Trends in dairy farm sales and factors influencing dairy land prices in New Zealand (1990-1997).

G P Rauniyar and A E Dooley - College of Sciences, Massey University, Palmerston North, New Zealand,
W J Parker - Dairy and Beef Division, AgResearch (Ruakura), Hamilton, New Zealand

 Dr G P Rauniyar, Senior Lecturer, Institute of Natural Resources, Massey University, Ag/Hort Science Building,
Level 3, Private Bag 11222, Palmerston North, NEW ZEALAND.
Phone: +64 6 350 6182. Fax: +64 6 350 5680.
Email: G.P.Rauniyar@massey.ac.nz

ABSTRACT

Dairy farm sale data (1990-1997) published by Valuation New Zealand are examined in this paper. Trends and interrelationships involving the number and size of dairy farm sales, land prices, milksolids prices and dairy farm productivity are explored on a national and regional basis. Buyer-seller relationships in dairy farm sales are also analyzed. An econometric model, constructed to explain dairy farm land prices (expressed as the logarithm of the net sale price in 1990 dollars: r²=0.78) indicated each 10% increase in farm area would decrease the sale price by 1.5% and that a 10% increase in milksolids would improve the price by 8.7%. In real terms dairy land sale values are projected to increase.

Keywords: dairy farm sales, land prices, farm size, New Zealand, dairy farms.

 1. INTRODUCTION

The outlook for the milksolids price, new technology, the relative profitability of alternative land-use enterprises to dairying, interest rates and the motivation of herd owners all impact on dairy farm land values, and the number and size of herds. The 1990-1997 period was particularly dynamic for the dairy industry with significant growth in national milk production (NZDB, 1997) due to expansion of existing herds, conversions of sheep and beef cattle farms to dairying and improved productivity per cow (LIC, 1996). The signing of the GATT agreement, lower interest rates, ‘bullish’ projections for dairy products and relatively good prices for milksolids from 1992/93 to 1994/95 all contributed to an optimistic outlook for dairy farming (Kilsby et al., 1998). These factors were capitalized into dairy land values. The optimism for dairying began to abate from around 1996 when milksolids prices declined, the need to own shares in dairy companies proportional to the milk supply was foreshadowed for the 1997/98 season and interest rates edged upwards. This paper presents an analysis of dairy farm sales for the 1990-1997 period. Changes in land values by farm type and size are reported.

2. DATA AND METHODOLOGY

Valuation New Zealand (VNZ) six monthly data were obtained on size of property sold, annual milksolids (MS) production, sale price and buyer-seller relationships. The VNZ definition of net sale price of dairy farm implies sale price net of buildings and chattels on the property. Production was recorded as kg milkfat up to 1996: this was converted to milksolids using a conversion factor of 1.74. For each farm sold, MS/ha was calculated by dividing the total production by effective area.

Dairy farm sales for the period 1990 to 1997, a total of 5,542 properties, were included in the initial data set. The VNZ data subset was subjected to trend, bivariate and multivariate analysis using the Excel and SAS packages. Net sales prices were converted to 1990 constant dollars using the Consumer Price Index (CPI). The total dairy farm sale data records included farms of all sizes. The analyses of farm size, sale price and productivity were based only on the data for economic units (n= 3513), defined by VNZ as "a farm unit capable of supporting at least one family". Economic units producing less than or equal to 5,000kg MS/year (n=24) were excluded from the data set. Nine other outliers, comprising farms over 120 ha and producing less than 20,000 kg MS, were excluded as well to give a final data set of 3480 sales.

Buyer-seller relationships, as defined by VNZ, were examined for 1990, 1993 and 1997. Farm buyers were classified as either expanding or not-expanding property owners, while farm sellers were classified as either subdividing or not subdividing. The farm buyer was classified into one of the four categories: existing farmer, new farmer, businessperson and other (Government/Local Authority and other). Similarly, the ownership structure of the buyer was divided into one of the four types: individual, partnership, company and other (co-operative, trust, Government/Local Authority and other).

