BP’s Failure to Debias (Page 4 of 6)

Apart from potential conflict of interests, we suggest that behavioral biases played a major role in analysts’ recommendations on BP’s stock. Figure 2 illustrates the time series for analysts’ recommendations between August 4, 2009 and September 17, 2010. We use a traditional23 five-point scale to code recommendations, where

  • 5 denotes buy and strong buy recommendations
  • 4 denotes add, overweight, outperform and accumulate recommendations
  • 3 denotes hold, perform, or neutral recommendations
  • 2 denotes reduce, underweight, and underperform recommendations
  • 1 denotes sell or strong sell recommendations

Figure 2. Analysts’ Recommendations: Ranks (Buy=5, Add=4, Hold=3)

Figure 2. Analysts’ Recommendations: Ranks (Buy=5, Add=4, Hold=3)

Notice from Figure 2 that there are no negative recommendations during the period. Indeed, there are no negative recommendations after April, 20 2010, the date of the Deepwater Horizon explosion. Even more interesting is the fact that the number of reports featuring ranks 3 or 4 recommendations declines after the accident. To highlight this point, we graph a second degree polynomial trendline to display the upward trending pattern in analysts’ recommendations over the period.

Effectively, Figure 2 indicates that the degree of herding in analysts’ recommendations for BP stock increased after the April 20 explosion. Kim and Zapatero (2009) propose a theory suggesting that the star-system leads analysts’ recommendations to display less herding in lower volatility stocks than in higher volatility stocks. Their theory also leads to the prediction that an increase in the volatility of a particular stock will cause an increase in herding for that stock. In this regard, Fodor and Stowe (2010) report that option market implied volatility (IV) for BP stock indeed increased after April 28, from below 0.38 to an average level of 0.65.24

Notably, analysts did reduce their target prices following the explosion of Deepwater Horizon. See Figure 3, which shows a positive trend until April 20, and then a decline, albeit with a lag. After June 1 2010, almost all reports feature target prices in the range £4 to £6.25

Figure 3. Analysts’ Target Prices in Sterling Pounds (£)

Figure 3. Analysts’ Target Prices in Sterling Pounds (£)

Figure 4. Market Prices versus Target Prices in Sterling Pounds (£)

Figure 4. Market Prices versus Target Prices in Sterling Pounds

Figure 4 displays time series for both target prices and market prices. We add two trendlines (polynomial, second order) to highlight the trends. Notice that the two trendlines feature similar patterns. However, the trendline of target prices (solid line) is always above the trendline for market prices (dashed line). Interestingly, the gap between the two seems to widen over time, perhaps because of the greater uncertainty and consequent difficulty in forecasting after the accident.

In any event, after April 20, the combination of declining target prices in conjunction with more favorable recommendations suggests that analysts viewed the decline in BP’s market price in response to the explosion of Deepwater Horizon to have been an overreaction. In this regard, consider Figure 5, which displays the percentage premium calculated as the difference between the target price and the current market price over the current market price and multiplied by 100. This premium is often used by analysts to calculate the future percentage upside or downside potential of the stock. Brokers usually have premium cut-offs which they use in issuing their final recommendations.

Figure 5. Target Price Premium over Current Market Price (in %)

Figure 5. Target Price Premium over Current Market Price

Notice that the trend for premium is upward sloping, similar to the one for recommendations. We note that after April 20 no premium is negative, and that premia seem to be more dispersed, reflecting the greater uncertainty analysts faced after the accident.

To gain more insight into analysts’ thinking, both before and after the explosion of Deepwater Horizon, consider some content from a sample of reports. On April 20, 2010 security analyst Dougie Youngson at Arbuthnot Banking Group Plc, London, initiated the coverage of BP with a buy recommendation. In justifying his recommendation, he stated that BP’s strategy and cost reductions were yielding the expected results. In respect to competitors, he claimed, “Safe, reliable and compliant operations remain the first priority… With its new strategy, board and aggressive cost cutting programme, we feel BP is much better-positioned relative to most of its peers in 2010 … When comparing BP’s progress in strategy development and cost cutting with Shell, we believe it is much more advanced in both areas. Consequently, in our view BP has a significant competitive advantage over its adversary” (Youngson, 2010, p. 1).

Notably, Youngson’s report identifies key issues: safety, cost cutting, and profitability. However, in our view his analysis failed to assess these correctly. For example, he appears to have downplayed if not ignored the accidents that continued to occur in Alaska in 2008 and 2009, and the fact that in March 2010, OSHA proposed US$3 million more in penalties after finding 62 violations at BP’s Ohio refinery. In addition, Youngson’s comparison of BP to Shell is striking, given the difference in choice of well design, as discussed in section 3.2.

Youngson’s report on BP is fairly typical of analysts’ assessments. We examined 33 reports that were issued after April 20, 10 of which by UBS. In the report of April 28, UBS’s analysts clearly underestimated the effects of the accident, stating “[…] we think these costs are more likely to be in the hundreds of millions rather than billions and hence, ultimately, unlikely to be material to the long-term investment case.” They were recommending purchasing BP stock, setting a target price of £7.25. Between April 20 and May 10, BP lost 17 percent of its value, corresponding to US$32bn of market capitalization. In UBS analysts’ view the decline corresponded to “a substantially exaggerated reaction although less so in the context of weak markets.”

