It's a pleasure to be talking to you today, although I regret that I could not join you in person. Still, to give technology its due, this is almost as good.

Let me first take this opportunity to pay tribute to the Chief Statisticians at this conference and, indeed, to all statisticians around the world. You are truly the unsung heroes behind central bankers and other policy-makers. You constantly labour under tight resource, financial, and time constraints, to satisfy our insatiable appetite for data. As users of those data, we tend to focus on what can be done to improve them, rather than on recognition for what is already there.

But I can assure you that central bankers hold statisticians in high regard for the breadth and quality of information they provide. The data that you painstakingly produce form the base for the analysis and research that informs the formulation of monetary policy and our understanding of trends in financial systems and markets. So it really all starts with the data!

Your efforts to ensure the accuracy, integrity, relevance, timeliness, and international comparability of statistics are vital to our decision-making process.

Having told you how much we value your contributions, I must also tell you that I won't pass up this opportunity to make still more demands of you, on behalf of all of us in central banking! But, at the same time, I hope to be helping you, by giving you a sense of priority and some general principles to guide your efforts in addressing those demands.

So, what drives us central bankers to put more and more demands on you?

The Statistical Needs of Central Bankers

Our statistical needs are fundamentally shaped by what we are expected to do under our mandate.

The primary goal of most central banks today is to conduct monetary policy so as to achieve and maintain price stability. Low, stable, and predictable inflation is the means to our ultimate objective of solid economic performance over time.

In addition to price stability, we are charged with promoting the safety, soundness, and efficiency of our national financial systems.

Some among us are also expected to regulate financial institutions.

I should add here that, in the process of carrying out these responsibilities, we are collecting and generating significant amounts of statistical information ourselves. Data are a public good. We need to work together to make them accessible to researchers and the public in a convenient format. From our side, we could do more to share with you the data we collect and make the most of limited national statistical resources. Central banks that are regulators, in particular, produce significant amounts of data that could be shared. But for that, a robust legal framework would have to be in place, to allow the exchange of information.

Let me now turn to what we need from our statisticians. In this context, I will focus on the two central bank functions I just mentioned: the pursuit of price stability and the promotion of financial system stability.

In either one of those cases, the objectives have not changed. But our economies are continually changing and becoming more interdependent under the effect of globalization and large and growing trade and capital flows.

As policy-makers, it is extraordinarily important that we understand how, and to what extent, these forces are likely to affect the achievement of our objectives, so that we can adjust our policies accordingly.

Identifying the sources of potential challenges and threats to the achievement of our objectives, and determining how we should adjust to changes, is quite a task. But finding ways to measure the effect of changes in our economies and in our financial systems is no less important or complicated a job. For that, naturally, we turn to you!

So, what are the broad trends and challenges facing those of us concerned with price stability, and those concerned with financial system stability? And what do these challenges mean in terms of what we need from our statisticians?

I will deal first with those related to inflation control.

Data Challenges Related to Price Stability

In conducting monetary policy, we aim to keep the economy operating close to the level of its production capacity, so as to maximize output, employment, and income gains over the longer term, while keeping inflation low. When making interest rate decisions, we always try to gauge the extent of demand pressures—current and prospective—on production capacity (or aggregate supply). And we try to assess the actions required to maintain or restore balance between aggregate demand and supply in the economy, so as to keep inflation under control.

As you can appreciate from this description, the range of statistical information that we have to look at, for purposes of our decision-making, is quite wide. But for those of us, in particular, who are operating with explicit inflation targets, the most important indicators, besides price indexes, are measures of output, productivity, and capacity. These measures directly affect our ability to estimate the production potential of the economy and to assess the balance between demand and supply.

I have made these general comments to serve as a marker for what I have to say next about the broad trends and challenges relevant to the goal of price stability and the associated statistical needs. Much of that discussion will, one way or another, hark back to price and productivity measures. (That's what you get when you talk to central bankers!)

