:^ cos. Actually, the same results may be approximated (due to rounding) by Subtracting Unes 12-14 of Table A-i. :;ne _ 3 frnni l ner. I4-19 : Taú%' ó. r'1C)-rES. crmerai" a cpe=ldltus+s d stributeC propertian4Hy to proad i .cºme. sxaendltuces distrlbured proportioraily to the nt;mL-e- of fami'ies. c --:al" expen itureS d•_eributs0 propdrt unatty to d1s005st r0 nlonpy ;nco»ce.

13541 PUBLIC FINANCE QUARTERLY

to fall upon, and in its final column gives the distributive series in Table A-1 which is employed) Note should be taken that it is postulated that certain public expenditure benefits may be exported to nonresidents. The pattern of public expenditure incidence is presented in Table 6, from where it is readily appreciated that the overall ex penditure structure (lines l 4- l 9) is essentially regressive' " over the entire range of income classes. Those specific components which have the greatest influence on the regressive pattern are expenditures on housing, social welfare, health, education, and federal OASDHI. These results are not really unexpected, for outlays in the former two areas are carried out with the express intent of aiding low-income families, while OASDHI programs principally benefit retirees or families with few alternative income sources. Expenditures on education and health in large part show a regressive structure simply because high-income families do not use public facilities. NET FISCAL INCIDENCE

Both sides of the Puerto Rican budget are now combined to estimate the incidence of the entire fiscal. system. Defining the absolute net. fiscal distribution as the distribution among family income groups of the costs incurred on behalf of each group less the distribution of taxes paid by each group, the net effective fiscal rates presented in Table 7 are merely these absolute figures expressed as a percentage of the corresponding group broad income. Table 7 presents the overall data under three alternative "general" expenditure assumptions in a fashion in which OASDHI benefits and federal social insurance contributions may be excluded if so desired) It is easily seen that the general pattern of net fiscal incidence is regressive, i.e., the implication is clearly that the f cal structure is favorable to low-income classes, which experience a net benefit. In contrast, the .hígher income groups suffer a net burden (denoted by a negative sign). Under all "general" expenditure assumptions

Mann / FISCAL SYSTEM AND INCOME Ó1~tU UlrlbN [3351

families in income groups up to a level of $5,000 realize net benefits (lines 1, 4, and 7), whereas those "average" families above $5,000 experience a net burden. The exclusion of all social insurance programs (lines 2, 5, and 8)' 6 or only federal

social security (lines 3, 6, and 9) has the effect of raising to $7,500 the family income level at which families begin to suffer a net fiscal burden. Thus, a redistribution of income is effectuated through the fiscal system by rechanneling real income from higher to lower income groups. The "break-even point," that income level at which net benefits turn to net burdens, cannot be specifically identified, and it would be better to conceive of it as a "break-even range." For the overall fiscal system (including social insurance) this "range" probably

falls between $4,500 and $6,500 (see Figure 1).

04.3 192.1

72.6

60 4

FJtpenditvres

Net . -20

Taxes

~ -40 -60 I

14 .

20 and over

Inccme(51,000) SJURCES: Table 3, line 12; Table 6, line 14; Table 7, line 1.

Figure 1: Effective Tax, Public Expenditure, and Net Fiscal Rates (broad income basal

13561 PUBLIC FINANCE QUARTERLY

Underlying the net fiscal incidence analysis is the existence of a budgetary deficit. In other words, since this study has distributed $1,124.4 million of public expenditures against only $$89.2 million of taxes, the gains experienced by the lower income classes do not equal the losses suffered by the upper income classes, i.e., the net aggregate gains are greater than the net aggregate losses. And that portion of net gains that does not consist of "transfers" from the upper income classes can be traced to the excess of expenditures over taxes. The deficit has several causes: (I) all expenditures, regardless of their source of financing, were allocated to income classes, whereas only tax incomes net of bond income were allocated iii a similar manner; (2) the tax amounts assumed exported are greater than the exported expenditure amounts;" (3) income taxes withheld from nonresidents and nontax incomes (from, for example, sales of public assets) are excluded from the tax burden, but do appear in the outlay totals. As a co:isequence,'the net fiscal rates are higher or greater than they would be in the absence of the deficit. Moreover, the presence of the deficit affects not only the level but the pattern of net fiscal incidence. In order to appreciate the net fiscal pattern sans deficit, it can be removed by distributing additional "taxes" of $235.2 million according to the overall distribution of the tax Burden. 8 If this is done the general fiscal incidence pattern remains regressive over the entire income secie, but the "break-even range" now found at a slightly lower family income level. sunnMAn<< An:r) CONCLUSIONS Table 4 indicates that the tax structure. although progressive, has much less influence on the redistribution of income than does the expenditwe structure. For example, the first family decile received 1.7% of total incon e. before the fiscal impact, 1.9% after taxes, and 3.8 af: er th.e fiscal impact. Those principal factors that are critical in determining these results are, on the tax side, the postuFitcd nonshiftabi'ity of the entire

Mann / FISCAL SYSTEM AND INCOME DISTRIBUTION 13371-

personal income tax and of 50% of the corporate income tax, the progressive structure of selective excise taxes until relatively high income levels, and the progressive pattern of Commonwealth social insurance taxes both before and after federal social security contributions turn regressive. On the expenditure side the distinctly pro-poor results are dependent upon the assumed and implicit utility fum:tions underlying the distribution of public goods benefits among the family income groups. That this is an extremely critical assumption has been pointed out by Aaron-McGuire (1970) and Maital (1973). The former experimented with arbitrary utility functions which posit that each family "should be imputed a fraction of the total value of the public good proportional to the reciprocal of its marginal utility of private good expenditure" (Aaron-McGuire, 1970: 911); the latter used independent estimates of the value of the inverse of the elasticity of substitution between public and private goods. Application of the two Aaron-McGuire functions and the Maital function to Puerto Rican data under. a "low" quantity of pure public goods assumption does not significantly alter the net distributive pattern; net benefits merely turn to net burdens TABLE 8 The Family Income Distribution by Deciles, Puerto Rico, 1970

Family percentiles Lower 10% 11-20% 21-30% 31-40% 41-50% 51-60% 61-70 71-80;4 81-90% Upper 10%

Absolute Provortione of Total Income Initial Broad After-Tax Income After Income Income Fiscal Impacts 1.7 1.9 3.8 3.9 3.5 5.9 6.8 5.5 5.5 6.2 6.3 7.6 8.2 7.6 7.8 9.0 9.2 9.4 9.7 10.3" 10.7 10.6 10.9 10.3 13.6 15.8 15.5 28.4 29.5 24.9'

a. Represents the arithmetic average of the three "generat" expenditure assumptions.

f 3581 PUBLIC FINANCE QUARTERLY in a slightly lower income range ($3,500-$4,500 in contrast to the original $4,500-$6,500). However, distributing the Puerto Rican data under a "high" quantity of pure public goods assumption by the second Aaron-McGuire function (1/Y2 ) and the Maital function does drastically change the net fiscal pattern, generating net fiscal benefits for families in the income classes over $15,000 while r:ducing the real incomes of middle-income families (in the $2,000-515,000 range). While there is no doubt as to the validity of the Aaron-McGuire contention that the income redistributive effects of fiscal systems do depend upon implicitly or explicitly assumed utility functions, the form of the utility function(s) appears far from attaining a solid empirical base (with all due respect to Maital's contribution). This is a topic which certainly warrants further research, but is beyond the scope of this paper.

