The Corporate Design of Investments in Startups: A European Experience

Abstract

This is the first European study to conduct an extensive empirical research of startup charters. Our aim is to test whether the significant reforms of the law on the Italian società a responsabilità limitata (the GmbH-type limited liability company) were successful in making Italian corporate law more amicable towards startups and venture capital contracting techniques. We explain why, in the Italian context, charters provide significant information on financing deals, and we analyse more than 5000 charters of Italian startups. We find almost 200 charters that reflect the features predicted by the financial contracting theory, albeit with some significant variations in comparison to the US experience. The main one is that convertible preferred shares are not used. We report the large use of (non-convertible) participating preferred shares but also the increasing adoption of preferred shares that are functionally equivalent to US convertible non-participating preferred shares. The absence of convertibility mechanisms also explains the different structure of antidilution clauses in the Italian market. Hybrids are used to provide SAFE- and KISS-like contractual solutions. Co-sale clauses (tag-along and drag-along) are widespread and also highly standardized. US-like vesting schemes are equally observed. Some of the peculiarities we report depend on Italian law idiosyncrasies that are mainly the product of doctrinal constructions. However, corporate practice is pushing the envelope in its efforts to adapt Italian charters to startuppers’ and investors’ needs. From this standpoint, the Italian reforms look, though not completely, successful. Startup law appears to be transforming the European corporate law tradition.

1 Introduction

In this article we conduct an empirical investigation of the charters of 5095 Italian companies that are registered in the Italian register of innovative startups and so enjoy the related special benefits accorded by Italian law. Amongst these companies we identify a restricted number whose charters exhibit some of the features of outside financed startups that one would expect to find in accordance with the financial contracting literature. We do not distinguish between types of outside investors, which can be business angels, venture capital funds or industrial firms. Accordingly, we will refer to those companies as ‘startups with outside investors’.

We wanted to test the hypothesis that Italian corporate law is not fully amicable towards venture capital contracting techniques identified by the financial contracting literature, which largely refers to the US experience. This hypothesis has been advanced by the first Italian scholars that analyzed in depth this specific and increasingly important field of law.¹ To this end, we wanted to understand if and how outside investors are transplanting US techniques in the Italian market or are adopting different but functionally equivalent instruments, and to what extent the differences with the US world, if any, are to be ascribed to corporate law restraints. Thus, our article is part of the small empirical literature that analyzes whether investments in startups outside the US have features that reflect US practice, as well as identifies and seeks to explain divergences. However, our research is somehow different in design, being much more focused on corporate law than many other papers of the same genre. As far as we know, it is the first research in Europe that analyzes such a large number of charters and goes into such in-depth, granular review of the charters’ clauses of startups.

The restricted group of companies that exhibit the features of startups with outside investors shows significant familiarity with US contractual techniques, but also some important differences. The most striking is that convertible preferred shares are virtually absent, as seems to be true also for the Chinese market.² Participating preferred shares are widely used, while (non-convertible) non-participating shares are rare; a specific form of preferred shares is used as a functional equivalent of US (convertible) non-participating stock. Also, co-sale clauses are becoming widely popular in startups with outside investors. The same is true with regard to SAFEs, KISSes and work-for-equity, even though charters are insufficient to ascertain the extent US-like practices are adopted across Italy.

Antidilution clauses are starting to be implemented and represent something truly new in Italian charters; accordingly, our research documents the emergence of an instrument that was almost never used in Italian corporate law before the emergence of venture capital financing. However, since Italian charters do not use convertible preferred shares, antidilution clauses are not focused on the conversion ratio, but on the attribution of adjunctive shares to the investors protected by them.

In our paper we have not devoted any attention to outside investors’ directors or veto rights, because they were already well known in Italian corporate practice and, thus, were not an effective instrument to identify startups with outside investors.

Our analysis confirms that the reform of Italian corporate law concerning the società a responsabilità limitata (the GmbH-type limited liability company) has made this form the preferred one in the market.³ From a wider, European perspective, this also confirms that Italy’s choice to reframe the LLC was correct and that the European heavily regulated joint-stock company form might be too burdensome for promoting the creation of startups.⁴

We report how corporate practice is pushing the envelope in order to satisfy market needs, assuming a liberal interpretation of the relevant Italian rules that in our view courts should adopt when litigation comes to the new corporate world of startups with outside investors and, more specifically, VC-backed startups. This courageous position shows the mechanism through which Italian law is subject to the influence of Delaware law, which in a previous paper we addressed as a key driver in the transformation of Italian corporate law during the last decade.⁵ We think that the Italian data and experience are important for any jurisdiction in Continental Europe.

Our paper proceeds as follows. In Sect. 2 we review the economic as well as the law & economics literature on financial contracting theory, and how VC financing reflects that theory in the US and outside the US. In Sect. 3 we review the law literature that, outside the US, analyzes whether and how the economic issues raised by VC financing are dealt with in different legal environments. In Sect. 4 we present our research methodology. Sect. 5 presents the data with the characterizing clauses. Section 6 analyses and discusses the different clauses and the main differences with US practice. Section 7 offers a summary, and Sect. 8 concludes.

2 Economic Literature Review

A well-known characteristic of VC financing is its articulated contractual framework. It is richly studied both by the economic⁶ and law literatures.⁷ When contracts are state contingent and incomplete, meaning that the parties cannot negotiate ex ante over all the states of the world that will be affected by their contractual relationship, contractual incompleteness can prevent a party from getting the ex post return needed to compensate her ex ante investment,⁸ especially if the parties have different objectives and wealth constraints.⁹ This situation is typical when a penniless entrepreneur seeks funding from an outside investor (usually a venture capital fund). In this situation, the entrepreneur has interest in handing over some control, leaving some residual rights to the investor that would be otherwise exposed to its opportunistic behavior.¹⁰ The outside investor, instead, can keep the entrepreneur on track through incentives in the form of equity compensation and share vesting.¹¹

This theoretical framework fits well with what happens in the venture capital world, at least in the US environment. Indeed, venture capital funds and entrepreneurs always agree to separately allocate cash-flow rights, board rights, voting rights, liquidation rights, and many other control rights.¹² The contractual instruments that are used to achieve the separate allocation of rights that characterizes VC financing can be easily found in the model legal documents of the US National Venture Capital Association (NVCA).¹³ First of all, the model certificate of incorporation contains the provisions concerning the convertible preferred shares, whose adoption is probably the most renowned contractual feature of VC financing. Those convertible preferred shares can be paid with cash or result from the conversions of previously paid SAFEs or bridge notes. They can have a right to dividends on an as-converted basis with common shares, or a preference on non-cumulative dividends, or carry an annual cumulative dividend and participate in other dividends on an as-converted basis. As to liquidation preferences, they can be non-participating preferred shares or participating preferred shares (full or with a cap on participation). With regard to voting rights, they vote together with common shares but are entitled to elect a separate class of board directors (preferred directors), and usually enjoy veto powers in the form of protective provisions that require the holders’ written consent in a series of events such as liquidation, winding-up, amendment or alteration of provisions of the charter or bylaws, issuance of any other security, etc.

