CNMI Consumer Price Index: SAIPAN

The 2016 Commonwealth of the Northern Mariana Islands Household Income and Expenditure Survey (HIES) report presents current information on demographic, social and economic characteristics of the Commonwealth of the Northern Mariana Islands (CNMI) population as well as some key expenditure details of what families and residents spent their money on. The survey results will provide our government decision makers as well as our business community leaders the tools they will need to make sound policy decisions. The survey will also allow our leaders to be able to have a better assessment of our communities and the CNMI population from a treasure trove of data points. The 2016 HIES is the third survey of the expenditure series since the first one in 1998. The main objectives of the survey are:

• To provide new expenditure weights and identify items for a new item sample (a basket of goods and services) as well as identifying an outlet sample people purchase their goods and services for the revision of the (CNMI’s) Consumer Price Index.
• To supplement personal consumption expenditure in compilation of National Accounts and Gross Domestic Products.
• To provide population count, population characteristics on labor force, income, employment and other key population and housing characteristics needed for policy making decisions.

Over a six-month period from July to December 2016, the Central Statistics Division (CSD) of the Commonwealth of the Northern Mariana Islands’ Department of Commerce conducted the HIES. This survey used two types of survey methods to gather data.  For the more expensive and less frequently purchased items, such as major appliances, electronic goods, new and used vehicles, etc., a pre-printed, survey questionnaire was completed by an interviewer.  For those items which are more frequently purchased, such as food, tobacco, drinks, etc., the household respondent was given a diary to record all purchases during a two-week period; this was increase from 2005 to capture data more accurately.

The Central Statistics Division (CSD) selected a 10% sample of all occupied housing units on Saipan and 20% of all occupied housing units on Rota and Tinian. The households interviewed were selected with equal probability from the household list on each island using a systematic random sampling technique.  The CSD also selected a substitute list of housing units using the random sampling technique for each of the islands. The substitute housing units were for housing units that were found to be vacant, have been converted to a business or no longer existed.  As is done in previous HIES, the CSD used both a diary and a household questionnaire to gather household income and expenditures data. The diary portion of the survey was left with each of the interviewed households for two weeks.

In December 2016, Dr. Michael J. Levin processed the data with the assistance of Mr. Fermin C. Sakisat and Justin H. Andrew. The CSD hired Mr. Brian Hannon, an International Consultant in Survey Statistics to assist the CSD in revising and then rebasing the CPI. The survey items and expenditures were arrayed using the U.S. Using the Bureau of Labor Statistics CPI classification system as a frame of reference, the CSD arrayed the expenditures data into 9 groups.  Then, using a statistical technique called “Probability Proportional to Size” or PPS; the CSD selected a new item sample.  Using the PPS technique to select the item sample ensured that those items with the highest monthly household expenditures were selected in the sample.  The sample included a total of 158 items for Saipan, 116 for Tinian, and 82 for Rota.

The Major groups retain the nine (9) major groups from the previous CPI Rebase in 2007. Having the same groups allows the users to understand better which groups are responsible for the increases/decreases of the CPI.  Prior to the 2007 CPI Rebase, for example, alcoholic beverages were classified under Health and Recreation Group.  Now, it is one of the 4 new groups.  The others are:  Medical Care, Education and Communication and Other Goods and Services.

The previous CPI collected prices from 170 different retail outlets in Saipan, 30 in Tinian and 60 in Rota. The updated CPI collects prices from 122 different retail outlets in Saipan, 68 in Tinian, and 41 in Rota which are geographically dispersed in Saipan, Tinian, and Rota. This is a 28% decrease in Saipan, a 27% increase in Tinian and a 32% decrease in Rota from the previous CPI.

Mr. Brian Hannon made improvements to the previous Data Collection Manual. It covers the main three aspects of CPI Data Collection. Retail Outlet Initiation, Sample Item Selection Rules, and Sample Item Re-pricing. This enabled CPI to improve data collection quality and accuracy.

In the 2^nd Quarter of 2017, data collectors collected price data for both the old and revised CPIs and the two indexes were “spliced” together at that point.  For two indexes to be linked together, there must be at least one period where prices are collected for both the old and the new indexes.  This overlapping period was the 2^nd Quarter 2017.  Beginning in the 3^rd Quarter 2017, the CSD collected price data for only the items in the revised CPI.  Therefore, up to the 2^nd Quarter 2017, the index change is based on the old basket of goods, whereas from 3^rd Quarter 2017 onwards, the index change is based on the new basket.

Another change in the CPI was rebasing it from the 4^th Quarter 2008 = 100 to the 2^nd Quarter 2017 = 100.  This is a mathematical change in the CPI, but the more recent base period will make the CPI easier for users to understand.  There is no difference in the percent changes from the old CPI from one period to another.

As with most CPIs around the world, the CSD will continue to use the Laspeyres method to calculate the CPI.  This is the most widely-used and internationally-accepted method of calculating price change for a CPI.  The Laspeyres formula is shown below:

Laspeyres Formula:

This method of index computation uses the quantities of commodities purchased in the base period as the basis for computing the value of the “market basket.”  First, there are n number of items in the market basket with each having a base quantity.  Each of these items has a base price, the p0 and current price, pi.  Multiplying p0 with q0 gives the total expenditure for a particular commodity for the base period.  Current expenditure for an item is obtained by multiplying pi with q0.  The sum of current expenditures for all items in the market basket is expressed as ? piq0, while the sum ?p0q0 is the total market basket expenditures in the base period.  Notice that in both summations the term q0 remains constant, which means by using the same quantities only price changes are reflected in two pricing periods.

For those unfamiliar with statistical notation, trying to read and understand the above explanation can be an intimidating experience.  In layman’s terms, the above simply means that CSD data collectors price the same quality of all sample items in the CPI every quarter and the prices for each item in the current quarter are compared to the same prices in the previous quarter. The base period is the 2^nd Quarter 2017 when the CPI was set equal to 100.  That is when the CPIs for each of the 9 major groups and the All Items CPI = 100.

The CSD uses a slight variation to this formula.  Each quarter, the average of the prices that are collected for each item are compared to the average of the same prices in the previous quarter – not the base period.  This is only done for items which are priced in both quarters.  The total of the comparable prices of an item in the current quarter is divided by the total of the comparable prices in the previous quarter to calculate a price relative.  This price relative is multiplied by the item’s weight in the previous quarter to update the item’s weight in the current quarter.  This product is then divided by the item’s weight in the base period to calculate an item’s CPI.   By summing the updated weights for all of the items in a group and then dividing that sum by the same weights in the base period will calculate the group’s current CPI.