by Dave Clark, scientist

The past decade has seen some heated debate concerning the expansion and intensification of New Zealand’s dairy industry. Two aspects of this multi-faceted debate have been consistently present. The economic returns available to farmers from intensification, expressed as feed input per farm, have been subject to different forms of analysis by farmers, rural professionals and economists. A wider debate has concerned the effect of expansion and intensification on the environment. Key issues have been increased water loadings of nitrogen, phosphorus, sediment and bacteria, and increased atmospheric loadings from greenhouse gases (methane and nitrous oxide). The size and complexity of these issues, together with the real and perceived threats to agricultural viability and the quality of our environment, have meant that it is difficult to arrive at an analytical position that can be generally agreed as a starting point for behavioural change.

The two aspects are closely linked. If greater profits are consistently available from an increase of say 25-50% in feed inputs above a farm’s baseline pasture production, then major technological advances and/or regulatory changes will be needed to meet Government and Regional Council legal obligations to the environment. If profits can be maintained and economic risks reduced by, say, keeping feed inputs to 10-20% of a farm’s baseline pasture production then there will be a less urgent requirement for technological breakthroughs or stringent regulations.

This paper will concentrate on the farm-based issues around feed inputs, specifically – what are the likely economic and farm management consequences of reducing feed inputs (de-intensification)? Note that it is possible to de-intensify in other ways, for example by adopting full or partial lactation once daily milking.

Before tackling the issue of de-intensification it is useful to understand how different commentators or analysts have approached the issue of feed inputs into farm systems. A key paper by Hedley et al. (2006) modelled Systems 1-5 for a Waikato farm with baseline annual pasture production of 19 t DM per ha. They concluded that if each system was run under ‘best practice’ conditions the operating profit per ha and return on assets (RoA) (%) increased progressively as feed input per ha increased. They pointed out that System 5 could achieve more consistent production because of less reliance on pasture. However, they also noted how sensitive the higher feed input systems were to an increase in supplement price, and/or a decrease in milk payout. Recently, Ma et al. (2018) cautioned that this approach, while it “…may provide an indication of relative-farm performance, these differences cannot be interpreted as the net impacts of adopting one farm system relative to another, because there may be other factors (e.g. milking area and irrigation intensity) influencing the outcomes…” However, Hedley et al. (2006) were careful not to have such factors confounding their modelled analyses.

The criticism identified by Ma et al. (2018) is relevant to some DairyBase® analyses because the simple effect of feed input (i.e. System classification) on operating profit cannot be separated easily from other factors (e.g. System 5 farms might be clustered around large, irrigated land whereas System 1 farms might be clustered around smaller non-irrigated farms. When Ma et al. (2018) statistically corrected for such differences they could find no evidence that high input systems performed better financially than lower input ones.

This is such an important issue from both an economic and environmental perspective that it deserves further critical, independent analysis. For this paper I will accept that there is enough evidence to model and discuss the effects of moderate de-intensification on some Waikato dairy farms.

Some data relevant to the expansion and intensification of the New Zealand and Waikato dairy industries are summarised in Table 1 and Table 2 respectively. Feed demand over the past 28 years has increased by 300 and 140% for New Zealand and the Waikato, respectively. In that time the Waikato region has lost 40,000 ha to other land uses, while the number of cows milked has increased by 40,000. Stocking rate has increased slightly from 2.8 to 2.95 cows/ha. Annual milksolids has increased from 264 to 358 kg per cow leading to an increased dairy feed demand for the region from 3.9 million t DM to 5.45 million t DM. There is little evidence that baseline pasture DM yield has increased so that the extra feed has come from an increased use of N fertiliser, maize silage, and particularly palm kernel expeller (PKE).

Table 1. Expansion and intensification of New Zealand dairy industry (1990-2018).

¹ 1990 data estimated from milksolids production

Sources: MPI (2016), DairyNZ (2018), LIC (1993a), LIC (1993b)

Table 2. Changes in the Waikato dairy industry (1990-2018).

