Political power and the Argentine statistical system: The case of INDEC 1

1. Introduction

Argentina has a long tradition in statistical matters which led, in 1968, to the sanction of a Law that established the basis for the National Statistical System, regulated by the National Institute of Statistics and Censuses (INDEC). The system was inspired on the concept of normative centralisation and operative decentralisation, which provided INDEC with ample authority concerning the different statistical services of the National State and the Provincial Statistics Offices.

In practice, however, INDEC has concentrated on the development of its own statistical programs. Those which involved field surveys at a national level, were elaborated to include the active participation of the provinces through specific agreements and the financing of human resources undertaking such tasks.

For many years, INDEC was considered, within the country and abroad, to be a body characterised by its high professionalism, its technical accuracy, and the reliability of its statistical output. In addition, the Institute remained active at an international level, participating in diverse specific forums and in international exchange programs.

Despite this positive record, INDEC did not gain institutional strength. Without purporting to produce an exhaustive analysis of this aspect, it is necessary to highlight at least two important features. On the one hand, INDEC always was (and still is) a non-autonomous body, that is, it does not possess the authority to independently administer its human and material resources, although it does have a certain degree of administrative autonomy. A diverse number of legislative projects proposing a change of status never achieved a successful outcome. Despite this, the different governments respected, throughout the years, the independence of the Institution regarding its methodologies, production processes and results, until 2007.

The second relevant issue was that – similarly to what happened in the rest of the Argentine National Public Administration – the different positions were no longer distributed through public contest. In the case of INDEC, there were no qualification contests to fill the available positions. By the end of 2002, and within the framework of one of the most profound crises of Argentine economic history, a norm was pronounced for all of the Public Administration that enabled the recruitment of temporary personnel. Over the years, this mechanism was increasingly used by different offices, and in the case of INDEC, it became of great importance in contributing to the destruction of any type of professional career outline.

2. Political intervention in 2007

In this context, the economic policy of the National Government, which arose from the 2003 elections and continued with its 2007 successor, prompted an economic policy of expansion of public and private spending that, together with other factors, generated a rising inflation towards 2006.

In 2007, INDEC was virtually “intervened” by the National Government, more specifically by the Foreign Commerce Secretary, who was not even administratively in charge of the Institute.

At first, this intervention was concentrated on the Consumer Price Index, which meant that this indicator started to show evident deviations from reality. A number of alternative private sources started to develop.

Additionally, this intervention resulted in the replacement of directors and the relocation of a great number of technicians from their usual tasks, as well as the resignation and dismissal of a large number of them between 2007 and 2015. Their assigned replacements were employees who seldom complied with the minimum education requirements needed to perform the technical tasks of the Institute.

From 2011 onwards, the IMF started to question the accuracy of the CPI and GDP data, and in September 2012 and February 2013, the IMF issued a Declaration of Censure and Statement of Concern stating that Argentina had violated Article VIII, Section 5 of the Articles of Agreement.

3. The state of the institute at the beginning of the new government (December 2015)

In December 2015, INDEC’s challenges could be grouped into three categories: 1) Its institutional and administrative problems, 2) The credibility issues of its indicators, and 3) INDEC’s loss of leadership over the National Statistical System (NSS).

One of the main institutional problems of INDEC was the loss of a considerable number of its technical and professional staff, despite having been renowned during decades for having an array of specialized and highly trained human resources. This process was essentially due to the dismantlement of the most sensitive areas of indicator production by displacing the most qualified professionals from their specific tasks for not complying with instructions that were contrary to good statistical practice. Many of them were reassigned, suspended, or dismissed. An appreciable number of employees chose to resign. In many cases, the displaced personnel waited for years without being reassigned to new tasks.

Another detected problem was the great influence exerted by one of the trade union organisations on different administrative proceedings. In addition to exercising leadership roles in different technical areas, there was also evidence of their interference in recruitment of personnel and promotion, reward and overtime protocols, among others.

