Initial Public Offerings: Status, Flaws and Dysfunctions

2. Costs of Stock Offerings

2.1 Overview and literature review

An effective issuing system requires access by companies at reasonable cost that does not unduly inflate the expense of this financing approach. Primary issues have two main types of costs, tangible and intangible costs on the one hand and initial underpricing on the other. It is also important to have a viable primary issues market: investors in primary issues have to be able to fetch an acceptable average yield, given the risk involved.

Robinson (1997) stresses that it usually costs a company $65,000 to $150,000 in regulatory costs for a primary issue in Alberta. The underwriter's commission represents an additional amount that varies between 7.5% and 10% of the capital raised. In Ontario, minimum regulatory costs for a primary issue range from $100,000 to $150,000. Jog (1997) estimates the total average cost of an issue on the Toronto Stock Exchange at $300,000 to $400,000, to which must be added a commission varying from 6% to 8% of the proceeds of the issue. Higgins (1994) estimates the direct costs of 68 primary issues by companies listed on the Toronto Stock Exchange in the period 1992-1993 and notes that the average direct cost weighted by issue size is 6.74%, which suggests that direct costs have gradually declined in Canada. These studies used relatively old data and samples were often limited to issues followed by TSE listings. An analysis based on more recent data was therefore needed.

2.2 Direct and tangible costs

2.2.1 Components

The direct costs of primary issues may be broken down into two main components: the costs incurred to comply with regulations and the commission paid to the underwriter.

The costs of compliance with listing requirements relate to preparing the prospectus, hiring various professionals and paying fees. These are partly fixed costs and, for given regulations, are proportionately higher for small issues.

The underwriter's commission is a percentage of the gross proceeds. It may come with various benefits such as stock options, the cost of which cannot be estimated here.

2.2.2 Measurement problems

Average costs are less meaningful, however, as the amounts disbursed in relation to issues are not proportionate to their size: the smallest issues are proportionately much more expensive than the larger ones. This is the conclusion of Shutt and Willams (2000), who estimate the direct costs of 49 primary issues of companies listed on the Toronto Stock Exchange in the period 1998-1999. The average total cost of primary issues varying from $1 to $10 million was 11.7%, whereas it was 9.3% for issues of $10 to $50 million, 7.7% for issues of $50 to $100 million, and 5.9% for issues worth between $100 and $200 million. Comparing direct costs in Canada and the US, Shutt and Williams (2000) show that the average direct cost weighted by issue size is 5.5% in Canada, compared with 8.7% for NASDAQ listings and 6.6% for NYSE listings. They conclude that the costs associated with primary issues are lower in Canada than in the US. However, the limited size of the sample and the fact that the analysis was limited to a single stock exchange seriously limits the conclusions that can be drawn from this analysis.

2.2.3 Statistical analysis of direct costs

An analysis was performed of 513 Canadian primary issues, including 295 CPC issues,^{Footnote 24} and 1,181 US primary issues conducted in 1997-1999. To this end, the various components of the direct costs of issues of comparable size in Canada and the US were calculated based on the issue prospectuses.^{Footnote 25} The findings are summarized in Table 8 for conventional issues and Table 9 for CPC issues.

Table 8: IPO costs by issue size, excluding issues under CPCs, 1997-1999

Size of Issue (US$ millions)	Number of IPOs	Broker Compensation (%)	Other Expenses (%)	Total Direct Costs (%)	Under-pricing (%)
* Significantly different from 0 at 1% significance level.
Canada
1.0 – 9.9	53	8.12%*	7.86%*	15.98%*	30.61%*
10.0 – 49.9	49	6.14%*	3.31%*	9.45%*	11.30%*
50.0 – 99.9	10	6%*	2%*	8%*	10.76%*
100 and over	16	5.53%*	1.75%*	7.28%*	8.88%*
Average		6.88%*	4.9%*	11.78%*	18.95%*
Weighted average (by size)		5.35%*	1.84%*	7.19%*	5.11%*
United States
1.0 – 9.9	119	9.29%*	8.7%*	17.99%*	9.05%*
10.0 – 49.9	532	6.93%*	3.70%*	10.63%*	26.15%*
50.0 – 99.9	300	6.88%*	2.12%*	9%*	55.57%*
100 and over	237	6.09%*	1.2%*	7.29%*	67.19%*
Average		7%*	3.3%*	10.30%*	37.5%*
Weighted average (by size)		5.79%*	1.43%*	7.22%*	38.38%*

