1
Forestry profitability panel dataset
Wei Zhang, Motu Research Data Documentation
Motu Economic and Public Policy Research
Date accessed/created: August 2010 Motu Ref ID: U9970
Suggested Citation:
Zhang, Wei. "Forestry Profitability Panel Dataset and GIS layers," obtained by Motu Research in 2010. Unrestricted data 9970. Details online at http://www.motu.org.nz/building-capacity/datasets.
Raw or derived data: Derived dataset Restrictions: Unrestricted
Can Motu put this data on our website? Yes Can Motu put this document on our website? Yes
Contact Details
Motu Economic and Public Policy Research Level 1, 97 Cuba St, Te Aro,
PO Box 24390 Wellington New Zealand Email [email protected] Telephone +64 4 9394250 Website www.motu.org.nz
User Responsibilities: Normal principles for the attribution of sources of information are expected to apply in any resulting publications. However, Motu cannot be held responsible for results obtained from applications of this data, or derivative versions of this data, by outside individuals. The results of any analysis based on this data by outside parties are not endorsed by Motu. It would therefore be inappropriate for outside users to suggest or infer that these results or interpretations attached to these results can in any way be attributed to Motu or its researchers. © 2010 Motu Economic and Public Policy Research
2
Data Documentation
(Please note that this is informal documentation intended to help users.
It is not a polished document. Additions/corrections are welcomed at [email protected].) 1. Main Motu contact for this data:
Wei Zhang, [email protected] 2. Data keywords:
Forest, profit, cashflow, GIS, slope
Dataset abstract:
Forestry profitability is measured as both net present values (NPV) and annualised return (treating NPV as an annuity), and classified by 10 Wood Supply Regions, 4 different slope levels and 7 different distances to port. We have linked the panel data with GIS maps and created an annual time series of potential forestry profitability maps for the whole of New Zealand from 1972 to 2009.
3. Motu Working Papers using this data set. n/a
4. Variables:
Most variables are self-explanatory. “Profit_pv” is the difference between the revenue and the total costs measured in present value. “AnnualizedReturn” is the annualized return derived after treating “Profit_pv” as an annuity (over 28 years with discount rate 4%). Both variable are measured in dollar per ha in 2009 dollar.
5. Additional notes.
Raw data:
Log price data was obtained from MAF
(http://www.maf.govt.nz/forestry/statistics/logprices/logprices1.xls). This series only traces back to 1992. One could certainly use the data to create a panel dataset and maps from year 1992. What if someone needs to create a longer series? In this case, we used the “Real Export Log Price Index”. This index was provided by David Evison from the Forestry School Uni Canterbury. According to his documentation, the index was derived by deflating the total export log unit value with the PPI all industry outputs. Using this index, we are able to create log price series tracing back to 1972. The drawback of recreating the log price is that the price series lost their individual variation. For example, Figure 1 shows that different types of log
3 have their specific price variations whereas Figure 2 presents that the variation is lost once we use the price index to recreate log prices. Nevertheless, both series are congruent with each other in terms of general pattern.
Figure 1 Real log price series from MAF, 1992 to 2009, deflated by using CPI, dollar per cubic meters
Figure 2 Real log price series deriving using the “real export log price index”, 1972 to 2009, dollar per cubic meters
Wood yield data is obtained from MAF (Ministry of Agriculture and Forestry, dataset, 1996). Area by regime and wood supply region (WSR) data is from the National Exotic Forestry Description
4
Figure 3 Wood yield data, cubic meters per ha
Figure 4 Area by regime and WSR from 2002 to 2009
On the cost side, there are operational cost and harvesting cost which consists of logging and roading cost, and cartage cost.
Operational cost data came with the wood yield data shown in
5
Table 1 Operational cost by Regime: Pruned with/without thinning (above) and unpruned with/without thinning (below), dollar per hectare
Age 1 4 6 7 8 9 14 27 Land Preparation_Fencing_Tracking 356 Planting_fertilising_releasing 945 Prune 1 678 Prune 2 550 Prune 3 520 Waste Thin 1 218 Waste Thin 2 295 Dothistroma spray 29 Mapping 10 Mid-rotation inventory 27 Pre-harvest inventory 84 Age 1 4 6 10 14 27 Land Preparation_Fencing_Tracking 356 Planting_fertilising_releasing 945 Prune 1 Prune 2 Prune 3 Waste Thin 1 454 Waste Thin 2 Dothistroma spray 29 29 Mapping 10 Mid-rotation inventory 27 Pre-harvest inventory 84
The logging, roading and cartage cost data are provided by David Evison from AgriFax report which we do not have assess.
6
Table 2 Logging and cartage cost, from the Regional Log Price and Cost Report 2010 by Agrifax Ltd
Cost data is “one-off”: they are not time series data. The Producers Price Index (Figure 5) is used to create time series of cost data.
Figure 5 Producers price index
Note: PPI Forestry and logging inputs, long term linked series, deflated by PPI All industry inputs
GIS data:
WSR map is created by dissolving the boundary of Territorial Authority map according to the definition of each WSR provided from NEFD report.
There are 12 seaports identified -- Whangarei, Tauranga, Gisborne, New Plymouth, Napier, Wellington, Picton, Nelson, Lyttleton, Timaru, Dunedin and Bluff. We then locate each port on
7 a raster map and calculate the Euclidean distance from each pixel to the closest seaport. We use this map as a proxy for the true (road) distance to market.
