STATISTICS (AMENDMENT) BILL

STATISTICS (AMENDMENT) BILL: Order for Second Reading read.

3.11 pm

The Minister for Trade and Industry (Mr Lim Hng Kiang): Mr Speaker, Sir, I beg to move, "That the Bill be now read a Second time."

Sir, the Bill proposes two key amendments to the Statistics Act. The first set of amendments allow the Department of Statistics (DOS) and Research and Statistical Units (RSUs) which are gazetted under the Statistics Act to provide researchers access to anonymised microdata. The second set of amendments clarifies DOS’ legislative powers to obtain information from public sector agencies.

Providing anonymised microdata access

Today, DOS and the gazetted RSUs, collect, analyse and disseminate a wide range of socio-economic and demographic statistics. Under the existing Statistics Act, such statistics can only be provided in aggregated form.

However, to undertake more sophisticated analysis of behaviour across

Column: 1884

different types of firms or households, researchers need access to microdata. Microdata refers to data on a person, household or establishment at an individual level. Many advanced National Statistical Offices (NSOs) like the US, the UK, Netherlands, Finland and Australia provide microdata in anonymised form, often masking the identity of the individual party by applying appropriate statistical techniques.

Microdata analysis will help us better understand the impact of public policies or external factors on firms and households with different characteristics. In countries where microdata are available, researchers have used it for a variety of valuable and useful research. For example, researchers used anonymised microdata from the Canadian Labour Force Survey of Statistics to investigate whether the government’s mandated maternity leave had any impact on the labour force behaviour of mothers. In Singapore’s case, income and expenditure microdata from the Household Expenditure Surveys could be used to study the consumption behaviour of different household types. Such studies will complement analyses based on aggregated data, and provide greater insights to aid policy formulation.

The proposed amendments will allow the Department of Statistics and gazetted RSUs to provide access to anonymised microdata for research purposes. We will however adopt a calibrated approach in opening up access to such data. To start off, only public agencies that require anonymised microdata for policy research will be allowed access to such data. These agencies can, on their own, engage consultants to undertake research using the anonymised microdata. To protect the confidentiality of individual data, all researchers, including consultants, will be subject to stringent security protocols. We will monitor this carefully. Access to a wider group, such as other research

Column: 1885

institutions, will be considered at a later stage.

_Protection of confidentiality
_

Mr Speaker, Sir, I will now elaborate on the various safeguards to protect the confidentiality of individual data. DOS will anonymise the microdata before access is provided. In instances where it may not be possible to completely mask the identity of the individual party, such as in the case of industries with only a small number of firms, then DOS will not provide access to this information.

To guard against any leaks, users will have to come to an on-site Data Lab at DOS or the gazetted RSUs to use the anonymised microdata. All data will be examined prior to removal from the Data Lab to ensure that only statistical tables and output without any anonymised microdata are taken out.

DOS will set up two committees to implement these changes. A Microdata Review Panel will look at the technical issues on anonymisation and subject matter sensitivities and will scrutinize each request for microdata. An additional Microdata Advisory Committee will advise the Microdata Review Panel on policy issues and processes relating to anonymised microdata access.

The proposed amendment will also increase the maximum fine for breaches of confidentiality, from $5,000 to $10,000, besides the existing provision for imprisonment for a term of up to 12 months. This serves as an additional deterrent to prevent abuse or misuse of anonymised microdata.

_Obtaining information from public agencies
_

Mr Speaker, Sir, let me now turn to the second amendment and explain the

Column: 1886

rationale to explicitly allow DOS to obtain information from public agencies. DOS was established to collect, compile, analyse and disseminate key official statistics. To minimise the burden on respondents, DOS has been using administrative data already collected by other agencies, instead of requesting the same information anew directly from members of the public or private establishments. This is a common practice among the many National Statistical Offices.