The convertible preferred shares are protected, in the event the company issues additional securities at a price below the preferred conversion price, by antidilution provisions that work at the expense of the common shares.¹⁴ They are subject to mandatory conversion in the event of a public offering and their holders can be subject to pay-to-play provisions, but they can also be redeemable for a certain period of time, granting a way out to the outside investor. Other clauses are contained in the investors’ rights agreement, amongst which lock-up obligations in the event of an IPO, management and information rights, rights to participate pro rata in future financing rounds, matters requiring preferred director approval, and employee stock option vesting schedules. Other clauses are contained in the right of first refusal/co-sale agreement, and others in the voting agreement, which also contains the drag-along provision.

This rich palette of contractual provisions probably makes VC financing the most complex form of alternative investment. For sure, it requires a very malleable contract and corporate law, able to accommodate freedom of contract and private ordering in ways that are not common in more traditional forms of investment.

It comes to no surprise, therefore, that the economic literature, after having investigated how financial contracting theory fits US practices of venture capital investments, moved on to analyze whether this contractual framework can be or has been actually exported outside the US, how and with what differences, and finally what institutional factors might explain the discrepancies from the US settings. This literature is linked to that exploring the relationship between the quality of law, law enforcement and the financing of firms, postulating that law matters,¹⁵ and differentiating law and ensuing economic outcomes on the grounds of legal origins.¹⁶ Many papers have investigated the capacity of different jurisdictions to sustain complex, state-dependent contracts like those involved in US venture capital transactions. The assumption is that some legal systems might not adequately enforce certain types of contracts or clauses and, therefore, those constraints can be reflected in the deal structure. Kaplan and others found that European VC-financing contracts are less likely to use contingencies and more likely to use common shares instead of convertibles, but they also found that more experienced and successful VCs adopted US-style contracts in the various jurisdictions where they operated, and that all the funds in the sample that used both non-US and US-style contracts switched from the former to the latter during the sample period—a sign that institutional impediments might not be sufficient to hinder the implementation of US-style contracts in different legal regimes.¹⁷ This conclusion does not fully coincide with that of an influential paper by Lerner and Schoar that investigated developing country private equity investments, finding that differences in legal regimes bring different deal structures, with investments in low enforcement countries and in civil law jurisdictions tending to rely more on common shares and debt than on convertible preferred shares with covenants.¹⁸ Cumming and others find that better laws facilitate investor board representation and favor negotiations and deal efficiency.¹⁹ Bonini and Alkan investigated the political and legal determinants of cross-country differences in VC investments, finding that legal system rigidity of each country plays an important role in explaining cross-sectional variance.²⁰ Similarly, many other articles seem to support the conclusion that the quality of law and law enforcement have a material impact on shaping financial decisions and governance structures and thus on the efficiency of VC deals.²¹

However, none of these economic articles contain a granular country-by-country analysis of the legal constraints that prevent venture capital financing from flourishing.

3 Law Literature Review

The analysis of the legal constraints that in each jurisdiction might prevent the adoption of US-style contracts can ultimately be found in the national law literature, though usually this literature is not connected to the international economic literature briefly summarized in the previous paragraph. Needless to say, language barriers and text accessibility prevent international researchers from getting a more precise view of the legislative obstacles to US-style contracts and the local, different routes that drafters may take to accommodate the needs of the outside investors and the entrepreneur. Nevertheless, there are a few exceptions. For instance, Lin describes the use of a contractual mechanism that is common in the Chinese venture capital sector—the valuation adjustment mechanism (VAM), which entitles investors, typically venture capital funds, to adjust the portfolio company’s original valuation or to get compensation by cash or equity from the company or its shareholders upon the occurrence of certain future events. The author attributes this peculiarity to Chinese law not allowing limited liability companies to issue convertible preferred shares, as well as to weak protection of minority investors and other factors.²² Giudici and Agstner describe the transformation of the Italian law of the limited liability company, addressed at accommodating the needs of venture capital financing, and analyze the shortcomings of the Italian law reforms and the cultural limits that hinder their transformative effects.²³

In the national literatures, Agstner and others²⁴ as well as Nigro and Enriques²⁵ analyze in detail the constraints of Italian law.²⁶ Both articles point out that many of these constraints are self-inflicted, since scholars and courts tend to infer by analogy or through other interpretative techniques mandatory rules from other provisions of corporate law, thereby limiting the attempts to render corporate law more enabling.

In the German literature, especially Kuntz evaluated in depth the possible implementation of US venture capital contracts in accordance with the law on joint-stock companies (Aktiengesellschaft or AG), characterized by the principle of Satzungsstrenge [§ 23(5) AktG], and the law on limited liability companies (Gesellschaft mit beschränkter Haftung or GmbH).²⁷ By examining every contractual provision typically regulating the VC-backed startup firm, Kuntz comes to the conclusion that German corporate law allows the adoption of most of the US-like arrangements,²⁸ though in the AG, due to the mentioned Satzungsstrenge, it is sometimes necessary to include the relevant provisions in a shareholder agreement.²⁹ This suggests that corporate law does not explain the less developed VC market in Germany.³⁰ Furthermore, the implementation of US-like founder and employee vesting schemes in German VC-backed startups is broadly analyzed by Ockert in her doctoral thesis.³¹

The French literature devotes almost no attention to such specific legal transplant issues. Legal scholars focus mainly on the plain description of the key terms of the agreements that are entered into between VC funds and their investors,³² as well as, especially in the standard textbook literature, on the analysis of the typical contractual design of VC operations (capital à risque).³³

