There are almost as many opinions on tawarruq as there are people who know what the term means. With questions being raised about whether a recently launched Goldman Sachs murabaha sukuk was a tawarruq or not, the issue has been thrust into the forefront of discussions within Islamic finance*. A discussion of some of these issues are available from Camille Paldi (posted as a PDF).
This article had a link to a page from ISRA, which listed a number of different fatawa from different Shari'ah boards about the permissibility of, and structure of, tawarruq transactions. I, of course, cannot say whether a tawarruq is permissible or not. However, reading Paldi's article, and the Shari'ah opinions about tawarruq, there are some points which I find noteworthy about tawarruq.
The basic structure of a tawarruq (as used in the financial industry) is that the bank will buy a commodity (e.g. nickel from the London Metal Exchange). Once ownership transfers to the bank, it will resell the metal the the customer for the price paid, plus a profit, with repayment either in installments or as a lump sum. In either case, repayment will occur in the future. The customer then takes ownership of the metal and sells it to a metal broker to get cash.
In terms of economic outcome, neither party owns the metal, but the customer has X dollars in cash, with an obligation to repay the bank X + p dollars (p is the profit). In this example, the bank and the client can be switched for a tawarruq-based deposit product. During the entire transaction, the only time the metal involved makes an appearance is in the initial purchases and sales. One fatwa said: "It is compulsory to base this practice on the purchase of assets, which are then sold to the party who desires financing, i.e. the customer. The customer will then sell it to another party at a price that is agreeable to both parties. The delivery and payment is concluded on the spot".
When the final sale is completed, another fatwa reads: "One of the conditions of a legitimate tawarruq contract is that the bank must not guarantee the customer a specified price in the market but sell the goods at the best price, in accordance with the forces of supply and demand at the time of sale." One of the issues raised by multiple fatawa was avoiding instances where the bank (or its broker) is both selling to the bank and buying back.
These Shari'ah issues are interesting--I cannot offer an opinion about what they mean for the Shari'ah-compliance of any individual product--but they do raise a few issues on the financial side (e.g. if one were looking at the way transactions were actually carried out). First, the basis of the organized tawarruq product relies to some degree on ensuring that the price paid for the commodity is equal to the price at which it can be sold. The commodity markets on which these tawarruq transactions are based can be volatile, even on an intra-day basis, so an Islamic bank acting on its own behalf to buy the metals and later as agent for the customer to sell them, is exposed to the risk that the price of the commodity rises or falls.
Theoretically, I assume that the bank is supposed to pass that price risk on to the customer, who owns the metals between the initial purchase and the final sale. However, many customers, particularly retail customers, will not want to assume the price risk of the underlying commodity. There may even be regulations that limit the ability of the bank to change the effective profit rate after the contract is signed (the effective rate would change if the markup is fixed and the price realized is different than the original cost of the metals; if the price fell, the rate would increase while if the price rises, the rate would decrease). Yet, the bank cannot guarantee the price realized on the sale of the commodities lest it become both the seller and the buyer (as an intermediary).
Besides the operational issues with tawarruq, there are more theoretical arguments about its use. Initially, tawarruq was used for short-term inter-bank financing, where there was no other alternative available (there still are only a few, which was the rationale for forming the International Islamic Liquidity Management Corp.). Then the product became a consumer financing product. Now it is being used to structure deposits. The theoretical question is whether tawarruq ties the hands of Islamic banks to structures that (at slightly higher cost) replicate conventional banks.
I am generally in favor of Islamic banks using products that meet a banking need from Muslims who would not otherwise engage in the financial system. I think that makes sense. However, I also believe that creating an exact replication of the conventional financial system, or even the banking system, albeit with more transaction costs, is not the best goal for Islamic banking. There is a middle ground between making products available to meet consumer need and turning the entire exercise into a problem-solving exercise for clever bankers and lawyers.
There is, I think, a valid place for tawarruq in the development of Islamic banking. It is a useful tool where other tools are not yet designed. For example, there are few products available for education financing from Islamic banks and the benefit from providing this type of financing to people who will not use interest-based financing is high, both for the individuals and for society as a whole. This is the proper role of organized tawarruq (in my opinion) within Islamic banking. It is a fine tool, but Islamic bankers should not let tawarruq be the only tool (as the expression goes, when all you have is a hammer, everything looks like a nail).
* I wrote an article on the sukuk for The Islamic Globe, which was followed by a critical article about the sukuk program, which was followed by another critical article with a rebuttal by the firm which provided Shari'ah consulting services to GS.