Selected Issues

Abstract

Selected Issues

Infrastructure Investment and Firm Performance: Evidence from Kazakhstan’s “Nurly Zhol” Program1

Infrastructure investment can facilitate export diversification and boost economic growth, but the overall impact of large and costly infrastructure projects is context specific. Given the significant resources involved, it is important to analyze the effects of infrastructure investments on economic outcomes. This note examines differences in firm performance arising from exposure to roads and railroads built under Kazakhstan’s “Nurly Zhol” program. Results suggest that increased exposure to transport projects is associated with a rise in firm revenues and profits; the effect on employment is not significant. The results can be used to inform future decisions on investment projects.

A. Introduction

1. Investment in infrastructure can be an important lever for promoting economic diversification and inclusivity and raising growth potential. In the short term, infrastructure investment can boost domestic demand and in the long run, it may enhance the economy’s productive capacity and potential growth. Transport infrastructure plays a special role, as the density and quality of the road and railroad networks are directly linked to the cost of trading (travel time, fuel use, vehicle maintenance). In addition to reducing costs and facilitating access to markets, increased connectivity may improve labor mobility and support regional development and social inclusion.

2. However, the overall economic impact of infrastructure projects is context specific. Research suggests that many factors influence the impact of infrastructure investment initiatives. These include the type of infrastructure built—railroads (Atack et al., 2009, Donaldson 2018), electrical grids (Dinkleman, 2011), or mobile phone towers (Jensen, 2007)—as well as factor mobility (Banerjee at al., 2012) and the timeframe, with immediate benefits if the infrastructure removes bottlenecks (Jensen, 2007) and longer-term impact from urbanization (Jedwab and Moradi, 2016). However, there may also be political economy issues. Warner (2014) finds a weak association between investment spending and growth and attributes it to debt financing, poor project selection, and incentive problems. In some cases, large infrastructure projects have contributed to the rapid accumulation of domestic and external debt, raising fiscal sustainability concerns.

3. Efficiency is key to maximizing benefits from investment. Investment inefficiencies are prevalent in developing countries. The gap between the unadjusted and efficiency-adjusted public capital stock can be quite large, implying significant potential gains by closing it (Gupta et al., 2014; Crivelli, 2017). The issue seems more acute at times of investment booms when absorptive capacity constraints, manifested in declining marginal returns to investment, become an important factor. Studies suggest that in periods of sharp scaling up of public investment, projects undertaken are less likely to be successful (Presbitero, 2016).

4. Kazakhstan has large infrastructure needs. The country’s vast territory and relatively small population present major challenges to building and maintaining an adequate transport network, but the proximity of two large markets—China and Russia—creates opportunities. Infrastructure development features prominently in the authorities’ main strategic plan, “Kazakhstan 2050”, which identifies two main goals: (i) integration of the national economy into the global economy; and (ii) increased connectivity among regions within the country. These high-level goals were operationalized in early 2013 in a detailed State Program for Transport Infrastructure Development 2020 (SPTID-2020), spanning 2014–20. The program set specific (and ambitious) targets for increases in cargo and passenger transportation (81 percent and 85 percent over 8 years) and outlined measures to achieve them, including scaling up of public investment and increased participation of the private sector.

B. Overview of “Nurly Zhol”

5. In 2015, the government of Kazakhstan launched a large fiscal stimulus package, “Nurly Zhol,” as a countercyclical measure. Similar to other countries in the region, Kazakhstan was hit by adverse external shocks in 2014. Growth decelerated noticeably following a steep decline in oil prices and slowdown of external demand, and imbalances emerged. Since shocks were perceived to be long lasting, the authorities responded by launching a sizeable stimulus program. The program combined anti-crisis and structural measures aimed at supporting specific sectors of the economy—providing affordable housing, modernizing infrastructure, promoting entrepreneurship, and increasing the competitiveness of domestic firms. Financing was secured mainly from the National Fund of the Republic of Kazakhstan (NFRK), with an allocation of $9 billion for 2015–17. Other sources of funding included central government and local budgets, borrowing from international financial institutions, and funds from state companies.

