Electronic Journal of Polish Agricultural Universities (EJPAU) founded by all Polish Agriculture Universities presents original papers and review articles relevant to all aspects of agricultural sciences. It is target for persons working both in science and industry,regulatory agencies or teaching in agricultural sector. Covered by IFIS Publishing (Food Science and Technology Abstracts), ELSEVIER Science - Food Science and Technology Program, CAS USA (Chemical Abstracts), CABI Publishing UK and ALPSP (Association of Learned and Professional Society Publisher - full membership). Presented in the Master List of Thomson ISI.
2017
Volume 20
Issue 4
Topic:
Economics
ELECTRONIC
JOURNAL OF
POLISH
AGRICULTURAL
UNIVERSITIES
Suchodolski P. , Idzik M. 2017. WOOD SUPPLY VARIABILITY, EJPAU 20(4), #15.
Available Online: http://www.ejpau.media.pl/volume20/issue4/art-15.html

WOOD SUPPLY VARIABILITY

Przemysław Suchodolski, Marcin Idzik
Faculty of Economic Sciences, Warsaw University of Life Sciences – SGGW, Poland

 

ABSTRACT

The study deals with  the issue of wood supply variability in Poland illustrated with an example of the Płock Forest District. The aim of this paper was to identify and evaluate the supply variability of the selected wood assortment. According to the rules of time series decomposition based on the CENSUS II X-11 method, eight wood types were tested. Data came from the Płock Forest District collected from 2004 to 2014 in monthly intervals. Based on the conducted study, it was found that wood supplies in the Płock Forest District are not characterized by a clear share of systematic changes, trends and cycles. The results of this study have shown a significant impact of the seasonal and accidental fluctuations on timber prices. The trends and cycles were irrelevant for all 8 assortments. The dynamics of seasonal fluctuations were different depending on the assortment. Seasonal fluctuations had the biggest influence on wood supply of most of the tested assortments. The analysis has shown that there were also a lot of random fluctuations, which had a high amplitude of deviations.

Key words: National Forest Holding, trend, cycle, seasonal fluctuations, random fluctuations, multiannual management plan.

INTRODUCTION

The National Forest Holding plays a key role in the production and distribution of wood in Poland. Its main target is not only the production and sale of wood, but also the restoration of the cut areas, registration of forest animals and the protection of sites of natural value which are recognized as protected [16].

The supply and demand factors have the main influence on the wood price levels. While the supply of wood remains within the competence of the National Forest Holding, which must take into account the recommendations of the Minister of the Environment, the demand factors, in general, are influenced by the changes in the economic situation in the country. The supply policy of the National Forest Holding is strongly connected with their market position, and is affected by the forest management plan of the Forest Law of 28 September 1991 [16].

The analysis of the phenomena over time contains different aspects, starting with defining the basic characteristics of the individual time series, through the analysis of the variability, to the multidimensional analysis of the time series. One of the goals of such analyses is to detect the nature of the phenomenon represented by the sequence of observations, while the second goal is to predict the future values. Both of these objectives require identifying and describing, in more or less formal terms, the components of the time series. The conclusions of such analyses can be based on the choice of methods for short-term forecasting and on the nature of dependence between the phenomena. At the same time, the calculated characteristics of the changes over time of the individual components allow us to estimate the risk magnitude, depending on the horizon of decision making.

The analysis of the changes in wood supply over time is important because it allows us to identify the main characteristics of the wood supply patterns over time. The knowledge of these facts enables one to prepare short-term timber supply forecasts and is helpful in assessing the current volume of raw material traded, giving the opportunity to anticipate the wood supply levels.

The aim of the study is to identify and evaluate the variability of supply of the selected wood types. The following questions were also asked: What is the significance of the changes in timber supply in the long term? What is the significance of the cycles, seasonal impact and accidental fluctuations? What is the influence of the systematic changes on the changes in the supply of wood that allow forecasting, and what is the share of random changes due to the influence of incidental, one-off factors? To what extent do short-term changes obscure the actual picture of the long-term trends in the changes in wood supply, hindering their correct interpretation?

METHODOLOGY

In analytical works, one of the first steps is to isolate individual components of time series and measure their magnitude. A time series is a series of elements characterizing an element or group, ordered chronologically, i.e. according to units of time [10]. Among the basic regularities of changes in wood supply over time, the analysis includes trends, cyclical fluctuations, seasonal fluctuations and accidental fluctuations.

In order to identify the variability components of the supply time series, the method of multistage decomposition of Census II X-11 time series was used. The research was done using STATISTICA, Gretl and Microsoft Excel. The empirical data came from the Płock Forest District and covered a time period between 2004–2014 in monthly intervals.

The Census II X-11 method allows one to extract components responsible for the time series dynamics. These include: seasonal fluctuations (S), random fluctuations (I), cyclical fluctuations with or without a trend (TC or C) and long-term linear or non-linear trends (T).

