3.2 E MPIRICAL S TEPS IN E STIMATING THE DRC
3.2.8 Social valuation of capital
3.2.8.1 The estimation approach
Capital input is measured as the sum of depreciation, i.e. the reduction in the value of assets arising from wear and tear, and the forgone return on financial capital tied in the value of assets (Corden, 1984; Griliches, 1960). Estimation approaches differ according to method of calculating the depreciation and the data constraints. In this study capital assets are assumed to provide a constant stream of services over the useful (economic) lifetime (Griliches, 1960). Algebraically, this method points to calculation of capital recovery factor, i.e. a number which, when applied to the initial value of assets yields a constant annual payment (user cost of capital). This annuity represents the value necessary to repay the initial value of investment in the assets over a given period of time (equal to the economic life of assets) and at a given interest rate (Tsakok, 1990 and Brandes and Odening, 1992).
The underlying formula is:
V - initial value of capital assets (investment) i - real rate of interest
For the private cost of capital the value of assets should represent private prices of investment goods and the private interest on financial capital engaged. The private interest rate i.e. private opportunity cost of financial capital reflects: (i) the costs of borrowing, in the case of the credit, and/or (ii) the foregone returns from the best alternative, possibly a return to long-term bank deposits or other financial assets, in the case of own financial resources engaged. Calculatory interest rate can be derived as the average of the two interests weighted by their shares in the financing of initial investment.
In social valuation of capital cost one needs to take account of distortions in the market for capital goods, e.g. distortions resulting from tariffs or taxes, and the divergence between private and social interest rates. The procedure to estimate social prices of capital goods is the same as in the case of tradable inputs, namely, it has been assumed that social prices diverge from the private ones by the effect of policy distortions: import tariffs and taxes. The social value of capital assets has been
3 Recent Comparative Advantages 71 derived from the private values by applying a conversion factor based on the information about the applied tariffs and taxes. The social interest rate corresponds to the marginal productivity of capital in the economy, i.e. the return forgone by the economy by engaging the scarce financial resources in the activity under consideration, rather than in the best alternative (e.g. Monke and Pearson, 1989 or Mashayekhi, 1980). Alternative approaches to estimating social opportunity cost of capital funds include (i) estimation of the marginal productivity of capital on the basis of macroeconomic or microeconomic data, and (ii) use of rates on borrowing from representative foreign capital markets e.g. Eurocurrency rates.
After having identified private (Vp) and social (Vs) values of capital assets, and private (ip) and social (is) interest rates on financial capital one can estimate private and social capital costs. This can be done by an application of formula (3.18) for private and social values of capital assets and interest rates. For technical reasons, it is more convenient to use a conversion factor for capital costs, a number which when multiplied by the private capital costs (UCp), yields the social value for the capital cost (UCs). The conversion factor for capital (CFC) is calculated as:
CFA CFI
The conversion factor for capital costs (CFC) is a product of the conversion factor for capital assets value (CFA) and the conversion factor for capital recovery factor (CFI). Assuming that the divergence in the value of capital assets stems solely from the tariff rate, theCFA is defined as:
)
The conversion factor for the capital recovery factor (CFI) can be estimated directly from private and social values for interest rate.
Treatment of foreign direct investment
Since the DRC serves to evaluate whether an activity represents an efficient (potential) use for domestic scarce resources the presence of foreign capital in the industry does not affect the calculation provided the domestic and foreign capital assets are substitutable in the technological sense. The domestic capital may not be a perfect substitute for FDI if the observed level or efficiency is only possible due to the presence of FDI. Such dependence takes place if the technology used by the FDI is not purchasable/attainable for the (potential) domestic investors, i.e. when FDI contributes some sort of firm-specific assets affecting factor productivity. In such a case the capital attributable to FDI may need to be treated as an imported (tradable) input of which the social cost is determined by the profits repatriated (profits net of corporate taxes paid in a host country). At the same time benefits from the FDI accrued by the host country (here Poland) should be accounted for by the use of
FDI-3 Recent Comparative Advantages 72 influenced technical coefficients and/or by the (higher) social revenues from the quality gains, or from improvement in the bargaining position on the world market.
