Quarterly Estimation of Agricultural Value Added by mhk16044


									 Quarterly Estimation of
Agricultural Value Added
                 Issues covered…..
•   Agriculture as a special case for national accounts compilers
•   data sources for QVA
•   imputation of work-in-progress for long production cycles
•   seasonal adjustment
•   comparison of time series
•   reference to OECD member country practices
       Agriculture as a special case
• Source data are not ideal
   – policy users are priority
   – annual data are priority - agricultural statisticians consider
     quarterly data for some commodities to be meaningless
   – high degree of informal activity
• But large amount of quantity and price data
• Homogeneous commodities
• Extreme seasonality in some activities
• Production cycles stretch over several quarters for some
  goods - both seasonal and smooth examples
• Output and seasonality can be unpredictable due to weather
            Quarterly data sources
Many countries can use double indicator and double deflation
  - output quantity and price data often available quarterly at a
  detailed level
  - but IC may be derived using I:O ratios (O:VA)
• starting with output in constant prices
• apply IO ratio to give IC in constant prices
• difference = VA in constant prices
• reflate (revalue) both output and IC with their own prices/indices
  to give current price output and IC
• difference = VA in current prices
    From constant to current prices
Quantity revaluation often used for output

Combination of quantity revaluation and deflation/reflation using
  price indices for IC

Data fairly homogeneous at most detailed level - i.e. no significant -
  - changes in mix of sub-products
  - changes in quality over time (minimised by updating base year
             Data sources - output
In OECD countries, output often measured indirectly,
               i.e. use rather than supply
Farms not surveyed quarterly
…...but, enterprises using the products are surveyed
So, commodity flow methods are used:
   - surveys give use (IC)
   - adjusted with any FC
   - plus exports less imports
   - plus changes in inventories (if possible)
Data sources - intermediate consumption
Although supply data for feed, seed and fertiliser may be quarterly,
   may not be used for IC
       ….maybe not be available by activity?

IO ratios often used, based on previous benchmark data (farm
…..but IO ratios fail when exogenous effects such as weather,
  disease, etc. affect IO
             UK example
Farm surveys :

 annual censuses

 twice yearly crop production survey,
 quarterly cereal stocks survey,
 monthly survey of grain fed to livestock,
 monthly survey of the production and marketing of
  hatching chicks and eggs.
                          UK - Use surveys

Enterprise acti vi ty      Data collected                                           Frequency

Dairies                    Milk purchased,                                          Monthly
                           Output of dairy products
Slaughterhouses            Nos. of cattle, sheep, pigs, poultry slaughtered and     Weekly, monthly and
                           average weights                                          quarterly
Egg packers                Nos. of eggs graded and packed and average prices paid   Monthly
                           to producers
Bacon and ham producers    Weight of bacon and ham cured                            Monthly
Millers                    Cereals used, flour produced                             Monthly
Brewers and distillers     Use of wheat and barley                                  Monthly
        UK - Estimating output from inputs

   Agricultural economists and statisticians closely monitor stocks
   and flows of feedstuffs….to use in modelling output

   Surveys of sales and stocks, combined with exports/imports data

Enterprise activity   Data collected                                                       Frequency

Feed manufacturers    Feedstuffs‟ production, stocks, raw materials. Value and volume of   Monthly
                      feed sold to farmers
Poultry farmers       Production of poultry feed for own use                               Monthly
                UK - Estimating IC

Although figures on supply of seed, feed and fertilisers are available
   monthly, they are not used directly for IC
               (no activity breakdown?)