Table 9: Direct and indirect issuing costs for CPC IPOs, 1997-1999

Number of IPOs	295
Total gross proceeds ($ millions)	$53.47
Average gross proceeds ($ millions)	$180
Broker compensation (%)	10%
Other costs (%)	12.95%
Total direct cost (%)	22.95%
Initial underpricing (%)	161.52%
Total issue cost (%)	184.47%

Size being equal, the direct costs of a Canadian issue are lower than the US average. This finding is consistent with earlier studies of much more limited samples and invalidates the argument that associates the decentralized structure of securities regulation in Canada with higher issuing costs and obstacles to public financing of businesses.^{Footnote 26}

We also note that, on average, the underwriter's commission in Canada is slightly lower than the US standard of 7%. Initial underpricing seems to be higher in the US, given the very high initial return on a number of US issues by Web businesses, although this trend may change and so there is a need to longer-term analysis. While lower than the US average, Canadian issuing costs remain very high, especially for small issues.

2.2.4 Effects of size, sector and broker

Table 9 summarizes CPC issuing costs. Paradoxically, these issues, which go through a simplified review process and are designed to alleviate financing costs for small businesses, involve higher costs than small conventional issues. These costs average 22.95%, compared with 15.98% for conventional issues worth $1 to $10 million. This may well be accounted for by the very small size of many CPC issues, as the average gross proceeds for CPC issues are only $180,000.

There are economies of scale in the direct costs of primary issues. These involve direct outlays on top of the intangible costs to the company, i.e. the opportunity costs to the senior officers who work on planning the issue. These direct costs, however, depend not only on the province where the listing takes place, but also on the method of subscription used. Moreover, with small issues, these costs make up a relatively minor part of total costs, which largely relate to initial underpricing.

2.3   Initial underpricing

2.3.1 Overall measurement

Underpricing, which is seen in many countries, represents an additional cost for issuing companies and a transfer of wealth to investors who are able to buy the securities at the issue price. This anomaly associated with primary issues has never been satisfactorily explained, although many theories have been advanced. As Canadian studies deal only with large issues listed on the Toronto Stock Exchange, an update based on all issues was needed.

To quantify initial underpricing, researchers generally use the initial return to the new shareholders expressed as:

where: P_m is the market's closing price on the first day of trading in the primary issue securities, and P_e is the issue price. This measures the return to an investor who was able to buy the shares from the underwriter and sell them by the close of the first day of trading.

The use of stock exchange data limits the number of observations available for empirical study of this phenomenon. The final sample includes 971 Canadian primary issues for the period January 1991 to December 1998, or 878 initial public offerings of common shares (including 433 CPC issues) and 93 initial unit issues. The latter have been included because very little research has been done on them (How and Howe, 2001), yet they are a major source of financing for a number of Canadian small businesses.

The issues are organized in three subsets: traditional common share issues, CPC issues and unit issues. Table 10 provides a breakdown of initial underpricing distribution for the 971 primary issues studied and shows them as generally underpriced. The equal-weighted mean return on the first trading day was 73.65% for all issues, with median underpricing of 30%, though the level of underpricing depends on issue type. The initial return was 20.57% for common shares excluding CPC shares, but rose to 135% for issues made through that system. The results thus confirm the findings of numerous researchers in various countries,^{Footnote 27} but differ from those of Jog (1997), who put initial underpricing on Toronto Stock Exchange issues at 7.89% for the period 1984-1992. This discrepancy in findings is probably connected with our inclusion in this study of smaller issues on other Canadian stock exchanges.

