Oil Price Indexing-Versus Large Price Shocks: Macroeconomic Impacts

International Finance Discussion Papers Number 180

May 1981

QIL PRICE INDEXING-VERSUS LARGE PRICE SHOCKS: MACROECONOMIC IMPACTS

by

Laurence Jacobson and Stephan Thurman*

NOTE: International Finance Discussion Papers are preliminary materials circulated to stimulate discussion and critical comment. References in publications to International Finance Discussion Papers (other than an

acknowledgment by a writer that he has had access to unpublished material) should be cleared with the author or authors.

Oil Price Indexing-Versus Large Price Shocks: Macroeconomic Impacts

by Laurence Jacobson and Stephan Thurman™

Introduction

Following the dramatic oil price increases by the Organization of Petiroleum Exporting Countries (OPEC) in late 1973 and 1974, average prices rewcined fairly stable through 1978 in nominal dollar terms (see table 1). However the purchasing power of OPEC oil as measured by the U.S. implicit GNP deflator or the price deflator for consumption fell substantially between the first quarter of 1974 and the end of 1978,

and the value of the dollar fell significantly versus a weighted

average index during the same period. This contrasts strongly to the experiance since the end of 1979 in which OPEC price increases have greatly outpaced the rate of inflation and exchange rate fluctuations, causing a more than doubling of the real oil price during the subsequent eighteen months.

As a consequence of the recent price shock (commonly referred to as OPEC II) and the resulting acceleration of inflation and macroeconomic adjustments, many observers have bemoaned the fact that real oil prices were allowed to fall after the 1973-74 price shock (OPEC I), and that the painful adjustments required as a result of OPEC II would have been substantially mitigated if real prices had increased gradually between 1974 and 1978. However, while much economic work has focused on the negative effects of large sudden oil price increases, little or no work has examined the effects of smaller but continuous price changes over ~*/ International Finance Division, Federal Reserve Board, and Wharton Econometric Forecasting Associates, Inc., respectively. This paper represents the views of the authors and should not be interpreted as reflecting the views of Board of Governors of the Federal Reserve System

or other members of its staff, or Wharton Econometric Forecasting Associates, Inc.

-2-

time, on which basis one might consider the desirability of indexation versus historical experience. The purpose of this study is to shed some light on the issue by examining simulation results using the Federal Reserve Board MPS model and comparing the 1974-1981 period with actual (and projected) oil prices versus several alternative oil price indexing formulations.

The indexing formulations were influenced by the long run pricing strategy which has been the subject of discussion within OPEC during the last two years, as well as by the theoretical literature on optimal extraction of depletable resources. We have experimented with several pricing schemes based on the premise that the price of oil would have increased in real terms following OPEC I, after adjusting for inflation and fluctuations of the dollar in terms of other currencies. All of the indexing formulations yield a large gap between the indexed and historical price by the end of 1978, but none of the indexing schemes yield as high an oil price as has already been experienced as a result of OPEC II. Therefore, we also introduced a modified shock scenario, in which oil price increases with indexing accelerate starting in 1979 and are equal to OPEC II levels in mid 1980. The simulation results suggest that a series of small oil price increases have smaller negative effects on overall prices and income than discrete jolts of equal cumulative magni-

tude.

The Energy Sector in the MPS Simulations

The energy sector has been expanded from that used in earlier

MPS oil price shock simulations (See Thurman and Berner (1980)). Pre-

-3-

viously, domestic petroleum demand was modeled as a function of real income and the exogenous wholesale price index for energy (PWIFE) relative to the implicit GNP deflator. The equation used in these simulations projects total primary energy demand in BIUs (BQET) as a function of real GNP (XSNP) and a thirty two quarter price lag specification. Cooling

and heating degree day data (DEGC and DEGH) were used to capture seasonal variation and a dummy variable was used to capture the effect of the 1973-74 oil embargo (EMB). PFW is the gross domestic sales fixed weight deflator. wis the observed error term, and the positive coefficient reflects partial correction for autocorrelation. The estimation technique used was Cochrane-Orcutt with a second degree Almon polynomical distributed lag constrained to zero at t-n. The sample period was 1966 through the

last quarter of 1979 (quarterly data).