An economic model was specified to explain dairy farm sale prices. Operational farm size, production performance, location of farm, buyer-seller attributes and time of sale all play important role in determining sale prices. Farm sale prices were hypothesized to be function of farm size (-); production per ha (+); location (Waikato in relation to other regions); buyer attributes (operation type, new buyer and enlarging); and a seller attribute defined as subdividing. Changes in technology and management practices were captured in a variable representing time. The first half of 1990 was assigned a value of 1, the second half a value of 2, and so on. The time variable also captured changes in the market environment, including the MS price received by farmers. Farm specific MS payments that reflected individual company performance and milk composition effects were not available.

Exploratory analysis was undertaken to determine data-appropriate functional form that was consistent with economic theory and hypothesized conceptual relationships. In the end, a log-linear functional form provided a better fit for the data and was found superior over a quadratic functional form. The parameters in log-linear models can be conveniently used as elasticities and easier to interpret without further computations. The empirical model was specified as in Equation 1. In the econometric model, farm level revenue attributes were not included because information on actual revenues from milksolids sale was not available. Furthermore, because of closure or merger of dairy companies regional milksolids payout could not be ascertained. The authors are aware of this limitation in the model specification and results should be interpreted accordingly. However, we contend that relative profitability from dairy farming is partially reflected in farms being sold.

LRSALEHA = ß 0 + ß 1 LSIZE + ß 2 LPDPHA +  3 NORTHLND +

ß 4 AUCKLAND + ß 5 BAYOFPL + ß 6 TARANAKI +

ß 7 LOWERNI + ß 8 SISLAND + ß 9 FSBN + ß 10 FSSN +

ß 11 BTSFRMI + ß 12 BTVFRMN + ß 13 MYR + e (1)

where:

The Waikato region was treated as a reference farm location, and thus was not included in the regression model. All explanatory variables specified in Equation 1 were subjected to collinearity diagnostics (Besley et al., 1980). None of the variables used in the model posed a collinearity problem.

3.1 Trends in dairy farm sales

The overall national trend in the number of dairy farm sales from 1990 to 1997 is shown in Figure 1. Economic units accounted for 55% to 68% of total dairy farm sales per annum during this period.

Figure 1: Number of "economic unit" and "other" dairy farms sold between 1990 and 1997.

The ‘economic’ units tended to be much larger in area than ‘other’ units, although this difference narrowed over time (Figure 2). The number of economic unit sales peaked in 1992 at 587, and thereafter trended downwards. The number of dairy farm sales, as a proportion of the total number of national herds, ranged from 2.7% in 1990 to 1.5% in 1997, with a peak of 4.1% in 1992. The least dairy farms were traded in 1997, at nearly half those sold in 1996, however this may be partly because the database had not been completely updated for 1997. The number of dairy farm sales peaked in 1992 in the North Island and 1994 in the South Island. Regional dairy farm sales were consistent with the national trends. When farm sales were expressed as a percentage of total sales, proportionately more farms changed hands in the North Island than the South Island (90% vs. 10% in 1997).

Figure 2: Size (ha) of "economic unit" and "other" dairy farms sold between 1990 and 1997.

3.2 Trends in farm sales by farm size

Dairy farm sales, cross-tabulated by four farm size categories (<40, 40-59, 60-79 and >79 hectares), are presented in Figure 3. On average, during 1990-1997, one in seven of the farms sold were less than 40 ha, while 32% were between 40-59 ha. Slightly more than half (54%) of the farms sold were 60 ha or larger. Over the study period the proportion of farms sold in the smallest (<40 ha) farm category increased (12.7% to 18.3% of total farm sales in 1990 and 1997, respectively). Nationally, the proportions of sales across farm categories were relatively stable (coefficients of variation less than 20%). However, variations were larger when data were disaggregated by regions. For example, in Northland only 2% of total sales were under 40 ha while two-thirds were 80 ha or larger. In 1993, 77% of farms sold in Northland (n=65) were larger than 79 ha. This possibly reflects the larger farms in Northland (Appendix A).

Figure 3: Number of economic dairy units by farm size sold between 1990 and 1997.

