Presentation of Francois Bourguignon

Senior Vice President and Chief Economist, World Bank

at the National Bureau of Statistics of China

May 18, 2005

Commissioner Li,  Ladies and Gentlemen,

It is with great pleasure that I address this audience of experts and friends from NBS and other distinguished guests.  Over the past 25 years, the National Bureau of Statistics has been a teacher for the World Bank in many ways, and without our close cooperation, our understanding of China, and our comprehension the country’s remarkable development success would have been far from what it is today.  I believe our cooperation has also furthered statistical work in China, and has supported transition of the Statistical System from one suited for a planned economy to one that is increasingly suited for a market economy.  Our joint research has also furthered our global knowledge of poverty and poverty reduction and many other issues relevant for development.  I am pleased to note that the cooperation between the NBS and the World Bank is becoming steadily closer.  Indeed, after my speech here today, Commissioner Li and I will sign a Memorandum of Understanding which expresses our mutual desire to work ever closer together.

Today, I would like to talk about the role of statistics in China’s development.  Specifically, I would like to talk about its role in China’s new development paradigm, the Scientific Approach to Development.  I will first review some of China’s development challenges as I see them, and then reflect upon the new approach to development that China’s leadership has proposed, and I will end with the implications for statistical work in general, and the work of NBS in particular.

China has been highly successful in development and poverty reduction over the last 25 years. Integration in the world economy, together with the remarkable transition from planned to market economy the country went through in the last quarter of a century, has been a powerful engine of growth and poverty reduction.  Over the last quarter century China’s GDP grew on average by over 9 percent , and more than 400 million people were lifted out of absolute poverty.  Trade grew even more rapidly than GDP, with an unprecedented 12 percent over the same period.   It is now the third largest trading nation in the world.  If recent trends continue, China will emerge as the largest economy somewhere between 2030 and 2040, and it will be the largest trading nation before 2020.  The country has over $600 billion has the second-largest stock of international reserves after Japan.

The prospects for China’s continued rapid growth are good.  It has favorable demographics with a rapidly growing labor force and a still-declining ratio of dependents (children and the aged) to workers for more than a decade to come.  China has until now also maintained a high savings rate, needed to finance a high level of investment.  As I have already noted, China has been remarkably open to FDI and trade, two factors that bode well for the adoption of new technology that can spur growth.  And among developing countries, China has created a relatively good investment climate – attracting foreign investors and its own private entrepreneurs alike.  All of this makes China’s emergence as a major international player highly likely. 

But China’s continued growth and development success are not a given.    In my view, China faces major challenges going forward.  These range from the crucial issue of how  China can maintain a high growth rate that is needed to continue to reduce poverty and increase prosperity for all, to the issue of how the country will manage its growing resource and environmental constraints, and the emerging issue of how China can more equally share the rising prosperity by providing equitable access to all.   How China will meet those challenges will determine whether the country will achieve its desired Harmonious society, and continue to alleviate poverty as effectively as it has done in the past 25 years. 

This is not the occasion to discuss these issues in depth, but what I would like to note here is that these are extraordinarily complex issues to tackle for any country, and the required reforms are likely to be of a higher level of sophistication that the first two phases of reforms China has undertaken―marketization and opening up.  This third generation of reforms will require new thinking, new tools, and new approaches to policy making. 

China’s New Development Approach.

China’s policy challenges reflect an increasingly complex society.  Indeed, it is not just China’s economic and social circumstances that have changed dramatically over the last few decades, China’s approach to reforms is evolving as well.  Deng Xiaoping’s famous approach of “Feeling the Stones to Cross the River” by experiment, trial and error, and scaling up of what works has been exceptionally successful in China’s reforms, and remains relevant in many areas of reforms.  However, a new approach has gained prominence in many government speeches, including the ones of the prime minister:  “The Scientific Approach” to development. 

