For the first time in more than two decades, Pakistan has produced a national economic census. It had more than 7.1 million establishments and nearly 40 million PK Press Club-marked structures, making it the largest digitization exercise in South Asia. Government representatives presented it as a historic step towards building a reliable statistical framework for the economy. But behind the figures of the titles, the experts claim that the report also exposes serious dead angles and contradictions which raise questions about the usefulness of the set of data for the development of policies.
The latest economic census in Pakistan was carried out in 2000. Since then, the country’s economy has been transformed, but decision -makers had to rely on obsolete estimates and unequal surveys.
This is why Bilal Gilani, executive director of Gallup Pakistan, considers the new census as a historic. “It is wonderful that it finally happened after 23 years. The world has radically changed during this period, but we are still counting on obsolete figures. This provides a digital reference base that Pakistan has never had before,“” He said.
The census was integrated into the 7th population and housing census of 2023. The enumerators equipped with digital tablets marked 38.3 million structures, 79% of which were residential and 13.4% were economical. A total of 7,142,941 establishments were documented, employing 25.3 million workers.
Dr. Lubna Naz, professor of economics and director of the Center for Business and Economic Research of IBA, described the extent of the effort as both impressive and unfinished. “”It is both an important step and a work in progress. The integration of seven million establishments in a single digital framework and the 40 million PK Press Club-marginary structures is unprecedented, but the data must be strengthened before being able to serve as a robust basis for politics, “she said.
However, how the census was made, its methods, tools and checks is as important as the figures themselves.
Methodology, divergences and role of AI
The census marked Pakistan’s first attempt to fully digitize the economic list. By integrating the exercise into the population census, the enumerators were able to collect data on companies while recording household information, using digital tablets for PK Press Club-marquerie.
However, reliability has become a challenge. More than half of all the entries, 52%, were initially recorded as “others” due to descriptions of waves of activities, errors of spelling or Roman Ourdou. To solve this problem, Pakistan Bureau of Statistics (PBS) turned to artificial intelligence, automatic learning and natural language processing to clean and classify data.
The tests have shown uneven results. In Lahore, the first trials revealed that the algorithm had only obtained a precision of 24.68% in correspondence establishments with their correct industrial codes. After the revisions, the rate improved at 85.23% in Karachi and 86.50% in Faisalabad.
Nationally, the final model has been implemented with an estimated precision at around 80%. On a set of data of more than seven million establishments, this margin means up to 1.4 million could be poorly classified.
Dr. Naz questioned the approach: “The process depended on the limited tests in three urban centers, which restricts its representativeness. Large parts of the linguistic and regional diversity of Pakistan may not have been sufficiently captured, “she said.
Gilani adopted a more indulgent view. “What matters is that a baseline has been created. Even with certain erroneous classifications, it is a leap forward compared to any census,” he said.
What the census leaves aside
Beyond the classification challenges, experts say that the biggest problem lies in what the census has excluded. The methodology covered establishments in fixed structures and household activities such as embroidery, sewing, poultry, tuition fees and food production at home. But that did not count mobile sellers, street stands, small beauty salons or the rapidly growing sector of freelancers and digital companies.
Dr. Naz has warned that this will distort the image: “Many unstructured activities are distributed in services, retail or small -scale production.
The report recognizes this weakness. Household activities were recorded, but 40.49% fell into a “diverse” category. This means that almost half of the household economy remains unspecified and actually unknown.
Contradictions and anomalies
A deeper examination of the report reveals inconsistencies between different data tables and classifications.
-
Education of education: Adding the types of units gives a total of 290,729 educational establishments (242,616 schools, 11,568 colleges, 214 universities and 36,331 Madrassas). However, as part of the industrial classification, educational establishments number 326,868. This leaves more than 36,000 unexplained establishments.
-
Délia Lutin: Hospitals are counted as 119,789 establishments, but in industrial codes for human health and social work, the number goes to 123,973. The report does not provide ventilation to reconcile this gap.
-
Mosque labor anomaly: The census has 600,403 mosques using 2.06 million people, more than the entire factory sector. But it is not clear if it includes volunteers, guards or teachers, which makes the figure difficult to interpret.
-
Structures VS establishments: The census documents around 6.3 million economic structures but more than 7.1 million establishments. This implies on average 1.13 companies per structure, confirming that many places, markets and buildings on several floors house several companies. However, the report does not analyze this density, missing an opportunity to shed light on the commercial grouping in the cities of Pakistan.
These contradictions indicate the weaknesses of how the data has been cleaned, tabulent and presented.
Geo-Tagging: a first digital
One of the most announced achievements of the census was PK Press Club-marquerie. The enumerators recorded the GPS coordinates of nearly 40 million structures across Pakistan, allowing for the first time a PK Press Club-referencing mapping of companies.
Gilani considers this as a digital breakthrough: “It is a major achievement. For the first time, each store and establishment is literally on the map. This can change the way Pakistan plans to grow urban and commercial “,” He said.
Dr. Naz has urged caution: “Geo-Tagging has a real promise for town planning and targeting of investments, but it is as good as the classifications attached to it. If the activity codes are inconsistent, then bind them to the coordinates does not solve the reliability problem, “she said.
Missing depth: finance and sex
Another major difference is the absence of financial and sex data. The census records the number of establishments and their workforce, but does not measure the sources of income, investment or financing. This makes it impossible to assess the economic value of household companies or their sustainability.
Dr. Naz underlined the implications: “By not monetting household work, Pakistan lacks crucial information on home companies, limiting the ability of political decision-makers to design effective credit, training and business support programs,” she said.
Lack of gender data is also a concern. Without depression of male and female workers in establishments, the census cannot precisely understand the role of women in the economy. This makes their contribution both visible in number but invisible in value.
The risk of staying motionless
Globally, economic censuses are carried out regularly. India and Bangladesh already use them to capture funding, investment and digital companies. The exercise of Pakistan, although important, risks being a if not institutionalized.
Dr. Naz warned: “If Pakistan does not repeat and does not improve its economic censuses, it is confronted with dead political angles and to lose competitiveness in emerging sectors such as electronic commerce and freelance, where women and young people are the most active.”
Gilani has agreed that continuity is essential, although it has stressed the importance of progress already made. For him, the census marks the start of a digital statistical system on which Pakistan must now rely.
A foundation on which to rely
The 2023 economic census is undeniably historic: the first in addition to two decades, the first digital and the largest in the genre in South Asia. It provides a reference image of the Pakistan economy which did not exist before.
However, its faults are just as important. The contradictions between the tables, the anomalies in number of staff, the exclusions from the informal and digital sectors, and a strong dependence on the imperfect AI all raise questions on the quantity of weight policies that decision -makers must place on its conclusions.
While Dr. Naz comes down, he is better considered as a breakthrough and a work in progress. The real test will be whether Pakistan can refine, repeat and extend this effort in a reliable system. For the moment, the country has a digital framework for its economy. The question of whether this framework is solid enough to support good economic decisions remains uncertain.
