Estimating Future Renewal Costs for Road Infrastructure and Financial Burden in Japanese Prefectures

Transcription

1 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March Estimating Future Renewal Costs for Road Infrastructure and Financial Burden in Japanese Prefectures Nobuo Akai Professor, Osaka School of International Public Policy, Osaka University Toru Takemoto Associate Professor, Faculty of Economics, Tezukayama University Abstract While the fiscal conditions of local governments will become increasingly severe along with the population decrease, the renewal costs they will incur for social infrastructure that has reached the end of its service life are expected to expand. In order to analyze the impact of the future increase in such costs, this paper estimates the road stock managed by each prefecture from super-long-term investment data from road management authorities, and estimates the necessary road renewal costs up until FY2040. The estimates show that if new investment is not reduced, renewal costs in FY2040 will be 2.6 times larger than current costs, exceeding the total fiscal expenditure in FY2012. If future fiscal expenditure is held at the current level, about half of all prefectures will be unable to renew 100% of the roads under their management, while the remaining half will be able to make new investment in roads. Extending the service life of infrastructure will be the effective solution for closing such a gap. However, these analysis results do not reflect the aspect of population decrease. Estimates made with consideration to this aspect show that the extension of the service life of infrastructure only has limited effectiveness. Thus, it is necessary to curtail the total amount of infrastructure through integration while carrying out reforms to the city structure, as typified by the concept of compact cities. Keywords: Road infrastructure, renewal, population decrease JEL Classification: H54, H72, R4, R5 I. Introduction While the fiscal conditions of local governments will become increasingly severe due to population decrease and an aging society, renewal costs for social infrastructure that has reached the end of its service life are expected to expand. In order to analyze the impact of the future increase in such costs, it is necessary to accurately estimate its costs for infrastructure. Especially roads, focused on in this paper, are one of the most important stock items for economic development and life. However the word renewal has not been defined even in the law of the Road Act in Japan and its cost has not been presented in road statistics clearly. In addition, the roads are managed by central and local (prefecture and municipality) governments and the resources for the management by the local governments are financed

2 96 N Akai, T Takemoto / Public Policy Review by not only their own budget but also through transfers from the central government. Therefore, due to the complex systems, the future renewal cost financed by each local government is very unclear. In developing countries facing the needs of the expansion of infrastructure, future renewal costs are be taken into consideration. However, urban and city planning (for example, compact city and city spread prevention) with infrastructure design and the optimal maintenance for minimizing the total life cost are important at this moment. Ignorance of renewal costs at a developing stage adds unnecessary costs in the future when the economy has matured and population declines. The Japanese experience should be regarded as a bad example. This paper estimates the future road stock and the necessary renewal costs for the roads that have been managed by each prefecture. The necessary renewal costs defined by the costs for keeping the road stock level are estimated by depreciation loss. As described above, data for renewal investments does not exist, so investment data must be collected retrospectively. Some related literature does exist, for example the following: The Director General for Economic and Fiscal Management, Cabinet Office (2001), Nakahigashi (2012), Ministry of Land, Infrastructure and Transport (2010), Kanba (2012), Nishimura and Miyazaki (2012), Nagano and Minami (2003), Local Public Finance Bureau, Ministry of Internal Affairs and Communications (2012), etc. However this literature estimates costs focusing on the region and not on management bodies, such as the prefectural government. Therefore, it is difficult to link due to the financial complexity of each government body. The problem the government faces is one of financial constraint due to population decrease and an aging society. This problem cannot be ignored when considering the future management of roads and the future renewal costs. In this paper, the tightness of the local budget is incorporated by considering the decrease in population. However the challenge for estimating the cost to the management body (prefecture, in this paper) faces a big wall in that each set of investment data is not summarized in road statistics. The Director General for Economic and Fiscal Management, Cabinet Office (2012) adopts a method that divides the cost of the total investment in each region into each government body proportionally to the share of the road area and distance. In this case, it has been assumed that the investment cost per area or distance is constant for all types of roads in each region. However this assumption cannot be accepted because the investment costs for national highways are different from that of local areas. This paper collects data on the investment for each government body from the Annual Report of Road Statistics, focusing on regionally-managed national highways and prefectural roads, managed by prefectural governments, and estimates road stock using the Perpetual Inventory Method (PI method). 1 This is the first contribution of this paper. This new data enables us to estimate accurate necessary renewal costs that should be financed in the future in 1 The tool road is not considered in this paper because this type of road is managed based on self-supporting accounting.