The growing importance of services
An important feature of all major economies these days is the growing share of services. But service output is less tangible than that of the traditional goods industries. Indeed, in some service industries, such as banking, there is not even agreement on the appropriate definition of output. And in many countries, certain services, such as health care, are not delivered through the market, which makes it even more difficult to measure. As if all this were not enough, now, we also have to contend with "virtual" service output—that is, software, music, movies, and data that people can download from the Web. You certainly don't need me to tell you how tricky it is to measure output, prices, and productivity in the services sector. Or, how tentative some of those estimates can be.

With the importance of services growing in our economies, the need for better output and price measures for this sector is becoming increasingly pressing. So, if you were looking to allocate limited statistical resources according to priority, my view would be that this particular area merits a higher ranking than in the past. What I'm saying is that, if you are trying to determine whether additional resources should be put in, say, quality adjustments for consumer goods or quality adjustments for services, the choice ought to be services.

Better price and output measures for services would improve our ability to assess overall economic conditions by providing better information on current trends in aggregate output and prices. They would also help us identify other emerging trends or structural changes in the economy. For example, these days, an important issue is to evaluate the productivity gains from the use of information and communication technology (ICT). Knowing the effects of ICT would help us predict future growth in the production capacity of the economy. One way to identify the contribution of ICT to overall productivity is to examine whether productivity gains have been strongest in those sectors that are heavy users of ICT. We know that services is one such sector. But if we do not have the right price deflators for services, we will never get reliable measures of productivity growth and of the ICT effect in that sector. And so, it will be more difficult to judge the contribution of ICT to total productivity and to production capacity.

Risk and insurance
The increased occurrence of unpredictable events (conflicts, new diseases, and natural disasters) means that the world has become a riskier place; or, at least, that our perception of the risk has increased.

In this type of environment, insurance and hedging have assumed a bigger role than before. Not only has the provision of insurance become a more globalized operation; it has undergone significant structural change.

There has also been significant repricing of insurance worldwide, for both consumers and businesses, mainly because of marked increases in the value of claims. In Canada, we have recently seen large effects from higher insurance premiums in our inflation data. The United Kingdom and Australia have had similar experiences.

It is for these reasons that I am identifying insurance separately from other services. The challenge here is how to measure something that we hope never happens. I can see that we will have to devote more statistical resources to this issue. We need to think conceptually about how to measure the economic value of insurance and how to correctly measure the price of the service. In addition, central bankers have to think about the implications for economic activity and potential output of the increased risks and higher costs of mitigating those risks. And from the perspective of financial system stability, we must see to it that those who hold the risk price it correctly and that they are in a position to carry it.

But let me now move on to the next issue of relevance to the goal of price stability.

The implications of rising trade flows and firms operating globally
Changes in the way certain economic activities are carried out in today's "global village" have been so rapid that all of us have had a hard time keeping up. In many cases, it is not so much the nature of the transaction that creates difficulties for measuring what is happening, as it is the location and the price at which it gets done.

International trade is a primary example. Thanks to trade liberalization and falling transportation and communication costs, the share of international trade has been rising in most economies and, within that, the share of services. The activities of firms that are operating globally have also been expanding, as have the number of mergers. National frontiers are thus blurring, and there is a whole lot more intra-firm trade going on.

As encouraging as this growth in global trade is from the perspective of bettering the lot of more people around the world, it is not without challenges for policy-makers and statisticians alike. From your perspective, the more direct challenge is that it has become harder to collect information and to accurately measure economic activity within, and outside, national boundaries.

These are not necessarily new issues, and neither are those related to services that I discussed earlier. But, with the growing importance of trade in services and of multinational firms, these issues are coming to the fore, and the need for better information becomes more pressing.

What are the most important statistical needs with respect to trade and multinationals?

First, clearly, we need better information on intra-firm trade and on trade in services in order to get better aggregate trade statistics. Second, central bankers will always tell you that they are especially interested in better data on prices in different currencies. That is because we need to assess the economic effects (for example, which margins "get squeezed") when exchange rate movements are passed, or not passed, on to domestic prices. This is what we call the exchange rate pass-through.