In BQET = -4.939 -.0657 EMB +.0001 DEGC (-14.8) (-2.7) (16.7)

+.00008 DEGH +1.08 1n XGNP (4.1) (22.6)

PWIFE -.301 (35) In ( PEW ) + .238H_, + (7.3)

Dw =1.910 Re =.971 SE = .0246

where cumulative four quarter lag sums from the lag distribution L35 are:

Cumulative Q Sum Q 4 - .0073 v3 8 - 1415 /, i 12 - .1934 i=1 16 ~ .2340 20 - .2642 24 - .2849 28 - .2969

32 - .3013

-4-

The energy demand sector in the MPS model can be used to determine relative shares of petroleum, natural gas and coal at the primary level as a function of prices of the different energy components. relative to the aggregate energy price index (PWIFE) with a ten quarter lag specification. However, at the time of the simulations undertaken for this study, the complete system had not been finalized, and a constant

share assumption was utilized;

= * * DQFL Yager ~ BOET * Kagey

where DQFL is domestic petroleum consumption (mb/d), Vagel is the historical ratio of petroleum consumption to total energy consumption in each time period in quadrillion BTUs, and Kagel is a unit corversion factor from quadrillion BTUs at an annual rate to millions of barrels per day (mb/d).

Under this assumption the percentage change in petroleum consumption is equal to the change in total energy consumption, although this would be true in the complete system only in the restrictive case of no change in relative prices of petroleum, natural gas and coal. In mltiplier simulations, petroleum consumption falls somewhat more than with the constant share assumption in response to a petroleum price shock, if there is no change in coal or natural gas prices. A complete description of the new energy demand equations is available in Jacobson, Loftin and Thurman (1980).

Imports of petroleum products (EMP) in billions of dollars at an

annual rate are then determined by the identity

EMP = (DQFL - SQFL) * PUVFL * k

-5-

where SQFL is exogenous domestic production (mb/d), PUVFL is the average dollar import price of oil per barrel and k is a unit conversion factor from mb/d to $billion/year.

In the reference case, the projections beyond the second quarter of 1980 for the domestic wholesale energy price are calculated as described in Thurman and Berner (1980). The simulations were run prior to the immediate decontrol of domestic crude oil prices which occurred in early 1981, and assumed the then existing phased decontrol plan in which domestic wellhead prices reached the equivalent import price in the fourth quarter of 1982. For the purpose of these simulations, the reference case assumes continued nominal oil import price increases of about 4 percent per quarter after the second quarter of 1980.

In all of our indexing schemes, we make the assumption that the pecentage differential in domestic petroleum prices equals the percentage difference in import prices when comparing the reference and index cases. The nonpetroleum components of the domestic wholesale price index for energy are not assumed to increase at the same rate as the domestic petroleum price component of that index.” This contrasts the two possible extreme cases; no passthrough to domestic producer petroleum prices (due to effective controls), and absolute increases in domestic petroleum prices equal to the increase in the import price of petroleum. Our assumption is equivalent to asserting that the domestic/ foreign petroleum price ratio is unchanged by the price path of petroleum

imports in the alternate regimes.

Alternative Oil Price Scenarios

Results from five oil price indexing schemes are presented in this paper. The spirit behind the alternative mechanisms is to allow the price of oil to rise gradually in real terms measured in dollars subsequent to the 1973-74 oil price shock. While the pattern of increase and total cumulative increase differs in the various cases, all but one show considerably higher prices through 1978 after which the differential between the index cases and the reference case closes rapidly and becomes negative by mid-1980. The one exception is an exogenously determined price path which contains a price acceleration during 1979 and 1980 so that oil prices in the reference case reaches, but does not exceed, prices in the alternate regime.

Of the five oil price indexing scenarios, two are exogenously determined (including the modified shock scenario described above), and three are endogenously determined within the model simulation. All indexing schemes are calculated in terms of the U.S. import unit value for petroleum, rather than the OPEC average crude price. As can be seen’ in Table 1, the U.S. import price historically is slightly higher than the OPEC price (due to higher average. quality composition), except in periods of rapid price increase (due to a two month transportation lag). While the choice of starting point is somewhat arbitrary, the first quarter of 1974 seems reasonable for OPEC prices, although the large oil price increases of 1973-74 were not fully reflected in the U.S. trade data until the second quarter of 1974. To reconcile this

problem, the second quarter U.S. import unit value in the indexed cases

was calculated as if OPEC prices had been indexed since the first quarter.

-7-

In the exogenous cases, this resulted in an increase in the second quarter 1974 U.S. import price (6.74). We then used this contrived second quarter 1974 U.S. import unit value as the starting point in the three ‘remaining endogenous simulations.

Our choice of oil indexing schemes was heavily influenced by

the 'Summary of Recommendations’ that was presented to the conference

of OPEC Oil Ministers in Taif, Saudi Arabia in May, 1980 by the Committee on Long-Term Strategy 2 This document suggests that a pricing floor should be established to maintain rising real prices combining three elements:

a) An inflation adjustment, in accordance with which prices would be adjusted by an index reflecting the impact of inflation on international trade.

b) An exchange rate adjustment on the basis of 11 currencies plus the U.S. dollar.

c) Oil prices should rise in real terms proportionately to the growth in real GNP of OECD countries.

