One of the greatest challenges of lawmaking is recognizing when a beautiful theory must succumb to an ugly fact. The purity of conceptual policymaking must take the real world into account and acknowledge that things that work well “in theory” sometimes fail spectacularly to meet expectations “in practice.”

It would be wonderful if a policy met its intended goals 100 percent of the time and there were never any unintended consequences. Unfortunately, we live in a reality where complex tax policies that work well “in theory” can have a very hard time when the rubber meets the road.

One instance of this is the challenge that the Organisation for Economic Co-operation and Development (OECD) has created for itself with the global tax deal, also fondly known as OECD Pillar 1 and OECD Pillar 2 (global minimum tax).

In the last month, model rules have been released on parts of Pillar 1 including how taxable profits will be allocated to different countries and how the tax base will be designed. The OECD is expecting those model rules to be agreed to and a multilateral convention (tax treaty) to be ready for signature by midyear.

However, the rush to finalize the rules is likely to sacrifice the work necessary to make Pillar 1 workable.

Let’s start with the goals of Pillar 1 Amount A, part of the new allocational rules changing where some large multinationals pay taxes. The new taxing rights under Amount A apply to companies with more than €20 billion in revenues and a profit margin above 10 percent, a definition that likely captures fewer than 100 companies. For those companies, 25 percent of profits above a 10 percent margin will be taxed where final consumption occurs. After a review period of seven years, the €20 billion threshold will fall to €10 billion.

Amount A is partly inspired by what is known as “formulary apportionment.” Such a system would take taxable business profits and allocate those across jurisdictions according to a formula. For Amount A, that formula relies on the number of sales where final consumers are located.

The general point of sales-based formulary apportionment is to move from taxing companies where they have employees and operations to taxing them where their products are consumed. For example, if a company makes a product in the United States, but then ships that product to French customers, the profits from that transaction would be taxed in France rather than in the U.S.

In theory, the various allocational rules Amount A incorporates follow the logic of achieving sales-based apportionment. In practice, however, they would be very difficult to implement, and in some cases nearly unworkable, particularly given the rule’s narrow scope.

Policymakers want to apply a sales-based formulary apportionment rule that only applies to a small set of companies while trying to allocate profits to the jurisdiction of final consumers. The challenge comes in requiring companies to identify the location of the customers of their customers for every transaction that is made.

Imagine that instead of only applying the rules to a small set of companies, policymakers choose to apply the rules equally to all companies and all profits. Each company would identify the location of its immediate customers and allocate a portion of its total profits to the jurisdictions where those customers are located. If a company were simply selling to another business (say a component for a car), it would not have to identify where the final car purchase is made. That responsibility is for the companies further down the supply chain and closer to the customer.

This is how U.S. states apportion taxable profits under sales factor apportionment. (All states with corporate income taxes use sales as at least one factor in formulary apportionment, though some also take the location of payroll and property into account.) Companies look to the location of their immediate customers and owe taxes in the jurisdictions those customers are. Because all companies are required to play by the same rules, no single company in a supply chain would have to know where all the final sales are made. This can still be complex in its own right, and, depending on the state, occasionally involves some “look-through” analysis when a consumer uses a service in multiple locations, but it never requires a company to follow their good or service all the way through the supply chain.

Policymakers have chosen a different path for Amount A by requiring fewer than 100 multinational companies (at least initially) to file taxes in alignment while still attempting to meet the goal of taxing profits in the location of final customers.

To reach that goal, a business might have to ask its clients for data on the businesses or individuals it sells to. Or ask its clients’ clients for that data. And on down the supply chain. This presents a real challenge for compliance.

The rules recognize this challenge and provide companies with a “proxy” that can be used when identifying the location of a final customer is not possible. If a company cannot identify the location of its final customers, it can use macroeconomic data on consumption expenditure to allocate its taxable profits.

Imagine that a U.S. company sells its products to a distributor based in Ireland, and that Irish distributor delivers products to countries throughout Europe. The distributor may rely on further distributors and ultimately may not have the data on the location of final consumers or be able to share that data with the U.S. company. In such a case, the rules allow the U.S. company to allocate Amount A profits to European countries based on each country’s share of consumption expenditure.

This could simplify things considerably. Rather than tracking down individual customer data all the way through the supply chain, the company can just look at UN data on final consumption expenditures and allocate taxable profits to France, Germany, Spain, or other European countries based on their shares of consumption.

The proxy is not perfect, however. The U.S. represents 28 percent of global consumption spending according to UN data. This is the largest share for any one single country, and this could benefit the U.S. Treasury.

However, total consumption represents something slightly different than what Amount A is trying to capture. Amount A is focused on cross-border transactions that mainly involve imports and exports. The U.S. has a much smaller share of world imports at 13 percent. Consumption of domestically produced goods and services inflates the share of Amount A that the U.S. might claim relative to imports or economic reality.

The OECD should not create a false sense of reality simply for the sake of providing a proxy for identifying where final customers are located. It is obvious that some companies will not be able to specifically identify the location of their customers, particularly if the final consumer is many steps further along in a supply chain. This provides justification for a proxy. But a proxy should at least capture something closer to a measure of cross-border trade than domestic consumption. The proposed rule, moreover, creates substantial uncertainty for taxpayers, as these companies will be forced to make judgment calls on when location data are truly unavailable—and the use of a proxy permissible—and when costly efforts to ascertain ultimate customer locations must be undertaken to comply with the rule.

The clock is ticking on these negotiations. Policymakers are quickly ironing out the details on Amount A and other related issues.

It is difficult to make the transition from a theoretical option to one that works in practice, but the OECD is making that challenge even worse by setting up a formulary apportionment system that is unlike even the ones that work even halfway decently as in many U.S. states.

Moving too quickly could easily create more trouble down the road.