“Soviet Planning and Spatial Efficiency”
The analysis of the preceding chapters shows that cementmaking technology did not dictate a concentrated industry. Raw materials of comparable quality were found extensively in the USSR, minimizing regional cost variation. Scale economies are modest in comparison with high real transport costs. Therefore, an efficient location pattern for this industry is one which approximates the geographical demand pattern and a widely dispersed industry should be expected. Did the Russians plan new capacity consistently with this optimal location pattern?
One indication that the Russians did not plan cement location well is the extreme rise in the average length of haul. Table IV-1 shows the fluctuations in this measure for the cement industry over the last fifty years. During the 1930’s the average haul varied from 100 percent to 300 percent greater than it was before the revolution, and transport input came to account for more than half of true delivered cost. In this chapter we will determine how much of this transport cost arose out of inefficient location patterns.
A. HISTORICAL SHIFTS IN LOCATION
Table IV-2 shows the changes in regional shares in consumption and production between 1913 and 1940. Information on consumption is less easily available than that on production and can be given for only a few years.
The data shown in Table IV-2 refer to hydraulic cement production and consumption as a whole. Information on the distribution of individual cement types is very limited. The salient characteristics of what is known are shown in Table IV-3, which gives a regional breakdown of planned portland-slag capacity in 1931 and estimates of blended Portland output by region in 1940.
aColumn 2 contains estimates of the total industry all-mode average, which differs in some years from that given in column 3 because of incomplete coverage of the latter. For example, in 1931–36, the column 3 average refers only to cement marketed by Soiuzstromsbyt, which handled cement of Soiuztsement and the State Administration of the Metallurgical Industry only. There are also discrepancies because of different methods of treating truck shipments in the calculations. Truck shipments are not included in the adjusted average.
bHauls for 1912 and 1924/25 estimated from data in V. Sobolev, “Znachenie estestvennykh mergelei v tsementnoi promyshlennosti,” SM (1932), No. 11–12, p.44. The open-ended class in that source (shipments over 1500 kilometers) is considered to have a midpoint of 5000 and 1800 kilometers in the two years respectively. The first estimate is based on information regarding regional distribution of output found in V. Shneider and G. Brodskii, “Itogi i perspektivy razmeshcheniia tsementnoi promyshlennosti,” PKh (1939), No. 2. (These authors’ estimate of the average haul—350–400 kilometers—for the period seems a little low.) The midpoint of the second year is assumed to be much lower because of the different shape of the distance frequency distribution.
c1927/28 all-mode average excludes local industry. It is taken from Kazanskii 1936, p. 34. Rail average haul is from NKPS, Materialy po statistike putei soobshcheniia Vyp. 108, Svodnaia statistika perevozok po zheleznym dorogam za 1927/28 operats. god, Tom. H, Transpechat’ (1930), pp. 297–298, and is based on information for 72.5 percent of rail shipment in that year.
dInformation in columns 3—7 from Brodskii 1935a, pp. 8–9. Separate hauls are not given for water and rail/water, and the figure given here (536 2) was calculated as a residual. Railroad hauls for 1931 and 1932 only are from A. Smurov, “Bolezni tsementnoi promyshlennosti,” NS (1935), No. 14, p. 32. All information for Soiuztsement only.
eBrodskii 1937a, p. 36. Information in columns 3–7 for Soiuzstromsbyt only.
fBrodskii 1939, p.4.
g”Tsementnye zavody maloi moshchnosti,” Ts (1947), No. 2, p. 1. The figure for water is the weighted average of hauls reported there for river and sea. The rail haul and the amount shipped by rail are believed to reflect the rail portions of joint rail-water hauls.
hBelov 1957, p. 7.
iBelov 1959, p. 6.
Regional Production and Consumption of Hydraulic Cements in Thirteen Regions: Selected Years
NOTE: The regions listed correspond to those shown in the map in Figure V-1. Central Asia and Kazakh SSR, shown separately in the map, are consolidated into the Central Asia/Kazakh SSR macroregion for many purposes in the investment analysis of this chapter.
1940 information from Z. I. Loginov, Tsem. prom. SSSR (Moscow, Gosplanizdat, 1960), pp. 134–135. Loginov combines Belorussian consumption with that of the Baltic republics. Estimate given here is based on 1960 proportions in E. G. Gutsev, “ Perevozka mineral’nykh stroitel’nykh materialov,” Chapter II in E. G. Gutsev et al., Ratsionalizatsiia perevozok massovykh gruzov y BSSR, IzdateTstvo Akademii Nauk BSSR, 1960.
1937 data from Promyshlennost’ 1957, p. 279.
1936 production data from Brodskii 1938, p. 11.
1936 first half production and consumption from Brodskii 1937a. Brodskii does not give consumption for 1936 as a whole. The full-year basis estimates given here preserve the relative halfyear consumption shares, while the ratio between the full-year estimate of consumption and the annual production volume maintains the relationship existing between Brodskii’s first-half data.
