Matt's Five Points, October 24: Not Your Father's Divided Government
Welcome! I’m Matt Glassman, Senior Fellow at the Government Affairs Institute at Georgetown. Here’s my weekly newsletter that always includes five points at the intersection of DC politics and political science, and often some links to interesting things I’ve been reading or listening to this week.
In this issue, I discuss divided government, the electoral college, whether the size of a Dem House majority matters in 2019, Trump’s continuing weakness as president, and turnover in the House.
Feedback? Drop me a line or just respond to this email. And you can always catch up with me on Twitter.
Divided Government Comes in Multiple Flavors.
And the politics of each flavor are somewhat distinctive.
Forecasts of the 2018 election indicate the 116th Congress will probably result in Democrats controlling the House and the GOP controlling the Senate, so let’s talk about divided government.
First off, it’s something that is more or less distinctive to the so-called American System of democracy. Generally, in parliamentary systems—which are, by far, the most common form of democracy in the world—it’s not possible for the executive to be controlled by a different party or coalition than the legislature, as the executive is selected by legislature. Only in the American System—with its separation of powers—does the executive have an independent electoral base outside the control of the legislature.
Second, divided government in America is rather common, particularly in the modern age. Since 1933, there have been 42 congressional elections. Exactly half of them (21) have produced clear unified governments, with the same party clearly in control of the House, Senate, and Presidency. Twenty of them have produced divided government, and one (2000) produced a tie in the Senate, which led to unified government for a few months, and then divided government after a member of the majority party switched affiliations and began caucusing with the other party.
These topline numbers, however, mask the fact that divided government comes in more than one flavor in the United States. Given that there are three relevant political institutions, there are actually three arrangements of divided government. The party that doesn’t hold the presidency can hold either the House, the Senate, or both the House and the Senate.
If you ask people to think about divided government, they might bring to mind President Clinton squaring off against Newt-Gingrich in the mid-90’s. Or they might think about President Obama squaring off against the John Boehner-led Tea Party in 2011. But notice that those were two different arrangements of divided government. In the Gingrich case, the GOP controlled both chambers of Congress. During the 112th and 113th Congresses (2011-2014), the GOP only controlled the House. The Senate was controlled by the Democrats, under the leadership of Harry Reid.
Traditionally, same-party control of both chambers of Congress is a much more common variety of divided government. Of the 21 federal elections that produced divided government since 1933, 15 of them featured the dynamic of the Clinton-Gingrich years: one party-controlling both chambers of Congress and the other party controlling the presidency. The other six included five where the president’s party controlled the Senate but not the House (Reagan’s first three Congresses and the two Obama Congresses after the 2010 Tea Party election) and just one where the president’s party controlled the House but not the Senate (the 2000 election, after Jeffords switched parties).
Are there meaningful distinctions between these various forms of divided government? Absolutely. While divided government in America inherently pits Republicans vs. Democrats, there are both structural features built into the legislative process and political dynamics operating in the public sphere that make unified congressional control different than one-chamber congressional control. The most obvious of these is the direct constitutional link between the presidency and the Senate: nominations. As the constitution provides no formal role for the House in the confirmation of judges and executive branch officials, control of the House alone does not actually create divided government in regard to the nomination process (or the treaty process). Particularly now that the filibuster has been eliminated on all nominations, GOP control of the Senate implies that the record-setting pace of judicial nominations under president Trump will continue in the 116th Congress.
A more subtle distinction is the location and control of political conflict. When one party controls both chambers, there’s a stronger possibility of legislation emerging from Congress that the president feels the need to veto. Indeed, of the twelve vetoes President Obama issued, ten of them came in the final Congress of presidency, when the GOP controlled both the House and the Senate. During the two Congresses in which the GOP controlled only the House, Obama didn’t veto a single bill. His other two vetoes came in his first Congress, when the Democrats had unified control.
It’s pretty obvious what is going on here: the Democratic Senate was refusing to approve (or even consider) legislation that Obama might have to veto. In effect, instead of the high-profile divided government vetoes (like the ACA repeal sent to Obama or the appropriations bills the GOP sent to Clinton in 1995), the political conflict was taking place between the chambers, rather than between the branches. This doesn’t mean the White House wasn’t involved, both privately and publicly, but it does mean that the opposition party couldn’t force veto-level confrontations into the public sphere, where they’d likely be magnified.