Here these effects have been neglected. Even if the scope for the gains from FDI due to transfer of non-purchasable technologies were high, they have not been exploited to a significant degree in the year 1997. The share of FDI in the industry was relatively low and the existing FDIs concentrated on the selected product-lines (ice creams, milk drinks). There is some evidence that even in product-lines with the most up-to-date technology and with relatively high FDI share (like the yoghurt production - see Chapter 6) there existed successful enterprises based on domestic capital, which may suggest that, as far as the milk processing technology is concerned, the investment capital of FDI has been substitutable by the domestic one. In this case, however, the contributions of FDI may have a less tangible character as they may involve a range of indirect effects such as: demonstration effects and offsetting of institutional and policy constraints on the development of domestic-capital-based firms (e.g. by exerting competitive pressure that forces them to take active business strategies despite the mitigating effects stemming from trade and investment policies).
3.2.8.2 The data and estimation results
The gross value of assets taken from GUS (1999c) has been assumed to represent the cost of replacement investment for which annuity value has been calculated. The structure of assets (i.e. the share of buildings, machinery equipment and vehicles) has been used to take account of specific economic lifetimes of various types of assets.
Economy-wide, capital goods belong to the least protected category of imports. The nominal protection rate declined from 9.2% in 1995 to 3.35 % in 1998 (Marczewski, 1998). These rates are even lower for some specialised machinery used in the milk processing industry (see Annex Table 16). The average rate of protection and taxes has been used to calculate the conversion factor for value of asset in the industry (CFA).
Private interest rate
Estimates of the private real interest rate have been based on the information about nominal interest rates, inflation and interest rate subsidy operated under the Programme of Restructuring and Modernisation of the Dairy Sector (see also Annex Table 18). The share of own capital has been estimated by combining income statement data (value of long-term credits) (GUS, 1998b) with the data on interest rates under the subsidy scheme. By using the subsidised interest rate as private interest rate (i.e. adjusted for the own versus borrowed capital fund structure), it has been implicitly assumed that all future payments during the considered economic lifetime will proceed at the same interest rate (Table 3.4). This assumption is consistent with the projection of future development of
3 Recent Comparative Advantages 73 real interest rates in Poland (Chapter 4). The private interest rate for indirect capital (capital embodied in the non-tradable inputs) has been assumed to equal the market interest rate.
Table 3.5. Lending interest rates faced by the milk processing industry, 1993-1998 (in %).
1993 1994 1995 1996 1997 1998 1996-1998
Nominal interest rates:
Official discount (1) 29.0 28.0 25.0 22.0 24.5 18.2 21.6 Commercial credit (2) 45.0 41.0 36.0 25.0 24.9 24.5 24.8 Subsidised credit (PMRM) (3) 10.9 10.5 9.4 8.3 9.2 6.8 8.1
CPI (4) 35.3 32.2 27.8 19.9 14.9 11.2 15.3
Real interest rates (IR):
Market real IR (5) 7.2 6.7 6.4 4.3 8.7 12.0 7.8
Private IR:
Subsidised credit (PMRM) (6) -18.1 -16.4 -14.4 -9.7 -5.0 -3.9 -8.3 Share of sub. cred. in funds (7) 0.0 18.3 24.3 24.1 25.2 23.0
Private IR (8) 7.2 2.4 1.3 0.9 5.3 8.3 4.0
(1) GUS (1999b).
(2) Twelve moth moving average of simple averages from lowest and highest rates for commercial credits.
(3) Equal to 0.25% of 1.5 of official discount rate (MAFE, 1998).
(4) GUS (1999a).
(5) IR(real) =(100+IR (nominal))/(100+CPI) -1; see Brandes and Odening (1992). pp. 131.
(6) See point (5).
(8) Average of market and subsidises weighted by shares in investment funds.
Source: author’s estimates.