6. A major component of “Nurly Zhol” was upgrade of the transport infrastructure. Analysis revealed significant constraints to transportation among regions, creating bottlenecks for cargo traffic and limiting labor mobility. It was recognized that a forward-looking approach to the provision of necessary infrastructure would also need to take into account important trends, in particular, Kazakhstan’s growing population and urbanization. Thus, the infrastructure component of “Nurly Zhol” was centered around the idea of macro regions, with large cities serving as hubs (Almaty, Nur-Sultan, Aktobe, Shymkent, and Ust-Kamenogorsk). Increasingly, resources (capital, human) and economic activity would be concentrated in these cities, and accordingly, they would receive more of the infrastructure investment. Implementation of “Nurly Zhol” would seek synergies with other state programs (e.g., SPTID-2020) and China’s Belt-Road Initiative.

Figure 1.
Figure 1.

Kazakhstan: Map of “Nurly Zhol” Road Projects

Citation: IMF Staff Country Reports 2020, 038; 10.5089/9781513529288.002.A002

Source: KazAutoZhol.

7. Selection of “Nurly Zhol” projects was based on the “ray” principle. Kazakhstan’s road and railroad networks were largely designed and built during the Soviet period and aimed at connecting the North with the South of the country, and as a result, transport links between other regions remained underdeveloped. Consistent with the concept of developing hub cities, transport infrastructure projects under “Nurly Zhol” were selected on a “ray” principle, whereby roads between hubs and roads connecting hubs with other large cities received priority. Projects involved major upgrades and expansion of existing roads or construction of new ones where warranted.2 As a result of the program implementation, the average travel time between hub cities was targeted to decrease by over one-third.

C. Empirical Strategy and Data

8. In view of the sizable resources dedicated to the implementation of “Nurly Zhol,” it is important to evaluate its impact on economic outcomes. Specifically, it is of interest to explore whether the construction of new roads and railroads has had a significant effect on output, employment, and firm profits. The timing and distribution of transport infrastructure projects create significant differences in the exposure of cities to the program. This makes it possible to exploit spatial variation in access to new infrastructure in a difference-in-difference design. The expectation is that during the sample period, the performance of firms situated in cities closer to completed infrastructure projects would differ from that of firms operating in more remote or less exposed locations.

9. Analysis is based on micro-level data. A balanced panel of 1,379 firms (anonymized survey data sourced from the National Bank of Kazakhstan) containing location and quarterly data on operating revenue, cost of sales, assets, and employment was used to evaluate performance over the sample period Q1:2014–Q1:2019. The firms in the sample comprise about 5 percent of total employment and value added in Kazakhstan and provide a sufficiently-broad geographical coverage. Firms operating in the oil sector were excluded, given their size and the industry specifics. Sources of (annual) data on road and railroad projects commissioned under “Nurly Zhol” through 2018 are the Kazakhstani road agency KazAutoZhol and the Ministry of National Economy, respectively. In total, 1,769 km of roads and 1,376 km of railways were included in the analysis.

10. Road and railroad information was used to construct an exposure variable. As a first step, Google Earth was used to geo-code points 10 kilometers apart along a newly renovated or constructed road and rail segment. Then, for each city where a firm was located, the distance to the geo-coded point was calculated (see text figure). A city c is considered to be exposed to the project if it is within d kilometers of any point along the segment, where three different values are tried for d— 50 km in the baseline and 25 km and 100 km in alternative specifications. In general, a location may be exposed to more than one road/railroad project p, so the cumulative exposure variable across all years y of the “Nurly Zhol” program is defined as:

RoadExpdc,y=Σp1{CitycwithindkmofProjectpcompletedbyyeary}(1)

