Even based on observed pK changes, CMA,pK=0.044. If weighted by the greater increases p*K, the MA demand augmenting impact of capital price changes would be C*MA,pK =0.056. The implied higher growth of virtual compared to measured price of capital could result from various factors. Its drivers could include substantive and rising adjustment costs , environmental or safety standards, or taxes, that are not effectively captured in the measured user cost of capital. These capital costs motivate a substitution effect toward primary agricultural products. In turn, growth in the scale of production, or output demand, has had a greater-than proportional effect on the augmentation of MA demand; eMA,Y=1.095 on average for the full sample, implying CMA,Y=0.024.24 And although eMA,Y>1 implies scale effects are MA-using, they are even more MF-using, so in this sense they are relatively MA-saving. By contrast to the positive substitution and scale influences on MA use, disembodied technological shift impacts on MA demand have been negative, and in a direct sense, quite large. That is, an input-cost-diminution impact associated with MA demand is evident = – 0.008 on average, that is typically interpreted as deriving from disembodied technical change.
This trend is statistically relevant; the eMA,t estimates are significantly different from zero for most individual observations.25 And this tendency was augmented post-1980 = – 0.021. The direct t- and t2- impacts are, however, much greater in magnitude than these total measures,big plastic pots since much of the direct trend effects are counteracted by effective price trends that may be interpreted as embodied technical change or adjustment costs, as alluded to above. These patterns can be seen from the decompositions of the total trend and structural change impacts in the first section of Table 2, that arise from the inclusion of t- terms in the p*MA and p*K specifications.For our scenario, although eMA,pMA < 0, since the trend component of p*MA is negative , the indirect p*MA effect on MA demand is positive – as is the p*K effect since K is a substitute but p*K is rising .This evidence is consistent with the embodied technical change interpretations of the timpacts on effective prices implied by the discussions of the p*MA and p*K as compared to pMA and pK changes above. Declines in effective as compared to measured pMA, and the reverse for pK, both tend to augment MA use. Escalation of the equipment-to-structure ratio, representing another form of embodied technical change, also had a positive impact on the demand for MA; CMA,ES = 0.014. Total Cost Implications In addition to the specific MA impacts, the total cost effects of adaptations in the economic and technological climate are of interest individually, as well as providing indications of input biases .
The cost effect most directly associated with the use of MA is represented by the eTC,pMA = 0.025 elasticity, indicating the impact on costs of pMA changes, which depends on the input intensity or average share of MA for industries that use agricultural commodities.This is larger than the corresponding elasticity for any other input; rising pMA has a substantive positive impact on production costs, and thus on output production/price, in the food processing industries. Note, however, that the overall pMA contribution to total cost increases of CTC,pMA=0.014 is not only smaller than that for capital , but is also is even lower if the smaller increase in effective pMA is recognized within this measure . The eTC,Y estimate of 0.868, which implies significantly increasing returns to scale, also deserves attention. This evidence is largely driven by a very small capital-output elasticity, that counteracts the eMA,Y elasticity of slightly more than 1, and an eMF,Y elasticity that is even higher , which suggests scale expansion is somewhat MA-using, and significantly K-saving and MF-using. This is of particular interest since this conclusion is closely linked to the inclusion of t in the lK and lMA specifications. When t is not included as an argument in these specifications , output increases instead appear MA-saving , and both eK,Y elasticity and eTC,Y elasticity estimates are much closer to 1, implying close to constant returns to scale. These patterns highlight two issues alluded to above. First, apparent declines in the MA-input-intensity of output production in the food industries are partly associated with increases in effective or quality-adjusted MA-inputs, perhaps due to embodied technical change. Second, adjustment costs for capital implied by a higher and more quickly rising p*K than pK may mean that these estimates should be interpreted as short-run, or at least capital-adjustment-constrained estimates.