On May, 25, UBS analysts eventually reduced the target price to £6.30, and to £5.80 on June, 1, but always maintained the buy recommendation. Interestingly, in the latter report, they state, “Our forecasts, which look at recurring net income and exclude specials (we assume the costs of the spill are ’special’) […]”. On June 7, the analysts calculate in a very detailed way the potential costs of the spill. However, they keep the buy recommendation, and the previous target price. In the reports of June 11 and 17 they stated that the market reaction was mainly driven by political factors. Interestingly, they maintained the buy recommendation and did not change their target price, both of which remained more or less stable through the UBS report of September 9, 2010.

By and large, analysts’ reports prior to the explosion of Deepwater Horizon emphasized costs and risks in the Gulf of Mexico which were associated with weather and price swings, rather than oil spills and related operational accidents. These reports strongly emphasize potential performance related to cost cutting. In 2009, BP exceeded analysts’ expectations thanks to its aggressive cost cutting policy. For example, in the report by Unicredit dated December, 17 2009, analysts claim that BP has a good operational momentum because of its “first-mover advantage in cost cutting.”

Interestingly, reports that provide a comparison with competitors often (though not always) feature a recommendation of hold. For example, the report issued by Collins Stewart on February, 2 2010 states, “BP’s shares outperformed their major peers substantially in the past year, driven by sharply improving performance, cost reduction and volume growth. We think the shares will struggle to outperform from here, given the outlook for slightly lower volumes in 2010, and a slower pace of cost reduction. We continue to recommend a switch into Royal Dutch Shell (Buy, TP 2150p/sh) where we see significant cost reduction potential, a major turnaround in free cash flow on a 2-year view and much better valuation upside (35% upside to SoP vs 16% for BP).”

Another example is a Morgan Stanley Europe research report on BP, dated March 11, 2010, and the last before April 20, assigns an Overweight/Buy recommendation to BP stock. In justifying the recommendation, the report states, “management focus on costs and execution over the last 18 months is undiminished.” A notable exception is Datamonitor, whose report dated April 7, 2010 includes a SWOT analysis in which one of the weaknesses described pertains to the Texas City accident and associated OSHA violations. In concluding the discussion on this point, the Datamonitor report states, “Such events causing environmental damage could result in heavy financial penalties for the company, eroding its profits. In addition, such law suits could also tarnish its brand image.”

What, if any, psychological pitfalls were at work in explaining why most analysts missed the signals? In our view, availability bias and confirmation bias loom large. Analysts focus heavily on earnings trajectories and company narratives, as these are readily available and salient. It is well known in the behavioral finance literature that security analysts tend to rely on management’s stories. See Montier (2005).

An illustrative example can be found in a report by Raymond James, issued August 27, 2009 by analyst Pavel Molchanov. He discusses the incidents at both Texas City and Prudhoe Bay, beginning with Texas City, stating, “Still, early signs are encouraging. BP’s companywide ‘recordable injury frequency’ in 2008 was below the level of 2005 and less than half the level of 2000 although it must also be pointed out that the Texas City refinery itself had another fatal accident in 2008. Perhaps most importantly, senior management’s increased emphasis on developing a safety culture appears clear. In its annual strategy presentation in March 2009, BP stated, ’Safe, compliant and reliable operations: our No. 1 priority.’ As with refinery safety, it is difficult to definitively gauge the level of BP’s progress in this regard. What is visible is that the company’s number of oil spills (above one Bbl) has continually declined over the past decade, and in 2008 it was more than 10% lower than in 2006.”

In conference calls with analysts, BP repeatedly stressed its focus on safety. In its 2009 Annual Report and Accounts, the following passage appears, “Competition puts pressure on product prices, affects oil products marketing and requires continuous management focus on reducing unit costs and improving efficiency.” (BP, 2009, p. 18) Following the release of the Baker report in January 2007, analysts accepted at face value the newly appointed CEO Tony Hayward’s assertions about accepting and implementing all the report’s recommendations. In line with confirmation bias, they appear to have underweighted subsequent information about OSHA violations.

1 | 2 | 3 | 4 | 5 | 6 | Next Page

Hersh Shefrin and Enrico Maria Cervellati

23. Commercial databases commonly use a reverse scale where 1 denotes a strong buy whereas 5 a strong sell recommendation. However, we find it more intuitive to use a 5 for strong buy, so that an upgrade, for example, is represented by an increase in rating score.

24. See figure 4 in Fodor and Stowe (2010).

25. Two notable exceptions are represented by the reports by ING dated June, 22 and August, 23, with target prices maintained at £7.12. However, as the analyst (Jason Kenney) highlights, these target prices should be considered on a 1-3 years horizon. Therefore, we do not include them in our analysis, since they cannot be compared with other target prices with a 12-months horizon.


NYSSA Job Center Search Results

To sign up for the jobs feed, click here.


NYSSA Market Forecast™: Investing In Turbulent Times
January 7, 2016

Join NYSSA to enjoy free member events and other benefits. You don't need to be a CFA charterholder to join!


CFA® Level I 4-Day Boot Camp

Thursday November 12, 2015
Instructor: O. Nathan Ronen, CFA

CFA® Level II Weekly Review - Session A Monday

Monday January 11, 2016
Instructor: O. Nathan Ronen, CFA

CFA® Level III Weekly Review - Session A Wednesday

Wednesday January 13, 2016
Instructor: O. Nathan Ronen, CFA

CFA® Level III Weekly Review - Session B Thursday
Thursday January 21, 2016
Instructor: O. Nathan Ronen, CFA

CFA® Level II Weekly Review - Session B Tuesday
Thursday January 26, 2016
Instructor: O. Nathan Ronen, CFA