As you all know, in October 2003, the 3rd Plenary Session of the 16th Party Central Committee of the CPC proposed the guideline for building the Xiao Kang Society in an all-around way, by achieving “five balances”―balancing urban and rural development, coastal and inland development, economic and social development, man and nature development, domestic development and opening up to the outside world.  A new development concept, “the scientific approach to development”, was proposed by President Hu Jingtao. This new concept calls for "people-centered development, which is comprehensive, coordinated and sustainable, for the promotion of overall harmonic development of economy, society and human beings." From a development perspective this is a very exciting new concept.  It is rare that a government explicitly calls for a specific approach to its development, let alone a scientific approach to development. 

As a scientist myself, I am keen to better understand this new development concept and what its implications are for China, and because I am Chief Economist of the World Bank, I am also eager to understand the implications of this approach for our work in other client countries.  I think it is fair to say that the Scientific Approach to development is still an evolving concept, and the domestic debate on its implications is far from settled.  I have discussed this approach already with some of China’s leading scientists and policymakers during my trip this time, and would like to explore some of my own thinking here today. 

The scientific approach to science is fairly settled by now: the cycle of developing theories, deriving testable hypothesis and testing hypothesis with the available data or data to be created.  This method dating back to the great Greek thinkers, and formally established as the fundamental methodology of science by Karl Popper in the last century, has become the norm.  Theories whose testable hypothesis seems most in conformity with the data become accepted knowledge until a better theory comes along.  Alternatively, new available data reject the old theory and force scientists to develop new theories.   A somewhat unsettling implication of this approach is that there are no absolute truths, but only hypotheses that are not yet rejected by the data.  At the same time, this approach has proven to be a most efficient way of advancing human knowledge.  It was Newton’s observations of falling objects that established the theory of gravity, and it was the observation that planets bend light that rejected the same theory in favor of Einstein’s theory of special relativity.  Less dramatic progress in knowledge is being made everyday in a cycle of induction, deduction, and testing.  Until I joined the Bank this is was what I did every day in university and at the DELTA research institute in Paris.

At the more practical level of development policy making, a scientific approach is not yet an established concept, but the very idea of such an approach is an exciting one, and I take this opportunity to propose some ideas on what it could imply.  In my view, a scientific approach could mean that policies should be under a constant empirical scrutiny, much like theories in science.  So it becomes the task of the policy adviser to develop those policies that best meet the political objectives of the country, based on the available facts.

 In addition, the scientific approach should also mean that policies under implementation should be under empirical scrutiny to assess their impact.  In other words, monitoring and evaluation of policies, programs, and projects is an essential part of the scientific approach the way I see it. 

Finally, much as scientific debate uses conferences, peer reviewed journals, and competition among researchers as tools to further knowledge, new and existing public policies should be under constant pressure from competing policies and policy proposals that better meet the goals of the political leadership or better address the realities that policies have to live up to.  “Seeking Truth from Facts” as Deng Xiaoping’s famous saying goes is in my view a good summary of the proposed scientific approach to development.   Note that it is my belief that in a scientific approach, the political ends in themselves are not subject to the same scrutiny: a policy adviser can only argue in terms of instrumental rationality, and only give advice conditional upon the goals: “IF these and these are the political objectives political leaders want to achieve, then, to the best of our knowledge, A, B, or C are the policies that best meet those goals”

The implications for the business of government of the Scientific Approach I have just proposed are profound.  Frankly, it puts the bar for the policy making machinery of a country very high indeed, and I am not sure whether there is any country in the world that would fully live up to such a high standard of integrity in policy making.  But I believe that we can all strive to live up to the standard as much as possible, and doing so would be of great benefit to the quality, effectiveness and impact of public policies.  Let me reflect further on some of the elements that a Scientific Approach to Development would require.