¹1990 data estimated from milksolids production

Sources: DairyNZ (2018), LIC (1993a), LIC (1993b)

• McMeekan found that stocking rate was the KEY determinant of pasture utilisation in all- grass dairy farming. He was careful to stress that there was an optimum stocking rate for maximum milk production at any given level of annual pasture yield. Rates above optimal would initially lead to higher pasture utilisation but lower milk production, as energy required for cow maintenance decreased that available for lactation. Eventually over-grazed pastures led to lower pasture yield. Rates below optimal led to high milk production per cow, but not sufficiently high to overcome the effect of fewer cows, so milk production per ha declined and pasture was wasted.
• McMeekan’s analysis assumed that because fixed costs were more important than variable costs in 1950’s New Zealand dairy farming, maximum profit per ha would follow from stocking rates that maximised milk production per ha.
• But as variable costs increase, particularly those associated with cow numbers, then the optimal stocking for maximum profit will be LOWER than that for maximum production, (Wright & Pringle, 1983).
• Any increase of feed SUPPLY into a system will change the optimal stocking rate in a complex way.
• A key factor in this complexity is SUBSTITUTION. When a herd is offered 1.1 t DM of palm kernel expeller in addition to its normal ration of pasture – about 1 t of PKE is consumed (10% wastage) but pasture intake by the herd decreases. If pasture intake by the herd decreases by 0.5 t DM then the substitution rate is 0.5 – in other words 0.5 t of pasture DM has been potentially wasted.
• Under increased feed supply the optimal stocking rate for profit will likely increase determined by: substitution rate, cost of the extra feed, milk payout, and many other costs associated with the extra cows and the method of delivery of this extra feed.
• From 1990, dairy systems diverged in terms of breed (mature live weight) and feed inputs such that stocking rate expressed simply as cow number per ha became a less accurate way of expressing feed demand in relation to feed supply. The concept of COMPARATIVE STOCKING RATE (cow LW (kg/ha)/ total feed offered (t DM/ha)) was developed.

The main purpose of this paper is to identify the opportunities and challenges associated with a decision to de-intensify a farming operation, especially in relation to feed supply and feed demand (stocking rate). It will concentrate on systems 2-3, i.e. those where 5-20% of total feed supply is imported. It will re-interpret several major Waikato farmlet experiments in order to demonstrate that questions about feed supply and demand for commercial farms need to be carefully framed if we are to reach the correct conclusions about optimal stocking rates.

The opportunities and challenges of low input, all grass farming are well-described in Table 3. There are reduced operating expenses and also fewer infrastructure costs, together with simpler feed management. However, there is increased risk from climate variation and less opportunity for taking advantage of high milk prices or low supplement costs.

Table 3. Summary of advantages and disadvantages of low input, all grass milking system (Southland farmer) (Fleck & Fleck, 2014).

De-intensification can be accomplished in many ways but this paper will mainly examine the effect of reducing feed inputs, or reducing stocking rate or a combination of these. An important point is that any ONE stocking rate will not be optimal for all possible combinations of milk payout: feed cost ratios. We will use examples from: Waikato farm systems experiments, dairy system models and economic surveys. First, we will examine the effect of removing the least expensive supplement, N-boosted pasture, from a dairy system.

Nitrogen fertiliser is usually the least expensive way to increase feed supply on a non-irrigated farm. Therefore, any decision to de-intensify by reducing N fertiliser will present a serious challenge. At DairyNZ’s Scott Farm, Newstead, two farmlets, one receiving 181 kg N/ha nitrogen (Control) and the other receiving no nitrogen (No-N) per year, were established on 1 June 2001 and continued until May 2011. Although some management changes occurred during this time these were not sufficient to alter the results from the main comparison. Table 4 summarises the treatments, full details are in Glassey et al. (2013).

Annual rainfall over the 10-year period varied from 915 mm (2007/08) to 1387 mm (2010/11). Annual pasture DM yield from No-N farmlet varied from 20 t DM/ha (2002/03) to 12 t DM/ha (2004/05). The range of climate conditions over the 10 years means that the results are a robust indicator of what can be expected if nitrogen fertiliser is completely removed from the farm system. Of course, this experiment was a very severe test of de-intensification. Commercial farmers with good pasture monitoring could reduce stocking rates but still use some N fertiliser, as dictated by their feed budgets.

Table 4. The design, physical and financial results averaged over 10 years for Control farmlet receiving 181 kg N per ha per and No-N farmlet receiving no nitrogen fertiliser for 10 years. Full details are in Glassey et al. (2013); operating profits per ha are slightly different because milk prices have been rounded.

¹ 2.40 milkers + 0.16 replacement heifers.