In the long run, the trend is characterized by systematic unidirectional changes in the level of the analyzed phenomenon [9].

A cyclical character is an element of the time series variability, characterized by the presence of regular patterns of movements around the trend over one year [9].

Seasonal fluctuations refer to short-term changes, not exceeding one year. Seasonal fluctuations are the changes in the value of a variable around the trend, characterized by a repetition within the same time interval [9].

Random fluctuations are irregular variations around the trend, resulting from random events [7].

The above components can occur in different time series in different combinations, different relationships, and have different roles in explaining past changes. At the same time, the assumption is not always fulfilled; there is no connection between the individual components.

The Hodrick-Prescott filter is used to separate the cyclical character and the trend. This method separates these components from the time series of stochastic nature and changes smoothly over time [5].

An essential part of the decomposition of the time series is to clean it from accidental components. For this purpose derandomization is carried out according to the concept of Months of Cyclical Dominance [5]. It consists of applying a moving average with a smoothing width equal to the number of months needed to extract the systematic changes.

The study covered the supply of 8 types of wood: total wood, total timber, coniferous and broadleaved timber total, broadleaved and coniferous medium-sized wood for industrial and mechanical processing (S2A and S2B), broadleaved (WA0, WB0, WC0, WD) and coniferous (W0, WD, WK) sawmill wood.

RESEARCH RESULTS

Studying the time series of wood supply of various sorts using the Census II X-11 decomposition has provided information about the significance of the individual components of the variability of the time series. Graphical results were presented for total wood and medium-sized wood.

Graph 1. Total wood supply together with a long-term trend (trend-cycle) and trend [m3]
Source: Own elaboration based on data from Płock Forest District

The analysis of the time series of total wood supply (Graph 1) shows the presence of seasonal and accidental fluctuations. The trend line shows a constant level throughout the period. Also noticeable is a small amount of cyclical fluctuations expressed in mild fluctuations in supply around the long-term constant level. The average amplitude of the cyclical fluctuations in total wood supply is 4.75%. The lowest cycles were reached in November 2005, February 2008, and June 2012. The cyclical fluctuations were recorded in January 2009, September 2011 and February 2013. The cycle intervals shortened over the survey period from 4 to 2 years.

Graph 2. Seasonal and random variations in total wood supply as % of long-term trend deviations [%]
Source: Own elaboration based on data from Płock Forest District

In the course of the total wood supply there is a clear seasonal effect (Graph 2), which has been changing since 2009, decreasing the amplitude of fluctuations. On average, the amplitude of the seasonal total wood supply is 60% in relation to the average supply level. The highest annual harvest is in March and it is about 25% higher than the annual average. On the other hand, the lowest annual harvest occurs in January and is approximately 36% lower than the annual average. A particular feature of the studied series is the increasing frequency and amplitude of the random fluctuations. Since 2008, the fluctuation amplitude has increased, which means an increase in incidence of random events and their magnitude of total wood supply volume, which can significantly distort the quality of the potential predictions. In the time series of total wood supply, the long-term changes described by the trend and cycle together account for only 6% of the fluctuations of the time series, the seasonality is responsible for 50%, and the casual fluctuations account for 44% of the total variation.

Graph 3. Total broadleaved S2A, S2B timber supply together with long-term trend (trend-cycle) and trend [m3]
Source: Own elaboration based on data from Płock Forest District

An analysis of the course of the temporal supply of broadleaved medium-sized timber (Graph 3) indicates the presence of seasonal and accidental fluctuations. The trend line also indicates the increase of the long-term trend throughout the period. It is worth noting that on average, the trend is rising by 1.5%; hence it can be stated that the long-term acquisition grows from year to year by 159 cbm. The cyclical fluctuations, expressed in systematic deviations from the long-term trend, can also be observed. The cycles reached their minimum in September 2009, August 2011, and February 2014. The cyclical fluctuations peaked in February 2005, March 2009, and January 2013. The average fluctuation amplitude throughout the study period was 14.8%. the cycle length was 4 years.

Graph 4. Seasonal and random variations in total broadleaved timber supply as % of long-term trend deviations [%]
Source: Own elaboration based on data from Płock Forest District

In the course of the supply of medium-sized broadleaved wood S2A, S2B over time, there is a clear influence of seasonality (Graph 4), whose pattern from 2010 has changed. The seasonal fluctuations have diminished. Only the extremes, where the lowest point was in January and the highest in March, did not change throughout the study period. On average, the amplitude of the supply of medium-sized broadleaved wood S2A, S2B due to seasonality by 2010 is 100% and after 2010 – 55%. This means that by 2010 the fluctuations reached the value of entire monthly gain, and after 2010 this value fell by half. This indicates that the fluctuations in the volume of supply per year have decreased by half.