3.3 Milksolids production

The productivity (kg MS/ha) of farms sold was similar from 1990 to 1994 (approximately 526 kg MS/ha), then increased by up to 13% (575 to 597 kgMS/ha) during the following three years (Figure 4). Small farms had the highest per hectare production (685 to 760 kg MS/ha for farms less than 40 ha) and as farm size increased, MS/ha declined (416 to 515 kg MS/ha for farms 80 ha or more). This trend was consistent across all years but the gap in MS/ha across farm size groups narrowed over the years. On average, milksolids production was 92%, 82% and 62% of the production for farms less than 40 ha, for farms 40-60 ha, 60-80 ha and 80 ha and over, respectively (i.e. 732, 669, 599 and 452 kg MS/ha, respectively).

Figure 4: Sale price ($/ha) and milksolids production (kg MS/ha) for dairy economic units sold between 1990 and 1997 in 1990 dollars.

A regional comparative analysis of the productivity of dairy farms sold (Appendix A), was performed relative to the Waikato, where most New Zealand dairy farms are located. All regions had lower MS/ha than the Waikato in all years, except in 1997 when Taranaki produced 1.3% more MS/ha, and in 1990 and 1997 when Canterbury produced 9-14% more MS/ha. However, the productivity gaps between Waikato and other regions were typically large in all years.

3.4 Trends in dairy farm net sale price

The per hectare net sale price increased substantially from 1990 until 1995 (Figure 4). The inflation-adjusted net sale price for dairy farm land increased from $6,616/ha in 1990 to $11,060/ha in 1997, with a peak in 1995 of $12,158/ha. In some regions, the price change was quite dramatic from one year to another. For example, in the Nelson/Marlborough region the average dairy farm land price increased from $3,916 in 1994 to $8,021 in 1995. This was probably a farm size effect since the average farm size sold in 1995 was 75 ha compared with 130 ha in 1994. In addition, annual sales in this region were small (Appendix A). A similar effect between 1996 and 1997 was evident for this region (72 ha and is $8246/ha vs. 150 ha and $4252/ha).

The South Island results were highly variable because of the small number of farms sold per year (Appendix A). This ranged from 22 to 50 farms/year (3 to 20 farms/region/year) compared to 195 to 542 farms/year in the North Island (7 to 209 farms/region/year). Prices for South Island dairy land, on average, were $4570/ha lower than in the North Island, ranging from $2936/ha less in 1990 to $5886/ha less in 1994. Although farm land prices started to decline in North Island after 1995, prices continued to increase in the South Island. The latter trend reflects the relatively higher land prices in the North Island and the greater remaining potential for the conversion of sheep and beef cattle farms to dairying in the South Island.

In terms of size effects, land prices increased proportionately more for large farms than small farms (an increase in real price per hectare between 1990 and 1997 of 36% for farms less than 40 ha, compared to 96% for farms 80 ha and over). This trend could be associated with the greater demand for larger, and more economic, farms.

3.5 Trends in capitalisation of land values

The trend in the capitalisation of land values, as reflected by the land price per kg MS, mirrored those for the land price per ha, but increases in the former were consistently lower than the increase in per hectare price (Figures 5 and 6).

Figure 5: Sale price ($/ha and $/kg MS) for dairy economic units sold between 1990 and 1997, milksolids price¹, and interest rates². All prices are in real values (1990 dollars).

2 Interest rates are for a 15 year term loan, from the Rural Bank (1990-1993) and the National Bank (1994-1997). Rates are given for January/February (Lincoln University, 1990-1997).

Thus, gains in land productivity did not keep pace with the increase in dairy land price. Land values increased with falls in interest rates in the 1992-1994 period (r = -0.98 for annual interest rate by average per hectare price for the eight years).

Land prices ($/kgMS) in the Waikato were generally higher than for the other regions, except for Auckland for all years (5.5% to 61.9% higher), Taranaki for four years (3.1 to 6.4% higher) and Wanganui/Wellington in 1992 (5% higher). Land prices per kg MS could be expected to be similar in Taranaki and Waikato, both traditionally strong dairy farming areas, with price differences per kg reflecting the size and quality of the farms sold. High land prices in the Auckland region reflects proximity effects of the city’s expansion.

