DISCUSSION PAPER Report No.: uRD':? 2 OIL w'1~DFALLS I~ A CO~TROLLED E(O~O~iY: A "FIX-PRICE" EQUILIBRI1.;"!i A.'iALYSIS OF ALGERIA Patrick Con~ay and Alan Ge'b 1!-._ _ _ _ _ _ _ _ _ _ _A_U_gus_t_l_9_8_4_ _ _ _ _ _ _ _ _ ~ Development Research Department Economics and Research Staff World Bank . • .. 'f - . • • • The views presented here are those of the author, and they should no~ be interpreted as reflecting those of the World Bank . .-.;. . Oi~ Windfalls in a Controlled Economv: . ~\~price· Equilibrium Analysis of Algeria .b! . Patrick Conway and Alan Gelb ,. University of North Carolina at Chapel Rill and Development Research Department, World Bank August 1984 'f ... - • - Disclaimer The World Bank does not accept responsiblity for the views expressed he~ein which are th03e of the author{s) and should not be attributed to the World Bank or to its affiliated organizations. The findings. interpretations. and conclusions are the results of research supported by the Bank; they do not necessarily represent official policy of the Bank. The designations employed. the presentation otmaterial. and any maps used in this document are solely for the convenience of the reader and do not imply the expression of any opinion whatsoever on the part of the World Bank or its affiliates concerning the legal status of any country. territory. city. area. or of its authorities. cr concerning the delimitation of its boundaries. or national affiliation. We are grateful to Nicolas Mathieu and Sweder van Wi1nbergen for cocments and suggestions on various drafts and Shahrzad Goharl for computer programming • • Abstract Algeria ir. distinctive among developing countries for its highly controlled, planned economy and the overwhelming weight of its public sector in produetion. Because of these char3ct~ristics the Algerian response to increased oil revenues after 1973 differed considerably from that of more market-oriented 011 exporters, in terms of sectoral evolution and real exchange rate movements. This paper reviews the Algerian development stratp.gy and develops and estimates econometrically a fix-price equilibrium model of the Algerian economy. The model emphasizes spillovers of purchasing power from rationed wage goods onto domestic financial assets as the major equilibrating variable. The results are shown to account for Algeria's unusual adjustment pattern after 1974. They also have implications for a transition towards a flex-price system. - • - of Contents Table _ Page No. I. lntroduction .................................................... 1 II. Origins of the Fix-Price Economy ard the Al~erian Dp.veloPQent Strategy ••••••••••••••••••••••••••• L • • • 5 III. Th~ Response to Higher Oil Prices .............................. 9 lV. A "Fix-price- Equilibrium Hodel for .Ugeria ................... 19 V. Econometric Resultb .........................................., .. . 34 VI. Conclusions .................................................... 46 References ........................................................... 50 Appendix 1: The Individual CClOsUlller's Decision Appendix 2: List llf Tables Table la ...................... Public Investl!le!lt Programs by Sector 12 Table Ib Public Industrial Investment .............................. Table 2 Estimation of Consumption and Real Demands ........ Balanc~ 44 Table 3 Ra~ioning in the Goods Market ............................. 44 List of Figur.!!,. • ngure 1 The:Jll WIndfall and Its Use: 197;·-81 2 Figure 2 Currency. Inflation and Seigniorage: 1970-81 •••••••••••• 16 Figure 3 Fix-Price and Walrasian Equilibria ........... . ............ . 28 Figure 4 Updating Administered Prices ................. .. .•......... " 40 I 011 Windfalls in a Controlled Economy: A "Fix-price" Equilibrium Analysis of Algeria by Patrick Conway and Alan Gelb University of North Carolina at Chapei H11l and Development Research Department. World Bank I. Introduction Half of Algeria's exports came from 011 in the early 1960s. In the aftermath of the prolonged war of indp.pendence which B~erel.)I dLsrllpted the. nono11 economy, this proportion had risen to three quarters' b,.~t-he'E1!ft".l1 · 19705. In 1972 ~il exports were equivalent to 16% of GDP. The quadrupling ~f h eir world 011 prices over 1973-74 and t - redoubHn~ over 1979-80 thereforp conferred large windfall gains on Algeria as it did on other oil eXl~rters. Taking into ar.count both price and volume changes, the windfall, expressed relative to the size of the nonmineral economy in Figure I, peaked in 1974 at the equivalent of 42% of nonmining GOP.!! It then declined to 26% and later to 18% as world oil prices eased and the volume of Algerian hydrocarbon output grew more slowly than its ~noil economy, before rebounding to 34% ~th the' second oil price rise. - • • ­ , 1/ Underlying Figure 1 is a hyrothesized projection of the Algerian economy in balanced growth as between mineral (overwhelmingly oil and gas) and nonmineral components at relative prices set at their 1970-72 base levels. The Figure depicts the windfall as the proportional deviatiou of hydrocarbon value-added from its balanced-growth trend. Figure 1 The 011 Windfall ond Its Use: 1973-81 Shore d Noo-Mlnlng GOP 0.3 0.2 0.1 O~~~~~~ ____ ~ ____ ~ ____- L____-L-­__ ~ ____ ~--~ .~ 1Q~n '73 '74 75 76 77 '78 1981 Unweighted Mean of Comporctors: 0.5 Ecuador. Indonesia. NIgeria. Trinidad ond Tobago ond Venezuela 0.4 . 'f • O~~~~ ____ ~ ____ ~ __ ~-L ____ ~ ____ ~ ___ -__ - ~!L- ____L-__ ~ 19~n '73 74 '75 76 78 '79 '~ 1981 While the relative size of the oil windfall was not untypical for a capital-importing oil-exporting developing country, Algeria is exceptional among developing countries because of the highly controlled nature of its socialist economic system. All prices and formal wage scales are administratively determined. A system~f Treasury subsidies to public enterprises essentially removes the constraint of even administered market forces on firms, which therefore face a ~soft" budget constraint.!! Private enterprise has been marginal outside a part of agriculture ~nd certain minor urban activities. The state has a virtual monopoly on the allocation of investible resources through its control of the banking system and the Treasury acts as a maj or financial interiDec:ft·al~:y~, ,~,a;:1~.:,t~~ea,.,~&~~..,;-:-~U\:iii'i official channels only, imports are cont~,~n~ fo.'· .2ign excha.-!l~i :·nc.!J&htly--· allocated. While "black markets" do u~d~ubtedly exist, by all accounts they are minor. Despite moves after 1978 to liberalize and decent.alize production and free certain markets (notably in agricultural products), Algeria is still well characterized as ha~ing a socialist, centrally planned "fix-price" economy 1i The broad outline of h\.·-., a market economy should respond to increased domestic spending out of oil (or other) windfalls is well established and follows the analysis of Bruno (1975).11 Relative prices playa major It equilibrating role, particularly the price of nontraded relative to traded ~ ... 11 Kornai (1978) analyses the "soft" budget constraint of the production .. 'f sector in the context of Eas~ European socialist countries. - .