EXPORTS AND IMPORTS OF R&D[1]

This paper stresses the need of collecting data on R&D exports and imports; lists the main sources for it; analyses part of the many forms that may adopt international explicit and implicit transactions on R&D and their implications for data collection; and, explains the convenience of using the special surveys on R&D of the Frascati Manual’s compilers to collect these data, in collaboration with national accountants, making R&D data more useful for R&D analysts and policy makers and improving the quality of these data in the national accounts.
The need of data on R&D exports and imports
The analysis of the role and impact of R&D in the economy requires data on the total supply of R&D and its different uses. Also the possible capitalization of R&D in the central framework of the national accounts or in satellite accounts, is facilitated by applying a supply and use balance method. For this method to be applied, ideally, independent estimates on the supply of domestically produced and imported R&D, as well as on its different uses, must be generated and balanced in the framework of an R&D supply and use account. When each one of the components is independently estimated, it is necessary to balance the sum of the supply components with the sum of the different uses. Capitalizing R&D requires, obviously, data on exports and imports of R&D, since not all the domestically produced R&D will be capitalized in the country (also other uses exists, including exports), and not all the capitalized R&D will be produced in the country.

Main sources for international R&D transactions.

R&D is defined and included in various international classifications. As an activity it is included in ISIC and other activity classifications; as a purpose of the non-profit institution serving households, in COPNI; as a function of the government, in COFOG; as an outlay of producers in a specific purpose, in COPP; as a product, in CPC and other products classifications. The existence of correspondences among different classifications permits to follow and link R&D data.

International R&D transactions, are explicitly recorded in the following statistical systems and classifications:

In the Balance of Payments, fifth edition (BOP5), R&D is included, but only as a supplementary item to its standard components. It appears

1. Current Account

A.  Goods and services

-----

b. Services

----

9 Other business services

---

9.3 Miscellaneous, business, professionals, and technical services

----

R&D is explicitly included in a separate table on selected supplementary information (BOP5, p 50), as one of four components of “Miscellaneous, business, professionals, and technical services”. Being only a supplementary item, many countries do not register separate R&D data.

R&D is described in BOP5 as covering “those services that are transacted between residents and non-residents and associated with basic research, applied research and experimental development of new products and processes. In principle, such activities in the physical sciences, social sciences and humanities are covered, including the development of operating systems that represent technological advances.” (BOP5, par. 264). The description is based on the one given by the Central Product Classification (CPC).

R&D is included in the Joint OECD-EUROSTAT Trade in Services Classification, which is a more extended classification of international trade in services than that of the BOP5. This classification was developed in 1996 by OECD and EUROSTAT, in consultation with the IMF.

R&D is also covered by Extended Balance of Payments Services Classification (EBOPS), which is based on the mentioned joint classification, and included in the Manual on Statistics of International Trade in Services (MSITS), published in 2002.

National accounts supply and use tables, containing a separate row for R&D, shows R&D exports and imports, respectively recorded as a use and a supply component. For reasons given below, data on these flows may be undervalued.

R&D exports and imports may be recorded also in some of the special R&D surveys carried out by FM’s compilers, and also in some satellite accounts on R&D.

Difficulties of collecting data on international R&D transactions

R&D may be defined and included in various classifications on international trade, but detecting and recording cross-border flows of R&D is very difficult, due to the many forms that they may adopt. Among them:

1 Direct international acquisitions of produced R&D, through international contracts to perform R&D or other forms.

The ownership of the outcome is transferred by the producer to the buyer, being one party a resident and the other a non-resident. The parties to the transaction are independent one from the other.

2 A subsidiary is created in the country by a non-resident corporation, to perform R&D activities, outcomes of which are transferred to the parent corporation.

The foreign parent corporation becomes the legal owner of the produced R&D. It might patent the acquired R&D or maintain its exclusive ownership through secrecy. The name of the subsidiary’s employee that created the R&D outcome may be registered when patenting it. The resident subsidiary will usually not use the transferred R&D in its own production processes.

This form has growing importance in Israel and other countries. Multinational corporations establish subsidiaries to produce high-tech R&D activities, due to the availability of highly trained researchers and the relatively low salaries they are paid in the subsidiary’s country. The R&D output is generated in the country and provided to the parent corporation through the transfer of the corresponding new knowledge, to be patented and/or applied by its establishments in its country of residence or by subsidiaries established in other countries.

The production costs of the resident R&D subsidiary are covered by the parent corporation. In some cases, the subsidiary reports -in its financial statements- that R&D activities are charged to the parent corporation on the basis of costs plus a mark-up. Most of the observed mark-ups fluctuate between 7.5 and 15 percent. In these cases the balance of payments registers commissions received for exported services. In cases in which the transferred R&D is not recorded, the funds received by the subsidiary will be registered as direct investment from the foreign parent corporation. Output is generated but, not being sold, it will probably be recorded as increasing the subsidiary’s inventories. If an estimated value of the implicit exports of R&D services is now included in BOP, duplications should be avoided when registering the financial flows.

3. R&D generated by teams of researchers working together and residing in different countries but employed by the same multinational corporation.

Like in the previous case, the owner of the R&D output is the foreign corporation and the output generated by the subsidiaries is sent abroad. The generated output, and corresponding exports, must be allocated among the different countries.

4. Joint R&D, through collaboration between independent units located in different countries.

The jointly generated R&D output must be allocated among the participating units and the way in which the ownership of the R&D outcomes is shared by participants should be taken into account.