3.6 Seasonality in dairy farm sales

The average number of dairy farm sales per month for the period 1990-1997 is shown in Figure 7. Farm sales were typically greatest during the summer-autumn (February-April) and least during July-September. The seasonality in farm sales was consistent across years. These results are not unexpected, as sales occur at the end of lactation to facilitate the traditional June takeover. Few sales occur during the busy spring period when lactation commences in seasonally calved herds.

The average number of dairy farm sales per month for the period 1990-1997 is shown in Figure 7. Farm sales were typically greatest during the summer-autumn (February-April) and least during July-September. The seasonality in farm sales was consistent across years. These results are not unexpected, as sales occur at the end of lactation to facilitate the traditional June takeover. Few sales occur during the busy spring period when lactation commences in seasonally calved herds.

Figure 7: Total number of economic dairy unit sales per month for 1990 to 1997.

3.7 Buyer-seller relationships

The proportion of buyers who planned to expand their farm operation increased in 1997 compared to 1990 (15% in 1990 vs. 25% in 1997) (Table 1). The proportion who were expanding was more pronounced in the South than North Island (14% vs. 24% in 1990 and 23% vs. 45% in 1997, for the NI and SI, respectively). On the other hand, when seller type was examined, the number of farms with subdivisions was 9.9% of all sales in 1990 and 13.8% of farms sold in 1997, a 30% increase in proportional terms.

Table 1: Buyer-seller relationships for not-enlarging (NE) and enlarging (E) dairy land sales.

Of the 3,480 dairy farm sales, one in seven farms were purchased by a new owner, while 78% were purchased by an existing owner. One in 16 buyer was classed as a business operation. Similarly, slightly more than one-fifth (22%) were bought by an individual, three fifths by a partnership, one in eight was a company and one in 20 was some other type of operation (Table 1). The majority (85.4%) of farm sales were as going concerns (not expanding), and in most cases (91.9%), farms were sold in their entirety rather than subdivided (Appendix B).

3.8 A model of dairy farm sale prices

The regression model parameters are summarized in Table 2. The model explained 78% of the variation in the dependent variable (natural logarithm of net sale price in 1990 dollars). There was a low degree of autocorrelation (r =0.32, DW=1.36). Of the 13 explanatory variables employed in the model, only a Taranaki farm location (with respect to Waikato) and buyers seeking to enlarge their farm were not statistically significant. This is consistent with the smaller differences between Taranaki and Waikato land prices, compared to those between Waikato and other regions.

The model shows a 10% increase in the size of dairy farm sold is likely to decrease the net sale price per ha by 1.5%. The time trend (MYR) indicates that, ceteris paribus, dairy land prices have declined in the past. However, a 10% increase in milksolids (MS) production would improve dairy land prices per ha by 8.7%. The sale of economic dairy farm units in Northland, Bay of Plenty, lower North Island and South Island would force dairy land prices to decline, while farm sales in Auckland region would increase land prices. Individual buyers, and sellers subdividing, would decrease the net sale price. In real terms, the model suggests dairy land values will continue to increase.

Table 2: Determinants of dairy land prices in New Zealand (1990-1997).

4. DISCUSSION

Dairy farm sales between 1990 and 1997 reflected, in number and value, both optimism for future, as well as current, MS prices and interest rates. While total sales peaked in 1992, the price paid per hectare did not peak until 1995. The relatively low prices in 1990-1992 had carried through from the late 1980’s when milkfat prices had declined sharply and interest rates were high (Back, 1995). At this time established farmers with strong equity, and access to improving milk prices and falling interest rates, could expand production on significantly cheaper non-dairying land, especially in Canterbury and Southland, with no requirement to purchase shares in the industry’s processing or marketing assets. Nationally the milk supply grew at 5.3%, 2.3%, 12.8%, 0.3% and 8.0% respectively, or a total of 32% between 1990/91 and 1995/95 (LIC, 1996).

This placed enormous pressure on processing capacity in the traditional as well as the "rediscovered" dairy regions (e.g. Southland had 1026 dairy farms with 20 or more cows in 1950 (Census and Statistics Department, 1953)). To fund this expansion and to shift the burden of cost to those increasing supply, the requirement to purchase company shares was introduced (initially by Southland in 1994/95, and by all companies from 1997/98). This unbundled some of the industry’s processing and marketing value from dairy land values, and contributed, along with lower MS price forecasts, to the decline in dairy land values from ca. 1995.

