Do we gain from collecting price data more than once per month?

1
Do we gain from collecting price data more than
once per month?

• Nicholas N.N. Nsowah-Nuamah
• AND
• Edward Asuo Afram

2
Introduction

• NATIONAL STATISTICS OFFICES (NSO) THAT COMPUTE
  INFLATION NEED TO COLLECTS MARKET PRICES. IT IS A
  MAJOR DATA REQIREMENT

3
Market Prices collection

• NSO NEED TO DECIDE WHETHER TO COLLECT PRICES
• ONCE A MONTH OR
• TWICE A MONTH

4
Advantages of once-a-month market reading

• Reduces cost in terms of money, human resources
  and time
• It also enables the release of inflation on time

5
Advantages of twice-a-month market reading

• Inflation is less vulnerable to short-term price
  fluctuations and better reflects the average
  monthly prices
• Increases precision

6
Short notes on collection of Market Prices in
Ghana

• Started from March 1963 and had been one-a-month
  (2nd week).
• Covers the whole country, both urban an rural
• Since May 2000 has been twice-a-month (1st and
  3rd week)

7
GSS experience

• GSS experience on twice-market-price collection
  in Ghana is that
• supervision is weak during the second price
  collection
• High cost involved
• There is pronounced interviewers and interviewees
  fatigue.

8
Pertinent questions

• Has the GSS gained from collecting price data
  twice in a month instead of once, since 2000?

9
Objective of the paper

• The objective of this paper is to assess the
  deviation of the CPI based on collection of
  prices once or twice per month so as to enable us
  make an informed decision on the number of price
  collection.

10
Benefits

• The analysis provides a statistical/scientific
  basis for the GSS to take a decision on whether
  to discontinue the two times data collection and
  as to which of the weeks (first or third) to
  undertake CPI price collection.

11
Methodology

• The task was to calculate the rate of inflation
  for year 2007 using the two sets of data and
  comparing corresponding levels to see whether
  there are large differences. The work centred on
  the following
• 1. Assembling the two sets of data
• 1st week data collection for 2006 and 2007
• 3rd week data collection for 2006 and 2007
• Calculating inflation rate from the 1st week
  prices data with 2002100
• Calculating inflation rate from the 3rd week
  prices data with 2002100
• Comparing the results with the existing method of
  twice-a-month collection
• Perform test of analysis for the significance of
  the inflation figures

12
Results
Table 1 CPI ands rate of inflation using first and second readings CPI reading Table 1 CPI ands rate of inflation using first and second readings CPI reading Table 1 CPI ands rate of inflation using first and second readings CPI reading Table 1 CPI ands rate of inflation using first and second readings CPI reading Table 1 CPI ands rate of inflation using first and second readings CPI reading Table 1 CPI ands rate of inflation using first and second readings CPI reading Table 1 CPI ands rate of inflation using first and second readings CPI reading
  CPI CPI CPI Rate of inflation Rate of inflation Rate of inflation
  1st 3 weeks 1st week 3rd week 1st 3 weeks 1st week 3rd week
Jul-05 206.08 209.4 203 10.89 11.99 9.77
Aug-05 208.62 211.56 206 10.42 11.37 9.46
Sep-05 211.28 214.07 208 10.19 10.92 9.45
Oct-05 214.55 215.83 213 10.50 9.99 11.02
Nov-05 218.41 218.34 218 11.02 9.57 12.51
Dec-05 220.02 220.92 219 10.69 10.89 10.48
Jan-06 222.59 223.48 222 10.14 8.97 11.35
Feb-06 223.02 223.58 222 10.41 9.72 11.11
Mar-06 222.52 224.05 221 10.19 9.95 10.44
Apr-06 222.15 226.17 218 10.14 10.16 10.13
May-06 225.69 232.19 219 11.40 11.52 11.28
Jun-06 229.79 236.36 223 12.75 12.14 13.40
13
Results cont. graph of CPI
14
Results cont. graph of CPI
15
Results test of hypothesis
ANOVA ANOVA ANOVA ANOVA ANOVA ANOVA
Inflation          
  Sum of Squares df Mean Square F Sig.
Between Groups 0.43 2 0.215 0.213 0.809
Within Groups 33.249 33 1.008    
Total 33.679 35      
16
Conclusion

• From the test results, we conclude that
• There is not much to gain from the two readings
  vis-à-vis cost and quality of data
• A single reading would be cost effective.
  Firstly, cost of data collection will be reduced
  by fifty percent. Secondly, quality data can be
  guaranteed from a smaller investment. Supervision
  will be more effective.

17
Conclusion (Contd)

1. The single reading be taken in the middle of the
   month
2. However, owing to the volatility in the price of
   the petroleum products, we suggest that petroleum
   prices should continue to be taken twice or more.