3 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March the prefecture. The second contribution of this paper is to estimate the necessary renewal cost until FY2040. This paper estimates the future of the renewal investments under multiple scenarios until FY2040, for example: keeping the current level of new investment, reducing the investment, and extending the service life of infrastructure. In addition, we estimate the extended case in which population decrease in each prefecture limits local budget. The estimates show that if new investment is not reduced, the renewal investments in FY2040 will be 2.6 times larger than the current costs, exceeding total fiscal expenditure in FY2012. If future fiscal expenditure is held at the current level, about half of all prefectures will be unable to renew 100% of the roads under their management, while the remaining half will be able to make new investment in roads. Extending the service life of infrastructure will be the effective solution to close such a gap. However, these analysis results do not reflect the aspect of population decrease. Estimates made while taking this aspect into consideration show that extension of the service life of infrastructure alone has limited effectiveness, so it is necessary to curtail the total amount of infrastructure through integration while carrying out reforms to city structure, as typified by the concept of the compact city. This paper is organized as follows. Section 2 considers how road stock is accumulated by road investment and Section 3 considers how road investment and maintenance are responsibly and financially supported. In Section 4, the method of estimation is explained. In Section 5, road stock is estimated. In Section 6, the renewal investment is estimated along with the impact of the change of the investment plan. In Section 7, considering the tightness of the local budget, an investigation is made into how the population decrease effect on renewal investments affects the budget for investment and renewal. Section 8 investigates how the extension of the service life of infrastructure affects renewal investments. Finally, the result of this paper is summarized in Section 9. II. The relationship between expenses for road projects and capital stock of road infrastructure II-1. Classification of the expenses for road projects The administrations under the road act to be performed by road administrators are construction of roads, reconstruction of roads, maintenance of roads, repair of roads, disaster recovery projects for roads, and the other administration. (The left side figure in Figure 1) In the Council on Studying the Maintenance and Management of National Highways in Ministry of Land, Infrastructure and Transport, terms about the maintenance and management of the road are defined as follows. First, all administrations under the road act to be performed by road administrators (construction of roads, reconstruction of roads, maintenance of roads, repair of roads, disaster recovery projects for roads, and other administration) are defined as administration. Next, the maintenance and repair of administration are defined as follows.

4 98 N Akai, T Takemoto / Public Policy Review Figure 1 Administrations to be performed by road administrators Maintenance: daily administration for the purpose of the maintenance of the function and the structure of the road (patrol, cleaning, weeding, pruning, snow-removal, patching of the pavement, etc.). Repair: administration to restore the damaged structure of roads to their original state and administration for the purpose of adding necessary functions and increasing the structural strength (repair of deteriorated and damaged parts of bridges, tunnels, and pavement; aseismic reinforcing work; slope reinforcement; snow protection measures: etc.). Generally the term renewal is used frequently, but is a term that is not defined by the road act. Therefore we define renewal as follows. Renewal: administration to replace aging facilities reduced in function and to restore functions to the same level. It falls under either reconstruction or repair depending on the specific contents of the construction. However all reconstruction is not included in renewal. Reconstruction may be performed to cope with an increase in traffic even if the facility is not aging. In this study, that kind of reconstruction and new construction are together called new investment. In other words, fiscal expenditure except for maintenance is defined as investment and investment except renewal and disaster recovery projects is defined as new investment. Therefore, in this study, the administration is classified into new investment, renewal, maintenance, and disaster recovery projects. (The right side figure in Figure 1) After section V, we analyze it according to this classification. II-2. Factors in increasing the capital stock of road infrastructure The factors that increase the capital stock of the road infrastructure are new construction, reconstruction, repairs and disaster recovery projects. Maintenance is necessary to maintain

5 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March the capital stock of road infrastructure but does not increase the capital stock of the road infrastructure. Figure 2 shows the relation between these terms in the Director General for Economic and Fiscal Management, Cabinet Office (2012) and this study and data in the Annual Report of Road Statistics. In the Director General for Economic and Fiscal Management, Cabinet Office (2001), new construction and reconstruction, maintenance, repairs and disaster recovery projects are related to the data classification in the Annual Report of Road Statistics as follows. New construction and reconstruction => Maintenance => the value, subtracting the amount of compensation for land to be expropriated from the expenses, for construction in the Annual Report of Road Statistics the expenses for local independent projects of the expenses for maintenance in the Annual Report of Road Statistics Figure 2 Relationship between terms and data in the Annual Report of Road Statistics

6 100 N Akai, T Takemoto / Public Policy Review Repairs => the expenses for projects directly carried out by the central government and subsidy projects on the expenses of maintenance in the Annual Report of Road Statistics Disaster recovery projects => the expenses for disaster recovery projects in the Annual Report of Road Statistics In this study, these are revised in the following way. New construction and reconstruction => Maintenance => Repairs => the value, subtracting the amount of compensation for land to be expropriated from the expenses, for construction in the Annual Report of Road Statistics maintenance in the details of construction type, which are the expenses for maintenance in the Annual Report of Road Statistics other than the above, which are the expenses for maintenance in the Annual Report of Road Statistics Disaster recovery projects => the value, subtracting the amount of compensation for land to be expropriated, from the expenses for disaster recovery projects in the Annual Report of Road Statistics II-3. Factors in decreasing capital stock of road infrastructure The factors in decreasing capital stock of road infrastructure are aging and disaster. The Director General for Economic and Fiscal Management, Cabinet Office (2001) assumed a bell-shaped distribution, that is, the capital stock of road infrastructure due to aging is decreasing according to survival functions such as in Figure 3. This study also adopts the same survival functions. In the Director General for Economic and Fiscal Management, Cabinet Office (2001), the service life of road bodies, bridges and pavement are set to 10 years, 52.5 years and 60 years, respectively, taking into consideration the Ministerial Ordinance concerning the Useful Life, etc. of Depreciable Assets. (Finance Ministry Ordinance No.15 of 1965) However, the Director General for Economic and Fiscal Management, Cabinet Office (2001) used 50 years as the only average service life, which is the weighted average of these three service lives from each investment, as it did not estimate the capital stock of road bodies, bridges and pavement separately. In contrast, we used 10 years, 52.5 years and 60 years as the average service life of road bodies, bridges and pavement, respectively and estimated the capital stock of road bodies, bridges and pavement separately.