­ 2/ • "Fix-price" does not imply that prices are constant; it rather means that price variations are not a major endogenous mechanism for clearing • ... markets. The terminology is due to Hicks (1965). 3/ The model has, in fact, been known since the ninetenth century when it was applied by John Cairnes to analyze the impact of gold discovery on Australia over 1851-73: See Bordo (1975). goods (the real exchange rate) which appreciates to shift factors of production to nontraded sectors and deflect demand onto nonoil exportables and imports. Such mechanisms cannot operate in a fix-price economy, which must substitute various rationing mechanisms (either on the demand or supply sides of individual markets) for the process of clearing through relative prices. If these mechanisms can be enforced, there is no presumption that the evolution of major price and resource indicators will iollow the free-market paradigm. If, however, costs of enforcement rise or its effectiveness declines as the degree of divergence between the rationed and unrationed equilibria increases, it is likely that the state will make some adjustments in the setting of wages and prices and the allocati'on -of- frivestible l:eso~ft albeit reluctantly and with a lag. This paper consi~ers the Alge~ian government's response to the oil -windfalls and the consequent evolution of the nonoil economy. Row has the Algerian strategy differed from strategies characteristic of more mar~et- oriented economies: How has the nonoil economy equilibrated in response to changing levels of public activity", and has it been able to avoid the "Dutch Disease" characterised by "overvalued" real exchange rates, contracting nonoil tradeable sectors and large resource shifts towards construction and services? .. To comprehend a development strategy so, consistent as that of Algeria it is necessary to appreciate the historical and institutional roots of the fix-price economy and the role of t~ public sector. Section II outlines "f - these, and provides a set of indicators describing the structure of the .. - economy beforO! the first oil price Ibcrease. of the strategy are brought out in Section III: The main, distinctive elements they reflect (a) the priorities of the Algerian government and (b) the natural resource endowment of the countty. The section also analyses the evolution of structural indicators after the oil price increase. and notes the unusual development of the controlled Algerian economy by comparison with other oil-expor:..ing countries. Section IV develops a fix-price equilibrium macroeconomic mod~l of the Algerian economy. tailored to refiect its institutions and the objectives of government. The model is used to investiga:~ the importance of rationing in the goods market, and the consequent involuntary holding of cash balances by the private sector, liS an explanation of the Algerian pattern. In ~ection '.J the model is estimat.ed econometrically using techniques developed b" Fair and Jaffee (1972) an.! Amemiya (1974) to incorP.Q.rate .. -splllove~s.~uC"'pCr·cha6tri'K ­ ,,' ~ - ' .' power due to the rationing of conlJumer demand. The results are of . f..~t-e[,I!-s-t: . for two reasons: they offer evidence ~f significant -spillover- effects on demand for real money bslances, and they provide coefficient estimates for a price-responsive model of Algerian demand behavior. These, in turn, permit an assessment of how Algeria might have adjusted in the absence of rationing and price controls. Section VI offers concluding comments. II. Origins of the Fix-Price Economy and the Algerian Development Strategy. The Algerian development strategy was first enunciated during th~ II eight-year revolution which led to Algerian independence, and its substance was shaped by the bitter and costly nature of th~ war. Although th~ Algerian leadership has been characterized by factionalism reflecting differences in political outlook, generation and role,l! the pssential e.ements . of 11 For analyses see Quandt (1970), Zartman (1975). development policy appear t~ have been accepted by almost all groups. These I elements can be summarized in tr:ree general principles: .lI (1) Algeria should develop from a ~ocialist perspective; income equalization was to be encouraged across regions and individuals. Growth and empl?yment creation were important objectives. (2) Economic poUcy shodd promote self-reliance and minimise long­ run dependence on foreign financiAl and technical assistance. (3) Economic planning was t~e responsibility of the State but should include SGw~ participation by workers. The economic history of post-il;dependence Algeria can be divided into three phase3. The constitution of Sept~ber 1962 established Algeria as a one party socialist republic vi th Ahmed Ben Bella as president. After three .yeus. he was toppled in a bloodless coup ·which placed Houar·t Boumedierre (l9~5-78) in the presidency. Boumediene was followed by ChadliBenjeG'!. Each president implemented a different policy regime, but all are best considered as ~~phasising different aapects of the three general principles. Ben Bella adopted a pragmatic approach towards the major goal of reconstituting the econo~ after the revolution. He focused on the promotion of light industry (especially consumer goods) and on agrarian reform in the wake of a massive exodus of French colons during and after the revolution. Management of the abandoned industrial and agricultural firms was entrusted to • r workers (the autogestion pro~ram). An assessment of Ben Bella's contribution suggests that economic performance was disappointing anci that problems of -4i chronic rural and urban labo~ surplus in particular had not be,n resolved. (Mallarde, '-1975». While estimates are unreliable, perhaps t7o-thirds of the • • • 11 The documents enumeracing the Algeriah ' principles I)f development art;;: Program of Tripoli, May 1962; ConBtitution, September 1963; Charter of Algiers, April 1964; National Cha~~~!~ June 1976; Constitution, November 1976; Three-Year Plan, (1967-1969); Four-Year Plan (1970-1973); Four-Year Plan (1974-1977); :ive-Year Plan (1980-1984). rural and one-fou~th of the urban labor forces were un-or under-employed.ll But the period carkci the emergence of "la nouvelle c'lasse" - a middle class of public sector bureaucrats. technocrats an~ ~anagers - ~hich ~as ~o play an increasin&ly important role in future years. Boumedicne's accession in June 1965 ~arked a major shift in Algerian develorment policy. Influen~ed by Francois Perroux and G. Destanne de Bemis, a strategy was formulated of industrialisation, integration of domestic industry and closing off the economy to foreign influences. Drawing on the work of Hirschman, "industrialising" industries - those vit~l strong back....ard and forw.ud linkages -- were to develop the economy 21, Heavy producer-goods industries (steel. petrochemicals w- !eT'Cl-rne'rsy · were believed by Destanne de Bernis to _promise the strongest linkages. The role of the hydrocarbon sector was considered particularly crucial: a !r industrialising industry·, it - pr~sed a source of investment demand, "premi- raw materials for downstream processing a.::t1vi ties and rt:"ource rents to fund industrialisation. Perroux's study of "growth poles" s~ge~ted that regional