NOTES

-'vis

1. For example, see ['rest (1955) and Shoup (1969: 11). 2. For .excellent criticisms of these and many other limitations see Bird and De Wulf (1973) and De Wulf (1975). 3. Two other exportation coefficients specifically related to corporate capital are also used, and are found under the corporate income tax discussion. 4. Although it can be rationally argued that excises are neutral from both the income-uses and income-sources sides, in the case of Puerto Rico it is difficult to argue that they are geographically neutral vis-á both sides. The absence of trade barriers and the use of a common currency between the island and the United States, its principal trading partner, might lead one to analytically treat Puerto Rico as simply another region within a larger continental system. Thus, although tax importation is ignored in this paper, the magnitude of tax exportation may be greater than that assumed under Coefficient Ill. however, this effect is probably of little import since most excise-related revenues are derived either from imported goods of from levies imposed at the producer level on goods almost entirely consumed domestically. For references on this point see McClure (1964 and 1969). 5. The original study offered four alternative cases in addition to this "standard" or "reasonable" cash. it is to be emphasized that no empirical support for these assumptions exists. 6. Of course, if the firm is a subsidiary of a large firm with market power such an effect might be counteracted. 7. this viewpoint is extensively developed in Mieszkowski (1972).

Mann / FiSCAL SYSTEM AND INCOME DISTRIBUTION [3591

all;

8. Brittain (1972) has developed an elaborate theoretical and empirical structure to support his argument. Deran (1967) specifically studied the Puerto Rican case, but her results are based on questionable statistical methodology. 9. Not given in this paper arc the effective rates applicable to specific excises, which vary from totally regressive on cigarettes to principally progressive on motor vehicles. I0. Even under orthodox assumptions the property tax is progressive in the higher income levels (but regressive in the lower groups). 11. In reality, ail public goods combine both characteristics. However, outlays on education (and health) are the only ones so treated in this paper, as it is assumed here that one can clearly distinguish between "specific-type" public goods and pure public goods. 12. Reference may be made to pages 77-109 of the original study for precisely this discussion. 13. The empirical justifications for the proportional allocations within each functional category are also omitted from this paper. Suffice it to state that most of them are based on statistical information taken from published and unpublished public documents. 14. Regressive in the context of public expenditures has a "pro-poor" connotation; it merely refers to the fact that the effective rates decrease ,is family income increases; a progressive structure (of taxes, expenditures, or net fiscal benefits) finds the effective rates rising as income increases. 15. This may be done in order to isolate the redistributive properties of the Puerto Rican fiscal system alone. 16. This is done for the benefit of those who might object to the inclusion of social security "contributions" as taxes. 17. One source of central government revenue, off-shore excises, does not burden residents at they are collected by the federal government under federal excise statutes on sales of Puerto Rican rums in the United States, and are returned to the Commonwealth treasury. Thus, they buoy expenditures but are not reflected in domestic taxes (in 1970 they amounted to $79.6 million). 18. That this procedure is replete with deficiencies is obvious. For an excellent discussion of this poini see Gillespie (1966: 1.64-173). To merely point out one of the limitations, that component of the deficit arising due to Treasxuv bond emissions may be equivalent to an inflation tax. Since the incidence of such a tax works from both the income-sources and income-uses sides, it quite unlikely to be the same as the average tax burden distribution.

REFERENCES AARON, H. and M. McGUIRE (1970) "Public goods and income distribution." Econometricº 38 (November): 907-920. BIRD, R. M. and L. DE WULF (1973) "Taxation and income distribution in Latin America: a critical review of empirical studies." 1MF' Staff Papers 20 (November): 639-682.

1360] PUBLIC FINANCE QUARTERLY BRITTAIN, J. A. (1972) The Payroll Tax for Social Security. Washington: Brookings. DEBAN. E. (1967) "Changes in factor income shares under the social security tax." Rev. of Economies and Statistics 49 (November): 627630. DE WULF, L. (1975) "Fiscal incidence studies in developing countries: survey and critique." IMF Staff Papers 22 (March): 6] I31. CILLESPIE, W. I. (1966) The Incidence of Taxes and Public Expenditures in the Canadian Economy. Study No. 2 of the Royal Commission on Taxation. Ottawa: Queen's Printer. MAITAL, S. (1973) "Public goods and income distribution: some further results." Econometrica 41 (May): 561-568. MCCLURE, C. (1964) "C.ommodity tax incidence in open economies." National Tax Journal 11 (June): 187-204. — --- (1969) "The interregional incidence of general regional taxes." Public Finance 24 (September): 457-484. MIESZKOWSKl, P. (1972) "The property tax: an excise tax or a profits tax?" J. of Public Economics 1 (April): 73-96. PREST, A. R. (1955) "Statistical calculations of the tax burden." Economiea 22 (August): 234-245. SIIOUP, C. S. (1969) Public Finance. Chicago: Aldine.

TABLE A-1 D'13tributiye Series, Puerto Rico, 1970 (perc¢n*3ges)

Lire

¡ 4h

Serte

Fa'P11y ir.co.+.e Claaa I+eo than $t,C^0- $2.000- 35,CO211.999 ,Q?9. I2j.2'

• 8a4eo aad Salarice 0.4 0. reter ri 3erefSts 3. tcome frcm Persots C:ntatde Hc,ce 6.? ♦. Or.elplcy'_._rt c-,ge==aticn ' 3,1 5. Fede si '>oc.~i eca-'.y 6. pv'1i1c Penslen0.6ec01*•ed ,.4 20,i ' Pti111c As:.i:ta-ce 8, Pt-c Prcdoc5d A Cor.4um!d or. 7am 9.9 0.9 9. Lottery Prizes 10. Or).^.ôrporate4 Erro Ir.eome 0,6 t1. Ae❑tai :oree,, 0.9 0,2 12- -ivider.d 'rcme 0.2 13. Partr.ershio iacoae 14. Intereot ' 0 0.1 15, 0,7 ~p.t~ .*~r* 16, kesldentiai Property 'ax 0.2 17. ôntrl*ntlon; to Pen4lon Punde ...1 15, L;tcreat i`-'4 or. Fersorsl Pabt 19. Psrao'si i4.1cre. 'ax — 20. Ir,coae t:hiver4ity Stºdents 4ccclving Aid 9.6 0.7 21. "Covere4• aa4,e ar.d Saiaifea 1.3 "' Ôtal Caasu.mpt.cn 0.4 23. Capital Inrome 0,7 24, ASr lo port £xperces 1.0 25. 0vr.er-C,nsd 11 0c<vpisd 4cnss 1.1 26. Re' a1 Feymentr. ` 0.2 27. :eieph0ne 3 7e146raph 0,1 loo 28. Irleate luto Op 19.1 <'9. r're-1Jr.:verlty 5tud0rtt 30, ;;n cers'_ty Stud,nts 2,4 31, 2ee-7oolis 4020th F.1aíi:ttee 26.4 0 808414.,-, 19.5 32. P' 33, :a1l1 hmiiiee 35,7 34. Fare Paîly IncMxs 6,5 1,9 35. Vue!1er nf =amllien 1,1 36. Famil7 :,oney lrco4:e 1.0 37, Panliy 207.84 Income 38. 'rnrsported C,0de 8xper.Eeo 1.5 8,0 39, bac'1 4?ented Dse:a 1C? 1.2 40. Diepoa,2le Inenme 41, lir:Jn 9118 Expsnees 1.0

4,000-

i5.t00-

s7'u0t3.99í

$40,OtA112,499

7'2.50074

8,5,000att9,999

220.020 ªná¡1 o*7.r

4qá

9.7 5.2

100.0 100,0

1.1

100.0 100.0 100.0 100,0 100.0 100.0 100,0 100.0 104.0 100.0 100.0 740.0 700.0 100.0 100.0 100,0 100.0