7 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March Figure 3 The survival function (Weibull distribution) Source: The Director General for Economic and Fiscal Management, Cabinet Office (2001) III. Road administrators and their burdens III-1. Road administrators and their burdens All roads under the law act are classified into state-managed national highways, regionally-managed national highways, prefectural roads and municipal roads. 2 Table 1 shows length of each class, road administrator, and its burden sharing. National highways are divided into state-managed national highways and regionally-managed national highways. The former s road administrator is Minister of Land, Infrastructure and Transport. The latter s road administrator is prefecture or ordinance-designated city. The road administrators of national highways are different. In addition, the road administrators of prefectural roads located in ordinance-designated cities are ordinance-designated cities. Otherwise, the road administrators of prefectural roads are prefectures. IV. Estimation method IV-1. The method for estimating capital stock of road infrastructure This section describes the method for estimating capital stock of road infrastructure 2 Roads other than roads under the road act are ordinary motorways under the road transportation act (33 routes by 28 companies including Izu Skyline, 323.3km in total), forest roads under the forestry partnership act, park ways under the national park act and fishing port roads under the fishing port act (See Nagano and Minami; 2003).

8 102 N Akai, T Takemoto / Public Policy Review Table 1 Road administrators and their burdens Source: Annual Report of Road Statistics 2014 managed by prefectures. Like the Director General for Economic and Fiscal Management, Cabinet Office (2001), we use the PI method. The investment is largely divided into the sum of new construction, reconstruction and repair and expenses for disaster recovery projects. xk s is the investment of the former in the road managed by prefecture k in the financial year of s. yk s is the real investment of the latter. These are the real value with base year 2005 deflated by the deflator. (See Appendix 1) There are two types of retirements. One is retirement due to aging and the other is retirement due to disaster. It is assumed that the amount of retirement due to disaster in each year

9 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March is equal to the expenses for disaster recovery projects in the same year. To estimate the amount of retirement due to aging, we assume the mortality function f h (h is an age) of the Weibull distribution is: m h f h = ( ) m-1 ( h m m=4, ƞ: satisfies that the cumulative mortality rate ) η η e- η in the average service life is Table 2 is an example of the retirement due to aging. In regard to the capital stock invested in the financial year of t, the amount of retirement due to aging in the financial year of t+1 is ( xk t + yk t ) f 1, and the amount of retirement due to aging in the financial year of t +2 is ( xk t + yk t ) f 2. In the financial year of t+2, the amount of retirement due to aging of the capital stock invested in financial year t+1, ( x t 1 k + + y t 1 k + ) f 1, is added. Therefore, the amount of retirement due to aging in the financial year of t+2 is d t+2 t+1 k = s=t (x s k +y s k ) f t+2-s. Consequently, the amounts of retirement in the financial year of t due to aging of the capital stocks invested in each financial year until the financial year of t-1 are calculated. The sum of these is the amount of retirement due to aging in the financial year t. However, Table 2 An example of the amount of retirement due to aging without considering retirement due to disaster 3 In the case that the average service life is 60 years, the scale parameter ƞ is ; in the case that the average service life is 52.5 years, the scale parameter ƞ is ; in the case that the average service life is 10 years, the scale parameter ƞ is ; in case that the average service life is 70 years, the scale parameter ƞ is ; and in case that the average service life is 80 years, the scale parameter ƞ is

10 104 N Akai, T Takemoto / Public Policy Review in the above-mentioned method, we don t take into account decreases in the capital stock due to disaster. Considering retirement due to disaster, the capital stock becomes zero faster than the case of only the retirement due to aging. Even though the capital stock has become zero, the positive amount of retirement due to aging is sometimes calculated in the above-mentioned method. Therefore, the amount of retirement due to aging is regarded as zero after the capital stock becomes zero when taking account of both retirement due to aging and retirement due to disaster. The amount d t k of retirement in the financial year of t due to aging of the capital stock of road infrastructure managed by prefecture k can be expressed as follows. d t t-1 k = s=1905 (x s k +y s t k ) f t-s χ Z(s)k t Where χ Z(s)k is the characteristic function, which is defined as: = 1 if Z(s) t k > 0 { t χ Z(s)k 0 if t Z(s)k = 0 t Where Z(s)k is surviving capital stock in the financial year of t among the capital stock of road infrastructure invested in the financial year of s, which is expressed as follows. { t-1 Z(s) Z(s) t k= max{ k t-1 t-1 u=1905 Z(u) k,0} Z(s) t-1 k -(x s k +y s t k ) f t-s -y k if t > s x s s k +y if t = s k The second term on the left-hand side of max is the amount of retirement due to aging and the third term is retirement due to disaster. The third term is calculated by multiplying the amount of retirement due to disaster in the financial year of t by the proportion of the t-1 capital stock of road infrastructure invested in the financial year of t-1, Z(s) k to all capital t-1 stock of road infrastructure existed in the financial year of t-1 (= s= 1905 Z(s) t-1 k ). This is based on the assumption that all capital stocks of road infrastructures invested in each financial year suffer from disasters, and the degree of damage of all capital stocks is equal (the Director General for Economic and Fiscal Management, Cabinet Office; 2001). The capital stock R t k of road infrastructure managed by prefecture k in the financial year of t can be expressed as follows. R k t = t s=1905(x k s +y k s )-(d k s + y k s )=R k t-1 +x k t -d k t IV-2. The estimation method for necessary renewal cost The amount of investment which is equal to the sum of retirement due to aging d t k and retirement due to disaster yk t in the financial year of t is defined as the necessary renewal cost n t k in the financial year of t, which is expressed as follows. 45 nk t = dk t + yk t