integration would best be achieved by establishing new industrial centres at Arzew, Sklkda and Annaba although they had scal! existing industrial bases. Finally, the objective of becoming a closed economy was heavily influenced by . the work of Samir Amin, which identified a dependent economy as one vith export and luxury consume::- goods Jgerian planners, but is abstracted from here. If, for example, aggregate demand were less than full capacity output, the constraint on labor would be the first term in parentheses. Private-sector Behavior. Private consumers are assumed to make their consumption and money balance holding decisions through intertemporal utility oaximisation taking ; \ ;' , . ·,.Jnt present constraints on labor demand and goods supply. This ·' .i~" . : ' is reformulated in each period t.l! Individual decis10n-1Uking is aggregated to obtain macroeconomic functions ror estimation. Consumption and money-balan~e holding choices for individual i will then depend on initial money holdings (mit-I)' present after-tax real wages w (I- Xit0-T t » and t expected future after-tax wages. U>. A utility-maximizing individual will have desired consumptf01t " ie.Uu~d by the following function: (5) where T is the time horizon of the individual and r the rate of time preference. The first term is consumption out of present after-tax income; (liT) is the fraction of total disco~~ted income to be consumed in anyone period, while . . 6 reflects the relations~ip between expecled future lubor income • and present realised income. The second term represents consumption out of . real balances • . 'f - . • - l! A mathematical presentation of the Appendix 1. deci~ion-mak1ng process is given in • 11 (vt/P t ) is the real annual wage in period t for a fully employed individual. Xi is the fraction of desired annual ~mployment which the individual spenas working. T is the ad-valorem tax on labor income. t Individual demand for real ~alances is determined analogously. Note that the two summed together equal total disposable wealth in period t: The individual has two choices for allocation of wealth -- consume. or save to the form nf money balances. Aggregate consumption and real balance demand behavi:Jr is found S ',. through sUMming across private individuals. There are L individuals 1n f " t L8 period t desiring full employment; t t xi t 18 t he -1'IV:lOO~ 9 L~~.P t . . .l' ~ 'Y ' i-1 employed individuals equivalent to perio~-t employment . .., S L (7) C - t «6/T)(w /Pt)(1-tt)X + «1+r)/T)(a _ /P » t t it it 1 t i-1 S L (8) Ht/P t - r (mit/p t ) - (1-6/T»(wtLt/Pt)\I-~t) + (1-(I+r)/T)H t _ 1/P t i-I Labor supply is assumed to be inelastic at any point in time, but to gro_ at a constant rate (n). 1! Unemployment iG the difference between labor supply and de~and u ­ ... .. 'f - - • ­ 1! In the Algerian case, extensive government control of em?loyment, wages and pri~es has reduced incentives to increased productivity and may have eliminated the poq~~~ility of a labor-leisure tr~de-off in the formal sector. The ass~~ption that the labor fcrce either works full time or is unemploy~d seems a reasonable approximation. (9 ) .. L e nt o Government. The government's operations are separated into hy~rocarbon and non­ .­ hydrocarbon compon~nts; each has associated e,.enditures and revenue3. Ye abstract froo hydrocarbon investments, which r~present essentially an effort • to stretch out the time profile of rents by effecting a transition from oil to nAtural ga~. Revenues in excess of ~xpenditures in the hydrocarbon sector (q E }l! are available for ~se by the government in the non-hydrocarbon t t sector. Public revenues also come from foreign borrowing (B t ), taxation of . w "'t. .. . . . t labor income (T ~ L ), public-enterprise profits (w t - Yt - (-pt,tt) a~ t t t money creation (at). Net hydrocarbon receipts and foreign borrowing ability are considered autonomous since they are mainly determined by external factors. Money cr~ation is the residual source of government financing, as defined below~ Government expenditures in the non-hydrocarbon sector are either investment (It) or co~sumption (G t ). The government investment program aies to eliminate unemployment. It responds to the availability of foreign exchange since a large proportion of capital goods is imported. In the simplest formulation: 11 Hydrocarbon production is assumed to be fonstant .~ . in physical volume for each period in the theoretical section •• Actual volumes are used for econometric purposes, with net revenues from the sector following from subtracting hydrocarbon investments. 2/ Internal bond financing, a minor feature of the Algerian economy, is assumed away. AGD/oil/8-2J-84 - 24 ).(0) .. 0 (10) More generally, it can be assumed that investment and the accumulation of capital creates its own dynamic, so that planned investment 1s related to the capital stock (whose rapid growth is, in the last resort, financed by oil receipts), but that fluctuations in the availability of foreign exchange cause deviations around that trend: (lOa) In both cases A is a speed of adjustment dependent on the number of unemployed. Investment therefore proceeds at a steady growth rate, Ail' except in periods of increased or decreased foreign financing. i1 represents the dependence "of the investment program on the availability of foreign exchange. As i2 increases, reliance on foreIgn fln~clng sources rises, so that fluctuations In those sources have larger repercussions on the level of f~vestment undertaken. Greater dependence coul~ ~ither be due to higher !~port intensity In investment projects. or greater complementarity betwee~ domestic and foreign capital goods; 'f ... Government consumption (G~) is assumed to represent a constant fraction (g) of non-hydrocarbon oull>ut • ... (11) G tot "" g + gY Mt -M t - 1 Money creation (St - )' is the final source of government M _ l t financing, and is determined residually by the excess of expenditures over other sources of financing 6S the budget deficit is monetised. AGD/011/8-23-84 - 25 ­ M 1 'It (12 ) (~) P • (G t - . t - T t P L t ) + It - (B t + q t Et ) -- t t The first right-hand term represents the current expendi~ure deficit; the second, spending for investment; the third. foreign exchange financing sources. In Algeria. current non-hydrocarbon government revenues have roughly equalled current govern3en~ expenditures. Capital investment expenditures have thus been financed mainly through fo.~.!ll~i "lX~ll"nge or IIIQNt~ .~ F.:.~~.i sources, and, to a degree. domes tic bot:r,Qv.i,ng • . The non-hydrocarbon balance of payments, which is also under exclusive government control, is then: (13) (x -1M ) + (B + Q_ E ) - 0 t t t ~ t where (X -1M) is the net export of non-hydrocarbon products. t t Aggregate Demand. Aggregate demand can be assembled from these elements: • One decomposition of this relationship (using (7), (10), (Ii> and (13» indicates both the dependencl! of aggregate demcnd on equilibrium values of Yt and Lt and the principal exogenous variables which induce demand: w M (14a) Yd _ ! (- t)(l-T)L + gY + (l+r) t-1 + h _ 'i + ' h t + ('i iiK h 2- 1)( qtEt+ Bt ) t T P t t t T P 0 t t AGo/011/8-23-84 - 26 ­ Increased real balances, planned investment and labor income all serve to stimulate aggregate demand~ The effect of (q E +B ), however, is ambiguous. t t t Through raising investment it stimulates Y:, but through firiancing imports it permits demand to be switched abroadr1i .