PensionReforms' summary and comments
This 2007 report looks at the investment strategies of a sample of employees who worked for employers that are listed on the Taiwan Stock Exchange (TSE).

"We utilize comprehensive data from Taiwan to show that employees of TSE-listed companies invest about one half of their portfolios in their employers' stocks."

Specifically, that is 47% of their share portfolios (rather than their total portfolios).  Taiwanese apparently tend to have more of their household assets invested in shares (24% in 1998) than in bank deposits (12%).  PensionReforms notes that this may have changed since.  1998 now seems a long time ago and it will be interesting to understand what might be different today.

Employees who invest in the shares of their employer magnify the risks associated with their employer's failure.  Not only is there an occupational risk (loss of job) but there is also an investment risk.  The report also identifies a return risk:
"The economic cost of [investing in employer shares] is considerable.  Investors on average give up 4.89 percent of raw returns per annum by holding their employers' shares, equal to 39.74 percent of their 1998 salary income."

There seems to be no direct reason for the employees' willingness to take on the additional risk and, as well, give up investment returns:
"Such allegiance to employer stocks cannot be attributed to executive option compensation, [Employee Share Ownership Plans], sponsoring policies by employers to own company stocks, plan designs, or private information.  Instead, behavioral biases, such as availability and salience heuristics, inertia, over-confidence, and over-extrapolation are possible reasons behind the phenomenon."

PensionReforms notes that the apparent over-investment in employer shares is less significant because the employees' "portfolios" are a relatively small proportion of total assets (about one quarter).

The data used in the report come from tax filings that are matched to listed employers. For the measured year (1998), there were very few formal share-purchase schemes or share option schemes run by employers so the employer share investments were direct holdings.  So, what the data show is only direct investments in the employers' shares compared with other direct investments.  Indirect investments (such as pension scheme holdings - apparently uncommon in 1998) or even the entitlement to the Tier 1 pension would not be covered.

The report brands the apparent preference for the employer's shares as "severe under-diversification" and "sub-optimal".  It then suggests that this apparently natural preference could undermine moves to privatise state pension arrangements.

"Any attempt to 'privatize' social security must be based on very careful consideration of individual behavioral biases and potential mistakes.  Unnecessarily risky investments can result in a loss of security after retirement and impose consequential problems to financial markets and social stability.  The lessons from Taiwan clearly stress the need to provide well-diversified alternatives for individuals to invest for their retirement, given their behavioral biases and failure to diversify in their autonomous accounts."

The report concludes with a call for further work on why employees seem to be so inclined before we can discuss ".the best way to help investors avoid the bias."  Apparently, the employees' "welfare" should be the main aim of such research.

And then there is the seemingly inevitable need to protect employees against themselves.   

"Finally, future research should come up with specific mechanisms that can limit employee's enthusiasm for their employers' stocks.  It is worth emphasizing that employees hardly have any information advantage on their employers compared to other investors and investing in other vehicles such as mutual funds should be advocated."

PensionReforms notes that privatising public pensions is seen inherently as a good thing.  But seemingly, there can be too much of a good thing.  The report cites with approval others' characterisation of the potential downside of savers' making inappropriate decisions as "higher risks and grave consequences".  But isn't that the point of markets?  The more relevant public policy question here is whether governments should be directly involved in devolving responsibility for a base level of income support and, instead, exposing citizens to those risks?  Individual levels of risk should be for individuals, not governments, to decide.  (File size 152 KB; 50 pp)  352