11 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March As can be seen from this equation, all investments in each financial year until t-1 t-1 (d t k = s= 1905 (x s k +y s t k ) f t-s χ Z(s)k ) and amount of retirement due to disaster in the financial year of t ( y ) affect the necessary renewal cost n t k t k. This is especially important in estimating the necessary renewal cost in the future. This is because assumptions about how to set amounts of investments in each financial year after FY2013 affect the necessary renewal cost in the future. It is assumed that retirement due to disaster in each financial year after FY2013, is equal to the average value for FY (y k ). 6 y t k IV-3. Data We used the nominal amounts of investment and the deflator in FY Regarding the nominal amounts of investment, A) data of FY are based on from Annual Report of Road Statistics 1958 to Annual Report of Road Statistics 2014, B) data of FY are based on table in Annual Report of Road Statistics Regarding the deflator, see Appendix 1. A) Regarding the expenses for general road projects and the city planning street project in the national subsidy project and the local independent project about National Highway, Principal Prefectural Road and General Prefectural Road, the nominal amounts of investment, which is the sum of new construction, reconstruction and repair of road bodies, that of bridges, and that of pavement are calculated as follows. Road bodies: the value from subtracting the amount of compensation for land to be expropriated from the sum of road improvement, other repairs, investigation and other in the details of construction type Bridges: the total amount of construction of bridges and repair of bridges in the details of construction type Pavement: the total amount of construction of pavement and repair of pavement in the details of construction type In addition, regarding expenses for disaster recovery projects concerning National Highways, Principal Prefectural Roads and General Prefectural Roads, the nominal expenses for disaster recovery projects for roads and that for bridges are calculated as follows. 4 It is necessary to renew not only bridges, tunnels and pavement but also road bodies. According to a survey of road administrators in Ministry of Land, Infrastructure and Transport (2010), almost half the administrators answered that they renewed all road body structures and 10% of administrators answered that they renew no road body structures. 5 Ministry of Land, Infrastructure and Transport (2010) assumed that the proportion of renewal costs to new investment cost of roads, that of bridges and that of pavement are 90%, 150% and 110%, respectively, on the basis of a survey of road administrators. However, it is a survey that lacks in objectivity. Therefore, this study assumed that renewal costs are the same as new investment costs in regards to all three sectors. 6 FY2011 is excluded as it is the year when the aftermath of the 2011 Tohoku earthquake and tsunami occurred.

12 106 N Akai, T Takemoto / Public Policy Review The nominal expenses for disaster recovery projects for roads: road disaster of total in the details of construction type. However, in the case of National Highways, the value subtracting road disasters of the project directly carried by the central government in the details of construction type from road disasters of total in the details of construction type. The nominal expenses for disaster recovery projects for bridges: bridge disaster in the details of construction type. However, in the case of National Highways, the value subtracting bridge disaster of the project directly carried by the central government in the details of construction type from bridge disaster of total in the details of construction type. Because the classification of data until FY1974 is different from the others, these are calculated as Appendix 2. B) The nominal amounts of investment, which is the sum of new construction, reconstruction and repair of road bodies, that of bridges, and that of pavement are calculated by multiplying the sum of expenses for construction and expenses for maintenance by the following proportion. Road bodies: the average value for FY of the proportion of the nominal investment of road bodies above A) in each prefecture to the sum of the expenses for general road projects for all roads in Japan and the expenses for the city planning street projects for all roads in Japan Bridges: the average value for FY of the proportion of the nominal investment of bridges above A) in each prefecture to the sum of the expenses for general road projects for all roads in Japan and the expenses for the city planning street projects for all roads in Japan Pavement: the average value for FY of the proportion of the nominal investment of pavement A) in each prefecture to the sum of the expenses for general road projects for all roads in Japan and the expenses for the city planning street projects for all roads in Japan In addition, the nominal expenses for disaster recovery projects for roads and that for bridges are calculated by multiplying the expenses for disaster recovery projects by the following proportion. The nominal expenses for disaster recovery projects for roads: the average value for FY of the proportion of the nominal expenses for disaster recovery projects for roads above A) in each prefecture to the sum of the expenses for general road projects for all roads in Japan and the expenses for city planning street projects for all roads in Japan