~other decompOSition of (14) defines aggregate demand in terms of its sources: w (I4b) yd _ C + 1r + T ~ L + e (M Ip) t t t t Pt t t ~ L.x_, ~ The first right-hand side variable is private consumption demand; the second ~nd third are together government demand out of current income; and :he final component is government demand financed by the monetisation of the budget deficit • Non-Walrasian Equilibrium-Regimes. d .s d The levels of Yt' Y ' and L will depend upon the economy's t t equilibrium output and labor use levels. Four equilibrium conditions, or regimes, are possible. There could be excess supply of both labor and output d (Y t - ydtand L - L ); t t thi~ is called the regime of Keynesian unemployment (KEY). This could be excess demand in both markets (Y t .. yS t and L - Ls t); .. t this is the Repressed Inflation (REP) regime. There could .. b. excess demand for goods and excess supply of labor (y • Classical Onemployment (CLA) regime. t _ yS t and L t Finally, balance in bot'!\ markets - Ld ,:t; this is the " . Y This formulation differs from the conventional Salter-Swan model underlying most analysis of oil economies because increased foreign exchange availability is not allowed to bid up non-traded good prices; the government shifts domestic resources to non-traded good production ~nd meets traded-good demand through increased consumer imports. AGD/oil/8-23-84 - 27 ­ (y _ y~ _ yd and L - LS - Ld ) is the ~alrasian equilibrium (WAL)~ t t t t t t . - Thesc four regimes span the set of possible co~~inations of Y and L. t-\_l Transforming variables into per capita terms (m t - - ---s-' k t 1 L .­L K t S t t and so forth) permi ts , a diagrammatic examination of the regimes as in Picard (1983). There are thrce boundary conditions for the econo~ -- the combinations where one or both markets clear wi th,)ut r~tioning. ~alrasian equilibrium is characterized by unrationed equilibria in markets for both labor and output and a government budget constraint. 1mces Manipulation of these conditions yields the \lalras1.tn valul!.g...;.o f' real- ba!- and the capi tal stock sholo71l in Figuce 3: Y OS) k* ­ as (16) m* • T [.!..::s. - ! z* + (bt+qtet)] • (T-6 ) [.!..:B. + b + q e ] PaT 1-6n' a t t t These values are invariant with respect to time so long as there are no shifts in government poUcy or in external financing. The equiUbrium level of real balances depends negatively upon the government consumpticn decision (g); as government consumption rises, private consumption must fall through lower consumption out of real balances. An increased nonoil trade deficit (b +q e ) t t t .. not offset by increased public spending has the opposite effect, since it increase! leakages via imports of consumer goods. The variable z~ is full • employmetrt per capita disposable wage incolile. The second equality follows • 11 'rhe other combination -- excess supply or go~s and excess demand for labor -- could not occur in this model, which does not allow inventory accumulation. It also seems a remote possibility for Algeria. 11 Equations un~erlying the following discussion ~re provided in Appendix 2. - 28 ­ Figure 3 Ax-Price and Walrcsian equilibria m m" :t • - - • • - 'AGD/oil/8-23-84 - 29 ­ from the time-invariance of mt and P t ,!" Walrasian equilibrium. In steady­ state money creation occurs at the labor growth rate which imposes a long-run government budget constraint. The Walrasian output per capita is fixed. (17) .,* - a1 (18) pIp - 0 (l9) a- n Away from Walrasian equilibrium, the regime CLA is characterised by underemployment and Mtoo highM levels of real balances leading to excess demand for output; the regime KEY mayor may not be characterized by sufficient capital stock to employ everyone, but has insufficient demand to generate full capacity employment due to Mtoo lowM levels of real money balances. Regime REP represents excess decand for goods and excess productive capacity leading to excess demand for labor. The level of capital where the labor' force is fully employed recains as- 1 in non-Walrasian equilibrium, and defines the vertical boundary between REP and CLA. The level of re4t money balances per capit~ which sets supply .. and demand for product equal a: given prices, and separates REP and KEY, is increasing in,k as a result of!desired disinvestment. The diagonal boundary betveen eLA and KEY is defined by the condit.ion that increased capacity creates its own demand for output; as consumption is a fraction of output, real balances must be higher to ensure the equality. AGD/oil/8-23-84 The Government's Role. - 30 ­ I Government policy is an important determinant of the equilibrium regimes. l"i.rst, it controls the capital intensity of production through the planning process. Increasing as vill decrease the capita1 intensity of production; this would make full employment more accessible, by moving the Walrasian equilibrium point in Figure 3 to th~ left. (This could be effected by a shift in the mix of investment from highly capital-intensive sectors to less capital-intensive ones; for example, from cement production to textiles.) Second, it can regulate government consumption (g) or taxation (T) during the adjustment process. Raising g or lowering T causes incre«lred demand for product and a higher rate of mcney creation (9), leading to '· a lower Walrasian level of real money balances and a greater probability of excess demand for goods (the CLA and KEY regimes). Third, it can regulate the speed with which it pursues a goal of full employment through varying the planned growth rate of investment (H ); higher H 1 increases the slope of CLA-KEY and KEY-REP boundaries. l Fourth, the government can alter its demand for domestic goods by varying the dependence of its investment on foreign sources of financing. Increasing i2 represents an increase in reliance on foreign sources: for fixed qtEt+B , this will lead to increased aggregate demand as in 04a). It t vill also increase the variability of It around planned levels. The CLA-KEY and REP-KEY boundaries in Figure 3 become less ste~. • The government cannot, however, regulate money creation (6) • directly; that depends on its ftnanci~g needs which are determined by the instruments I, T. g and external factors. A strongly capital- and import- intensive development strategy corr~sponds to a lov aBo AGD/011/8-23-84 - 31 ­ Algeria: a Classical Unemployment E~uilibri~. Data are insufficient to test whether Algeria is i.n REP, KEY or CtA in Figure 3, or at the ~alrasian equilibrium.!! H~wever, the available evidence strongly suggests that the economy has been in CU, since widespread un- and under-employment has coexisted with excess demand for a broad range of consumer goods. This is closely linked to increased 011 income and .:fonHgn borrowing over the past decadE, as may be seen from the effect of higher net hydrocarbon revenues on the boundsries of Figure 3. By equations (15) and (16) the ~alrasian capital stock per head k* is constant but m* rises; with desired investment ra'ees I1m!ted !tr labor ~~e. growtl. 