Table 10: Initial underpricing of Canadian initial public offerings, 1991-1998

	Common Shares (Non-CPC)	CPC Common Shares	Units	Total
Sample includes 971 IPOs by companies subsequently listed on the Toronto, Montreal, Vancouver or Alberta stock exchanges.
Number	445	433	93	971
Mean	20.57%	135.41%	40.06%	73.65%
Standard deviation	0.56	1.73	1.46	1.41
T-statistic	7.65	16.21	2.63	16.2
Median	5%	100%	1%	30%

The underpricing of CPC issues already highlighted by Robinson (1997) is likely related to their very small size and low share value. The major underpricing of the shares of companies with cash and a plan as their only assets can be explained only by the gradual disclosure of information often known by the proponents from the outset. This phenomenon calls for in-depth analysis that is not our purpose here. Unit issues are also highly underpriced compared with traditional common share issues, at 40.06%. These findings are similar to those of Schultz (1993) and How and Howe (2001). Schultz (1993) points out that companies electing to issue units are relatively younger and less stable than companies issuing common shares.

2.3.2 Effect of size and other characteristics

Table 11 shows that initial underpricing, like the issuing operation itself, is not static. In 1993-1994, average initial underpricing and the number of primary issues increased over earlier years. In 1996, the number of primary issues rose by 107.31% over 1995 and initial underpricing increased as well. In 1997, average initial underpricing of common shares reached a peak of 39.30%. That year also saw a peak in activity, with companies raising $3328.5 million.^{Footnote 28} In 1998, primary issue and initial underpricing figures declined.^{Footnote 29}

Table 11: Annual distribution of initial underpricing of Canadian IPOs of common shares, 1991-1998

Year	Common Shares (Non-CPC)		Common CPC Shares		Total
	Number	Initial Underpricing	Number	Initial Underpricing	Number	Initial Underpricing
Sample includes 971 IPOs by companies subsequently listed on the Toronto, Montreal, Vancouver or Alberta stock exchanges.
1991	11	-4.85%	2	0.55%	19	8.07%
1992	25	31.72%***	8	85.42%**	38	34.92%**
1993	78	15.87%**	55	159.54%*	147	79.40%*
1994	70	17.85%*	55	127.79%*	141	64.88%*
1995	41	16.87%**	52	53.73%*	101	35.34%*
1996	85	13.25%*	68	68.41%*	162	36.24%*
1997	88	39.30%*	96	194.23%*	209	107.93%*
1998	47	13.79%*	97	165.49%*	154	111.78%*
1991-1998	445	20.57%*	433	135.41%*	971	73.64%*

The years 1993, 1994, 1996 and 1997 were periods of intense activity, while 1991, 1992, 1995 and 1998 were slow. Ibbotson and Jaffe (1975) found that periods of intense activity were typified by high issue volumes and strong initial underpricing followed by declining initial underpricing.

Table 12: Distribution by size of gross proceeds of common share IPOs, 1991-1998

Size ($ millions)	Number	Initial Underpricing 1	Initial Underpricing 2
Sample includes 878 IPOs of companies subsequently listed on the Toronto, Montreal, Vancouver or Alberta stock exchanges. Initial underpricing is calculated in two ways: 1 – (average closing price for first five days of trading minus issue price) / issue price; 2 – [(average closing price for first five days of trading minus issue price) / issue price] – [average TSE 300 index for first five days of trading minus TSE 300 index at time issue price was set / TSE 300 index at time issue price was set]. * Significant at 1% level; ** significant at 5% level; and *** significant at 10% level. P = t-test for the difference in means of the two groups (Size < $1M and Size ≥ $20M), significant at 1%. (N) = Kruskal-Wallis and Mann-Whitney tests for differences in means of the two groups (Size < $1M and Size ≥ $20M), significant at 1%.
Size < $1M	540	119.07%*P(N)	124.11%*P(N)
$1M ≤ Size < $5M	101	31.03%*	32.94%*
$5M ≤ Size < $10M	27	11.54%***	11.22%***
$10M ≤ Size < $20M	65	3.6%	4.28%
Size ≥ $20M	145	-1.36%	-0.78%
Total	878	20.57%*	22.11%*

2.4   Conclusion

Size being equal, direct issue costs in Canada are lower than those in the US. The cost of primary issues, however, seems high for small businesses. This cost may have a significant impact on their competitiveness. In the US, collusion among brokers has been suggested as the reason for the high direct costs (remaining at 7%). The situation may be the same in Canada, where there is also a concentration of issue marketing activities.