First, such an approach would require an open debate on proposed policies to address given policy objectives.  Such a debate would not only be necessary to sort out good policies from inferior policies, but also to generate the very proposals themselves―it would only be worthwhile to formulate competing policies if there is a chance of acceptance of those policies in the market place for policy ideas.  I must say that in following the more recent policy debate in China, I am very encouraged by the openness with which many of the policy issues are discussed nowadays―in newspapers, in the numerous forums and conferences organized every day in China, and even at high level policy seminars that involves the leadership itself.

Second, the approach requires presence of the analytical capacity to develop alternative policies and to test their effectiveness in a rigorous way.  Again, I am encouraged by the enormous progress that China has made in this area.  China is home to a number of world class universities, has a large array of think thanks and policy research centers at all levels of government.  In contrast to the past, China now has the capability to attract promising Chinese scientists trained overseas back to China.  Nowadays many of the ministries also attract internationally trained talent, and with it the capacity to absorb scientific findings in the business of government has greatly increased.

Third, the approach requires the availability of and access to high quality data with which policy proposals can be developed, and policy effectiveness can be tested.  Indeed, data, information, and statistics are the lifeblood of the Scientific Approach to Development.  It is on this topic that I want to focus on for the rest of my speech.

The Implications for Statistical Work

The implications of China’s new development paradigm for statistical work are very large indeed.  The Statistical System will become a cornerstone of the new approach, and will have to carry the flag of the approach by supplying an increasing diversity and quantity of information.  The statistical system will also have to supply that information not just to the traditional users, but also to a growing array of users that will have to be involved in policy in order to make the Scientific Approach to Development work.  This includes policymakers, policy analysts, academics, the general public, and increasingly the outside world, as China is more integrated in the world economy, and becomes a more prominent player in leading international policy fora.  

China’s leadership is well aware of the need for better statistics. Indeed, the 3rd plenum that introduced the Scientific Approach also called for the following:

The statistical system should also be improved to enhance monitoring of economic performance and facilitate information sharing among economic regulation departments of the Government

It is my belief that the importance of those words cannot be overestimated.  

Fortunately, China’s statistical system is already a strong one compared to many other developing countries, and it has made major progress in the last decades.  The system has recovered from the heavy blow dealt to it during the Cultural Revolution. It has introduced a host of international standards into its work, with perhaps the adoption of SNA in 1993 as the single most important step.  It has also absorbed internationally accepted techniques for core statistics, notably surveys for collecting macroeconomic as well as household information, and the household and enterprise surveys are true treasures of information that can be put to use for a whole array of policies China is searching for.  Departments and organizations, first and foremost NBS, have also made tremendous investments in human capital and physical infrastructure, which is clearly paying off.  Finally, since 1984, China has a statistical law that provides the important legal basis and institutional framework for statistical work in the country.

Despite this progress, the Statistical System still faces a number of tremendous challenges, which China needs to address in order to make the vision of a scientific Approach work.  The World Bank has had the pleasure to work with the NBS in recent months to analyze the challenges and find practical solutions appropriate for China’s circumstances, and I believe there is growing consensus on some of the most critical issues.  Let me highlight a few here.

l        First and foremost, coordination and leadership in statistical work needs to be assured.  The statistical system as a whole remains highly fragmented across agencies and levels of government.  This is in part the result of the way China’s government is organized, which results in imperfect control of central offices, such as NBS and line ministries that have offices at the provincial, city, county, and grassroots level.  In turn this creates issues in quality and methodology for data generated.  But at the central government as well, there is a lack of coordination among the various departments on statistical work, and funding for such work remains fragmented and often too low.  Overcoming this lack of organization requires not only tighter management across the vertical lines within line ministries and bureaus, but also coordination among agencies as well.  In other countries, a Statistical Council led by the highest levels of government to coordinate statistical work, methods, and definitions across agencies and levels of government has proven to be useful to overcome the coordination issues, and it is certainly an idea that China could further pursue.