It should be noted that production in 1936 rose sharply from the previous year largely because of an NKTP Bureau of Standardization Decree of March 1936 permitting the addition of hydraulic additives to Portland cement without changing its designation (see Chapter II, section A.1). The resulting product appears not to have been successful and to have led to a cutback in blending in 1937. In that year with portland-slag and pozzolan production almost equal to that of 1936, there was a reduction of about 12 percent in portland production at the same time that kiln productivity was almost unchanged (533 kilograms of klinker per day per cubic meter of kiln volume in 1936 vs. 531 in 1937). E. S. Shatalov, “Zadachi tsementnoi promyshlennosti,” Ts. (1938), p. 10. 1932 data from Promyshlennost* 1957, p. 279.
1929/30 information from “Novye metody sostavleniia balansa stroitel’nykh materialov i ikh raspredeleniia mezhdu stroiiashchimi oranizatsiiami,” NS (1930), No. 7–8, p. 274. Siberian information, there aggregated, here is divided between Western and Eastern Siberia in accordance with trends since 1927/28 (NKPS, Materialy po statistike putei soobshcheniia, Vyp. 108, Svodnaia statistika perevozok po zheleznym dorogam za 1927/28 operats. god, Tom II, Transpechat’, 1930), and 1933 (Brodskii 1935a).
1913 information from V. Shneider and G. Brodskii, “Itogi i perspektivy razmeshecheniia tsementnoi promyshlennost,” PKh (1939), No. 2. The authors left undistributed 10 percent of total production and 23 percent of total consumption. These quantities have been distributed here in proportion to the balances, except for Northern and Belorussian consumption, which has been based on the trends, and the Central industrial regions, which have been lowered slightly to allow for the consumption of the two aforementioned regions.
Portland cement blended with pozzolanic additives did have an advantage over straight Portland in certain types of structure, as noted in Chapter II. This may have necessitated the shipment of blends even to regions which were cement-surplus areas when all types were considered together. This in turn would have adversely affected the calculated average length of haul. However, as Table IV-3 shows, almost all major regions had their own source of blended Portland, the major exceptions being Leningrad and the Far East. But the need for blends should not be overstressed. As noted earlier, the main parameter was production cost, and the special characteristic of lower pozzolan hydration heat was rarely taken into account in the Soviet journals of the prewar period.
One important special-purpose cement was that used for lining and sealing oil wells. Most or all of this was probably produced at the Baku plant, near where it was consumed.
That the Soviet cement industry was unsatisfactorily located in the 1930’s came to be recognized and discussed after 1937—that is, after the end of the Second Five Year Plan. Some of the industry’s observers correctly point to the very long average haul as evidence of this inefficient location. Others1 point to the average haul as evidence of inefficient marketing, and in Chapter V this will be seen to be an incorrect view. Still other writers deny the validity of a comparison between the values of this measure in the 1930’s and in the prerevolutionary period, because of the radically altered demand patterns. But to the last objection may be counterposed the argument that any capacity distribution is supposed to represent the adjustments required by the demand pattern which has developed, and, because cement production is not seriously restricted by materials or scale, there is nothing to keep these adjustments from being easily made. An average haul which for ten years remained two to four times longer than that previously existing, while production increased by about 250 percent over the Tsarist period, strongly suggests that these adjustments were not being made.
Many Soviet writers do not even concede that the Tsarist industry was well located with respect to its own demand pattern. For example, Balzak, Vasiutin, and Feigin complain that the industry was concentrated largely in the Northern Caucasus (Novorossiisk area) and the middle Volga (mainly in the area of Vol’sk, whence it was shipped all over the country).2 P.N. Stepanov, writing in 1958, says that 80 percent of the prewar industry was concentrated in the aforementioned two regions and the central Moscow region combined.3 A third combination of regions is offered by D. V. Gorskii, who wrote that the Tsarist industry was 80 percent concentrated in the three regions of the Northern Caucasus, the middle Volga, and the Donets Basin.4 In truth, however, there were five important cement-producing regions: the middle Volga, the industrial center, Ukraine, the Northern Caucasus, and the northwest Baltic region. And contrary to the three sources cited above, there was also some hydraulic cement capacity in the Urals, Siberia, and the Far East.
Sources: aPlanned portland-slag capacity data from M. Bogachik, “Ispol’zovanie domennykh shlakov v tsementnoi promyshlennosti,” SM (1931), No. 10, p. 10.
bTotal production of portland-slag and Portland-pozzolan in 1940 from Promyshlennost’ 1964, p.320. Regional estimates based on 1931 plans and on fragmentary evidence in Brodskii 1937b; S. Shchukin, “Karadagskii trass i proizvodstvo trassovogo portlandtsementa,” SM (1930), No. 2; and D. V. Gorskii, “ Tsementnaia promyshlennost’,” SM (1937), No. 21.