The proposed amendment makes it explicit that the Chief Statistician is allowed to obtain data from public agencies for statistical purposes. The data obtained will be protected under the provisions of the Statistics Act. This amendment will further strengthen DOS’ role as Singapore’s National Statistical Authority and will enhance DOS’ ability to provide more timely and relevant statistics to the Government and to members of the public.

As a safeguard, the Ministers responsible for the agencies that collect, or are in possession of the data will have the power to exempt sensitive information from being provided to the Chief Statistician.

Sir, I beg to move.

Question proposed.

3.18 pm

Mdm Ho Geok Choo: Mr Speaker, Sir, I welcome the Statistics (Amendment) Bill as it would give more teeth to the Department of Statistics, our national statistical authority.

Although statistics is a simplification of things in a real world, the process of gathering data is a tedious one. This Amendment Bill changes the law to allow the Department of Statistics and the

Column: 1887

Research and Statistical Units to provide researchers access to anonymised microdata, when they previously only had access to data in aggregated form. The greater precision and detail of information provided will thus empower researchers in delving deeply into the significance and meaning of the data, and can only be beneficial to the policy formulation process.

This provision under section 3 of the Amendment Bill gives these units the authority to collect and analyse socio-economic and demographic statistics. This change is to be welcomed, and will hopefully be of great use in coming up with policies that cater to the needs and wants of our populace. Anecdotal evidence and expert opinions, together with a more detailed set of data and statistics, will go a long way towards identifying the source of the issues that we face, as well as pre-empt problems before they arise.

And now with the Integrated Resorts soon coming onstream, the data collected will be even more significant. However, there is the concern of accidentally revealing the identity or sensitive information about a particular respondent, with the provision of microdata relating to individuals. At the same time, it is important to ensure that the anonymisation process does not distort the data set, resulting in policies formulated based on mistaken information.

Sir, I am glad to note that there will be established a Microdata Review Panel to examine technical issues on anonymisation and subject matter sensitivities, as well as scrutinise each request for data. I would like to query if this Panel will effectively be making the decision as to whether to release the data, and who will be sitting on the Panel? Similarly, who will make up the Microdata Advisory Committee, which has the duty of advising the Review Panel on policy issues

Column: 1888

and processes relating to anonymised microdata access?

Sir, at the same time, it is imperative that Singapore safeguards the data obtained and ensure that it does not fall into unauthorised hands. I would like to commend the Ministry for having proper safeguards in place to ensure confidentiality of information and the doubling of penalties for the breach of confidentiality is reassuring and helpful in safeguarding the status of Singapore as an international business hub.

With that, I support the Bill.

3.22 pm

Ms Sylvia Lim (Non-Constituency Member): Sir, there is a need to balance the public interest in gathering information and the right to individual privacy. Thus, I would like to ask whether the Bill will facilitate more data gathering, but makes compromises on information safeguards?

Specifically, the Bill has gotten rid of a safeguard which is present in the existing section 7(6) of the Act. Under the existing section 7, information that may have been provided under confidential circumstances to the Census Department, CPF Board, Comptroller of GST and Comptroller of Income Tax, had two layers of checks. The first check was that the Minister-in-charge of the department which gathered the data, has the power to put restrictions on what the statisticians can disclose after obtaining the information. There was an additional check under section 7(6) that no information furnished by them to the Chief Statistician or Director of any research unit, could be disclosed without the written permission of the agency itself.

Sir, the departments referred to in the current section 7(6), indeed collect sensitive tax information, personal

Column: 1889

data and a person's employment details. They may also have gathered the data under obligations of secrecy. The importance of this additional safeguard on privacy by the agency itself was, in fact, highlighted in Parliament in 1990 and again, in 2004, when amendments to the Statistic Act were debated. It was stressed by two different Ministers that due to the sensitivity of those data, the approval of the agency itself, whether the CPF Board, Comptroller of GST or Comptroller of Income Tax, etc, should be obtained before statisticians release it. This is over and on top of the Ministers' right to restrict the disclosure of the information.

