On Innovation

On structural changes to encourage risk-taking -- a proposal and a rejection

i. Introduction

It seems reasonable to believe that we underinvest in innovation. Attempting to find new production processes creates a lot of value which is not appropriable by the investor, even in the presence of legal constructs, such as patents, to try and correct this. The full value of innovation includes not only the direct benefits of the invention, but also some part of every invention informed or inspired by it — a hopeless task. Moreover, the risk of failure cannot be insured against – if you knew that if you don’t succeed you get a payout, it greatly increases your incentive to shirk. 

So, there is a strong case for government intervention to increase innovation. But how? The government cannot tell ex ante what lines of inquiry will be successful — though, to be fair, no one can — but its difficulties are especially strong since it lacks the tacit knowledge that firms and individuals have. It runs the risk of subsidizing a target which loses its correlation with the output. Its attempts to keep people focused on productive research, such as grant applications, are expensive and cumbersome. 

ii. The proposal

Perhaps, however, the government could subsidize risk-taking per se, without regard to the form that those risks take. The government could grant favorable tax treatments to payment in the form of options to buy assets at a rate above the prevailing rate. Suppose the price of company X is $100 dollars per share, and its CEO received as part of his compensation an option to buy at $110, on some date in the future. (A so-called “European” option). This option is valueless, if the price doesn’t rise above $110, so all else being equal, the executive will prefer greater risk. This also works with an option to sell at a price below, although not as well — in the limit, the company could go bankrupt and there be no one to sell it to. For that reason, I will examine only the case of subsidizing the giving of buy options as a form of compensation, as I do not wish to extend my advocacy to the uncertain. 

iii. Some problems, and solutions

The trouble is that firms could simply make all compensation, for all employees, options, and thus avoid taxes entirely. Second, employers could offer options which mature immediately at a price below the present market value, thus being functionally equivalent to pay, but tax free. What we can do is require that it be only a buy option at a price some fraction above the current market price. How would we figure out the market price, though? Couldn’t the shareholders sell it to another entity at an artificially low price? And how are we to value assets for which no liquid market exists? 

What one could do is a variation of what has been proposed for the assessment of property taxes, and to provide public goods in ancient Athens. Executives will state what they believe the asset has value p at the beginning of time period t, and if they are concurrently given an option to buy at price p+x% at time period t+1, their profit from that transaction is non-taxable income. The catch is that they are required to immediately sell the asset to anyone who wishes to buy it at their stated price (valid only on the day that the declaration is made, which need not be the same for all companies). You do not need to make such a declaration, but then you are ineligible for the tax credit on options. You need not put up the whole stake, although you only have one opportunity per time period to declare the stocks value, but the assets must be homogenous, and can be limited to just a percentage of the company with the same privileges.

iv. Ancillary Advantages

This sort of forced sale can have efficiency advantages outside of just increasing the level of risk-taking toward the socially optimal level. The following is drawing from the tradition of George Akerlof’s “The Market for Lemons”, (1970), and Milgrom and Stokey’s “Information, Trade and Common Knowledge”, 1982. Suppose that one can invest money in finding information, such as finding which stocks are undervalued or could be more profitable under new management. You acquire private information, which leads you to make an offer to buy the stock. Making the offer, however, reveals the private information to some degree. It encourages the owner to raise their price to extract some of the surplus that the buyer would be able to get – which discourages ever finding that information to begin with! All people, both buyers and sellers, would prefer a world in which there is undiminished effort into finding information, and the buyers make offers which the seller would find acceptable – the individual opportunism of sellers puts us in a worse equilibrium. This goes in the opposite direction too – if someone wishes to sell out of the blue, it indicates a lack of confidence in their position. By having a reason for them to put a price out, you can make buyers more confident in buying. 

v. The final problem

Another difficulty lurks, which I do not have a solution to as of yet. I post this blog as something of an open problem, and will attempt to formalize what the tradeoffs appear to be. Suppose someone is the sole owner of the assets of a company. They could create a second, essentially fictitious company, and value it at a level clearly far above the fundamental value of the company. Then, upon receiving the tax preference, you transfer the assets of the first company over to the second. Another person could attempt to buy the company when it is proposed for sale, but there is nothing keeping the first person from simply not transferring the value over.

We could make it so the period t is shorter than the period in which assets can be transferred. We run into a problem of transaction costs, however. The period I had in mind originally was perhaps a year – we want assets to be transferable extremely quickly, so the periods would have to be even quicker. It is costly to create a prediction of price; besides, we want people to take risks which may take years to implement, such as research and development. Perhaps we could regulate it, such that massive transfers of that sort are discouraged. This would, however, mean that this option is only available for a minority of firms, or that the small firms can evade taxes with near impunity – an option no one would want. 

vi. Takeaways

So a system of subsidizing buy options may not work directly. Still, we must find some way of incentivizing innovation. It is plainly a massive public good, and we should be willing to accept some inefficiencies to get more of it. The world depends on it.