Table IV-4 compares the regional distribution of hydraulic cement production before the revolution with the distribution of other major industries at the time and shows that cement production was more evenly distributed than total industrial production or any of the other major industries compared.
aAll totals exclude Poland but include the Baltic region.
b Cement shares for 1913 based on F. A. Shober, ed., Fabrichno-zavodskie predpriiatiia Rossiskoi Imperii, Sovet s’ezdov predstavitelei promyshlennosti, 1914. The Caucasus share includes 23.6 percent in Northern Caucasus and 2.0 percent in Azerbaijan (Tauz). All figures are percentages of reckoned capacity, believed to include almost all industry.
cCoal production for 1913 from Ministerstvo finansov, Narodnoe Khoziastvo v 1913 g., quoted in R. S. Livshits, Razmeshchenie promyshlennosti v dorevoliustionni Rossii, Izdatel’stvo Akademii Nauk (1955), pp. 276–277.
dIron and steel production for 1913 from Gornyi departament, Obschchii obzor glavnykh ostraslei gornoi i gornozavodskoi promyshlennosti, quoted in Livshits, pp. 284–285.
eCotton looms in 1910 from various sources quoted in Livshits, pp. 202–203.
fTotal industrial production for 1910 from various sources given in Livshits, Table 37. The entry there for western region (1.9 percent) and that for the Central Black Earth Region (3.0 percent) have been included here along with the Central Industrial Region under the heading “Center.” The Far East has been included by Livshits in Eastern Siberia, which, with Western Siberia and the Urals, has been put here into a single group.
B. LOCATION DECISIONS IN THE PREWAR PERIOD
1. Indicators of Regional Cement Demand
Table IV-5 shows the RSFSR cement requirements projected for 1932/33, the last year of the First Five Year Plan, and two proxy indicators of cement demand during the first two plans. Study of this table shows that rather substantial volumes of cement consumption in the east were foreseen, and production should have been located accordingly. The evidence suggests that Soviet planners did not expect to cover this need through additions to eastern capacity. For example, as Table IV-5 shows, it was anticipated in 1929 that by the end of the First Five Year Plan the eastern regions (Western and Eastern Siberia and the Far East) would raise their consumption of cement by 639 percent. At the same time, plans called for an increase in these regions’ production of only 256 percent (see below. Table IV-6). The result of these planned changes is that while the eastern region was 99 percent self-sufficient in 1927/28, only 46 percent self-sufficiency was planned for 1932/33 !5
The other major deficit region was Central Asia/Kazakh SSR, and here, also, the planners seem not to have intended to build the new capacity that seems to have been warranted. Although annual consumption in that part of the region which was in the RSFSR at the time—the Kazakh republic and Kirgizia—was scheduled to increase by 852 percent to 182,400 tons, total production of only 93,000 tons was planned for the terminal year of the First Plan.6 As it turned out, actual production in the whole Central Asia-Kazakh SSR macroregion in 1932 amounted to only 69,000 tons, as is shown in Table IV-2, and this took place not in the Kazakh SSR or Kirgizia but in the Uzbek SSR. The title list of new construction called for three new cement plants, but only one of these was undertaken during the First Plan.
The 1932/33 consumption projections in Table IV-5 cover only the RSFSR, and we have no comparable information for the Ukraine, the Transcaucasus, the Belorussian SSR, and part of Central Asia. (The RSFSR covered about 80 percent of national consumption.) However, we have an alternative indicator of cement requirements for the First Plan period which does embrace these regions—namely, capital construction plans for heavy industry (VSNKh). This is not a precise gauge, of course, since building requirements and materials input coefficients, even for the same type of structure, vary from one region to another. In addition, there are the usual problems connected with the use of Soviet prices. Finally, this series reflects only the construction plans of heavy industry, and the possibility must be recognized that construction of municipal infrastructure, housing, canals, and other facilities will not necessarily vary in proportion to heavy industry construction, and next to nothing is known about the distribution of cement use by sector (agriculture, transport, etc.) or by type of structure (dams, buildings, canals, railways, etc.).7
Indicators of Cement Demand by Region during the First Two Five Year Plans
Sources: a RSFSR Plan, Vol. II, pp. 170–174. Before 1933, Central Asia-Kazakh SSR excludes Turkmen and Uzbek Republics.
b USSR First Five Year Plan, Vol. III, pp. 568–569.
c Second Five Year Plan, Vol. II, pp. 260–261.
Column 5 of Table IV-5 shows the regional breakdown of planned construction for RSFSR state industry for the entire First Plan, and Column 6 shows the shares for the USSR as a whole. The RSFSR shares parallel closely the consumption projections for 1932/33 and, although the periods covered are respectively five years and one, this approximation tends to justify our use of VSNKh construction plans as an indicator of regional cement demand for the USSR as a whole. The most notable discrepancy between 1932/33 projected consumption and total planned construction for the whole period is in the Ural region. This may be due to the great outlay required directly for steel-making processes in the Urals. A large discrepancy in the other direction is observed in the Far East, where projected consumption in the final year is 28 percent greater than investment over the period, probably because of the different time periods considered.