4 Research Methodology

In US VC-financing models, the contractual structure of transactions is articulated in a wide range of interrelated contracts (Voting Agreement, Term Sheet, Stock Purchase Agreement, Right of First Refusal and Co-Sale Agreement, Model Legal Opinion, Management Rights Letter, Investors’ Rights Agreement, Indemnification Agreement, Certificate of Incorporation).³⁴ A comprehensive analysis aimed at investigating if and how international VC-financing techniques have been implemented by Italian startups would ideally require the examination of all the documents of the financing transaction.³⁵ This would imply conducting the research through a top-down approach, directly acquiring from VC funds and business angel networks the documentation of each transaction. This approach would raise significant problems. First, it would require the identification of the business angel networks and VC funds that have financed Italian firms. This is probably not an unsurmountable task, but researchers in Italy cannot rely on databases equivalent to those used by US researchers. For instance, we have used the Crunchbase.com database, but it appeared that it does not offer a comprehensive panorama of the Italian market. Moreover, we would not have been able to trace transactions where an industrial partner acted as an outside investor. Second, and most important, it is doubtful that VCs active in Italy would be willing to share confidential documents. Indeed, in Italy there is nothing similar to the NVCA model legal documents, and from our previous researches we learned that there is only a low level of standardization in startups’ constitutional documents.³⁶ We therefore decided not to follow this route and, conversely, to adopt a bottom-up approach.

We collected, for a group of selected provinces, all the charters of all the Italian companies registered in the dedicated section of the company register in a given interval of time, in order to verify if and how the relevant clauses are concretely transplanted into the Italian legal environment or adapted to achieve similar economic results.³⁷ Thus, the examined dataset is free from any self-reporting bias. We think that the charters offer sufficient information about our research topic, for at least four reasons.

First, we are mainly interested in corporate law and whether and how it is hospitable to VC financing techniques. Accordingly, the clauses we are most interested in from a corporate law perspective are contained in the charters.

Second, we believe that charters are even more important in Italy than they are in other countries, such as the US.³⁸ Indeed, shareholder agreements that constitute a significant part of the financing deals are less important, in Italy, as instruments of corporate governance. This is not because they are not used or are novel to the Italian experience. On the contrary, they are widely used in the corporate governance of Italian companies, at the level of both listed and private companies. Thus, we do not expect any resistance from Italian shareholders, whether investors or entrepreneurs, to enter into these types of agreements. However, Italian courts and scholars firmly hold that only agreements that are incorporated in the charter have an erga omnes effect, in the sense that the obligations stemming from those agreements are not owed by shareholders bilaterally or multilaterally, but are entered into also in the interest of the corporation and are therefore opposable to third parties.³⁹ This makes it much easier for shareholders to enforce duties that are commanded by the charters rather than suing for breach of contract under a shareholder agreement and claim for damages or fixed sums as penalties.⁴⁰ For this reason, for instance, it is typical to find right of first refusal clauses in charters rather than in shareholder agreements: if a shareholder does not comply with the clause and sells the shares to a third party, the sale has no effect towards the company. The same is true with regard to tag-along and drag-along clauses, as well as to the other contractual clauses that characterize VC financing. Accordingly, it is well known among Italian corporate lawyers that it is better to put those clauses in charters than in shareholder agreements only. For this reason we expect to find in Italian charters as many clauses as possible, and all the main ones we are interested in.⁴¹

Third, contrary to the German experience, where—as we noted—shareholder agreements enjoy an almost unlimited freedom of contract,⁴² in Italy it is highly debated whether shareholders can insert in shareholder agreements clauses that circumvent mandatory corporate law provisions.⁴³ These uncertainties, coupled with ineffectiveness vis-à-vis third parties, make shareholder agreements less attractive. It was no surprise, then, to find in the charters of our sample clauses affecting the relationship between founders and outside investors, such as co-sale clauses and (various) preferential right clauses.

Finally, charters are important also because they offer evidence about the interpretation of corporate law by specialized lawyers and public notaries. Indeed, under Italian law, public notaries have the duty to verify that charters’ clauses are in compliance with the law. In order to facilitate these controls, notaries’ associations (consigli notarili) have drafted guidelines (massime notarili) with comments and bibliographic references to the benefit of their associates and the business community. These indications have become very important among practitioners because they provide reliable suggestions about what can be done and what cannot be done. Of course, they do not cover any possible clause, and in our data we find many provisions that are not straightforwardly covered by them. In any event, those provisions have been drafted by a specialized law firm or notary, and the notary has implicitly assessed that they are in compliance with the corporate law’s mandatory rules. Even though courts can always declare those provisions null and void, and therefore the notary’s assessment does not limit the discretion of courts in a litigation setting, corporate charters offer a very good and qualified view about what specialist attorneys and notaries believe are innovative provisions, not yet tested in litigation, that are nevertheless in compliance with corporate law.

5 Data

Italian law was amended in 2012 to favor the creation of innovative startups. Accordingly, newly created companies with certain formal prerequisites can qualify as ‘innovative startups’ and enjoy a significant set of benefits related to taxation, labor law, corporate and bankruptcy law.⁴⁴ These companies are recorded in a special section of the companies register held by the Chamber of Commerce in any province. We collected the charters of 5095 of these companies in the period from 1 January 2015 to 31 March 2021 in the provinces of Bozen (105), Florence (195), Genoa (181), Milan (2465), Naples (477), Rome (1232) and Turin (440).⁴⁵ The data show a steady annual increase in the sample, evenly distributed among the provinces surveyed. The examined sample consists of the charters of the 5095 innovative startup companies in pdf format,⁴⁶ 5021 established in the form of limited liability companies and 74 in the form of joint-stock companies (società per azioni or s.p.a.). The statistics therefore show the prevalence of the limited liability company over the joint-stock company, with the latter constituting only 1.4 percent of the sample. This confirms the low attractiveness of the joint-stock companies for early-stage startups due to the excessive costs of incorporation and management,⁴⁷ thereby supporting Italy’s choice to transform the GmbH-type limited liability company in order to promote the creation of startups.

The pdfs of the charters resulting from optical scanning were first made searchable thanks to standard OCR (Optical Character Recognition) software. Then, using indexing software, we identified 1137 charters of companies established online in accordance with legislation no longer in force,⁴⁸ which—following a tick-the-box approach—enabled a choice between the different options set out in a standard model prepared by the Ministry of Economic Development.⁴⁹ To identify them, we searched our data sample for charters containing a specific expression used by the mentioned ministerial model (‘choose one of the following options’). The results showed that online incorporation with the adoption of the standard model was used by just over 22 percent of the overall cases.⁵⁰ Since the scope of our research is to understand what types of instruments outside investors are choosing in the Italian market, our interest does not lie in companies that adopted this ministerial model, even though we cannot exclude that some of them have outside investors that do not know of (or care about) the contractual features of VC financing, or that have chosen, for the reason explained, shareholder agreements to govern their relationship.

