A basic consumer price index for Cambodia.
The Cambodia Development Resource Institute (CDRI) has been collecting prices of basic consumer items in twelve provincial markets in Cambodia, nine urban and three rural, since July 1993. It was decided that the CDRI would not collect consumer prices in Phnom Penh, the capital and major urban centre, as the National Institute of Statistics (NIS) was already collecting consumer prices there on a monthly basis. Instead, the CDRI would focus on collecting prices in provincial markets, with the aim of constructing a national price index as well as individual provincial basic price indices (Kannan 1995). Prices were collected for fourteen items: twelve basic consumer items; the riel-U.S. dollar exchange rate; and gold. (1)
This paper has its genesis in a study commenced in 2006 to review the methodology and data used by the CDRI to calculate the basic price indices. That preliminary study uncovered a number of shortcomings in the approach that had been used to calculate the price indices prior to 2006. Those shortcomings were corrected, the existing price data were cleaned and a full data set of the available prices was constructed. Some of the collected price data had been misplaced at various times in the past when staff or computer systems had been changed. Despite some missing data, it proved possible to build a comprehensive provincial price data set going back to 1993. This paper reports the outcome of that study, which led to the construction of a national basic consumer price index for Cambodia from 1993 to 2009. (2)
In making adjustments and corrections to the various provincial basic price indices, this study applied updated expenditure weights as new data became available from the various Cambodian Socio-Economic Surveys (CSES). The individual provincial indices were also weighted by the appropriate population shares--either urban or rural--to construct urban, rural, provincial and national basic price indices. The provincial weights were updated as new population data became available from successive national censuses (1998, 2000 and 2006). Thus, the adjustments and corrections made for the current study applied standard price index techniques to Cambodian data to ensure that the resultant indices were representative of the conditions existing in Cambodia at the time the indices were constructed.
The rest of this paper is organized as follows: the next section discusses some issues in relation to constructing basic price indices. The third and fourth sections discuss, respectively, the expenditure and population weights used in constructing the indices. The fifth section describes the provincial and national indices while the sixth section discusses the construction of a Phnom Penh basic consumer price index using price data from the NIS Consumer Price Index (CPI) for Phnom Penh. This section is followed by some observations on the constructed price indices together with a graphical analysis of price movements over 1993 to 2009. The eighth section outlines some policy implications from the work undertaken before finishing with a concluding section. II. II.
II. Basic Price Index
The ultimate aim of the study was to construct a national basic consumer price index (National Basic CPI) for Cambodia that would be broadly comparable with the NIS Consumer Price Index for Phnom Penh (NIS Phnom Penh CPI). The constructed national index would use the same base periods as the NIS Phnom Penh CPI.
The National Basic CPI was constructed in four steps. First, basic price indices were constructed for each of the twelve markets where consumer prices were collected by the CDRI. These market indices were then combined to construct urban and rural basic consumer price indices according to the type of markets these prices were collected from. In the third step, these urban and rural indices were combined to construct a provincial basic consumer price index (Provincial Basic CPI), with "provincial" referring to all areas outside the urban area of the capital, Phnom Penh. The final step was to combine the Provincial Basic CPI with a basic consumer price index constructed for Phnom Penh (Phnom Penh Basic CPI) using price data from the NIS Phnom Penh CPI to construct a National Basic CPI. (3) In all cases, the individual indices were weighted by the appropriate population shares taken from Cambodian census data.
The items for which prices were collected by the CDRI are basic goods needed for daily living, so the resultant price indices are more properly called basic price indices rather than true consumer price indices. A consumer price index covers most items or groups of items purchased by a typical urban consumer, and the NIS Phnom Penh CPI is of this type. The CDRI price indices are intended to measure the prices of goods used by the masses living in the provinces outside the metropolitan area, and not the relatively more affluent urban dwellers of Phnom Penh.
In July 1993, when the price data collection began, Cambodia was a post-conflict society, under the administration of the United Nations Transitional Authority for Cambodia (UNTAC), which did not leave Cambodia until October 1993. The CDRI was primarily concerned with maximizing the geographic breadth rather than the depth of the coverage, so price collections were undertaken for fourteen items considered crucial to the livelihood of the rural poor. These items covered seven major consumption groups: rice, meat and fish, vegetables, fruit, clothing, domestic cooking fuel and transportation (see Table 1).
The provincial coverage for the CDRI data collection did not follow any particular sampling criteria nor was it systematic; but that approach reflects, in part, the security situation in Cambodia in 1993. Some parts of the country were still under the control of the Khmer Rouge and it was not possible to travel to those areas. The provincial coverage also reflects areas where the CDRI had contacts that were willing to collect prices in local markets.
The NIS Phnom Penh CPI, on the other hand, is a true consumer price index; it covers the range of goods used by the more prosperous urban dwellers living in the capital and is not representative of the goods used or the prices applying to those living outside the capital, especially the rural poor. However, data used in the NIS Phnom Penh CPI can be used to construct a basic consumer price index for Phnom Penh, and this basic price index is also included in this study.
It is not necessary for the NIS Phnom Penh CPI and Phnom Penh Basic CPI to move together, as will be shown below. During the period of rapidly rising food prices in 2008, triggered by rising food prices worldwide, the NIS Phnom Penh CPI increased but not as much as the Phnom Penh Basic CPI. This was because the CPI contains a much larger weighting of non-food and non-fuel items, which were relatively price stable during that period.
The formula used to construct each basic price index is a standard weighted price relative index using expenditure weights:
where: [I.sub.t] price index at time t; i ith good; n number of goods in the index; [w.sub.i] weight of ith good in the index; [p.sub.i0] base price of ith good (i.e. at time 0); and [p.sub.it] price of ith good at time t.