13 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March The nominal expenses for disaster recovery projects for bridges: the average value in each prefecture for FY of the proportion of the nominal expenses for disaster recovery projects for bridges above A) in each prefecture to the sum of the expenses for general road projects for all roads in Japan and the expenses for city planning street projects for all roads in Japan V. Changes in investments in roads managed by prefectures and the capital stock of road infrastructure V-1. Changes in investments in roads managed by prefectures Figure 4 indicates the changes in investments in roads managed by prefectures from the financial year of 1956 (FY1956) to FY2012 and their breakdown. This is an accumulated stick graph consisting of five items, which are the sum of new construction, reconstruction and repair of road bodies, that of bridges, that of pavement, expenses for disaster recovery projects for roads, and that for bridges. In addition, all the values are deflated by the deflator with base year 2005 (See Appendix 1). The investments in roads managed by ordinance-designated cities are omitted from the data of the investments in roads managed by prefectures. Therefore, Figure 4 can be compared across different financial years. The investments were 82,600 million yen in FY1956 and 1,390,700 million yen in Figure 4 Changes in investments in roads managed by prefectures (FY1956 FY2012) Source: Annual Report of Road Statistics ( )

14 108 N Akai, T Takemoto / Public Policy Review FY1972, a significant increase of approximately 17 times. The investments were stable for around 1,500 billion yen between FY1977 and FY1984. After FY1985, it grew again and was 2,944,000 million yen in FY1996. After FY1997, it was decreasing and 1,377,800 million yen in FY2012, which was almost half of its peak and even under the investment in FY1972. In addition, the breakdown of investments was changing. The sum of new construction, reconstruction and repair of pavement was 37.4% of the total investment in FY1969, the highest percentage ever. After that, it was decreasing and 10.8% in FY2012. In contrast, the percentage of the sum of new construction, reconstruction and repair of roads to the total investment increased from 29.7% in FY1956 to 70.5% in FY2012. The percentage of the sum of new construction, reconstruction and repair of bridges to the total investment decreased from 28.9% in FY1957 to 13% in FY1967 and was stable between 12% and 15% after that. V-2. Changes in the capital stock of road infrastructure Figure 5 indicates the changes in the capital stock of road infrastructure estimated by authors from FY1956 to FY2012 and their breakdown. The capital stock of road bodies increased by 116 fold from 485,400 million yen in FY1956 to 56,287,400 million yen in FY2012. The capital stock of bridges increased by 26 fold from 438,300 million yen in FY1956 to 11,198,400 million yen in FY2012. The capital stock of pavement increased by Figure 5 Changes in the capital stock of road infrastructure (FY1956 FY2012)

15 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March fold from 101,500 million yen in FY1956 to 1,722,500 million yen in FY2012. Therefore, the capital stock of road infrastructure increased by 68 fold from 1,251,000 million yen in FY1956 to 69,208,300 million yen in FY2012. The capital stocks of road bodies and bridges have consistently increased in that period. In contrast, the capital stock of pavement reached its peak, 3,106,700 million yen, in FY1995 and was decreasing after that. The rates of change of the capital stock of road bodies and bridges from the previous financial year were highest in FY1965 at 23.0% and 10.4% respectively. After that, these were consistently decreasing and reached the lowest in FY2011 at 1.5% and 1.0% respectively. The rate of change of the capital stock of pavement from the previous financial year reached its peak, 29.0%, in FY1961. After that, the rate fell sharply, and was around 0% from FY1980 to FY2000 and around -5% from FY2000 to FY2012. The capital stock of road infrastructure has consistently increased since FY2012. However, the rate of change of the capital stock of road infrastructure has been decreasing. The capital stock of road infrastructure will reach its peak in the near future. VI. Results of estimates of renewal investments until FY2040 We clarify the changes in the renewal investment from FY2013 to FY2040 in the following three cases. 1) the case in which the new investment is set as equal to the current level, 2) the case in which the new investment is set as zero (all the necessary renewal investments are made), and 3) the case in which the total investment in roads is set as equal to the current level. Note that new investment in the future affects the cost required for renewal in the future (the necessary renewal investment). It is assumed that the expenses for disaster recovery projects in each financial year after FY2013 are equal to average value for FY (see Section VI-2). VI-1 The case in which new construction investment is equal to the current level First, we estimate the new investments and renewal investments in the case in which new investments from FY2013 to FY2040 are set as equal to the average of the period from FY2008 to FY2010. In this case, the renewal investment for each year is set as equal to the necessary renewal cost of each year. Figure 6 indicates the changes in new investments and renewal investments estimated by authors from FY2013 to FY2040. The renewal investment of road bodies will increase by 3.6 fold from 280,700 million yen in FY2013 to 1,004,600 million yen in FY2040. The renewal investment of bridges will increase by 2.9 fold from 83,600 million yen in FY2013 to 240,700 million yen in FY2040. The renewal investment of pavement will increase 1.1 fold from 192,500 million yen in FY2013 to 206,100 million yen in FY2040. Therefore, the total renewal investments will increase by 2.6 fold from 556,700 million yen in FY2013 to 1,451,400 million yen in FY2040. By adding the annual new investment of 1,065,100 million yen, total investment in roads will increase by 3.6 fold from 1,621,800 million yen in

16 110 N Akai, T Takemoto / Public Policy Review Figure 6 Changes in new investments and renewal investments from FY2013 to FY2040 in the case in which new construction investment is set as equal to the current level FY2013 to 2,516,500 million yen in FY2040. This is close to the highest peak, FY1992 FY2002. VI-2. The case in which the new investment is set as zero (all the necessary renewal investments are made) Next, we estimate the renewal investments in the case in which the new investments from FY2013 to FY2040 are set as zero and all the necessary renewal investments are made. In the case, the renewal investment of each year is assumed to be equal to the necessary renewal cost of each year. Figure 7 indicates the changes in the renewal investments estimated by the authors from FY2013 to FY2040. Unlike the case of the previous section, the capital stock of road infrastructure is same as the level in FY2013. The renewal investment of road bodies, that of bridges and that of pavement in FY2040 will be 977,400 million yen, 234,200 million yen and 165,200 million yen, respectively. This is a decrease of 27,200 million yen (3%), 6,500 million yen (3%) and 40,900 million yen (20%), respectively, compared to the case of the previous section. Therefore, the total renewal investments in FY2040 will be 1,376,800 million yen and a decrease of 74,600 million yen (5%) compared to the case of the previous