7.8 ñ,4

5.9 11.6

8.5 11,8

11,7 .7.7

25.2 22.1

15.8 12,9

1 1.6 10,1

6.5 5.2

2,4 2.0

14.3 23.3 17,7 2.9 '.2,2 1],1 2.1 4.0 4,5 1.2 4,1 0.5 0.2 ' ,9 :.5 ^.,5 0.2

79.5 23.4 21,1 7.8 19,3 21,0 5.' 5.1 4.0 1,7 7,s+ 1.9 0.8 2,2 3.2 1,8 1.0

15.2 15,9 15,3 /4.2 11.5 1F.2 7.7 8.0 8.0 1.1 3.5 4.3 7.6 3,7 5.4 4,3 2,6

11•7 14.7 13,9 '5.7 12.1 14.9 9,4 8,7 «,á 2.0 4.0 8.2 1.5 5.4 1•7.0 8.2 4.4

1 1.7 10,7 1f,.4 14.4 4.8 13.0 13.9 17.8 15.3 4.4 9.6 22.7 3.9 16,8 23.5 22.7 11.9

10.7 7.7 5.9 11,4 6.2 12.2 15.7 12.1 4.7 7.1 23.2 4.1 17,7 71,4 23,2 14.8

5.9 2.3 4,2 9.7 2.8 11.4 8.5 9.4 4.9 7,0 1;,6 4.2 13,9 14.0 !4.6 13.9

1.1 0.7 1.2 9.1 0.7 9.8 5.6 7.4 5.4 4.2 7.8 =.7 9.3 7,9 7.8 11.2

0.2 0.3 2.1 3.8 0.9 5.1 4.1 5.5 7.4 12,0 2.3 0.4 11.3 7.4 7.3 7.0

14.1 3.9 5.3 2.5 3.0 3.9 4,0 0.9 0.2 14.7 3.9 21,0 79,4 21.2 12.5 12,4 4.8 4.♦ 5.8 17.9 5.4 7.7

32.7 7.3 8.0 5.8 4.4

20.7 10.6 9.5 5.4 4,1 7,'. 9,6 3.9 4.3 14.3 10.4 19,9 17.7 9.6 13.4 12,7 9.7 8.7 9.7 12.8 9.6 16.4

14.1 13.5 71.3 6.3 5,6 8.6 11.6 5.8 8.9 11.3 11,7 4.7 t:.5 5.3 ?.6 t1,8 10.8 10.5 11.4 17,3 11.1 17,4

6,4 27.9 22.5 14.3 14.8 27.1 25.2 21.5 24.0 15,3 25.8 1.0 5.8 7.4 18.2 20,8 22.1 21.6 22.8 19.4 22.5 21.3

1,7 15.5 14.5 12.5 12.0 15.2 16.2 19.8 19.9 6.? 13.4 0.9 2.9 9.9 9.4 14.2 14.7 14.4 5.5 13.8 IO.i

0.7 10,8 10.2 8.8 11.9 10.6 10.0 15.8 15.8 2.4 8.9 1,1 4.9 5.3 10.2 10.1 10,0 4.9 9.7 5.1

4.9 5.6 6.3 6.2 6.5 3.4 7.9 7,7 0.6 5.6 0.6 3.0 2.5 5.8 5.8 5.2 2.5 5.4 2.0

2.0 2.3 6.0 2.8 2.E 1.8 3 5 3.4 0.2 5.4 0.4 3.0 0.8 2,4 3.0 2.1 0.9 2.2 *

6.1

6.2 2.3 1.5 16.3 9.1 ^6.6 18.2 15.5 15.5 13.2 7.6 7.2 8.3 13.0 8.3 14.6

10,5 0.3 22.4 23.9 20.9 66.3 48.5 14.4 ?1.7 18.6 4.6 9.4 33.0 2.9 9.5 33.8 34.4 10.7 8.7 18.2 14,1 3.0 0.3 3.5 3.2 tt,e 75,6 8.8 ♦,a 11.0 -

100,0 400.0 100.0 100.D 400.0 100,0 100.0 100,0 100.0 100.0 100.0 700.0 100.0 100.0 100,0 100.0 100.0 100.0 100.0 100.0 /00.0 100.0

(362] PUBLIC FINANCE QUARTERLY NOTES to Table A-7 SOURCES: Lines 1 to 9--Each one of these series is developed from work sheet data that, in their final published form, became part of the Puerto Rican Department of Labor's study entitled ingresos y Gastos de les Famiilas, Puerto Rico, 1963. Since the data refer to calendar year 1963 they had to be extrapolated to 1970 for use in this study. The following extrapolation procedure was developed in consultation wilts Professor Richard A. Musgrave of Harvard University: (a) The 1963 series, presented in abs--'lute dollar amounts, were cºnverted into cumulative percentage series by tart ly income class. (b) The 1963 groups were converted into 1970 family income groups by multiplying the upper limit of each class by the increase in average family income over the time interval. {c) The cumulative percentage series were graphically located together with the 1970 family Income classes; the resulting curves yielded a 1970 percentage distribution based on "unequal income classes. (d) In order to return to the "even" income classes required by this study, interpolation between data points was carried out. The result is a percentage distribution by "even" or "equal" income groups in terms of 1970 dollars. During the entire process the implicit assumption is maintained that the relative family income position remained constant over ttre seven year period. A priori this seems to be a reasonable assumption, and it is partially corroborated by noting that the Girl coefficients (as calculated from the Census of Population) for 1959 and 1969 are 0.5403 and 0.5311 respectively. In order to test the validity of this methodology the same procedure was applied to the 1953 study data, and the results were checked against actual 1963 data. Although the chanyes which occurred in the socioeconomic structure of Puerto Rico between 1953 and 1963 were perhaps greater than those which occurred between 1963 and 1970, the"derived" 1963 series very well approximated the "real" 1963 series. Lines 10 to 19—The raw data for these serles were taken from work sheets provided by the Office of Economic and Financial Studies of the Puerto Rican Treasury Department. Unfortunately they are grouped according to adjusted gross Income levels, for what is needed in this study are data grouped according to family money Income levels. The principal differences between these two income concepts consist of the exclusion from adjusted gross income of public transfer páyments and the income of those low-income families or individuals who are not legally required to file returns. Since both items represent a large portion of family income In the lower income classes, the Treasury data must be adjusted before being used. The adjustment procedure employed here is essentially that ueveloped by Peter Newman in conjunction with the Survey of Consumer Finances of the University of Michigant see Peter Newman, "An Empirical Study of the Distribution of the Tax Burden in the United States, 1955.1.959" (unpublished paper, dated September, 1961). Briefly, the procedure involves calculating gross factor income by family income class, with gross factor income being roughly equivalent to morrey income plus Imputed rent and home-produced food less public transfers: the average family income-in-kind is then subtracted from gross factor income, and the average family transfer payment received plus an estimate of legally unreported income are nxpressed as a proportion of average money income. These proportions, calculated for each income class, are then used to adjust the Treasury data. For example, the Income corresponding to adjusted gross income class 51,000-51,999 is reduced by 50% and added to the next highest category ($2,000-$2,9949), because on the basis of family money income (including transfers) those in this category would fail approximately in a $1,500.$3,000 income class--if such a class existed. Although the $1,000-$1,g99 clac.:: suffers an initial reduction, its net lot_ .s less, for the portion removed from the "less than $1,000" class is added to it. This type of adjustment is carried out in a similar fashion for all the adjusted gross income categories given by Treasury data. Line 20 —Adapted from Coilege !_nuance Examination board, Study of Student Financial Aid in Higher Education In Puerto Rico (March, 1971).