011 revenues imply a higher nono11 trade deficit for extet"nal" balane-e.· which must be achieved by rel4Y.t~~ import restrictions. To ~aintain aggregate demand at full-employment level higher m* is needed to raise consumption. As shown in Appendix 2, the CLA/KEY and REP/KEY boundaries become steeper and at some ievel of konses wer.e allowe,:i; these are neglected here. specifica~ion of the labor supply deCision, any unemplo)~ent also represents a level of leisure off the supply curve. Involuntary consumption and employment levels have i~plications for estimation arid interpretation of consumer demar.~ hehavior. In e Walrasian world, notional consumption and real balance decands are observed as functions of disposable income and the level of real money balances. In ~lassical unemplo~ent, supply-side factors and resource constraints dominate. Realised decands are observed as functions 01 government spending, productive capacity and net exports. Higher investment or current government expe~diture crowds out private consumption one-for one. Real balanca--a>.~~~m':lla~ion is partl.~ , · !nvoluntary; it will rise with public spE'nding and fall with rea! output. (29) This will increase with government expendi~ures. The investment program, rather than the -Dutch Diseaee-, determines the pattern of resource allocation. Sectors favored by j.\1blic investment grow; sectors serving rationed private demands shrink • • v. Econometric Re3uits. For reasons described above, we mainta.in the hypothesis that the . . ­ economy is in CLA and eXaIt~e the lmplicat!.ons econometrically. The above model defines a simultaneodS economic system. Estication is complicated by the fact that two endogenou~ • variables -- effective consumption and effective demand for money balances -- are unobserved. A technique by Amemiya (1974). is used to redefine these variables in terms of their observed, realised levels. Error terms are assumed multivariate and normally distributed. Estimation of the equation system is done by two-stage least squares <2SLS).l! Estimation has two purposes: (1) Determination of the accuracy of the fix-price equilibirum construct as applied t~ the Algerian economy; (2) Measurement of the impact of ratiooing in the · Algerian econocy, including both the size of the ~h~rtfall in ~vailable goods and the increased inflation which vould have been necessary to reach a flex- price equilibrium. Data. The data used in estimation are annual observations over the period 196 7 -1980, and have been drawn froa official Alger!3fl' ~C;j!8, 8-nc:lf>?". '·~:rlJ"" · Bank reports'; many of the time series are describ_ ed. in detail in Conwa~-- (1982). All variables, with the exception of the money supply and pl.'1ces, are in real terms, wi~h 1974 as the base year. Ml (currency + decand deposits) is used as the measure of the m~ney supply. Y The non-mining GOP deflator is used as tha index of the price level (p). Th~ U.S. consumer price index (P*) is used as a proxy for world prices. The non-hydrocarbon capital stock {K) is formed through the perpetual inveltory method from a base Ko' depreciation rate, and subsequent investment levels; Ko is found by positing an average non-hydrocarbon . capit~l/output ratio of 2 for 1967.11 The tax r~te (T) is derived through dividing non- hydrocarbon taxation receipts by non-hydrocarbon domestic incom~. Wage income II Only 14 annual observations are available for,all variables, which precludes the use of system estimation t~chni~ues. y In 1974, the components of Ml represent~d94 percent of t:1e c:om~: . I.:~ts of M2; in. 1980, 90 percent. 3/ 'Non-hydrocarbon f1gu!"e~ are approximated by use of official non-mining statistics. Hydr~carbons represent about 95 percent of the mining sector. --_._._---'""=''==:=:::..;.;, 1s based on time series for the public sector scaled to be compatible on average with national value ad~ed. Production Decisions. The model presented in section IV suggests simple 11n~ar forms for the production function and labor dema~' equations under classical unecployment~ Because OLS resi~uals are highly correlated, estimates ~re also given with first-order autoregression correction p: (a) uncorre.:ted: Yt • 1.03 + .268 ~ R2 • .99 (13.ln (35.25) ~ '...... ~tt~;·· ( b) cor,ected: "it • .97 .. .275 Kt -2 a: "" ".58 (8.29) (25.39) p • 0.53 (2.36) (a) uncorrect~d: L - 2.12 + .10 Ie . R2 - .94 (29.04) (14.35) tv - .92 (b) corrected: I.e - 2.\0 f" .10 Ke . (24.77) (12.52) p - .15 (.55) '.. 0 The equations have good explanatory power. Both intercepts are posi tiveand -.. significant, indicating an inp.lastic response of output and emplo1ient to •• investment. !I The point estimate of the incrP.mental capital-output ratio If This and all following statistical tests of significance are evaluated in a two-tailed test at the 95 percent level of confidence. 1 (6) as given by (1/a1) is 3.6 for the corrected estimate; thi~ is high by cross-country standards, "but expected in view of the capital intensity of Algerian production. b 1 represents the labor-capital ratio aB; the implied average estimate of elasticity of labor demand IoI1.th respect to output is 0.40. Technological progress ~rowth terms were tried in both equations as independent regressors, but were statistically insignificant. Government Demand. Policy functions for non-hydrocarbon investment and governme~t spending are drawn from section IV. Where residuals are highly der correction autocorrelated, estimates are also given IoI1.th first-or" p • (a) uncorrected It - .52 + 1.12 (qtEt+Bt) - .70 (2.79) (5.54) OW - 1.10 2 (b) corrected .86 + .78 (qtEt+Bt) i - .75 (3.61) (3.69) p - .40 (1.50) -2 .14 + .11 ~ + .30 (qt Et + Bt ) R - .89 (1.02) (4.87) (I.45) ow - 1;.61 .. "f In equat.on (32). the estimate of ~i2 demonstrates that t~ - - hypothesis that the level of investment res~)nds one-for-one to the contemporant!ous availability of foreign flnanc ing cannot be rej ected. Eqcation (33) separates investment into "planned" capital accumulation at rate AGD/oil/8-23-84 - 38 ­ Ail (itself related to the trend of q E +B over the period) and the, t t t deviations froo plan due to foreign financing fluctuations (Ai ). AI though 2 the magnitude of the coefficient and its explanatory po~er (8 value • .22) are considerable, it is significant only at the 30: level. This is not unexpected given the shift in investment policy implemented by the Benjedid regime after 1978. Introducing a dummy variable for 1979-1980 to capture this policy shift yields similar and more precise estimates: (34) I - - .08 + .14 K + .36 (q E +B ) - .60 I)IDC. · • ::;'6 · ~. :o.. .'