The stocking rate of milking cows was reduced by 21.5% (total stocking rate was reduced by 16.3% if replacements on the No-N farmlet are counted). This led to a 17% decrease in milksolids per ha and a 25% decrease in operating expenses per ha for the No-N farmlet. At milk prices below $5.10 per kg MS the No-N farmlet option was more profitable than the Control, but above this figure use of N fertiliser became profitable. From this important dataset we can identify some opportunities and challenges associated with quite radical de-intensification.

• Major cost savings in wages, rearing replacements and N fertiliser.
• Increased profit in lower payout years.
• Improved pasture quality through higher white clover content in mid-late lactation.
• Farm nitrogen surplus decreased by the difference between N fertiliser reduction and N fixation increased by more white clover.

• Capturing the cost savings, especially in wages.
• Change of management ‘style’ – having to accept an early cull and dry off in summer droughts.
• More vulnerable to climatic risk, especially a cold, wet spring.
• Profit opportunity foregone in high payout years.

But of course there is no need to reduce stocking rate by over 20% and stop N fertiliser use altogether – there are plenty of intermediate options that might fit your farm much better. This experiment, however, showed that even major de-intensification did not necessarily lead to financial gloom if we look at the volatility of milk prices over the past two decades.

We will now turn to a re-assessment of a DairyNZ farmlet stocking rate experiment that has been used to justify the need for high stocking rates in order to optimise operating profit per ha. The experiment was conducted at Ruakura No. 2 Dairy from 1999-2001. Full details are provided by Macdonald et al. (2008; 2011). Brief details are given in Table 5.

Table 5. Ruakura stocking rate trial – design, physical and financial results at different milksolids payouts. All treatments received 200 kg N per ha per year. (Macdonald et al. 2008; 2011).

This experiment concluded that financial performance was optimised at 3.3 cows per ha or a CSR of about 77 kg of live weight per t feed DM available. This conclusion held at milk prices from $4.30 -$6.30 per kg MS. As seen in Table 5 once the milksolids price increased to $7.50 per kg MS stocking rates above 3.3 cows per ha became more profitable.

The authors followed the standard scientific practice of varying only the parameter of interest – stocking rate – and leaving all other inputs the same. This meant that all treatments were scheduled to receive 200 kg N per ha per year (actual = 181). But even with zero N fertiliser input the 2.2 cows per ha had surplus pasture, adding N simply meant more topping and making silage that was not required. Conversely, even with N fertiliser the cows at 4.3 cows per ha had insufficient feed for much of the year.

The experiment answered the question – how does stocking rate effect milksolids production and financial performance when feed inputs are held constant? But no rational farmer would spend money to grow extra feed when a feed budget showed it would never be needed. So a better question would be “What stocking rate optimises financial performance when a farmer is allowed to vary the input of the cheapest feed source?”.

Given this question, it is possible to reanalyse the Ruakura experiment in a simple way to see if different conclusions are reached. To do this N fertiliser inputs were varied from zero at 2.2 cows per ha to 200 kg N per ha per year at 3.1 cows per ha and 400 kg N per ha per year at 4.3 cows per ha – based on the approximate amounts of pasture required at each stocking rate (Table 6). The extra feed at the lowest stocking rate was also sufficient to graze all replacement stock on the milking platform. Financial results were recalculated based on decreased or increased spending on N fertiliser and saved grazing costs; and extra milksolids production at the higher stocking rates. Optimal stocking rates at different payouts could then be compared with the original experiment.

Decreasing the stocking rate by 13% from 3.1 to 2.7 cows per ha increases operating profit by $56/ha at $4.50 per kg MS; decreases profit by $54 per ha at $6.00 per kg MS and $168 per ha at $7.50 per kg MS (Table 6 and Figure 1). The decreases in profit at the two higher milk prices are hardly catastrophic. At $5.17 per kg MS, profit is the same for ALL stocking rates. Obviously at that milk price it would be logical to choose the lowest stocking rate!

At a $6.00 per kg MS price with zero N fertiliser use and grazing replacements at home, 2.2 cows per ha increases profit by $398 per ha compared with the original experiment; and at 2.7 cows per ha profit is increased by $177 per ha with zero N fertiliser use. In the original experiment, profitability at lower stocking rates was compromised by spending on N fertiliser that was not needed; and at higher stocking rates by not spending enough on N fertiliser.

