The Cost of Common Ownership to the Singapore Government
Common ownership around the world has risen through the rise of intermediated financial investments. The jury is out on whether the increase in common ownership distorts the behavior of firms competing in the same markets. Using publicly available data from over 22,000 public auctions, my co-author Kentaro Asai at the Australian National University and I find that companies with common registered shareholders are more likely to submit identical bids.
In addition to investigating the correlations between common ownership and identical bidding, we simulate a counterfactual world where the relationship between common ownership and bid rigging were truly zero. The challenge with this simulation is to take into account the strategic interaction of all the firms participating in the same auction. We find that reducing the effect the collusion related to common ownership would decrease government expenditures in those auctions by almost 5%.
At first, we suspected that the relationship between common ownership and identical bids that we found may not be collusive. It could simply be due to rounding to the nearest dollar or hundreds of dollars. But even when we control for this effect, we obtain the correlations between common ownership and identical bids.
The findings in our recent working paper have been covered by the Business Times this past January and quickly drew a response from the Ministry of Finance, the body that oversees the Government Business Expenditures (GeBIZ) platform where about half of Singapore government expenditures take place. The spokesperson stated that “participation of related firms does not prevent others from submitting bids”.
But the spokesperson is mistaken. The microeconomic theory behind collusive identical bidding behavior does not rely on assumptions or produce results about any deterrence from market entry. It simply means that if the identical bid happens to win, the average price that the government pays would be higher than if the collusive firms had acted independently. A long line of academic research has shown that barring side payments, identical bidding is indeed the optimal collusive behavior for companies participating in the same auctions.
To see the economic intuition, suppose we consider an auction where only members of a cartel bid. The cartel’s goal is to maximize the profits across all firms. Suppose the cartel submits the identical bid and the government randomly allocates a contract to one of the lowest bidders. But the cartel could raise the price by agreeing to submit the same bid at a higher price. Because no rival exists, they can raise the price without losing demand. In this example, the cartel could theoretically raise its bid up to the maximal admissible price. The government’s expenditure increases with the submission of identical bids even though cartel members do not lose the probability of winning a contract.
The intuition continues to apply even when we introduce non-cartel members. In this case, the cartel would still submit higher identical bids than if they were competing, but it would be lower than if there were no non-cartel members in the auction. This is because the cartel doesn’t know how the non-cartel member will behave. None of this intuition relies on related firms deterring the entry of other firms. The end result remains that a collusive identical bidding ring would tend to raise prices if the identical bid happens to win compared to a world where all bidders are competitive.
Submitting identical bids has been widely viewed as an indicator for the potential presence of an agreement among suppliers. In fact, the Organization of Economic Development (OECD) adopts this stance with even more precision: recommending governments “avoid splitting contracts between suppliers with identical bids and investigate the reasons for the identical bids and, if necessary, consider re-issuing the invitation to tender or award the contract to one supplier only”.
On its website, the Competition & Consumer Commission of Singapore states “encourage all businesses to set their prices independently.” But it does not explicitly outline any penalties from collusion.¹ Although identifying whether identical bidding is collusive or accidental is difficult, the implication for government agencies is much simpler.
Individual government agencies should at least look at the ownership structure of the firms in the auction. Even if those firms are not actively deterring others from participating, identical bidding imposes a higher cost for taxpayers. The burden of checking for identical bids and asking for explanations of identical bids lies on the officials evaluating the different bids.²
Our findings also have broader implications beyond Singapore.
The common-ownership concentration of publicly listed firms has increased due to the increased intermediation of ownership through both active and passive mutual funds. A burgeoning academic and practitioner research finds that measures of ownership concentration are correlated with anti-competitive behavior, like increases in product prices and mark-ups. A difficulty for regulators around the world was pinning down the mechanism for the anti-competitive behavior. In our setting, we document a robust set of empirical results that are grounded in economic theory. Although people may argue that a cost of 5% to the government is not large, any reasonable government should still attempt to minimize costs.
To see all of our results in detail, including all regression analyses and details on our counterfactual simulation, see our freely available paper here.
¹Moreover, only half of the auctions are those where the lowest bids win. We try to understand the government’s preferences using a machine learning algorithm, which suggest that prices still matter, although a government agency may choose to award a contract to a non-lowest bidder if it has other auctions going on at the same time.
²Of course, rather than collusion, companies can directly corrupt the government official as well. Although we do not focus on this, it could certainly occur. In our data, around half of the auctions are not awarded to the cheapest bid. However, this could simply be due to quality differences in the products and services.
