This paper estimates the capacity utilization rate for Russian manufacturing. We also propose a way to build continuous production capacity time series and indicators to describe the basic characteristics of production capacity. The data come from form 1-natura-BM of the Russian Federal State Statistics Service. Our findings on the trends and structural characteristics of production capacity are shown to be significant for economic policy since we found that in recent years capacities utilization rate in Russian manufacturing industry has been not extremely high and that there is a strong correlation not only between capacities utilization rate and inflation rate but between capacities utilization rate and capacities commissioning intensity as well.

production capacity, manufacturing, capacity utilization rate, capital investments, fixed assets, Russian economy

This paper investigates the opportunities and industry structure for non-capitalintensive industrial production growth in Russia and the utilization and structural characteristics of production capacities (PC) in the manufacturing industry. We will formulate proposals for monetary policy and industry priorities for stimulating structural policy and for improving the industry statistics.

This paper's relevance follows from a lack of detailed post-Soviet estimates of manufacturing industry modernization by the expert community and public agencies, the fragmented nature of the basic concepts for measuring production potential, and insufficient investigation into the real utilization of PC. At the same time, the level of capacity utilization is directly or indirectly referred to (through the output gap estimates) by those discussing the optimal parameters for monetary policy.

In reviewing global practice, detailed analyses of PC trends and characteristics are usually made for individual products as part of specific industrial sector research. Meanwhile, there is a lack of cross-sectoral industrial capacities analysis. The main reason for the current state of affairs is apparently the lack of comparable data gathered across a wide range of sectors using a single methodology.1 At the same time, direct and indirect estimates of production capacity utilization rates have been widely used for many years as a significant business cycle indicator.

Three main approaches to measuring PC2 utilization are typically used according to global practice.

The first approach is based on surveying managers who estimate company's capacity utilization as a percentage of a given level. This is a common approach used by many countries and international organizations for promptly monitoring the economic “health” of the manufacturing sector. For example, in surveys by the U.S. Census Bureau, the respondent must measure a company's production output by fully utilizing the PC but without changing the working mode (Federal Reserve, 2016). In European Commission surveys (European Commission, 2016, p. 28), respondents specified PC utilization as a percentage of the maximum level. In the OECD surveys (OECD, 2003), the PC utilization rate is measured as a percentage of “normal” utilization rate.

The second approach is focused on measuring PC utilization for specific products as ratio of actual and maximum levels where maximum level is determined by characteristics of the equipment used (“engineering” approach). For example, the U.S. Federal Reserve uses both the Census Bureau surveys and PC utilization data in physical terms (for individual industries), gathered by industry associations and institutions. Similar production capacity estimates are also used in Japan (METI, 2015) and India (Reserve Bank of India, 2011).

The third approach uses an indirect measurement of capacity utilization rates based on certain models and assumptions. The simplest methodologies for this approach include linear interpolation of values between local maximums and interpretation of those values as a potential output, the Wharton Business School Index (Klein and Summers, 1966), a comparison between current output and the output calculated as a ratio of capital and minimum capital-output ratio within a local neighborhood,3 and a comparison between the average and maximum hours worked and capital used (Taubman and Gottschalk, 1971; Beaulieu and Mattey, 1998). In the framework of this approach, more complex methodologies should be mentioned: the optimal utilization level would occur if either short-term costs are minimized (Klein, 1960; Friedman, 1963; Hickman, 1964; Morrison, 1985; Prior and Filimon, 2002; Ray, 2015), short-term output is maximized (Johansen, 1968; Fare et al., 1989; Ray et al., 2006), or profit is maximized (Coelli et al., 2002). Recently, the structural and cyclical components of economic trends have been isolated using filters or structural vector autoregression methods (Andrle, 2013; Havik et al., 2014; Apokin et al., 2014; Bank of Russia, 2014; Sinelnikov- Murylev et al., 2014; IMF, 2014), along with econometric analyses of the correlation between changes in the core inflation rate and capacity utilization rates (McElhattan, 1978; Oomes and Dynnikova, 2006; Mironov and Kanofiev, 2014). Finally, a number of papers estimate capacity utilization rates using indirect indicators that describe the consumption of electricity and other basic resources (Foss, 1963; Jorgenson and Griliches, 1967; Anxo and Sterner, 1994).