The rail exposure variable (RailExpdc,y) is defined in a similar way. These exposure variables vary by year and location. The main explanatory variable of interest, transport exposure (TransExpdc,y), is obtained as the sum of the road and rail exposures. An alternative approach is also considered, where a binary (non-time dependent) variable is introduced, which takes the value of one if a location has ever been exposed to any project during the sample period and zero otherwise:

TransEverExpdc=1{TransExpdc,20182}3(2)
3

Table 1 illustrates these definitions using the example of the city of Uralsk, and Figure 2 shows the evolution of exposure across regions and over time. There is significant variation in the cumulative transportation exposure, with Pavlodar and Nur-Sultan being exposed to 8 and 4 projects by 2018, respectively, while no major infrastructure upgrades have taken place near Shymkent or Taraz (Figure 3). 4

Table 1.

Kazakhstan: Uralsk—Transportation Exposure Variables

article image
Source: IMF staff.The table presents the evolution, over time, of the main transportation variables of interest, for a distance d = 50km.
Figure 2.
Figure 2.

Kazakhstan: Transportation Exposure: Evolution over Time

Citation: IMF Staff Country Reports 2020, 038; 10.5089/9781513529288.002.A002

Source: IMF staff calculations.The figure plots whether a location was exposed to at least one transportation project by a given year. Each point represents a city that was exposed to at least one newly constructed transportation project by year y, where exposure is an indicator variable that equals one if a city was located within 50 km of a transportation project.
Figure 3.
Figure 3.

Kazakhstan: Cumulative Transportation Exposure: Evolution over Time for Selected Cities (50 km Definition)

Citation: IMF Staff Country Reports 2020, 038; 10.5089/9781513529288.002.A002

Source: IMF staff calculations.

D. Discussion of Results

11. Various empirical specifications were estimated. The baseline empirical specification involves a regression of the outcome variable of interest (output, gross operating profit or employment) on the (one period) lag of the cumulative exposure variable with d=50 km, a firm fixed effect (αf) and a time fixed effect (μq,y)5.

Yf,c,q,y=βTransExpdc,y1+αf+μq,y+ϵf,c,q,y(3)

12. Since the information on roads and railroads commissioned is available only at the annual frequency, the lag of the exposure variable ensures that the project is completed before the results are measured. An alternative specification involving the binary exposure variable defined in (2) was estimated as well. This specification is completely non-parametric in that it compares the relative performance of firms that were ever exposed to a transportation project to firms that did not receive one, without considering the total number of projects that “treated” firms were exposed to. Therefore, this approach addresses concerns that results may be overly influenced by firms located in certain cities where a disproportionate number of roads or railways were built.

Yf,c,q,y=Σq,yβq,yTransExpc×1{Period=q,y}+αf+μq,y+ϵf,c,q,y(4)

13. The identifying assumption behind these empirical specifications is that in the absence of the infrastructure projects, firms in both exposed and unexposed cities would have followed the same trend. This assumption could be violated if, for instance, the decision to launch an infrastructure project in a particular location was endogenously linked to that location’s higher growth potential. Two considerations mitigate potential endogeneity concerns. First, most of the road projects under Nurly Zhol involved renovating existing, Soviet-era roads that were built prior to Kazakhstan’s independence.6 Second, the analysis is not limited to the main cities that were connected by a new transportation project, but rather includes all smaller cities that happen to be along the project’s route. It is unlikely that road building decisions factored in the economic potential of these en route cities.

14. Results suggest significant impact on output and profits and no effect on employment. The baseline regression reveals a significant relationship between a firm’s proximity to a major transport infrastructure project being completed and its sales revenue and gross profits (Table 2). As reported in column (1), being exposed to an additional transportation project is associated with a KZT 63 million increase in revenues and a KZT 22 million increase in gross profits, which represents a 4 percent and a 10 percent increase relative to the mean, respectively. Employment, on the other hand, does not seem to react to the construction or upgrade of roads and railways. This could be a timing issue but also could signal structural problems with labor mobility. Results from the alternative specification (4) suggest that the impact of transport infrastructure exposure tends to increase over time (Figure 4).