Mann / FISCAL SYSTEM AND tNCOME DISTRIAUTION (363) NOTES to Table A-i (Continued) Line 21—Constructed by assuming that all income derived from wages and salaries In family income classes below $7,500 was covered by OASOHI regulations in 1970 (the actual maximum was $7,800). Toe figures in the higher income categories were computed by rnuitiptying the number of taxable and exempt returns in each category by $7,800. The raw data were takerr from the same source as lines 10-19, and were subjected to the same adjustment procedure. Line 22 and lines 24 to 28--The raw data 'O- these serles (and many others which are not given here) were tai e -i frorn the 193 Department of Labor study entitled Ingresos y Gastos da tas Familias, Puerto Rica, 1963 (actually the unpublished work sheets were used). The procedure employed to update these data to 1970 was developed in consultation with Professor Richard A. Musgrave. From tire survey data it was found that there existed significant dltterunces in consumption patterns between rural and urban areas. Thus, the following procedure was separately applied to the rural and urban sectorial data, (a) The average family money income by income class in 1963 was calculated. (b) The average family • expenditure on each specific item in 1963 was additionally caiculated. (c) From (a) and (b) 1963 average propenslties to consume out of money Income were computed for each income group. (d) These propensities to consume were then graphed against average family money income in each income class. (e) Alter adjusting 191i3 average family income for the 1963.1970 inflation (using the consumer price index), the relation between the dverage propensities to consume and 1970 adjusted family incomes was read off the curves. (fj The rural ano urban propensities to consume, until this juncture separate, were reaggregated using population weights for each income class. (g) 'the derived propensities to consume were teen multiplied by a distribution of money income estimated from q Planning Hoard anaiy.sis done especlally for this study. And from these absolute expenditure series the resulting inn, 22 and fines 24-28 were derived. percentage distributions .hat appear i Line 23--Includes unincorporated firm inco, re, corporate profits net of tax and net of that portion accruing to nonresidents, money and imputed interest and rent income, realized capital gains, and r artnership profits. Line 29—it was assumed that i r urban tones toe distribution of Cl ildren between the ages of 6 and 18 is similar so 'he distribution of urban families by income level (data tram the Census of Population); an analogous assumption was made with respect to the rural zone and cnildren between toe ages of 7 and .17. Usíng the totals thus derived, urban and rural matrices were developed. ii which appeared the number of children corresponding to ear.h age group and .:come class. Subsequently a series of hypotheses was made in relation to sr'rocl er'roiiment by zone, age, and income class. For example, it was assumed that only 50% of urban children between ages 12 and 14 antl in Stir; family income class of "taw than S1,00,:á" were enrolled. Finally, further reductions were made to take into account private school enrollment. Line 30-- Consejo de Educac ión Superior, Grupo 1e T abajo de Planificación Integral, Propuesta de urn Plan Integral de Ayuda a tot cstua;antes de tas Instituciones de Edueaeión Superior en Puerto Rico —Con Alternativas (Pia Piedras- 1969), Appendice 4, Tabla E. Line 31--Adapted from data p'avided I.,' the C'ffice. of Information Systems Development of the Puerto Rican lepartmem c' Heal,,?,. Line 32—Adapted from data provideni by fhe r'~.,crtn i?:can t?epartment of Housing, Und Secretary of Planrring. Lines 33 and 34 —Adapted frorn u.; . Depa'tment al Commerce, Bureau of the Census, 1970 Census of Population, G,otera) Social ;,nd Economic Characteristics, Puerto Rico, Final Report -ái'(1)--t53 (Washington: U.S. Govt. Printing Office, 1972), en. 53.206.

[3641 PUBLIC FINANCE QUARTERLY

NOTES to Tibie A-1 (Continued) Line 35—Pianning Board, Division of Economic Projectionsi these data were subjected to a series of Statistical procedures described In Table A-19 of the original Study. Lines 36 and 37--From lines 9 and 18 of Tablo A-2, Line 38—Derived by the procedure described in lines 24-28. The proportions of expenditures considered transported goods were estimated logically (i.e., they are not Supported empirically). For example, 40% of expenditures on housing were considered as expenditures on transported goods, white the corresponding percentage for food was 100%. Lino 39--The original percentage distributions were taken from the Puerto Rican Labor Department publication Ingresas y Gastos de las Familial, Puerto Rleo, 1963, informe 1 , p. 48. They were extrapolated to 1970 and weighted according to data in U.S. Department of Commerce, 1970 Census of Housing, Detailed Housing Characteristics, Puerto Rico, Final Report i-IC(1)—B53 (Washington: U.S. Govt. Printing Office. 1972), pp. 53.94. Line 40--Derived by subtracting taxes paid from fine 9 of Table A-2. Line 41—Derived by the procedure described in iines 24.28 with reference to only the urban sector. NOTE. Details may not sum to total due to rounding.

• FABL

Te lnco~,i

ie ..~ .

~. . 7. .

/^. '1, 1 .

Ia, iV. 26. r'. IF.

tnmi,'y ;neomt Cast l.eea 68322 51,000- i2 r 000• .cene Itew #1,020, 51.99Q 52.99Q Wnfea :iad 5;:1ar!s4 5.7 5^.; f43,1 1'-.1.^corporated ñtiFiwers 19.1 ,s.0 0.7 0.1 P.*.t D?.i dende ' 2.1 ., let Healiked C.pitai Caioe pn N.nvrobiy Profito r., r,nter Tx~rmer,la 05.5 Ronny 7nta2, 36.1 IneOne .wen'-s tO Non'!}, 2ncCm4 an-klnd ltagea 6 &a3ariea 0.4 Tood 2'rcRvoed 6 Co22enaed 3,7 on ?ere pnf.neni-Pnrated &:einese 1t'cowe-in-kind H.•t e,-nt:+1 Yalae Rceider.ces d.red tgr Oeonpier 0.2 Int^ro,.t Incowe-in-kiM 4p`alnnd rirningó lecivinó to itceidente 0.) Corporate TaX Iwlnstetion :^c2,i Inlpt2nce If2»ntati0n 1.6 7^1.,1, Broad lnc9ae 44,)

3.= 0.3 0.5 0.2 0.1 c'.f.

6.8 2.2 0.7

A•2 Peto Rico, 1970 (mans of dollars) Conronents,

1~3.000-

j,,.) 206.1 29.9 6.0 2.9 0.7

S4.C-00 ;'a +3 271,6

,nü0~:'_. u .,V

ií7,500$9,995 ;a83,2

3i(..Ot~^$12,499 28 3

! ;,~g14.?9á 157.6

614.u93

3?.5 5.2 5.5 0.8

:,r,.F 2 1.;

51.3 9.1 ¶ .5 1.9

.'.ts 7.1

15.3 4.2 1.3 2.9

1,6 0.5 9?.i 60;'.s

1.4 0.4 _ 39.4 513,5

9.7 2.0 1.7 0,4 34.E 368.2

21,0 5.6 5.2 2.1 1.5 0.2 1LQ 209.0

:5.1 2,8

tfi,r-,,;;t,058.2

120,000 and over 235.2

gytal 2.424,9

419.4 15.8 9.6

374.1 75.4 66.7

3.4 0.6 6y •?2.7

26.4 28.7 2.5 02 .1 427,7

39.9 40.0 5.1 590.! 3.626.4

0.3 5.1

0.6 0.2

102.4 273.7

329-^

0,6 0.2 75,4 392,9

2.6

5.5

7.9

'0.4

.•''t.4

13,7

10.5

6.0

2.2

9.0

93,0

5,2

7.8

6.0

5.3

1,8

7.3

1,0 .