In 1987, the federal and provincial governments amended the legislation so that banks, trust companies and foreign brokerage houses could own stock brokerage firms. This led to a wholesale restructuring of ownership in the brokerage industry. The big Canadian banks bought major brokerage houses or created their own. According to Strategis,^{Footnote 30} the Canadian stock brokerage industry came out of 2001 with 198 houses, compared with fewer than 120 in early 1990. Three major types of brokers coexist: integrated brokerage firms (71% of sales in the sector), institutional brokers (9%) and retail brokers (20%).^{Footnote 31} In 2001, the seven main integrated firms, including brokerage subsidiaries of the six big Canadian banks and a major US brokerage house, generated more than 70% of industry sales. These were:

• BMO Nesbitt Burns, owned by the Bank of Montreal,
• CIBC World Markets, owned by CIBC,
• National Bank Financial, owned by the National Bank of Canada,
• RBC Dominion, owned by the Royal Bank of Canada,
• Scotia Capital, owned by the Bank of Nova Scotia,
• TD Securities, owned by the TD Bank Financial Group,
• Merrill Lynch, which had taken over a major portion of the Canadian retail brokerage market, withdrew from the retail trade, keeping only its institutional clients. The CIBC acquired Merrill Lynch's Canadian retail brokerage operations in late 2001 and its stock and mutual fund services in 2002.^{Footnote 32}

The biggest banks have made several attempts to consolidate. In 1998,^{Footnote 33} these attempts resulted in new legislation on mergers of financial institutions. By late 2001, all of the big integrated brokerage firms in Canada, generating about 70% of industry revenue, belonged to banks, six of which accounted for over 90% of total bank assets.

Underpricing may reflect a lack of competition among brokers, but other factors are probably involved, such as companies' lack of experience and negotiating power, valuation and book building approaches and the role of institutional investors.

^{Footnote 24} For the details on this study and its findings, see Kooli and Suret (2003a).

^{Footnote 25} The direct costs of Canadian primary issues are from final prospectuses that were not available on the SEDAR until 1997. The direct costs of US primary issues are from final prospectuses available on the SEC site.

^{Footnote 26} These arguments are defended in various reports including Sawiak et al. (1996) and the letter of November 15, 2002 from H. MacKay to the Deputy Prime Minister and Finance Minister of Canada, available at http://www.collectionscanada.gc.ca/webarchives/20071122012618/http://www.fin.gc.ca/news02/02-094e.html.

^{Footnote 27} Initial underpricing for primary issues is 23.87% in Italy according to Guidici and Paleari (1999), 13.23% in France according to Derrien and Womack (2000), 61.8% in Malaysia according to Paudyal et al. (1998), 14.3% in the UK according to Lévis (1993), and 15.8% in the US according to Ritter (1998).

^{Footnote 28} Lowry and Schwert (2000) find it puzzling that companies choose to go public when primary share issues are sharply underpriced.

^{Footnote 29} In 1998, the gross proceeds of primary issues also fell, compared with 1997 (a hot period in the issues market). This may have been due to the fact that, as more and more companies go public (as in 1997), doubts about their value fade. Accordingly, their subsequent initial returns go down (1998). sic – TR.]

^{Footnote 30} Canada's Securities Industry

^{Footnote 31} The integrated brokerage houses serve institutional and retail markets. Institutional brokerage houses work with pension funds, insurance companies, mutual funds, banks and trust companies. Retail brokerage firms offer products and services to retail investors.

^{Footnote 32} http://www.cbc.ca/news/business/story/2001/11/22/merrill_011122.html

^{Footnote 33} Cookey (2001), p. 1.