The Netherlands have a large group of farms with grassland as the sole form of land utilization and cattle as the only economically important animal species. The major part of the food for the cattle on these farms consists of homegrown grassland products. In most cases a significant amount of supplementary feeding such as concentrates and other purchased feed-stuffs are administered in winter only.
There are several causes for which the structure of the dairy-grassland farms can be more variegated than it seems at first sight. Some of these factors being of fundamental importance to dairy-grassland farms form the theme of the present study.
In the first place the question arises whether roughage production and its turn-over into animal products should be integrated into one firm or divided between two firms. Even if it is economically desirable to combine roughage production and dairy cattle husbandry, the question has to be answered whether the combination should be complete or incomplete. Integration is complete if grassland production is attuned to the requirements of dairy cattle husbandry. In this situation the course of grass growth (fig. 5) and feeding requirements of the cattle during one year implies that slightly less than half of the grass production is fed as conserved food when the cattle pass the winter indoors. The combination of grassland production and dairy cattle husbandry per farm is incomplete when stocking density per ha of grassland is so high that all the grass or practically all of it is grazed during summer. Then all the feed for the winter season or nearly all of it is purchased.
In the second place a choice should be made as to the intensity of grassland production, which can be varied considerably by application of nitrogen fertilizers. This choice has its consequences for the herd size and for the ratio between purchased and homegrown feed-stuffs per animal.
In the third place it should be considered whether some differentiation can be effected within the dairy cattle husbandry i.e. the raising of young cattle and milk production by separate farms.
When the possibilities of varying the organization of dairy-grassland farms are explored on short run, the relation between available production factors per farm should be regarded as cardinal, particularly between those which cannot be acquired or repelled any time in any quantity, or if so, only at heavy financial sacrifices. Examples are accomodation and in particular the relation between grassland and working capacity, the latter being governed by the factors labour, equipment and traction. The optional production processes for designing the optimum farm plan vary markedly concerning the ratio in which they require grassland area and working capacity. It may be assumed that in a particular situation the production processes will be selected in such a way that the requirements of grassland and working capacity in the optimum farm plan - dependent on other factors such as price ratios - are more or less the reflection of the ratio between available grassland and working capacity.
As appears from the literature consulted, no exhaustive study has yet been made of all alternatives in mutual relation. Besides, in many cases models have been used giving a rather simplified picture of the actual situation. This is why it was considered desirable to investigate the various interrelated alternatives in a series of co-ordinated calculations which should approach reality as close as possible. This study was carried out for dairy-grassland farms employing a fixed labour force of two men (including the farmer). The grassland area was varied considerably. This variation enabled not only to calculate the optimum plan for each farm area but also which farm size is optimum for the given labour force. Different situations are investigated concerning buildings, implements and machinery, production level of the cattle and price ratios. Beef production is not considered. Chapter 11 goes further into the technical and economic relations on which the calculations are based. The estimation of costs and yields per category of cattle could not be made as accurately as for a herd with a given ratio between young stock and dairy cows. So the first investigations were based on a constant herd composition. Chapter III deals with the results of this part of the study and chapter IV with those recorded for a variable herd composition. A final discussion is given in chapter V.
II. INPUT/OUTPUT RELATIONS AND RESEARCH METHODS
The technical and economic relations on which this study is based (more precise data have been published in a separate documentation report) apply to the conditions of the peat grassland district of Friesland. They primarily concern the financial year 1964-65.Grassland
As far as the grassland area is concerned exploitation by a tenant-farmer was initially considered. Later on exploitation of farmer owned land was also taken into consideration.
The grassland was assumed to be of good quality with proper lay-out and drainage. Nitrogen fertilization may vary between 50 kg and 250 kg pure fertilizer-N per ha. P and K are applied according to the standards, depending on type of use which may range from grazing only to cutting only. Besides conserved grassland products can be purchased and sold.
For a particular soil quality net SE production of the grassland is supposed to depend on amount of N and on type of use. With all types of use the marginal return of grassland production to nitrogen fertilization decreases as the amount of N-fertilizer increases. This decrease is most rapid in case of exclusive or practically exclusive grazing. If a slight amount of fertilizer is applied, it is assumed that types of use with partial or exclusive cutting initially lead to a lag in net SE production. When the amount of N-fertilizer is increased this lag is reduced and even converted into a lead as compared with exclusive or practically exclusive grazing (table 1 and 2).