The original experiment concluded that the optimal stocking rate was 3.3 cows per ha (CSR = 77) for any milk price from $4.30- $6.30 per kg MS. In contrast, the revised analysis shows that at less than $5.32 per kg MS the stocking rate could decrease from 3.3 to 2.7 cows per ha without any decrease in profit per ha. It is also important that the Ruakura experiment was conducted from 1999-2001, in the 20 years since New Zealand dairy cows have been bred for more efficient feed utilisation so optimal stocking rates will have decreased for any given amount of feed available per ha. Also of importance is the fact that the inflation-adjusted price of $5.32 in 2000 is now $7.94.

Table 6. Revised Ruakura stocking rate trial – design, physical and financial results at different milksolids payouts. Treatments receive N fertiliser at rates appropriate for the different stocking rates (Macdonald et al. 2008; 2011).

SR = 2.2 cows/ha – deduct $222/ha for grazing at home and $175/ha for zero N = $397/ha saving.  Expenses=2732-397= $2335/ha.

SR= 2.7 cows/ha – deduct $175/ha for zero N = $175/ha saving.  Expenses=3090-175= $2915/ha.

SR = 3.1 cows/ha – No deductions.

SR = 3.7 cows/ha – add $105/ha to expenses for an extra 100 kg N/ha.  Expenses =3633+105 = $3738/ha.  Assume marginal N response above 200 kg N/ha is 8:1 gives 800 kg DM/ha, 85% utilized at 15 kg DM/kg MS gives an extra 45 kg MS/ha.

Figure 1. The effect of stocking rate on operating profit per ha at milk prices of $4.50, $6.00 and $7.50 per kg milksolids for re-analysed Ruakura experiment (from Macdonald et al. 2008, 2011; experimental data and prices from 1999-2001).
SR= 4.3 cows/ha – add $210/ha to expenses for an extra 200 kg N/ha.  Expenses = 3993 + 210= $4203/ha.  Assume marginal N response above 200 kg N/ha is 6:1 gives 1200 kg DM/ha, 85% utilized at 15 kg DM/kg MS gives an extra 68 kg MS/ha.

We will now examine a more recent Waikato experiment that sought to use a suite of management techniques to reduce nitrate leaching by about 50% and maintain profitability. Here we will concentrate on the financial aspects of this experiment and again do a revised analysis to look at whether some different management might have improved the financial outcome. The experiment at DairyNZ, Scott Farm, Newstead was part of a national Pastoral 21 series designed to improve environmental performance without compromising profitability. Full details of design and physical and financial performance are given in Clark et al. (2019) and some preliminary environmental results in Selbie et al. (2017). Table 7 summarises the design, physical and financial results (annual average from 2008-2013) from the Waikato Pastoral 21 (P21) experiment. A CURRENT farmlet was designed using average stocking rate, N fertiliser input, and cow genetics for the Waikato. On a FUTURE farmlet stocking rate was reduced from 3.2 to 2.6 cows per ha (-19%), annual N fertiliser from 135 to 60 kg N per ha and a standoff pad used to capture effluent when cows where removed from pasture to reduce soil urinary load at critical times. To counter the expected milksolids loss higher BW cows were used on the FUTURE farmlet and maize grain offered up to 300 kg per cow per year.

Table 7. Pastoral 21 experiment comparing CURRENT and FUTURE farmlets –design, physical and financial results (from Clark et al. 2019).

In summary, the higher stocked CURRENT farmlet produced 50 kg MS/ha more than the FUTURE farmlet, with extra expenses of $71 per ha, providing an increased annual operating profit of $279 per ha averaged over the five years at an average payout of $6.08 per kg MS (range of $3.92- $7.69). Although profit from the FUTURE farmlet could not match that from the CURRENT; annual N leaching was reduced by 43% from 54 to 31 kg N per ha. Factors that reduced the profitability of the FUTURE farmlet included: the depreciation charge for the standoff, the non-return of some N associated with standoff wood chips, and the cost of grain. The reduction in stocking rate by 19% also created a challenge in utilising pasture efficiently.

We will reanalyse the P21 experiment, with the benefit of hindsight, to see if it would be possible to de-intensify, reduce N leaching but achieve similar levels of profit to the CURRENT farmlet (Table 8). A simple reanalysis involves a FUTURE stocking rate of 2.9 cows per ha (10% lower than CURRENT) and using the same amount of supplement per cow but swapping maize grain for PKE. Labour and stock expenses per ha were increased to reflect the increased stocking rate; feed expenses decreased because of cheaper supplement; working and overhead expenses were unchanged. Operating expenses were reduced by $87 per ha compared with the original FUTURE farmlet. Milk income increased because an extra 26 kg MS per ha was produced, and livestock income increased by $66 per ha (compare Table 8 vs Table 7). This conservative reanalysis showed that profit increased by $216 per ha ($6.00 per kg MS) from the original FUTURE farmlet. The difference in operating profit between CURRENT and FUTURE-Revised for milk prices of $4.50, $6.00 and $7.50 per kg MS favours the CURRENT by only $14, $50 and $84 per ha respectively (Table 8).