What this Amendment Bill appears to do is to repeal the third shadow altogether, which means that these five sensitive departments are now grouped together with the entire government machinery as public agencies. Under the new section 6, the Ministers-in-charge can exempt their agencies from disclosing certain information to their statisticians. However, the safeguard of requiring the five agencies to specifically give written approval before release of their data, has been removed. Instead, in the new section 7(3), it is the Chief Statistician or the Director of the research unit, who can imposed conditions on the use of the information supplied.

Is this change a compromise on the safeguards of the confidential data gathered? With due respect, the Chief Statistician or the Director of the research unit, may have no specialist-knowledge of the subject matter nor the public expectations at the time the data was gathered. I would like the Minister to clarify, why it was necessary to remove this additional check by the five departments which clearly gathers sensitive information?

Column: 1890

3.25 pm

Assoc. Prof. Paulin Tay Straughan: Thank you, Mr Speaker. This Bill gives the Chief Statistician a lot more authority to amass data collected through Government agencies.

The assumption is that relevant data will be used where appropriate to advice and inform policy. In principle, I agree with the amendments spelt out in this Bill for two main reasons. First, as research commissioned by Government agencies is done using public monies, there should be better consolidation of knowledge and information to avoid replication and inefficiency in research. And second, as we grow a keen appreciation of evidence-based research and acknowledge the relevance of data in informing policy, we should enable the Chief Statistician to facilitate the use of data to advance the important work of policy makers.

While I stand in support of this Bill, I have three comments. The first relates to the quality of data. As we move towards a culture that embraces quantitative research and the outcomes of surveys, we must reconcile the quality of data, especially those from survey research. For example, large scale survey projects that rely on quota sampling or convenient sampling like straw polls to obtain data are not able to generalise trends that are representative of community.

May I ask the Minister, what plans his Ministry has in place to ensure the quality of data before it is used to inform policy? For example, will research conducted by Government agencies be governed by instructions from institutional review boards?

My second concern relates to the implications of this Bill on ethical considerations in research. For example, a respondent agrees to participate in a research commissioned by a public

Column: 1891

agency and is given the assurance that the data collected will only be used to advance appreciation of a specific set of social issues. Should this data be later used to reflect other concerns that were not spelt out in the original research agenda, this will compromise the issue of informed consent that in pivotal in any ethics consideration.

To overcome this possible ethical dilemma, may I suggest that data obtained from public agencies be used within the context of the original research objectives. Let me illustrate with an example. Health Promotion Board (HPB) commissions a research on teen smoking. So teenagers are encouraged to participate and HPB assures all respondents that the aim of the study is simply to appreciate the motivators that push a teen to take up smoking. But if the data is later used by law enforcement agencies to locate hot-spots where underaged smokers are likely to hang out, then that would be a serious breach of ethics. Subsequently, there will be a greater reluctance, if this should happen in a community to participate in research commissioned by public agencies.

My final point relates to growing a conducive culture for evidence-based research. If we are going to move towards a culture where we are going to use more data to advance policy and inform policy, then we have to ensure that there is good quality data that is amassed by researchers on the ground and this has a lot to do with cultivating a positive culture and appreciation for research by Singaporeans. For us to be able to leverage on the rich details present in the data that is collected through a professionally conducted research, we must nurture a community that is appreciative of this kind of social research. So it is critical that researchers in Singapore are able to establish a relationship with potential respondents that is based on trust and respect.

Column: 1892

Currently, in most of the researches done, there is no direct feedback to respondents on the outcome of the studies that they participated and co-operated in. To overcome this, we should have better accountability to the public on how the statistics are used. One simple way is to achieve a meaningful connect between respondents and research outcome is the provision of an Executive Summary. So, maybe the Minister might want to consider making this mandatory in the research that is funded by Government agencies. If you have this kind of direct feedback to the public, then it will result, I am sure, in an increase in voluntary participation and it would invoke a more conducive research environment in Singapore.