If the herd size is constant the seasonal cutting pattern is of course markedly determined by the shape of the curve of grass growth (fig. 5). Besides it depends on the percentage of cut grassland area'. On the basis of data from the farm accounts kept by the Agricultural Economics Research Institute in The Hague quantitative estimations have been made of the seasonal cutting pattern at varying percentages of cut grassland area (fig. 6). The same data were used in determining hay and silage yields per cutting in different months (table 3).Buildings
The farm buildings are supposed to be of an efficient conventional type (including a barn with a store room, an implement shed and a built-out doublerange cow-shed). Most calculations include building costs per animal based on replacement value. This applies to the situation in which no selection has yet been made for buildings and to the situation in which buildings are a limiting factor, although they can be enlarged. Other serial calculations have been made, marginal buildings costs being put at zero. This concerns the situation in which the buildings are no limiting factor.Cattle and feed rations
As stated before the herd composition is kept constant in the calculations of chapter III. According to the average conditions of the peat grassland district in Friesland the selection of size of young breeding stock has been suited to the own dairy herd requirements and have some young animals left for sale. On average 140 animal units are kept per 100 dairy cows. In the calculations of chapter IV, in which herd composition may vary, the cattle have been divided into three categories. In addition to dairy cows two categories of young cattle have been distinguished, of which one relates to calves up to the age of about 6 months (at the end of the first grazing season). The other category includes the age group of six months and up to the end of the first gestation. Initially milk production has been put at 4100 kg per cow per year and later on the effect of a higher production level has been studied.
With the aid of linear programming seven rations have been designed for milking cows. These rations meet feeding requirements of maintenance and production of 10 kg milk a day. The ratio between homegrown and purchased feed-stuffs in the rations varies between the maximum amount of homegrown roughage governed by intake capacity and 100 % purchased feed throughout the year. A proper concentrate mixture meets the feeding requirements of that part of milk production which exceeds 10 kg per cow per day. Rations with similar proportions between homegrown and purchased feeds as have been designed for milking cows have also been adopted for young cattle of different ages.Labour
The labour required by the various activities is closely associated with available equipment. In the first place there are four optional sets of implements and machinery which vary chiefly in hay and silage making capacity (table 8). Furthermore calculations have been made in which it is assumed that milking can be speeded up by adjustment of the four sets of implements and machinery adopted previously. Also a set of implements and machinery is considered by which it is made possible to employ labour-saving methods for milking and hay and silage making.
It is assumed that no casual labour can be attracted, but a contractor can be hired for several activities.Research methods
The study was carried out linear programming being used. In the first set of calculations the grassland area was treated as a variable resource restriction. Later on the normal type of linear programming was used. In this case the grassland area was fixed consecutively to several sizes.
In the basic computational tables the processes were disaggregated to a high extent to obtain better information as compared with a more aggregated design.
Basic computational table no. 1 is given in appendix no. 1a. Appendix no. 3 contains the modified part of this model concerning the situation in which herd composition may vary.
III. THE RESULTS FOR A CONSTANT HERD COMPOSITION
In this chapter the relation is studied between available grassland area, farm plan and financial outcome with a constant herd composition.No grassland available
As appears from the results a complete separation between milk and roughage production is not advantageous in the assumed conditions. In dairy farms without grassland the gross returns outweigh the variable costs only if higher yields per cow are achieved than was adopted initially (III.3.1), or if marginal building costs per animal are zero (III.3.2). But even in these situations the labour income' of dairy farms without grassland would turn out to be strongly negative.
Therefore dry lot farming is not taken into consideration. So if grassland is available, a linear relation will initially exist between farm area and production rate, and this will not cease until labour becomes restrictive.Grassland area as the only restrictive factor
In the plans applying to the initially adopted conditions (III.2) about 2.30 milking cows with their young replacement stock are kept per ha of grassland till labour becomes restrictive. This stocking density can safely be considered high, even though it could have been raised still further. The high stocking density per ha of grassland requires a heavy nitrogen fertilization up to the maximum level for these calculations (250 kg pure fertilizer-N), and large quantities of purchased feed. Throughout the grazing period all the grass is browsed by the cattle, except in the period of the quickest growth. Only in this period (May and June) a cutting of forage, amounting to forty percent of the grassland area, is harvested. Then almost all the required feed for the winter season must be bought. In this case feed costs amount to about f 780 per milking cow with young stock.