For the Second Five Year Plan, total investment rather than VSNKh construction projections is all that is now available, and presumably Soviet planners at the time did not have better information. Column 7 shows total investment in the economy as a whole—investment in railways, municipal facilities, roads, residential construction, machinery, industrial and agricultural equipment, and so on. Objections similar to the heavy industry construction series may also be raised to the total investment series when used as a proxy indicator of cement demand, although the much broader coverage of the total investment series, which includes cement-intensive works such as dams, suggests that it would be more appropriate than the VSNKh construction series alone.
The total investment series for the Second Plan shows accelerated development in the remote regions—the Far East, Eastern and Western Siberia, Central Asia-Kazakh SSR, and the Transcaucasus. Thus the authors of the Second Five Year Plan should have foreseen a great need for cement in these developing regions; a total of 12.3 percent of total investment was planned for the three eastern regions, and 9.9 percent for Central Asia-Kazakh SSR. We will see in the next section that not enough new cement capacity was planned—much less put into place—to satisfy the growing demand of these two macroregions.
2. Production Capacity Decisions
As in the case of demand, so with supply, information is incomplete. For the First Five Year Plan there is a series on total planned investment in cement production facilities by region and also information on the location and cost of cement plants scheduled for construction during that period. There is no information on capacity (either clinker capacity or cement capacity). However, since at this time almost all cement was still pure Portland, inconsistency between rubles of investment and plant potential is minimized. For the RSFSR only there is also information on planned production increments by region.
For the Second Five Year Plan as a whole there is information on planned production increments by region. There are also lists of plants for construction. Many of these plants are carryover projects from the First Plan. For the Third Five Year Plan there is no information on regional investment or production plans for the cement industry.
Table IV-6 lists the available indicators in the various regions of the plans for expansion of the capacity for cement production. Analysis of this information for the period of the first two Five Year Plans and comparison with the developing pattern of demand which we have just analyzed shows that proper attention was not paid to expanding production in the eastern regions. This is revealed in several ways. In the first place, consumption for 1932/33 was projected at 349,020 tons, but production for that year was projected at 162,000 tons. The First Five Year Plan envisaged no new plant for the Far East (actually, around 1931 a large plant was begun at Spassk, and the first of two stages was completed in 1935). But a plant of 75,000–100,000 tons almost certainly could have been completed within the period if provision had been made for it, and even smaller plants based on the shaft kiln could have been introduced still earlier.
There were no projections of consumption for the Second Plan. The share of total national investment planned for the three eastern regions, however, was 12.3 percent (see Table IV-5), and this should have signaled the need for larger cement-making capacity. And indeed, a 150,000 ton plant was planned for Eastern Siberia, the second stage of the Spassk plant was to be completed (designed capacity: 350,000 tons), and large new works were scheduled for construction in Western Siberia near Novosibirsk and Stalinsk with a total capacity of 340,000 tons. As it developed, cement was first produced in Eastern Siberia only in 1945—eight years after the conclusion of the Second Plan; Spassk production was only 164,000 tons in 1937 and did not reach the 1937 target of 280,000 tons even by 1940, when production totaled only 234,000 tons. Production in Western Siberia, which was to reach 470,000 tons in 1937 with the new plants, actually was only about half that amount, and even in 1940 reached only 263,000 tons. In other words, in the first two plans, even when a limited extent of the need for eastern capacity was foreseen, the plans were not executed.
The first two plans provided for less self-sufficiency in the Central Asia-Kazakh macroregion than in the eastern region. Failure to provide for local production in the First Five Year Plan was mentioned above. This disproportion continued in the Second Plan, during which 9.9 percent of total national investment was scheduled for this region. By contrast in 1937, the terminal year of the Second Plan, only 4.4 percent of the nation’s cement was to be produced in this region. As it turned out, production of 2.3 percent of the national total did take place here, but that was just over half the total planned for the region. Moreover, since the planned share relates to the planned volume of 7,500,000 tons and the actual share relates to 5,454,000 tons, actual production was evidently less than 40 percent of the volume planned for this region in 1937. By 1940 the region was providing all its own consumption, but the consumption level of 267,000 tons was almost 20 percent below the production volume of 325,600 tons that had been planned for the region in 1937. Reductions therefore seemed to have been caused by industry bottlenecks rather than by independently determined reductions in demand.
