Finally, there are some related papers that consider how entry costs maybe necessary
to complement the gains from reputation in markets where there is free entry. Klein
and Leffler (1981) shows that if there are moral hazard issues, firms need to earn
price premiums in order to incentivize them not to cheat. With free entry, such price
premiums cannot be sustained in equilibrium, and so it is important to have fixed entry
or investment costs to support these premiums. Atkeson et al. (2012) also consider a
reputational model with free entry. In their paper, producers make quality decisions
at the time of entry, and these qualities are then fixed afterward. The market learns
over time about quality via signals as in our paper, and the government can impose
fixed entry costs for entry into the market. These costs do not produce information as
in certification, but do allow for higher-quality firms to enter as in Klein and Leffler,
leading to greater social welfare. While we do not explicitly model free entry in our
paper, certification is costly and can be thought of as an entry barrier to a market.
Indeed, we show that in some markets certification will act as a de facto license,
with non-certifying agents unable to ever find work. However, there are two main
differences between certification and the entry costs in these other papers. The first is
that certification provides information to the market, and so serves more than just a
signaling role since not all agents can pass certification. The second is that certification
can also happen after an agent enters the market, with the agent choosing to certify
only if it sends bad signals. Thus, very high-quality agents may not need to pay the
certification cost if they produce good results, leading to a potentially more efficient
outcome than one where all agents must pay an entry cost.