So don’t expect many high-stakes vetoes from Trump next year; most of the things that might be on that list are going to be buried in the GOP Senate, even in cases the president might relish the opportunity to veto them. Conversely, you shouldn’t take veto threats from the president all that seriously under one-chamber divided government; it’s not particularly credible for the president to threaten to veto something that overcame a filibuster in a chamber in which his party also has agenda control, save for non-partisan issues that squarely line up Congress vs. the president.
Perhaps even more subtle is that one-chamber divided government is likely to result in the lack of a budget resolution in Congress. Indeed, Congress failed to pass a budget resolution in any year between 2011 and 2014, the four years in which chamber control was split between the parties. This takes on outsized importance in contemporary politics, because most of the high-profile legislation moving through the Senate right now is doing so under expedited procedures: nominations are proceeding under the nuked filibuster rules that allow bare majorities to gain cloture, Obama-era regulations were repealed via the debate-limited Congressional Review Act process, and both the failed ACA repeal and the December 2017 GOP tax cuts were moved through the reconciliation process. Absent a budget resolution and reconciliation instructions agreed to by both chambers, however, there will be no reconciliation process offering expedited procedures in the Senate in 2019.
Contrast this to ACA repeal in the 114th Congress. . Because the GOP had unified control of both chambers in 2015-2106, they were able to produce a budget resolution, reconciliation instructions, and ultimately the ACA repeal bill that could ride along the reconciliation process. As everyone expected, the bill was ultimately vetoed by president Obama, but the public sphere operation of that bill differed quite significantly from the previous 50-odd times the GOP had moved ACA repeal votes through the House only, just to see them die in the Senate.
In sum, the likely divided government in 2019 is not going to operate like the common “textbook” divided governments of the Clinton era. Instead, divided chamber control produces its own distinctive form of politics. It may also be the arrangement most amendable to total gridlock. There is some political science evidence that divided-chamber control results in less of the policy agenda being enacted than even traditional divided government. Perhaps because of the blunt visibilty and counter-majoritarian nature of the veto, combined with the public expectation of presidents to govern, traditional divided government results in more compromises than chamber-division, where decentralized legislators with less individual responsibility to the whole find it easier to dig in against one another when their policy positions are polarized against each other.
Trump continues to be a weak POTUS, part XXIV.
And it’s his fault. He’s not good at guarding power.
I continue to believe Donald Trump is as weak president in danger of a failed one-term presidency. He just doesn’t seem to be learning how to amass or guard his political power. I subscribe to the general Neustadt argument on presidential power, and I’ve written about it plenty before: on Trump’s general Neustadtian weakness, his trouble managing the White House, his cheerleader role in congressional agenda-setting and legislation, and a host of tweetstorms about his various poor moves that have sapped his power.
Now, all presidents are powerful in an absolute sense, but we differentiate them by their ability to influence public policy outcomes. Presidents compete with lots of other actors in the government—elected members of Congress, appointed administration officials, civil servants, White House staffers—for influence over legislation and executive branch administration. Ultimately, a president’s success or failure rests on their ability to persuade others that the costs of opposing him are too high.
And, in this sense, Trump just looks supremely weak in DC. He can’t get the GOP to do anything legislatively that is on his agenda but not theirs. He constantly complains about his own cabinet officials, who appear to ignore him regularly, but is too boxed in politically to fire them. Private sector “allies” loudly walk away from him the minute he crosses them. And he can’t even stop the record-setting departures and the legendary-level leaks at his own White House, where power-hungry staffers engage in endless intrigue, precisely because they know it is so easy to manipulate the president if you can get the face time.
If you believe, like I do, that Neustadt was correct the path to sustained power in the presidency is the developing and guarding of this informal influence by having a stellar professional reputation in Washington as someone who gets his way and punishes those who stand in his way, Trump has spent two year flailing around.
One aspect of this that continues to amaze me is Trump’s impulsive policy pronouncements. Despite his supporters’ belief that he is “tough” and doesn’t bend to political pressure, quite the opposite seems true. Trump constantly reverses his positions and backs down from stands that he is allegedly taking. This happens all the time in foreign policy. He questions NATO Article V. He’s going to close the border. He says he believes Putin over U.S. intelligence services. He says he believes MBS about the killing of a journalist. And on and on and on.
On all of these things, Trump eventually reverses course, usually within a a few days, after the political pressure from his enemies and allies grows. That is the essence of how a president can make himself look weak. You just keep backing down from public positions, and pretty soon everyone knows that if they just apply enough pressure, you’ll back down next time too. His word—even his public word—is garbage. And everyone in Washington knows it. And that’s very dangerous for a president.