0.3

0.2

37.2

0,3

0.4

0.6

0.7

1,4

1.2

0.7

0.4

0.3

1,9

8.0

2,0 0.1

2.3 0.3

3.9 1.1

5.7 2.2

17.7 6.0

26,7 6.1

14.7 3.9

9.8 2.1

11.9

19,6

1,9

2.5

103.5 26.4

2.0

2.8 1,0

2.8 1.3

3.3 1.7

8.8 1,5

2.3

108.3 8.9

163.3 24.8

18.9 312.9

...i2=f 381.0

456.6

7.7 2.8 •s.z 624.9

8.0 2.1

_j.! 293.0

'l,.2 3.7 76.3 242.9.

12.1

0.5

?ZaP 254.9

-'iIZ 130.9

?3sí 393.1

4,569.2

172.9

57.z

33si 443.1

[366) PUBLIC. FINANCE QUARTERLY NOTES to Table A-2 SOURCES: The gross totals of tines 1-9 and 7-8 were taken from Junta de Planificación, Negociado de Pianificac)ón Económica, Ingreso y Producto, Puerto Rleo, 1972 (San Juan: 1973), pp. 30-31. From the gross totals of lines 1.4 were subtracted the Imputed Items which appear in line. 10-14, the latter entries being the product of estimates done by the Division de Proyeccion de Factores Economlcos of the Puarto Rican Planning Board. Line 1--Includes money wages and salaries less employee and employer contributlons to social insurance; distributed by line 1, Table A-1. Lines 2 to S--Distributed respectively by lines 10, 11, 14, and 12 of Table A-1. Line 6—Represents realized capital gains, and was developed by the author from work sheet data provided by the Office of Economic and Financial Studies of the P. R. Treasury Department. Conceptually it would have been appropriate to add to the broad income base all capital ,.sins, both realized and accrued, but no estimate for Puerto Rico is available; distributed by line 15, Table A•1. Line 7—Includes the partnership income tax; distributed by tine 13, Table A-1. Line 8—The totals for all the different public and private sector transfers received by families were taken from work sheets provided by the Junta de Planificación, Negociado de Planificaclón Económica. Due to space constraints the breakdowns are not presented in this paper; suffice it to state that each component was distributed by the appropriate series of Table A-I. Line 9--Sum of lines 1-8. Lines 10 to 13—Distributed respectively by lines 1, 8, I0, and 16 of Table A-1. Line 14—Distributed by line 18, Table A-1; it Is assumed that this serles reasonably represents services rendered to persons by financial intermediaries, but the serles itself inctudes only personal interest payments on personal debt. Line 15—Total taken from Junta de P!aniticaci6n, Negociado de Planificación Económica, Ingreso y Producto, Puerto Rico, 1970 (San Juan: 1911), p. 1, and distributed by line 12, Table A-I. Line 16—This imputation refers to the "standard" case assumption$ regarding the incidence of the corporate profits tax which are described in the body of the text. Since wage-earners and stockholders experience a reduction in their before-tax incomes due to the tax, this logs is returned to them by this item. Line 17--Those fractions of social insurance taxes which fail on the wage-earner in reality form part of his before-tax income, and are returned to him by this item. Line IS—Sum of lines 9.1Í. NOTE: Details may not sum horizontally to totals due to rounding.

Arthur J. Rlann Lc on the faculty of the (iniversirl' of Puerto Rico in Mayaguez. He has published articles in the Melds of public expenditure, tax reform, and income distribution. Ile has also servt'd Cl consultant to the gover;;ments of Puerto Pico and the Dominican republic.

SEPARATA Número 6

Arhtur J. Mann y Donald Perlis (U.P.R., Recinto Universitario de Mayaguez)

UTILITY FUNCTIONS, PUBLIC GOODS AND INCOME DISTRIBUTION

/Public Finance Quarterly, 5 (January 1977):

9-22/

The results of fiscal incidence studies depend upon the underlying utility functions pertinent to the distribution of public good benefits among income classes Generally, the use of such functions is not explicitly recognized. By taking, in the first instance, an extended approach to utility based on an original Aaron-McGuire formulation and, in the second instance, a new approach, it is concluded that the income-equalizing tendencies of fiscal activities muy be overstated when more conventional analysis is employed.

UTILITY FUNCTIONS, PUBLIC GOODS, AND INCOME REDISTRIBUTION ARTHUR J. MANN DONALD PERLIS University of Puerto Rico at Mayaguez

Over the past several decades, but especially since the early 1960s, a growing number of fiscal incidence studies have appeared.' Despite the recognized methodological defects of such analyses, they are duly produced and published with little apparent heed paid to their multitudinous limitations. Tax (and expenditure) policies are partially formulated on the basis of their conclusions, the raison d'être for doing so evidently being founded on the presumption that the educated guess of the economist is better than that of the layman. It is generally the case that these studies find a redistribution of income from higher to lower income groups as having been effectuated by the national (or state) fiscal system. Moreover, the expenditure side of the public budget is normally targeted as the prime mover toward greater income equality. However, the "benefits" of public disbursements on both specific-type PUBLIC FINANCE QUARTERLY, Vol. 5 No. 1, January 1977 ©1977 Sage Publications, Inc.

[91

1101 PUBLIC FINANCE QUARTERLY goods (e.g., public liousir►g) and pure public goods (e.g., defense) may be distributed by rather poor proxy series; a further deficiency is that those series employed to allocate pure public goods outlays among income classes entail the use of implicit average utility functions. Aaron and McGuire (1970 and Maital (1973) have ingeniously attempted to come to grips with this problem (and, as a corollary, with the general equilibrium quandary) by working in terms of a Lindahl solution. The former, positing a set Af eight simplifying assumptions, demonstrate that each family "should be imputed a fraction of the total value of the public good proportional to the reciprocal of its marginal utility of private good expenditure" (Aaron and McGuire, 1970: 911). Nevertheless, to carry out their analysis, they use two arbitrary utility functions apparently not supported by data. Maital (1973) attempts to inject operational content into the functions by using independent estimates of the crucial parameter related to the marginal rate of substitution between private and public goods (he actually finds estimates of the value of the inverse of the elasticity of substitution between public and private goods, labeled O). Apportionment of pure public goods benefits among income classes on the basis of Maital's derived utility functions reveals that public goods provision does tend to redistribute income from high to low income classes but in a magnitude far less than suggested by more conventional studies. In this paper, we attempt to develop further this approach in two directions. First, the Aaron-McGuire argument is extended to include specific good benefits as dependent on a utility function. Second, there is developed an alternative utility model to Maital's fundamental assumption of constant elasticity of substitution of income for public goods (an assumption made apparently for the sake of convenience rather than on the basis of some economic rationale). We have used Puerto Rican data and spending patterns, and c mparisons with United States data suggest a close similarity of trends.