There are four data sources for capacity utilization rates in Russia today (Table 1). Despite the variety of data sources, it's very difficult to get the reliable estimates for capacity utilization rates (and certainly for other capacity characteristics).

Table 1Download as CSV 

Basic characteristics of data sources.

	Rosstat (Business Activity of Organizations in Russia)	Russian Economic Barometer (REB)	Gaidar Institute for Economic Policy	Rosstat (Form 1-natura-BM)
Type of data	Surveys	Surveys	Surveys	Statements
Level	Sectors	Overall industry	Overall industry	Products
Frequency	Month	Month / quarter	Month	Year
Beginning of publications	Jan 2006 (OKVED) 1995 (OKONH, Russia Classification of Economic Sectors)	Dec 1991	Jan 1993	1992

Source: Compiled by the authors.

Rosstat estimates are compiled based on the results of surveys among executives at large and medium-sized companies; the aggregated capacity utilization ratio for the manufacturing industry is published in Business Activity of Organizations in Russia (as a percentage of the maximum4). Estimates by sectors are available as well.5 However, there are two major flaws in Rosstat's estimates. First, we identified a stable shift in estimates, they are understated because Rosstat assumes that non-responding companies had a zero capacity utilization rate. Second, Rosstat receives aggregated estimates across activities, without taking into account the size of the companies and sectors.

REB data on capacity utilization rates are compiled based on managers surveys. It should be stressed that, unlike the Rosstat capacity utilization ratio (percentage of the maximum possible utilization), the REB indicator measures utilization against the “normal” level, determined independently by respondents (it may exceed 100%), which is why these ratios cannot be directly compared. Speaking about drawbacks of the indicator, we should note that it is calculated on a monthly basis, but the bulletin is published quarterly, reducing the opportunity to promptly use the estimates. Moreover, the results are only presented for the industry as a whole and look distorted, while the calculation does not account for the size of companies and sectors (indicator for the industry is calculated as simple arithmetic mean).

IEP (Institute for Economic Policy) data (Monitoring of Russia's Economic Outlook monthly bulletin) are compiled from industrial company directors in response to the following question: “What is the current level of production capacity utilization in your company (%)?” However, they are not published regularly and are available publicly only as separate values for different periods, which is insufficient for a regular analysis. Combined with the limited openness of the methodology this does not qualify as an indicator for use in this paper.

Rosstat's statistical form 1-natura-BM “Information about the Production, Shipments, and Production Capacity Balance”,6 represents the most complete source of information about capacities. The form cannot be used for real-time monitoring, as it is published annually, and only six months after the end of the reporting year. However, it offers highly detailed data about capacities for several hundred products disaggregated by the regions of the Russian Federation.

The PC of a company in the form are determined according to an “engineering” approach as the maximum possible output of product per unit of time at full utilization of the industrial equipment. The PC balance contains approximately 20 items (Table 2).

Table 2Download as CSV 

Production capacity balance indicators.

No. Line description
1 PC at the beginning of the year
2 PC increase, including:
3 Creation and expansion of facilities
4 Reconstruction of facilities (improving the technical and economic level of production through expanding or reconstructing existing production areas in a company)
5 Technical upgrades for facilities (improving the technical and economic level of production without expanding or reconstructing existing production areas in a company)
6 Leased equipment
7 Changes in the composition of products manufactured (conversion, decreasing labor intensity)
8 Other factors
9 PC reduction, including:
10 Changes in the composition of products manufactured (conversion, increasing labor intensity)
11 De-commissioning (due to the old age of equipment, transition to new types of raw materials, calamities, etc.)
12 Equipment lease-out
13 Other factors (sale, bankruptcy, or liquidation of companies, etc.)
14 PC at the end of the year
15 Annual average capacity during the reporting year
16 Product output during the reporting year
17 Utilization of annual average capacity during the reporting year (%)
18 Product output using non-core facilities
19 Product output during non-working hours

Source: Compiled by the authors based on Rosstat Order No. 320 dated July 15, 2015.