A particular feature of the studied series is the increasing frequency and amplitude of the random fluctuations. Since 2008, the fluctuation amplitude has increased, which means that the incidence of accidental events and their magnitude increases, which can significantly distort the quality of the potential predictions. Their amplitude regularly exceeded 60%, and the maximum deviation reached 100%. In the time series of the supply of S2A broadleaved wood and S2B coniferous wood, the long-term changes described by the trend and cycle together account for only 12% of the fluctuations of this time series, the seasonality is responsible for 44%, and the seasonal fluctuations account for 44% of the total variation.

Graph 5. S2A and S2B coniferous supply along with long-term trend (trend-cycle) and trend [m3]
Source: Own elaboration based on data from Płock Forest District

The evaluation of the time series of medium-sized S2A and S2B coniferous timber (Graph 5) indicates the predominance of seasonal, cyclical and accidental fluctuations. The trend line indicates the long-term changes that can be parabolic with the maximum reached in 2010. After 2010, the average annual rate of decline is 2% on average, and the supply of medium-sized coniferous timber S2A, S2B fell by 250 cbm per year. The share of cyclical fluctuations, which were regular throughout the period under review, can also be noted. The cycles reached their minimum in August 2004, January, 2008, June 2010 and May 2012. The maximum cyclical fluctuations were recorded in November 2006, January 2009, June 2011 and July 2013. The average fluctuation amplitude throughout the study period was 7.38%. The cycle length has decreased from 4 years to 2 years, and the border year was 2009.

Graph 6. Seasonal and random variations in S2A, S2B coniferous timber supply as % of long-term trend deviations [%]
Source: Own elaboration based on data from Płock Forest District

In the course of supply of medium-sized S2A and S2B coniferous timber in time, there is a clear influence of the seasonal and accidental fluctuations (Graph 6), whose course since 2009 has begun to change rapidly. The amplitude of seasonality changes has decreased. On average, the amplitude of the supply of medium-sized coniferous timber S2A and S2B conifer seasonality was 35% before 2009, and after 2009 it was 20%. Throughout the considered period, the largest supply of wood was in March and the lowest in December.

A particular feature of the studied series is the increasing frequency and amplitude of the random fluctuations. Since 2008, the amplitude of fluctuations has decreased, despite the occurrence of one-off high amplitude effects. By 2008, the annual fluctuation amplitude was 4.3%, and after 2008 it was 3.17%. In the time series of medium-sized S2A and S2B coniferous timber supply, the long-term changes described by the trend and cycle together account for 14% of the fluctuations of the series, the seasonality is responsible for 28%, and the seasonal fluctuations account for 58% of the total variation.

SUMMARY AND CONCLUSIONS

Application of CENSUS II X 11 method enabled an implementation of following main aim of this paper: identification and evaluation of the variability of supply of the selected wood types throughout extracting components responsible for the time series dynamics. Applied method had been chosen by authors basing on comprehensive analysis of the time series. As disadvantage of CENSUS II X 11 method, shortage of separating events affecting the stationarity of the series is mentioned. However, this drawback had no impact on presented research results.

The share of the individual variations in the supply of the studied wood grades indicated the highest average share of random fluctuations (48.25%) and seasonal fluctuations (43.73%), which predominated the trend and cycle, whose average share in the overall variability of the examined time series amounted to 8.12%. In the case of the coniferous S2A, S2B and WA0, WB0, WC0, WD broadleaved timber, the accidental fluctuations (58%, 71%, respectively) constituted the largest share , which indicates the high impact of factors that are not systematic and are due to one-offs. The share of the trend and cycle was on a low level for each type. This means that during the period under review, the long-term direction of the changes in the supply of wood did not alter significantly and its dynamics were small. This is proof of a stable level of acquisition over a long period of time.

The supply of these general varieties indicates the largest share of seasonal variations in total wood (50%), total timber (48%), total broadleaved timber (49%), total coniferous timber (48%) and coniferous W0, WD, WK timber (60%).

The share of the accidental fluctuations in all types exceeds 40%. The share of the trend and cycle in the supply is negligible and is 5–14% for all types. Such distribution of the individual components of variability indicates the sensitivity of the supply of wood, understood as acquisition, to climate change and short-term fluctuations in economic conditions. Demand changes in the short term may decide to increase the logging only within the annual plan.

Table 1. Share of individual components of supply of selected wood assortments in their total variance based on change over time [%]
Assortment
Supply
Assortment
Supply
I
TC
S
I
TC
S
Total wood
44%
6%
50%
S2A, S2B broadleaved
44%
12%
44%
Total large timber
47%
5%
48%
S2A, S2B coniferous
58%
14%
28%
Broadleaved large timber
45%
6%
49%
WA0, WB0, WC0, WD broadleaved
71%
7%
23%
Coniferous large timber
44%
8%
48%
W0,WD, WK coniferous
33%
7%
60%
Source: Own elaboration based on data from Plock Forest District

The results of the research indicated the highest share of seasonal and random fluctuations in the variability of the supply of the tested varieties. Such an important contribution of these factors results primarily from the specific features connected with the acquisition of raw material which is particularly sensitive to weather and climate change.