III. Expenditure Weights
The bases of the CDRI basic price indices were changed on each occasion to match the three bases used for the NIS Phnom Penh CPI: 1994; 2000; and 2006. The expenditure weights used by the NIS were taken from the Cambodian Socio-Economic Surveys. Table 2 lists the sources of the expenditure weights and the base price periods. The CDRI indices use the same sources as the NIS for the expenditure weights and the same base price periods are used for the construction of the provincial basic price indices.
For each of the CSES, the results of the expenditure survey were reported slightly differently. Consistency in classification was maintained by examining the questionnaires used in each survey and classifying outputs in a consistent manner, so that the items used by the CDRI matched the categories used in the questionnaires and the outputs. The expenditure data were directly obtained from the NIS for each expenditure survey for the items listed as subcategories in Table 1, which were not reported in the published tables. Phnom Penh presents a special case. For some reporting purposes, the NIS classifies Cambodia's population locations using three broad classifications: Phnom Penh, other urban and rural. This approach is used in reporting the CSES results. For census purposes, Phnom Penh itself is further categorised into urban and rural.
For the CSES 1994, expenditure data was reported in terms of average monthly household expenditure. Table 3 lists the spending on the items used as a proportion of the total spending on all items and for each sub-classification of total spending. Thus for "rural", the items used in the basic price index covered: 68 per cent of all spending; 82 per cent of spending on food, beverages and tobacco; 87 per cent of spending on food alone; and 37 per cent of spending on nonfood items.
For CSES 1999, expenditure data were reported in terms of average monthly per capita consumption. Table 4 lists the spending on the items used as a proportion of the total spending on all items and for sub-classifications of total spending.
For the CSES 2004, expenditure data were reported in terms of household monthly expenditure. Table 5 lists the spending on the items used as a proportion of the total spending on all items and for sub-classifications of total spending.
From May 2008 until January 2009, the Royal Government of Cambodia suppressed the publication of the NIS Phnom Penh CPI. This suppression followed a period of rapidly rising food prices; particularly rice prices. The suppression of the publication of the NIS CPI coincided with the lead-up to national elections held in July 2008, however, the data collection process itself continued. The CDRI, too, continued to collect and publish its provincial price indices during this period. When the NIS re-commenced publication of its CPI in February 2009, the missing data for the months when the publication was suppressed were obtainable from the report of the index components for the previous year, i.e., 2008. As a result, the Phnom Penh price data series is unbroken, and it is possible to construct a comprehensive Phnom Penh Basic CPI.
IV. Population Weights
To build each of the basic consumer price indices (for urban, rural, provincial and national markets) from the Provincial Basic CPI, the individual market indices were weighted by the share of the Cambodian population in that province (either urban or rural). Thus for urban markets, the weight used was that province's urban population as a share of the total urban population of Cambodia (sans the urban population of Phnom Penh municipality). The same approach was used for rural markets but with the inclusion of the rural population in the Phnom Penh municipality. (4) The assumption is that the prices collected by the CDRI from a particular market are representative of prices in the same type of market in the rest of the province.
Two aggregate basic price indices were constructed: an urban index and a rural index. Where data for individual markets were not available, the total weights were adjusted by removing those markets from the weighting system. The need for this occurs particularly with the indices with a 1994 base, when data on some markets are either not available or are available only for some time periods. This problem is not as acute for the indices with 2000 and 2006 bases as price data is consistently available or the missing individual indices can be interpolated. For example, price data were not available for Kompong Speu, a rural market, from January 1999 to March 2000 so the rural weights were adjusted for the non-availability of data for that period.
The Provincial Basic CPI was constructed by weighting the urban index by the "other urban" share of Cambodia's population and weighting the rural index by the "rural" share. As was mentioned above, the approach used was to divide the Phnom Penh municipality's population into its urban and rural census components and, for the purposes of the Provincial Basic CPI, the "rural" population of Phnom Penh was included in that BPI's rural share. (5)
The National Basic CPI was then constructed by weighting the Provincial Basic CPI and the Phnom Penh Basic CPI by their respective population shares. The construction of the Phnom Penh Basic CPI is discussed below. Table 6 shows how the components of the population were used to construct the various basic consumer price indices used in this paper.
At the time this paper was being prepared, data from both the 1998 and 2008 population censuses, divided into urban and rural populations for each province, were available. With two sets of census data, it is possible to interpolate population values for the intervening years, assuming that the rate of change was spread evenly over the inter-censual period. An appropriate provincial population share was used for each index base period to construct the relevant basic price indices. For the 1994 base period, the 1998 census shares were used; while the interpolated population shares were used for the 2000 and 2006 base periods.
This abovementioned census data was also used to weight the components of the urban, rural, provincial and national basic price indices. At the time of the 1998 census, then existing security concerns meant that some parts of the country were not enumerated, namely the rural parts of Preah Sihanouk, Kep and Pailin. The 2004 Cambodian Inter-Censal Population Survey (National Institute of Statistics 2005) estimated that these areas contained 45,000 people. For the purposes of constructing the basic price indices, these 45,000 people were added to the rural population and to the total population for the 1998 census; this adjustment has a marginal impact on the urban-rural split and the Provincial-Phnom Penh split. No account was taken of the undercounting in the 1998 census (estimated at 1.78 per cent) nor of the Cambodian refugees in Thailand (estimated at 60,000). The population weights for each provincial market used to construct urban, rural, provincial and national basic consumer price indices are shown in Table 7.
V. A Provincial Basic Consumer Price Index and a National Basic Consumer Price Index
The work discussed above culminated in the construction of basic consumer price indices for each individual market, either urban or rural, where the CDRI had collected prices of basic commodities. As was mentioned above, these individual price indices were then combined to produce either urban or rural basic consumer price indices, using the provincial population weights shown in Table 7. In a subsequent step, the urban and rural basic consumer price indices were combined to calculate a Provincial Basic CPI using the population shares of the "other urban" and rural populations as weights. This Provincial index was combined with a constructed Basic CPI for Phnom Penh (see below) to produce a National Basic CPI for Cambodia.