We directed our main efforts at compiling PC balance data based on Rosstat's highly disaggregated form 1-natura-BM: we gathered data on 388 product items for the 2010–2015 period (OKPD classifier, All-Russian Classification of Products by Economic Activities) and 411 product items for the 1995–2009 period (OKP classifier), with a breakdown by Russian regions for the 2000–2014 period. There are approximately 2.5 million primary data points.

It is advisable to use time series to analyze the structural characteristics of PC. However, none have been built, until recently, for the Russian economy. We decided to fill the gap.

The key issue in building PC time series is interpreting the annual deviations between statements (between the year-end data of year t statements and the year- beginning data of year t + 1). To verify the approach, we studied a considerable number of actual deviations of this type and their representation in statistical reports and can confidently state that — at least in the case of Russian economy — it would be most correct to interpret these deviations not as sample shifts but as real movements of capacities (opening or closing down a factory, or fully converting it). Regretfully, the limitations of the article do not allow us to include a detailed rationale behind the approach used by citing the full range of these types of cases7.

In building the PC time series, we accounted for the change in the product type classifiers (the transition from OKP to OKPD at the beginning of 2010). To this end, we improved the existing OKP-OKPD transition key (it lacked a number of product items for PC balances). Changing the classifier did not allow us to interpret the deviations between the end of 2009 and beginning of 2010 as a real change in capacity.8

In the new series, we corrected for the gaps in the statements, which is especially important for regional data.9 Upon identifying a regional and national gap, we interpolated all data; upon identifying only a regional gap, we interpolated just the regional data and added the regional PC estimate obtained to the national value.10 Missing data on PC at the end of the year were treated as equal to the data on PC at the beginning of the next year (equivalent corrections were made for missing data on PC at the beginning of the year). The difference between PC at the end and the beginning of neighboring years was interpreted as the increase or reduction of capacities. This decision is based on studying a large set of particular cases of changes in capacity reported in the statistical form 1-natura-BM.

The widespread approach is to review the PC increase, reduction, renovation, and liquidation ratios. For Russian industry, capacities estimates were calculated recently (Zamaraev and Marshova, 2015).

The PC increase ratio kv and PC reduction ratio kw are calculated as follows:(1)where c(i)t is the i-th indicator (i-th PC balance line in Table 2) per year t.

The renovation ratio kr and liquidation ratio kl are calculated from the following PC balance lines:11(2)

Certain doubts arise regarding this approach, as the renovation and, particularly, the liquidation ratios are quite small even in sectors with well-known high investment intensity (e.g., in food production). At the same time, the PC increase and reduction ratios are extremely high, so much that Zamaraev and Marshova (2015) noted their “degenerated” nature (authors noted that such indicators fail to reflect the real changes in PC).

Relying on sufficiently detailed data on changes in PC at the regional level, media publications and company press releases, we analyzed a significant number of cases of real movements of PC by individual products. As a result, we confirmed that:

1. a change of ownership is inversely reflected in increase and reduction of PC and is reported as the change in PC due to “other factors”;
2. a company's bankruptcy is reported as the reduction of PC due to “other factors”;
3. PC leases between companies (usually within a given region, between adjacent regions or within a single company) are reflected in PC balances as “inverted” increase and reduction of PC.

We also discovered a number of new phenomena that are not in line with the classic methodology. We found that the emergence of a new factory is often reported as the increase of PC due to “other factors”12 or as a discrepancy between data at the beginning and at the end of adjacent years. Shut-down of a factory is often reflected in the reduction of PC due to “other factors” or as a discrepancy between data at the beginning and the end of adjacent years. Moreover, real growth in the PC is often reflected in the increase of PC due to “other factors.” Finally, a top-down conversion of PC accompanied by modernization may be reflected as a “change in composition” (accounts 7 and 10 of the balance).