The employment of labour-saving methods (III.3.3) does not affect this farm organization. If gross returns minus variable costs per cow rise by more than f 56, or if feed prices fall by more than five percent as compared with the initially adopted price level, an even higher stocking density is profitable till labour becomes restrictive. Then 3.05 milking cows with young cattle are kept per ha of grassland. Sufficient grass for grazing in this situation is only available in May and June, the months with the most rapid grass growth. The cattle receive supplementary feed during the other summer months. So all the feed required in winter must be bought. The feed costs are nearly f 1000 per cow. The raised margin per cow which results in the adaptation of the optimum farm plan, may arise from a higher milk yield per cow, a higher price level of the milk or a higher amount of net livestock sales per cow (III.3.1) or from the availability of sufficient housing space (III.3.2). In these two paragraphs the margin per cow is supposed to be f 130 higher than in III.2.
The point at which labour force becomes restrictive is not situated at one and the same holding area in the respective calculations. In the calculations applying to the initially adopted starting points this area amounts to well over 18 ha. As a result of the higher stocking density in those calculations the labour force in III.3.1 and III.3.2 becomes restrictive at well over 14.50 ha. In the calculations of III.3.3 this point lies at about 21 ha owing to the larger working capacity.Grassland area and labour force as restrictive factors
On the whole, herd size is not increased after labour has become restrictive. At increasing grassland area homegrown forages are substituted for purchased feed and grassland for fertilizer nitrogen. The increased labour requirement is met by extending the equipment for hay and silage making and by hiring a contractor increasingly more. The income improvement caused by these changes in farm organization, diminishes per additional ha of grassland for three reasons.
In the first place the substitution of grassland for nitrogen becomes less favourable as the nitrogen dressing decreases due to the non-linear input/output relationships. In the second place the substitution of homegrown forages for purchased feed becomes less profitable as the amount of expensive components of bought feed per cow diminishes. In the third place at increasing grassland area a contractor must be commissioned to do more expensive activities. From the marginal values of grass and forages it can be concluded that it is far more important to reduce the pasture losses for a small area per worker than for large holding areas. On the contrary reduction of losses in grass conservation is nearly as important for large holding areas as for small ones, as appears from the marginal value of silage.
For small holding areas it turns out to be advantageous to buy part of the required hay or even all of it. Farms with optimum or larger holding areas may profitably sell some hay. However the extent of this hay trade between grassland farms with different holding areas is modest. The reasons for this small extent of hay trade are discussed in III.2. Price sensitivity of hay purchase and hay sale is studied in the same paragraph.
The above mentioned basic pattern of the organizational changes, taking place at increasing holding area after labour has become restrictive, is similar in all calculations. With this basic pattern the differences in starting points, however, lead to farm plans which vary more or less at the same holding area.
In III.3.1 and III.3.2 the starting-point differs from III.2 because stocking density is higher than in III.2 as long as labour is not restrictive. Consequently the point at which labour requirement equals labour supply is reached at a holding area which is smaller than that in III.2. In the beginning herd size is considerably larger than in the calculations of III.2. The magnitude of this difference fluctuates rather much, chiefly because the choice of a larger set of implements and machinery is not made at the same holding area as in III.2. Yet the difference referred to decreases clearly and at well over 38 ha it is reduced to zero. Because of the higher stocking density as compared with III.2 more feed is required below the point of 38 ha. Nitrogen application however, is about the same as in III.2. Therefore less hay and silage is made at the same holding area because of the larger amount of grass required for grazing. So rather much feed has to be bought per cow in excess of that in III.2 below the point of 38 ha.