Mann,Perlis / UTILITY FUNCTIONS 111 J THE CONVENTIONAL AND AARON-IMCGUIRE APPROACHES

Table 1 contains the statistical results of the most recently completed conventional fiscal incidence anaiysis (Mann, 1973) of Puerto Rico's tax and expenditure system (see the subheading "Conventional Assumptions"); the last six columns are based on the Aaron-McGuire (1/Y, 1/Y2 ) and Maital (1/Y' 5 ) functions relating to the marginal utility of income. These three functions have been applied to two rather arbitrary assumptions as to the proportions of public good costs that are allocable between pure public goods and specific public goods (see Table 1, notes b and e). TABLE 1 Net Fiscal Distribution: Puerto Rico, 1970 (dollar mounts per family) Conventional Asauaiptlone" Featly Incoee Class

A•

Lees than ><1,000

1918 2277 1887

$1,000-*1,999

1641 1907 1547 1523 1711 1352

*2,000-82,999 $3,000->➢3,999

1054 1184 533 578

804

Lowb uantlty Pure Public Goode

h.t6hc.civantity Pure Ptt ].ic Coodg

1 Y

1/Y1.5

iLl

1 Y2

1ír 1.5

2455

2322

2375

637

414

474

798

678

723

697

549

614

283 -65 301 -194

175 -511

252

240 -337

-77

$7,500-$9,999

176 -354 -57 -466 -816 -960 -746 -954-1441 -1486 -1526

*10,000-$12,499

-1259-1633-2169 -2304 -2198

*12,500-$14,999

-1894-2445-3074 -3165 -2818 -2750-3629-4591 -4756 -3209

$4,000-!;4,999 $5,000-$7,499

$15,000-$19,999

-41

$20,000 and over -3773-5129-5684 -5214 -2377

-428

33 -578

-267

-875 -1472

-23 -604 -278 -534

-269

-2205 -2942

-605 -329 -991 157

-277 -198

-4007 -3968

-1485 3905 -706 9633

1223

-272

3949

SOURCE: Data in first three columns from Mann (1973). NOTES: a. Under A, pure public goods benefits were distributed according to "broad" Income; under 8, they were distributed according to the number of families in each Income class; under C, the "analytical" budgetary deficit was eliminated by distributing it according to the overall tax burden, and the pure publh goods were then allocated as in A. b. The "low" quantity of pure public foods corresponds to the original study assumptions—l.e., 25% of expenditures on education and all expenditures on general government, police, and economic development. C. The "high" quantity of pure public goods corresponds to the "low" quantity assumptions in addition to the following specific goods cost proportions assigned as pure public goods: education-0.8; health and sanitation--0..8; welfare-0.5; labor relations--0.5; agriculture—0.5; highway construrtlon—0.5; housing--•C.6; interest payments on public debt--0.2; social insurance—{1,5.

(121 PUBLIC FINANCE QUARTERLY

The Aaron-McGuire approach may he summarized as follow,. (1) Y' = average family income (net of the personal income tax) In the ith income group. (2) SGB' = average specific goods benefits to a family in the ith group, as calculated by the two series (low, high). (3) P = total pure public goods expenditures (low, high). (4) D' = Y' + SGB' = net disposable family income in the ith group. (5) M(D') = marginal utility of disposable income Di (M is selected by independent means). (6) PB' = P/ [M(D') E' /M(D'{)] (k taken over all families). (7) NB' = SGB' + PBi — Taxi = net family benefits in the ith group. These are the data presented in Table 1. Use of the marginal utility of income function 1/Y does not appear to vary drastically the distributive pattern of net fiscal benefits among family income classes; the only marginal change its use generates is that, under "low" quantity assumptions, net benefits turn to net burdens in a lower income group. Much the same can be said with respect to functions 1/Y2 and 1 /Yl .s under the "low" quantity premises. However, combining the latter two functions with a "high" quantity of public goods does significantly affect the net benefit pattern, yielding net fiscal benefits to upper income families while reducing the real incomes of middle income families. Moreover, the net benefit accruing to the upper income families appears to grow in both relative and absolute terms as income levels increase. These results do not generally differ significantly from those applicable to the United States found in Aaron-McGuire's Table HI and Maital's Table I. This, of course, is not at all surprising, since we have used their marginal utility functions. Thus, the differences may be accounted for by the distinct character of the Puerto Rican tax and expenditure structure,: but by the same token the resemblance attests to the similarity of those structures.

Mann, Perlis / UTILITY FUNCTIONS 1131 THE EXTENDED APPROACH

In Table 2, where we have carried out the process of distributing the costs of public goods to a further degree, the statistical results do offer interesting differences (relative to both the Aaron-McGuire—Maital results for the United States and to those in our own Table 1 j. The same functions employed to generate the Table 1 data are used to calculate net fiscal benefits under an alternative approach in which specific public goods benefits, weighted according to utility, are distributed in a fashion similar to the benefits of pure public goods. Referring to our mathematical summary of the Aaron-McGuire approach, in the altered version we take: (4) ¡Y = Y' instead of Di = Y + SGB'. Then, in (S), M(D) becomes M(Y'). We then convert (6) into: SGB' B' = P/[M(Y4) E 1/M(Y' )l + S M(YI) E SGBk tc M(Yk} TABLE 2

Net Fiscal Distribution—An Extended Approach: Puerto Rico, 1970 (dollar amounts per family) Lowa Quantity Pure Public Goode Family Income

Clase Lees than $1,000 $1,000-$1,999

1/Y2

1/y1.5

237 145

-91 -270 -355 -431

$2.,000-$2,999

297

$3,000-$3,999 $4,000-$4,999

257 -173

$5,000-$7,499 $7,500-$9,999

-556 -747 -1433 -2009 -1156 -3054

$1 ,000-$12,499

$12,500-$14,999 $15,000-$19,999 $20,000 and over

Highb Quantity Pure Public Goode IL!!

-19 -130

69 93

-103 -291

-69 -172

-81

153

-694

164 46

-437 -542 -681

-210

-96 -401

-886 -750 -943 -733 6823 5042

-670 -602 -1026 -1202 2692 479

42 -76 -326 -563 -103 128

-613 -416 -45 744 7019 11973

-237 -86 42 362 3526 5494

SOURCES: See source for Table 1. NOTES: a and b above correspond to notes b and c for Table 1.

1/Y1.5

ji!

-230 -315

[141 PUBLIC FINANCE QUARTERLY

which denotes the total imputed benefits (S is total government expenditure on specific public goods). Finally, (7) now reads: NB = B' — Taxi, since SGB' is already included in B. Such an approach may be justified on the grounds that specific good benefits are not usually given in the form of direct cash payments (with the exception of some welfare payments), and hence may be translated into cash only via utility. Only income itself (i.e., cash income net of the personal income tax) is utilized in the marginal utilities. The average specific good expenditure on each family in each income class is calculated (under assumptions of various distributive series in each of the major functional expenditure areas: education, health, housing, and transportation, but excluding welfare or other cash income), and this expenditure is then translated into a cash benefit exactly as if it were a pure public good. In other words, the pretense is that the actual government service received by income class i is received by all income classes, and the imputed benefit to class i is computed precisely as in Aaron-McGuire. Such an assumption, though clearly false, would seem in no way to affect the actual benefit to those families genuinely receiving it. Thus, this technique would appear to be as sound as the basic Aaron-McGuire idea of distribution via utility upon which it is founded, and seems more appropriate than their implicit assumption that specific goods are valued by families at direct cost. Of course, total benefits over all families do not have to sum to (approximately) P + S — T (public expenditures minus taxes), for there is no reason that they should; subjective expected benefits, which utility may be interpreted to measure, need not necessarily correspond to proportional cash flows. However, we do "normalize" the benefits by supposing the total "value" of public goods equals its cost, which is the assumption Aaron-McGuire n lke. The results of applying our extended approach to the estimation of the net fiscal distribution, as presented in Table 2, demonstrate striking comparisons to those of Table 1. Under