As from 2001, the number of staff increased almost by a third, and without consideration of the technical and professional requirements of the Institute. Around 75% of the new employees had no further-education qualifications. Thus, by December 2015, only 38% of the staff had technical or professional training, while the remaining 62% was comprised of employees with either a full or an incomplete secondary education.

The decisions made by the previous administration regarding modification of methodology and proceedings and, in some cases, specific measuring and/or publications eventually resulted in the qualified users’ lack of credibility and mistrust towards published information. This mistrust later spread across the rest of the population, who became sceptical of INDEC data and began using alternative indicators to make decisions. Hence the growth of private consulting companies and of university, guild, and non-governmental organisation social and economic observatories.

As a consequence of this, INDEC lost leadership over the National Statistical System (NSS). Added to the selective disappearance of supporting documentation for the published data; the violation of statistical confidentiality; the constant and significant methodological changes without the corresponding publications of explanatory reports; and the lack of publication of essential indicators – such as measurements of poverty and extreme poverty – was the lack of working relationships with the Provincial Statistics Offices and with the sectoral bodies which integrate the National Statistical System. Thus, INDEC was weakened in its leadership of the National Statistical System and as the main source of official Argentine statistics.

At close range, the situation of each and every area by December 2015 reflected the general diagnosis of the Institute – the working difficulties, the lack of credibility of indicators – but it also showed details that permitted a better understanding of those deteriorating years.

One of those areas is the National Directorate of International Accounts (DNCI, for its Spanish acronym), which is in charge of generating the statistics for the foreign sector. Historically, the DNCI functioned with a staff of 25 employees, nearly all professionals, which allowed it to produce high-quality statistics and studies. In December 2015 it had 16 employees, only ten of which were qualified with a university education. The combination of low salary and scarce professional incentives had caused the Directorate to lose a great part of its qualified staff. During those last seven years, the staff had received very little training, which deteriorated the quality of the estimations and led them to conduct routine data collection and processing tasks. Furthermore, the available hardware and software had become obsolete, which obstructed the implementation of the latest advances in international good practice.

In 2013, the Directorate of National Accounts, responsible for measuring the Gross Domestic Product (GDP), modified the base year used to measure the economy of the country. The base year, which until then had been 1993, was now 2004. The peculiarity of this extremely significant modification in the measurement of national accounts was the absence of a corresponding methodology to explain such changes. Many specialists expressed their misgivings, which, together with the lack of information on the used criteria, cast doubts on some of the results obtained in the new series produced with the base year 2004. This rendered the need to make the decision to completely revise the series at current and constant prices for each sector of economic activity.

The Directorate of Secondary Sector Statistics comprises the industry, construction and energy areas. The statistics produced in this area were based on a structure that was considerably out of date. The Monthly Industrial Estimator (EMI, for its Spanish acronym), for instance, included a high proportion of estimations. This lead to the decision to reobtain the original data from companies for the years 2014 and 2015. With regard to the EMI base year, deficiencies were detected in pricing procedures both in the last change of base (2012) and in the former (2006). The Installed Capacity Utilization indicator registered a very low response rate from companies during those past years.

The Monthly Industrial Survey presented similar issues. Only 2500 out of the 4400 consulted businesses still existed. The reason for this was that the base being used was from 1997. Moreover, in order to construct the physical volume index, sales in pesos were adjusted by the wholesale price index, which had several inconsistencies. The pharmaceutical industry database was also out of date: less than 40 companies remained out of the 75 of the original sample. Additionally, there was a shortage in human resources necessary to process this information. For example, during the National Survey on Technological Innovation and Behaviour, there were only two people dedicated to surveying almost 2000 companies. Other indicators had a very low level of progress in information gathering. For instance, in the statistics on construction permits, which includes participation of the Provincial Statistics Offices, only 41 of the 188 municipalities involved regularly provided information on the constructed areas covered by their building permits.

The Directorate of Tertiary Sector Statistics and Prices featured the same human resources problems. Only a small part of the personnel that dealt with the development and analysis of indicators had specific training for the task. This resulted in an array of issues with the indicators. The System of Wholesale Prices presented dispersions and irregularities in the reported prices. The imputed values for materials and labour costs in the Construction Cost Index System were noticeably lower than the real values.