If labour-saving methods are applied in milking (III.3.3.1) stocking density is the same as that in the original calculated plans until labour becomes restrictive. In this case labour is not restrictive below 21 ha. From about 19 ha the herd is larger than that in the plans of III.2 and this difference remains about constant at increasing farm area, in contrast to III.3.1. and III.3.2. Besides, the extra feed requirements for the herd larger than that in III.2 are met in a way different from III.3.1 and III.3.2, for the nitrogen fertilization at the same holding area is significantly higher than in the original plans of III.2. Consequently more hay and silage can be made from 25 ha, so that the amount of feed to be bought per cow is only slightly more than that in the plans of III.2. Labour-saving methods for milking do not only result in a larger herd and in a higher hay and silage production at the same holding area but also entail lower costs for contract work in grassland production.
The purchase of an extensive set of implements and machinery, enabling reduction of labour requirements both for milking and hay and silage making, is not advantageous in view of the adopted technical and economics relations (III.3.3.2).Optimum farm area
In all calculations the optimum farm area is reached at the stage when nitrogen fertilization per ha of grassland and the amount of bought feed per cow have been reduced to their minimum levels. In the calculations of III.2 through III.3.2 the optimum farm area does not vary much. It always ranges between 37 and 40 ha of grassland for exploitation of land in tenancy. A difference of some hectares to the optimum farm area only slightly affects the financial results, since marginal cost and marginal value of grassland only slightly diverge near the optimum. In the calculations of III.2,3.3, where labour-saving methods are adopted, the optimum grassland area amounts to about 45 ha being significantly higher than in the preceding calculations. If the grassland is owned by the farmer, the optimum farm area is little affected or not at all by the higher land costs (f 220 per ha of owned grassland in stead of f 160 for tenancy).Financial results
In III.2, in which paragraph building costs were based on replacement value and land costs initially applied to tenancy, the labour income of the farmer' is rather low even at the optimum farm area. It amounts to f 10160. If the farmer owns the grassland (III.2.3.3) net revenue is even negative. In case of higher returns per cow, labour income is distinctly higher (III.3. 1). This is largely effected directly by the higher margin per cow. The rest of the income rise, by adapting the farm organization of III.2, is rather small, especially at larger farm areas. The same holds good if farm buildings are supposed to be unrestrictive (III.3.2). The maximum labour income which is attainable can also be raised by applying labour-saving methods for milking (III.3.3.1). If a higher margin per cow (as in III.3.1) and lower labour requirements (as in III.3.3.1) are combined, the labour income of the farmer will amount to about f 20000 a year at an optimum farm area of about 45 ha. If casual labour is available, all calculations indicate that this could be profitably employed by farms with relative large areas, especially in the months when the most hay and silage is made.Results of other studies
It was not possible to compare the results with those of other studies in all respects. The first reason is that not all alternatives are considered in most studies. The second reason is the difference in the planned situations. In addition to differences in natural conditions this concerns especially differences which are connected with the developments in the fields of mechanization, working methods and prices. The differences in farm organization and financial results found to exist between our results and those of other studies, could be attributed to the factors mentioned above.
IV. THE RESULTS FOR A VARIABLE HERD COMPOSITION
In addition to the factors studied in chapter III, chapter IV includes the variation of herd composition. In this way it was possible to determine how far it is advantageous to effect a differentiation between dairy-grassland farms with diverging farm area with regard to the categories of cattle kept.
In this scope calculations were made in IV.2 in which the average prices of cattle markets were used. For every farm area studied the herd calculated in this way is exclusively composed of milking cows. This is certainly an unlikely situation. Even farms with a relatively large grassland area per man can only raise heifers profitably if the price would rise considerably.
The assumption seemed to be justified that the average quality of the animals bought through the cattle market will be lower than the average level of young stock. For this reason the average market prices for pregnant heifers and for six- month-old calves have been left out in the further calculations (IV.3). It was studied in that paragraph whether a price ratio can be found by raising the prices for the categories referred to, so that a differentiation sets inbetween dairy-grass-land farms with diverging grassland area per man as regards the categories of cattle kept. As appears from the calculations there is a price level of pregnant heifers at which it is advantageous for farms with relatively little land per man to buy heifers, whereas it is advantageous for farms with relatively much land per man to raise the heifers to be sold. This is still more profitable in spite of the costs charged for intermediate trade and transport. The course of the level of nitrogen fertilization and of making hay and silage presents a picture which is very similar to that of the calculations with a constant herd composition and the same starting points in other respects. Milking cows occupy an important place in all farm plans. A plan without milking cows shows much more unfavourable results in case of farms with relatively little grassland per man. On the other hand a farm organization without milking cows can almost compete with the optimum farm plan including milking cows in case of farms with relatively much grassland per man.
