An Analysis Of The V-Dem Data On The Decline of Democracy In India Reveals Interesting Insights—About V-Dem

Recently, the Swedish organisation Varieties of Democracy (V-Dem) came out with an impressive looking ranking of the extent of liberal democracy across the world.

Their distinctive approach is data-driven, with a very large number of variables. All the data is available online. The set of parameters are impressive, and report the data from as far back as 1789 to present. It reports minimum, maximum and standard deviation of the indices.

Tragically, according to this data India is an Elected Autocracy.

Autocracy is defined as “a system of government by one person with absolute power”. The classification means that elections in India simply assert that power.

The report has, not unexpectedly, energised the Indian media elite who are super-sensitive to cues emerging from the Euro-American intellectual/academic set. Nothing sets them off on a democracy defending frenzy like some fresh data from this set that paints the government of Prime Minister Modi in a poor light.

Still, the common man, as it were, takes a great deal of pride in our noisy, festive democracy. So how could this have happened? How did India descend into a dictatorship unnoticed by the hundreds of millions who voted for the government barely two years ago?

Let us investigate the V-Dem data to understand. After all, as it comes from a Swedish think tank, we have been conditioned to assume that it must bring the weight of a well-constructed and irreproachable moral architecture with it.

For purposes of comparison, we will evaluate India’s current reprehensible position against what every Indian knows as the Emergency, the only period post-independence when India was dictatorially ruled by decree - by Indira Gandhi.

This was a period when India was under what V-Dem ought to classify as Complete Autocracy. There were no elections, the opposition and press were jailed, all personal freedoms stood curtailed, and more.

Just how close is India today to that nadir? Even lower, on some parameters, according to V-Dem. Join us on this revelatory journey, and bear with us patiently, as we walk together through the eye-opening and mind-boggling data*.

(All parameter descriptions below are quoted verbatim from the V-Dem codebook without alteration).

Parameter: Electoral Democracy Index

Description: The electoral principle of democracy seeks to embody the core value of making rulers responsive to citizens, achieved through electoral competition for the electorate’s approval under circumstances when suffrage is extensive; political and civil society organisations can operate freely; elections are clean and not marred by fraud or systematic irregularities; and elections affect the composition of the chief executive of the country. In between elections, there is freedom of expression and an independent media capable of presenting alternative views on matters of political relevance.

Average (1952-2019): 0.664

1975: 0.415

1976: 0.407

2018: 0.476

2019: 0.450

Observations: The average belies the substantial gains made in recent decades when the index rose to the mid 0.700s. This happened in the 1990s, when India had four general elections and six governments, clearly indicating a very high level of electoral democracy.

Under the BJP, V-Dem shows that India reports an index value close to that of the Emergency. This is particularly interesting because there were no polls held then at all. Clearly, the experts who generated/analysed the data must have had some deeper rationale than just elections.

Parameter: Liberal Democracy Index

Description: The liberal principle of democracy emphasizes the importance of protecting individual and minority rights against the tyranny of the state and the tyranny of the majority. The liberal model takes a "negative" view of political power insofar as it judges the quality of democracy by the limits placed on government.

Average (1952-2019): 0.522

1975: 0.302

1976: 0.280

2018: 0.369

2019: 0.340

Observations: V-Dem concludes that the zenith of Indian liberal democracy was in 1997-98, i.e. after the Vajpayee government was undone in 13 days, followed by a revolving door of H D Deve Gowda, I K Gujral, and the fragile Vajpayee government that was pulled down again.

However, today, if we are to understand V-Dem right, under the stable BJP-led government India ranks nearly as badly as it did during the Emergency. This might be because the leading opposition politicians are currently imprisoned in their own ideological and dynastic enclosures, and cannot get out; try as they might to swim by sea or drive from pillar to post on land.

Parameter: Participatory Democracy Index

Description: The participatory principle of democracy emphasises active participation by citizens in all political processes, electoral and non-electoral. It is motivated by uneasiness about a bedrock practice of electoral democracy: delegating authority to representatives.

Average (1952-2019): 0.412

1975: 0.227

1976: 0.228

2018: 0.288

2019: 0.278

Observations: As per V-Dem, participatory democracy peaked in 1998-99 too. This makes sense because the electorate had two governments in span of a single year. However, again the current period is ranked almost on par with the Emergency. This is despite the fact that 2014 and 2019 had the highest turnouts in history. One could argue that the experts have some explaining to do.

Parameter: Deliberative Democracy Index

Description: The deliberative principle of democracy focuses on the process by which decisions are reached in a polity. A deliberative process is one in which public reasoning focused on the common good motivates political decisions—as contrasted with emotional appeals, solidary attachments, parochial interests, or coercion. According to this principle, democracy requires more than an aggregation of existing preferences. There should also be respectful dialogue at all levels—from preference formation to final decision—among informed and competent participants who are open to persuasion.