Significantly, however, the OPEC price calculated by such a formula was proposed to be a ‘minimum floor price’ with the explicit qualification that 'as markets swing from surplus to shortage ... the pricing strategy must be flexible enough to go beyond the minimum floor price level’. After any possible shortage is over, it was proposed that prices be frozen in real terms until the calculated floor price catches up with the new market price, or else that the floor price be immediately

raised to a higher level. Thus, an OPEC indexing scheme as formulated

could not insulate the Western economics completely from an oil price

shock caused, for instance, by an oil supply disruption. Therefore,

-8-

we decided to examine a modified shock, or a rapid acceleration of oil prices between the fourth quarter of 1978 and the second quarter of 1980 which would yield the same ultimate effect on the world oil price level

as in the reference case.

In some formulations we used only U.S. variables rather than

weighted average industrial country variables in calculating indexed

oil prices. In the MPS sector, the determination of U.S. income and prices is much more developed than foreign GNP and prices (some of the foreign sector is still in experimental stage). In all cases in which the oil price is determined endogenously, the weighted average foreign exchange rate is exogenously set at historical values (with the assumption of no change between 1980-Q2 and 1981-Q4) ; that is, unchanged from the reference case.

The general indexing formula used in the simulations is as follows:

W W W ; W PUVFL = PUVFL o PGNP . ) 1 ( FPCI, ) 2 (ee \ 1 (ae: Wo (Re ) t t-1 l

FPCL FGNP_ PGNP, 4 te XGNP) tel WFER, 1 where PUVFL = unit value price per barrel of oil PGNP = Implicit U.S. GNP deflator XGNP = U.S. GNP in 1972$ FGNP = Foreign real GNP FPCI = Foreign consumer prices, multilateral weights WFER = Weighted foreign exchange value of the dollar

and Wy + Wy = 1.

Thus the oil price is indexed multiplicatively by changes in prices,

GNP and the exchange rate. In the first exogenous oil price scenario

(Scenario 1), W, = 0 (mo foreign variables) and the GNP adjustment is

2

replaced by a constant 3 percent annual rate of growth. U.S. prices

~9-

are derived from historical values (Table 1). The second exogenous case (Scenario 2), is identical to the first case through 1978. Starting in the first quarter of 1979, a 6 percent per quarter increase is applied through 1980-Q2, after which oil prices are set equal to those in the reference case.

fu the first endogenous case (Scenario 3), Wo = © (no foreign variables) and an exchange rate adjustment is not included. In effect, the oil price is indexed to changes in nominal U.S. GNP. Since U.S. prices and incomes are derived within the simulation, feedback effects of oil prices to macroeconomic variables and back to oil prices again occur throaghout the scenario. The next simulation (Scenario 4) includes the exchange rate adjustment; an appreciation of the dollar (increase in WFER) ‘reduces the oil price, and conversely for a depreciation. The last simulation (Scenario 5) uses foreign as well as U.S. variables in

determining oil prices, with W, = .213 and Wo = .787 (the actual °

1

multilateral trade weight shares in 1976).

Comparative Simulation Results A summary table showing simulation results for oil prices, GNP and the consumer price index in the five different scenarios, is shown in table 2. More detailed results for individual scenarios are shown in tables 3 through 7. Charts 1 through 3 display the paths of oil prices, real GNP and consumer prices. Note that the reference case 4

consists of historical data through the second quarter of 1980', and

a model forecast during the second half of 1980 and 1981.

- 10 -

Common to all of the endogenous indexing schemes is a steadily widening gap between the indexed oil price and the historical price, which reaches 20 to 30 percent by the end of 1977 and increases dramatically in 1978. By the end of 1979 the indexed prices are lower than in the reference case (due to the almost doubling of oil prices during 1979) and become increasingly negative during 1980, reaching over -30 percent in the endogenous cases. The exogenous oil price scenarios (Scenarios 1 and 2) show a greater peak price differential by the end of 1978 due to the constant 3 percent real price growth assumption. Since the U.S. growth rate averages less than 2.5 percent during the simulation period, this accounts for most of the discrepancy. By assumption, Scenario 2 shows no difference in oil price from the control case during the last six quarters of the simulation.

Scenario 3 shows a much smaller peak oil price differential than the other cases, due to the ommission of an exchange rate adjustment (as seen in Table 1, the dollar depreciated almost 10 percent between 1977 and 1978). However, the relative stability of the dollar (and actual appreciation during 1975 and 1976) results in lower prices before 1978 in Scenario 4 (with exchange rate adjustment) compared to Scenario 3. Scenario 5, which indexes oil prices to a weighted average of foreign and U.S. price variables, shows generally higher prices than the other endogenous scenarios due to the relatively rapid growth of

foreign to U.S. prices in the model.