This pass-through seems to have been less pronounced recently than in the high-inflation years of the 1970s and 1980s. There may be more pricing to markets, that is to say, firms may increasingly be setting prices based on what local markets can bear. A thorough analysis of the exchange rate pass-through issue is complicated by the fact that national statistical agencies often use a mechanical approach to convert foreign currency prices into domestic ones, rather than collecting actual import prices.

Transfer pricing by multinationals adds yet another dimension to the problem. I know that it is virtually impossible to get good information on transfer pricing. But better data on intra-firm trade and work with global enterprises may still help us get a better fix on import prices.

Real estate prices
Fluctuations in asset markets have become a more prominent feature of modern economies in recent years. Considering that property is by far the world's biggest single asset class, it is not surprising that movements in the real estate market are drawing a lot of attention. In many countries, housing prices, in particular, have been rising rapidly, raising some concerns about a possible sharp correction at some point.

Given that investment in housing represents a big chunk of household spending, and that for most people their homes represent their most valuable asset, it is surprising that, in many countries, there are no comprehensive quality-adjusted data on housing prices and rents. In its recent survey of global property markets, The Economist commented that "official statistics offices typically collect more information about the price of shoes or cement than housing, despite its far greater importance."

There is a need to expand the current limited international experience in constructing standardized housing price and rent indexes. And so, I am encouraged that the International Monetary Fund (IMF) and the Bank for International Settlements (BIS) are convening a joint conference of experts in the autumn to support work to improve data availability in this field.

Given how often real estate booms have triggered banking crises around the world, this issue is clearly important from the point of view of financial system stability as well—to which I would now like to turn.

Challenges Related to Financial System Stability

Financial market liberalization has led to tremendous growth in global financial activity in recent years and to more integrated financial markets. The demand for financial services has also risen in response to the growth in the operations of cross-border firms. And in many countries, there has been a merging of the traditional segments of the financial sector and a proliferation of financial products, including derivatives and hybrids.

Unfortunately, with globalization and with increasingly complex financial markets, the effects of any disturbance have tended to reverberate around the world. Concern about these spillover effects has led to efforts to strengthen the analytic capability of many central banks with respect to financial system stability. Central banks and investors now demand more and better information about the financial behaviour of both industrial and financial enterprises. After the Asian crisis of 1997-98, which highlighted the lack of transparency and proper surveillance of financial systems in several countries, more resources were dedicated by national authorities, including central banks, to understanding the workings of the financial system and to communicating that understanding publicly.

The policy objectives of the various national agencies that are involved in setting standards and codes for the financial system are safety, soundness, and efficiency. No economy can function properly, unless supported by a robust, efficient financial system and sound financial institutions that can help to appropriately channel savings and investments.

By their very nature, central banks take a systemwide approach to financial stability. So, our focus is on the nature and causes of vulnerabilities with potential systemwide implications. After all, we are paid to worry about these things! Once such vulnerabilities are identified, we would work with other standard-setting bodies to find ways to prevent or contain them.

In addition to the real estate price indexes that I discussed earlier, there is at least one other area where we could use your help to achieve our goal for the financial system.

To better understand financial behaviours and their implications for systemwide vulnerabilities, we need to link financial market data (new issues of bonds and equities, secondary pricing of bonds and equities, etc.) to industry or sector characteristics and to economic activity. We have found this to be a rather cumbersome and time-consuming exercise. This is where you can help us, by undertaking to link financial data to the firms involved, in a more transparent, systematic, and consistent way. In most cases, these would be firms to which you have already assigned an industrial classification code, and for which you have other relevant information (such as characteristics and surveys on their economic activities). The ability to cross-reference financial and economic data would allow us to explore in a more scientific manner questions of financial vulnerabilities from a systemwide perspective. I am, of course, aware of the privacy concerns, particularly in this area, and, hence, the need to find ways to deal with them.

It would also be useful if central banks and national statistical agencies could work together to define needs and to share expertise on financial statistics. In Canada, advisory groups of this nature have worked well in a number of areas.

