Prosperity Index should be aimed at regional disparity

Central Bank initiative needs improvement:

Q: How do you explain that the SLPI is conceptually very weak and what is the theoretical basis upon which you build up your argument against it?

A: Initially, the fundamental mistake has been made in the definition of prosperity. The introduced index has borrowed the definition for prosperity from the so-called development indices developed by the West targeting material prosperity in industrial and service sectors. Prosperity is defined by different people differently

 in different contexts. For some people, it means "Freedom from all of those things that decrease the experience of well-being, and the freedom to choose situations that increase our sense of well-being. For some others it is an expansion of experience of the life that fulfils at all levels-mind, body and spirit. If the prosperity index is to be applied to inter-provincial status on a comparative platform, prosperity should not be defined from the view point of societies which are totally alien to us.

'Alienated perceptions'

For example, the number of 'elderly homes' in a given region as an indicator is a reflection of the social security system prevailing in the West.

Higher the number of 'elderly homes', better the prosperity of the Western society. However, in our society, we look after our parents and relatives and more 'elderly homes' is an indication of the erosion of our moral and cultural norms and values. It is not at all a suitable, positively correlated indicator to measure the prosperity in our societies. In my opinion, our prosperity should be defined as perceived by our own people with respect to our own unique economic, social, cultural and religious setting.

Further, the SLPI has identified three main areas in its composition of indicators, namely Economy of business climate, well-being of the people and Infrastructure. These three main pillars on which SLPI stands, are out of balance with respect to their own standings of their importance. In order to keep the SLPI in balance, the main pillars should rise to the equal level of status in terms of their value in the society. Well-being of the people is the ultimate goal of any rational policy, programme or development activity.

Infrastructure is the general term describing the basic physical systems and facilities such as roads, bridges, sewers and sewer plants, water lines, power lines, and other sites and facilities necessary to the functioning of an area.

The relevance of the economy of business climate as a main pillar for the SLPI is questionable. It may be important for a country where industrial or service sector dominates its economy, but not for a country like ours where predominant agriculture sector should take the proper share of the economy. The Western bias of the indicators is very obvious when no explicit indicator to cover the contribution of the agricultural sector has been included in the set of indicators while industrial density has been given undue prominence. In our ancient kingdoms, prosperity of the country was mostly based on the agricultural productivity and the freedom from foreign invasions.

Even at present, we have a great leader as our king who stages a war against the invasions on separatist agenda and neo-economic and political colonialism of so-called Western democracies while promoting the agriculture sector in the country with the introduction of the national food drive Api Wawamu, Rata Nagamu. The government's vision for a new Sri Lanka has been ignored in the selection of the indicators for the SLPI.

Q: How do you justify that the SLPI ignored the very foundation of our civilization?

A: As a nation, we inherit a great hydraulic civilization and a proud agricultural history where the protection of the environment was given due consideration. Today, environmental education is an integral component of our education. The Ministry of Environment and Natural Resources has embarked upon several projects and taken several measures to prevent the pollution of the environment realizing the health and sanitation issues associated with the poor environmental conditions. However, in the Prosperity Index, there is no reflection of the bio-physical environment in its indicators which is considered to be a prime factor in the prosperity of a nation.

In fact, there are indicators which are contradictory to our national environmental policies if implemented and promoted by provincial administration as recommended by the Central Bank. Per capita electricity consumption is one such indicator. As a nation, we must endeavour to curtail per capita energy consumption since more than 91% of our energy is derived either from petroleum or biomass according to the data published by the Sustainable Energy Authority of Sri Lanka.

'Major lapses'

Although I am not an economist, I still remember my early university days where we learnt the definitions for economic growth and development which brings in the issue of equity of resource allocation. A society can never be prosperous unless the individuals are having an equitable stake in sharing resources. In the SLPI, there is no consideration whatsoever on the equity of basic and common resources within the spatial units.

Apparently, some other indicators such as telephone density and banking density have been rated for unit area since these are not evaluated as per capita availability like electricity consumption. Telephone and banking are services for the people and are provided to meet the demand and needs of the people. These indicators are obviously lower in value in other provinces where the population density is lower compared to the Western Province.

Any such service should be counted on the basis of population density rather than proportioning it to the extent of the land in each of the province.

Information was not available as to how a composite indicator like law and order is assessed. In my opinion, law and order should reflect a separate category since lawlessness is an evil against the prosperity of the country.

If it at all is to be included under the three main pillars of the SLPI, I would rather include it in the category of well being of the people rather than it to be in infrastructure.

All the above facts point to the conclusion that the SLPI is conceptually weak in its development, division into three main areas and selection of indicators.

Q: How do you explain that it is said to be statistically incorrect? *

A: A main goal of inferential statistics is to quantify or measure uncertainty using fundamental concepts and theories of probability and randomness when the underlying process is not perfectly deterministic.