As appears from the comparison of the financial results for a constant and a variable herd composition respectively and the same starting-points in other respects, the advantage of the differentiation in the dairy process is not large at the costs charged for intermediate trade and transport. These costs however can be lowered if a regular and direct trade is effected between farms with diverging grassland area per man. In that case the advantage of such a differentiation would become important, especially for the farms with up to 20-25 ha of grassland and for those with more than 35-40 ha for which it would amount to f 1250-f 1500 per farm. It% that situation farms with relatively much land per man can obtain the best results by a farm plan without milking cows. On the other hand farms with relatively little land per man will keep many milking cows and buy the required heifers. The dairy cattle husbandry of farms with an intermediate grassland area will be little differentiated. In the optimum plans of these farms the herd is composed of dairy cows and young stock which is raised to replace the own dairy herd.
Based on data from the Herd Books and on those of state-managed farms it has been studied how far the calculated prices agree with the real prices of heifers with about the same qualities. It appeared that the latter price level is indeed higher than the average market price, but still about f' 100 per heifer lower than the price calculated in this study. As long as this continues it will be an extra stimulus for farms with little land per man to abandon replacement breeding.
The results of this chapter have been compared with those of other studies. Allowing for differences in research methods and starting-points, the publications discussed presented about the same picture for herd composition at increasing grassland area per man as the plans of this chapter.
V. FINAL REMARKS
In the first part of this chapter the results of the study based on linear programming models are compared with the actual situation. In view of the difficulties in isolating the significance of a certain factor from farm records (e.g. recorded in farm accounts) this comparison between the theoretical farm model and the practical situation had to be rough. The picture of the practical situation was based mainly on farm comparison investigations of the Agricultural Economics Research Institute. Backgrounded by these farm comparisons which have been made chiefly by factor analysis, the farm accounts of some more recent financial years were also taken into account. This indicated that dairy-grassland farms with relatively little land per man may improve their income in two ways at a given holding area and a given labour force. In the first place stocking density might profitably be raised, even if the results of the calculations should be used with care in view of some doubt regarding to the level of net S.E. production of the grassland at a high stocking density and nitrogen fertilization. This conclusion also holds good if farm buildings must be enlarged owing to the expansion of the herd. In the second place farms with little land per man may profitably switch herd composition to more milking cows. Good prospects for this differentiation in the dairy cattle husbandry exist particularly if arrangements could be made with farmers raising young stock. Such agreements would reduce the costs of intermediate trade and transport and particularly lead to more certainty about heifer quality. In that case farms with relatively much grassland per man could profitably extend young stock raising at the expense of the dairy herd. Apart from the possibilities of improving the income of farms with little land per man at a given holding area and a given labour force, additional grassland has a high marginal value for these farms. So specific priority should be given to improving the ratio between grassland and labour force on these farms.
A renewed critical discussion deals with the accuracy of the technical and economic relations which are used in the linear programming models and with the extent of simplification of the models as compared with reality. In this context it is discussed in which directions further technical research is desirable and how the farm models could be extended to enhance adequacy of farm planning to reality.
The second part of this chapter describes how the investigations should be adapted to future practice, in view of the continuous developments in the field of working methods and price ratios. A systematic co-operation between technical and economic research workers has been advocated, to be able to incorporate the various most recent developments in the programming models. The structure of these models might be largely similar to those used in this study. In the first place comparitive-static linear programming might be used just like in this study. Afterwards application of dynamic linear programming would be recommendable to refine the results obtained. These investigations for the sake of farm management, may also be useful for agricultural policy and technical research policy. As far as the latter purpose is concerned, more information can be obtained by linear programming calculations, including still unrealized technological progress in different fields. These calculations, quantifying the significance of technological progress in the respective fields, could be an aid in indicating the most promising direction for technical research.