We have so far been concerned with the deficit regions. But misplanning of location also implies errors in the allocation of investment to some regions. The surplus regions in the 1930’s were Ukraine; the Northern Caucasus, with production at Novorossissk; the mid-Volga; and, in some years, the central region. No new facilities were planned for the Northern Caucasus during the first plan, but production was scheduled to increase from 550,000 tons to 960,000 tons. There is no indication of how this was to be secured. The production of pozzolan at the time was still in its earliest stages (by 1930 total national production of pozzolan reached only 196,000 tons, and by 1933 it fell back to 110,000 tons).8 Actual production in the Northern Caucasus in 1932 reached 1,050,000 tons, but in view of the failure to build new facilities in this region during the First Plan, this increase is difficult to explain except by the possibility that some low-strength cement of the roman type was produced, increasing the throughput of the kiln. The suspicion is confirmed by a statement in the Second Five Year Plan regarding improvement of the product in this region together with a planned increase in production to 1.3 million tons.9 Actually, there was a small decline in the second plan, though perhaps quality did improve.
Approximately the same picture of rising and then falling production is observed in the mid-Volga region. But it was a mistake to cut back production there after 193 7, since the region was about 1,300 kilometers closer to the eastern markets than was the Ukraine, where production held steady after 1937. Ukrainian production should have been curtailed instead.
Finally, capacity in the central region (Moscow and environs, the central Black Earth belt, and the European west) was quite justifiably built up. Not only were there good raw materials and slag reserves in the various subregions, notably Briansk, Belgorod, Tula, and Lipetsk—just as there were in Ukraine—but also demand, as measured by capital investment plans, increased rapidly (it remained an approximately constant share of a quickly growing total), and this was not true of Ukraine.
The foregoing manifestations of the developing imbalance, between supply and demand for cement in the 1930’s resulted partly from failure to execute investment plans which had been adopted and partly from failure to provide for changing demand patterns in the first place. But since, as we have seen, the changing demand patterns were usually themselves foreseen, why was provision not made early enough for local supply in some of the regions most in need, especially the three eastern regions and Central Asia-Kazakh SSR?
The search for the underlying causes of the disproportions between supply and demand is frustrated by the absence of Soviet discussion of the location of the cement industry before 1937. Several hypotheses may be put forth to explain this development.
a) Lower Western Costs. The first explanation for preferential investment in the West is that costs were lower there—and indeed they were—on the average, over 50 percent lower than in the East. However, cement that was shipped very long distances to the Far East had to be packed in bags or barrels to guard against dust loss and damage en route. The packing and transport cost combined to raise the delivered costs of Ukrainian cement in the Far East to levels about four times the cost of locally produced products. For example, from 1932 to 1935 the price of the middle grade of portland in the Far East was over 51 rubles a ton while that of Ukrainian cement was 22 rubles. In Chapter III evidence was adduced showing that prices reflected regional production costs. Packing the Ukrainian cement in bags added 14 rubles a ton, and the loading and unloading costs of packed cement were a little higher, narrowing the cost differential to just 14 rubles, or one third of the Far East production cost. Then adding the 115 rubles of total transport cost for the 9500-kilometer journey raised the total delivered social cost of the Ukrainian cement to 152 rubles—200 percent more than locally produced cement.10
Indicators of Planned Cement Supply by Region during the First Two Five Year Plans
Sources: aRSFSR Plan, Vol. II, pp. 170–174.
bSecond Five Year Plan, Vol. II, pp. 335–579, and USSR First Five Year Plan, Vol. III, appendix pp. 10–72.
cSecond Five Year Plan, Vol. II, p.279.
b) Gigantomania. The gigantomania that afflicted many Soviet planners throughout the 1930’s caused no or only slight locational imbalance in the cement industry.11 As noted in Chapter III, construction of only one of the huge 465,000-ton plants projected in 1930 was undertaken—the Gigant plant near Moscow. Most plants were of moderate size, and although all of the capacities of the forty-five plants listed for construction during the first plan are not known, their planned average capacity was around 90,000 tons. However, though gigantomania as it is usually conceived did not show itself in this industry, it is fair to say that the potential usefulness of the very small plant based on the shaft kiln was not exploited as it might have been. Small plants based on the shaft kiln might have been a useful technological variant in some areas, including the major centers of the Kirgiz, Tadzhik, or Turkmen republics in Central Asia, all of which had an average haul of 1000–2000 kilometers in the 1930’s,12 as also in the Kazakh republic, with its actual average haul of over 2500 kilometers. Small plants could have been placed in several areas in this republic. One other region where the small plant might have been utilized was the European North around Archangel, consuming around forty to fifty thousand tons, which at the time it was importing from Novorossiisk. (Importing from Moscow might have been an even better solution.)