l        Second, a major investment in statistical infrastructure is needed.  Agencies at all levels have heavily invested in information technology, which has greatly benefited the statistical system.  However, investment in what I would call “soft infrastructure” is lagging.  The soft infrastructure of China’s statistical system needs a boost in terms of definitions and standards.  On definitions, for example, the remnants of the planning system are still present in the statistical system even today, and moving to a full adoption of SNA, including a redesign of the survey system that should generate the necessary information for such a move should be high on the agenda, and the NBS’ ongoing efforts on the economic census will be a major milestone in achieving the move.   On methods, in many areas, the importance placed on comprehensive reporting compared to survey work distracts enormous amounts of resources into areas that are simply less productive for assembling useful statistics, in fact, they may be counterproductive.    In the area of household surveys, moving to a unified household survey for all of China’s households―which would enable better estimates of poverty and inequality to just name a few applications, would be highly desirable.  New areas of interest to policy makers requires attention as well, including the ones closely linked to the Harmonious Society like environmental statistics, and statistics on equality of income and opportunity among households.  And finally, reliable information on sub national governments below the provincial level is required by policy makers who desire to monitor performance of local leaders.  This requires scaling up of the sample sizes used in some of the common surveys implemented in China today.

l        Third, consolidation and elimination of overlap remains a key challenge.  There is still a considerable amount of overlap in statistical work, and the suppliers of information―households, enterprises and grass root organizations, are showing increasing signs of “survey fatigue,” a symptom that could seriously undermine the integrity of the statistics itself.  This forebodes an increasing number of non-responders to surveys.  Moving away from comprehensive reporting will certainly help in fighting survey fatigue.  Moreover, consolidating surveys and putting single surveys to use for multiple departments will also be important,  This again emphasizes the importance of coordination at the highest levels of government.

I would like to highlight two other aspects of the statistical system that require attention if the scientific approach to development is to succeed.  The first is monitoring and evaluation, and the other is access to data. 

The collection of information on the policies of the government itself―monitoring and evaluation―is only nascent in China.  I have just been to the Qingba Mountains in Sichuan Province to study a highly successful poverty program that was implemented there to draw remote communities out of poverty by means of community investment, and by enabling people to move to more productive jobs in the cities.  I was very impressed by the results of the project, and as a scientist I was particularly taken by the excellent monitoring and evaluation effort done to track the impact of the project.  This effort, which was led by NBS itself in cooperation with the Leading Group for Poverty, is truly exemplary for other departments as they implement policies in the context of the Scientific Approach to Development. NBS monitored the impact of the project by means of collecting data on the beneficiary group of villages as well as on a control group of villages not under the program.  Within program and non-program villages, household income, poverty, migration patterns, health, education, and other quality of life indicators were collected and analyzed, and although more thorough analysis of the data is called for, the efforts has already paid off in terms of some important findings, of which I would like to highlight two.  First, the data show that the program villages caught up with non-program villages in terms of per capita income and other welfare indicators―so overall the program indeed seems to work to the benefit of reducing poverty in remote mountainous areas.  Second, the labor mobility component, which was considered to be an innovative part of the Qinba project, does not seem to have resulted in more labor migration from the program villages compared to the non-program villages. 

This type of monitoring and evaluation work is key to understanding the impact of the government’s own policies, and for the improvement of those policies.  Indeed, in some advanced countries, being able to demonstrate the effectiveness of a policy, program, or project, is what counts for the annual budget allocation from the Ministry of Finance.  Monitoring and Evaluation―and follow up in terms of budget allocation―requires more attention in China as it does in many countries around the world, and I congratulate the NBS, and the Leading Group for Poverty Alleviation for spearheading such important effort in China.