The really stupid thing, for Trump, is that this is readily avoidable. All presidents get hemmed in politically at times, and all presidents have to take positions they might not actually want to take, because they won’t be able to sustain themselves politically on another course. But for most presidents, these concerns are teased out in private, with their advisors. Trump just blurts them out.
And that’s costly. Because instead of seeing the president take a strong position, stick with it, and get backed by a wide array of actors, we see the president blurt out a dumb position, defend it for a day or two, and then come around to the Washington consensus about what the position should actually be. You end up in the same place, except everyone knows you didn’t lead and, even worse, you had to weakly give in to the influence of others.
On a different dimension—the backing of electoral candidates—you can see the president learning how this works. At first, he was sort of just backing people willy-nilly, perhaps because he liked them or thought his influence could help them. This led to repeated dumb positions in the Alabama special election, where Trump first backed Luther Strange in the primary and then Roy Moore in the general election. Both lost. That’s the same bad feedback loop as changing course on foreign policy pronouncements.
More recently, Trump appears to have begun to judiciously hand out his endorsements and aid to electoral candidates, focusing on backing candidates who are likely winners, but who might be seen as having really benefited from his endorsement. Now that’s a good presidential power strategy, and it’s one place where I think Trump has grown a bit into he presidency in terms of guarding power. But he’s got an awful long way to go.
Does the size of a Democratic House majority matter?
Not as much as you might think.
The current “classic” forecast at FiveThirtyEight has the Democrats with an 85.9% chance of winning control of the House of Representatives next month, with an average gain of 40 seats, giving them 235 seats in the House. As with any probabilistic election model, there is significant variance. It suggests a 20% chance of Democrats gaining fewer than 25 seats, but also a 20% chance of them gaining more than 54. If we assume the FiveThirtyEight model is correct and the election results fall between those two poles—which they will, well, 60% of the time—does it make a difference if the Democrats gain 25 or 54 seats?
Yes and no. Let’s start with the case for yes. Picking up a greater-than-expected number of seats would send a market signal that voters were more annoyed with the president/GOP/115th Congress than previously thought. This would likely induce some subset of moderate Republican politicians to become more vocally critical of the president, more likely to make a clean break from the Trump administration, and more likely to endorse, support, or become primary challengers to the president in 2020. Likewise, a surprisingly large number might embolden Democrats to become more combative on both policy and politics. In Bayesian terms, a larger Democratic majority would indicate that the true state of the world was worse for GOP policies and Trumpist governance than we thought.
Internally to the House Democratic caucus, a large influx of new members might shift the balance of power. The existing party leadership has been in place for over a decade, and a sizeable number of new members might be the trigger for entrepreneurial would-be leaders to seek to mount a serious challenge for the leadership. Regardless of the leadership dynamic, a larger House majority would also provide an easy cushion for chamber votes, allowing the leadership room to put their priorities on the floor and still give passes to conservative members who needed to vote against them. There would be very little reason/need to bargain with the Republican minority.
But beyond the market signal we get about the voters and some power dynamics internally in the House, the bottom line is that the size of the Democratic House majority in the 116th Congress isn’t really important, and pales relative to winning control of the chamber. On at least three important dimensions, it basically doesn’t matter if the Democrats have 220 seats (a 25 seats gain) or 249 seats ( a 54 seat gain). The most obvious reason, of course, is that control of the House by any margin denies the Republicans unified government and the ability to move the GOP agenda through the House. The House floor operates in a context of strong procedural partisanship and pure majoritarian rules; control of the agenda resides almost exclusively with the majority party, and each marginal vote beyond 218 does little to augment that. It’s like ice turning to water. At 32 degrees, it just happens, and it doesn’t get less solid if you make it 38 degrees or 75 degrees.
Likewise, the committee system in the House operates in a majoritarian partisan manner. The majority, regardless of size, gets the chairmanships of all of the committees, and almost all of the power in the committee system resides with the chairmen, backed by the majority part having more seats on each committee. They set the committee agenda, they control the committee budgets, they call the hearings and mark-ups, and they (mostly) issue the subpoenas. If the Democrats take control of the House, there will be a large number of investigations into the practices of the Trump administration; if they do not take the House, there will not be. The size of their majority will not influence this.