The National Directorate of Statistical Planning and Co-ordination was functioning in a manner that was entirely detached from its original objective. The last plan that this area had drafted had been produced more than ten years before. The areas subordinated to the Directorate were leaderless or dismantled. One particular group, dedicated to the development of a multidimensional measurement of poverty, no longer received support from the authorities and had no assigned tasks. In the Directorate of Institutional Relations and Technical Co-operation, neither the director nor the more than 100 employees were dedicated to tasks concerning that area.

The two departments that make up the National Directorate of Social and Population Statistics were leaderless. As a consequence, they were subordinated to the National Leadership and had many middle-management positions vacant, such as undesignated coordinator positions. In practice, the Directorate did not meet the regular information production requirements. In general, information was elaborated according to specific demands and not based on a comprehensive plan. Many unfinished documents were found, corresponding to projects which were aborted because of the volatility of the conveyed instructions and the absence of a medium-term plan.

The situation at the National Directorate of Living Conditions Statistics was similar. Three out of the four leadership positions were vacant. The information produced was temporally inconsistent and the duties and responsibilities of the different areas were blurred: there was an overlap of tasks and structural flexibility.

From February 2007 onwards, the Consumer Price Index which INDEC had calculated since the 1920s started to drift away from the perceptions of large sectors of the population on price development. In time, the divergence started to be seen, as well, in the price indices produced by the Provinces. Different private consulting companies and institutions began to produce their own inflation calculations. The National Congress even gathered together a group of these sources in the denominated CPI-Congress. Multilateral bodies, such as the IMF, began to include warnings about the quality of the information provided by Argentina. It was precisely the IMF who, in 2013, issued Argentina a Declaration of Censure as a result of the inaccuracy of its indices.

The head officials of the area responsible for producing the CPI resigned a few days before the change of administration, leaving no information that would allow a reconstruction of the protocols followed or the methodological procedures. Not only was there lack of information within the Institute, but also lack of transparency in the technical and methodological documentation, which was not made public.

There was an attempt, in 2012, to develop a new national indicator: The Urban Consumer Price Index. However, the values produced by this new indicator were still significantly lower than the ones produced by most Provincial Statistics Offices. The suspicion of deficiencies in the application of the methodological protocols could never be completely avoided. After the change of government, it was found that the field device and the solidity of the information were in a precarious state.

There were also methodological irregularities in the Permanent Household Survey, the main social and employment data source of the country. One of them was the modification of the sampling frame based on questionable projections from the 2010 population census. There is very little methodological information on that modification, and there is no documentation that analyses how these changes impacted on the main indicators of the survey. Furthermore, there are other inaccuracies as of 2008: a typical career paths and a systematic lack of answers to questions about income.

Another vastly questioned measurement of INDEC was the poverty measurement, which fundamentally depends on three sources: The Permanent Household Survey, the National Survey on Household Expenditure, and the Consumer Price Index. The irregularities in these sources, especially the first two, lead to a great gap between the official measurement and studies produced by universities and private organisations. Eventually, in 2012, the publication of this measurement was discontinued.

With regard to administrative matters, the Directorate of Assets and Procurement lacked the management control to monitor contracting and keep track of the needs of the Institute. Additionally, the Institute did not have a purchasing plan approved by administrative act. In consequence, contracting was not programmed and there was no information concerning budgetary formulation. The approved budget for 2016, for example, was improperly formulated in technical terms. Many items displayed excessive budgetary provisions, while others displayed insufficient budgetary provisions.

Most circuits of the administrative area of the Institute did not have any proceedings manuals or instructions. The Procurement and Contracting Procedure, which had started its ISO 9001 quality certification process, was never followed-up and had expired by 2015. The inventory system was also obsolete, as was the storage system. The Institute’s material assets were not properly maintained; a great part of the movable and immovable property had fallen into disuse or had been rendered unusable.