IC is derived using IO ratios but this is under review
               (effects of disease….?)
           Australia - Data sources
wheat and barley      - marketing boards
livestock             - slaughterers
wool                  - trade association

IO ratios not used due to weather conditions
seed                    - area sown x rate
feed                   - quarterly supply
marketing costs        - base year value extrapolated by constant
                                price output
    Types of agricultural output
• Continuous, with short production cycles
   – milk, eggs, chicks, poultrymeat
• Single-use plants and livestock - they produce output
  once only, after a long production cycle
   – arable crops, livestock fattened for slaughter
   – (poultrymeat is single-use but a short production cycle)
• Cultivated assets
   – dairy cattle, breeding livestock, draught animals, bees, vines,
     fruit trees
   – production of these assets covers rearing them to productive
     maturity and usually takes > 3 months
     Production of cultivated assets
• When assets are produced for eventual own use, output
  of unfinished goods is recorded as own-account GFCF
• When produced under contract, the sale has been agreed
  and is considered to be taking place continuously -
  recorded as sales of unfinished goods for producer and
  GFCF for purchaser
• When assets are produced with the intention of eventually
  selling them, but the purchaser is not identified, output
  of unfinished goods = work-in-progress. There is no
  GFCF associated with this transaction
                      Dairy cattle…...
• If a dairy farmer breeds his own replacement milking cows, the
    production of the calf and then rearing her to maturity is recorded as
    own-account GFCF i.e. simultaneously output and GFCF
• If he buys year-old heifers from a breeder at a market, he records the
    purchase as GFCF and subsequent rearing to maturity as own-account
    GFCF (as above). The breeder records the birth and first year of
    rearing as work-in-progress
• If the breeder produced and reared calves under contract to the farmer,
    then the breeder would record all production as sales, and in parallel,
    the farmer records GFCF
i.e. in all cases, output should be recorded in each quarter on a
    continuing basis. This avoids negative value added in earlier
               Long production cycles
SNA93 accounting rules state that flows recorded on accrual basis:
• when economic value is created, transformed or extinguished, or when claims
     and obligations arise, are transformed or cancelled.
i.e. production should be recorded at the time it takes place, and valued at the
     prices of that time.
So, the production of unfinished products is recorded
               as work-in-progress.
Principle is     - in line with business accounting,
                 - better reflects economic reality,
                 - and avoids recording of holding gains.

               But does it make sense for agriculture?
             Long production cycles
                 SNA93 (10.107) states…..
“In the case of single-use plants or livestock, any such crops or
   livestock that have not yet been harvested or slaughtered at the
   end of the accounting period constitute work-in-progress, as
   follows: (a) When accounts are compiled quarterly, the value
   of the finished output of an annual crop - i.e. the value of the
   grain or other crop actually harvested - may be distributed over
   the quarters in which production has taken place in proportion
   to the costs incurred each quarter. The value of the output
   produced in the quarter in which the crop is harvested is then
   equal to the value of the harvested crop less the value of the
   additions to work-in-progress produced in the previous
         Example - arable crops…..
• At the beginning of the production cycle, output is forecast, in
  volume/constant price terms
e.g. area planted x expected yield x base year average price
• .an allocation of costs across quarters is determined, e.g.
  Q1=0.20        Q2=0.25 Q3=0.30 Q4=0.25
• .the constant price (volume) forecast is allocated across quarters
  using the costs profile;
• .the forecast output (WIP) in each quarter is revalued to the
  current price of that quarter;
• .all quarters are revised after the harvest.
NB harvest value (current prices) will not equal value of WIP due
  to holding gains
        Eurostat alternative method
                      (Handbook 3.35 +)
• Impute output (WIP) as equal to costs in the pre-sale quarters
• In the sale quarter, record balance of output (beware of holding
• No revisions made to earlier quarters, but output, and values
  added in last quarter could be strange, and very different from
  year to year - last quarter being used as “residual”
• Also, the pattern of output, and value added, across quarters
  depends on how costs are defined
    Eurostat - effect on value added of
    different cost estimation methods
If “full” valuation used for costs:
output =      intermediate consumption
              + compensation of employees
              + taxes on production
              + operating surplus/mixed income
              + consumption of fixed capital
then, for all quarters:
value added = compensation of employees
              + taxes on production
              + operating surplus/mixed income
              + consumption of fixed capital
But if costs (and thus output) are estimated as:
output =      intermediate consumption
              + compensation of employees

then, for pre-sale all quarters:
value added = compensation of employees

and for the sales quarter:
value added = compensation of employees (for the sale quarter)
              + taxes on production (for whole year)
              + operating surplus/mixed income (for whole year)
              + consumption of fixed capital (for whole year)

SNA68 recommended allocating all output to sale quarter
      Problems with imputing WIP
  SNA93 method:
     - reasonable QVA profile? (non-negative, smoothed)
     - but big revisions possible

Eurostat alternative:
       - no revisions
       - but possible exaggerated harvest peak
       - negative output possible if harvest fails?
                    IMF concerns….
Although accrual approach is paramount….SNA93 (6.100) :
“There may be circumstances in which the uncertainties attached to the estimation
   of the value of WIP in advance of the harvest are so great that no useful
   analytical or policy purpose is served by compiling such estimates”.