Reducing stocking rate by 10% from 3.2 to 2.9 cows per ha when accompanied by changes in cow genetics, decreased N fertiliser, standoff management, and the use of cheap supplements, can achieve very similar levels of operating profit regardless of milk price up to $7.50 per kg MS AND reduce annual N leaching losses by about 20 kg N per ha (37%).

This five-year experiment covered a period of important variations in climatic and economic conditions. The original experiment showed that the CURRENT system was robust under summer rainfall varying from 427 mm (2011-12) to 165 mm (2012-13); and milk prices per kg MS varying from $7.69 (2013-14) to $3.92 (2015-16). However, the FUTURE system was more susceptible to very low summer or autumn rainfall (2012-13 and 2015-16, respectively). In fact 52% of the five-year total difference in milksolids yield was due to reduced production on the FUTURE farmlet in 2015-16. The decision rules used meant that either expensive grain was used initially to maintain lactation followed by culling and drying off to protect body condition. The FUTURE farmlet white clover levels did not increase as much as expected so that summer-autumn pasture was not of sufficient quality to support high per cow yields. Also not enough high quality pasture silage was made to support autumn lactation.

The economic analysis (Clark et al. 2019) assumed that a 19% reduction in stocking rate on a 100 ha farm originally milking 320 cows would only save $5300 per year in labour costs if 60 fewer cows were milked, i.e. about 31,000 fewer cow-milkings per year. Obviously on a range of commercial farms this figure might vary from zero to $20,000 per year. Another way to approach this is to ask “What would you pay to be absent from milking one day every five through the season?”.

The anticipated problem of managing excess spring pasture growth at 2.6 cows per ha on high quality Waikato dairy land did not arise. This is likely a result of weekly pasture monitoring and the use of decision rules that helped identify genuine pasture surpluses for conservation. It is also likely that modern cows are capable of high DM intakes if they are provided with high pasture allowances. But of some concern is the high response rates to N fertiliser of 15-20 kg DM per kg N applied. This suggests that Waikato soils are more responsive than formerly, perhaps because of the lack of biological N fixation by white clover over the past 20 years resulting from clover root weevil damage. If low N rates are to be used greater efforts must be made to increase the annual clover contribution from 5-10% currently to 20+%.

Table 8. Reanalysis of Pastoral 21 experiment comparing current and future farmlets –design, physical and financial results (based on Clark et al. 2019, but with stocking rate decreased by only 10%).

Ho et al. (2013) argued that “… the intensity of the production system that a farmer chooses to run relates to what kind of risk they prefer to be exposed to, their goals and their available resources.” Therefore any decision to transition from Systems 4/5 should be: consistent with risk assessment, goals and resources, it should be well-planned, and not done when current or immediate forecasts for milk price are low. This allows surplus cows to be sold more profitably. The psychological impact of lower per cow production because of lower feed inputs has to be accepted. Over time there may need to be a shift in cow genetics. Transition gives an opportunity to reduce unprofitable feed inputs because Neal & Roche (2019) report that “On average, for every $1 spent on imported feed, total costs increased by $1.66…for the Waikato…” This analysis was based on 12 years of Waikato DairyBase data. These authors concluded “…that maximising pasture harvested, and minimising reliance on supplementary feed, and effective cost control (minimising expenditure) are the key factors that lead to profitable businesses which are also resilient to the low milk prices that occur in a volatile market.” Such a transition also positions a farm better to handle future environmental regulations.

An argument against transition is that the opportunity to take advantage of higher milk price and/or lower feed inputs costs. However, Neal & Roche (2019) found that from 2005-06 to 2016-17 there were only three years when the milk price was above $7.50 per kg MS that DairyBase farms were able to generate higher profit from using more imported feed. It is doubtful that this could offset the years of medium –low milk prices. Further, Shadbolt et al. (2017) analysed nine years of New Zealand dairy farm data and found that even when operating profit per ha increased, operating return on dairy assets showed no difference between systems. The latter measure acknowledges that when the stocking rate increases extra co-operative shares may be required, and a significant investment in assets, such as feed pads, herd homes, machinery, and sometimes housing for labour. However, although shares may be traded when systems become less intense, some assets may be left ‘stranded’ with little chance of recouping the investment.