The sample of 3958 non-ministerial charters that form the core of our research sample was subjected to further indexing on the basis of a list of 94 keywords (Table 1). The purpose was to identify the charters containing all the typical elements of VC-backed startups: convertible preferred shares; convertible debt; antidilution provisions; liquidation preferences; co-sale clauses (in particular, drag-along and tag-along); lock-up provisions; clauses on the appointment of directors and/or on reserved matters, with enhanced quorum requirements or veto rights of the outside investor or the appointed directors; and IPO clauses. Of course, some of those clauses do not reflect per se the presence of outside investors. For instance, clauses on the appointment of directors and on reserved matters are generally common in the charters of closed corporations where minority shareholders want to have some form of control over the management of the company.⁵¹ The same is true for co-sale clauses, which are also contained in the ministerial model and are widely adopted.⁵² In fact, in our sample of non-standard charters, drag-along and tag-along provisions are vastly overrepresented compared to all other clauses.

Table 1 Keyword list

It was therefore necessary to examine other provisions that, in the light of financial contracting theory, would signal with more precision the existence of an outside investor such as a business angel or a venture capitalist. Among those provisions we focused on three in particular: convertibility, antidilution and liquidation preferences. We would have liked to focus also on class of shares, but charters almost invariably make reference to the possibility of creating different classes of stock, and are not always sufficient to understand if different classes of stock have been actually issued.⁵³ According to financial contracting theory, these characterizing clauses should be considered a reliable proxy for the use of typical VC-financing techniques. Only 183 charters (4.5 percent of the non-ministerial ones) contained at least one of these three provisions.⁵⁴

Work-for-equity (vesting) schemes deserve separate consideration. Despite the great importance of such employee incentive programs, given certain limitations of our data sample,⁵⁵ we encountered only a small number of charters that reproduce qualitatively equivalent US vesting arrangements.

6 Clauses

Here we examine the archetypical clauses contained in the surveyed charters of the startups with outside investors.

6.1 Convertibles

In US venture capital practice, convertibility is a key feature. The popularity of this instrument has often been explained as a result of its ability to respond to conflicts of interest and agency costs characterizing VC-backed firms, especially at the time of exit.⁵⁶ In the charters we collected, convertibility is attached to: (i) instruments most typically subscribed by business angels in the early stages of financing [i.e., convertible bonds, warrants, ‘Simple Agreements for Future Equity’ (SAFEs), ‘Keep it Simple Securities (KISSes)];⁵⁷ (ii) convertible preferred shares;⁵⁸ (iii) pay-to-play clauses; and (iv) automatic conversion clauses triggered by IPOs.⁵⁹

6.1.1 Convertible Preferred Shares

The most surprising result of our empirical research concerns the almost total absence of convertible preferred shares (CPS), which we found in only 6 charters. In US practice, VCs, which bear the cost of the evaluation process, receive preferential governance rights⁶⁰ and a downside protection of their investment through a set of preferences and privileges, mostly as non-participating liquidation preferences.⁶¹ At the same time, they can still participate in the upside converting into common stock at any moment. The mix of liquidation preferences and conversion rights leads to what US observers consider an efficient balance between the interests of the founders and the VC. Each time a liquidity event occurs, the VC, at least where a non-participating liquidation preference has been chosen, will have to assess whether it is economically convenient to exercise the preference and get (at least) the return of the invested capital or, instead, convert to common stock and thus participate on an equal footing with the common shareholders. Thus, the founders have a strong incentive to find exit solutions that incentivize the outside investors to exercise their conversion rights, since conversion rights are usually exercised at liquidity events and because after conversion all economic preferences disappear.⁶²

Our finding reveals an apparent difference between Italian and US practice that needs to be explained. In our sample the majority of preferred shares enjoy a full participating liquidation preference.⁶³ We also report mechanisms where preferred shares work in a way which is functionally equivalent to convertibility, giving the VC both downside protection and upside participation. Accordingly, there is no need for the convertibility mechanisms that are adopted in US practice, where non-participating liquidation preferences actually prevail,⁶⁴ and the only way for the VC to participate in the upside is to convert. When participating preferred shares or functional equivalents to conversion are adopted, convertibility into commons becomes financially useless.⁶⁵ The absence of CPS is also closely linked to the peculiar configuration of antidilution provisions found in our sample.⁶⁶

6.1.2 SAFEs, KISSes and Other Early-Stage Hybrid Convertible Securities

In 21 LLCs’ charters (0.5 percent of the non-ministerial ones) we find clauses regulating a specific type of convertible hybrid securities⁶⁷ which fully reflect the US SAFE-KISS model, with advanced conversion ratio settings and valuation cap techniques.⁶⁸ In US practice, under SAFE or KISS terms the initial contribution does not have to be reimbursed by the company, except in the case of liquidity events (takeover, merger, IPO), dissolution, or bankruptcy (often with a preference over shareholders). Instead, the contribution is intended to be automatically converted into share capital in the event of future financing rounds, at favorable conversion rates.⁶⁹

The fact that this financing technique has been found only in LLCs and not in joint-stock companies might be due to the statistical irrelevance of joint-stock companies in our large sample or to the fact that this form of company is usually not adopted by early-stage startups. At the same time, in joint-stock companies the same functions of the SAFE can be fulfilled by convertible bonds, which are well known and regulated in detail by Italian law;⁷⁰ for LLCs, however, the obstacle posed by Art. 2483 Civil Code to the issuance of bonds in favor of early-stage investors⁷¹ has probably encouraged the adoption of hybrids.

Italian convertible hybrids present complex features aimed at tackling some statutory issues. The first obstacle is the mandatory principle of necessary correspondence between the total value of contributions and the nominal amount of the legal capital,⁷² which is dealt with by using so-called personalized reserves.⁷³ The second obstacle is the need for a shareholder vote to amend the charter and issue the shares the hybrids are to be converted into: only 12 charters require that concurrently with the issuance of the hybrids the share capital be increased by an amount sufficient to serve the conversion rights, while the remaining 9 only foresee an obligation for the shareholders to vote, at the exercise of the conversion rights, the occurrent capital increase. The last approach raises the problem of the preventive waiver of the withdrawal right granted to shareholders dissenting with a legal capital increase.⁷⁴ This risk could be considered and dealt with in shareholders’ agreements – even though this withdrawal right represents a serious obstacle to the optimal transplant of US-style transactions in Italy.