Mann,Perils / UTILITY FUNCIIONS(151

marginal utility function 1/Y, the low income groups receive substantially reduced net benefits from the fiscal system's operation, whereas the upper income groups enjoy a much decreased net burden; little change is observed in the middle income groups ($5,000-$9,999). Under function 1/Y2, the alternative approach in both high and low columns produces negative benefits for the lower income classes, while providing larger positive benefits for the upper income categories. Under the Maital (l/Y 15 ) function, which might well represent the most plausible or "moderate" case (for marginal utility functions of the constant elasticity of substitution type), governmental provision of a "low" quantity of pure public goods creates net fiscal burdens for the lower income classes and benefits for the upper categories, but to a lesser extent than the 1/Y2 function. THE CONSUMPTION APPROACH

Two major faults can be found with the Maital function. Firstly, it is based on a rigid constant elasticity of substitution (of public goods for income) a umption, which appears to have been arbitrarily chosen for its mathematical convenience alone. The second defective aspect follows from this, for his function therefore has the form BY; the economic significance of this is to postulate infinite marginal utility at zero income, a truly implausible premise.3 For example, if one ounce of soup per day will not prevent starvation but ten ounces will, then it hardly seems reasonable that the ten ounces are "worth" less than ten times the one ounce. Indeed, it might well be argued that marginal utility should increase slightly as a movement is made away from zero. Furthermore, if we distribute goods via a marginal utility function that is infinite at zero, we obtain infinite benefits at zero income, and this certainly cannot be what Aaron-McG•iirc and Maital intended. In an attempt to over.,;me these deficiencies, we have tried to "reconstruct" utility from data on spending behavior, making, as much as possible, only economic (rather than

116) PUBLIC FINANCE QUARTERLY

mathematical) assumptions. We consider utility in terms analogous to physical mechanics. Thus, total utility (U) can be conceived as a "potential" and marginal utility (U') as a "force"—i.e., a change in potential U produces a force dUJdY that is realized as a change in spending. Therefore, U now takes on cardinal significance, indicating not only preference but the extent of action taken toward the preferred good. To be sure, this does overlook important factors, especially that of the satisfaction expected from savings (for security or progeny); savings could be included as a sort of "inertial" term, but (worthwhile as it might be) the model then becomes correspondingly difficult to handle. For this reason vie have ignored the utility of savings and in this paper simply ,point out that further study of this problem might be beneficial.4 Moreover, there occurs an unjustified mixing of an implicit time variable with the income variable Y, but it can be imagined that families "utilize" additional income by increasing consumption according to the "force" to do so—i.e., U'(Y). Thus, we simply define U'(Y) = E'(Y)—i.e., , the marginal utility of income is equal to the marginal propensity to spend. A curve based on this postulate does empirically have the TABLE 3

Net Fiscal Distribution—A Consumption Approach: Puerto Rico, 1970 (dollar amounts per family) Pure ?yblic :.00de Via Dtil tv i.or usntity High" Quantity Public Goode Public Goode

Partly Incoes Clara Leas than $1,000 $1,000-$1,999 $2,000-$2,999 $3,000-$3,999 $4,000-$4.999 $5,000-47,499 $7,500-$9,999 $10.000-$12,499 $12,500-114.999 $15,000-$19,999 $20,000 and over

2611 966 800 373 -371 -891 -1707 -2604 -3491 -4286 -5084

1661 1102 808 473 -26 -358 -1184 -'882 -2427 -2866 -810

All Public Goo ep V a 1)t1115!

Lows Quantity 11L€hb Quapfty Public Goode

Public Goods

2609 949 718 275 -447 -936 -17411 -2584 -3360 -3495 -3846

1655 1088 782 440 -56 -376 -1216 -1879 -2418 -2070 -366

above

SOURCES: See Table 1 source. Additionally, wore sheets from the Puerto RICan survey of family income and expenditures were used. correspond to notes b and c for Table I. NOTES: a and

Mann,Perlis / UTILITY FUNCTIONS [17J

classical shape expected for a marginal utility function (although it should be noted what was previously stated with regard to a more sophisticated model incorporating the utility of savings).5 Then, using E'(Y) under the Aaron-McGuire approach, we derive the corresponding redistribution of income. Since E(Y) is not given a priori in analytical form as in the case of Aaron-McGuire and Maital, but is instead given in tabular form as taken from empirical data, we used a least-squares technique to fit a polynomial to the data.6 The results are shown in Table 3. Comparing the new data of Table 3 with those in Tables I and 2, it is found that the net fiscal patterh under. our "consumption" approach approximates most closely that of the Aaron-McGuire 1/Y function. Most striking is the distinct contrast with the Maital approach, which yields quite opposite results. Moreover, in comparison with the conventional approach (see the Cst three columns of Table 1), the results for both our "extén d" approach and our "consumption" approach with a "high" quantity of pure public goods demonstrate a much-reduced magnitude of income redistribution from high to low income classes. A GRAPHICAL SUMMARY

Figure 1 performs the function of visually contrasting the different utility approaches to the measurement of net fiscal incidence. The curves refer to effective net fiscal incidence rates and are computed by simply taking data from selected columns of Tables 1, 2, and 3 and expressing them as percentages of average family income in each family income class. Curve AA is a graphical exposition of the second column of Table I and as such is representative of the pattern of effective rates generated by conventional analysis: in this case pure public good benefits are distributed proportionally to the total number of families, an assumption tantamount to presuming constant marginal utility of income. The curve, not unexpectedly, continually descends.

A(26E) E* (194)

óc _ g 50

B'tx

CODE: -- - -

1 ~

40 3-i ~ 2310-• 0, '

~ `

E£

~ #

M BB

~k ~t et ~

_ -10 D

_.:._.`

' 1--

rD ~B

iY: E

P ...-.`.~

-30 -40 -50 2i

Inoome(51,000) Figure 1: Net Fiscal Incidence Rates

X26 and over

Mann, Aerli, / UTILITY FUNCTIONS (191

Curve BB, corresponding to the 1/Y' -5 function with a "high" quantity of pure public goods (Table 1), bears the appearance of a "skewed V." The shape of curve CC (function 1/Y~- 5 and a "low" quantity -Table 2) is not easily characterized, but falls entirely below the "zero redistribution line" until high income levels. In diametric opposition to AA, curve DD (function 1/Y1 .5 with a "high" quantity—Table 2) reveals constantly ascending rates. Finally, curve EE, corresponding to the second column of Table 3 (our "consumption" approach), behaves quite similarly to curve AA with the significant exception of the two upper income classes, where it ascends instead of continuing downward.

CONCLUSION

We have attempted to develop further the Aaron-McGuire contention that the income redistributive effects of public budgets depend on implicitly assumed utility functions. By distributing all public goods via utility, we have demonstrated that the apparently strong "pro-poor" tendencies of most fiscal systems that emerge from conventional analyses may be even more misleading than were shown by Aaron-McGuire and Maital. Whether in fact the tendency is toward or away from greater equality will not be clear as long as the empirical basis of utility itself remains clouded. We hope that future efforts to find appropriate economically based utility models will be fruitful. As an added inference, it might also be concluded that the undertaking of further conventional fiscal incidence studies is generally subject to sharply diminishing returns, and, as has been pointed out by others, time might be better spent analyzing the incidence of limited and specific public expenditure programs.