The numerous findings prove that traditional liquidation and renovation ratios understate the real volume of new equipment de-commissioning and commissioning. Therefore, we suggest modifying formula (2), considering the actual specifics outlined above:(3)where klt is the original liquidation ratio per year t.

Accordingly, the renovation ratio should be calculated as(4)where krt is the original renovation ratio per year t.

In addition, we introduce the net commissioning ratio and de-commissioning ratio of old facilities:(5)(6)

The suggested modification makes a considerable correction to the understanding of the actual PC renovation processes in the Russian economy. For example, according to the standard approach, only one-third of manufacturing facilities were less than 10 years old in 2015, while, according to the adjusted methodology, such facilities accounted approximately for one-half of all facilities.

Other specific features of the methodology used in the paper include the following. PC at the industry level were obtained by aggregating the measurements by product type, using annual average prices of 2012.13 All estimates are provided only for civil products (dual-purpose or military products are excluded from the analysis). Mining and quarrying was evaluated without oil and gas production. Vehicle production accounts mainly for automobiles and railway rolling stock, as data on airplane and helicopter production has only been published since 2010.

We also calculated some additional indicators. First, we broke down output growth by factors (the commissioning of new PC and additional utilization of existing PC), separately by products and regions, based on the hypothesis that new PC are utilized first. We assumed that new PC, commissioned in a given year, can be utilized additionally during the next three years. Second, we calculated the capacity utilization ratio of newly commissioned capacities (the level of capacity utilization ratio which stimulates companies to increase their PC) as an average weighted ratio for products and regions in year t – 1, with weights corresponding to the PC values in year t – 1 and commissioning in year t across all regions where new capacities were commissioned.

A comparison of PC utilization estimates from different sources (Fig. 1) shows that, despite slight differences between them, the estimates appear to be very highly correlated in time. This fact supports the reliability of the PC utilization estimates obtained.

Fig. 1

PC utilization level in manufacturing companies (as a percentage of maximum level).

Source: Hereinafter in figures, Rosstat data and estimates by the Center for Macroeconomic Analysis and Short-term Forecasting (CMASF).

Since 2000, the capacity utilization ratio grew steadily, peaking in 2007 (at 72%). After the fall during the 2008–2009 crisis and the subsequent rebound, the ratio hovered at approximately 70% in 2011 and 2012, followed by a gradual decline to 66% in 2015.14

These first results alone show that there were evidently some excess PC, at least from the beginning of 2013: in this year, the ratio decreased not just below the maximum of 2007 but below the level of the two preceding years. Moreover, in 2014, the capacity utilization ratio continued to decline, so that some excess PC definitely existed. Based on this tentative and extremely simplified estimate (an “upward estimate”, as shown below), excess PC account for at least 3–5 p.p. of PC, which corresponds to a possible increase in output for 4.5–7.5% in case of fully utilizing these PC.15

During the first phase of the post-crisis16 development (2001–2005), the PC level hardly increased until 2006 (Fig. 2), while the intensity of both PC de-commissioning and PC renovation rates did not exceed 3–4% (Fig. 3); on the whole industry level one can conclude that the output increased through additional utilization of existing PC (as we found, the process was more complicated: new PC were commissioned and launched while existing PC were utilized more and old PC were utilized less, see below). During the second half of the 2000s, the situation changed considerably: the intensity of commissioning increased sharply while the intensity of de-commissioning remained the same, which gave rise to a new trend, i.e., PC growth until 2015. At the same time, the 2008–2009 and 2014–2015 crises only slowed but did not stop the intensity of PC growth. The deceleration was caused by less commissioning; the intensity of de-commissioning hardly changed since 2005.