In 89 charters (2 percent of the non-ministerial ones), in the event of the issuance of hybrid securities, the bearers would be entitled to the right to subscribe, together and pari passu with the incumbent shareholders and with priority over third parties, any future capital increase through the conversion of their hybrids into shares. The relevant clauses seem to work similarly to warrants in the US-like VC financing world.⁷⁵

6.1.3 Automatic Conversion: Segregating Conversion, Pay-to-Play and IPO Conversion

In our sample, 56 charters (1 percent of the non-ministerial ones) contain a form of ‘segregating’ automatic conversion clause to prevent a shareholder from acquiring shares of a class different from the one already held. For example, if a shareholder holding A shares buys or subscribes common shares, the latter automatically convert to class A. This use of convertibility is instrumental in keeping the positions of the founders and outside investors always separate, in order to prevent conflict of interests which arises when the same shareholder owns different classes of shares.⁷⁶

In 7 charters (0.1 percent of the non-ministerial ones) we found a clause inspired by the same ‘pay-to-play’ logic of incentivizing participation in multiple financing rounds. Under these clauses, in the event of non-participation in a ‘qualified financing’, all or some of the Series A preference shares held by the outside investor are automatically converted into common shares.⁷⁷ These clauses, however, have been adapted considerably to the context of Italian company law where the shareholder preemption right is a default rule.⁷⁸ In the companies in question (all LLCs), the clause assumes a staggered capital increase reserved to the VC, with the VC’s obligation to pay for the shares at each step and conditional on certain events.⁷⁹ If the VC defaults, its preferred shares automatically convert into common equity.

Finally, in 17 charters (0.4 percent of the non-ministerial ones) we found an automatic conversion clause in the event of an IPO. In all these companies the outside investor was a VC.

6.1.4 Plain Convertible Bonds

We found no trace of the issuance of regular convertible bonds in our dataset. Charters only make a generic reference to the possibility of issuing such securities in compliance with legal rules. Nevertheless, we do not consider this finding to be significant, because a simple analysis of the charters does not necessarily reveal the actual recourse to this method of financing.

6.1.5 Work-for-Equity Hybrid Securities

In 4 charters (0.1 percent of the non-ministerial ones) we found a use of convertible hybrid securities by LLCs to implement vesting schemes. The securities, which are expressly non-transferable, are given to the company’s employees and consultants and accrue a fraction of the right to profits and liquidation from year to year. Upon achievement of targets identified by the directors, the possibility of conversion into common shares is granted at a negligible amount (a few hundred euros).

6.2 Antidilution

One of the most interesting results of our research concerns antidilution clauses. European company law, as far as we are aware, is unfamiliar with antidilution clauses of the type used in US VC financing. In Continental Europe preemptive rights are traditionally seen as the instruments that protect shareholders from dilution.⁸⁰ If the company issues new shares at an undervalue, any shareholder has the right to subscribe the new shares before any outside investors. If the shareholder does not use its preemption right, it can be diluted.

In the VC world, antidilution clauses deal with down-rounds, i.e., capital increases at a price below the one paid in the previous round. In a down-round, if the shareholder has no preemption right, other investors can become shareholders at a price different from the one the shareholder originally paid. From an economic standpoint, this effect simply reflects the fact that the company value is decreased and that previous shareholders have lost value. If the Founder and Fund A paid EUR 500,000 each for common stock (the Founder) and convertible preferred shares (Fund A), and after a year the company needs a further 500k of fresh money but no outside investor is willing to pay that amount without receiving at least a 50% stock in the company, it means that the company’s value has become EUR 500,000 before the investor B’s payment (pre-money) and will be EUR 1,000,000 after investor B’s payment (post-money). The Founder’s common stock and Fund A’s convertible preferred shares have thus lost half of their value.

US venture capitalists, however, want to be protected from this type of dilution in the event of a down-round, because they want the founder to pay, in part or in full, for the loss of value suffered by their shares. The logic is that they paid too much, relying on the founder’s promises that remained unfulfilled, and therefore the terms of the exchange have to be readjusted. The readjustment affects the conversion rate of the convertible preferred shares. US antidilution clauses deal with and modify the conversion rate because US practice treats convertible preferred shares on an ‘as-converted basis’, thereby making irrelevant the actual conversion of the convertibles. Since the company value is less than expected, through the conversion the VC will get more common stock than initially agreed and more rights on an as-converted basis.

Antidilution clauses increase the rate at which Fund A’s preferred stock converts into shares of common stock. How many more depends on the formula that has been agreed during the negotiations. The most protective formula is the ‘full ratchet’ (in our sample no. 15, 0.3 percent of the non-ministerial ones): it works as if Fund A had originally paid for its own convertible stock at the price Fund B is paying for its own convertible stock at the down-round. The most used are ‘weighted average’ formulas (in our sample no. 23, 0.5 percent):⁸¹ they work as if Fund A had originally paid a price which is the average between Fund A’s original evaluation of the company and Fund B’s subsequent one.

In Italy, the structure of deals differs markedly. As we have noted, in our sample we found almost no convertible preferred shares and, in any event, no use of the ‘as-converted’ mechanism. Accordingly, the charters that do tackle the problem have to adopt a different approach. Since the issue is brand-new and there is as yet no full standardization, there are clusters of charters that adopt very different classes of solutions. The most widespread (in 38 charters, 0.9 percent) is a clause that gives the VC a veto right regarding any capital increase that could dilute its participation. Common are also clauses (in 30 charters, 0.7 percent) pursuant to which shareholders have a duty, in case of a down-round, to unanimously vote a second, parallel capital increase (waiving their preemption rights) to be paid at par value by the protected shareholders. The purpose of this second capital increase is to readjust the protected shareholders’ stock. Of course, this mechanism can work with low-par shares, where protected shareholders pay almost nothing for new shares. Moreover, it requires unanimity, meaning that if one shareholder does not vote or does not waive its own preemption right, the down-round cannot take place (or the dissenting shareholder has a statutory exit right at fair value). This is because, under the Italian law governing the LLC, the preemption right cannot be eliminated unless each individual shareholder consents to it or is allowed to exit at fair value.⁸² Thus, the mechanism is not truly effective because of the hold-up problem.

