If it were possible the benefits of replication would be lower fees and therefore higher returns to the investor. Today, a real benefit is as a benchmark to differentiate strategy alpha from strategy beta. While traditional managers have generally had a tough index to compare themselves to, hedge fund managers have been able to have until now the relatively easy to outperform benchmarks of zero or a risk free hurdle. Replications might allow identification of which funds are skilled in running a strategy and those that are not. That is the primary existing advantage for investors.  

3. How can hedge fun returns be replicated synthetically if the essence of the hedge fund industry is to provide managers with the freedom to add value through unfettered skill?
 
I have seen little evidence that they can be replicated synthetically. The marketing literature for all the existing and proposed hedge fund clone products leaves much to be desired as to the statistical rigor and robustness of their models. It is easy to find factors that predicted the past but it will be many years before satisfactory evidence is available as to the utility of these untested products. The essence of hedge funds is dynamic varying dependence on market factors and the ability of good funds to generate alpha has been proven over many years. The capability to research and innovate new strategies, trade new assets and derivatives and adapt to changing markets is something clones are unlikely to have.
 
 4. What are the constraints and challenges of replication?
 
The main constraint is that clones have not been battle tested particularly in difficult market conditions. A significant and sustained bear market rather than a quick correction will form a better acid test to see if these can still generate absolute returns like a proper hedge fund. Other challenges include investor education and the poor performance of previous hedge fund "innovations" like investible indices. There is also the question of which hedge funds to copy; with alpha a zero sum game the performance of the average hedge fund is unlikely to be impressive. Transparency is often the enemy of performance in both strategy and position disclosure terms therefore clones may end up emulating weaker hedge funds that provide too much information. A typical hedge fund can probably be cloned but not the best hedge funds.