How insurers inflate bond values to mask underperformance

A new study finds that insurance companies investing in bonds are actively changing their external sources that value their investments, but not for good reasons

By Putro Harnowo, Senior Content Manager, China Business Knowledge @ CUHK

There are many avenues for insurance companies or insurers to invest and generate profits while fulfilling their financial obligations. Among various options, bonds are particularly appealing due to their relatively stable returns and low risk. As a matter of fact, bonds play a crucial role in the overall investment strategy of insurers.

The US National Association of Insurance Commissioners (NAIC) points out in its Year-end 2023 capital markets update that the US insurance industry has a total investment of around US$2.85 trillion in corporate bonds, which accounts for 35 per cent of the industry’s total cash and assets. The European Central Bank also reported in September last year that US insurers hold nearly 40 per cent of US corporate bonds.

In many countries, insurers are highly regulated and mandated to report the value of their investment in financial statements to regulators and investors. Considering its significant portion, an estimated value or “fair value” of invested bonds would affect the insurers’ balance sheet and overall financial health. While insurers can value their own bond investments, many outsource this task to third parties to enhance transparency and comply with regulatory requirements.

However, a new study found that insurers strategically switch between external sources to seek out favourable fair value estimates of their bond investments. This practice is called fair value opinion shopping, which involves selecting sources that provide higher estimates and potentially inflate the real value of bond investments.

“There is a very common way of valuation where insurers switch the pricing sources, called third-party source switching behaviour,” says Koo Minjae, an Assistant Professor of Accounting at the Chinese University of Hong Kong (CUHK) Business School.

“There might be some opportunism going on when switching the pricing sources to get a more favourable estimate that is more consistent with what the insurers want: boosting their assets. After looking at certain instances where different sources generate conflicting estimates, we found that the opinion shopping motive dominates.”

Opinion shopping could impair financial reporting transparency and lead to mistrust among investors and policyholders. Policyholders may also misallocate and make inefficient investment decisions if the financial statements are misstated or inflated.

Opinion shopping vs. objective valuation

In the research paper titled Third-party source switches: Objective valuation or fair value opinion shopping? Professor Koo, along with Konduru Sivaramakrishnan of Rice University and Zhao Yuping of the University of Houston, analysed documents submitted to the NAIC by US insurers on their investments in bonds and other securities. The researchers obtained a sample comprised of 662,528 security-insurer-year observations from 1,852 unique insurer-years from 2014 to 2017.

The result confirmed that insurers strategically switch between third-party sources to either provide objective valuations or inflate fair value estimates by opinion shopping. While both motives exist, opinion shopping tends to be prevalent. The study highlights that opinion shopping is more widespread among financially weaker insurers with lower risk-based capital ratios. Opinion shopping is also more common among illiquid bonds or securities that are not traded frequently.

Furthermore, the researchers found that insurers engage in block-switching behaviour, where they strategically switch the price source for groups of securities rather than individual ones to make it less obvious to the regulators or the auditors. This can result in changes in fair value estimates for multiple securities within a block.

“We found evidence of opportunistic opinion shopping among insurers, even when they are regulated to disclose the source and pricing vendors at the security level.”

Professor Koo Minjae

As US insurers are regulated at the state level, with each state having its own regulations, Professor Koo and the team further interviewed several regulators. Many admitted that some insurers may opportunistically inflate their value estimates.

“What the regulators typically do is compare the same securities owned by an insurer with those of other insurers,” says Professor Koo. “If they think that the price is deviating quite a lot from other insurance companies, the regulators would advise or recommend the relevant insurers to readjust their financial statements.”

This comparison method is quite normal for valuing level-two assets that have no regular market pricing, such as bonds that are traded in not very active markets. Companies usually infer bonds at this level with other bonds that have similar interest rates or maturities.

Meanwhile, level-one assets are the easiest to value for their readily observable and transparent prices, such as listed stocks. On the other side of the spectrum, level-three assets have no observable market prices and can only be valued based on internal models or “guesstimates”.

Indicative signs of opinion shopping

In an additional analysis of 280 unique price sources, Professor Koo and her team found that the “Big Five” external price sources accounted for 63.2 per cent of the market share during the sample period, dominated by Intercontinental Exchange and followed by Bloomberg, Thomson Reuters, S&P, and Markit. These top players have more capital and the ability to better estimate the value of the securities.

Surprisingly, after re-estimating the effect of switches to Big Five and non-Big Five sources, the researchers found that switches to non-Big Five are associated with a greater deterioration in fair value estimate quality than switches to Big Five price sources. This result is consistent with opinion shopping.

“If we see some cases where the securities had been measured by Big Five but not anymore, there is a likelihood of manipulation ongoing,” says Professor Koo. “In such cases, the insurers might have switched to a vendor that can give them more favourable estimates.”

Enhancing fair value estimates

There is much literature on a fair valuation, and according to Professor Koo, the general view is that companies manipulate or engage in opportunistic inflation of the fair valuation to boost their assets. For example, opinion shopping can also be found among hedge funds, which are not mandated to disclose their security holdings, giving them incentives to inflate their value estimates.

Although insurers need to maintain their risk-based capital in a highly regulated industry, this does not apply to the less-regulated mutual fund industry.

“We found evidence of opportunistic opinion shopping among insurers, even when they are regulated to disclose the source and pricing vendors at the security level,” says Professor Koo. “Public companies are not mandated to disclose at the security level, so in that case, we are suggesting that there can be more opportunistic opinion shopping going on.”

To enhance the fair value estimates, Professor Koo recommends using multiple pricing sources for the same securities, instead of relying on only a single source. Regulators could also help by carrying out cross-validation and verification. For instance, state regulators can observe in real-time how the same security is valued by different states or different insurers in different states.

“The disclosure itself should be more transparent, in the sense that there may be something going on when insurance companies switch from a certain source to another,” she says. “Therefore, there should be more mandatory disclosure on why insurers are switching to another source.”

Lastly, more frequent external auditory reviews and regulatory reviews may be necessary. Currently, regulators only review an insurance company once every three to five years. “This frequency should be increased to verify and cross-validate the insurers’ estimates,” Professor Koo adds.