Finally, this Bill supports the principle that data should be shared – between Government agencies and Chief Statistician. I would like to leverage on this, and propose that perhaps, a central repository of data be created where data sets derived from commissioned research can be shared, not just within the Government but with academics, students and interested members of the public. If we share non-sensitive data, it will result only in a win-win situation when you have many pairs of eyes scrutinizing the data set. Through this, we will only empower the community to rely on obtaining facts to inform opinion and decisions.

I would like to leverage on this, and propose that perhaps a central repository of data be created where data sets derived from commissioned research can be shared not just within the Government, but with academics, students and interested members of the public. If we share non-sensitive data, it will result only in a win-win situation when you have many pairs of eyes scrutinising the data set. Through this, we will empower the community to rely on obtaining facts to inform opinion and decision, and if

Column: 1893

Singaporeans can reap direct benefits from research outcomes, they will also be more likely to contribute to research that is commissioned by the Government.

With this, I support the Bill.

Mr Lim Hng Kiang: **** Mr Speaker, Sir, first let me thank Mdm Ho and Prof. Straughan for their support of the Bill, and also to Ms Lim for raising some very important clarifications. Let me respond to the issues raised.

First, on the microdata review panel. This will be chaired by Chief Statistician and will include all the relevant members from the various agencies providing the source as well as the policy agencies. This panel will evolve, and depending on the requests put up to the mircodata review panel, we will have the composition of the relevant people to review the requests, and also to scrutinise the mircodata that will be released.

Ms Sylvia Lim asked whether by amending the Act, we have reduced the safeguards. The answer is "No". The amendment essentially formalises the current arrangement and, as I mentioned in my Second Reading speech, the Ministers who are responsible for the agencies that collect the data or that are in possession of the data, will have the power to exempt the sensitive information from being provided to the Chief Statistician. So the gatekeeper remains the agencies that collect the data, and the Ministers responsible for these agencies. So in the six agencies that she mentioned – CPF, IRAS, etc, – if the data is considered sensitive, the agencies and the Ministers can still exempt these data from being given to the Chief Statistician.

Sir, let me next answer the three comments raised by Prof. Straughan. First, on the quality of the data, I agree

Column: 1894

with her that this is very important. I think the various statistical agencies are very concerned and they have international standards to abide by when they collect the statistics. Nevertheless, Statistics Department will review these quality guidelines and formalise this set of guidelines for the various RSUs. Second, on informed consent and ethical issues raised – again, because we are involving the agencies, both the policy agencies as well as the statistical agencies, in the microdata panel, if requests for such data compromises the informed consent or ethical issues, then the agencies involved in collecting the data, can veto and not release the information. So that is the check and balance.

Third point about how we can inculcate a conducive culture of evidence-based research – this is indeed our objective and we are taking these initial steps to release anonymised microdata in order to facilitate this culture of evidence-based research. But we do so in a careful calibrated manner. I think eventually we hope to have it extended to all important research agencies. For a start, we are extending this only to these agencies and the researchers appointed by these agencies. But once people gain confidence that even though we are releasing anonymised microdata, this is done very carefully and that confidentiality is safeguarded, then I think we can extend this to more researchers. I agree with her suggestion that there ought to be a good feedback mechanism, so that people feel that the data which they provided for various statistical purposes, are put to good use and that policies are increasingly based on proper research and proper evidence.

Question put, and agreed to.

Bill accordingly read a Second time and committed to a Committee of the whole House.

Column: 1895

The House immediately resolved itself into a Committee on the Bill. – [Mr Lim Hng Kiang].

Bill considered in Committee.

Mr Lim Hng Kiang:

The Chairman: The citation year "2009" will be changed to "2010", as indicated in the Order Paper Supplement.

Clauses 1 to 14 inclusive ordered to stand part of the Bill.

Bill reported without amendment; read a Third time and passed.