Three other factors probably had some effect on dairy prices. First, the high log prices in 1992-1994 that had prompted strong investment in land for forestry, thereby underpinning sheep and beef cattle land values, began to abate (MAF Policy, 1995). Second, horticultural crop (e.g. apples, vegetables) returns were generally poor from ca. 1995 and land demand by this sector was low. Third, the costs of resource consents due to the 1992 Resource Management Act increased, especially for dairy shed effluent disposal (McShane, 1998).

The productivity (kg MS/ha) of dairy land sold increased every year from 1992, except in 1996. The lower levels of productivity in the 1992-94 period incorporate the effects of conversion of sheep and beef cattle farms to dairying (estimated at 261, 226, 365 and 151 farms, respectively, from the 1994/95 to 1997/98 seasons (NZMWBES, 1995 - 1997)). The MS per ha of farms sold was 82-100 kg/ha less than the national average for the season concerned, and mirrored the national trend for increased MS output per hectare over the study period (LIC, 1996).

Nationally, the proportion of small (<40 ha) dairy farm sales to total sales (12.7% to 18.4%) increased over the study period. The ratio of sales:farms in this size category was 1.77% in 1990, and 2.23% in 1993 (Fairweather, 1997). Corresponding turnover rates for 40-60 ha farms were 4.29 and 5.49%, and 1.93 and 2.41% for farms >60 ha. In terms of total dairy farm sales:total dairy farms the percentage annual turnover was 2.69, 2.97, 4.04, 3.43, 3.40, 2.80 and 2.88% for each year from 1990. The 1997 data were not available at publication but turnover rates have declined from their 4.04% peak in 1992.

The future of the small dairy farm (<60 ha and 150 cows) was reviewed by Parker (1998). These farms represented 39% of dairy herds and 24% of the national herd in 1995/96 and were unable to service debt at a MS price of $3.50 kg MS. Limited published data suggests these farms are owned by older farmers (Hughes et al., 1989).

Thus, the conditions seem set for many of these farms to exit the industry over the next decade, either through amalgamation with an existing dairy farm or another land-use enterprise. The sales data to 1997 provides some evidence that this is occurring - 27.6% of farms less than 40 ha and 16% in the 40-60 ha category were brought for expansion purposes versus a maximum of 11.5% in the other farm size categories (Appendix B). More of the small farms were purchased by individuals and partnerships than other forms of ownership. While partnerships dominated buyer type in all farm size categories, company farms were most prevalent in the >120 ha category. This is consistent with the growth of corporate dairy farm businesses, such as Tasman Agriculture, during the study period.

Regional variation in land prices between years reflected both productivity and the number of sales (Appendix A). Land prices were highest, but not necessarily more consistent in the dominant dairying regions: Auckland (CV = 22.4 %), Waikato (CV = 23.1%) and Taranaki (CV = 25.4 %) versus Wanganui/Wellington (CV = 16.6%). The greatest variability in dairy sales prices between years was in the South Island regions where a small number of farms changed hands each year.

5. CONCLUSIONS

This paper has reported on some, but not all, of the factors that have an impact on dairy land values. These factors, as illustrated during the 1990-97 period, can result in significant, and relatively quick, changes in the dairy land market. They are generally consistent with the model of rural real estate cycles developed by Hargreaves and McCarthy (1994). Our expectation is that nominal values for dairy farms will fall through to 2000 as the full impact of unbundling of off-farm investment in processing and marketing from land values occurs; as the status of the New Zealand Dairy Board and its relationship with Companies is determined, interest rates for on farm lending fall (possibly by 3-4% by 2000) on the basis of a lower exchange rate for the New Zealand dollar and lower demand for housing finance; and as the long-term impacts of sustained low inflation rates reduce the expectation for capital gain from land ownership.

REFERENCES

Parker, W.J., The future of the small New Zealand dairy farm. Dairy Farming Annual, Massey University 50: 43-56 (1998).

Appendix A: Statistics for dairy economic units (means).

Appendix B: Buyer and seller type by farm size.