The various indexing schemes yielded a reduction in domestic

energy demand of 6 to 9 percent (about 1 to 1.5 mb/d) by the end of

-~ ll -

1978 when the price differentials between the index and reference cases

are at their peak. By the end of 1981, the index schemes yield slightly

higher petroleum consumption levels than the reference case (0 to 6

percent) due to the substantially lower prices. In the modified shock

case (in which the price differential equals zero during the last six

quarters but does not become negative), petroleum consumption remains

3.4 percent lower at the end of 1981. This differential is affected by the

lag structure of the energy demand equation, and would be partially eliminated with a longer simulation period. In all cases expenditures on energy product

imports are sharply higher by the end of 1978 (17 to 33 percent), and

substantially lower by the end of 1981 (10 to 25 percent).

Reductions in real GNP are present in all cases through 1978, with end 1978 values lower by 1.3 percent to 1.8 percent. The amouni:s vary directly with the cumlative oil price shock, with the change in level of GNP per 10 percent cumulative price increase ranging fron .24 to .33 percent. However, between 1978 and 1981, the negative impact of the oil price shock in the reference case yields a level of GNP 1.7% to 2.9% higher than the reference case. Clearly, the MPS model yields stronger real effects from large price shocks than from a series of small shocks spread over a longer period of time. This :is most clearly evident in the case of the modified shock scenario, in whiich real GNP is 1.7 percent higher at the end of 1981, even though prices are identical to the reference case during the last six quarters.

In the case of domestic prices, the effects of the difference

of small versus large oil price shocks is even more dramatic. By the end

-12-

of 1978, mltiplier values per 10 percent increase in oil prices «ange from .10 to .16 percent for the GNP deflator and .15 to .20 percent

for the consumption deflator. Change in price levels at the end of 1978 average about 1 percent in the various scenarios. However, the price levels at the end of 1981 are 4 percent or more lower in Scenarios 1, 3, 4 and 5. In the modified shock case (Scenario 2) all price indexes are 3 to 4 percent lower by the end of 1981.even though oil prices are always

greater than or equal to those in the reference case.

It should be noted, however, that the price multiplier effects from a constant oil price shock are larger over the first few years of the simulation and then become smaller in subsequent years. The decline in the price mltiplier is a corisequence of thte energy price induced ‘reduction in aggregate demand having a strong negative effect on prices.? This results in relatively flat, or in some cases declining, price level multiplier between 1975 and 1978. This relationship also exaggerates the ultimate effect on prices in the reference case relative to the indexed cases at the end of 1981.

The unemployment rate gradually increases in all scenarios to a level .85 to 1.25 percentage points higher than the reference case.

The peak differential occurs about one year after the peak oil price differential, due to the strong lagged effects of unemployment to reductions in GNP in the MPS model. By the end of 1981 the unemployment rate is .3 to .8 percentage points lower in the index cases, and .15 points lower in the modified shock case.

Interest rates are not strongly affected in any of the scenarios

(all simulations assume unborrowed reserves identical to the reference

- 13 -

case). Interest rates are generally slightly higher in the first two or three years and drift lower in later years, but show no changes greater than one percentage point.

Foreign GNP effects are generally similar to U.S. GNP effects, showing somewhat stronger negative effects in the earlier years and lower percentage increases at the end of 1981 (although always higher than the reference case). Foreign price effects are much stronger than U.S. price effects, but this is largely due to the structure of the foreign price equation (discussed above), which is extremely sensitive to oil price increases,

The current account balance shows a pattern similar to expenditure on energy imports, with deterioration in the earlier years, and improvement after 1978. The end of 1981 current account improvement in the index cases is within $5 billion of the reduced expenditure on energy imports. In the modified shock case the gap is higher ($9 billion) due largely to different effects on U.S. and

foreign GNP levels.

Summary

In the comparison of the mltipliers from these alternative oil indexing schemes, the nonlinearities within the econometric model (ané. in the economy it replicates) play an important role. By smoothing the rate of increase in world oil prices, the simulated economy is spared the discrete jolts which occurred in this commodity price since 1978. Many behavioral relationships in the model and within the economy:-- the trade-off between employment and prices chief among

them -- will react more strongly to a sudden impulse than to an

- 14 -

impulse of equal magnitude occurring over time. The differences in

reaction of the model to smoothly indexed oil price increases explain

much of the improved economic performance even in the case of Scenario 2

where world oil price levels are made equal by the end of the simulation. It could well be that the relative improvement in the economic

indicators stemming from the various indexing schemes is understated.