The Climate Change Response (Zero Carbon) Amendment Bill (Zero Carbon Amendment Bill) sets a target to reduce biogenic methane emissions by 24-47 per cent below 2017 levels by 2050, including a 10% reduction by 2030 (Ministry for the Environment, 2019). The analyses in this paper show de-intensifying dairy systems can reduce DM intake per ha by about 5-15% and therefore methane emissions should be reduced by a similar amount. The 2030 target is economically feasible.

The P21 Waikato research experiment shows that commercial farms could reduce their stocking rates by up to 10% over a short period without compromising profitability and with substantial environmental benefits. Those with herds of lower genetic merit and more susceptible than average to summer-autumn droughts should take a more cautious approach. Like any major management change a full preliminary economic analysis should be done and advice sought, especially from those farmers who have already successfully made this transition. Those with excellent genetic merit herds, standoff facilities, and a history of good pasture management could envisage replicating the results obtained from the 19% reduction obtained at Scott Farm from the FUTURE farmlet.

However, if major reductions in greenhouse gas emissions (30-50%) were required from dairy farms over the next 30 years then farm systems research faces major challenges in supporting commercial farmers. The story so far – without any supporting analysis – is that traditional management of soils, pasture and animals will find this an impossible task. Therefore, we must rely on science in the form of methane inhibitors, methane vaccines and nitrification inhibitors. But these are unlikely to be available before 2030. Some believe that reducing cow numbers would lead to financial ruin. Even if we have misgivings about this option we should at least understand what might be required to develop a system where stocking rates are decreased by 30+%.

The major reasons for seeking lower stocking rates are to avoid high DM intakes per ha (> 18 t DM per ha per year) and to reduce the number of replacement stock required. These stock spend two years emitting methane and depositing urine without producing milk. Stocking rates of 2 cows per ha with cows weighing 550 kg and yielding 1 kg MS/kg LW (550 kg MS per year) would produce 1100 kg MS per ha, from about 12.6 t DM/ha eaten (assuming 12 MJ ME/kg DM). Currently, the Waikato average stocking rate is 3 cows per ha, yielding 360 kg MS per cow or 1080 kg MS per ha, from about 15 t DM/ha eaten (assuming 11 MJ ME/ kg DM). Given that methane output is very closely related to DM intake, the lower stocking rate with 2.4 t DM per ha less eaten should emit about 16% less methane, or about 17% less if the lower replacement rate but with heavier heifers is considered.

To achieve such a target would require pastures with about 10% higher concentration of ME than currently available, but more importantly, pastures (probably ryegrass-white clover) would have to support cow DM intakes of > 18 kg DM per day for 290 days of lactation and > 13 kg DM per day for 75 days when dry. Such levels imply the following: ryegrass breeders should concentrate on increasing the grazing intake potential of grass, and provide a grass that supports an association with white clover that allows 20-30% white clover content in summer-autumn pastures. Pasture managers can assist this goal by keeping N fertiliser inputs below 100 kg N per ha as a long-term average. With these provisos dairy farming could still be very profitable; if the P21 management protocols were used N leaching would be substantially reduced; but methane emissions would still be far short of the 2030-2050 targets.

New Zealand may have to accept genetically modified ryegrasses if such targets are to be reached. Targets for a key dairy region, such as the Waikato, to reduce methane emissions by 24-47% by 2050 would require the development of methane inhibitors and/or vaccines and a change in land use.

• De-intensifying Waikato dairy farms by up to a 10% reduction in stocking rate can be accomplished without major financial losses.
• De-intensification needs to be well-planned and consistent with your financial and social goals, and with your values and resources.
• De-intensification is not risk-free. Pasture monitoring and feed budgeting assume even greater importance. There may be more exposure to climate risk. The gains from a high milk price are harder to capture.
• Conversely, the system is much more robust to a low milk price.
• The pressure to de-intensify increases if input costs increase, the availability of well-trained, committed staff decreases, and if environmental regulations become more stringent.
• De-intensifying offers opportunities to meet likely regulatory standards for nitrate leaching.
• De-intensifying offers opportunities to meet current methane emissions reduction targets for 2030, but not beyond, with present management and technology.