Fig. 2

PC volume and age structure (2000 = 100).

Fig. 3

PC renovation and liquidation ratios (%).

Our approach allows obtaining the decomposition of output growth into the two components: growth due to commissioning of new PC and growth due to additional utilization of existing PC.17 It is usually considered by Russian experts that during the first half of the 2000s, output increased mainly due to additional utilization of existing PC. The estimates we obtained demonstrate that this factor has been overestimated for the period of early 2000s: according to our calculations, it accounted for slightly less than a half of the output growth, while PC commissioning accounted for the remainder (Fig. 4). At the same time, the role of additional utilization was underestimated in recent years (it is usually held that the additional utilization was no more important during the second half of the 2000s). Our research shows quite the contrary: in 2012–2014, around one-third of output growth was caused by additional utilization (this figure was only 1.5 times higher in 2001–2004).

Fig. 4

Decomposition of output growth by factor (RUB trillion, 2012).

PC changes in physical terms are highly differentiated by economic activities (Fig. 5). Four groups should be identified.

Fig. 5

Changes in annual average PC volume, 2015 compared with 2000 (%).

Group A. Superleaders — PC more than doubled in 15 years. This group includes manufacture of rubber and plastic products, as well as manufacture of electrical equipment and consumer electronics. However, PC in the latter economic activity grew mainly due to the new assembling lines used to produce home appliances.18

Group B. Leaders — PC growth by 1.5–1.7 times. The group includes three sectors: coal mining and processing, manufacture of food products, and manufacture of motor vehicles and other transport equipment.19

Group C. Followers — PC growth by no more than one-third. The group includes almost all raw materials production, as well as non-fuel mineral extraction. The limited PC growth in this segment was caused by reduced incentives: as shown by the analysis, the age of PC in these sectors is not crucial factor to use them (as a rule, the old PC utilization ratio differs nonsignificantly from the new PC utilization ratio in these sectors).

Group D. Outsiders — PC reduction. The group includes mostly manufacture of machinery and equipment (without the military-industrial complex), as well as manufacture of textiles and wearing apparel.

This paper estimates the average age of PC. It is differentiated considerably by sectors and reflects the intensity of PC renovation. On the whole, there is a logical pattern: older PC (older than 15 years20) are concentrated in raw materials production, while newer PC (younger than 10 years) — in end-use production (Fig. 6). Significant progress should be noted in the PC renovation — in manufacture of food products, manufacture of plastic products, and wood processing. On the other hand, PC renovation is not intensive in the majority of mechanical engineering sectors (in manufacture of electrical equipment and consumer electronics, new PC are largely assembly lines). In manufacture of textile and wearing apparel a significant portion of the new PC was commissioned for a narrow group of products in conditions of output stagnation and absence of de-commissioning of old PC.

Fig. 6

Average PC age by sector in 2015 (years).

We have estimated the share of noncompetitive PC (we define such PC as PC older than 10 years and not utilized for at least the past five years).21 Note that it can be reserved PC in a number of industries.22 Our estimate shows that the weighted average share of noncompetitive PC in the manufacturing industry (excluding oil refining) is 13–14% (Fig. 7). It is especially high in the investment segment: from 17% in manufacture of motor vehicles and other transport equipment to 26% in manufacture of machinery and equipment. A large share of these capacities is allocated in the dynamically developing chemical production (a number of old PC have not been de-commissioned and are not utilized across a wide range of both raw materials and the final products). In the other industries, the share of noncompetitive capacities is low, no more than 10–11% (in the manufacture of food products, estimation of their share is 14%; however, those are mainly reserved PC).

Fig. 7

Share of noncompetitive and old PC in 2015 by sector (%).