Another group of charters requires, in case of a down-round, a parallel free (nominal) capital increase aimed at creating additional shares to be distributed to the protected shareholders (11 in total, 0.2 percent). Thus, Fund B pays the capital increase with consideration (the down-round) and gets its stock, while the protected Fund A investor gets its new antidilution adjunctive shares through a second, for-free issuance of shares. This mechanism is only possible with no-par shares or with low-par shares. Moreover, it assumes that only some shareholders can benefit from free capital increases and, therefore, that the rule according to which free capital increases equally benefit all shareholders is not mandatory.⁸³

Few charters mix the two systems (2). If the free capital increase aimed at creating new shares in favor of the protected shareholder is not possible, the company has to issue new shares that the protected shareholder can pay for at par value to readjust its participation.

Other charters (8, 0.2 percent) make express reference to the possibility of having the new antidilution shares issued with the consideration paid by the new investor (Fund B), a situation that Italian law explicitly makes possible and which was introduced to be used in these types of transactions.⁸⁴ Only one charter explicitly foresees the readjustment of the stock owned by the protected shareholder, increasing it as a consequence of the down-round. This mechanism is the only one that works as a US-like conversion rate adjustment, exploiting the no-par value of shares.⁸⁵

6.3 Liquidation Preferences

Liquidation preferences, in the paradigmatic US startup financing model, belong to the standard contractual repertoire. Their function of protecting the financial expectations of VC investors and business angels at the time of exit (i.e., downside protection) is unanimously acknowledged in scholarship and practice.⁸⁶ The impact of such preferential rights on the founder’s interests is marginal if the business turns out to be successful, while significant conflict of interests may arise in the event that the initiative turns out to be a living dead.⁸⁷ In fact, in this scenario the proceeds from the sale are likely to be equal to or slightly higher than the liquidation right assigned to the investor, with a consequent disincentive for the founder to invest additional resources and energy in the business project.

The liquidation rights incorporated in preferred shares are triggered at the occurrence of a liquidity event, such as a winding-up, merger, change of control, or sale, with subsequent allocation of the gained proceeds among outside investors and founders. Liquidation rights can take two forms: participating or non-participating. In the latter case, sometimes also called simple preferred, the preferred shares carry only the initial liquidation preference (usually a capped amount or a multiple of the original purchase price, e.g. 1.5X, 2X, etc.);⁸⁸ instead, in fully participating preferred shares the initial liquidation preference is coupled with an additional pro rata participation right in the distribution of the proceeds available—on an as-converted basis—after satisfaction of the holders of preferred shares.⁸⁹ Thus, as stated in VC jargon, participating preferred shares ‘double-dip’ in the liquidation proceeds. Since in current US practice most preferred shares are of the non-participating type,⁹⁰ the conversion option into shares of common stock at the time of exit is essential to get the upside protection. Differently, in participating preferred shares upside protection is provided by the same participation right, making conversion less important.⁹¹

In our data we find, from a quantitative point of view, a significant presence of liquidation preference clauses. From a qualitative standpoint, the contractual design is well conceived so as to allow the achievement of the objectives typically underlying such provisions, as briefly illustrated above. In this perspective, at least with reference to the two standard types of preferential rights (i.e., participating and non-participating liquidation preferences), it seems noteworthy that no significant deviations from the US model clauses, as prepared by the NVCA, are encountered. Looking more closely at our results, there are 146 charters, 132 of which relate to LLCs and 14 to joint-stock companies.⁹² The characterizing feature of this core sample is the considerable degree of standardization of the provisions in question, which suggests positive network and learning economies among the professionals involved.⁹³ This also testifies to the existence of a ‘Delaware effect’ on Italian corporate law by virtue of a transnational circulation of contractual models.⁹⁴ Not differently from what is observable in the Silicon Valley market, this high level of standardization is probably due to the involvement of a small circle of professionals specialized in VC and private equity operations.⁹⁵

In particular, the encountered contractual standardization concerns the definition of the liquidity or distribution event,⁹⁶ as well as the distribution order and mechanisms. The great majority of the relevant charters (no. 105, or 72 percent) provide a full participating liquidation preference,⁹⁷ while no. 3 (2 per cent) charters contain a pure non-participating liquidation preference. The more interesting class, however, is a third one. It concerns provisions (no. 32, or 22 percent) that foresee a preferential distribution to the VC investor when the proceeds from the liquidity event are below a certain target amount, while in the case of proceeds exceeding such a target amount the distribution is made pro rata among all shareholders without distinction.⁹⁸ This class works, accordingly, as US convertible non-participating preferred stock. Indeed, below a certain target amount the US instrument gets the preferential distribution, while above the target amount the shareholder converts into common stock and participates in the upside. The Italian instrument reaches similar results because the stock offers two alternative rights to the holder and, therefore, does not need a formal convertibility mechanism: preferential distribution (below the target) and pro rata participation with the commons (above the target).

US convertible participating preferred shares get a ‘double dipping’, since they cumulate the preference with the subsequent, eventual pro rata participation, while US convertible non-participating preferred shares offer an option—they are non-participating because below the target amount they have a debt-like payoff and in order to participate in the upside they need to be converted into common stock. The end result is that they do not enjoy the double-dipping of convertible participating preferred shares, with the conversion right being central instead. The Italian third genus is actually ‘participating’, because the pro rata participation right is incorporated into the instrument. However, they do not enjoy the double-dipping and, for the reasons explained, have a pay-off which is similar to US convertible non-participating preferred stock. Accordingly, they are referred to by Italian practitioners as ‘non-participating’.⁹⁹

Finally, some other charters (no. 6, or 4 percent) foresee that the VC investor, after receiving the initial liquidation preference, participates once more in the distribution of the liquidation proceeds only after full repayment of the investment made by all other shareholders—a weak form of potential double-dipping.

Importantly, in several charters the liquidation right operates only in correspondence with the first liquidity event, whereas the proceeds deriving from subsequent events are indiscriminately distributed on a pro rata basis among all shareholders, without any distinction between founder and VC. On other occasions, however, it is specified that the liquidation preference, if not fully satisfied at the time of the relevant liquidity event, works cumulatively on any further distributions that may be made until the amount due has been fully paid to the preferred investor.