-'. t t t t t . ,' (.73) (8.53) (2.66) (4.04) ~. - 1.8,1 Equation (34) suggests a planned investment growth rate of 14% annually until 1979. At that time, the dummy coef fic1ent suggests a 21% fall in inve~~~ent below trend follo~~ng the Benjedid policy revisions. This shift is also reflected in total investment, as sho'Jn in Figure 1. The estimate for Ai2 indicates roughly one··third of fluctuations in foreign financing are passed through to investment demand. As shown below, docestic involuntary saving in money balances provides the main alternative source of financing. Government consumption is treated as a linear function of non­ . . hydrocarbon output. which is treated as a simultaneously deternined variable in 2SLS: ... (35) G - go t 1 gY t (a) uncort,ected: • Gt - -.10 + (2.70) .23 Y (22.91) t R2 DW - .98 - 1.18 • -2 (b) corrected: Gt - -.10 + (2.24) .23 Yt (19.29) R - p - .97 .20 (.76) The estimate indicates a marginal propensity to spen~ from non-hydrocarbon domestic product of .23, and an elasticity of 1.17. AGO/oll/S-23-S4 - 39 ­ Private Demands. Rationing in the goods oarket implies that the observed levels of private consumption and holdings of money balances are realised, not effective. A method for estimating the underlying effective behavioral equations was first suggested by Fair and Jaffee (1972). In a command economy, prices are often adjusted to reflect the level of excess demand for goods in the previous period. The government in effect follows a slow tatonnement rule. Assuming that only. C'emt'OXllet'li'" ~'P'~ '-£:~W~~d: 1 - - (36) (P t - p - J ) - t y (C t - C t ) wherey is the speed-of-adjustment parameter. As y becomes smaller, price adjustment approa~hes that of an auction market; as it becomes larger, prices are increasingly unresIX't!sive to excess demand. If such a relationship holds exactly, the observed price change will serve au an indicator of the rationillg deviation between observed C and effective C. l! The government's price-updating equ~tion (36) may be used to identify effective demand (C t ): (37) .. - where C is specified in equation (23), and is observed only with a random t error. This relationship is illustrated in Figure 4. In classical Excess demand may be only one of a n-umber of causes of increases in prices. Another might be a desire ~ reconcile stable real and nominal . exchange ratas by -passing through- ~rld price increases to the domestic market. In this case, the left-hand variable would be «Pt-Pt-l) ­ (P~ -P~-l» and the expanded price adjustment equation would be: (Pt- P t-l) - (P~ - P~-l) + (l/Y)(C t - Ct ) This formulation has been used in some estimations. - 40 ­ Figure 4 Updatlng Administered Prices P,~------------------~ Po - - - - ~------~----__-.;..--.... ­ . c, ... c, c, c, 'f ... - .. • • •­ AGD! oil!8-23~84 - 41 ­ unemplo~ent in period I. ~nly CI may be available for private-sector consumption. The previous-period price (PO) creates excess consumption demand through its effect on real balances. Excess decand is giVen by (C 1- C1 ). ' The government then raises prices to PI according to the excess demand rule (36). Knowledge of the chan6e in Pt for all p~riods and the government's rule allows calculation of '- C I and the C schedule. t Effective demand for money can be treated silJlilarly. The observed lIIoney holdings then include both effective demand holdings (M t ) and involuntary saving. whose magnitude is measured ,.by . th.{L . ?c;,1J:, ~ ", a~justment. (38) ' . (39) Equations (7) and (8) of Sectt"on IV indicate the functional relationships between effective consumption and real balance demand a~~ their determinants. Combining these with (37) and (39) yields estimable ~xprp.~~!ons for realised consumption and real balance demands: (40) + e ct M (41) -:II + (l-fl)("'tLt(l-Tt)!P ) + (l-f )(- ~~I) - f (P -P - ) + e . o t 2 3 t t l mt - - • The cross-equation restrictions on coefficients demonstrate tie linkage between the two demands in the private-sector decision process. The individual has two options for the use of present after-tax labor income or past savi!',g: to spend on ,'onsumer goods or to add to saving held in real balances. The restrictions on the coefficients of labor income and lagged money holding reflect this choice by summing to one across equations. AGD/011/8-23-84 - 42 ­ (p -p 1) represents the effect of r~tioning; the cross-equation restriction t t­ on its coefficient indicates that increased rationing must decrease consumption and incre~se real balance holdings by the same factor. Labor income plays a dual role in estimation. It is the labor income ~f the present period, but it is also a proxy for the expectation of the income stream over future periods. fl is then expected to lie between zero and one, most probably in the range .• 6 - .9. f2 is an indication of the time horizon of the consumer; as it approaches zero, it indicates a lengthening of the time horizon. A consumer with an infinite horIzon ' ~uId ~ aD effe.c:.t.i\i~ propensity to consume from real balances which is infinitely small', and-Idrig horizons may result in f2 being statistically indistinct from zero 1n small samples. f3 nea~ zero indicat~s that ~he market for goods is not rationed; the larger it is, the more important rationing becomes in determining realized demands. The p point estimate of autocorrelation of consumption equation errors was slightly greater than one. and indicated that (40) should be estimated in first-difference form. The estimate of the consumption equation intercept 1s then lost, but the~y predicts zero intercepts in both equations; This specification of the private-sector decision demonstrates high explanatory power. . Comparison cif the results with unconstrained .tstimation indicated a1moss no difference in explanatory power for consumpti~ demand, but a loss in e'p1anatory power in the real balance equation. l! ~ This appears to be due to the fact that when estimated s~parate1y, the real balances equation confounds supply and demand characteristics in the money market. The cross-equatio~l restrictions imposed here then act to identify the demand-equation parameters. AGD/01l/8-23-84 - 43 ­ The most striking feature of these results is the importance of the rationing coefficient. 11 Using the estimate of -f 3 - y- 2.36. the level of rationing implied by the government price-updating rule (35) is shown in Table 3. and appears to have been a si& ..,Liicant fraction of realised private consumption demand. Although government outlays were sizeable throughout the period under consideration. in the pre-1973 period they tended to be financed through taxation. thus leaving the private sector closer to its demand curve. In the post-1973 period. the government relied more and more on rationing and seigniorage to finance expendit\lreSi this led to larger involuntary saving. By the start of the Benjedid period in 1979-. the government was shifting away from its high-investment-growth and high-saving strategy. This is reflected in lower rationing levels in 1979 and 1980. Subsequent years would likely indicate still less rationing. as government policy has diminished the scarcity of consumption goods. The other coefficients in Table 2 are consistent with theory. f1 lies in the interval between .6 and .9. and accords remarkably well with estimates from other countries of the marginal propensity to consume from disposable income1i f2 is SlIIa11. and is insignificantly different from zero.which suggests that Algerian consumers make decislons for a long time horizon. mo is also insignificantly different from zero. as theory suggests. Fin~IY. the link ,between investment and rationing can be drawn from • the above . res~ts • Consider the residuals of equation (33). (RESt)' which ' represent the~component of investment not explained by contemporaneous foreign • l! The alternative specification of the government price-updating rule which incorporated a "world price pass-through" yielded poorer results in terms of explanatory power. possibly because so small a share of consumption is traded. 