And both of these impacts, if ignored, affect estimation of the scale- or output-effects. Finally, the elasticities associated with disembodied and capital-embodied technical change deriving from t and ES changes, and with structural changes in the 1980s , suggest other technological forces have contributed to cost diminution. The negative values for both CTC,t = -0.004 and CTC,t2 = -0.012, augmented by the embodied technical change impact CTC,ES = -0.041, highlight such trends, and their enhancement in the 1980s, and from technological advance embodied in equipment. However, the total disembodied technical change impact becomes positive – CTC,t = 0.0004 – when the higher cost of capital is recognized, even though the analogous effect for p*MA is in the opposite direction . By contrast, CTC,t2 is even more negative than its direct counterpart, since CTC,p*K,t2 = -0.0025 outweighs CTC,p*MA,t2 = 0.001. Note also that the input-specific CMA,t = -0.0525 measure is much larger than the associated overall input declines captured by CTC,t = -0.004, and the total MA effect CMA,t is negative whereas that for TC, CTC,t is positive, indicating that “technical change” has been both relatively and absolutely, MA-input-saving. Over time there has been a technical change bias toward reducing MA use more than other inputs for a given level of output.27 Marginal Cost and Output Price To move toward consideration of the pass-through of MA prices to output price, as well as its impact on scale economies, we can compare these estimates to those for marginal cost in the third panel of Table 1. Note that the input price effects for the materials and labor inputs are slightly larger for MC than for total cost,growing berries in containers implying a depressing impact on scale economies . The reverse is true, however, for the pK and pE elasticities, supporting the notion that capital is subject to adjustment costs, and “lumpiness”, that are driving forces for returns to scale. This is also consistent with the virtually nonexistent MC impacts of changing output. And with the fact that marginal cost has decreased significantly over time, both in terms of the direct and indirect effects, largely due to the smaller impact of pK on MC than on TC. Comparing these measures to those for pY provides some insights about markup behavior, and its determinants. The average epY,pMA = 0.272 elasticityis larger than either eTC,pMA, or the eMC,pMA. So a 1 percent increase in pMA drives a somewhat larger increase in AC than MC, and an even greater adaptation in pY than MC. This implies a higher markup pY/MC associated with a rise in pMA, but also an increase in the scale economies that support such markups .Note also that pY decreases somewhat more than MC as time progresses, primarily due to the larger p*MA effect. Temporal and Industrial Variations In addition to the indicators for the data averaged for the entire sample, it is useful to briefly consider variations in the estimates over time and by industry, which are presented in Tables 3 and 4, respectively.
The temporal decompositions presented in Table 329 show a much smaller depressing contribution of pMA increases to MA demand post-1980, that results from low pMA growth; the measured eMA,pMA elasticity is actually larger later in the sample. Also note that the trend in the effective price of MA is actually downward for the post-1980 period, so the full contribution of own price changes to MA demand is positive. This tendency is particularly worth highlighting since measured pMA changes that occurred after the end of our sample period actually dropped, which implies that the implications from these measures may have been exacerbated. It also appears that although the growth rate of MA demand in the 1980s was larger than in the 1970s, the individual input price contributions were generally smaller, with less of the growth arising from output increases. In fact, a large proportion of MA demand expansion seems to have arisen from t-effects. In particular, the indirect p*MA effect has increased over time to the point where CMA,t is positive post-1980, although the direct impact, CMA,t, reported in Table 2, remains negative in the later time period. The TC measures for the 1970s as contrasted to the 1980s, presented in Table 3, indicate a much smaller average annual percentage increase in total costs for the food processing industries overall post-1980, that is only in part due to a slower output growth rate . All the contributions of individual TC determinants are smaller , although they remain statistically significant. In particular, the eTC,pMA elasticity is slightly lower in the 1980s, but the contribution falls more since pMA increased so little . The -estimate of the actual TC change in the 1980s seems to be driven by capital price effects, which appear in the CTC,pK measure of 0.014, as well as a positive CTC,p*K,t measure of 0.009 which augments the direct CTC,t = 0.004 .30 Although a full analysis of the 3-digit industries within the food processing aggregate is beyond the scope of this study, it is worth briefly considering the differences in MA demand that are apparent across these sub-samples, as reported in Table 4. First note that for the meat products industries very little substitution is apparent, as might be expected.Note also that the t-effect is very small, at only about 10% the magnitude of that for these industries as a whole. For the dairy industry, the own and cross-substitution responses seem similar to those for the overall food processing industries. But the t impact in total is very slightly positive, since the indirect adjustment – particularly the CMA,p*MA,t component – is quite large. The vegetables sector of the industry seems to be fairly responsive to the own price of MA. The p*K contribution, as well as the t elasticities are also large. The substantial t impacts on p*MA and p*K in fact suggest a particularly significant amount of embodied technology in the primary agricultural vegetable inputs, as well as high and increasing adjustment costs, likely due to the great scale and processing expansion in this industry. The grain mill and oil industries have exhibited quite different patterns.31 We find a negative output impact on MA demand for grains, both due to the very low eMAY elasticity , and observed output declines for some observations. Responsiveness to other factors seems generally low in this industry, except perhaps for ES. For the oil industries, we find the own contribution to be smaller than for most industries, and even less responsiveness to prices of other inputs, and thus substitutability; the cross-demand contributions are only about half those for the food industries as a group. By contrast, the output response is the largest of any other industry on average. For sugar and confectionary products the own price contribution is by contrast very large, although other substitution effects are somewhat small relative to the other industries. The pK impact is slightly more minor, and the CMA,t impact more major, than for the industry as a whole. And industries in the miscellaneous category have exhibited similar substitutability patterns to those apparent for the overall industry, except for very small capital/energy and technological contributions. Impacts of MA Price Changes Finally, in Table 5 we report elasticities that facilitate an evaluation of responsiveness to pMA changes, which may be thought of as a converse experiment to the evaluation of MA demand changes that began our discussion of empirical results. These measures facilitate investigation of the potential implications of the declines in pMA that were experienced by the food industries during the remainder of the 1990s not represented by our data sample.