Investment expenditure planning, for example, would be facilitated through

the increased degree of certainty which result from the oil price

indexing formulas. This, and other expectational relationships within

the economy, are inadequately captured in this and most econometric

models.

-15-

Footnotes

1. The model was reestimated in January, 1981, following a revision in historical GNP data by the Department of Commerce, and including actual data through the first quarter of 1980. The latest estimates show a slightly lower income elasticity (1.065), and a larger cumulative

price elasticity (-.33).

2. Coal and natural gas prices generally are determined through long term contracting procedures and thus the competing energy price effect on these nonoil energy prices occurs with a considerable lag and will not increase as much as petroleum prices. See Thurman and Berner (1980).

3. Text presented in Petroleum Intelligence Weekly, May 12, 1980.

4. The historical data for GNP and prices prior to the Department of Comnerce revision (see footnote 1).

5. This effect is stronger than results derived in earlier oil price shock simulations using previous versions of the MPS model.

References

Thurman, S. and Berner, D. Analysis of Oil Price Shocks in the MPS Model.

Paper prepared for the CEPR Conference on Energy Prices, Inflation and Economic Activity. Cambridge, November 7, 1979.

Jacobson, L., Loftin, A. and Thurman, S. Energy Demand and U.S. Petroleum

Imports International Finance Division Discussion Paper. Federal Reserve Board. (Forthcoming).

SST°T 8ST 'T €cr't TEO'T €20'T 7L0°T €90°T 180°T

("ang *20q/$) aqey esueyoxy uster0g oesersay “PIM

. (¢)

LT°LT 6€° IT 8L°8 €€°6 66°8 06°8 LE°6 VIE COOT# CE) /(Z)) aot ag

UNSTOIRWSg SATIETOY

(4)

WLLT 6° €9T 9° TST SCT T° S€T 9° 821 @° LTT 6° SOT

TOFETFOQ SoTaAg

AND 3FotTTdwy

(€)

97° 0€ L£9°8T Te €T 62° €T 1° ZT Sv TT 86° OT 3 a

——TBaREGT EY

enqTeA Fun qaodwI Ito s‘a (Z)

SHNIVA TVOTHOLSTIH ‘TVILLOV

T WIV.

18°0€

£9°81

€6°Z1

88° ZI

LL°IT

8L° OT

8Z° IT

oly —Treaaea ey aotag

Jqd0 e8er2ay (1)

O86T 6L6T 8L6T LL6T 9L6T SL6T YLET €L6L

ITBOR

1974 1975 1976 1977 1978 1979 1980 1981

1974 1975 1976 1977 1978 1979 1980 1981

aa bo WODUR HE EE

Scenario #1:

Scenario #2:

Secnario #3:

Scenario #4:

Scenario #5:

Scenario #1

Lae]

Scenario #3

TABLE 2

OIL PRICE INDEXING SCENARIOS Summary of Multiplier Results (percent difference fron control)

(1974-1981), 3% real P oi increase, exchange rate adjusted.

il (1974-1978), 3% real Pod (1979-1981), 6% P oil equal to reference scenario thereafter.

1 increase, exchange rate adjusted.

increase per quarter through 1980-Q2,

P_., increases with endogenous U.S. P_, Y. oil y

Pol increases with endogenous U.S. Po Y, and adjusted for exchange rates,

Poil increases with endogenous U.S. and foreign Py? Y, and

adjusted for exchange rates.

Scenario #2

oil Y Po Foil x Po 15.7 -0.2 0.5 1974 15.7 -0.2 0.5 23.7 -0.7 1.1 1975 23.7 -0.7 1.1 21.1 -1.1 0.9 1976 21.1 -1.1 0.9 28.2 -1.3 0.7 1977 28.2 -1.3 0.7 64.5 -1.8 1.0 1978 64.5 -1.8 1.0 4.3 -1.5 -1.2 1979 16.8 -1.7 -0.8 -16.6 -0.1 -3.8 1980 0.0 -0.7 -3.0 -19.3 2.9 -4.9 1981 0.0 1.7 -4.0 Scenario #4 Scenario #5 yf, Four = Fe Pour *% Pe -0.2 0.4 1974 11.5 -0.2 0.3 1974 12.7 -0.2 0.4 -0.5 0.8 1975 17.0 -0.5 0.7 1975 20.4 -0.5 0.8 -0.9 1.0 1976 17.2 -0.8 0.7 1976 25.7 -0.9 1.0 -1.1 0.9 1977 22.6 -0.9 0.7 1977 29.5 -1.1 0.9 -1.4 0.8 1978 54.2 -1.3 1.1 1978 57.4 -1.6 1.1 -1.0 -1.4 1979 -4.9 -1.1 0.8 1979 -0.9 -1.4 -0.9 0.1 -4.1 1980 -31.0 -0.3 -3.5 1980 -25.6 -0.5 -3.5 2.8 -5.4 1981 -32.9 2.1 -4.9 1981 -30.3 2.1 -5.2 Pil = Price of imported oil (PUVFL) Po = Consumer Price Index (PCON) PY = GNP deflator (PGNP) Y = Real GNP (XGNP)