2/ See page 45 for footnote. AGD/ 011/8-2 3-84 - 44 ­ Table 2 Estl~ation of Consumption and Real Balance Demands Parameter Estimate T-statistic E not obtained .82 . 4.92 .19 .70 -2.36 2.58 .07 .55 -? ­ bl r (C) - .97 il2 (M/P) •• 98 "-' '- '.;,, '. lY The critical T-va1ue for the 95 percent level of confiden~~ ~IU degrees of freedo~ is 2.23. bl First-differencing was accomplished by placing all lagged values on the right-hand side of the equation. Table 3 Rationing in the Goods Market Year Rationing as percent of C ill 1967 2 1968 4 1969 1 1970 4 ... 1971 6 ... 'f 1972 1973 197 .. • 3 5 10 - ..­ ­, 1975 11 1976 8 1917 9 1978 15 1979 7 1980 8 AGD/011/8-23-84 - 45 ­ financing. From the public budget constraint these should be positively. associated with the degree of rationing in Table 2 (RAT t ) as dra~ from equation (41). In the absence of other financing, as pointed out in equation (20), investment residuals should completely e~plain rationing. Rationing and foreign credit should together account for the financing of the investment program •..:' In fact, the two variables are highly correlated: 6.64 + 7.43 RESt -2 R - .43 (8.55) (3.30) .;. 1.26 Support for the above view of the Algerian budgetary process . ~-l-5;),'" ',.r comes from the fact that variations in (qt E +B ) and RAT capture 92% o.f th.e.. t t t variance of investment. l.! As shown in the budget equation (13), a perfect correspondence would require that current government expenditures equalled current revenues. It would also require perfect consistency between the national, fiscal, monetary and balanc~ of payments accounts. (footnote from page 43) 1/ The results of estimates for the U.S. of the consumption function are - almost identical. The Ando/Modigl1ani results reported in Froyen (1983) for the U.S. included the following statistical relation: • where ~ is the value of presently held assets, and wtL t is pre-tax labor income. The comparable regression result for Algeria, using pre­ tax labor income and real balances, yields: Ct - .26 (Mt-I/P t ) + .73 VtL t 1 2 • 29 (Pt-P t - 1 ) (2.14) (7.87) !-2.78) • There is a striking similarity In the marginal propensity to consume; the asset coefficient is Algeria is larger in point estimate, perhaps because of the exclusion of non-fin3ncial assets, but it is indistinguishable statistically from the Ando/Modigliani result. The rationing term repr~sents the distance the consumers are from their demand cur-.;e. AGO/oll/8-23-84 - 46 ­ VI. Conclusions. The Algerian pattern of development since the ea:-ly 19608 has . differed from that of other developing countries in two major respe~ts. First, Algeria's economy, under the influence of its massive investment program an~ its commitment to socialism, has been centralised and state- controlled. The public sector undertakes nearly all formal invest~~nt and production, and the government exercises strict control over prices in goods, labor, financial asset and foreign exchange markets. Second, Algeria stands out among 011 exporters' in--t:li'e ' vay it - has. managed to channel its crude 011 revenues since 197) · in.to . 1n.v.~neit't~ · " . without bringir.g about soaring inflation or a -Dutch disp.as~- shrinking in the relative size of nonoil tradeable sectors. The theoretical and empirical analysis of this study indicates that Algeria's performance over 1967-80 has been largely due to its commar.d, or fix-price, economic structure. This represents an alternative to flex-price adjustment for oil exporter or other countries experien~ing windfalls gains. It also offers insight into the economic forces with which the Algerians will have ~o contend as they move to,a more flex-price market economy. II Over the past decade Algeria appears to have been in the classical unemployment variant of fix-price equilibrium, with excess demand for goods and excess supply of labor. In this framework, involuntary ~oldings of money - balances result from goods market rationing. . Econometric relUlts reported . here otfer strong support to this view of the Algerian economy. Estimation of the theoretical model indicates both stror.g explanatory power and the impc'rtance of rationing in explaining realise.d consumption and real balance holdings. AGD/011/S-23-S4 - 47 ­ , :.~ . The~e results le~d to both theoretical and policy con('lusions. I First, in the fix-price economy, efforts to estimate consumer demand parameters through Walrasian demand equations will yield biased results. For example, past studies (including the aut~ors') have concluded that the Algerian economy has a low propensity to consume. However, as noted above, estim~tes taking rationing int~ account lead to effective consumption propensities almost identical to those in the U.S. Second, these results have important implications for the design of economic policy in Algeria. The impressive historical Algerian ability to accept high national saving rates and to defer consump~dn does not imply rh8~ such behavior will carryover to a liberal1sed econocy vt,t.h ~lttBttet-clearing prices. The effective demand functions indicate that higher'income, and more goods available for purchas'l, will lead, to demand responses siallar to those found in market economies. Planning for a shift from a fix-price economy towards a liberalised one should not extrapolate past trends without correcting for this rationing effect. Consumption could increase more quickly, and lead to higher inflation, than might have been expected from the historical evidence. Indeed, the rationing estiaates above indicate that without rationing the Algeiian price level would have had to increase by.some 40% more than its historical trend over 1974-78 to hold consumption at its hi3torical levels, even under the favorable assumption that «esired real .. 'f balances are invariant to inflation. - With any appreCiable sensitivity of desired reBl balances to inflation, the Algerian strategy o~he 1970s would • ­ have been infeasible without far tighter wage control or higher nono11 taxes since it would not have been possible for the state to have secured so large a share of purchasing power. Have the distinctL'e features of Algerta's controlled economy been an advantage or a liability in dealing with the variation:. in its t~ros of trade after 19737 On the one hand, there are indications that Algeria has 1I0t used its investment as productively as planned, both in the hydrocarbon and non- hydrocarbon econo~ies 3nd that it has failed to stimulate nonhydrocarbon exports, but the same might be said for certain other oil producers. An (:qually serious feature is the capital intensity of the public investment program in the face of widespread uuderemployment. On the other hand, an excessi'/e shift to nontraded goods was avoided at the cost of postponing . private consumption. S::eps have also been taken to divers..l!!f a::oa~£; f>r\.rCf' \J1'1.