VI. A Basic Consumer Price Index for Phnom Penh
Although the CDRI collects prices for only basic commodities, the NIS Phnom Penh CPI is a comprehensive consumer price index currently covering some 225 items, of which 82 are classified as "food" items and 93 are classified as "food, beverages and tobacco". Therefore, calculating a National Basic CPI that includes data from Phnom Penh requires the calculation of a basic consumer price index for Phnom Penh itself.
Each month, until December 2007, when the Phnom Penh CPI was published by the NIS, the published output also included the current prices for the consumer items sampled by the NIS for its CPI. These prices can be used to construct a Phnom Penh Basic CPI. Most items for which prices are collected by the NIS are similar to those collected by the CDRI, with the exception of wax gourds. (6) The items from the NIS CPI data used to construct a Phnom Penh Basic CPI are shown in Appendix 2. (7)
In January 2008, the NIS changed the method of presenting the Phnom Penh CPI. The base of the index was changed to "October to December 2006", and the prices of goods in the index were no longer provided as part of the output. Instead, price relatives for groups of goods were provided. This approach provides a degree of flexibility to the NIS, because the individual goods used in the CPI are no longer fixed, though the weights for the various sub-categories of goods are fixed. For the period subsequent to January 2008, selected sub-indices from the NIS CPI that aligned with the goods prices collected by the CDRI for the provincial market indices were used to calculate a Phnom Penh Basic CPI. The sub-indices used from the NIS CPI post 2008 and their relationship to the items used in the CDRI indices are shown in Appendix 3.
The NIS published prices are available only from October 1994 and the NIS Phnom Penh CPI is based on the average price level in the three months from July to September 1994. Some additional prices were found for earlier periods; these prices had been collected by the Ministry of Commerce and by the National Bank of Cambodia before the NIS took over the compilation of consumer price indices. Taken together, these additional prices covered most but not all of the items entering into the CDRI index. These additional prices were used to extend the basic consumer price index back to January 1993, with appropriate changes in weights to allow for the missing items.
VII. Some Observations
This section discusses some broad trends that emerge from the provincial basic consumer price data through an examination of the various episodes of price changes since 1993, as revealed in the changes in the National Basic CPI. (8) This section concludes with a discussion of the movements in the urban and rural basic consumer price indices, relative to the Phnom Penh Basic CPI.
In examining the movements in the price indices, it needs to be borne in mind that a particular index number at a point in time relates only to the base of that index. Comparing a price index number to its base tells only whether overall prices have risen or fallen relative to the base; if the two time periods are known, then it is also possible to determine the rate of change of prices. It is not possible to make inter-index comparisons even if the indices have the same base.
Inter-index comparisons require a different approach. At the base period, each individual index must be expressed relative to the index chosen as the numeraire to allow for the differences in the initial price levels in each province. Such an approach and an inter-index comparison conclude this section, using the Phnom Penh Basic CPI as the numeraire.
Broad trends in price movements are evident in Figures 1 and 2. The Provincial Basic CPI is a population-weighted average of the urban and rural provincial basic price indices. The data is displayed in two graphs for ease of presentation, as the provincial price index tends to closely follow the rural index, given the overwhelming weight of the rural population. The rural markets have a greater weight in the provincial index because the rural population comprised 81.26 per cent of Cambodia's population, while the "other urban" population (i.e., excluding Phnom Penh municipality's urban population) comprises only 10.32 per cent. The rural markets sample, however, is in provinces that account for only about 24 per cent of the total rural population, while the urban markets are in provinces that account for about 51 per cent of the total non-Phnom Penh urban population.
Figure 3 compares the NIS Phnom Penh CPI with the Phnom Penh Basic CPI constructed by the CDRI.
Figure 3 shows that for Phnom Penh, until the beginning of 2005, the basic price index moved broadly in line with the NIS Phnom Penh CPI. Subsequently, the Phnom Penh Basic CPI has risen faster than the NIS CPI and has remained higher. This is not surprising, because price rises since 2003 have been mainly in food and fuel, items which comprise most of the basic price index. This outcome suggests that the poor have been impacted more by recent price rises, because those rises have mostly been in commodities that feature more highly in their spending. Price movements in the NIS Phnom Penh CPI have been moderated by the stability of the non-food and non-energy items, which form a greater proportion of that index. This same pattern can be seen in Figure 4, which shows the NIS Phnom Penh CPI and the National Basic CPI. Price increases have been more moderate in Phnom Penh.
The outcomes shown in Figure 4 can be seen more starkly in Table 8, which compares changes in selected periods in the National Basic CPI with those in the NIS Phnom Penh CPI for the same periods. For each of the sub-periods, the movements observed in the National Basic CPI have been about twice those observed in the NIS Phnom Penh CPI.
The overall pattern of price changes in Cambodia can be seen in Table 9 and Figure 5, which show movements in the National Basic CPI for July 2003 to December 2009.
The National Basic CPI shows that prices fell from July 1993 to February 1994. This period of falling prices was followed by a lengthy period of price rises from February 1994 until July 1998, when the basic price index rose by 90 per cent, which is an annual rate of almost 16 per cent. These rises were followed by a two-year period of price stability from July 1998 until June 2000, when the index fell by 9 per cent. From June 2000 to October 2000, the index rose by 13 per cent (an annual rate of 45 per cent) and then fell from October 2000 to March 2001 by almost 2 per cent. These rises and falls in the index were followed by a three-year period of stability that lasted until February 2004, though the index rose by 9 per cent over this period.
The changes over 1993 to 2003 can be seen more clearly in Figure 6, which has a different vertical scale from Figure 5, so that the changes in this period are not swamped in the graphical presentation by the extreme changes that occurred in 2008 and 2009.