17 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March Figure 7 Changes in new investments and renewal investments from FY2013 to FY2040 in the case in which the new investment is set as zero section. In this case, the assumption that new investment is set as zero is strict, but total renewal investments are at least 70,000 million yen less than in the case in which the new investment is set as equal to the current level. However, this is just a mere part of the effect of no new investment because the average service life is more than 50 years and the amounts of retirements for 20 years after investment are small. This must be effective in reducing total renewal investments in FY2040 and later. VI-3. The case where the total investment in roads is set as equal to the current level Even though the assumption of the previous section that new investment is set as zero is not realistic, total renewal investments in FY2040 will be 1,376,800 million yen and approximately equal to the total investment in FY2012. In addition, this is a comparison of the total amounts of all prefectures. The situations may be more serious in parts of prefectures. Accordingly, we estimate the total investment in the case in which the total investments from FY2013 to FY2040 are set as equal to the current level. Each prefecture s total investments in three sectors (road bodies, bridges and pavement) from FY2013 to FY2040 are set

18 112 N Akai, T Takemoto / Public Policy Review as equal to the average of the period from FY2008 to FY2010 (hereafter referred to as the fixed total investment). If the necessary renewal cost of a sector in a prefecture is less than the fixed total investment of its sector, the renewal investment is set as equal to the necessary renewal cost and the remaining amount is applied to the new investment. If the necessary renewal cost of a sector in a prefecture is more than the fixed total investment of its sector, the renewal investment is set as equal to its fixed total investment and the new investment is set as zero. In the latter case, all capital stock retired due to aging and disaster cannot be renewed because the renewal investment is less than the necessary renewal cost (hereafter the amount of capital stock that cannot be renewed is referred to as renewal deficit ). Figure 8 indicates changes in new investments, renewal investments and renewal deficits estimated by authors from FY2013 to FY2040. The renewal investment of road bodies, that of bridges, pavement and the total renewal investments in FY2040 will be 921,200 million yen, 178,800 million yen, 157,200 million yen and 1,257,200 million yen, respectively. They represent a decrease of 56,200 million yen (6%), 55,400 million yen (24%) and 8,000 million yen (5%) compared to the case in which the new investment is set as zero (the section VI-2), respectively. Therefore, the total decrease costs are 119,600 million yen (9%). This includes not only the effect that all the necessary renewal investments are not made in Figure 8 Changes in new investments, renewal investments and renewal deficits from FY2013 to FY2040 in the case in which the total investment in roads is set as equal to the current level

19 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March FY2040 but also the effect that the renewal investments of the capital stocks which have not been renewed in FY2039 and before are not incurred. However, the renewal deficits of road bodies and that of bridges in FY2040 will be 78,400 million yen and 59,600 million yen, respectively. Therefore, the total renewal deficits in FY2040 will be 138,000 million yen and 18,400 million yen more than the amount of the decrease in the renewal investments compared to the case in which the new investment is set as zero. This is because the situations are different depending on the prefecture. The total renewal deficits in FY2040 will be the total amounts of all prefectures and the positive value does not show that renewal deficits occur in all prefectures. Focusing on the situation by prefecture, there are some prefectures in which the necessary renewal cost of a sector is less than the fixed total investment and new investment will be made. In fact, the amount of the new investment will be 229,900 million yen in FY2040. The capital stocks in some prefectures are more than in the case in which new investment is set as zero. Renewal investment in these prefectures is more by the part. This is because the total renewal deficit is more than the amount of the decrease in renewal investments compared to the case in which the new investment is set as zero. In other words, it is because some prefectures reduce the capital stock and other prefectures increase the capital stock. VI-4. New investments and renewal deficits by prefectures Here, we look closely at the situation in the previous section by prefecture. As mentioned above, if the necessary renewal cost of a sector in a prefecture is less than the fixed total investment of its sector, the renewal investment is set as equal to the necessary renewal cost and the remaining amount is applied to the new investment. If the necessary renewal cost of a sector in a prefecture is more than the fixed total investment of its sector, the difference between the necessary renewal cost and the fixed total investment is the renewal deficit. However, if a renewal deficit in any sector occurs, first its deficit is compensated for within the capacity for the new investment of the other sectors. Therefore, the value necessarily becomes a positive, which shows the new investment, or negative, which shows the renewal deficit. Figure 9 indicates the new investments or the renewal deficits by prefectures in FY2010, FY2020, FY2030, and FY2040. Positive values in the vertical axis show the new investments and a negative value shows the renewal deficits. Kagoshima prefecture is explained as an example below. The new investment in FY2010, in FY2020 and FY2030 are 25,100 million yen, 18,700 million yen and 9,000 million yen. In contrast, the new investment in FY2040 will be zero and the renewal deficit is 800 million yen. In both FY2010 and FY2020, all prefectures can renew all capital stocks that are retired due to aging and disaster. However, 5 prefectures and 21 prefectures cannot renew all capital stocks that are retired due to aging and disaster and produce a renewal deficit in FY2030 and in FY2040, respectively.