~ · ~ ;" rapidly depleting natural resource, to natural gas, one vi th afa1:- .1o,oger. > expected lifetime. The payoff to these hydrocarbon investments, a topic not explicitly addressed in this paper, depends o~ the relation£hip between gas prices and world intP-rest rates since the moderate element of natural rent in gas prices is sensitive to ch~nges in capital costs. Algeria was also been spared the sharp contraction occasioned ~i the massive outflows of private capital which characterised Venezuela, Mexico and other exporters with open financial markets after 1981. In sum, it will be possible to evaluate the Algerian strategy only whel the costs gnd benefits of the phase of rapid accumulation becomes clear. These will depend on three key factors, two external and one .. . .. internal:t (a) de'elopments in the world gas marke!: relative to (b) t~i~ levc'. of world ~nterest rates, a.ld (c) the efficiency with which noe-hydrocarbon investmen~• and domestic labor can be used. :f external factors are favorable • • and domestic policies encourage efficiency, the transformation of oil income and borrowing into domestic capital will permit substantial growth of reJl consumption. Otherwise, alternative strategies such as saving abroild or . o ~ increasing domestic consumption more directly through relaxing import I restriction, and easing nonoil taxation will appea~, in retrospect, to have been more promising. • f o • - · • • • ­ Bi bl10graphv Amemiya, T.: "A Note on a Fair and Jaffee Model." Econometrica, vol. 42 (July 1974), p~. 759-762. Amin. S.: Accumulation on a Wcrld S~~l~. London: ~onthly Rev1ey Press. 1974. Renassy. J. P.: "Nea-Keynesian Disequilibri~ Theory ~n a Mon~tary Economy." R=·,iey of EconoCli~ Studies. vol. 42 (October 1975). pp. 503-524. Bordo. M.D.: "John E. Cairnes on the Effects of the Australian Cold Discoveries, 1851-37: An Early Appreciation of the ~~th~ology of Positive Economics". History of Political Econ~m~ 7. No. 3 (1975). pp. 33J-359. Bruno, M., "The Two Sector Open Economy and the Real Exchange Rate", Fajk Institute, Hebrey University of Jerusalem (August. 197!) •.. Ch~nery H. and M. Syrquin, Pntterns of Devel0Peen~: 1950-r970. Oxford: Oxfor1 University Press, 1975. COnyal, P.: "011 Rents and Development Strategy: A Case S:.,dy of Al~e:-::-'\," World Bank mimeo, 1982. v .-' Drazen, A.: "Recent Developments in Macroeconoc.1c Disequilibrit.:' Thoeor . Econometrica, vol. ~8 (March 1980), pp. 283-306. Dreze, J.: "Existence of an Exch£nge Equilibrium under Price R1gidl~ies," International EconOtUic Revie", Vol. 16 (':une 19;5), pp. 3al-320. Fai~, R. and D. J&ffee: "Methods of Estimation for Markets in Disequilibrium, - Econometrl,=a, vol. 40 (972), pp. 497-514. Froyen, R., Macroeconoaics, Ne\l York: MacHilla~, 198~. Hicks, J.: Capital and Gro1o"th. Oxford: Oxford University Press. 1965. .. Komai, J.: 1he Econocics of Shortage. Aosterdam: North-Holland. 1978 • • Lewis, W. A., -Economic Development witl. Uolim1:ed Supplies of L.::bor" , Manchester Sch:>ol of I:conomics & Social Studies. 20 (Hay 1952), pp. 139­ 191. Mallarde, E.: .!:..:~&erie Oe":>uis. Table Ronde 1j75 • Paris: Le • Mal1."veud, E.: ~e Theory of Unemplovme!lt Reconsidered. Oxford: B:tsil olackYell, 1977. . Morgan, D.R., -Fiscal Policy in Oil Exporting Countries", EMF Staff Papers (March 1972), pp. 55-86. Muellbauer, J. and R. Portes: -Macrocconou.ic ~dels I.11th Quantity Rationing-, Economic Journal, vol. 88 (1978), pp. 788-821. Murphy, K.J , MacroproJ~~ts in the ~ird World. Ney York:~est~~nd Press, 1983. Picard, P.: -Inflation and G~")yth :.1 .1 DisequUibriu:n Macroeconooic :- : :~l. · Journal of Economic Theorv. v~ ·~., (1983), pp. Portes, R. and S. Winter: -The Demand for Money and Consumption Goods in Centrally Planned Economies-, Reviey of Economics and Statistics, vol. 60 (February 1978), pp. 8-17. Portes, R. and S. Winter: -DisequilIbrium Estimates for Cons~ption Goods in Centrally Planned Economies-, Revteyof Economic Studies. vol. 47 (January 1980), pp. 137-157. Guandt, R.: -Testing nypothescs in Disequilibrium Moaels,- Princeton Econometric Research Frogram 1197, (1976). Quandt, W.: Revolution and Political Leadct"sh1'pfAlgeria p 1~~~-196&. Cambridge, MA: MIT Press, 1971. Rllffinot M. and P. Jacquemoti Le Cap1talisme d'EtAt Algeria. Paris: ~~.· spero, 1977 • Zartman, I.W., -The Elites of the Maghreb: A Revue Article-, International Jcurnal of Middle East Studies 6, No.4 (December 1975). Ii ... .. 'f - - • • " • Appendix 1 The Individual Consumer's Decision Suppose a Bernoulli utility function for individual i (the subscript i is suppressed until summation across individuals) u - (1) to be maximised subject to a budget cocstraint for all t: lj" + i - (b+qe) (9) (iv) Combining the goods market and government budget: * 6 o - (T -l)z l+r +~ m - t 1 + (g'-l)y* + z* - (b+qe) . (10) . <5 l+r - =z* + ~~-l + (g'-l)y* -(b+qe) (11) T f o ... - c* + (g'-l)y* - (b+qe) - .. • - (12) .. ", Therefore 1 y* - (l_g'){c * - (b+qe)} .. - 1 a (13) By (11): T 6 * (l-g') m~_1 - l+r ( - r + a + (b+qe)) (14 ) Equation (13) indicates the relationship between c* and y* and (14) the Walrasian level of real balances dependent on parameters and policy variables. (v) In equilibrium m* is constant so that e- n; inflation is zer~ s d because y - y. Government spending and taxes are Ifnked by t~e budget and the steady state constraint on financing: mn - (g'-l)/a + *(I-r) - (b+qe) or (15) w g' - a[mn - ~l-r) + (b+qe)] + 1. w . Noting that j"{l-T) - z* and using (14) to eliminate z*: T-6 !.=8.' m* - (1+r+6n) [( a ) + (b+qe») (16) Together, (8) and (16) determine the Walrasian equilibriu~ in (k*,m*) space! - . B. The Boundary Conditions. • ­ " (i) Repressed inflation/classical unemployment (REP/CLA): s d d · (a) labor market is in balance: L a Land 1 - min{ay ,aSk} ( bl) product deman d excee d s supp 1:': Yd ) yS 101 hi c h iep 1 ies: (c) for the labor market, 1 - aSk, and (17) 1 <5 l+r (d) on the goods market ~ < rz* + g'/a + ~t-l - (qe+b) Therefore by (14) mt - 1 ) m*t_l. (18) (ii) Repressed inflation/Keynesian unemployment (REP/KEY): d s 1 (a) product market is in balance: y - y min(- Sk) a' . d (b) labor market is supply constrained: 1 ( min(ay ,aSk) which implies: 1 ( c) 1 ( aSk or as ( k , and (19) (d) in the product market, by (5) - 1 a <5 (t .-l)z* + 1. a + (8 + -r-)m l+r - • t 1 (20) Replacing em t­ 1 from the governcent budget constraint (3): <5 l+r .&=.!.. . o • (t -l)z* + ~t-l + i + a + z* - (qe+b) (21) l-rT' <5 =-z or ~. (22) a T -.,L* g or mt-l • T{ T~ + (la- ') + (b- ' q'e) - A( i i _ +i ( »} qe+b 1 t l 2 (23) - If investment is • adjuste~ to the labor shortage by reducing capital formation this is formally equivalent to allowing A to be negative so that m-1 ) m~_l: to huld the product market in balance consumption out of money t balances ~st be large enough to compensate for reduced investment. (iii) Keynesian unemployment/classical unemploycent (KEY/CLA): (a) the labor market Is demand constrained: 1 ) min{ayd ,aSk} (b) product market Is in balance" ys ~ y d - min{ !,Sk} . a This implies: (c) 1) aSk so that u ) 0, and (d) in the product market, by (1) where (25) The locus (m,k) in (28) has positive slope (except possibly for k»k*) positive since :~ - ST[6+~ta(i,it_1+i2(b+qe»)). If k-O, mt_1