The period of relative price stability that ended in February 2004 was followed by a sharp rise in the basic price index that lasted until September 2007; the index rose by 85 per cent in this period, an annual rate of 19 per cent. In the period from September 2007 until August 2008, there was an extreme rise in the basic price index, when the index rose by 35 per cent, an annual rate of 38 per cent, and this period was followed by a very
harp fall from August 2008 to January 2009, when the index fell by 19 per cent, an annual rate of 40 per cent, with the basic index ending back at almost the same level as it was before the period of the extreme price rises began. From January 2009 until December 2009, the end of the period covered by this paper, the basic price index has exhibited considerable volatility.
The rise in the National Basic CPI from September 2007 to August 2008 was driven by rises in the prices of rice, fish and gasoline, with the rice price increase being the major driver. From December 2007 until August 2008, the basic price index rose by 23 per cent. Rice, fish and gasoline accounted for 67 per cent of this change, with the contributors to the change being rice, 43 per cent; fish, 16 per cent; and gasoline, 8 per cent. Rice was the major contributor to the index increase over this period, when the world rice price almost doubled as a result of both supply shortages and demand increases.
Figure 7 compares the Phnom Penh Basic CPI to the urban and rural basic price indices.
In Figure 7, all the non-Phnom Penh prices indices are expressed relative to the Phnom Penh price level in the period October to December 2006, when the base of that index was set at 100 (the actual average value was 100.05). For the period October to December 2006, urban basic prices were about 1.5 per cent higher than Phnom Penh prices, i.e., effectively almost the same, using the expenditure required to purchase the goods comprising the index as the comparative measure; the urban basic consumer price index was 101.46. Rural basic prices were about 6 per cent higher than comparable Phnom Penh prices; the rural basic consumer price index was 105.65. Over the whole period, Phnom Penh prices have tended to be lower than either urban or rural prices for basic commodities. From 1993 to November 1997, rural prices were higher than urban prices. After December 1997, urban and rural prices have tended to move together, although since November 2006, rural prices have again tended to be higher than urban prices. The provincial and national basic consumer price indices are not shown in Figure 7, as they tend to mirror the movements in rural prices, given the preponderant weight of the rural population in the composition of both indices.
Figure 8 shows the impact of using the National Basic CPI as a deflator of GDP on per capita GDP to calculate the real per capita GDP, compared with alternative deflators.
The per capita GDP in current U.S. dollar terms grew by 8.14 per cent per annum from 1994 to 2008, whereas for the real per capita GDP, using the implicit GDP deflator with a base of 2006, the per annum growth rate was 0.88 per cent. Using the NIS Phnom Penh CPI as the deflator, the per annum growth rate in real per capita GDP was 1.63 per cent, whereas using the National Basic CPI as the deflator, there was a 1.32 per cent decline in real per capita GDP. This outcome suggests that for the great majority of the population over this fourteen year period, their real income declined while the benefits of the growth in the economy were largely restricted to or appropriated by the more affluent urban dwellers of Phnom Penh.
VIII. Policy Implications
The analysis in this paper addresses the measurement of changes in the general price level. The usual published measure of price inflation in Cambodia is the NIS Phnom Penh CPI, yet this measure covers less than 10 per cent of the population, a relatively affluent portion. The non-Phnom Penh urban population of Cambodia accounts for another 10 per cent of the total population, while the rural population accounts for about 80 per cent. The rural population is, by any objective measure, predominantly poor. Thus the NIS Phnom Penh CPI is unlikely to reflect the price movements impacting the majority of the population and particularly the rural poor.
The initial aim of the CDRI in collecting provincial price data was to measure price movements on a national basis in a post-conflict economy that was subject to bouts of price inflation in order to construct an effective measure of the national price inflation. The focus of the price collection was on basic consumer goods used by everyone, including the poor: food, clothing, household cooking fuel and transport fuel. The original price collections for these items covered about two-thirds of the total spending of the rural population, including over 85 per cent of their spending on food and over 35 per cent of their spending on non-food items, as measured by the 1994 CSES. (9) These price collections have resulted in the construction of a consistent measure of the general price level for basic goods used by the great mass of the population that is predominantly poor. When this measure is combined with a Phnom Penh Basic CPI, constructed using price data extracted from the NIS CPI data for Phnom Penh, a National Basic CPI can be constructed that is more representative of the prices paid by the whole population than is the NIS CPI.
The differences in the two measures can be seen in Figure 6 and Table 7. Up until about 2006, the two indices moved broadly together but subsequent to 2006, when the prices of food and energy-related items increased significantly, the National Basic CPI increased much faster than the NIS CPI and also remained at a higher level. Over the period of very rapid price inflation beginning in August 2006 until the end of the period covered by this paper, December 2009, the CDRI national basic price index rose by 56 per cent, while the NIS Phnom Penh CPI rose by 35 per cent. This very sharp rise in the national basic price index indicates that the poor, particularly the rural poor, suffered a substantial decline in their standard of living as a result of the increases in food and energy prices. While the Phnom Penh prices also rose, the effect on the index was moderated by the greater share of non-food and non-energy items in the Phnom Penh index. The changes in the NIS CPI for Phnom Penh did not fully reflect the impact of the rising prices on the great mass of the population, particularly the rural poor. Policymakers using only the capital city measure of inflation would not see the full extent of the price inflation that was occurring and, thus, they would not be aware of the impact of the price inflation on the great mass of the population.
The lesson to be drawn from this outcome is that to be effective as a policy tool, an inflation measure must cover the whole of the population or as much as is feasible to achieve a representative measure of the prices faced by the bulk of the population. The issue here is one of coverage, not the quality of the data used or of the technical underpinnings of the NIS Phnom Penh CPI itself. The NIS CPI is a well-constructed price index but it only covers a relatively affluent 10 per cent of the total Cambodian population. The CDRI national basic price index covers about 50 per cent of the non-Phnom Penh urban population and about 25 per cent of the rural population, so it is much more representative of the total population than the NIS CPI. This is not to say that the geographic coverage of the CDRI national basic price index is fully representative; it could be improved. The initial geographic coverage of the CDRI index was dictated by the ability of the CDRI to locate informants who were willing to collect consumer prices on a regular basis, as well as the internal security situation in Cambodia; at that time large areas of the country were not accessible.