Slope data is derived from Land Environments of New Zealand (LENZ) Map (Landcare Research, dataset, 2004). The original data is the raster form with the resolution of 25 by 25 m1.
We used the resample function from the Arc GIS to “condense” the map to a map of 500 by 500 m (25 ha per pixel) resolution which is compatible with LURNZ’s base map.
We have to categorize the slope data into 4 different levels (flat, easy, steep and v steep) identified from the AgriFax report. The Land Use Capability handbook (Lynn et al, 2009) provides a guideline of classifying slope.
Table 3 Slope group according to the Land Use Capability survey hand book (Table 26 of the book)
Table 4 Slope classification
Flat Easy Steep V steep 0 -7 8 - 20 21 - 25 >=26
8
Method:
We first calculate the revenue for each WSR at yeart. Revenues are generated from selling all types of logs and measured in dollars per hectare. For each wood supply region and each year we calculate the revenue as follows:
3 1 , , 4 1 , ( ) i j i t i j t j t W P Yield H R , (1)
where Rt is the revenues from selling all types of logs in a WSR at year t.
j
represents 4 tending regimes. Wj,tis the weight of regimej
in year t, which is calculated byt t j t j Area Area W, , , (2)
where Areaj,tis the area of regime
j
at year t and Areat is the total forestry area of theWSR at year t.
i
represents 3 log types; Pi,tis the price of log typei
in year t;H
is the harvest age, aninput of the model; Yieldi,j(H)is the log type
i
’s wood yield at the time of harvest (age of H) under tending regimej
. The present value of the revenue is calculated byH t t r R R PV ) 1 ( _ , (3)
r
is the discount rate, an input to our formulae, obtained from (Manley, 2007). The same discount rate is used to calculate the present value of costs.Secondly, we calculate the present value of cost which is measured in dollars per hectare. We first divide costs into two parts – operational costs and harvest costs. Operational costs include establishment, over head cost (annual cost), tending and managing a forest, excluding land costs (see
Table 1)2. On the other hand, costs occurred at the age of harvesting include the expense resulted from
harvesting and transporting. Hence, the costs will differ according the terrain (slope) and distance to market or ports. For we only have one year cost data, the producers price index data is used to generate time series data. We first present the formula for calculating the present value of the operational cost at yeart
2 We assume the production thinning operation breaks even, i.e., the cost of production thinning offsets the revenue from it.
9 H n n n t t r OPC PI OPCOST PV 1(1 ) _ , (4)
where PV _OPCOSTt is the present value of the operational costs under a regime (costs do
not distinguish across WSRs); PItis the producers price index at year t.OPCn is the operational costs at
timen.
The present value of harvest costs is presented as the formula below.
H i i t t r H Yield t CartageCos t LoggingCos PI t HarvestCos PV ) 1 ( ) ( ) ( _ 3 1 , (5)
wherePV _HarvestCosttis the present value of harvest costs;LoggingCostand
t
CartageCos are logging costs for different slopes and cartage costs for different distance to markets
respectively; they are measured in dollars per cubic metre of log. These costs vary by pixel as slope and distance are measured at that level.
Finally, the net present value of measured in dollars per hectare can presented as the formula below. t t t t PV R PV OPCOST PV HarvestCost ofit PV_Pr _ _ _ , (6)
The net present value is effectively a one-off payment at the time of establishment, for a single rotation forest crop. This measurement of profitability is however not directly comparable with those commonly used in the dairy and sheep and beef sectors. These sectors normally use a yearly measurement such as earnings before interest and tax (EBIT). By assuming forestry profit an annuity (constant discount rate and yearly payment, for
H
years), we could derive an equivalent annualisedreturn. 1 ) 1 ( ) 1 ( Pr _ return Annualised t H H t r r r ofit PV , (7)
For 10 WSRs, 4 different slope levels and 7 different distance thresholds, we effectively calculate 280 different values of profit each year.
Output
The panel data is in .csv form which is software-friendly, i.e., R, Stata, Matlab and even M$ Excel.
The GIS data is in .asc form. This form is ArcGIS friendly. The files can be directly imported into ArcGIS 9.3+. Otherwise, if you use ArcGIS 9.2 or -, you need to ASCIITORASTER first before visualizing the map. The resolution of the map is 25 ha per pixel or equivalently 500m by 500m.
10 References
Landcare Research. "Land Environments of New Zealand (LENZ) Map 2002," obtained by Motu Research 2004. Restricted dataset Details online at
http://www.motu.org.nz/building-capacity/datasets.
Lynn, Ian; Andrew Manderson; Mike Page; Garth Harmsworth; Garth Eyles; Grant Douglas; Alec Mackay and Peter Newsome. 2009. Land Use Capability Survey Handbook, 3rd ed.AgResearch Ltd, Hamilton; Landcare Research, Lincoln; Institute of Geological and Nuclear Sciences, Lower Hutt. Available online at
http://www.landcareresearch.co.nz/research/soil/luc/luc_handbook.pdf.
Manley, Bruce. 2007. "Discount Rates Used for Forest Valuation - Results of 2007 Survey", New
Zealand Journal of Forestry, 52:3, pp. 21-7.
Ministry of Agriculture and Forestry. "National Exotic Forest Description Regional Yield Tables as at 1 April 1995," obtained by Motu Research 1996. Restricted data Details online at http://www.motu.org.nz/building-capacity/datasets.