Members of such groups can also share knowledge on international initiatives related to financial system data. As you know, there are a number of initiatives underway to improve such data from various perspectives. For example, the IMF is coordinating a project to develop national Financial Soundness Indicators. The BIS is looking to provide data on ultimate risk. And the Financial Stability Forum (FSF) is monitoring the consistency and comprehensiveness of international financial standards and codes.

I am now coming to the last part of my remarks. Here, I intend to stay away from specific data needs. Rather, I will talk about the general principles that I see as relevant in guiding your efforts and in determining statistical priorities as we move forward.

Don't Let the Best Become the Enemy of the Good

Central bankers have to make policy decisions in real time and often not under the best of circumstances. For this, we need the best available information.

To be sure, it is important that the data we rely on for those decisions be of high quality. But this does not mean that we should let our quest for high-quality standards prevent the publication of potentially useful data. In other words, we should never "let the best become the enemy of the good."

So, this is my advice: if you have imperfect data, don't sit on them. Put them out, together with your professional assessment of their quality and vulnerability. Remember, as policy-makers, we are used to taking decisions under uncertainty, in less than perfect conditions. We would rather have imperfect data than no data at all.

Data comparability

Data comparability—now, there's a big issue! With national economies and financial systems becoming more and more integrated, national central banks increasingly have to rely more on information and concepts from other jurisdictions to read the trends, and to figure out what is going on in their own economies. But for this to be effective, statistical agencies need to collect and aggregate data on a comparable basis.

Comparability is also very important in that we use cross-country variation as a way of identifying and distinguishing between hypotheses as to what is working and what is not. But if we cannot really compare the data, then we lose a major source of identification of the cross-country differences that are relevant to the decision-making process. We also lose a valuable yardstick for measuring our performance relative to other countries.

For an example, I will refer again to productivity measures, which, as you may have gathered by now, are at the top of my list—for good reason. I have already talked about their relevance for estimates of the economy's production potential and the implications for capacity pressures and inflation. But our interest in productivity measures is also driven by the important link between productivity growth and improvements in living standards. Over the past few years, a hot issue for a number of countries, including Canada, has been to understand why the trend growth of productivity differs across countries. Basically, we need to understand the reasons for these differences in productivity levels and growth, if we are to formulate appropriate policy responses. But if the data are not comparable, then we do not know how much of a problem we really have to begin with.

I have used productivity measures as an example to make the point about the importance of comparability across countries. But this applies equally to other key data including, importantly for us central bankers, various price measures. In particular, we need to understand deviations from "the law of one price." And so we see merit in, and support, the ongoing program by the World Bank on the international comparison of purchasing-power parities.

Importance of Co-Operation Among Statistical Agencies

It is primarily in the context of, and in the interest of, cross-country data comparability that I will make my final remarks today.

This is where the importance of co-operation among national statistical agencies comes in—in a big way.

There are several reasons for that, and I have already touched on them as I went along. But since I attach a great deal of importance to this issue, let me summarize them again.

In the first place, as I said before, the demands on our national statistical agencies for additional, more timely, and more accurate information keep growing. Given staffing and financial constraints, they need to use their resources more effectively. All the more reason then to work "smart" by pooling resources on how to address common data challenges.

Another reason is that, as firms become global, they report to many statistical agencies. It would be helpful to gain better understanding of their operations by pooling our resources. Right now, we are a bit like the fabled blind men describing an elephant!

Still another reason for co-operation among national statistical offices is in the interests of better cross-country comparability of data.

Statisticians often face a trade-off between designing data that best fit the particular structure and circumstances of their country and adhering to international standards that allow for better comparability across countries. In a perfect world, we would all like to see both sets of statistics. But in the presence of resource constraints, I would be willing to give up a little on the best fit for my own country in exchange for better international comparability.

One way or another, it is extraordinarily important that the Chief Statisticians of the world have the opportunity to get together to discuss and agree on common concepts, definitions, and methodology. That's a key step on the road to more comparable data. Not to mention, that such meetings also provide excellent learning opportunities for the participants—as, I am sure, this conference will, too.

I wish you all much success and many fruitful discussions—today and in the future. I will now be happy to take your questions.