Uncertainty about output has stopped many countries implementing
Implementation of prices in non-market quarters may be very
Imputing WIP does not reflect economic reality in the case of
  agriculture - farmers do not behave as if income will be
               Allocation in practice
             (interpolation, imputation…)
• The ways in which countries treat the issue of single-use crops
  and livestock with a long production cycle (whether to
  interpolate, or not) vary considerably:
   – Crop production is assigned to the quarters in which harvest occurs –
     Austria, Germany, Norway
   – Crop production is assigned to the third quarter – Denmark, Finland,
     Netherlands (sugar beet in the fourth quarter)
   – A quarter of total output is assigned to each quarter – Sweden
   – Intermediate consumption for crops is distributed over quarters in fixed
     proportions – Denmark, Norway.
   Summary of allocation methods
• Output equal to harvest/sales (no WIP)                       SNA68
   – possible negative VA                     - no revisions
   – large peaks
• Output allocated equally to all Qs
   – no peaks                                 - revisions
   – steps between years
   – no relation to economic reality?
• Output allocated with costs profile                          SNA93
   – seasonal pattern but smaller amplitude   - revisions
• Output allocated equal to costs, except harvest Q            Eurostat
   – possible large peaks                     - no revisions
   – possible negative output?
                     Agricultural QVA - Selected OECD Countries
                              Constant prices referenced to 1995=100







         Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1
        1990   1991  1992  1993  1994  1995  1996  1997  1998  1999  2000  2001
                  Australia     France      Germany     Korea      New Zealand   Turkey
                    Agricultural QVA - Nordic/Scandanavian Region
                                   Constant prices referenced to 1995=100










         Q1    Q4   Q3    Q2     Q1    Q4     Q3     Q2    Q1      Q4   Q3    Q2    Q1    Q4   Q3
        1990                    1993                      1996                     1999
                         Netherlands        Norway        Sweden        Denmark      Finland
                                Agricultural QVA - Selected CCs
                                  Constant prices referenced to 1995=100







         Q1    Q3    Q1    Q3     Q1      Q3    Q1     Q3    Q1      Q3      Q1    Q3      Q1    Q3    Q1
        1994        1995         1996          1997         1998            1999          2000        2001
                                 Poland        Czech Rep.       Slovak Rep.        T urkey
                               Agricultural QVA - Baltic countries
                                     Constant prices referenced to 1995=100






       Q1    Q3    Q1    Q3    Q1    Q3    Q1    Q3    Q1      Q3      Q1    Q3    Q1    Q3    Q1    Q3    Q1    Q3
      1992        1993        1994        1995        1996            1997        1998        1999        2000
                                           Estonia           Latvia          Lithuania






         Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1
        1992   1993  1994  1995  1996  1997  1998  1999  2000  2001
                    Poland     Czech Rep.   Slovak Rep.   T urkey
                    Estonia    Latvia       Lithuania
              Seasonal adjustment
• Zero or negative values cause problems
• Multiplicative models not possible
• Additive models may not be appropriate for series
• Australian solution = to use pseudo-additive model
• Previously had „adjusted‟ by allocating VA equally across Qs
  („annual spreading‟)
• Only allows growth at annual joins (Q2-3 in Australia)
               Discussion points...
Do participants :

•   Survey farms quarterly, for output and IC?
•   Use IO ratios?
•   Survey processors for use data (commodity flow)?
•   Survey suppliers for IC data?
•   Impute WIP (allocate output) for long production cycles?
•   Have problems with seasonal adjustment of agricultural QVA?

To top