Dealing with uncertainty, one must select a valid statistical approach to follow in order to get the results endorsed by the scientific community.

The SLPI does not get support from any statistical valid methodology for its construction and is apparently lacking any valid statistical base for its standing as an index of prosperity estimation. The five point score given to each indicator is arbitrary and has no statistical connotation.

Although the sub indices which are merely nominal labels have been given differential weights not based on any statistical outcome but as a results of number of indicators arbitrarily classified under each sub index. A simple approach of multi-variate analysis could provide the necessary weights for the indicators.

There is a critical mistake in the use of some of these indicators as implied by their nomenclature. Some of the indicators are based on per capita estimation of the variables and some others like people above poverty, law and order, ownership of vehicles, industrial density are not derived to be comparable in view of the differences of the regional context in terms of land extent and population.

There are fundamental spatial statistical approaches to derive indicators to be used in the analysis of spatial diversity and to determine lower and upper bounds for the indicators. However, in the case of SLPI, I could not find any reference to how the lower and upper thresholds were predetermined and rescaled to 0 and 100. The only clue provided as an explanation to this is that indicator value at all times in the foreseeable future will be lower than the upper threshold.

It stages the pessimistic assumption that the prosperity in Sri Lanka will not be achieved in the foreseeable future.

Further, there are several highly correlated indicators in the list of the selected 20 indicators. This can be very easily shown by developing correlation matrix for these variables.

The common variance shared by several indicators leads to the statistical error that it gives an additional weight for the outcome indicators due to the duplication of common variance in the calculation.

For example, continuation of secondary education, O/L and A/L pass rate, school density, pupil teacher ratio are obviously highly correlated indicators which share a considerable proportion of common variance and contribution from individual variance is understood to be marginal.

Also, the health indicators such as hospital beds (government), doctors (government), infant survival rate, maternal survival rate are highly correlated and inclusion of these four indicators in place of a single indicator out of the four in the index would not add any significant new information into the formula.

A simple approach like Principal Component Analysis or Factor Analysis could extract the individual variance from communalities and avoid the inclusion of redundant indicators in the index.

Further, the entire sub-index of well-being of the people has been curtailed to education and health indicators except the ownership of vehicles.

Q: How do you perceive that SLPI has poorly interpreted the ground realities?

A: It was emphasized in the media that the SLPI could be used to make comparisons among the provinces in terms of prosperity status and also it can be used to compare the performance of the country as a whole. It implies that the SLPI is proposed to compare the performance of different countries, at least in the region.

Even if an index is developed to determine the true aspirations of the people towards prosperity at sub national level, such an empirical index is not valid to be applied in a different context due to the underlying assumptions made to derive the index.

It is common knowledge that the purpose built indices cannot be used outside its context or spatial boundaries.

In spatial statistics, this is identified as the boundary problem and where the derivation of data is based on specific boundary conditions which cannot be interpreted upon a different set of boundaries.

This has been highlighted even in the analysis of election results where the mandate given by people based on the district based electoral system cannot be interpreted based on the original boundaries of the individual electorates.

Further, it should be understood that the boundaries of the spatial units cannot be modified at a later stage to interpret the outcomes based on new spatial demarcations.

Q: Why do you maintain that the SLPI does not have any numerical significance?

A: The arbitrary allocation of fixed weights to the indicators makes it impossible to add the individual indicator value to a common scale running from 0 to 100. It appears that the number of indicators was constrained to be 20 in order to allow each indicator to have a maximum value of 5 thus restricting the 0 to 100 scale to a new scale of 0 to 5. Without extracting the actual communality of the data through a dimensionality reduction technique, it is not possible to just add rescaled variables together to derive an index having numerical significance. Each indicator has its own unit of measurement and a simple rescaling could only derive a set of ordinal data series which does not bear any numerical significance other than its comparative location in the ordinal scale data domain.

Further, each sub index is differently weighted not because of the relative importance of the each sub index towards the overall prosperity but the assigned number of indicators for the each sub index.

In addition, the three sub indices are merely nominal categories which are just ornaments to the prosperity index and do not add any value in terms of its numerical strength in the computation process.

The development of a true prosperity index should not be constrained by the limitations of the indicators being used in the process for easy computation. The index should take into account a diverse set of societal dimensions to reflect the regional disparities that have a spatial distribution.

In this exercise, indicators that are having poor spatial diversity or high cross correlation could be avoided. SLPI should be a purpose-built index which would indicate the causative factors for the poor performance on the prosperity scale rather than highlighting the outcome variables so that decision makers could easily understand the interventions that are needed to guide the region or province on the path leading to prosperity.

I am certain that the introduction of such an index would definitely draw the attention of the Chief Ministers of the provinces and other government officials to be based their development programmes to achieve prosperity of the people and the country.