In the Dissemination Directorate, which was found in a virtual state of paralysis, several strategies used by the outgoing management to reduce publications to a minimum could be identified. One of them consisted of harassing the employees who did not agree with the guidelines imposed by the directors. They would freeze their wages or discretionally assign overtime or survey hours; unilaterally remove or change tasks; physically isolate them by placing partitions dividing sectors; restrict the use of equipment or block the user licences on their work-station computers.

INDEC did no longer participate in national or international forums and fairs; the user services sectors (consulting rooms, remote service, library, publication and special works sales) suffered a reduction of information demand and response rates as a result of the lack of credibility towards the Institute.

The publication of data, in its varied formats (webpage, printed and digital publications) was drastically diminished in the last years, and all information was submitted to prior censorship by the authorities.

Thus, the majority of the areas were transformed into watertight compartments, with limited communication, and without responding to any global objectives or programs. In fact, without clear logic or proceedings in most dissemination tasks, duties were assigned according to each group or area leader’s criteria. Similarly, coordinators were assigned based on their affinity with the directors and not their abilities. In addition, the server used for mailing expedition was so slow that sending technical reports would sometimes take up to four hours.

The Informatics Directorate did not have an approved IT plan to specify long term objectives and allow a guideline for the development of activities, verification of goal fulfilment and optimisation of budgetary allocation according to dedicated projects. Regarding the infrastructure, the premises for the processing of critical or sensitive information (protected by Statistical Confidentiality) were not located in a protected area of the building. The physical security system of the computing centre was not adequate. There was no fire protection system in place in the Server Room or in the Computing Centre. Nor was there a power generator in place for the Data Centre to allow the operational continuity of the technological services during unforeseen interruptions of power supply.

4. INDEC in the new Argentine government

During the election campaign, the team that later became the new government had reviewed the situation of INDEC and conducted meetings with various experts from around the country. Given such a degree of destruction of the Institution, one of the considered alternatives was the creation of a new organisation.

Finally, it was decided that the best option was a process of reconstruction of the Institute that would simultaneously enable the restoration of the Official Statistics System within a reasonable period of time and design a profound institutional change in accordance with the recommendations of the main international statistical centres.

Since the restoration of reliable statistics would require immediate changes in the leadership of the organisation and the incorporation of experienced technicians, the National Government declared an Administrative Emergency for the National Statistical System and granted broad powers to the Director-General of INDEC in order to modify the leadership layout of the entity and make some changes to its organisational structure.

Additionally, the President of the Republic instructed the Director-General to begin an investigation on the irregularities of the previous years.

This emergency situation also permitted the suspension of the regular statistical series Release Calendar.

Administratively, the emergency situation and extraordinary powers of the Director-General concluded on 31 December 2016.

5. The process of restoring official statistics

The reorganisation of the Institute meant replacing about 30 leadership positions, which constitute almost all of its critical positions. The new directors were chosen from among the staff which had been relocated from their tasks, and technicians who had been resigned for not agreeing with the previous situation or who had been dismissed for the same reasons.

Moreover, this new leadership team managed to rally a limited but high-quality number of technicians who had experienced the same exclusion process.

This core of leaders began a process of reviewing all statistical series and historically reconstructing them whenever possible.

Naturally, the series with Public Record sources have been much easier to correct and reconstruct. Those which depend on field work, however, sometimes presented insurmountable difficulties, since there were strong deficiencies in almost all their respective surveys.

Such is the case of the consumer price (CPI) series, Labour Market Statistics, and the Poverty and Extreme Poverty Indicators.

As a result of this work, last September INDEC re-established the publication of all of the statistical series. (See Appendix 1)

6. Some critical developments

Two of the most strongly questioned statistics, especially by the IMF, were the CPI and the GDP.

In the case of the CPI, the previous government had developed a national indicator (IPCNu) which had serious deficiencies. As a consequence, the decision was made to return to the traditional measurement setting, which had produced this index for many decades, in order to find a more solid point of departure within a reasonable period of time. Under the IPCNu, data collectors worked largely unsupervised and without quality control checks. For the new index, supervisors once again focus on the supervision and management of data collectors. Reflecting best practice, supervisors make unannounced visits to observe data collectors and randomly select questionnaires for verification and validation. To further improve the collection of data, INDEC is developing a training program for relevant INDEC staff, and plans to hold regular meetings between data collectors and headquarters-based INDEC staff to discuss data collection issues and challenges.