*(queaand) UOTTITG § UT eoUereFJTp SRNTOSqe oy. ST UMOYS eBuUPYS 347 ‘g0ueTeq oper} oy} A0q ‘squTod e8equoeorzed uy ueATs ere seBuByod ‘soeqeAr T[Tq Aansvearq pue quowsoTdueun ey 20g x

sooTig uotjdunsuoyj uste1z0¥g dN) uster0g

cO°T- = 9ST GZ°9 evs 87°17 19°€ Le°€ 86°

ve" T 06° 90°%- = HOS 16° - IT*I~ 60°T- €T°- O€*9T OS°9E LT°9Z TIT'OT- €€°T- 92°S- c9°E- =O 7 Q9S°LT- ZL°L7- O8°ST- EEE c6°1IT = T?* TT es'vT = YI°4vT

yOOUBT BG opel], osTpusyoszoy sjioduy wneToijzed Jo ene,

€L° Go°y- =e B- —s LOB BE°S- GZ 88°t- 6L°- uoyzadwunsuos wnejTorzjzed oFAseuU0g O€°6T- O9°9T- LZ°y 45° 179 61°8Z O1'IZ 69°€Z ELST uneTorqeg Jo eotad yaodwy 9g°- L9°- Ly- TZ°- Il’- TO°- 80° Z0° %2781 TITq Aanseory LL’- 89° 91°T 90°T 8L° 6S° e* L0° ¥e9e21 QueuXoTdueuy 06°4- T8°E- T2°I- 66° ZL° 06° Il’ ¢¥ 1ozeT Zep uoftydunsuop O€'+- 99°%- 94°- z9° gs° GL° c3° 02° 103eTFep IND €6°2 90°- ZS°T- = SLZ°T- L@°I- | (21° 69°- Z@e'- ($2261) dND OS*I- @L°@- 86°I- LTI'T- 69°- 8E°- cT° TO°- ($ Jueaans) ANd

T86T O86T 6L6T 8L6T LL61 9L6T GL6T L6T

(seqea aevek Jo pus f[orqUOD WOAZ SdUSTEZITpP Jusd10g) (quoujsnf{pe ejei a8ueyoxe YIM (esPetoUT [Ret YE) UoFIe[NUTS votad snousZ0xqg i] OTazBUadSs 209TId TIO PTAOM PEexepur FO ASFFT

€ Flava

*(quezInNd) UOTTITG § UF soUeAeFZTP OANTOSqe oY} ST UMOYS B3URYdD 943 Ss90uReTeq eperq oy} A20q ‘SjUTOd Se8eqUeDZed UT UsATS Ore soOBuBYDS ‘sEeqeAl [ITq AAnsver} pue QuowAOTdWeUN Oy AO_ x

8T°Z% T8°€ Te°T Cv'l 87° T9°€ LO°'e 86° soofad uotydunsuo) uster0g or’ 12° Ge°Cc~ 70°C 06° TIT’°T- 60°T- €T°’- dN) uszer0g €c° Lt YI°TE = CT° 0% TIT°OT- €e°T- 92° S- o9°E- =O soUuBTeq opera osTpuRryorey] TT°OT- TZ°61- TS°8- eve 26° TT T2° rT €S° YT YI Vi sjiaodut wneTozjed Fo enTeA €v7°e- 88°9- 90°6- L0°8- 8¢°sc- (4 i bl 88°2- 6L°- uofjdunsuod uneToijzed oF 4sem0g 0 0 T8°9T 7°49 6T°8t OTT 69°€% ELST uneTorjzed Jo sotad Aaoduy 9S°- o9° oy°- To°- Tt*9 TO°- 80° Z0° ¥2781 TT Fq Aanseerzy, cT*- L6° €Z°T 90°T 8L° 6S° ve" L0° ¥27e81 QueufoTdueuy 66°E=" O0°E- O8*- 66° ZL’ 06° IT°T Sv 1ojeT Joep uoF zdunsuop e7°e- 66°T- €2'°- c9° 8s° cL* c8° 02° A0R38TFOP IND YL°T 89°- OL°T- 8ZL°T- Le°T- et’ t- 69°- (7 iad ($ 7261) dNd 9L°T- 99°@- €6°T- LTT 69°= 8E°- ST’ T0°- (§ JqueaIND) aNd