On the policy front, it may seem at first blush that a larger Democratic House majority would move overall public policy leftward in the 116th Congress. As a marginal effect, this could be true, in theory. With a razor-thin margin in the House (say 220-215), the GOP might be able to pick off some moderate Democrats to their policy positions, but that only applies if the GOP could get their priorities onto the floor for a vote. A thin Democratic coalition will use negative agenda control—keeping bills off the floor that would have the support of a majority of the chamber but don’t have the support of a majority of the Democratic caucus—to pevent those votes from ever happening.
This confuses people because they are used to the Senate model, where the filibuster creates the odd situation in which the majority party has agenda control but is often still looking to pick off members of the minority. That’s a totally different situation than the House, where the agenda setter needs only to pick up votes from their own centrist wing in order to achieve passage.
Finally, it’s also the case that a larger Democratic majority isn’t a more liberal Democratic majority. We don’t live in a proportional representation system with a party list. Winning more seats doesn’t put more median Democrats in the House. Winning more seats in 2018 means Dems will have won in redder and redder districts, places where Trump may have won in 2016 by a fair margin. The winners in the these last 20 marginal seats (pickups 35-54 in our thought experiment) will obviously be more liberal than their GOP alternatives, but they will hardly be foot-soldiers for the left. Some of them will be pro-life, many of them will support gun rights, and almost none of them wil be interested in impeaching the president. In a caucus meeting taking a majority vote for leadership positions, having 25 additional seats will effectively move the median vote signficantly to the right.
The Electoral College ignores suffrage.
A National Popular Vote Could Not.
There’s been some general interest in the Electoral College on Twitter this week, and for a while now there has been a focus on its mild distortive effects relative to a straight popular-vote election. The cases for and against the electoral college are varied and interesting, but until recently they have not been particularly relevant, because the popular vote winner has almost always matched the electoral college vote winner (this ignores the way that down-ballot races in blowout presidential states might be affected by turnout, but set that aside). This, of course, has not been true in 2 of the last 5 presidential elections, as the Democrats won the popular vote in both 2000 and 2016. And thus Concern.
I don’t have particularly strong feelings about the electoral college or the popular vote—there are legitimate competing values—but I do think the proponents of a national popular vote need to take seriously a huge issue that is buried inside the electoral college system: decentralized suffrage. There are minimum suffrage standards in the constitution—bars on racial, gender, or age (over 18) discrimination—but no ceilings. States are free to expand suffrage well beyond where it currently resides in most states.
This was undoubtedly in the back of the Framers’ minds when they adopted the constitutional provisions for the electoral college. The Founders were notably not big fans of democracy; much of the machinery of the orginal constitution was purposefully designed to place different parts of the government different distances from the people (modeling things after the 18th century British mixed system of monarchy, aristocracy, and democracy). Thus you find in the original constitution no federal voting right: you are eligible to vote for the House if you are eligible to vote for your state legislature. The Senate and electoral college are even further from the people; the original constitution directs them to be chosen by state legislature, and in a manner the state legislature directs, respectively.
This made some logical sense in the 1780s, as suffrage laws varied considerably by state. The electoral college, however, solved a particular problem in this regard. The presidency / vice presidency are the only interstate election in the constitution. That is, there is no other federal election where the voters in one state can affect a result that matters in another state. The electoral college thus solves a basic problem the Founders would have had to grapple with under a popular vote system: how do you construct a national election where you want each state to have influence proportional to its population, without disturbing their suffrage laws?
Consider the problem from the opposite point of view. If the Framers had chosen a national popular vote for president, states would have had strong incentives to loosen their suffrage laws in order to gain influence in the election. The iron logic of a national popular vote would have invariably led to universal white male suffrage, and perhaps even women voting. The only other option would have been uniform federal voting qualifications, but given the disparity in state laws and preference for state control of elections, that was a likely non-starter.
While universal adult suffrage seems like an obvious good idea to us today, in the 18th century it was a dangerous symptoms of runaway democracy, and something the Founders sought, quite purposely, to avoid. The electoral college, by decoupling state populations and voting populations, avoids a state suffrage race to the bottom. Each state gets the same number of electoral college votes, roughly proportional to their populations, regardless of how loose or tight their suffrage laws are.
This has ramifications for the contemporary debate over a national popular vote. The constitution remains with just floors, not ceilings, on voting rights. And indeed, throughout American history, states have had quite different voter qualification profiles. Free blacks could vote in some northern states but not others prior to the civil war. Women could vote in many western states in the later 19th century. Georgia and Kentucky lowered their voting age to 18 decades before the constitutional amendment. States have often let non-citizens vote. Though today the rules are more uniform across states than in the past, we still have widely different rules about disenfranchised felons.