The measurement of price inflation is a cumulative process. Price data needs to be collected in good times and bad; prices cannot be collected in times of rapidly rising prices and then abandoned in times of price stability or relative price stability. The present levels of prices can only be interpreted in the light of the past price history. Both the NIS CPI and the CDRI national basic price index meet this criterion; they both contain price data going back to the early 1990s. The benefit of this relatively long run of data became apparent when rapid price rises began after August 2006 and it was possible to place those rises in the context of what had gone before.
At times when prices are relatively stable, cost-cutting moves may be initiated by some interest groups in an attempt to cut back on the relatively resource-intensive basic data collections needed to construct comprehensive price indices. Giving in to such pressures and cutting back on data collections are counter-productive in the long term. If price data are not collected in the field at the time, they cannot be recreated later, thus creating a gap in the price series, and all future policy analysis is affected by this break in the data. Good data collection protocols require that consumer price data be collected regularly and frequently if measures of consumer price inflation are to have enduring meaning. This point was demonstrated very starkly in 2008 when the publication of the NIS CPI was suppressed. During that period, the CDRI was the only source of price index data in Cambodia. When the publication of the NIS CPI was suppressed, it was not known how long the suppression of the official data would last nor whether the past data would be released when publication resumed. When publication recommenced, these points were not a concern but if publication had not been resumed or if past data had not been available, there would have been a serious gap in the Cambodian price inflation data series.
This paper has used basic consumer price data gathered in provincial markets, both urban and rural, by the CDRI since 1993 to construct a National Basic CPI for Cambodia. A basic price index uses only a limited range of commodities, in this case twelve commodities, and is not intended to be a fully comprehensive consumer price index. Nonetheless, this index gives an indication of the movements in prices paid by the poor for basic commodities over this period.
Basic consumer price indices were constructed for nine urban provincial markets and three rural provincial markets using price data collected monthly by the CDRI. These basic price indices for each provincial market were used to construct urban and rural basic consumer price indices, according to the location of the market where the prices were collected. In this context, "urban" means "other urban", i.e., urban areas located outside the Phnom Penh municipality. The urban and rural basic consumer price indices were combined to produce a provincial basic consumer price index for Cambodia, with "provincial" referring to all areas outside of the NIS-defined urban portion of the Phnom Penh municipality. The population of the "rural" portion of Phnom Penh municipality was included with the provincial rural areas for the purposes of constructing the provincial basic price index. In producing the composite price indices from the individual market indices, the components were weighted by their appropriate population share, i.e., either urban or rural, of each province in the total Cambodian population.
A basic consumer price index was also constructed for the urban portion of Phnom Penh municipality using NIS price data for the same twelve basic commodities. Finally, the Phnom Penh basic consumer price index and the provincial basic consumer price index were combined, again using appropriate population weights, to produce a National Basic CPI.
All the constructed indices use the current NIS CPI base of October to December 2006 to report price movements over the period 1993 to 2009. Basic price indices that used the earlier base periods of 1994 and 2000 were incorporated into linked indices using the 2006 base.
The paper includes a brief graphical analysis and commentary that discusses the movements of all the basic indices over the period 1993 to 2009. The analysis includes a discussion of the relationship of the provincial basic price indices to the Phnom Penh Basic CPI. The paper concludes with a brief discussion of the policy implications of the analysis, noting that the NIS CPI for Phnom Penh does not give an adequate coverage of the total population of Cambodia, particularly the rural poor, and that the CDRI national basic consumer price index constructed in this paper is a better measure of the underlying price inflation in Cambodia for this period.
