HOW GOOD (OR BAD) ARE ECONOIC STATISTICS IN THE CARIBBEAN SECOND HIGH-LEVEL ADVOCACY FORUM ON...

HOW GOOD (OR BAD) ARE ECONOIC STATISTICS IN THE

CARIBBEAN

SECOND HIGH-LEVEL ADVOCACY FORUM ON STATISTICS

Radisson Grenada Beach ResortMay 26, 2014

Marc Prud’HommeReal Sector Statistics Advisor

Caribbean Regional Technical Assistance Centre

What is CARTAC

The strategic goal of the CARTAC macroeconomic statistics module is to:

Enhance the quality of macroeconomic statistics by promoting the use of internationally accepted concepts and statistical methodologies.To assist countries in developing, compiling, and disseminating national statistics in line with internationally recognized “best practices”.

IMF Statistics Department JVI13.11: 2

How

CARTAC STA provides technical assistance (TA) and training to most Caribbean countries for improving national accounts, price and international trade statistics.The key focus areas are:

Training and implementation of SUT, Expenditure-based and Production-based GDP, various price statistics, Balance of trade statistics, and so on. Assist in developing and finding new data sources.While maintaining robust programs for compiling and releasing GDP by production work would begin on identifying or prioritizing production of indicators for quarterly GDP price and volume compilation and dissemination.Promote regular rebasing of the national accounts, the CPI and other price statistics.


Why (Quality) Economic Data (Intelligence) Are Important

Economic data do 4 things:① Tells us if we are making progress or not.② It identifies the kinds of changes that make

things better for us as an individual and as a society.

③ When government policy is being assessed, it informs us about its effectiveness.

④ It helps policymakers make informed choices (not necessarily the right choices) that lead to an improved allocation of our scarce resources and higher standards of living.


How Badly Do We Need “Good Data”“Public servants risk becoming policy dinosaurs”

Technology and big data are turning the world of policy-making on its ear.The public service can no longer rely on traditional sources of “structured” and “cleansed” data produced by the likes of a downsized Statistics Canada to advise ministers in a world flooded with massive amounts of unfiltered information and less reliable data (sound familiar?).Ministers want information and advice faster and if the public service drags its feet because of outdated methods, tools and attitudes, the government will look elsewhere.Public servants have to get out of the “Ottawa bubble,” re-think how to analyze and manipulate data and speed up internal approval processes to get advice to ministers faster.


“Gresham’s Law” of Statistics

“Bad data will drive out good data”


What are Good Data?

RelevantAccurateTimely and punctualAccessibleInterpretableCoherentCredible


What is the old Engineering adaga?

What are Good Data?

As the economies grow more complex with time, the need for more and better economic data has never been as pressing.Paradoxically… ironically… unfortunately… government budgets are being squeezed including those of NSOs.


What are Good Data?

The (Charles) Babbage thesis?

“Errors using inadequate data are much less than those using no data at

all”


• More and more, a clear relationship is starting to emerge between reliable economic data and:• Foreign Investment• Economic Growth

10

What are Good Data?

CEF11.06_L6: 11

Real World Examples of Bad Data

IMF Statistics Department

CEF11.06_L6: 12



Nigeria’s economy largest in Africa as rebasing boosts GDP to $405bn!


Argentina says 2013 inflation nearly 11 percent, but… … analysts say it's more like 28 percent!

13IMF Statistics Department

CEF11.06_L6: 14



“That sounds minor—until you calculate that a 1% monthly inflation rate works out at an annual 12.7%, whereas 1.9% monthly compounds to 25.3%.”

“One was told that since the cost of domestic service was “a wage, not a price”, he should not have included it in his CPI calculations.”

CEF11.06_L6: 15



But, one thing that hasn't changed is that there are more questions than ever over the reliability of Chinese figures.

What is China's 'Li Keqiang Index'?

Electricity usage.

The amount of cargo freight.

Bank loans.

Doing so in just over 2 weeks is a fairly remarkable feat for any country, much less one with 1.3 billion people. That's a lot of data to process in a rather short period of time.

African Survey by the AfDB

Situational Analysis of the Reliability of Economic Statistics in Africa: Special Focus on GDP Measurement (2013) This report summarizes the situation regarding national accounts statistics in Africa. It shows

The availability of the basic GDP estimates.Considers the availability of survey data and price statistics from which the national accounts are derived.Describes how far African countries are able to follow the international guidelines given in the System of National Accounts (SNA).


The Motivation For the Survey

During 2012, concerns were raised in the international press about the quality of statistics in Africa and particularly about GDP estimates. The reliability of economic statistics is a crucial concern now that African development is gathering pace and foreign investors need reassurance that they can rely on African statistics.


The reliability of economic statistics in Africa: GDP Measurement

Forty-three of the 44 African countries that responded are now regularly publishing estimates of current price GDP from both the production and expenditure sides and almost all publish GDP at constant prices. Most countries publish GDP estimates with delays of only one or two years. This represents a significant improvement compared with the situation pertaining a decade ago, when many countries had not published GDP estimates for several years.

18


To obtain GDP at constant prices, the SNA 1993 recommends the use of annual chain indices, which means in effect updating the base year each year. If countries are unable to use chain indices, the SNA 1993 recommends that the base year be updated every five years.Deflation using price indices is the preferred method for calculating GDP at constant prices.

19


It would be reasonable to surmise that middle-income countries in Africa would have more reliable national accounts than poorer countries. However, our research suggests otherwise: rich countries are not guaranteed to have good statistics; neither are poor countries condemned to have bad data. Some poor countries appear to give high priority to their national accounts statistics while some richer ones have low-quality statistics.