A shift to a national popular vote for president would have to grapple with the small amount of diversity of suffrage standards across states, but also the possibility of states loosening their suffrage laws to gain influence in the election. Is it out of the realm of possibility that California could abolish the voting age (for presidential elections only) in order to gain an estimated 8 million voters (who, if they turned out at 50% and broke 60-40 for the Dems would provide 800K additional marginal votes)? Or (try to) allow some of their millions of non-citizens to vote?
One retort to this is that we could simply put in federal voting standards along with any constitutional amendment abolishing the electoral college, requiring voters for the presidency to be at least 18 and a citizen of the United States. That would certainly solve much of the issue (though disparate treatment of disenfranchised felons and the mentally-ill would still loom as a secondary issue). The problem, however, is that the likely mechanism for a national popular vote isn’t a constitutional amendment, but the interstate compact, in which states comprising 270 electoral college votes all agree to give their electors to the popular vote winner, essentially creating a national popular vote system within the electoral college framework. Such a compact between states has not ability to regulate the franchise, and could set off a race to the bottom in states looking to pad partisan totals.
Could Congress step in and regulate voting qualifications? Plausibly. The issue hit the Supreme Court in 1970, when Congress tried to uniformly lower the voting age to 18 via federal law. In a 5-4 decision, the Court ruled Congress could not regulate the voting age for state elections, but could regulate it for federal elections (thus the shift in 1971 to a constitutional amendment strategy). There were, however, four votes on the Court that did not believe Congress could regulate the voting age for any election, federal or state.
On a happier note, a national popular vote might encourage states to adopt non-qualification policies that encourage voting. One downside of the electoral college is that it doesn’t punish a state that seeks to generally suppress the vote. A national popular vote might reverse that trend, as states seeking to maximize their contribution to a national election (for state or partisan reasons) might ultimately adopt automatic registration, or develop easier absentee ballot systems, or just generally put resources into more polling locations and easier access to the vote.
Turnover in the House is greater than you think.
It’s true: 95%+ of Members get reelected. But it’s very misleading.
You have probably seen the statistic, perhaps in meme form: Members of Congress got reelected at a 96% rate in 2014. The reelection rate in other years is quite similar. This gives many people the distorted view that Congress is a bunch of old dudes, who have been there forever, sitting around doing nothing. Members of Congress are certainly old, and perhaps they are sitting around doing nothing, but I can assure you they have not been there forever.
The statistic is very misleading, because it only counts members who seek to be reelected. This causes a large distortion, because many people choose to leave Congress every year—to retire, to take a job in the administration, to seek other elected offices—and one set of people who often retire are members who know they are about to get crushed in their reelection. No one wants to spend the time and money it takes to run, just to be embarrassed on election day. So retirements are endogenous to the likelihood of losing. And thus statistics about how often members lose are inflated.
What we are really interested in is congressional turnover. How many new members of the House of Representatives will we have next year? There’s a simple formula: new districts created via apportionment + incumbent retirements + incumbents defeated = new freshmen. In non-redistricting years, the first term is zero. I’ve blogged extensively about this before and wrote about it regularly at CRS (click the link for all sorts of fancy charts), but I think it’s worth putting the 2018 election into context.
So far this year, 69 members of the House have voluntarily declined to return next year. Four more have lost primary elections. So there will be 73 freshmen in the House next year, plus how ever many incumbents are defeated on election day. I would guess the total number of freshmen will be well north of 90, meaning the turnover rate will be at least 21%. That’s a far cry from the 4% implied in the meme.
But is this an unusual year, with the alleged Democratic wave coming? A little bit, but not particularly so. The average number of new freshmen in the last 20 Congresses has been 61, with a high of 109 in 1992 and a low of 30 in 1988. The last three Congresses have featured 75, 59, and 52 new freshmen. A full half of the current Congress wasn’t here in 2008. That means, among other things, half of the House can’t even remember a semi-functional appropriations process. That has big implications when we talk about getting back to “regular order.”
Two Things I Enjoyed Reading This Week
Honest Graft: Why the 2018 Election Won't "All Come Down to Turnout"
Nationalized elections and the 2018 U.S. Senate midterms — LegBranch.com
Two Things I Enjoyed Listening to This Week
The Wave Podcast - The Pelosi Question | Free Listening on Podbean App
s
What Nate Silver's learned about forecasting elections
Thanks!
Thanks for reading. I’ll be back again next week! In the meantime, follow me online or reach out via email. I’d love to hear from you!