5. Acquisition of R&D enterprises by non-resident firms (mainly, but not only, successful start-ups) with the aim of acquiring the ownership of produced R&D and the potential future production of R&D, among other possible purposes.

This case is relatively similar to the creation of a subsidiary mentioned above. It differs by the fact that the already produced R&D is implicitly exported, since it is usually transferred to the new owner, and used abroad. As in the other case, the ownership rights of the parent corporation are established legally by patenting the transferred R&D outcome, or maintained through secrecy. The exported R&D is not registered as such by BOP, as the whole transaction is recorded as a direct investment in the financial account. The start-up could be producing for various years R&D output for sale that, being neither sold nor used, is accumulated as increases in inventories. When “sold”, the accumulated R&D output will have to be registered as exported, with a counterpart negative entry in the resident enterprise’s capital account.

6. R&D internationally traded on-line.

On-line trade is not registered in many cases, so that data on internationally traded R&D are undervalued.

7. Cross-border acquisition of patented R&D.

Acquiring a patent of an R&D outcome means acquiring the ownership of the produced R&D. If R&D would be capitalized, the acquisition of the R&D patent, a non-produced asset, implies the simultaneous acquisition of the produced asset protected by it. There is an ongoing discussion on the existence of one or two assets in cases like this. The relation, in the national accounts, between patents and the assets protected by them is one of the issues that the Canberra II Task Force will discuss.

8. Royalty payments and license fees for the right to use generated R&D.

A country may benefit from using R&D performed in other countries by paying royalties and fees. In this case no change in R&D ownership takes place. What is paid for is the right to use previously developed knowledge. The SNA93 states that the transaction is to be registered as intermediate consumption by the buyer of the rental service and as an output sale by the owner of the patented, or otherwise protected, R&D.

This type of international payment for the right to use R&D is practically impossible to observe as an isolated item in international classifications. BOP5 includes “Royalties and license fees”, as one single standard component, under goods and services. MSITS recommends a disaggregation of this component into “franchises and similar rights” and “other royalties and licensed fees”. The first component covering “international payments and receipts of franchising fees and royalties paid for the use of registered trade marks”; and the second “international payments and receipts for the authorized use of intangible, non-produced, non financial assets and property rights (such as patent, copyrights and industrial processes and designs) and with the use, through licensing agreements, of produced originals (such as manuscripts, computer programs and cinematographic works and sound recordings)” (MSITS, par. 3.121).

The above list does not pretend to be exhaustive. It is important to note that only part of the different forms that may adopt the cross-border flows of R&D is registered as international transactions by the existing statistical systems and other data sources. Also, not all of them imply international ownership changes of produced R&D output, that should be included as exports and imports in an R&D supply and use account. Another important topic to stress is the need to clarify the borderline between R&D and software (in many cases the direct outcome of R&D activities), in order to avoid duplications.

A possible new data source for R&D exports and imports

The implementation of the Frascati Manual (FM) by the OECD countries for many years provides the largest and most complete available database on R&D today. Unfortunately, it does not recommend to separately record exports and imports of R&D. Preferential attention is given to the measurement of the Gross domestic expenditure on R&D (GERD) and its financing sources. Special attention is given to the sectoral composition of the financing sources, but not to their disaggregation by economic categories. This affects particularly the possibility of recording exports and imports of R&D, as they are included as financing sources

The financial resources obtained by selling domestically produced R&D to non-residents (exports), are included in the financing flows of intramural R&D activities, together with received donations and other transfers. The financial resources provided to non-resident performers by acquiring R&D generated by them (imports), are recorded as extramural expenditures, together with donations and other transfers, financing the intramural R&D activities of non-resident performers. As a result no data on international trade on R&D are recorded.

Not recommending the separate recording of R&D exports and imports could be based on the difficulties to measure them, which is quite justified, given the many forms that international trade on R&D may adopt, as mentioned above. But the FM compilers should be asked to do it. In effect, they compile R&D data through special surveys carried out among R&D performers. They have to answer questions on how their intramural expenditures are financed. Since they should include the value of exported R&D in the funding sources of intramural expenditures, they are the best source for registering and valuing their R&D exports.

The same happens with imported R&D. In effect, the value of imported R&D is included as part of the funding provided by the country to non-resident R&D producers. It is true that the R&D acquisition reported by the buyer may overvalue the effective R&D component of what he buys, but it is doubtful that a better source could be found for measuring imported R&D.

In our experience, the special R&D surveys collect separate data on imported R&D. Sales of R&D abroad are not asked for in these surveys, but we intend to do it. Other countries may be collecting data on both. However, even if the FM compilers would collect these data from R&D performers, the same type of data for non-R&D producers, will not be covered. A different source will be needed for them, and the possibility of using for this purpose the current business surveys or other means could be explored and discussed by the TF.

Recommendations

1.  A questionnaire should be sent to the OECD member countries and some non-members, asking for: Available data on R&D exports and imports, data sources used and methods applied, evaluation on the quality of the collected data, and potential sources and methods that could be used to improve their quality. The questionnaire should be sent to compilers of the FM and the national accounts.

2.  The TF should suggest that FM compilers analyze, in collaboration with national accountants, the use of their surveys to R&D producers, to collect also data on R&D exports and imports, if they are not doing it. For non-R&D producers, not covered by the mentioned surveys, the TF could explore and suggest other means to collect data on these international transactions. It should be taken into account, that for some countries these transactions may not be relatively significant.

6

[1] Prepared by Pablo Mandler, Central Bureau of Statistics, Israel