c) Raw Materials Advantages in West. The neglect of the eastern regions might be explained by the absence of suitable raw materials, or the failure to discover them during the time of the First Plan. The map in Chapter II is taken from a recent source. Is it possible that the deposits of lime materials shown in the map were not known at the time so that planners were not aware of the possibility for undertaking local production? The answer is negative. In the first place there had been production before the revolution in the East in Yenisey and Transbaykal oblasts, at Spassk, and in Tomsk guberniia.13 In the second place, production at Spassk continued in the 1920’s and totaled 18,150 tons in 1927/28, as shown in Table IV-6. Finally, an article in 1939 shows that the reserves of cement materials in the East were known and estimated at 16-25 percent of total national reserves.14 Those in Central Asia-Kazakh SSR were also known and were estimated to be about 3 percent of the national total,15 which would permit substantial production in this region. This information is believed to be accurate for the period and must have been available for many years. Moreover, Kholin’s directory lists many plants built in recent years which are precisely the areas close to the urban centers of Eastern and Western Siberia, the Kazakh SSR, and the Central Asian Republics16 so that it cannot be said that the raw materials which did exist in the macroregions were located far away from the major consumption centers. And reference to the distribution of coal reserves in the 1930’s shows that the east and Central Asia-Kazakh SSR were well endowed with this fuel.17
d) Better Operating Conditions in the West. Although there were really no fundamental technological obstacles to location in the East, it might be argued that capacity was properly located in the West because slowdowns caused by materials or labor shortages arising in failure to make complementary investments would be minimized in the West. But from plant information on operating rates presented in 1936 by Khrennikov and Meliushev, it is evident that the placing of capacity in the West was no guarantee of uninterrupted operation.18
e) Cement Pricing Policy. The four technological factors just presented are inadequate to explain the locational imbalance of the 1930’s. There are two economic factors, however, whose effect in distorting location must be considered. The first of these is the pre-1941 pricing policy for cement which, a priori, does not appear to promote rationalization of location or distribution, since the user had no say about his source of supply. The cement industry paid no penalty for using a less than optimal distribution or location pattern, for the user had to pay all freight costs. This pricing policy was outlined in Chapter III, and only its implications need be recalled here.
f) Railroad Pricing Policy. A second economic factor conducive to inefficiency in location is the structure of railroad rates. The charge per kilometer before 1939 decreased rapidly with the increasing length of haul, although there are reasons to believe (see Chapter II) that the relevant real carrying costs did not decrease. Such a rate structure might lead to a locational pattern with substantially higher real costs than a second pattern even if the industry attempted to minimize buyers’ freight charges, for the industry might find that further scale economies in production could be gained at the expense of only slight added freight costs to buyers, because of the longer shipments. Although there is no record of discussion of the transport issue in relation to cement location before 1939, it is reasonable to believe that the planners themselves considered the trade-off possibilities. The understated cost would in that case have encouraged capacity increments in the West. There may have been some spillover effect from the Ural-Kuznetsk project as well; if the steel industry could be so cavalier about the long distances involved in the project, why should not the cement industry as well?19
Transport considerations did influence one locational decision early in the decade, but with perverse results. Because of the long haul of cement from Novorossiisk to Leningrad and the high demand in the Leningrad region, the planners felt an urgency to expand production there. It was decided to do this by expanding and modernizing the capacity of the Leningrad plant from 112,000 tons to 186,000 tons. The raw materials base was of poor quality and limited size, but that was not fully realized until 1934—several years after the project was undertaken20—and there is no coal in Leningrad. The First Plan also called for a plant at Pikalevo, about 250 kilometers from Leningrad, which had a superior base. This plant was still not in operation by 1939.21 Production here would have been relatively high in cost, owing to the shortage of fuel, but it would have been preferable to production in Leningrad, and the plant should have been constructed in preference to expansion of Leningrad facilities.
Before 1937, when transportation was discussed, it was usually from the point of view of rationalizing marketing—that is, the distribution of the output of fixed facilities rather than location. By the time of the Third Five Year Plan it was recognized that mislocation of the industry was having a serious impact on transport utilization. But the post-1937 analysis of the locational problem is not incisive; Brodskii and Shneider, for example, reveal an attitude of “build where production costs are low and also where there are shortages” that solves nothing, and their specific recommendations call for investment in every region except (understandably) the Northern Caucasus, and (inexplicably) the Far East.22 Kazanskii comes closest to anything resembling careful analysis in a table in which he tabulates average production cost and average length of haul for the output of nine plants. This is still a long way from anything definitive. The plants are believed to vary widely in their output mixes, and, since it may be part of an irrational shipping pattern, the existing average length of haul is not relevant to the location of new capacity. But even this small and inadequate table is not analyzed—it is simply presented with a few accompanying words.23
C. COMPARATIVE COSTS OF ALTERNATIVE LOCATION PATTERNS
In this section we will compare the total social cost of cement consumption in 1940 under the regional production pattern as it actually did develop and under an alternative optimal location variant.
Table IV-7 provides the basic data required for the analysis. We consider prewar boundaries and the same regionalization used in earlier tables. Columns 1–4 relate to the actual production pattern, while columns 5–10 are relevant to the calculation of the cost of an alternative pattern. Column 1 gives the regional production level in 1940 as shown in Table IV-2, except that since consumption (5,268,000 tons) fell short of production (5,578,000 tons), we have reduced production by 310,000 tons. Column 2 gives the regional mill prices for grade 00 portland as shown in Table II-3. These prices are believed to reflect accurately the cost of this grade Portland in the various regions (see Chapter II, section C).