Achieving high-quality data entails a political choice and a firm commitment to invest in statistics that will support informed evidence based decisions.

20

The AFDB Report

Note that the results for the Caribbean are still work in progress. Not to be quoted or reproduced.

21

Publication of Basic GDP Estimates

22

Basic GDP estimates: Questions posed Countries responding Yes

Africa%

Caribbean%

Do you publish annual estimates of GDP by kind of activity at current prices? (Production approach)

98 95

Do you publish annual estimates of GDP by kind of activity at constant prices? (Production approach)

98 95

Do you publish annual estimates of final expenditure on GDP at current prices? (Expenditure approach)

93 90

Do you publish annual estimates of final expenditure on GDP at constant prices? (Expenditure approach)

82 20

Base Year of GDP Estimates at K Prices

23

Supply Use Tables (Africa)

The AfDB has 54 countries44 countries responded to the survey.33 countries have compiled SUT since 2000• 61 % of the 54 countries or• 75 % of the 44 countries responding.

14 countries compile them every year.When the SUT becomes available, it is normal practice in Africa to revise the previously released GDP estimates.

24

Supply Use Tables (Caribbean)

In the Caribbean:45% of countries have compiled an SUT since 20001 (5%) country last compiled an SUT in 19991 (5%) country last compiled an SUT in 1987. 1 (5%) country compiles one every year.45% of countries have never compiled an SUT.

25

Base Year of SUT (Caribbean)

26

How closely do countries follow 93 SNA?

27

Adherence to new features of the SNA 1993: questions posed

Countries responding Yes

Africa%

Caribbean%

Do you treat expenditure on computer software by producers as capital formation? 59 30

Do you treat mineral exploration as capital formation?61 15

Do you treat government expenditure on military durable goods other than weapons as capital formation?

61 20

Do you allocate FISIM* to the institutional sectors that consume FISIM*? 36 55

Do you estimate acquisitions and disposals of valuables? 18 10

Coverage of Non-monetary production

28

Coverage of non-monetary production: Questions posed


Africa%

Caribbean%

Do you impute rents for owner-occupied dwellings in rural areas?

82 80

Do you impute rents for owner-occupied dwellings in urban areas?

89 95

Do you include own-produced crops and livestock in household consumption expenditure? 89 20

Do you include estimates for own-construction of dwellings? 80 25

Do you include estimates for own-construction of other buildings (e.g. storehouses, barns, cattle-pens)? 61 20

Do you include the net increase in herds of domestic animals in the gross output of agriculture? 75 5

Do you include consumption of fixed capital in final consumption expenditure of government? 70 25

Frequency of different surveys and/or censuses

29

Availability of survey/ census data: Questions posed


Africa%

Caribbean%

Have you had a household expenditure or budget survey since 2000? 98 100

Have you had an industry/enterprise/business survey or census since 2000? 82 95

Have you had an agricultural survey or census since 2000? 55 50

Have you had a household labor force survey since 2000? 70 65

Have you had a survey or census of dwellings since 2000? 59 90

Have you had a survey of the informal sector since 2000? 52 30

Have you had a population census since 1990? 93 100

Availability of Price Data

30

Types of price data collected: Questionsposed


Africa%

Caribbean%

Do you have a Consumer Price Index? 98 95

Do you collect any wholesale prices? 82 5

Do you collect any producer prices? 55 15

Do you collect prices of construction materials? 52 70

Do you collect data on wages of construction workers? 23 40

Do you collect data on unit values of exports and imports? 70 50 or 80

Ranking by All Quality Variables (1-71) Percentage scores

31

Recommendations

(a) National accounts system and methodology1. Countries should implement at least these four

features of the SNA 1993:a. mineral prospecting should be treated as capital

formation;b. software should be treated as capital formation;c. government defense expenditures should be treated as

capital formation; andd. allocation of FISIM to consuming sectors.

32

Recommendations

(a) National accounts system and methodology2. Rents should be imputed for owner-occupied

dwellings in both rural and urban areas.3. Rents should be imputed by the user cost method

if actual rents for similar dwellings are not available.

4. Imputations should be made for value added and final expenditures for at least these non-monetary transactions:

a. production of crops and livestock for own consumption;b. own-construction of dwellings; and c. own-construction of farm buildings

33

Recommendations

(b) Surveys3. Informal sector surveys should be

carried out every five years. Ideally, these should be mixed household/enterprise surveys and held in conjunction with the household labor force survey.

34

Recommendations

(c) Prices1. In addition to prices of consumer goods and

services, producer prices should be collected for major crops, livestock products, minerals and manufacturing output.

2. Unit value indices for imports and exports should be compiled from customs documents. If this is not possible, c.i.f. prices for major imports and f.o.b. prices for major exports should be collected directly from importers and exporters.

35

A Tale of Two (Different) Inflation Measures (Total ECCB – 2002 to 2013)


CP

I

Imp

lici

t d

efla

tor

CEF11.06_L6: 37IMF Statistics Department

Thank you!

HOW GOOD (OR BAD) ARE ECONOIC STATISTICS IN THE CARIBBEAN SECOND HIGH-LEVEL ADVOCACY FORUM ON...

Documents

Transcript of HOW GOOD (OR BAD) ARE ECONOIC STATISTICS IN THE CARIBBEAN SECOND HIGH-LEVEL ADVOCACY FORUM ON...