Appendix 1 Expenditure weights Expenditure Weights: 1994 Base Phnom Penh Item Item Unit Weight No. 1 Riel/US$ exchange rate riels/US$ 2 Gold riels/chi 3 Rice (Phkar Knhei) riels/kg 0.13307 4 Meat, Poultry, Fish a. Fresh fish (Trey Ros) riels/kg 0.11337 b. Pork 1st quality riels/kg 0.14354 c. Beef 1st quality riels/kg 0.03301 5 Palm sugar riels/kg 0.07077 6 Vegetables a. Cucumbers riels/kg 0.03941 b. Wax gourds riels/gourd 0.03526 7 Bananas (Namva) riels/hand 0.08610 8 Thai sarong riels/sarong 0.07368 9 Fuel a. Charcoal riels/sack 0.07717 b. Kerosene riels/l 0.01452 10 Gasoline riels/l 0.18010 1.00000 Other Urban Rural Item Weight Weight No. 1 2 3 0.18864 0.25551 4 a. 0.16545 0.18720 b. 0.13501 0.11747 c. 0.02161 0.02013 5 0.08196 0.09626 6 a. 0.05439 0.04530 b. 0.04548 0.04255 7 0.08261 0.05836 8 0.05510 0.05850 9 a. 0.05801 0.05091 b. 0.01165 0.01300 10 0.10008 0.05483 1.00000 1.00000 Source: Derived from CSES1994; NIS 1995. Expenditure Weights: 2000 Base Phnom Penh Item Item Unit Weight No. 1 Riel/US$ exchange rate riels/US$ 2 Gold riels/chi 3 Rice (Phkar Knhei) riels/kg 0.19002 4 Meat, Poultry, Fish a. Fresh fish (Trey Ros) riels/kg 0.11639 b. Pork, 1st quality riels/kg 0.13858 c. Beef, 1st quality riels/kg 0.02057 5 Palm sugar riels/kg 0.04038 6 Vegetables a. Cucumbers riels/kg 0.03815 b. Wax gourds riels/gourd 0.02836 7 Bananas (Namva) riels/hand 0.06651 8 Thai sarong riels/sarong 0.04988 9 Fuel a. Charcoal riels/sack 0.04276 b. Kerosene riels/l 0.05226 10 Gasoline riels/l 0.21615 1.00001 Other Urban Rural Item Weight Weight No. 1 2 3 0.28860 0.34663 4 a. 0.15993 0.16258 b. 0.12475 0.10962 c. 0.01496 0.01461 5 0.05331 0.06442 6 a. 0.05372 0.05917 b. 0.04371 0.04819 7 0.07537 0.05521 8 0.04596 0.04448 9 a. 0.04228 0.05368 b. 0.03309 0.01687 10 0.06434 0.02454 1.00002 1.00000 Source: Derived from CSES 1999; NIS (2000). Expenditure Weights: 2006 Base Phnom Penh Item Item Unit Weight No. 1 Riel/US$ exchange rate riels/US$ 2 Gold riels/chi 3 Rice (Phkar Knhei) riels/kg 0.22508 4 Meat, Poultry, Fish a. Fresh fish (Trey Ros) riels/kg 0.14357 b. Pork 1st quality riels/kg 0.12833 c. Beef 1st quality riels/kg 0.02841 5 Palm sugar riels/kg 0.03687 6 Vegetables a. Cucumbers riels/kg 0.02738 b. Wax gourd riels/gourd 0.04455 7 Bananas (Namva) riels/hand 0.08088 8 Thai sarong riels/sarong 0.05252 9 Fuel a. Charcoal riels/sack 0.02136 b. Kerosene riels/l 0.11204 10 Gasoline riels/l 0.09900 1.00000 Other Urban Rural Item Weight Weight No. 1 2 3 0.24421 0.29870 4 a. 0.17843 0.19648 b. 0.10172 0.09583 c. 0.02307 0.02550 5 0.05180 0.06501 6 a. 0.03479 0.03730 b. 0.04393 0.05294 7 0.06587 0.05388 8 0.04661 0.04208 9 a. 0.04254 0.05635 b. 0.06442 0.02718 10 0.10261 0.04875 1.00000 1.00000 Source: Derived from CSES 2004 data supplied by NIS. Appendix 2 Phnom Penh Basic Consumer Price Index: Commodities Used in 1994 and 2000 Base Indices Items used in the Phnom Penh basic consumer price indices with 1994 and 2000 bases compared to those used in the CDRI provincial basic consumer price indices Item Items Unit No. 1 Riel/US$ exchange rate riels/US$ 2 Gold riels/chi 3 Rice (Phkar Knhei) riels/kg 4 Meat, Poultry, Fish a. Fresh fish (Trey Ros) riels/kg b. Pork, 1st quality riels/kg c. Beef, 1st quality riels/kg 5 Palm sugar riels/kg 6 Vegetables a. Cucumbers riels/kg b. Wax gourds riels/gourd 7 Bananas (Namva) riels/hand 8 Thai sarong riels/sarong 9 Fuel a. Charcoal riels/sack b. Kerosene riels// 10 Gasoline riels// Description Items Food 8 Food, beverages & tobacco 8 All items 12 NIS Phnom Penh CPI 1994 base Item Name Unit No. 002 Rice Quality #2 kg 015 Mud fish (large) kg 009 Pork without fat kg on Beef No. 1 kg 064 Brown sugar kg 039 Cucumbers kg 041 Ridge gourds kg 053 Bananas hand (medium) 091 Sarong piece 132 Charcoal kg 130 Liquid fuels litre (kerosene) 171 Gasoline litre Items Description 78 Food 87 Food, beverages & tobacco 206 All items NIS Phnom Penh CPI 2000 base Item Name Unit No. 002 Rice Quality #2 kg 015 Mud fish (large) kg 009 Pork without fat kg Oil Beef No. 1 kg 073 Brown sugar kg 042 Cucumbers kg 044 Ridge gourds kg 058 Bananas hand (medium) 100 Sarong piece 145 Charcoal kg 143 Liquid fuels litre (kerosene) 192 Gasoline litre Items Description 87 Food 96 Food, beverages & tobacco 227 All items Appendix 3 Phnom Penh Basic Consumer Price Index: Commodities Used in 2006 Base Index Items used in the Phnom Penh basic consumer price index with a 2006 base, compared to those used in the CDR1 provincial basic consumer price indices CDRI Provincial Indices NIS Phnom Penh CPI 2006 base Item Items Item Item ID name No. ID 1 Riel/US$ exchange rate 2 Gold 3 Rice (Phkar Knhei) 101101 Rice, quality no. 2 White, Neang-Menh, Battambang 4 Meat, Poultry, Fish a. Fresh fish (Trey Ros) 118001 Mud fish (large) b. Pork, 1st quality 109001 Pork without fat c. Beef, 1st quality 110001 Beef no. 1 5 Palm sugar 145101 Brown sugar 6 Vegetables a. Cucumbers 137501 Cucumbers b. Wax gourds 137201 Ridge gourds 7 Bananas (Namva) 128001 Banana 8 Thai sarong 304003 Sarong 9 Fuel a. Charcoal 330101 Charcoal b. Kerosene 329001 Liquid fuels (kerosene) 10 Gasoline 506001 Gasoline CDRI Provincial Indices NIS Phnom Penh CPI 2006 base Item Items Coicop Coicop name No. 1 Riel/US$ exchange rate 2 Gold 3 Rice (Phkar Knhei) 01.1.1.1 Rice 4 Meat, Poultry, Fish a. Fresh fish (Trey Ros) 01.1.3.1 Fish (fresh) b. Pork, 1st quality 01.1.2.1 Pork (fresh) c. Beef, 1st quality 01.1.2.2 Beef (fresh) 5 Palm sugar 01.1.8.1 Sugar, jam, honey, chocolate and confectionery 6 Vegetables a. Cucumbers 01.1.7.2 Fruit vegetables b. Wax gourds 01.1.7.2 Fruit vegetables 7 Bananas (Namva) 01.1.6.1 Fresh fruits 8 Thai sarong 03.1.1.1 Clothing materials 9 Fuel a. Charcoal 04.5.4.1 Solid fuels b. Kerosene 04.5.3.1 Liquid fuels 10 Gasoline 07.2.2.1 Fuels and lubricants for personal transport equipment CDRI Provincial Indices NIS Phnom Penh CPI 2006 base Item Items Used Unique No. category 1 Riel/US$ exchange rate 2 Gold 3 Rice (Phkar Knhei) Item Unique 4 Meat, Poultry, Fish a. Fresh fish (Trey Ros) Coicop b. Pork, 1st quality Coicop c. Beef, 1st quality Coicop Unique 5 Palm sugar Coicop 6 Vegetables a. Cucumbers Coicop b. Wax gourds Coicop 7 Bananas (Namva) Coicop 8 Thai sarong Coicop 9 Fuel a. Charcoal Coicop b. Kerosene Coicop Unique 10 Gasoline Item Unique
Carpenter, Keith, assisted by Pon Dorina. "A Basic Consumer Price Index for Cambodia". Working Paper Series, no. 67. Phnom Penh: Cambodia Development Resource Institute, 2012.