Column 3 represents an attempt to estimate average regional production cost based on the 1932–35 price list which took effect in 1932. This adjustment takes account of each region’s product mix and makes use of the regional distribution of portland-slag and pozzolan production. The costs of these cements were estimated by adding to the cost of the portland component, which is 50 percent for pozzolan and 20 percent for portland-slag, the cost of processing and grinding the hydraulic additive. This cost for each of these additives was estimated to be 60 percent of the production cost of portland cement. The resulting costs estimated for pozzolan and portland-slag are 80 percent and 68 percent of the cost of grade 00 portland respectively. These ratios are somewhat lower than the relative mill prices as compared with Portland. Since slag is most important in Ukraine and pozzolans are most important in the West, this is a conservative approach. It helps to justify in these regions the increased production that developed during the decade. The reader might find it of interest to vary the estimated production cost of any particular region and observe the effect on total production costs.
Analysis of Delivered Costs of Cement in 1940
Finally, column 4 gives the cost of producing in each region the amount that actually was produced. The entries in this column are the product of the entries in columns 1 and 3. The total cost of producing by this regional production plan is 138.3 million rubles.
Next we must calculate the delivery costs incurred for this volume. In Chapter II an estimate of 1.2 kopecks was developed as the total ton-kilometer social cost of cement in 1932. It was reasoned that the approximate constancy of the unit charge imposed by the new rate structure in 1939 was characteristic of the real cost structure for hauling cement earlier in the period (this cost does not include the loading and unloading cost, which must be added). We can now multiply this cost by the number of railton kilometers recorded for 1940 in order to get a total rail transport cost and add to it the water transport cost.
The input coefficient for physical transport for 1940 is not known. It is known for 1939, however, as is the breakdown between water and rail use. This was shown in Table IV-1. From this information and from data available on the cost of shipping by water (see Chapter V), it is possible to estimate the total cost of transport in money. Total rail-ton kilometrage will be 6,728 million, which yields a rail-carrying cost of 80.7 million rubles. Transportation by water will amount to 2,278,000 rubles (the average cost, including handling or transloading, where necessary, of 1.35 kopecks per ton kilometer). To this must be added the loading and unloading cost for the 4,973,000 tons that went by rail, or 8,951,400 rubles, giving a total production and shipping cost of 230,306,200 rubles, including 138.3 million for production and 92.0 million for transportation.
The cost of an alternative regional location pattern is shown in columns 5–10. Column 5 gives the production required in each region to correspond to the consumption requirements listed in Table IV-2,24 with the exception that self-sufficiency is not planned for the North or Northwest. Instead, these two regions are to be supplied from the central region and Belorussia respectively. The reason for treating the North in this way is that the delivered cost from the Center is less than the estimate of the production cost in the North (48 rubles); the Northwest is handled in this way because Belorussian cement delivered there is somewhat cheaper than that produced in Leningrad (at the Vorovskii plant) and is believed to have been of superior quality. And since cement from Pikalevo in the Northwest delivered to Leningrad would have exceeded the cost of Belorussian cement, the alternative variant here presented calls for importation from Belorussia. Belorussian cement rather than cement from the Moscow region is proposed here to minimize congestion in the Center, although the run from Moscow is somewhat shorter than the haul from Krichev in the Belorussian SSR (see Table A-2 in appendix to Chapter V).
Column 6 contains regional production costs for this alternative locational pattern. The costs used there are those of column 3, which were used in estimating the cost of the actual production pattern. Eastern Siberian cost is estimated to be slightly higher than that in the Far East. The total production cost under this variant is 148.2 million rubles, 7 percent more than the estimated cost of the actual production pattern. The main differences between the two location patterns are seen to be in lower production levels of the Ukraine, the Northern Caucasus, and the Volga region and higher production in the eastern regions in the alternative pattern.
Column 7 contains the estimated average haul within each region. In some of the regions, for example, the Center, production originates at several points. The basis for the estimates of the average haul is explained in Chapter V. Column 8 is the result of multiplying the average hauls of Column 7 by 1.2 kopeks, which is the estimated transport cost as explained in Chapter II, note 32. This column gives the estimated average carrying cost per ton of cement shipped within the various regions. Column 9 is the product of multiplying column 8, the average rail cost per ton, by column 5, the relevant tonnage, and gives the transportation cost to supply each region. The total carrying cost is seen to be 24.6 million rubles, a saving of 70 percent over the actual carrying cost. Finally, column 10 gives the total production plus carrying cost. For the shorter distances involved under this variant, however, all movement could have taken place in bulk, which would have reduced the handling cost to 1.6 rubles per ton. Applying this cost to the 5268 thousand tons gives a handling cost of 8,428,600 rubles, making a total production and transport cost of 181,256,200 rubles, which in turn represents a saving of 21 percent over the corresponding cost of the actual pattern.