The long-term elements of variability are the result of the application of the management plan by the Polish National Forest Holding which provides the direction and the dynamics of the developing trend of the supply of particular wood types. The low dynamics of the long-term changes characterized by the trend indicates a stable level of quantity of raw material offered every year.

REFERENCES

  1. Biniek P., 2007. Drewno wokół nas i jego perspektywy w XXI wieku [Wood around us and its perspective in XXI century]. Muratorplus.pl [in Polish].
  2. Dagum E.B., 1988. X11-ARIMA/88. Seasonal Adjustment Method – Foundations and User’s Manual. Statistics Canada. Ottawa.
  3. Giefing D.F., Pazdrowski W., 2012. Szacunek brakarski i klasyfikacja drewna okrągłego [Missing and classification of round wood]. Uniwersytet Przyrodniczy w Poznaniu. Poznań [in Polish].
  4. Hill T., Lewicki P. 2007. Statistics: Methods and Aplications. Statsoft, Tulsa.
  5. Hodrick R., Prescott E.C., 1997. Postwar U.S. Business Cycles: An Empirical Investigation. Journal of Money, Credit, and Banking. Columbus.
  6. Idzik M., 2013. Projekcja dochodów na 2015 rok dla wybranych produktów rolniczych [Projection of revenue for 2015 for selected agricultural products] [in:] Skarżyńska A. [red.]. Nadwyżka bezpośrednia z wybranych produktów rolniczych w 2012 roku oraz projekcja dochodów na 2015 rok [Direct surplus from selected agricultural products in 2012 and projection of revenues for 2015], Instytut Ekonomiki Rolnictwa i Gospodarki Żywnościowej Państwowy Instytut Badawczy, Warszawa [in Polish].
  7. Idzik M. 2013. Szeregi czasowe i składowe ich zmienności [Time series and their volatility] [in:] Stańko S. [red.]. Prognozowanie w agrobiznesie, Wydawnictwo SGGW, Warszawa [in Polish].
  8. Jennings C.L., Kulahci M., Montgomery D.C., 2016. Introduction to Time Series Analysis and Forecasting, Wiley, wydanie 2, New Yersey.
  9. Jóżwiak J., Podgórski J., 2012. Statystyka od podstaw [Statistics from scratch], PWE, Warszawa  [in Polish].
  10. Kompa K., Tarczyński W., Witkowska D., 2013. Współczynnik Beta. Teoria i praktyka [Beta factor. Theory and practice]. Pielaszek Research. Warszawa [in Polish].
  11. Musgrave J.C., Shiskin J., Young A.H., 1967. The X-11 Variant of Census Method II Seasonal Adjustment Program, Bureau of Census, New York.
  12. Pułaska-Turyna B., 2011, Statystyka dla ekonomistów [Statistics for economists], wyd. Difin, wydanie trzecie, Warszawa [in Polish].
  13. Raport o stanie lasów w Polsce 2014 [Report on the state of forests in Poland 2014]. 2015. Centrum Informacyjne Lasów Państwowych na zlecenie Dyrekcji Generalnej Lasów Państwowych, Warszawa [in Polish].
  14. Ratajczak E., 2011. Popyt na drewno w Polsce – zmiany strukturalne oraz możliwości zaspokojenia [Demand for wood in Poland - structural changes and possibilities of satisfaction], Prezentacja z III Sesji Zimowej Szkoły Leśnej pt. Strategie rozwoju lasów i leśnictwa w Polsce do 2030 roku, Sękocin [in Polish].
  15. Surowiec drzewny, podział terminologia i symbole [Wood raw material, division of terminology and symbols], Polski komitet normalizacji, miar i jakości, PN-93/D-02002 [in Polish].
  16. Ustawa z dnia 28 września 1991 r. o lasach [Forest law], Dz.U. 1991 Nr 101 poz. 444.

Accepted for print: 30.12.2017


Przemysław Suchodolski
Faculty of Economic Sciences, Warsaw University of Life Sciences – SGGW, Poland
161 Nowoursynowska Str.
02-787 Warsaw
Poland
email: przemyslaw_suchodolski@sggw.pl

Marcin Idzik
Faculty of Economic Sciences, Warsaw University of Life Sciences – SGGW, Poland
161 Nowoursynowska Str.
02-787 Warsaw
Poland
email: marcin_idzik@sggw.pl

Responses to this article, comments are invited and should be submitted within three months of the publication of the article. If accepted for publication, they will be published in the chapter headed 'Discussions' and hyperlinked to the article.