To combat food deserts in Chicago, the city government is considering opening up a municipal grocery store. The Wall Street Journal recently had a piece about this initiative. In it Joe Barrett tells the stories of a municipal grocery store in Erie, Kansas, which last year was $123,000 in the red and a municipal grocery store in St. Paul, Kansas, which is profitable and has been in operation for 16 years.

Similar municipal grocery stores exist elsewhere in the country, including in Baldwin, Florida, which was profiled in the Washington Post four years ago. In each case, a generally conservative rural municipality opens up a grocery store because nobody else would and because they feared that not having a grocery store could result in a death spiral of depopulation.

This is the story of a lot of state-owned enterprises in the US and elsewhere in the world. Governments frequently resort to operating their own enterprises only when the private sector fails.

For advocates of public ownership, this is a little unfortunate as it means that SOEs are disproportionately concentrated in extremely challenging business situations, meaning that losing money is more likely than not, which then gives rhetorical fuel to those who are ideologically opposed to public enterprise. If governments were more inclined to focus on average or better-than-average business opportunities, then the average financial outcome of SOEs would be better than they generally are. Indeed, when governments do run SOEs in highly lucrative sectors, like oil and gas, they tend to be cash cows and the knock against them is just that they are playing on easy mode.

But beyond the discourse, it’s obviously understandable why SOEs end up concentrated in difficult business situations. Unlike private businesses, SOEs can decide that a particular enterprise serves a social or policy purpose that justifies a negative cashflow. These kinds of social-purpose SOEs still generally would like to have a neutral or positive cashflow, but it’s not strictly necessary so long as the negative cashflow is worth the achievement of the underlying social purpose.

In the US, the most prominent example of an SOE with a social purpose is the United States Postal Service (USPS). Depending on how you do the accounting, the USPS generally breaks even or runs a slight deficit. This is often touted as evidence of its failure, but the USPS does not operate with a solely commercial purpose. It also has the non-financial goal of ensuring that every address in the country can receive postal services at an affordable price, which is known as the universal service obligation (USO). The government has deemed that the achievement of the USO is worth the toll it takes on the USPS’s financial outcome.

The fact that SOEs can and often do have social purposes in addition to commercial purposes can also lead to some sloppy arguments from SOE proponents in which every SOE failure or inefficiency is waived off as being justified by the social purpose. But just as it is wrong to suggest that financial outcomes are the only measure of an SOE’s success, it is also wrong to suggest that they are totally irrelevant to an SOE’s success.

In order to evaluate whether an SOE is successful, governments like Chicago should first establish a general framework for state ownership and a specific purpose for each SOE. This framework and purpose can then be used to evaluate whether the SOE is meeting expectations while avoiding the trap of fighting about these kinds of things after the fact using ill-defined, shifting, and contestable criteria.

The best framework I’ve seen for state ownership comes from Norway. There, the government has placed each of its 69 SOEs into one of two categories. The first category is “companies where the State’s goal is the highest possible return over time in a sustainable manner.” These are generally cash-cow SOEs concentrated in natural resource and natural monopoly sectors. The second category is “companies for which the State’s goal is … the most efficient possible attainment of public policy goals.” This category is quite diverse and includes companies dedicated to ensuring universal access to things like arts and transportation even when doing so is money-losing in certain areas.

In addition to assigning companies to these two categories, Norway also outlines the specific public policy goal and special framework for each category-two company, which then allows it to effectively evaluate whether the SOE is achieving its public policy goal in the most efficient way. Within this framework, losing money is not necessarily inefficient, but losing more money — i.e. undertaking more costs — than is necessary to achieve the public policy goal is.

Chicago should use a similar approach for establishing what it is hoping to accomplish with a municipal grocery store. From what I have read so far, it appears that the government is conceptualizing the store as not having a merely commercial purpose where the goal is the highest possible return. Rather, the store is being assigned the public policy purpose of ensuring that all or most Chicagoans live within a certain proximity of a full-service grocery store. This makes it a category-two company within the Norwegian framework above and its performance should be evaluated according to whether the store achieves that purpose in the most efficient possible way.

Again, this does not necessarily mean that the store must have a positive cashflow. It may be impossible within the prevailing conditions to achieve the public policy purpose while also achieving a positive cashflow. Indeed, the fact that other stores are not operating in these areas suggest that it will be very difficult to do so. But it should be aiming to have the best cashflow outcome that is compatible with the achievement of the public policy purpose.

Defining the goals this way will not totally protect the project from bad-faith criticisms that attempt to declare the project a failure based on goals that the project never aimed to accomplish. But it should provide some protection from that and also serves a useful organization and managerial purpose. Clearly defined goals and purposes do not guarantee SOE success. But the lack of them more or less guarantees failure.