Average (1952-2019): 0.563

1975: 0.314

1976: 0.287

2018: 0.299

2019: 0.267

Observations: Yet again, deliberative democracy peaked in 1998, when the Vajpayee government was brought down, respectfully of course. But 2018-2019 constitutes the bottom of the barrel for deliberative democracy in India with the index showing a lower score than even during the Emergency.

One can only speculate that, when V-Dem’s experts were generating consensus on the matter, they concurred that during the Emergency nearly all participants of democracy were in jail and having a reasoned and respectful communication with the police, shorn of emotional appeals and coercion.

Parameter: Egalitarian Democracy IndexDescription: The egalitarian principle of democracy holds that material and immaterial inequalities inhibit the exercise of formal rights and liberties, and diminish the ability of citizens from all social groups to participate.

Average (1952-2019): 0.429

1975: 0.283

1976: 0.282

2018: 0.265

2019: 0.259

Observations: Impressively, according to V-Dem, the period of Emergency was not the worst period for egalitarian access to rights. Rather it is right now, when the largest number of people voted to democratically elect the ruling NDA; presumably egalitarianism was better served when all constitutional rights were suspended by decree for everyone, by the then Congress government of Indira Gandhi.

At the very least, it appears V-Dem has a Scandinavian sense of humour: The best period for egalitarian democracy as per their data was in 1998, when citizens were able to exercise their right to democratic process (and then soon after in 1999).

Parameter: Freedom of Association

Description: To what extent are parties, including opposition parties, allowed to form and to participate in elections, and to what extent are civil society organisations able to form and to operate freely?

Average (1952-2019): 0.772

1975: 0.605

1976: 0.589

2018: 0.661

2019: 0.606

Observation: Impressively, 1998 was again the best year for freedom of association. 2019 and 1975 are both at par when it comes to freedom of association; in 1975 the opposition were collectively in jail, able to associate easily in confined spaces.

Parameter: Clean Elections Index

Description: Free and fair connotes an absence of registration fraud, systematic irregularities, government intimidation of the opposition, vote buying, and election violence.

Average (1952-2019): 0.786

1975: 0.648

1976: 0.646

2018: 0.617

2019: 0.577

Observation: 1998 continues to be the zenith of Indian democracy. Compared to the present, which rates as the worst period of Indian democratic history, the environment during the years of the Emergency are rated free and fair, probably because there weren’t any elections then.

Parameter: Legislative Constraints on Executive

Description: To what extent are the legislature and government agencies e.g., comptroller general, general prosecutor, or ombudsman capable of questioning, investigating and exercising oversight over the executive?

Average (1952-2019): 0.806

1975: 0.797

1976: 0.667

2018: 0.710

2019: 0.648

Observations: Surprisingly, in 1976 when the legislature was suspended, they had greater ability to place checks and balances upon the executive under someone ruling by decree, as compared to 2019.

The best oversight was in the late 2000s during UPA-2. This must have been necessary because, at that time, multiple major scams necessitated a significant amount of oversight, at least in quantity if not quality.

Parameter: Deliberative Component Index

Description: The deliberative principle of democracy focuses on the process by which decisions are reached in a polity. A deliberative process is one in which public reasoning focused on the common good motivates political decisions—as contrasted with emotional appeals, solidary attachments, parochial interests, or coercion.

Average (1952-2019): 0.844

1975: 0.680

1976: 0.616

2018: 0.508

2019: 0.507

Observations: As with the deliberative democracy index, the component index is also indicative of the lack of freedom at present. The data suggests the deliberative process can be better conducted during an Emergency, in the absence of any democratic process with the opposition in jail.

During UPA-2, when the parallel NAC ran the government, this component registered an all time high rating. Since the NAC were unelected, they could clearly act without emotional appeal, parochial interests or coercion. Those with a longer memory will recall this was evident on the face of then Prime Minister Manmohan Singh through much of that period.

Parameter: Egalitarian Component Index

Description: To what extent is the egalitarian principle achieved?

Average (1952-2019): 0.582

1975: 0.598

1976: 0.598

2018: 0.433

2019: 0.482

Observations: This one is truly bizarre. The Emergency period ranks higher than full-period average for this index. In fact it ranks better than UPA-1 and UPA-2 eras. This is perhaps explained by the fact that everyone could get into trouble equally easily during the Emergency.

Many of the reported punishments were also applied impartially. V-Dem appears to be arguing here that substance is the key, not form.

Parameter: Equal Protection Index

Description: How equal is the protection of rights and freedoms across social groups by the state?

Average (1952-2019): 0.704

1975: 0.707

1976: 0.707

2018: 0.586

2019: 0.614

Observations: As with the egalitarian component, the Emergency was a period of greater equality of rights and freedoms for all, compared to the all time average. And certainly much more so than the current NDA government. In all likelihood, this was because fundamental rights were suspended at the time by decree for everyone equally.

Parameter: Equal Access Index

Description: The Equal Access subcomponent is based on the idea that neither the protections of rights and freedoms nor the equal distribution of resources is sufficient to ensure adequate representation. Ideally, all groups should enjoy equal de facto capabilities to participate, to serve in positions of political power, to put issues on the agenda, and to influence policy making.