Social interest rates
The social rate of return to capital can be estimated in an approximate way using the macroeconomic data on gross domestic product (GDP) distribution (Monke and Pearson, 1989). Such analysis has been conducted here for the period 1995-1998 (Table 3.6). Under assumption of competitive conditions the ratio of marginal to average value products should be equal to the share of the factor in the total income.43 Capital’s shares of GDP were on average 55 percent over the considered time period (GUS, 1999c). Gross marginal rates of return are therefore 55 percent of the average rate of return; the estimates vary between 10.0 percent in 1995 and 13.7 percent in 1998. After accounting for the capital assets depreciation the average net marginal rate of return (return to investment capital) in the years 1996-1998 equals to 7.7 percent.
43 This relationships is based on the marginal productivity theory and the ‘Euler’s product exhaustion‘ theorem (see e.g. Koutsoyiannis, 1979). The former predicts that in competitive conditions factor prices should be equal marginal value product. According to the latter, for production functions with constant returns to scale, factor payments (factor income); exhaust the total value of product. Consequently, for two factor case, capital and labour, and prices of: product, capital and labour denoted, respectively, as P, r and w, and quantities of:
product, capital and labour denoted, respectively, as Q, K and L it is true that:
P*Q = w*L + r*K , i.e. 1 = w*L/P*Q + r*K/P*Q , where r*K/P*Q = capital share of total value of product and r*K/P*Q = r/(P*Q/K) = ratio of marginal to average product of capital.
3 Recent Comparative Advantages 74 Table 3.6. Distribution of gross domestic product and rates of return on capital, Poland, 1995-1998.
GDP Labour Capital Land 1996 385448 167234 212895 5319 1086491 19.6 10.8 6.9 1997 469372 207323 257345 4704 1158653 22.2 12.2 7.7 1998 549467 239337 305822 4308 1243436 24.6 13.7 8.8 (1) and (5) GUS (1999c).
(2) Labour income calculated as the share of compensation of employees in GDP.
(3) Capital income calculated as the residual after subtracting labour and land incomes.
(4) Return to land calculated as the average rental used in tenure of 0.5 ton of wheat on a total of 18.5 million hectares.
(6)= (3)/(5).
(7) = (3)/(1)*(6).
(8) After subtraction of amortisation costs. Here assumed 11% of GDP (assumption based on Dornbush and Fisher, 1994), which corresponds to 4-5% of capital stock value.
Source: quoted above and own calculations.
The calculated rates may be influenced by the divergence in the wage rates. Such a wage rate effect may, however, be compensated by the change in the employment level, i.e. decreased employment in the case of policies, and other distortions, increasing unit labour cost. If, however, in the considered years the increased wages (see next section) have resulted in the higher labour share of total factor income (e.g. due to a lagged adjustment in employment level), then the estimated rates of return to capital underestimates ‘actual’ social marginal productivity of capital.
Relying on the above discussion it has been assumed that the social interest rate equals 8 percent, the value close to the estimate of the market lending rate and the marginal return calculated from macroeconomic data.44
Conversion factors for capital costs - results
Estimation results of conversion factors for the capital costs in the year 1997 are presented in Annex Table 18 (private and social capital costs for the entire industry) and in Table 3.7.
44 In a recent study on Polish agriculture Münch et al. (1997) used the 7.5% rate for the social interest rate for the years 1994-1995. In studies on developing countries the choice of the social opportunity rate is often made in a rather arbitrary way. These rates range between 8% (Pearson, et al., 1987, for Portugal) and 12%
(Barzelay and Pearson, 1982, for Brazil). Such relatively high values reflect the fact that less developed countries, which are the subject of most of the cost-benefit studies, are characterised by higher marginal productivity of capital compared to developed countries), e.g. due to the low relative endowments in this factor.
3 Recent Comparative Advantages 75 Table 3.7. Conversion factors for capital cost in the milk processing industry, 1997.
Symbol Values
Conversion factor for capital assets CFAd 1.03
Direct capital Conversion factor for capital recovery factor CFId 1.28
Conversion factor for capital cost CFCd 1.31
Conversion factor for capital assets CFAi 1.06
Indirect capital Conversion factor for capital recovery factor CFIi 1.00
Conversion factor for capital costs CFCi 1.06
Source: author’s calculation based to data and method described in the text.