20 114 N Akai, T Takemoto / Public Policy Review Figure 9 New investments and renewal deficits by prefectures in FY2010, FY2020, FY2030, and FY2040 in the case in which the total investment in roads is set as equal to the current level VI-5. Differences in the aging degree of the capital stocks of road infrastructure As mentioned in the previous subsection, the gap of capacity for new investment among prefectures is caused by the fact that the necessary renewal cost per capital stock in the future is different between prefectures. Namely, the aging degree of the capital stocks of road infrastructure is different between prefectures. Figure 10 indicates the breakdown by age of the capital stocks of road infrastructures. 7 In this figure, the capital stocks of road infrastructures, which are invested in FY1970 or before, are divided by all capital stocks in FY2010, from FY1971 to FY1980, from FY1981 to FY1990, from FY1991 to FY2000 and from FY2001 to FY2010 are shown by prefectures. For example, in Aomori, Tokyo, Yamanashi, Hyogo, Shimane, Ehime, Fukuoka, Saga and Oita, the proportion of the capital stocks of road infrastructures that are invested within 20 years is high. If the proportion of the capital stocks of road infrastructures whose age is young is high, the necessary renewal cost is less than otherwise, even though the amounts of the capital stocks of road infrastructures are the 7 Hamagata and Hitomi (2009) showed the stock-weighted average age of capital stocks by prefectures.

21 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March Figure 10 Age profile of capital stock of road infrastructure for prefectures in FY2010 same. In general, local allocation tax grants narrow the gaps in fiscal capacities among prefectures. However, local allocation tax grants for roads are calculated in accordance with road area. If the necessary renewal costs per road area are different with prefecture, local allocation tax grants cannot narrow the gaps in fiscal capacities among prefectures. Figure 11 indicates the necessary renewal costs per road area by prefecture. This shows that the necessary renewal costs per road area are different between prefectures and that local allocation tax grants cannot accommodate difference in aging among prefectures. VII. Capacity for total investment considering depopulation Japan s population began to decline in In the future, the country will face further aging of society with fewer children and move into a full-fledged era of population decrease (Ministry of Health, Labour and Welfare, Annual Health, Labour and Welfare Report ). If the total amount remains at the same level, the per capita expenses increase. In this section, we clarify the changes in renewal investment from FY2013 to FY2040 in the case where the per capita total investment in roads is set as equal to the current level.

22 116 N Akai, T Takemoto / Public Policy Review Figure 11 Necessary renewal costs per road area by prefecture in FY2010 VII-1. Population projection The following is an overview of population projection until FY2040. Figure 12 indicates the percentage of total population projection for each prefecture in FY2025 and FY2040 to the total population of each prefecture in FY2010 and the percentage of the working age population projection for each prefecture in FY2025 and FY2040 to the working age population for each prefecture in FY2010. Regarding total population, the percentage in Okinawa is the highest, 102% in FY2025 and 98% in FY2040. The percentage in Akita is the lowest, 82% in FY2025 and 64% in FY2040. The values about the working age population are less. The percentage in Tokyo is the highest in FY2025, 95%, and the percentage in Okinawa is the highest in FY2040, 84%. The percentage in Akita is the lowest, 72% in FY2025 and 54% in FY2040. VII-2. The case in which the per capita total investment in roads is set as equal to the current level We estimate renewal investments in the case of per capita total investments from FY2013 to FY2040 being set as equal to the current level. Each prefecture s per capita total

23 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March Figure 12 Population projection compared to FY2010 levels for prefectures Source: National Institute of Population and Social Security Research Regional Population Projection for Japan: (March 2013) investments in three sectors (road bodies, bridges and pavement) from FY2013 to FY2040 are set as equal to the value obtained by multiplying the average of the period from FY2008 to FY2010 by population projection from FY2013 to FY2040 (hereafter referred to as total investment in the case of the fixed per capita total investment). If the necessary renewal cost of a sector in a prefecture is less than the total investment of its sector in the case of the fixed per capita total investment, the renewal investment is set as equal to the necessary renewal cost and the remaining amount is applied to the new investment. If the necessary renewal cost of a sector in a prefecture is more than the total investment of its sector in the case of the fixed per capita total investment, the renewal investment is set as equal to its fixed total investment and the new investment is set as zero. In the latter case, all capital stock that is retired due to aging and disaster cannot be renewed because renewal investment will be less than the necessary renewal cost. Figure 13 indicates the changes in the new investments, renewal investments and renewal deficits estimated by authors from FY2013 to FY2040. The renewal investment of road bodies, that of bridges, pavement and the total renewal investments in FY2040 will be 735,900 million yen, 133,100 million yen, 109,300 million yen and 975,300 million yen, respectively. This is a decrease of 185,300 million yen (20%), 45,700 million yen (26%) and