Still other costs involved in the actual production pattern could have been avoided by the alternatives presented here. In the first place, with the relatively short hauls of the alternative plan, dust loss would have been lower. The average haul implied by the alternative scheme is around 400 kilometers. Since part of the dust loss is incurred in loading and unloading and only part en route, the saving would not be proportional to the reduction in average haul, which is about 70 percent. Rather, a reduction in dust loss of around 40 percent might have been achieved. Therefore, we estimate the average dust loss on bulk shipments as 6 percent rather than 10 percent, as in the actual pattern. However, with the shorter hauls all cement may be assumed to have been cheaper to transport in bulk than the 60–70 percent (see Chapter III, section C. 2) shipped in bulk in the actual pattern. This loss ratio must therefore be applied to total shipment in the alternative plan, and in the actual pattern the larger loss ratio must be applied to only 60 percent of the total. The losses calculated in this way are 16.2 million rubles and 11.4 million rubles for the actual and alternative plans respectively and are subtracted from the total delivered costs of the two plans in reverse order. That is, the location variant proposed here would have permitted a saving of 16.2 million rubles, which was the cost of the dust loss in the actual plan, and the actual plan should reflect a saving of the 11.4 million rubles that the alternative plan would involve.
As just noted, it would have been possible to ship all cement in bulk, given the shorter hauls of the alternative plan, except possibly over damp routes; but since the alternative location pattern has been based entirely on rail shipment, instances of loss by moisture would have been quite rare. Therefore, no packing costs need be added to the delivered cost under the alternative plan, and in the actual plan a cost of around 30 million rubles is indicated, on the basis of information presented earlier on packing procedures. This figure reflects the cost of barrels and bags and the use mix (see Chapter III, section C. 2). These several cost adjustments are shown in Table IV-7, where the total saving possible through the alternative-location pattern is seen to be 83.8 million rubles—33 percent of costs actually incurred!
The savings possible through the new location patterns are substantial. Is there any chance they have been overstated?
Possible causes of overstatement in the savings attainable through the alternative-location pattern include underestimation of the intraregional lengths of haul in column 7 and overestimation of the transport input involved in the actual pattern. Even if the intraregional hauls had been twice as high as our estimates, however, a substantial saving would still have been achieved—58 million rubles or 24 percent of delivered costs. But checks on the accuracy of all estimates of distance calculated for use in Chapter V indicate that the intraregional estimates are not likely to be off by more than a few percentage points. By varying the coefficients in column 7, the reader may easily see the effect on total cost of any assumed intraregional average length of haul. For example, if the Far Eastern intraregional average haul were assumed to be three times higher, total carrying costs within the Far East under the alternative-location plan would rise by 3.8 million rubles, raising total delivery costs by only 2.3 percent.
The total social overexpenditure of resources caused by the incorrect location of the industry in 1940 has been estimated at 33 percent. In addition, the alternative-location pattern here presented envisages more portland cement production (in the Far East and Eastern Siberia) and less portland-slag and pozzolan production which, as they actually were produced in the Soviet setting even if not in theory, were seen in Chapter II to be inferior in their concrete-making capacity.
During the period of the prewar plan the Soviet cement industry came to be very inefficiently located. The most marked indicator of this inefficiency is the industry’s average length of haul, which through most of the period averaged 70 percent longer than the 1927–28 average and around 200 percent longer than the average in the prerevolutionary industry. This increase resulted primarily from the increased demand of the Eastern and Central Asian regions, where capacity to keep pace was not added. As we concluded from evidence of contemporary cement consumption projections and capital construction plans, however, these needs ought to have been clearly anticipated, and, given the relatively simple nature of cement technology and cost functions, capacity should have been emplaced in these regions.
A precise notion of the cost of the developing locational imbalance was gained through our analysis of 1940. Using estimates of 1932 real social costs for production and transportation, we calculated that 33 percent of the cost in 1940 represented avoidable waste. This means an overexpenditure of about 50 percent over what need have been spent under a regime of economic efficiency.
Three factors seem to have contributed to the poor performance. First, the industry busied itself chiefly with the minimization of production cost rather than total social cost, and new output could generally be secured more rapidly in the west, although not at costs low enough to offset the high transport costs. In the second place, the railroad costs were understated, so that even when the industry may have been interested in minimizing total costs, it probably based its decisions on the railroad rates confronting it rather than the true social costs. There was a major discrepancy between the two before 1939, when the average ton kilometer costs declined with increasing distances much more than did average costs. Finally, even when observers were concerned with the total picture—that is, with minimization of costs to the economy—they chose the wrong target, as the writers quoted in the next chapter indicate. When they were concerned about the long average haul, they generally supposed that the solution lay in rationalization of distribution rather than in rectification of the locational imbalance, and this attitude undoubtedly begot complacency with respect to the planning of location in the industry. As we will see in Chapter V, however, distribution was not the real problem at all.
We use cookies to analyze our traffic. Please decide if you are willing to accept cookies from our website. You can change this setting anytime in Privacy Settings.