T86T O86T 6Z6T 8ZL6T LL6T 9L6T SL61T %L6T

(soeqei aveh Fo pus {[orqUOD WoAZ soUSTeTJIp WI9II0g) (O8-6Z6T UF YOouS petsTpow uaTM T OFzAPUeDS Se SUES) UOTIeTNUTS eoTAd snoussoxy :7 OFABUsDS 2°F4d TIO PTAOM pexepuy FO JO9FFA

b WTawt ¢ eka viw

SsourTeq eperq 3y} 310g

0€°9- WES

80°€- £0°€- SO°ST ev" V1 L9°€- 8T*e- TT" 42 S0° 22

LE*ST G8° z= c8°s ce LE~ cS*cB" €7°S- BL° 18°Z TT°?-

T86T .

LS°Z- T8°€ so°T Ze°T- IS°Ze G0°92 16° LE- 87° ¥- S6°~ 68°S- SL°SE- T6"TT- L9°- 1y"-

a 68° Ol’? - Le*Tt- €0°e*- 09°- OT’ T0°T- €6°@ 09°T- 086T 6Z6T

0e°9 z9°T-

OT’- Le°Lt LL°9- Sv er cT** 88° T8° 0L° 0v°T- TL*-

8L6T

*(qUezAND) UOTTITG § UF BdUSeTeFZTP eANTosqe oy} ST UMOYS oeZueYo oY4I

ss°¥ 80°T-

70° - 70° OT 76° Y 90°22 70°- L9° 88° LL° 80°Tce"

LL6T

c9°€ LY"? 60°T- SL’

z0* sO" 87° (Aas L6° €8° 9L° €9° 88°- yG*~ €T’- 60°- 9L6T SL6T

G WIaVi

68° €T’-

S8°?~ 98° OT TL°irae ZO* 90° 6€° co" 8T*- 70° -

SL6T

(soqevi azeek Jo pue {T[O1}qUOD WoOAZ DdUSTEFFTpP Juedi1eEg) (AND *S* [euTWOU UT eseetTOUT AJoOF QuoWySNf[ pe) UOTIETMITS eoFiag snousso0puq 20FId [TO PTIOM pexepuy JO F0eFFa

*squtod e8equeczed uz ueaT3 ore sosuryo ‘seer [Tq Arnsverq pue JuewfoTdmeun oy a0q x

seotiad uofydunsuog usfo10g

dN) u8ter104g

yOOURTeq oper} osTpueyoisy sqaoduy wneTorjzed jo enTteA

uot zdunsuos uneTorzjed oTjsseu0g uneTorqzed Jo aofad Aaodwy 4%2981 [ITq Aanseory,

yx 23e1 QuowAOTdueUp

Joe Jep uozjdunsuog

1oWeTJEPp ANd ($ ZL61) aNd (g JUeaInS) aNd

2€ OTTBUDDS

*(quezAINdD) UOTTTTA ¢$ U} eoUeLEFJTP SAN Tosqe ey. st uMoYys eSuRyo eyQ faoueTeq oper} oy} AOq “squTod e8eqUeOAed UT UeATS ore saeBuRYO ‘seqzeA [ITq Aanseery pue JQusewhoTduEUN xy AOq x

88° 47- O€*T- c9o°¥ Ze°9 6S°€ 08°Z 97°7 LL° sootad uotydunsuop uster04 cy" 19°- ve°T- Z6°- 8L°- €g°- cL°- Il’- aN) ustes04 ev Lt 70°C€ 09°72 oc*c- 6Z°- 0S*T- G6°T- 89°7- yOOURTEG Spel} osTpuRYyoISW GL°€t- 9S°9€- ZI°O0%@- = 79°" LZ 12°6 08°6 CZ°OT 91° OT sjiodut wneTorjzed jo enteA OL°E- 8Z°7- 77° 9= LE*9- 10°? cO°€- 66°T- Z9°~ uotadumsuos wneTorjzed oT3sSeu0G 88°7e- 66° 0€- 16° 9 02° 7S 6S°2Z ST°LT 16° 9T cvIt uneTorjzed jo edtad AAtoduy I¢*- gc°- 0€°- LO’- 0° 10°- c0° zO° ¥2781 TITq Aanseery Ly- 09° 98° 6L° cc° tv ze" c0° %o7 eI JuoufsoTdweup 88° + cv e- Z8°- LO°T S9° OL° €L° €¢° 1oVeTJop uotydunsuop 12° 17° 7- 8T°- 6L° 9¢° 8c° gc° ST 1OIeT FOP AND €T°Z ZEr= 80°T- O€*T- 68°< 9L°- Li’- 9T° ($ ZZ£61) aNd 8I°7- ZL°t- 9Z°T- (Ae ero oh ie OT’ zO° (§ JUeaaInD) AND ~~ T86T O86t oOl61 £SZ61 £ZL6T 9L6T cl6ét vwlOl