24 118 N Akai, T Takemoto / Public Policy Review Figure 13 Changes in new investments, renewal investments and renewal deficits from FY2013 to FY2040 in the case in which the per capita total investment in roads is set as equal to the current level 47,900 million yen (30%) compared to the case in which the total investment is set as equal to the current level (the section VI-3), respectively. Therefore, the total decrease costs are 278,900 million yen (22%). In addition, total new investment in FY2040 will be 64,500 million yen and is a decrease of 165,500 million yen (72%) compared to the case of section VI-3. Therefore, total investment in FY2040 will be 1,042,800 million yen and is a decrease of 165,500 million yen (72%) compared to the case of section VI-3. Just like the case of section VI-3, this is caused by the effect of all necessary renewal investments not being made in FY2040 and the effect of renewal investments in capital stocks not being renewed in FY2039 and before are not incurred. Considering depopulation, total investments in the case of the fixed per capita total investment are less than the fixed total investments. Therefore, the new investments and the renewal investments are less by the part. And the necessary renewal costs in FY2040 will be a decrease of 33,100 million yen (2%) compared to the case of section VI-3.

25 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March VII-3. New investments and renewal deficits by prefecture Here, we look closely at the situation of the previous section by prefecture. The assumption is same as Section VI-4. Figure 14 indicates the new investments or the renewal deficits by prefecture in FY2010, FY2020, FY2030, and FY2040. Positive values in a vertical axis show the new investments and negative value in a vertical axis show the renewal deficits. In FY2010, all prefectures can make new investments. However, a prefecture cannot renew all capital stocks that have been retired due to aging or disaster and produce a renewal deficit in FY2020, 17 prefectures in FY2030 and 44 prefectures in FY2040. The above difference among prefectures is caused by the necessary renewal cost per capital stock in the future being different between prefectures, such as the explanation in section VI-4, and by the pace of depopulation being different between prefectures. Because the fixed per capita total investment is the value obtained by multiplying the current per capita total investment recommendation degrees by population projection, the differences in the pace of depopulation between prefectures affect the gap of capacity for new investment. After all, the differences in both the aging degree of capital stocks of road infrastructure and Figure 14 New investments and renewal deficits by prefectures in FY2010, FY2020, FY2030, and FY2040 in the case in which the per capita total investment in roads is set as equal to the current level

26 120 N Akai, T Takemoto / Public Policy Review the pace of depopulation among prefectures give rise to differences in the new investments among prefectures. The local allocation tax grants for roads are calculated in accordance with road area. On the other hand, traffic density and population include correction coefficients of the local allocation tax grants. Even though local allocation tax grants cannot accommodate difference in aging between prefectures, depopulation reduces the amount of local allocation tax grants. VIII. The effect of prolonged life VIII-1. Re-examination of the assumption of average service life The results in section VI showed that some prefectures cannot renew the capital stocks sufficiently. In addition, considering the decline in capacity for total investment due to depopulation, the results in section VII showed that almost all prefectures produce a renewal deficit. It is necessary to re-examine the assumption of average service life. The average service life of road bodies and bridges is set as 60 years and 52.5 years, respectively. This is valid from renewal results in the past. However, many renewals are not related to the problem of service life. The proportion of improvement construction and the problem of function to the reason of replacing the bridges is high, from 70% to 80% in total. The proportion of damage to the reason of replacing the bridges is from 10% to 20% (Tamakoshi, Ookubo, Ichikawa, and Tatsuya Takeda; 2008). If the renewal of still-usable road infrastructure is stopped, service life is lengthened. In addition, the recently attended preventive maintenance can lengthen service life with technical ease. VIII-2. The effect of prolonged life Here, we clarify the effect of prolonged life in the case in which the average service life of road bodies and bridges is set as 70 years and 80 years, respectively. Figure 15 indicates new investments or renewal deficits by prefectures in FY2010, FY2020, FY2030, and FY2040 in the case in which the capital stocks in section VI-3 are set as long-life. In FY2030, all prefectures can renew all capital stocks that are retired due to aging and disaster. Furthermore, the number of prefectures that produce a renewal deficit in FY2040 will have decreased from 21 to 2. This result shows that prolonged life improves some prefecture s situation. Similarly, Figure 16 indicates the case in which the capital stocks in section VII-2 are made long-life. The number of prefectures that produce a renewal deficit in FY2030 and FY2040 are 1 and 19, respectively. In the case in which the total investments are set as the current level, the prolonged life

27 Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.12, No.1, March Figure 15 New investments and renewal deficits by prefectures in FY2010, FY2020, FY2030, and FY2040 in the case in which the capital stocks in section VI-3 are made long-life of road infrastructure is one of the effective solutions. However, in the case in which the per capita total investments are set as the current level, just the prolonged life of road infrastructure cannot sufficiently improve the situation. IX. Conclusion This paper estimates the future renewal cost for roads that have been managed by each prefecture. In addition, we calculate the effect on future renewal costs by reductions in new investment. The result shows that the renewal investment becomes 2.6 times of the current cost when new investment continues at the current speed and that 5% of costs can be saved when new investment is stopped in FY2013 and continues with no investment policy. It should be noted that the effect of a reduction of the new investment on future costs emerges over a long period of time and becomes larger when we consider the life cycle cost. In addition, the result shows that even when we adopt no investment policy, the renewal investment in FY2040 exceeds the total amount of investment in FY2012. The situation in different prefectures is expected to be different. Therefore, we simulate a situation in which the total amount of expenditure for roads is fixed at the current level. Then the result shows