The capacity utilization ratio as of 2015 is highly differentiated by sectors (Fig. 8). On the one hand, we identified a group of products with capacity utilization ratios exceeding 80%, for which the prospects for non-capital-intensive growth are very limited (most raw materials products). On the other hand, sectors dependent on investment demand (almost all mechanical engineering productions) and focused on durable products (which is natural in a crisis) demonstrated extremely low performance on this ratio. If noncompetitive capacities are considered in the calculation, the PC utilization ratio estimate will increase for the majority of sectors: the excess competitive PC in raw materials production are very small,23 but there is no deficit of PC in most domestically focused sectors.

Fig. 8

Capacity utilization ratio in 2015 by sector (%).

To estimate the capacity utilization ratio growth potential, we conducted a sectoral analysis relying on information about the top decile of the capacity utilization ratio, the “new commissioning ratio,” capacity limitations at the detailed level, etc. We found that for the majority of sectors the level of PC utilization in recent years has been at least 5 to 10 p.p. below the maximum capacity utilization ratio (Table 3). The literature (Oomes and Dynnikova, 2006; Mironov and Kanofiev, 2014) demonstrates that before the 2008–2009 crisis, Russia's industry operated close to a capacity utilization ratio that accelerated inflation. As our study shows, achieving a sufficiently high capacity utilization ratio at the same time stimulates investment activity. We found that an increase in the capacity utilization ratio by 5% results in 8% to 10% more commissioning if capacity utilization ratio is below 90%, whereas this figure jumps by 1.5 times if capacity utilization ratio is above this threshold (Fig. 9).24

Table 3Download as CSV 

Capacity utilization ratio growth potential: estimates for sectors.

Sector	Growth potential against 2015, p.p.	Capacity utilization ratio in 2015, %	Optimal ratio, %	For reference: capacity utilization ratio in 2013, %
Manufacture of food products	5–10	≈ 60	65–70	≈ 60
Manufacture of textiles and wearing apparel	≥ 10	≈ 67	75–85	70
Manufacture of leather and related products	< 10	51	75–80	≈ 60
Manufacture of wood and of products of wood and cork	≥ 10	≈ 60	70–75	≈ 60
Manufacture of paper and paper products	< 5	82	85–90	74
Manufacture of chemicals and chemical products	5–10	74	80–85	72
Manufacture of rubber and plastic products	15–20	56	75–80	56
Manufacture of other non-metallic mineral products	≈ 15	57	70–75	64
Manufacture of ferrous metals and fabricated metal products	≈ 10	78	85–90	78
Manufacture of machinery and equipment	25–30	35	65–70	39
Manufacture of electrical equipment and consumer electronics	≈ 20	34	65–80	47
Manufacture of motor vehicles and other transport equipment	> 30	38	75–85	60

Source: Compiled by the authors.

Fig. 9

Distribution of new commissioning based on the capacity utilization ratio (in previous year) in the manufacturing industry (aggregated results for 15 years, 200 products and 70 regions) (%).

Our results clearly indicate that there is some excess capacity in Russian economy: currently, potential of output growth in the manufacturing industry by improving capacity utilization is estimated to be at least 6–8%. This fact should be recognized by policymakers.

We do not question the findings of other authors that the industry operated close to a capacity utilization ratio that accelerated inflation before the 2008–2009 crisis. However, we have shown that a high capacity utilization ratio also accelerates the intensity of investment activity. Hence, a careful policymaker should seek a balance between accelerating inflation and accelerating capacity renovation.

Statistical form 1-natura-BM is an important and unique source of data on production capacities. The economists would benefit much from the open access to this data (that should be ensured by publishing full data at the regional level).41 The quality of primary statistics should be improved by expanding the number of products represented in the form (to ensure that the data on capacities for individual sectors are representative). Finally, the methodology used by Rosstat should be corrected in order to eliminate the detected shift in capacity utilization estimates.

The further research may address a number of issues. First, it is interesting to study further the correlation between accelerating inflation and accelerating capacity renovation (the current level of capacity utilization ratio may influence both inflation and future investment intensity). Second, it's viable to obtain estimates of the costs of capacity commissioning by sectors and products. Finally, it is important to figure out are the data is representative for conducting a regional analysis.