1

(seqea Aaeek JO pus fTorARUOD WoOAZ VoUeASTFTp Jusd1eq) (queujsn€pe o3e1 asueyoxXe pue GND °S°N Teuywou ut AagSPATAMIT TAT ATIamMASNn Fnr\ noatarptnmte aat3y snonagoniag : b oT arma nc

li it in ans ai Sy An ia i 2 | is ia <

20F4d TIO PI70M pexepul JO Joes FA

9 TIAVL

*(qUeTIND) UOTTITqG $ UT BoUSeTEFFTP SANTOSqe oY} ST UMOYS esuURYS 243 ‘s0ueTeq eperj ey} AOq_ ‘squtod osequecrzed uF ueATZS ore sosueyo ‘soezer [I Tq Aansverz} pue quouko,dueun oy} 30g x

G9°€- c0° 89°S ee°l GL’ 4 Go°€ Gs°z 98° soofzad uotydunsuop usto107 G9°T 60°T e°T- 98° T- 9T*T- €0°T- €8°- (A ie aN) ustet0g 06°02 98° VE L6° 9% €6°C- Tc°- 6€°E- TS°2~ 66°?" yoOoUBTeq opetlj osSTpuryors9y] 8L° 42- 19° SE- €S°0@- —s_- OE ° SZ SY TT c€° OL €9°7I 8e° TT sjaoduy wnejTorjed Fo ante, YI°?C ¥e°e- €S*L- GS°L- TI°G- LL°€- €2°C~ 89°- uoftjzdunsuos uneTorjed oT] seul0g 92° 0€- LS°S?@- c6°- Ge°ls 64°62 G9°SZ Ge°07~ eL°Z1 uneToijzed Jo edtad Aaroduy 09°- 79° Le°- OT’- €0°- TO° 90° Z0° 4%2981 [ITq Aansvery, ve"- 6L° LO°T L6° 69° os’ GZ" 90° ¥27e1 JQuewhoTdweup 61°S- yG°€- 68°- €T°T €6° 66° €38° L¢e° JOWVTJOp uot zdunsuog 77° ye 9v7'°C- 6T°- T6° 08° LL’ 79° Ze° AO7BTFJOP AND L0°? HG‘ 9€°T- LS°T- ZT°t- 16° €S°*- 8t°- ($ ZZ61T) dNd 97° 7- 86'°S~ cc°T- L9°- €e°- LT°- TT’ 20° (§ JUerand) ANd

LoL 8L6T LL6T 9L6T SL6T

(soqei azeek Jo pue {TorqUOD WOAF BOUSAEZZTP JUdIeg) (seqei o8ueyoxe pue smOoUT puke seoyzad uSTerz0F pue *s*n UT eseoIOUT AOF JuSwASsNf[pe) uot Ie[MMITS eoTrg snousZopuy :¢ OFABUSDS 2°TId TO PIAOM pexepuy JO JO9FFg

YLOT

L WIdvi

T86T 626T LL61 SL6L €L6T

0

OT

02

O€ — G OfTazeUs0S —_ y OTzeUDS

Ov -_— € OTzeUD0S — — Z OTAPUDDS sevcee ee T of ateusls

| aemmer «= OSED DOUD TOF OY | yer1eq/¢

THHAVE WAd AOLUd TIO

I 32eY9

T86T 6L6T LL61 O€TT

OLZ |),

OvET s 7

“ao pa , w2e> LA ei Oty 6) ~NQy ee Cf fe f Pad ‘f /. iy x 08? _ F OSST

dND ‘Ivaa

Z 34B4D

— canta

eeoaecve

€L61

G oTzBUAaDS 4 OT 1BUaDg € OTAzBUsNS Z OF ABUsNSg I Ofazeus0s

BSB sdUeTsTOYy

T86T 6Z6T LL6T SL6T €L6T

06 OIT O€T OST Pg OLT . oe ry “ —_ Zo “7 3 7 o' ea We oe O6TT (+ — on G OTazeusDsg ‘7. ——- % OF ABeUDNS — € OFzeusSg _—— Z OTABUsDS ecoceee . I OTzeus.¢ Oz as «= OSED DOUDIOFOY

XHQNI HOIYd NOTLdWNSNOO

€ 43zeyuD