Kannan, K.P., assisted by Be Kalyanna, Long Piseth and Chao Kimthy. "Construction of Consumer Price Index for Cambodia: A Review of Current Practices and Suggestions for Improvement". Working Paper 1. Phnom Penh: Cambodia Development Resource Institute, 1995.
NIS. Consumer Price Index, Phnom Penh, monthly. Phnom Penh: National Institute of Statistics, Ministry of Planning, various dates.
--. "Report on the Socio-Economic Survey of Cambodia 1993/94" (All Rounds). Phnom Penh: National Institute of Statistics, Ministry of Planning, 1995.
--. "Report on the Cambodia Socio-Economic Survey 1999". Phnom Penh: National Institute of Statistics, Ministry of Planning, 2000.
--. Kingdom of Cambodia: Statistical Year Book 2003. Phnom Penh: National Institute of Statistics, Ministry of Planning, 2003.
--. "Cambodia Inter-Censal Population Survey 2004: Demographic Estimates and Revised Population Projections". Phnom Penh: National Institute of Statistics, Ministry of Planning, 2005.
Keith Carpenter is an economic and financial analyst based in Sydney, N.S.W. Australia.
The author wishes to thank Dr Brett Ballard, a former acting research director at the CDRI, for his suggestion and encouragement to undertake this study, and Dr Hossein Jalilian, a former research director at the CDRI, for his ongoing encouragement to persist with the completion of the research on which this paper is based. Thanks are also due to Mrs Pon Dorina for her assistance with the provision of the provincial price data; Mrs Pon was, for many years, responsible for organizing the collection and storing of the provincial markets' price data at the CDRI.
(1.) The exchange rate and the gold price were not used in calculating the basic price indices.
(2.) For details of the study and the changes made to correct the shortcomings of the initial work, see Carpenter (2012).
(3.) Basic price indices were calculated for this study because prices of basic items were all that were available.
(4.) There may be some objection to this approach, as it could be argued that the prices in the rural parts of Phnom Penh municipality on the outskirts of the city are more likely to be influenced by prices in urban Phnom Penh than by prices in the provinces, either urban or rural. A judgement was made for this paper that the rural population of Phnom Penh more properly belongs with the provincial rural population, when constructing a Provincial Basic CPI, so that the municipality's rural population shares the pricing characteristics of the other rural areas. This decision also affects the construction of the National Basic CPI.
(5.) See note 4.
(6.) See Carpenter (2012), p. 12 for a discussion of the treatment of this point.
(7.) A complete set of index data and other appendices can be obtained from the author.
(8.) A fuller discussion of the index data is given in Carpenter (2012), p. 12ff.
(9.) See Table 2. See Tables 3 and 4 for the corresponding shares for later periods.