The 2012/2013 expenditure survey was discarded as a tool that provided the weights for the IPCNu. The 2012/13 survey was affected by an unusually high non-response rate. For the city of Buenos Aires, the response rate was 27 percent (53 percent in the previous survey conducted in 2004/05) and for the Greater Buenos Aires area it was 41 percent (69 percent in 2004/05). Given the lack of documentation on the criteria that were used to correct the survey for these low response rates, there were concerns about the survey itself and the resulting quality of the weights used in the construction of the IPCNu.

The new CPI index published on 15 June covered the Greater Buenos Aires area. The first priority for the new management was to reorganise and rebuild the price collection process in the greater Buenos Aires area, where the staff is under the direct control of INDEC (as opposed to provinces where price collection relies on the provincial statistics offices). Given the problems with source data and the lack of adequate information on calculation procedures, INDEC decided that it would take too long to produce a high-quality national index. Instead, the new management focused its efforts on the production of a CPI for the Greater Buenos Aires area (comprising the City of Buenos Aires and 24 surrounding districts, accounting for approximately 45 percent of national expenditure).

In July 2017, INDEC began to release the CPI with national coverage, which includes data for the national level and 6 regions of the country (GBA, Pampas, North-east, North-west, Cuyo and Patagonia). The data is collected in 39 agglomerations of the country, the City of Buenos Aires and the Greater Buenos Aires area (previous CPI).

Furthermore, the revision of the official GDP data produced a downward adjustment of economic performance since 2004. On 29 June 2016, INDEC released a second revision of the GDP data with the base year 2004. The new level of nominal GDP for 2004 was 9.5 percent below that estimated two years before. The new series showed that the nominal GDP increase accelerated since 2009. By contrast, real GDP growth was significantly lower than the previous series: on average for the period 2004–2015 real GDP grew by 3.8 percent against the 5 percent in the old series (cumulative real GDP growth was lowered by 18 percentage points in this period). It is worth noting that the new series showed that the economy suffered the 2009 global financial crisis: the revised real GDP growth - 6 percent, instead of the previous 0.1 percent. The positive year-on-year percentage change in GDP (Q4 2016, Q1 and Q2 2017) broke the former year-on-year decline.

The changes introduced in June 2016 to correct mismeasurement of the official GDP data are methodologically sound and in line with international standards. The GDP released on 29 June is now largely based on detailed industrial volume indicators drawing on various independent sources (such as the Chamber of Commerce and business associations). Such methodology is applied in many other countries, and appears warranted in the case of Argentina, as the high level of inflation complicates using price indicators for the calculation of volume estimates. The new GDP data are derived using a production approach (that is, as the sum of gross value added by economic activity plus net taxes on products) which is common in many other countries. The 2008 System of National Accounts manual recommends a detailed supply-use framework for GDP compilation, and the authorities plan to implement a more sophisticated and modern GDP estimation system in the next benchmark revision of national accounts tentatively scheduled for 2018.

In May 2017, the new organisation structure was approved by the Chief of Cabinet. One of the most important changes was the division of the organisation of the Institute into a Technical Directorate, which co-ordinates the producing areas, and a Management Directorate, which co-ordinates the supporting areas. This new organisation structure also raised the hierarchy of the Directorate of Statistical Methodology and the Directorate of Dissemination. They are both now National Directorates, four rank-levels under the President of the Nation. Another important change is the creation of the National Directorate of the National Statistical System, which will co-ordinate six new regional offices (Cuyo, Metropolitan, Northeast, Northwest, Pampas and Patagonia) both in the administrative management and in the monitoring of programmed statistical activities.

7. Guidelines for the future

Now that INDEC is again delivering official statistics on a regular basis and with transparency and complete adherence to international principles in the elaboration processes, it is necessary to update to a new and improved INDEC.