Average (1952-2019): 0.797

1975: 0.889

1976: 0.889

2018: 0.436

2019: 0.581

Observations: This is perhaps the most controversial component of this report, and it is mind-boggling in its implications for democracy worldwide. The data from the Emergency years shows the highest value in this index, far above long term average.

It indicates that when all rights and freedoms were curtailed and representative democracy was halted, when everyone had zero access to these - that was when things were at their most equal. An insight for the ages.

Parameter: Election Commission Autonomy

Description: Does the Election Management Body (EMB) have autonomy from the government to apply election laws and administrative rules impartially in national elections?

Average (1952-2019): 2.479

1975: 1.487

1976: 1.23

2018: 1.136

2019: 0.771

Observations: Surprisingly, the autonomy of the ECI somehow nosedives in 2014 after staying high for years. It likely had nothing to do with the fact that at long-last a non-Congress party got the majority vote. In fact during the Emergency when there were no elections, the ECI was considerably more autonomous since they were free to do anything else they liked, barring official duties.

Parameter: Election Voter Registry

Description: In this national election, was there a reasonably accurate voter registry in place and was it used?

2014: 1.136

2019: 0.732

Observations: Surprisingly, this parameter drops from an all-time high in 2014, to the lowest in decades five years later, for no apparent reason. It is something only V-Dem’s panel of experts can explain.

Parameter: Election Other Voting Irregularities

Description: In this national election, was there evidence of other intentional irregularities by incumbent and/or opposition parties, and/or vote fraud?

1962: 1.661

2019: -0.059

Observations: The 1962 elections rate as the best election in terms of lack of irregularities, as the country descended towards defeat against China. The 2019 election is by far the worst. So bad in fact, that it is the only negative index here. Indeed, all three elections in the 1990s before Electronic Voting Machines (EVMs) rate higher than 1.0.

Parameter: Election Free and Fair

Description: In this national election, was there evidence of other intentional irregularities by incumbent and/or opposition parties, and/or vote fraud?

2014: 1.691

2019: 0.594

Observations: As with previous parameters, this starts out from a near all time high in 2014 and drops to an all time low in 2019. It’s not clear why it didn’t drop below 0 because the irregularities parameters somehow managed to struggle across that line.

Parameter: Election Loser Accepts Defeat

Description: Did losing parties and candidates accept the result of this national election within three months?

2014: 1.358

2019: 1.163

Observations: While other data is marked in just how absurd they are, this one is in fact the most insightful (and hilarious) of all. The highest rating of this parameter was registered in 2014… and in 2019 it registered the lowest. No need to wonder why.

To summarise these observations

The data repeatedly suggests that the best time of Indian liberal participatory democratic history was between 1996-99.

This seems to be borne out by the fact that the electorate could repeatedly exercise their franchise - three times over four years. If you missed an election, you just had to wait a few months for the next.

Multiple career politicians and regional satraps could dream of becoming prime minister, and we know they did. This appears to be the ideal form of liberal democracy according to the V-Dem metric.

In comparison, from their perspective, the first stable democratic mandate in a generation given to the NDA is for the most part worse than the entire democratic process being replaced with rule by decree.

While the effort to collect so many interesting parameters and construct a mathematical model out of it is indeed praiseworthy, this exercise is also a great example of the phenomenon known as garbage-in-garbage-out (GIGO) - if you feed nonsense into a mathematical model, it will report rubbish.

Data Science endeavours to construct interpretable models out of fair and unbiased raw data. The effort to construct such a model is interesting, but is fatally flawed if the data is too. Several dozen more parameters exist, some reasonably sane, some just as poor as the ones listed above.

A professional ethic would endeavour to ensure that inference is performed on valid data. As demonstrated, this data fails the test. But it does provide sufficient hilarity while at it.

If one were to be charitable, it could be said that due to the nature of the data as described here, the conclusions about India by V-Dem are complete nonsense. Hopefully their data scientists will recognize the need to scrub their data better, and will make the effort to collect data that does not look completely out of touch with reality, and repeat their exercise.

After all it’s a matter of their own credibility - there’s no telling how many other countries have the same atrociously poor data and correspondingly bad inferences.

Having said that, we had a look at the International Advisory Board of V-Dem, presumably the people who must have had a say in the central thrust of the final report.

These are names listed on the V-Dem website: Niraja Gopal Jayal (Professor and currently Chair of the Centre for the Study of Law and Governance, Jawaharlal Nehru University, New Delhi); Pratap Bhanu Mehta (former president of the Centre for Policy Research, New Delhi); and Aitzaz Ahsan (Senior Advocate of the Supreme Court of Pakistan). Make of that what you will.

*(V-Dem helpfully provides reference material including a codebook, which describe each of the indices of their multivariate formula used to classify countries. The data is available here in multiple formats including CSV, R, STATA and SPSS. Different variables from the V-Dem raw data are listed and reviewed, for the years 1975, 1976, 2018 and 2019. Some data from other years is also presented).