TABLE 1 Consumer Prices Collected and Measurement Units Item Item Unit Number 1 Riel/US$ exchange rate3 riels/US$ 2 Gold (a) riels/chi 3 Rice (Phkar Knhei) riels/kg 4 Meat, Poultry, Fish a. Fresh fish (Trey Ros) riels/kg b. Pork 1st quality riels/kg c. Beef 1st quality riels/kg 5 Palm sugar riels/kg 6 Vegetables a. Cucumbers riels/kg b. Wax gourds (also known as Bottle Gourd riels/gourd or Winter Melon) 7 Bananas (Namva) riels/hand 8 Thai sarong riels/sarong 9 Fuel a. Charcoal riels/sack (b) b. Kerosene riels/l 10 Gasoline riels/l NOTES: (a.) Riel/US$ exchange rate and gold price were not used in the construction of the price indices. (b.) For some markets, at some time periods, charcoal prices were reported in riel per kilogram. Phnom Penh charcoal prices are reported in this way. Per kilogram prices were converted to per sack prices using the conversion rate of 40 kg per sack. TABLE 2 NIS Expenditure Weights and Base Periods3 Index base Expenditure Base Price Periods Weights Source 1994 CSES 1994 July-September 1994 2000 CSES 1999 July-December 2000 2006 CSES 2004 October-December 2006 NOTE: (a.) The results from the CSES 1994 and CSES 1999 have been published, under these headings: "Phnom Penh", "other urban" and "rural"; but the full results from the CSES 2004 have not been published (National Institute of Statistics 1995; National Institute of Statistics 2000). The appropriate expenditure data from the CSES 2004 were provided to the CDRI by the NIS to allow for the provincial price indices to be constructed on a 2006 base. Spending weights were constructed for the provincial basic consumer price indices from the expenditure survey data. The expenditure patterns under the heading "other urban" were used for provincial urban markets, while the patterns under the heading "rural" were used for rural markets. The expenditure weights used are detailed in Appendix 1. TABLE 3 CSES 1994--Expenditure Weights: Coverage of Items as Proportion of Total Spending on Those Items (%) Spending Cambodia Phnom Other Rural Penh Urban All items 62.84 49.18 55.54 67.68 Food, beverages & tobacco 79.33 66.84 75.02 82.30 Food only 84.21 71.97 80.84 86.94 Non-food 34.93 32.77 29.30 37.09 TABLE 4 CSES 1999--Expenditure Weights: Coverage of Items as Proportion of Total Spending on Those Items (%) Spending Cambodia Phnom Other Rural Penh Urban All items 51.86 37.93 47.78 57.71 Food, beverages & tobacco 67.44 57.42 64.58 69.99 Food only 73.22 64.20 70.99 75.38 Non-food 27.37 25.14 24.20 29.83 TABLE 5 CSES 2004--Expenditure Weights: Coverage of Items as Proportion of Total Spending on Those Items (%) Spending Cambodia Phnom Other Rural Penh Urban All items 49.17 42.42 52.45 64.31 Food, beverages & tobacco 76.28 73.63 75.57 82.48 Food only 90.06 89.25 90.00 91.58 Non-food 23.85 20.55 27.78 31.48 TABLE 6 Components of Basic Consumer Price Indices Census Categories Basic Price Indices Urban Rural Provincial Phnom National Penh Urban less Phnom X X X Penh Urban Phnom Penh Urban X X Rural less Phnom X X X Penh Rural Phnom Penh Rural X X X 1998 Census Shares 10.67 84.36 95.03 4.97 100.00 (%) 2008 Census Shares 10.23 80.49 90.72 9.28 100.00 (%) TABLE 7 Urban and Rural Population Shares for each Index Base (%) Population share Index base 1994 2000 Population base 1998 2000 Population share source 1998 Interpolation Census Province Number Province Type 1994 2000 02 Battambang Urban 7.79 7.62 03 Kompong Cham Urban 2.53 2.93 04 Kompong Chhnang Urban 2.32 2.19 06 Kompong Thom Urban 3.68 Not included 07 Kampot Urban 1.84 1.85 10 Kratie Urban 4.41 3.80 14 Prey Veng Urban 3.07 2.71 16 Ratanakkiri Urban 0.95 0.91 17 Siem Reap Urban 6.66 6.66 20 Svay Rieng Urban 1.18 1.08 05 Kompong Speu Rural 5.75 5.83 08 Kandal Rural 10.50 10.38 21 Takeo Rural 7.75 7.74 12 Phnom Penh Urban 4.97 5.83 Share of Urban less Phnom Penh Urban 10.67 10.58 total Rural 84.36 83.59 population Provinces 95.03 94.17 Phnom Penh Urban 4.97 5.83 Check 100.00 100.00 Population share Index base 2006 Population base 2006 Population share source Interpolation Province Number Province Type 2006 02 Battambang Urban 7.09 03 Kompong Cham Urban 4.12 04 Kompong Chhnang Urban 1.78 06 Kompong Thom Urban 07 Kampot Urban 1.85 10 Kratie Urban 1.98 14 Prey Veng Urban 1.63 16 Ratanakkiri Urban 0.78 17 Siem Reap Urban 6.66 20 Svay Rieng Urban 0.76 05 Kompong Speu Rural 6.07 08 Kandal Rural 10.03 21 Takeo Rural 7.71 12 Phnom Penh Urban 8.42 Share of Urban less Phnom Penh Urban 10.32 total Rural 81.26 population Provinces 91.58 Phnom Penh Urban 8.42 Check 100.00 TABLE 8 Comparing National Basic and NIS Price Indices--Price Changes: August 2006 to December 2009 (%) Period Change From To National Basic NIS CPI Phnom Penh Aug 2006 Jun 2008 85.24 39.80 Jun 2008 Feb 2009 -18.47 -9.46 Feb 2009 Dec 2009 3.55 6.90 Aug 2006 Dec 2009 56.40 35.30 TABLE 9 National Basic Consumer Price Index--Price change episodes: July 1993 to December 2009 (%) Description Begin End Months Index Annualized change rate Falling Jul 1993 Feb 1994 7 -26.40 -10.59 Rising Feb 1994 Jul 1998 53 89.55 15.60 Stability Jul 1998 Jun 2000 23 -9.12 -1.86 Rising Jun 2000 Oct 2000 4 13.23 45.05 Falling Oct 2000 Mar 2001 5 -1.69 -4.04 Stability Mar 2001 Feb 2004 35 9.08 3.02 Sharp rise Feb 2004 Sep 2007 43 85.14 18.77 Extreme rise Sep 2007 Aug 2008 11 34.36 37.99 Extreme fall Aug 2008 Jan 2009 5 -19.13 -39.77 Volatility Jan 2009 Dec 2009 11 3.55 3.89 TABLE 10 Real GDP per capita, 1994-2008, with Alternative Deflators Growth p.a. (a) GDP per capita 8.14% Real GDP per capita 2006 base 0.88% Real GDP per capita NIS CPI 1.63% Real GDP per capita National Basic -1.32% Note: a. 1994-2008.
|Printer friendly Cite/link Email Feedback|
|Publication:||Journal of Southeast Asian Economies|
|Date:||Apr 1, 2014|
|Previous Article:||Exploring the nexus between ICT, remittances and economic growth: a study of Vietnam.|
|Next Article:||Informal credit in Vietnam: a necessity rather than an evil.|