Second, I would like to highlight access to data.  Apart from the institutional and statistical challenges, China’s Statistical System needs to be able to cater for the whole range of new users that I mentioned before.  This is perhaps the most difficult challenge given the institutional traditions in China that are not conducive to sharing data, even internally inside one organization.  Indeed, it is likely to require  a cultural change if the statistical system is to live up to the requirements of the scientific approach.  Involvement of users of data needs to take place on at least three levels: 

(i)                  in asking for inputs from users about which statistics to produce.  The demand for statistics is changing all the time and those that produce statistics need to be aware of this shifting demand much like an entrepreneur should be aware of changing demands for his products.  Therefore, regular interactions with key users of statistics come at a premium.

(ii)                in providing “meta data” to ensure that the statistics produced are well understood.  While organizations such as NBS are increasingly publishing their statistics in regular publications, the description of the data and the methods of compilation of the data may well be enough for a general public, but it falls short of the requirements a scientist would have if she were to work with the data to derive policy implications from them.

(iii)               providing access to data by users for them to put the data to work in analysis without compromising the confidentiality that is not only of paramount importance to keep the trust in statistical survey work, but also to keep within the confines of the law.

Within the Bank we have also struggled with the issue of providing access to data.  The World Bank collects a wealth of data from around the world for its own work and for providing policy advice to its client countries.  Much of the data we publish in a range of publications, the best known is perhaps the World Development Indicators which has become a standard work for development economics.  The hardcopy of the data can be downloaded from the web, and a softcopy with historical data back to 1960 can be ordered on CD against a nominal fee that covers production and handling costs.   

More recently, we have stepped our own efforts in survey work as well―much like the NBS.  Access to these data is more sensitive than data derived from country statistical publications, as there is the important issue of respondents’ confidentiality to take care of.  Yet, providing access to data is key to getting the world academic community involved in analyzing the data.  Doing so simply means that thousands of researchers outside the World Bank can also put their efforts to work for our client countries, by pursuing their own research, testing their own hypotheses, and through this contribute to the scientific approach to development.  In other countries, access to record level data is available after the statistical office has removed any identifier in the record that could personalize the information.  For household data, such an approach is usually enough to discharge the statistical office of their legal obligation to confidentiality.  Only in some instances is this insufficient, for example in enterprise data it may be possible to identify large enterprises by sorting through the records and additional confidentiality measures may be necessary.

We faced such a challenge with the data of our world-wide effort in investment climate surveys.  These surveys contain a wealth of data on how enterprises perceive the investment conditions around the world, and data to test for the adversial or beneficial impact of these conditions on enterprise productivity and profitability.  To provide access while protecting confidentiality, the Bank has developed the Online Data Analysis Toolkit which provides the user the opportunity to work with record level data in their analysis without providing the record data themselves to the user.  This has been a very useful exercise to us, and one that is much appreciated by users around the world.  We would be happy to assist the NBS in developing a similar approach for China.

These are all important challenges to the Statistical System.  It is therefore very encouraging to know that the NBS, which is the leading agency within that system, has already embarked on wide-ranging reforms to meet the challenges within its own organization.  The organizational and institutional reforms that NBS is currently contemplating are very important indeed for the future of China’s statistical system and I believe they will receive all the support from the leadership that they deserve.  The ongoing work on the Economic Census and the efforts to fully adopt the SNA system are equally encouraging.  At the same time, NBS cannot do it alone.  The reforms needed to reshape the statistical system in light of the scientific approach to development go well beyond those that can be done within NBS.  I feel that the Statistical Council that I referred to earlier could play a crucial role to ensure that all departments at all levels of government give statistical work the priority it deserves.

Commissioner Li, Ladies and Gentlemen,

In closing let me reiterate the importance that I see for statistical work in achieving the Scientific Approach to Development, which, in turn, the leadership sees as the way to achieve a Harmonious Society for China.  I hope that the World Bank can play a small part in supporting China’s authorities to achieve these goals.  In particular, I hope we can have a strong partnership with NBS in achieving these goals, and I am very pleased that Commissioner Li and I will sign an MOU today that signals both parties’ intent to build such a partnership. 

Xie Xie.