The relatively high level of standardization we surveyed is surprising considering that some influential scholars envisage certain mandatory constraints on the reception of US-like liquidation preference arrangements.¹⁰⁰ Here, the obstacles to a full legal transplant are supposedly seen in the necessarily profit-driven causa societatis,¹⁰¹ or, similarly, in the ban against a societas leonine clause,¹⁰² which would both be undermined if the VC investor held, from the outset, the right to recover the total capital investment made, with the founder exposed to the risk of getting no substantial return. In particular, especially with regard to the non-participating preferred shares, these scholars argue that the non-participation in the upside, on the one hand, excludes the possibility to qualify the relative holder as ‘proper’ shareholder, resembling more likely a creditor or a partner in a silent partnership;¹⁰³ and, on the other hand, guarantees an ‘immunization’ from the entrepreneurial risk in contrast with the societas leonine clause.¹⁰⁴ We are not very convinced by these arguments. First, the idea of a corporation as a ‘peer group’, with shareholders all devoted to a common purpose, is losing ground also in Italy.¹⁰⁵ Second, the total or partial satisfaction of the liquidation preference (both in the downside and the upside) requires in any case a liquidity event with a positive cash-flow. Thus, already ex ante, the preferred shareholder, too, assumes an entrepreneurial risk, not differently from any other shareholder. Third, preferred shareholders are entitled to receive, upon occurrence of a liquidity event, a liquidation preference after creditors’ claims are satisfied. For this reason, they are always residual claimants and never holder of a subjective right to a fixed claim against the company or other shareholders. In case of insolvency, they cannot be treated as company creditors. For all these reasons, we do not think that liquidation preferences, including those of the non-participating type, result in a ‘deviation’ from the traditional shareholder paradigm.

Furthermore, it is interesting to observe that legal practice has somehow envisioned a means to bypass similar objections—even if there is doubt whether this is possible in a judicially enforceable way, should courts agree with such critics. In fact, very few provisions on liquidation rights state that before any preferred distribution shall be made, all shareholders (including common shareholders) have the right to receive first from the distributable proceeds a certain amount of money that ranges between the shares’ par value, EUR 100 and/or EUR 0.01.¹⁰⁶ It is certainly a formalistic ‘escape method’, but perhaps as formalistic as the societas leonine clause, which requires exclusion from any profit or loss participation. Obviously, as stated, in a hypothetical court case a judge following the mandatory constraints theory would check whether such a contractual provision is in fraud of the law and thus invalid.¹⁰⁷

6.4 Co-Sale Clauses

Co-sale rights are essential tools of the contractual architecture of VC financing.¹⁰⁸ Through drag-along clauses, the venture capitalist is able to force the exit of the founder (or of another early investor) and to liquidate its investment. Tag-along provisions, on the other hand, allow the investor to benefit from any liquidation of the investment promoted by the founder.¹⁰⁹ It is no coincidence that such clauses have been included in the model articles prepared at ministerial level¹¹⁰ and that, as a result, we are now witnessing a high degree of standardization of these contractual terms. However, this contractual drafting suffers from heavy regulatory constraint, represented by the majority interpretation, which we do not share, that equates the exercise of the drag-along right with the forced withdrawal of the (dragged-along) shareholder and thus requires that the purchase price offered by the prospective acquirer reflects the fair valuation of the shareholder’s stake.¹¹¹ This limitation evidently constitutes an obstacle to VC-financing transactions, in that it allows the dragged-along shareholder to challenge the price offered to the selling party, opening up the way to strategic dissent (i.e., holdout) by minority ‘cat and dog’ shareholders jeopardizing the speed or even feasibility of the exit.¹¹²

Overall, in our data, we have 1310 charters with a drag-along and/or tag-along clause (33 percent of the non-ministerial ones). As explained above, given these extremely high numbers, a refining restriction to the charters of the startups with outside investors (i.e., containing provisions on convertibility, antidilution and/or preferred liquidation rights) seemed fruitful. Accordingly, in our sample, 171 of the 183 startups contain co-sale provisions (equal to 93.95 percent). Except for eight, all of such provisions with co-sale clauses provide for a drag-along right (163, or 95.32 percent).¹¹³

With reference to the drag-along right, the relevant charters constantly refer to the fair value determination, so that no contractual attempt can be observed to escape a principle that is clearly considered imperative. Only in two charters does the relevant clause foresee that the equity value offered by the third party to the dragged-along shareholder shall not be significantly below—but without further specification or clarification—the fair value granted according to the applicable provisions.¹¹⁴ However, more than one third of the charters with a drag-along provision regulate the criterion for quantifying the exit price in more flexible ways. In this regard, most clauses require that the prospective acquiror shall offer a minimum price for the purchase of the company (e.g., a fixed amount or a multiple of the company’s revenues), with such minimum quantification sometimes to be replaced by the ordinary fair value estimation after the end of a certain reference period (i.e., X years after the incorporation); other clauses specify that the purchase price following the exercise of the drag-along right shall be the highest between the withdrawal price (or fair market value) and another reference value (e.g., X times the investment made or post-money valuation); many clauses provide that if the offered purchase price is not equal to the market value, the dragging-along shareholder him/herself can pay the relative difference or otherwise renounce the exercise of the drag-along right. One article of incorporation in particular dictates a very analytical mechanism for determining the price corresponding to fair value based on EBITDA, to which a minority discount and a control premium must then be applied.

6.5 Work-for-Equity (Vesting) Schemes

Vesting schemes are of crucial importance in the growth path of an innovative startup.¹¹⁵ In addition to positive loyalty effects, they contribute to mitigating opportunistic behavior by directors and other key personnel of the VC-funded company by making it more expensive to leave the firm (alignment of interest).¹¹⁶ In US VC practice, industry-standard vesting for early-stage companies is a one-year vesting cliff and monthly vesting thereafter for a total of four years.¹¹⁷ The vesting period may lapse earlier in the case of a (single or double) trigger acceleration, such as a sale, merger or listing of the company. Only vested shares are freely transferrable. On the other hand, unvested shares, upon termination of the employment of the shareholder, can be repurchased by the company or its assignee either at cost or at the current fair market value, depending usually on whether a bad leaver or a good leaver event occurs. Often, founders will get somewhat different vesting provisions than the rest of the employees.¹¹⁸

In Italy, although—to the best of our knowledge—solid statistical data are not available, work-for-equity incentive plans seem quite widespread, gaining new momentum thanks also to the reform of the law on LLCs.¹¹⁹ With specific regard to our empirical analysis some caveats are necessary. First, only marginal attention was paid to the many boilerplate provisions authorizing a share capital increase with allotment of the newly issued shares to the founder and/or employees and to be paid up by the professional services promised.¹²⁰ Second, there is a significant chance that more detailed regulation of incentive compensation is contained in non-publicly available shareholders’ agreements.¹²¹ Third, given that often the charters merely authorize the subsequent adoption of vesting schemes by the board of directors or a shareholder resolution, such arrangements are not captured in our dataset.