[20[ PUBLIC FINANCE QUARTERLY

NOTES I. For two of the most recent examples, see Musgrave et al. (1974) and Reynolds and Smolensky (1974). For an erceikn% critique of those relating to underdeveloped countries, see Bird and De Wulf (1973). 2. As a corollary of the lxec:eding analysis, it should also be noted that SGB' is no longer considered to represent disposr oIc income in the Aaron-McGuire sense. This provides a way to assess, in cash terms, the utility of market items. We imagine the government to be supptying the goon/ to all families in equal "utils" and then apply the Aaron-McGuire procedure to this new "pure public good." Admittedly, with some items ii may be quite difficult to estimate the number of "oils" provided by specific goods to those families in different income categories. Given this limitation, it is very tempting to assume along with Aaron and McGuire that most items considered together enter all persons (or families') utility functions similarly, however dubious such a premise may be. 3. To demonstrate this, let i MRSpy/

i p/Y

MRS ,/

ply

where MRS is the marginal rate of substitution between public goods (P) and income (Y). Then, MRSpy, ~ 4 n PIY -

PIY. ' MRSpy

If P/Y = x, then 4 _ MRS` • x¡MRS _ xlog (MRS). fj = flog` (MRS) C + 4logx

log MRS dU /au Kx = MRS f T = A ñY PY óU

_ cYx~~ _ ;. (PJY)`~ = BY~

oY 4. One result generated by the inclusion of savings that the tótal expe.iditure uIrvc is apparently not concave. but that it may take the following fot,n'

Mann, Perlis / UTILITY FUNCTIONS (211

expenditure

45O

Income Figure 1a it would be of interest to examine marginal propensities to consume for very high income classes, in order to see if 1/2 is the limiting tendency. 5. Indeed, our data even show the slope of E(Y) increasing slightly for incomes dose to zero. 6. Near the last data point (Y = $28,000), the derivative of the polynomial was badly behaved since no further data points were available to affect the least-squares fit; hence, we used the polvnomilid as an analytic form of G(Y) only for Y C $28,000, and we assigned, purely by inspection of the polynomial's derivative, the value of 0.2 to G'(Y) for Y - $26,000. We are grateful to Drs. John Comeila and Ramon Sarraga for providing and implementing a curve-fitting program from the Arecibo Observatory for use on our PDP-10 computer.

REFERENCES AARON, H. and M. McGU1RE (1970) "Public goods and income distribution." Econometriea 38 (November): 907-920. BIRD, R. M. and L. H. De WULF (1973) "Taxation and income distribution in Latin America: a critical review of empirical studies." International Monetary Fund Staff Papers 20 (November): 639-682. MAITAL, S. (1973) "Public goods and income distribution: some further results." Econometrica 41 (May): 561.568. MANN, A. J. (1973) "La carga de las contribuciones y los beneficios de los gastos publicns, Puerto Rico, 1970 (The tax burden and public expenditure benefits] ." Study prepared for the Puerto Rican Tax Reform Commission.

[221 FU13L:C i 1"A.yC3 A., 1'.. I). Cr_. ar '. il. LTL1A' `.,, ..:) Y. . j "Thc distzitutior: fiscal burdens and ` en..tits." R.n&c _nx.nce Q. 2(July): 2~ 2-3:1. REYNOLDS, M. and E. 5Mfl1.EN5KY (1974) "Pn.e ppst flse distribution: 1961 : 1970 compared." wstionas Ta\ .!. 27 ;T"cember): 5,5-530.

Ecanor+a:' ^ at the University of Arthur J. Munn is A.rsocii, Proles: or Puerto Rio at Maurguez. Ií'e has served as consultant to serem! -'rrerto Rhin gover'anent commissions and alter*eies, the government of the l)onrinican Republic, the Organization of A n.a-itsrn .ti'_atee (!J(A. ad private industry. Don 'r! Perlis is currently Asst .ant Pruf~sso' (A:...'"-narras at the University of . . Rion at Mayaeuex fle re ,; _i .his Ph,12. ' "'rnaties from New York ay ü. 1973. f. r'rn'r Inlere"ts inc1u i . ' iQticoT Iogie, the philosoph; of relence, end nwtl,e,. .:. "s.

rub1 H~iones de la Unidad _de. Trxe; Ligaciones Económicas

TEMAS SOBRE ECONOMIA DE PUERTO RICO Serie de Conferencias y Foros Núm. 1:

José A. Herrero, La economía de Puerto Rico: E1 presente crítico, noviembre 1976.

Núm. 2:

Jaime Santiago, Impacto de la recesión en Puerto Rico, febrero 1976.

Núm. 3:

Jaime A. Santiago, Objectives and Strategy for Economic Development: A Challenge, julio 1976.

Núm. 4:

Jaime A. Santiago, Conferencia sobre economía y población, septiembre 1976.

Núm. S:

Rubén A. Vilches, Current Labor Force Perspectives & Problems in Puerto Rico, octubre 1976.

Núm. 6:

Jaime A. Santiago, La situación presupuestaria de 1976 y 1977 y sus efectos sobre la economía, diciembre 1976.

Núm. 7:

Elías R. Asón, Comentarios sobre las recomendaciones del uso de los fondos de las corporaciones 936, mayo 1977. Serie de Ensayos y Monografías

Núm. 1:

Alejandro Asmar, Análisis breve del Informe Tobin, junio 1916.

Núm. 2:

Samuel Torres Román, La problemática económica de Puerto Rico, diciembre 1976.

Núm. 3:

Angel L. Ruíz, Enfoque de insumo-producto para determinar una política de sustitución ca importaciones en Puerto Rico, mayo 1977.

Núm. 4:

Angel A. Rivera, Modelo de coto-beneficio para la evaluación de proyectosque sustituyan importaciones, junio 1977. Serie de Tema, Relacionados

Núm. 1:

Walter Bruckman, Diseño y método de investigación en la ciencia económica, abril 1977.

Núm. 2:

Stephen J. Hiemstra, Persp~rctives on Food Program Research, mayo 1977.

Núm. 3:

Angel L. Ruíz, Las conferencias cimeras de Rambouillet y Puerto Rico, mayo 1977.

Evaluation and

(Continúa al dorso)

TEMAS DIVERSOS DE ECONOMIA

Núm. 1:

Fuat y Suphan Andic,,Public Finances .World_, d'ic.embre 1976.

Núm. 2:

Paul W. S_rassman, Housing Priorities in Developing Countries: a. A Plánniq Model enero 1977.

Núm. 3:

William A Price

Núm.

Angel L. Ruiz, La recuoeración.económica' norteamericana y perspectival para 1977, iuni_o 1977.

Núm. '5:

J. Vane , Selfmzna~em.ent, [1o. kers' Management and Labor ManapZement ín Theory, and ?ractice__A Oomparctive Stud julio 1977.

Núm. 6:

Trade Doctrine and J. Vanek, The Absurdity. of : he Rich Mc. Institutions for the '?resent I.ny 1or?.d Economy and an Attempt to Reformulate, rgocto,l977.

i Optiiaa1jy. Baumol, . em, marzo 19/7.

Development, and the Third

The Price We Must Pay for

SEPARATAS Núm. 1:

Richard A.

er, The Econo-iic Approach to Law, 1976.

Núm. 2: $ichard.A. Musgrave, E' ., OT.

std S.B.T., 1976.

Núm. 3:

Fuat M. Andic y Arthu; Mair., Secular Tendencies Irt the 'Inequality of Earnings in Pt'erto Rico, ieb".ero 1977.

Núm. 4:

Arthur J. M n, Public Expr diture Patterns in the Dominican Republic and Puerto Rico, 1:30-1970, septientire 1977.

Núm. 5:

Arthur J. Mann, The 'Fiscal ter i end Income Distribution: Case of Puerto Rico, septic.:Lre 1977.

The

Núm. 6: Arthur J. t'ann y Pcecld Pe_lic, Utilit~Functions, Public Goods, and Income Distributioa, cepti~mbre 1977.,