The deep changes that are necessary to achieve high levels of quality cannot be done without this fundamental step. New statistical capacities according to the needs of the society and the government need to be developed; participation in the international economic, social and statistical community also demands new answers. Human resources have been severely affected: only 44% of our personnel are professionals, of which 34% of total employees have a university degree. This percentage is an improvement from the original (December 2015) 38% professionals, but the goal is to reach 60%. To this end, 85% of new employees hold a professional degree.

The infrastructure is obsolete and the activities are dispersed in five different locations.

The present achievements are the result of the enthusiasm of a human group eager to participate in a new era in the management of public affairs and in a foundational process involving access to information and modern and reliable official statistics.

It would be somewhat naive to assume that this process has concluded. We are at the beginning of the road.

The new INDEC we are thinking of must be in line with the philosophical and organisational principles that predominate in the international statistical community.

Full adherence to the Fundamental Principles of Official Statistics and similar statements must be explicit at the highest level of our government and in the new law.

A new institutional structure is needed. Some of the highlights are:

INDEC is now a decentralised organisation. In terms of Argentine public rules, that means something like a hybrid with some independence at the administrative level, but without independent management of the budget, organisational structure and (not considering the emergency situation) previous presidential or ministerial approval of every appointment.

The National Statistical System needs revision, especially the relationship with the Provinces, but also within the executive power.

Since the beginning of the administration, INDEC has re-established relations with the international statistics and economics community which had been severely damaged and, in some cases, completely interrupted.

At present, a flow of cooperation has been developed with the United Nations Statistical Commission, the International Monetary Fund (IMF), the World Bank (WB), the Inter-American Development Bank (IDB) and the Economic Commission for Latin America and the Caribbean (ECLAC).

Furthermore, the Argentine Government decided to begin a journey of incorporation into the Organisation for Economic Co-Operation and Development (OECD) and the Minister of Economy of Argentina has expressed, on behalf of the government, the explicit desire for the Argentine statistical system to meet with the standards and good practices in this area.

Following close collaboration between INDEC and the OECD and an evaluation process, in August 2017, Argentina was announced as an Adherent to the OECD Recommendation of the Council on Good Statistical Practice.

8. Lessons from the experience

The process experienced by INDEC permits us to draw a variety of conclusions.

9. Appendix: Restored statistical information

Activity Level Progress Report. GDP.

Monthly Economic Activity Estimator (EMAE, for its Spanish acronym).

Balance of Payments, Argentina’s Investment Position (PII, for its Spanish acronym) and External Debt.

Argentine Foreign Trade Statistics.

Foreign Trade Prices and Quantities Indices.

Export Complexes.

International Tourism Survey (ETI, for its Spanish acronym).

Hotel Occupancy Survey (EOH, for its Spanish acronym).

Industrial Products Statistics (EPI, for its Spanish acronym).

Monthly Industrial Estimate (EMI, for its Spanish acronym).

Synthetic Indicator of Construction Activity (ISAC, for its Spanish acronym).

Synthetic Energy Indicator (ISE, for its Spanish acronym).

Argentine Pharmaceutical Industry.

Agricultural Machinery Industry Report.

Biofuels.

Consumer Price Index (CPI) with national coverage.

Construction Cost Index (ICC, for its Spanish acronym).

System of Wholesale Prices Indices (SIPM, for its Spanish acronym).

Installed Industrial Capacity Utilisation.

Survey of Shopping Centres.

Supermarket Survey.

Household Appliances and Articles Outlets Survey.

Main Indicators for the Labour Market.

Employed Workers, Hours Worked and Wages per Worker in the Manufacturing Industry.

Monthly Valuation of the Basic Food Basket (CBA) and the Complete Basic Basket (CBT).

Co-generation and Self-generation of Electric Power.

Employment Records Statistics. Jobs and Income of Registered Salaried Employees.

Poverty and Extreme Poverty Incidence.

Argentine Trade Exchange (ICA, for its Spanish acronym).

Provincial Origin of Exports.

Internet Access.

Synthetic Indicator of Public Services.

Salary Index.

Argentine Foreign Trade Statistics.

Socio-economic Indicators.

Evolution of Income Distribution.