Notwithstanding these limitations, we encountered some charters that exactly reproduce the contractual design of US vesting schemes.¹²² In particular, a call option is assigned to the company or some shareholders, which allows the holder to acquire the unvested shares or share options from the founder or employee upon the occurrence of either a good leaver or a bad leaver event,¹²³ with the repurchase price being equal to the fair value in the first case and to the face value in the second case. Generally, bad leavers lose the right to retain any shares or share options (or, at least, the vested but unexercised share options), while good leavers hold on to their vested share options after leaving the firm. Finally, noteworthy is that some charters permit the issue of (hybrid or quasi-equity) participatory financial instruments¹²⁴ in order to acquire professional services from the key personnel, provided also with a conversion right into equity once such services are correctly performed.¹²⁵

7 Summary

Italian startuppers and their financiers use contractual techniques that appear to be functionally in line with the needs highlighted in the literature on financial contracting. The most recurrent contractual features of startups with outside investors that we document are co-sale clauses and preference rights. According to our data, outside investors in Italy want co-sale clauses and liquidation preferences more than any other contractual instrument.

Co-sale clauses are almost a regular feature in startup charters, becoming popular thanks also to the standard ministerial model aimed at facilitating electronic startups’ incorporation. There is a high level of standardization, and virtually all clauses reflect the problem of fair value protection in the case of a shareholder exit. The need for fair value protection in drag-along clauses is not expressly mandated by the law, but is adopted by the majority of commentators, public notaries and courts. A certain amount of charters seek to mitigate this mandatory constraint through tailor-made criteria of evaluation, probably aimed at limiting the rigidity of a one-size-fits-all criterion based on fair value determination. These criteria try to limit the discretion of the expert witness that can be appointed by the court in case of discussion on the fair value of the dragged-along stock. Whether these attempts will be successful in case of litigation concerning their validity remains to be seen.

Preferred shares with liquidation preferences are less problematic. Liquidation preferences attached to preferred shares are highly standardized. They are already widely used in private equity practice and in restructuring. Hence, there is no major variation in the way they are drafted. We report a prevalence of participating preferred shares over non-participating shares. In addition, we also find liquidation preferences that work as functional equivalents to US convertible non-participating preferred shares. Both the prevalence of full participating preferred shares and the emergence of such functional equivalents explain the absence of convertible preferred shares – a striking difference with the US world.

Antidilution clauses are a new component of Italian charters emerging from the transplant of mechanisms adopted by VCs in US practice. Since convertible preferred shares are not used, antidilution mechanisms are not based on the conversion rate but on the attribution of additional shares to the protected shareholder. Given that these contractual techniques are brand-new in the Italian experience and there is still uncertainty about the best suited mechanisms to adopt, we observe a significant level of variability in the drafting of the relevant clauses. This is certainly a field that deserves further research at an Italian and European level, since the topic seems to be under-researched across Europe.

SAFEs and KISSes are transplanted in Italy also through hybrids, which are structured to comply with legal capital rules. However, charters are not fully informative about how market participants use them. The same is true with regard to work-for-equity incentive schemes, even though it appears that such vesting schemes are implemented in Italy and reproduce the current US standard. Our results concerning hybrids and vesting schemes show a limitation of our study, which relies exclusively on charters and therefore does not offer a complete picture of their provisions. Future research should expand the spectrum of investigation in order to include also the relative arrangements.

8 Conclusions

Italian startups adopt almost exclusively the LLC form, which was transformed through a series of statutory reforms that started in 2012 and ended in 2017. Those reforms were driven by the need to offer startuppers and venture capitalists a new company form. Curiously, the majority of Italian commentators analyzed the reforms only with reference to the phenomenon of crowdfunding and ignored the impact on venture capital financing. Still, a small group of researchers analyzed the reforms from a venture capital perspective and raised concerns about the viability of certain solutions adopted in international practice, or, more radically, expressed a very negative view on the overall suitability of Italian law to efficiently structure venture capital deals.

The empirical data we collected show that specialized lawyers and public notaries do not share the grim view expressed by some scholars. Startups with outside investors form a small percentage of the new companies that have been registered as ‘startups’ in the Italian register, but have charters that reflect in part or in full the features predicted by the financial contracting theory as well as the US experience. Where differences from US practice exist, the relevant clauses achieve functionally equivalent results. Indeed, we report market practices that are pushing the envelope of Italian corporate law. Startup law is transforming Italian corporate law, which is deeply rooted in the tradition of Continental Europe, Germany in particular. Market practices have turned out to be more advanced and courageous than we originally thought possible from the perspective of risk-adverse parties, and from this standpoint the Italian reforms of 2012-2017 regarding the LLC appear, though not completely, successful. Of course, there are issues that can be greatly improved. With regard to the topics we have investigated here, they mainly concern non-participating liquidation preference and the general principle of fair value determination in case of exit. From this perspective, full liberalization of the LLC regulation would eliminate the grey areas created by a legal environment where the boundaries between mandatory and default rules are sometimes unclear. This uncertainty heavily depends on scholars’ propensity to extract high-order principles from statutory materials in order to infer mandatory rules or to solve interpretative questions in favor of the mandatory alternative, usually out of concern to prevent abusive corporate behavior at the expense of the weaker contractual party. The adoption of a counter-Satzungsstrenge principle according to which any provision not explicitly qualified as mandatory is a default one would prevent this process of mandatory rules’ creation.

Nevertheless, deals are made and charters appear to be very advanced, and this marks a significant difference between the business world, exposed to international competition, and some national, idiosyncratic theoretical discussions that seem detached from business needs and economic theory. It is our hope that, through our empirical analysis, courts can learn about market practices and needs that are not adequately explored by traditional legal studies. In this way, an apparently new legal issue can be handled with greater awareness, and not considered as an extravagant outlier which might conceal, in hindsight, abusive behavior.