To Debbie McCallum, Karl Buhler, Stephen Canton, John Roche and Chris Glassey for enlightening conversations and access to pre-publication copies of papers.

Clark, D.A., Macdonald, K.A., Glassey, C.B., Roach, C.G., Woodward, S.L., Griffiths, W.M., Neal, M.B. and Shepherd, M.A., 2019. Production and profit of current and future dairy systems using differing nitrogen leaching mitigation methods: the Pastoral 21 experience in Waikato. New Zealand Journal of Agricultural Research 62:1-24.

DairyNZ 2018. Farm Dairy Statistics 2017-18. https://www.dairynz.co.nz/publications/dairy-industry/new-zealand-dairy-statistics-2017-18/ [Accessed 19 May 2019].

Fleck, W. and Fleck, K., 2014.  The advantages and disadvantages of a low input all grass milking system. South Island Dairy Event, 2014. https://side.org.nz/wp-content/uploads/2014/05/2.1-Low-Input-Milking-System.pdf [Accessed 22 April 2019].

Glassey, C.B., Roach, C.G., Lee, J.M. and Clark, D.A., 2013. The impact of farming without nitrogen fertiliser for ten years on pasture yield and composition, milksolids production and profitability; a research farmlet comparison. In Proceedings of the New Zealand Grassland Association 75:71-77.

Hedley, P., Kolver, E., Glassey, C., Thorrold, B.S., Van Bysterveldt, A., Roche, J.R. and Macdonald, K., 2006. Achieving high performance from a range of farm systems. Proceedings of the Dairy3 Conference 4:147-165.

Ho, C.K.M., Newman, M., Dalley, D.E., Little, S. and Wales, W.J., 2013. Performance, return and risk of different dairy systems in Australia and New Zealand. Animal Production Science 53:894-906.

LIC, 1993a. Dairy Statistics 1992/93. Pp. 34.

LIC, 1993b.  1993 Economic survey of factory supply dairy farmers. Pp. 29.

Macdonald, K.A., Beca, D., Penno, J.W, Lancaster, J.A.S., Roche, J.R., 2011. Effect of stocking rate on the economics of pasture based dairy farms. Journal of Dairy Science 94:2581–2586.

Macdonald, K.A., Penno, J.W, Lancaster, J.A.S., Roche, J.R., 2008. Effect of stocking rate on pasture production, milk production and reproduction of dairy cows in Pasture-Based Systems. Journal of Dairy Science 91:2151–2163.

Ma, W., Renwick, A. and Bicknell, K., 2018. Higher intensity, higher profit? Empirical evidence from dairy farming in New Zealand. Journal of Agricultural Economics 69:739-755.

Ministry for the Environment. 2019. Climate Change Response (Zero Carbon) Amendment Bill: Summary. Wellington: Ministry for the Environment. http://www.mfe.govt.nz/sites/default/files/media/Climate%20Change/climate-change-response-zero-carbon-amendment-bill-summary.pdf [Accessed 20 May 2019]

MPI Report, 2016. Feed Use in the NZ Dairy Industry. MPI Technical Paper 2017/53 Prepared for the Ministry for Primary Industries by DairyNZ Farm Economics Group. June 2016. https://www.mpi.govt.nz/dmsdocument/20897/send [Accessed 22 April 2019].

Neal, M. and Roche, J., 2019. Profitable and resilient pasture-based dairy farm businesses in New Zealand. In Animal Production Science. https://www.publish.csiro.au/AN/justaccepted/AN18572 {Accessed 19 May 2019}.

Selbie, D., Shepherd, M., Hedley, M., Macdonald, K., Chapman, D., Lucci, G., Shorten, P., Welten, B., Pirie, M., Roach, C. and Glassey, C., 2017. Following the nitrogen: explaining the reasons for decreased N leaching in the Waikato P21 farmlets. Science and policy: nutrient management challenges for the next generation. Occasional Report, (30).

Shadbolt, N.M., Siddique, M.I., Hammond, N., 2017.  Pastoral dairy farming systems and intensification, challenges in interpretation. 21^st International Farm Management Congress, pp. 16. https://ifmaonline.org/wp-content/uploads/2018/02/17_PR_Shadbolt_etal_w1_p1.pdf [Accessed 19 May 2019]

Wright D.F., Pringle, R.M., 1983. Stocking rate effects in dairying. Proceedings of the New Zealand Society of Animal Production 43: 97-100.