Table 2.

Kazakhstan: Transportation Exposure (Lagged) and Firm Performance

article image
Source: IMF staff estimates.Note: The table presents results from estimating equation (4). The sample consists of a balanced panel of firms from Q1 2014-Q1 2019. Revenues and gross profit are in KZT million. All dependent variables are winsorized at the 1 percent level. Robust standard errors (reported in parentheses) are clustered at the city level. **, **, and * indicate statistical significance at the 1 percent, 5 percent and 10 percent levels, respectively.
Figure 4.
Figure 4.

Kazakhstan: Estimated Impact of Transport Infrastructure Exposure on Revenue

Citation: IMF Staff Country Reports 2020, 038; 10.5089/9781513529288.002.A002

Source: IMF staff calculations.

15. The results are robust to alternative distance bandwidths. Using the same specifications as above but with different criteria for proximity (d=25 km and d=100 km) provides a useful test of the sensitivity of results under the baseline. In fact, the effects on revenue or profits are stronger when a shorter cut-off value for exposure is applied. For example, the point estimate of the coefficient in the revenue regression increases from 63.0 in the baseline to 108.2 when the 25 km benchmark is used, and it is significant at the 1 percent level. Moving to d=100 km, the effect disappears, suggesting that firms that are relatively far from new infrastructure do not react much in terms of output.

E. Conclusion

16. Overall, the transport infrastructure component of the “Nurly Zhol” program appears to have yielded positive short-term results. Firms have benefitted from improved connectivity and reduced transportation costs, which are manifested in revenue, and especially, in profit increases. As a rule, the closer a firm is to major new roads and railways, the larger the effect. Also, firms operating in locations where a larger number of projects has been completed seem to gain more. While facilitating labor migration has been one of the goals of the program, there is not enough evidence to support the conclusion that firms in exposed locations hire more workers. This could be due to the fact that most of the projects have been completed only recently, and more time is needed for the effects on employment to materialize. It might, however, be an indication of constraints on the labor supply side, reflecting factors other than cost of travel, e.g., traditions, underdeveloped housing markets, bureaucratic restrictions or obstacles, or skills shortages. More targeted policies would be needed in this case to promote the movement of people.

17. Increasing the efficiency of infrastructure investment would help to achieve better results within a given resource envelope. IMF analysis suggests that countries lose on average about 30 percent of the returns on their investment due to investment inefficiencies (IMF, 2015). In this regard, strengthening public investment management at all stages of the investment process is key. Project proposals should be subject to rigorous appraisal and decisions should be made based on sound economic and financial analysis and assessment of risks. This is particularly true when choosing among competing projects. In addition, putting in place effective and transparent procurement systems would reduce costs and improve the quality of implementation. Project management, oversight and ex post evaluations are also important to ensure that projects are delivered on time and on budget and to provide insights for future investment decisions. The IMF’s Public Investment Management Assessment (PIMA) framework provides a useful tool for assessment of the processes and institutions related to the provision of infrastructure assets.

18. Increasing private sector involvement in infrastructure could generate benefits but also entails risks. Both “Nurly Zhol” and SPTID-2020 envisage private sector participation in infrastructure provision and maintenance, mainly through public-private partnerships (PPPs). Many countries have resorted to this vehicle, including as a way to address tight fiscal constraints. If designed and implemented well, PPPs can offer advantages in terms of efficiency in use of resources, technology, and quality of service, in addition to budget savings and risk-sharing. However, weak PPP designs can unduly expose public finances to risks and result in substantial costs if, for example, contracts were based on overly optimistic assumptions about service usage or guarantees were provided by the government. PPPs are becoming increasingly utilized in Kazakhstan, with KZT 570 billion contracted from private investors and government obligations close to KZT 178 billion.7 This calls for a careful assessment and management of risks. PFRAM—an analytical tool, developed by the IMF and the World Bank—could be useful in evaluating fiscal costs and risks arising from PPPs.

References

  • Atack, J., Bateman, F., Haines, M. and R.A. Margo (2010), “Did Railroads Induce or Follow Economic Growth? Urbanization and Population Growth in the American Midwest, 1850–1860”, Social Science History, Vol. 34, No.2, pp. 171197.

    • Search Google Scholar
    • Export Citation
  • Banerjee, A., Duflo, E. and N. Qian (2012), “On the Road: Access to Transportation Infrastructure and Economic Growth in China”, NBER Working Paper No. 17897.

    • Search Google Scholar
    • Export Citation
  • Crivelli, E. (2017), “Efficiency-Adjusted Public Capital, Capital Grants, and Growth”, IMF Working Paper WP/17/168.

  • Dinkelman, T. (2011), “The Effects of Rural Electrification on Employment: New Evidence from South Africa”, American Economic Review, 101, pp. 30783108.

    • Search Google Scholar
    • Export Citation
  • Donaldson, D. (2018), “Railroads of the Raj: Estimating the Impact of Transportation Infrastructure”, American Economic Review, 108 (4–5), pp. 899934.

    • Search Google Scholar
    • Export Citation
  • Gupta, S., Kangur, A., Papageorgiou. C., and A. Wane (2014), “Efficiency-Adjusted Public Capital and Growth”. World Development, Vol. 57, pp. 164178.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund (2015), “Making Public Investment More Efficient”, IMF Policy Paper, Washington D.C.

  • Jedwab, R. and A. Moradi (2016), “The Permanent Effect of Transportation Revolutions in Poor Countries: Evidence from Africa”, The Review of Economics and Statistics, 98(2), pp. 268284.

    • Search Google Scholar
    • Export Citation
  • Jensen, R. (2007), “The Digital Provide: Information (Technology), Market Performance, and Welfare in the South Indian Fisheries Sector”, The Quarterly Journal of Economics, Vol. 122, No. 3, pp. 879924.

    • Search Google Scholar
    • Export Citation
  • Presbitero, A. (2016), “Too Much and Too Fast? Public Investment Scaling-up and Absorptive Capacity”, Journal of Development Economics, No. 120, pp. 1731.

    • Search Google Scholar
    • Export Citation
  • State Program for Infrastructure DevelopmentNurly Zhol” for 2015-2019, (in Russian), http://economy.gov.kz/ru/pages/gosudarstvennaya-programma-infrastrukturnogo-razvitiya-nurly-zhol-na-2015-2019-gody.

    • Search Google Scholar
    • Export Citation
  • State Program for Transport Infrastructure Development of the Republic of Kazakhstan till 2020, (in Russian), http://aviation.miid.gov.kz/ru/pages/gosudarstvennaya-programma-razvitiya-i-integracii-infrastruktury-transportnoy-sistemy-0.

    • Search Google Scholar
    • Export Citation
  • Warner, A. (2014), “Public Investment as an Engine of Growth”, IMF Working Paper WP/14/148.

1

Prepared by Faizaan Kisat and Rossen Rozenov.

2

For example, construction of a new direct road from Nur-Sultan to Aktobe was estimated to reduce travel time by 7 hours compared to an existing road passing through Kostanay.

3

The definition is based on the median transportation exposure in 2018 which is 2.

4

In fact, at the time of the launch of “Nurly Zhol”, the reconstruction of the road connecting Shymkent to Aktobe had already been completed as part of the project Western Europe-Western China, and work was ongoing on the road from Shymkent to Almaty.

5

q indexes quarter and y indexes year.

6

Nurly Zhol’s rail projects mainly consist of a more than 1,000 km railway line built through the sparsely populated center of the country. Rail exposure is therefore zero for most of the firms in the sample, and therefore its (potential) endogenous placement is not likely to bias the results significantly.

7

As of early September 2019, according to information of the Kazakhstani Public-Private Partnership Center (https://kzppp.kz/).

Republic of Kazakhstan: Selected Issues
Author: International Monetary Fund. Middle East and Central Asia Dept.